diff options
| author | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-04-16 15:20:36 -0700 |
|---|---|---|
| committer | Linus Torvalds <torvalds@ppc970.osdl.org> | 2005-04-16 15:20:36 -0700 |
| commit | 1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch) | |
| tree | 0bba044c4ce775e45a88a51686b5d9f90697ea9d /arch/i386/kernel | |
Linux-2.6.12-rc2v2.6.12-rc2
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Diffstat (limited to 'arch/i386/kernel')
124 files changed, 43777 insertions, 0 deletions
diff --git a/arch/i386/kernel/Makefile b/arch/i386/kernel/Makefile new file mode 100644 index 00000000000..933787a46b4 --- /dev/null +++ b/arch/i386/kernel/Makefile @@ -0,0 +1,71 @@ +# +# Makefile for the linux kernel. +# + +extra-y := head.o init_task.o vmlinux.lds + +obj-y := process.o semaphore.o signal.o entry.o traps.o irq.o vm86.o \ + ptrace.o time.o ioport.o ldt.o setup.o i8259.o sys_i386.o \ + pci-dma.o i386_ksyms.o i387.o dmi_scan.o bootflag.o \ + doublefault.o quirks.o + +obj-y += cpu/ +obj-y += timers/ +obj-$(CONFIG_ACPI_BOOT) += acpi/ +obj-$(CONFIG_X86_BIOS_REBOOT) += reboot.o +obj-$(CONFIG_MCA) += mca.o +obj-$(CONFIG_X86_MSR) += msr.o +obj-$(CONFIG_X86_CPUID) += cpuid.o +obj-$(CONFIG_MICROCODE) += microcode.o +obj-$(CONFIG_APM) += apm.o +obj-$(CONFIG_X86_SMP) += smp.o smpboot.o +obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o +obj-$(CONFIG_X86_MPPARSE) += mpparse.o +obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o +obj-$(CONFIG_X86_IO_APIC) += io_apic.o +obj-$(CONFIG_X86_NUMAQ) += numaq.o +obj-$(CONFIG_X86_SUMMIT_NUMA) += summit.o +obj-$(CONFIG_KPROBES) += kprobes.o +obj-$(CONFIG_MODULES) += module.o +obj-y += sysenter.o vsyscall.o +obj-$(CONFIG_ACPI_SRAT) += srat.o +obj-$(CONFIG_HPET_TIMER) += time_hpet.o +obj-$(CONFIG_EFI) += efi.o efi_stub.o +obj-$(CONFIG_EARLY_PRINTK) += early_printk.o + +EXTRA_AFLAGS := -traditional + +obj-$(CONFIG_SCx200) += scx200.o + +# vsyscall.o contains the vsyscall DSO images as __initdata. +# We must build both images before we can assemble it. +# Note: kbuild does not track this dependency due to usage of .incbin +$(obj)/vsyscall.o: $(obj)/vsyscall-int80.so $(obj)/vsyscall-sysenter.so +targets += $(foreach F,int80 sysenter,vsyscall-$F.o vsyscall-$F.so) +targets += vsyscall.lds + +# The DSO images are built using a special linker script. +quiet_cmd_syscall = SYSCALL $@ + cmd_syscall = $(CC) -m elf_i386 -nostdlib $(SYSCFLAGS_$(@F)) \ + -Wl,-T,$(filter-out FORCE,$^) -o $@ + +export CPPFLAGS_vsyscall.lds += -P -C -U$(ARCH) + +vsyscall-flags = -shared -s -Wl,-soname=linux-gate.so.1 +SYSCFLAGS_vsyscall-sysenter.so = $(vsyscall-flags) +SYSCFLAGS_vsyscall-int80.so = $(vsyscall-flags) + +$(obj)/vsyscall-int80.so $(obj)/vsyscall-sysenter.so: \ +$(obj)/vsyscall-%.so: $(src)/vsyscall.lds $(obj)/vsyscall-%.o FORCE + $(call if_changed,syscall) + +# We also create a special relocatable object that should mirror the symbol +# table and layout of the linked DSO. With ld -R we can then refer to +# these symbols in the kernel code rather than hand-coded addresses. +extra-y += vsyscall-syms.o +$(obj)/built-in.o: $(obj)/vsyscall-syms.o +$(obj)/built-in.o: ld_flags += -R $(obj)/vsyscall-syms.o + +SYSCFLAGS_vsyscall-syms.o = -r +$(obj)/vsyscall-syms.o: $(src)/vsyscall.lds $(obj)/vsyscall-sysenter.o FORCE + $(call if_changed,syscall) diff --git a/arch/i386/kernel/acpi/Makefile b/arch/i386/kernel/acpi/Makefile new file mode 100644 index 00000000000..ee75cb286cf --- /dev/null +++ b/arch/i386/kernel/acpi/Makefile @@ -0,0 +1,4 @@ +obj-$(CONFIG_ACPI_BOOT) := boot.o +obj-$(CONFIG_X86_IO_APIC) += earlyquirk.o +obj-$(CONFIG_ACPI_SLEEP) += sleep.o wakeup.o + diff --git a/arch/i386/kernel/acpi/boot.c b/arch/i386/kernel/acpi/boot.c new file mode 100644 index 00000000000..9ba0b957d11 --- /dev/null +++ b/arch/i386/kernel/acpi/boot.c @@ -0,0 +1,908 @@ +/* + * boot.c - Architecture-Specific Low-Level ACPI Boot Support + * + * Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com> + * Copyright (C) 2001 Jun Nakajima <jun.nakajima@intel.com> + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + */ + +#include <linux/init.h> +#include <linux/config.h> +#include <linux/acpi.h> +#include <linux/efi.h> +#include <linux/irq.h> +#include <linux/module.h> + +#include <asm/pgtable.h> +#include <asm/io_apic.h> +#include <asm/apic.h> +#include <asm/io.h> +#include <asm/irq.h> +#include <asm/mpspec.h> + +#ifdef CONFIG_X86_64 + +static inline void acpi_madt_oem_check(char *oem_id, char *oem_table_id) { } +extern void __init clustered_apic_check(void); +static inline int ioapic_setup_disabled(void) { return 0; } +#include <asm/proto.h> + +#else /* X86 */ + +#ifdef CONFIG_X86_LOCAL_APIC +#include <mach_apic.h> +#include <mach_mpparse.h> +#endif /* CONFIG_X86_LOCAL_APIC */ + +#endif /* X86 */ + +#define BAD_MADT_ENTRY(entry, end) ( \ + (!entry) || (unsigned long)entry + sizeof(*entry) > end || \ + ((acpi_table_entry_header *)entry)->length != sizeof(*entry)) + +#define PREFIX "ACPI: " + +#ifdef CONFIG_ACPI_PCI +int acpi_noirq __initdata; /* skip ACPI IRQ initialization */ +int acpi_pci_disabled __initdata; /* skip ACPI PCI scan and IRQ initialization */ +#else +int acpi_noirq __initdata = 1; +int acpi_pci_disabled __initdata = 1; +#endif +int acpi_ht __initdata = 1; /* enable HT */ + +int acpi_lapic; +int acpi_ioapic; +int acpi_strict; +EXPORT_SYMBOL(acpi_strict); + +acpi_interrupt_flags acpi_sci_flags __initdata; +int acpi_sci_override_gsi __initdata; +int acpi_skip_timer_override __initdata; + +#ifdef CONFIG_X86_LOCAL_APIC +static u64 acpi_lapic_addr __initdata = APIC_DEFAULT_PHYS_BASE; +#endif + +#ifndef __HAVE_ARCH_CMPXCHG +#warning ACPI uses CMPXCHG, i486 and later hardware +#endif + +#define MAX_MADT_ENTRIES 256 +u8 x86_acpiid_to_apicid[MAX_MADT_ENTRIES] = + { [0 ... MAX_MADT_ENTRIES-1] = 0xff }; +EXPORT_SYMBOL(x86_acpiid_to_apicid); + +/* -------------------------------------------------------------------------- + Boot-time Configuration + -------------------------------------------------------------------------- */ + +/* + * The default interrupt routing model is PIC (8259). This gets + * overriden if IOAPICs are enumerated (below). + */ +enum acpi_irq_model_id acpi_irq_model = ACPI_IRQ_MODEL_PIC; + +#ifdef CONFIG_X86_64 + +/* rely on all ACPI tables being in the direct mapping */ +char *__acpi_map_table(unsigned long phys_addr, unsigned long size) +{ + if (!phys_addr || !size) + return NULL; + + if (phys_addr < (end_pfn_map << PAGE_SHIFT)) + return __va(phys_addr); + + return NULL; +} + +#else + +/* + * Temporarily use the virtual area starting from FIX_IO_APIC_BASE_END, + * to map the target physical address. The problem is that set_fixmap() + * provides a single page, and it is possible that the page is not + * sufficient. + * By using this area, we can map up to MAX_IO_APICS pages temporarily, + * i.e. until the next __va_range() call. + * + * Important Safety Note: The fixed I/O APIC page numbers are *subtracted* + * from the fixed base. That's why we start at FIX_IO_APIC_BASE_END and + * count idx down while incrementing the phys address. + */ +char *__acpi_map_table(unsigned long phys, unsigned long size) +{ + unsigned long base, offset, mapped_size; + int idx; + + if (phys + size < 8*1024*1024) + return __va(phys); + + offset = phys & (PAGE_SIZE - 1); + mapped_size = PAGE_SIZE - offset; + set_fixmap(FIX_ACPI_END, phys); + base = fix_to_virt(FIX_ACPI_END); + + /* + * Most cases can be covered by the below. + */ + idx = FIX_ACPI_END; + while (mapped_size < size) { + if (--idx < FIX_ACPI_BEGIN) + return NULL; /* cannot handle this */ + phys += PAGE_SIZE; + set_fixmap(idx, phys); + mapped_size += PAGE_SIZE; + } + + return ((unsigned char *) base + offset); +} +#endif + +#ifdef CONFIG_PCI_MMCONFIG +static int __init acpi_parse_mcfg(unsigned long phys_addr, unsigned long size) +{ + struct acpi_table_mcfg *mcfg; + + if (!phys_addr || !size) + return -EINVAL; + + mcfg = (struct acpi_table_mcfg *) __acpi_map_table(phys_addr, size); + if (!mcfg) { + printk(KERN_WARNING PREFIX "Unable to map MCFG\n"); + return -ENODEV; + } + + if (mcfg->base_reserved) { + printk(KERN_ERR PREFIX "MMCONFIG not in low 4GB of memory\n"); + return -ENODEV; + } + + pci_mmcfg_base_addr = mcfg->base_address; + + return 0; +} +#else +#define acpi_parse_mcfg NULL +#endif /* !CONFIG_PCI_MMCONFIG */ + +#ifdef CONFIG_X86_LOCAL_APIC +static int __init +acpi_parse_madt ( + unsigned long phys_addr, + unsigned long size) +{ + struct acpi_table_madt *madt = NULL; + + if (!phys_addr || !size) + return -EINVAL; + + madt = (struct acpi_table_madt *) __acpi_map_table(phys_addr, size); + if (!madt) { + printk(KERN_WARNING PREFIX "Unable to map MADT\n"); + return -ENODEV; + } + + if (madt->lapic_address) { + acpi_lapic_addr = (u64) madt->lapic_address; + + printk(KERN_DEBUG PREFIX "Local APIC address 0x%08x\n", + madt->lapic_address); + } + + acpi_madt_oem_check(madt->header.oem_id, madt->header.oem_table_id); + + return 0; +} + + +static int __init +acpi_parse_lapic ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_lapic *processor = NULL; + + processor = (struct acpi_table_lapic*) header; + + if (BAD_MADT_ENTRY(processor, end)) + return -EINVAL; + + acpi_table_print_madt_entry(header); + + /* no utility in registering a disabled processor */ + if (processor->flags.enabled == 0) + return 0; + + x86_acpiid_to_apicid[processor->acpi_id] = processor->id; + + mp_register_lapic ( + processor->id, /* APIC ID */ + processor->flags.enabled); /* Enabled? */ + + return 0; +} + +static int __init +acpi_parse_lapic_addr_ovr ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_lapic_addr_ovr *lapic_addr_ovr = NULL; + + lapic_addr_ovr = (struct acpi_table_lapic_addr_ovr*) header; + + if (BAD_MADT_ENTRY(lapic_addr_ovr, end)) + return -EINVAL; + + acpi_lapic_addr = lapic_addr_ovr->address; + + return 0; +} + +static int __init +acpi_parse_lapic_nmi ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_lapic_nmi *lapic_nmi = NULL; + + lapic_nmi = (struct acpi_table_lapic_nmi*) header; + + if (BAD_MADT_ENTRY(lapic_nmi, end)) + return -EINVAL; + + acpi_table_print_madt_entry(header); + + if (lapic_nmi->lint != 1) + printk(KERN_WARNING PREFIX "NMI not connected to LINT 1!\n"); + + return 0; +} + + +#endif /*CONFIG_X86_LOCAL_APIC*/ + +#if defined(CONFIG_X86_IO_APIC) && defined(CONFIG_ACPI_INTERPRETER) + +static int __init +acpi_parse_ioapic ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_ioapic *ioapic = NULL; + + ioapic = (struct acpi_table_ioapic*) header; + + if (BAD_MADT_ENTRY(ioapic, end)) + return -EINVAL; + + acpi_table_print_madt_entry(header); + + mp_register_ioapic ( + ioapic->id, + ioapic->address, + ioapic->global_irq_base); + + return 0; +} + +/* + * Parse Interrupt Source Override for the ACPI SCI + */ +static void +acpi_sci_ioapic_setup(u32 gsi, u16 polarity, u16 trigger) +{ + if (trigger == 0) /* compatible SCI trigger is level */ + trigger = 3; + + if (polarity == 0) /* compatible SCI polarity is low */ + polarity = 3; + + /* Command-line over-ride via acpi_sci= */ + if (acpi_sci_flags.trigger) + trigger = acpi_sci_flags.trigger; + + if (acpi_sci_flags.polarity) + polarity = acpi_sci_flags.polarity; + + /* + * mp_config_acpi_legacy_irqs() already setup IRQs < 16 + * If GSI is < 16, this will update its flags, + * else it will create a new mp_irqs[] entry. + */ + mp_override_legacy_irq(gsi, polarity, trigger, gsi); + + /* + * stash over-ride to indicate we've been here + * and for later update of acpi_fadt + */ + acpi_sci_override_gsi = gsi; + return; +} + +static int __init +acpi_parse_int_src_ovr ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_int_src_ovr *intsrc = NULL; + + intsrc = (struct acpi_table_int_src_ovr*) header; + + if (BAD_MADT_ENTRY(intsrc, end)) + return -EINVAL; + + acpi_table_print_madt_entry(header); + + if (intsrc->bus_irq == acpi_fadt.sci_int) { + acpi_sci_ioapic_setup(intsrc->global_irq, + intsrc->flags.polarity, intsrc->flags.trigger); + return 0; + } + + if (acpi_skip_timer_override && + intsrc->bus_irq == 0 && intsrc->global_irq == 2) { + printk(PREFIX "BIOS IRQ0 pin2 override ignored.\n"); + return 0; + } + + mp_override_legacy_irq ( + intsrc->bus_irq, + intsrc->flags.polarity, + intsrc->flags.trigger, + intsrc->global_irq); + + return 0; +} + + +static int __init +acpi_parse_nmi_src ( + acpi_table_entry_header *header, const unsigned long end) +{ + struct acpi_table_nmi_src *nmi_src = NULL; + + nmi_src = (struct acpi_table_nmi_src*) header; + + if (BAD_MADT_ENTRY(nmi_src, end)) + return -EINVAL; + + acpi_table_print_madt_entry(header); + + /* TBD: Support nimsrc entries? */ + + return 0; +} + +#endif /* CONFIG_X86_IO_APIC */ + +#ifdef CONFIG_ACPI_BUS + +/* + * acpi_pic_sci_set_trigger() + * + * use ELCR to set PIC-mode trigger type for SCI + * + * If a PIC-mode SCI is not recognized or gives spurious IRQ7's + * it may require Edge Trigger -- use "acpi_sci=edge" + * + * Port 0x4d0-4d1 are ECLR1 and ECLR2, the Edge/Level Control Registers + * for the 8259 PIC. bit[n] = 1 means irq[n] is Level, otherwise Edge. + * ECLR1 is IRQ's 0-7 (IRQ 0, 1, 2 must be 0) + * ECLR2 is IRQ's 8-15 (IRQ 8, 13 must be 0) + */ + +void __init +acpi_pic_sci_set_trigger(unsigned int irq, u16 trigger) +{ + unsigned int mask = 1 << irq; + unsigned int old, new; + + /* Real old ELCR mask */ + old = inb(0x4d0) | (inb(0x4d1) << 8); + + /* + * If we use ACPI to set PCI irq's, then we should clear ELCR + * since we will set it correctly as we enable the PCI irq + * routing. + */ + new = acpi_noirq ? old : 0; + + /* + * Update SCI information in the ELCR, it isn't in the PCI + * routing tables.. + */ + switch (trigger) { + case 1: /* Edge - clear */ + new &= ~mask; + break; + case 3: /* Level - set */ + new |= mask; + break; + } + + if (old == new) + return; + + printk(PREFIX "setting ELCR to %04x (from %04x)\n", new, old); + outb(new, 0x4d0); + outb(new >> 8, 0x4d1); +} + + +#endif /* CONFIG_ACPI_BUS */ + +int acpi_gsi_to_irq(u32 gsi, unsigned int *irq) +{ +#ifdef CONFIG_X86_IO_APIC + if (use_pci_vector() && !platform_legacy_irq(gsi)) + *irq = IO_APIC_VECTOR(gsi); + else +#endif + *irq = gsi; + return 0; +} + +unsigned int acpi_register_gsi(u32 gsi, int edge_level, int active_high_low) +{ + unsigned int irq; + unsigned int plat_gsi = gsi; + +#ifdef CONFIG_PCI + /* + * Make sure all (legacy) PCI IRQs are set as level-triggered. + */ + if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) { + extern void eisa_set_level_irq(unsigned int irq); + + if (edge_level == ACPI_LEVEL_SENSITIVE) + eisa_set_level_irq(gsi); + } +#endif + +#ifdef CONFIG_X86_IO_APIC + if (acpi_irq_model == ACPI_IRQ_MODEL_IOAPIC) { + plat_gsi = mp_register_gsi(gsi, edge_level, active_high_low); + } +#endif + acpi_gsi_to_irq(plat_gsi, &irq); + return irq; +} +EXPORT_SYMBOL(acpi_register_gsi); + +/* + * ACPI based hotplug support for CPU + */ +#ifdef CONFIG_ACPI_HOTPLUG_CPU +int +acpi_map_lsapic(acpi_handle handle, int *pcpu) +{ + /* TBD */ + return -EINVAL; +} +EXPORT_SYMBOL(acpi_map_lsapic); + + +int +acpi_unmap_lsapic(int cpu) +{ + /* TBD */ + return -EINVAL; +} +EXPORT_SYMBOL(acpi_unmap_lsapic); +#endif /* CONFIG_ACPI_HOTPLUG_CPU */ + +static unsigned long __init +acpi_scan_rsdp ( + unsigned long start, + unsigned long length) +{ + unsigned long offset = 0; + unsigned long sig_len = sizeof("RSD PTR ") - 1; + + /* + * Scan all 16-byte boundaries of the physical memory region for the + * RSDP signature. + */ + for (offset = 0; offset < length; offset += 16) { + if (strncmp((char *) (start + offset), "RSD PTR ", sig_len)) + continue; + return (start + offset); + } + + return 0; +} + +static int __init acpi_parse_sbf(unsigned long phys_addr, unsigned long size) +{ + struct acpi_table_sbf *sb; + + if (!phys_addr || !size) + return -EINVAL; + + sb = (struct acpi_table_sbf *) __acpi_map_table(phys_addr, size); + if (!sb) { + printk(KERN_WARNING PREFIX "Unable to map SBF\n"); + return -ENODEV; + } + + sbf_port = sb->sbf_cmos; /* Save CMOS port */ + + return 0; +} + + +#ifdef CONFIG_HPET_TIMER + +static int __init acpi_parse_hpet(unsigned long phys, unsigned long size) +{ + struct acpi_table_hpet *hpet_tbl; + + if (!phys || !size) + return -EINVAL; + + hpet_tbl = (struct acpi_table_hpet *) __acpi_map_table(phys, size); + if (!hpet_tbl) { + printk(KERN_WARNING PREFIX "Unable to map HPET\n"); + return -ENODEV; + } + + if (hpet_tbl->addr.space_id != ACPI_SPACE_MEM) { + printk(KERN_WARNING PREFIX "HPET timers must be located in " + "memory.\n"); + return -1; + } + +#ifdef CONFIG_X86_64 + vxtime.hpet_address = hpet_tbl->addr.addrl | + ((long) hpet_tbl->addr.addrh << 32); + + printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n", + hpet_tbl->id, vxtime.hpet_address); +#else /* X86 */ + { + extern unsigned long hpet_address; + + hpet_address = hpet_tbl->addr.addrl; + printk(KERN_INFO PREFIX "HPET id: %#x base: %#lx\n", + hpet_tbl->id, hpet_address); + } +#endif /* X86 */ + + return 0; +} +#else +#define acpi_parse_hpet NULL +#endif + +#ifdef CONFIG_X86_PM_TIMER +extern u32 pmtmr_ioport; +#endif + +static int __init acpi_parse_fadt(unsigned long phys, unsigned long size) +{ + struct fadt_descriptor_rev2 *fadt = NULL; + + fadt = (struct fadt_descriptor_rev2*) __acpi_map_table(phys,size); + if(!fadt) { + printk(KERN_WARNING PREFIX "Unable to map FADT\n"); + return 0; + } + +#ifdef CONFIG_ACPI_INTERPRETER + /* initialize sci_int early for INT_SRC_OVR MADT parsing */ + acpi_fadt.sci_int = fadt->sci_int; +#endif + +#ifdef CONFIG_X86_PM_TIMER + /* detect the location of the ACPI PM Timer */ + if (fadt->revision >= FADT2_REVISION_ID) { + /* FADT rev. 2 */ + if (fadt->xpm_tmr_blk.address_space_id != ACPI_ADR_SPACE_SYSTEM_IO) + return 0; + + pmtmr_ioport = fadt->xpm_tmr_blk.address; + } else { + /* FADT rev. 1 */ + pmtmr_ioport = fadt->V1_pm_tmr_blk; + } + if (pmtmr_ioport) + printk(KERN_INFO PREFIX "PM-Timer IO Port: %#x\n", pmtmr_ioport); +#endif + return 0; +} + + +unsigned long __init +acpi_find_rsdp (void) +{ + unsigned long rsdp_phys = 0; + + if (efi_enabled) { + if (efi.acpi20) + return __pa(efi.acpi20); + else if (efi.acpi) + return __pa(efi.acpi); + } + /* + * Scan memory looking for the RSDP signature. First search EBDA (low + * memory) paragraphs and then search upper memory (E0000-FFFFF). + */ + rsdp_phys = acpi_scan_rsdp (0, 0x400); + if (!rsdp_phys) + rsdp_phys = acpi_scan_rsdp (0xE0000, 0xFFFFF); + + return rsdp_phys; +} + +#ifdef CONFIG_X86_LOCAL_APIC +/* + * Parse LAPIC entries in MADT + * returns 0 on success, < 0 on error + */ +static int __init +acpi_parse_madt_lapic_entries(void) +{ + int count; + + /* + * Note that the LAPIC address is obtained from the MADT (32-bit value) + * and (optionally) overriden by a LAPIC_ADDR_OVR entry (64-bit value). + */ + + count = acpi_table_parse_madt(ACPI_MADT_LAPIC_ADDR_OVR, acpi_parse_lapic_addr_ovr, 0); + if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing LAPIC address override entry\n"); + return count; + } + + mp_register_lapic_address(acpi_lapic_addr); + + count = acpi_table_parse_madt(ACPI_MADT_LAPIC, acpi_parse_lapic, + MAX_APICS); + if (!count) { + printk(KERN_ERR PREFIX "No LAPIC entries present\n"); + /* TBD: Cleanup to allow fallback to MPS */ + return -ENODEV; + } + else if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing LAPIC entry\n"); + /* TBD: Cleanup to allow fallback to MPS */ + return count; + } + + count = acpi_table_parse_madt(ACPI_MADT_LAPIC_NMI, acpi_parse_lapic_nmi, 0); + if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing LAPIC NMI entry\n"); + /* TBD: Cleanup to allow fallback to MPS */ + return count; + } + return 0; +} +#endif /* CONFIG_X86_LOCAL_APIC */ + +#if defined(CONFIG_X86_IO_APIC) && defined(CONFIG_ACPI_INTERPRETER) +/* + * Parse IOAPIC related entries in MADT + * returns 0 on success, < 0 on error + */ +static int __init +acpi_parse_madt_ioapic_entries(void) +{ + int count; + + /* + * ACPI interpreter is required to complete interrupt setup, + * so if it is off, don't enumerate the io-apics with ACPI. + * If MPS is present, it will handle them, + * otherwise the system will stay in PIC mode + */ + if (acpi_disabled || acpi_noirq) { + return -ENODEV; + } + + /* + * if "noapic" boot option, don't look for IO-APICs + */ + if (skip_ioapic_setup) { + printk(KERN_INFO PREFIX "Skipping IOAPIC probe " + "due to 'noapic' option.\n"); + return -ENODEV; + } + + count = acpi_table_parse_madt(ACPI_MADT_IOAPIC, acpi_parse_ioapic, MAX_IO_APICS); + if (!count) { + printk(KERN_ERR PREFIX "No IOAPIC entries present\n"); + return -ENODEV; + } + else if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing IOAPIC entry\n"); + return count; + } + + count = acpi_table_parse_madt(ACPI_MADT_INT_SRC_OVR, acpi_parse_int_src_ovr, NR_IRQ_VECTORS); + if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing interrupt source overrides entry\n"); + /* TBD: Cleanup to allow fallback to MPS */ + return count; + } + + /* + * If BIOS did not supply an INT_SRC_OVR for the SCI + * pretend we got one so we can set the SCI flags. + */ + if (!acpi_sci_override_gsi) + acpi_sci_ioapic_setup(acpi_fadt.sci_int, 0, 0); + + /* Fill in identity legacy mapings where no override */ + mp_config_acpi_legacy_irqs(); + + count = acpi_table_parse_madt(ACPI_MADT_NMI_SRC, acpi_parse_nmi_src, NR_IRQ_VECTORS); + if (count < 0) { + printk(KERN_ERR PREFIX "Error parsing NMI SRC entry\n"); + /* TBD: Cleanup to allow fallback to MPS */ + return count; + } + + return 0; +} +#else +static inline int acpi_parse_madt_ioapic_entries(void) +{ + return -1; +} +#endif /* !(CONFIG_X86_IO_APIC && CONFIG_ACPI_INTERPRETER) */ + + +static void __init +acpi_process_madt(void) +{ +#ifdef CONFIG_X86_LOCAL_APIC + int count, error; + + count = acpi_table_parse(ACPI_APIC, acpi_parse_madt); + if (count >= 1) { + + /* + * Parse MADT LAPIC entries + */ + error = acpi_parse_madt_lapic_entries(); + if (!error) { + acpi_lapic = 1; + + /* + * Parse MADT IO-APIC entries + */ + error = acpi_parse_madt_ioapic_entries(); + if (!error) { + acpi_irq_model = ACPI_IRQ_MODEL_IOAPIC; + acpi_irq_balance_set(NULL); + acpi_ioapic = 1; + + smp_found_config = 1; + clustered_apic_check(); + } + } + if (error == -EINVAL) { + /* + * Dell Precision Workstation 410, 610 come here. + */ + printk(KERN_ERR PREFIX "Invalid BIOS MADT, disabling ACPI\n"); + disable_acpi(); + } + } +#endif + return; +} + +/* + * acpi_boot_table_init() and acpi_boot_init() + * called from setup_arch(), always. + * 1. checksums all tables + * 2. enumerates lapics + * 3. enumerates io-apics + * + * acpi_table_init() is separate to allow reading SRAT without + * other side effects. + * + * side effects of acpi_boot_init: + * acpi_lapic = 1 if LAPIC found + * acpi_ioapic = 1 if IOAPIC found + * if (acpi_lapic && acpi_ioapic) smp_found_config = 1; + * if acpi_blacklisted() acpi_disabled = 1; + * acpi_irq_model=... + * ... + * + * return value: (currently ignored) + * 0: success + * !0: failure + */ + +int __init +acpi_boot_table_init(void) +{ + int error; + + /* + * If acpi_disabled, bail out + * One exception: acpi=ht continues far enough to enumerate LAPICs + */ + if (acpi_disabled && !acpi_ht) + return 1; + + /* + * Initialize the ACPI boot-time table parser. + */ + error = acpi_table_init(); + if (error) { + disable_acpi(); + return error; + } + +#ifdef __i386__ + check_acpi_pci(); +#endif + + acpi_table_parse(ACPI_BOOT, acpi_parse_sbf); + + /* + * blacklist may disable ACPI entirely + */ + error = acpi_blacklisted(); + if (error) { + extern int acpi_force; + + if (acpi_force) { + printk(KERN_WARNING PREFIX "acpi=force override\n"); + } else { + printk(KERN_WARNING PREFIX "Disabling ACPI support\n"); + disable_acpi(); + return error; + } + } + + return 0; +} + + +int __init acpi_boot_init(void) +{ + /* + * If acpi_disabled, bail out + * One exception: acpi=ht continues far enough to enumerate LAPICs + */ + if (acpi_disabled && !acpi_ht) + return 1; + + acpi_table_parse(ACPI_BOOT, acpi_parse_sbf); + + /* + * set sci_int and PM timer address + */ + acpi_table_parse(ACPI_FADT, acpi_parse_fadt); + + /* + * Process the Multiple APIC Description Table (MADT), if present + */ + acpi_process_madt(); + + acpi_table_parse(ACPI_HPET, acpi_parse_hpet); + acpi_table_parse(ACPI_MCFG, acpi_parse_mcfg); + + return 0; +} + diff --git a/arch/i386/kernel/acpi/earlyquirk.c b/arch/i386/kernel/acpi/earlyquirk.c new file mode 100644 index 00000000000..726a5ca4b16 --- /dev/null +++ b/arch/i386/kernel/acpi/earlyquirk.c @@ -0,0 +1,51 @@ +/* + * Do early PCI probing for bug detection when the main PCI subsystem is + * not up yet. + */ +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/pci.h> +#include <asm/pci-direct.h> +#include <asm/acpi.h> + +static int __init check_bridge(int vendor, int device) +{ + /* According to Nvidia all timer overrides are bogus. Just ignore + them all. */ + if (vendor == PCI_VENDOR_ID_NVIDIA) { + acpi_skip_timer_override = 1; + } + return 0; +} + +void __init check_acpi_pci(void) +{ + int num,slot,func; + + /* Assume the machine supports type 1. If not it will + always read ffffffff and should not have any side effect. */ + + /* Poor man's PCI discovery */ + for (num = 0; num < 32; num++) { + for (slot = 0; slot < 32; slot++) { + for (func = 0; func < 8; func++) { + u32 class; + u32 vendor; + class = read_pci_config(num,slot,func, + PCI_CLASS_REVISION); + if (class == 0xffffffff) + break; + + if ((class >> 16) != PCI_CLASS_BRIDGE_PCI) + continue; + + vendor = read_pci_config(num, slot, func, + PCI_VENDOR_ID); + + if (check_bridge(vendor&0xffff, vendor >> 16)) + return; + } + + } + } +} diff --git a/arch/i386/kernel/acpi/sleep.c b/arch/i386/kernel/acpi/sleep.c new file mode 100644 index 00000000000..28bb0514bb6 --- /dev/null +++ b/arch/i386/kernel/acpi/sleep.c @@ -0,0 +1,93 @@ +/* + * sleep.c - x86-specific ACPI sleep support. + * + * Copyright (C) 2001-2003 Patrick Mochel + * Copyright (C) 2001-2003 Pavel Machek <pavel@suse.cz> + */ + +#include <linux/acpi.h> +#include <linux/bootmem.h> +#include <asm/smp.h> +#include <asm/tlbflush.h> + +/* address in low memory of the wakeup routine. */ +unsigned long acpi_wakeup_address = 0; +unsigned long acpi_video_flags; +extern char wakeup_start, wakeup_end; + +extern void zap_low_mappings(void); + +extern unsigned long FASTCALL(acpi_copy_wakeup_routine(unsigned long)); + +static void init_low_mapping(pgd_t *pgd, int pgd_limit) +{ + int pgd_ofs = 0; + + while ((pgd_ofs < pgd_limit) && (pgd_ofs + USER_PTRS_PER_PGD < PTRS_PER_PGD)) { + set_pgd(pgd, *(pgd+USER_PTRS_PER_PGD)); + pgd_ofs++, pgd++; + } + flush_tlb_all(); +} + +/** + * acpi_save_state_mem - save kernel state + * + * Create an identity mapped page table and copy the wakeup routine to + * low memory. + */ +int acpi_save_state_mem (void) +{ + if (!acpi_wakeup_address) + return 1; + init_low_mapping(swapper_pg_dir, USER_PTRS_PER_PGD); + memcpy((void *) acpi_wakeup_address, &wakeup_start, &wakeup_end - &wakeup_start); + acpi_copy_wakeup_routine(acpi_wakeup_address); + + return 0; +} + +/* + * acpi_restore_state - undo effects of acpi_save_state_mem + */ +void acpi_restore_state_mem (void) +{ + zap_low_mappings(); +} + +/** + * acpi_reserve_bootmem - do _very_ early ACPI initialisation + * + * We allocate a page from the first 1MB of memory for the wakeup + * routine for when we come back from a sleep state. The + * runtime allocator allows specification of <16MB pages, but not + * <1MB pages. + */ +void __init acpi_reserve_bootmem(void) +{ + if ((&wakeup_end - &wakeup_start) > PAGE_SIZE) { + printk(KERN_ERR "ACPI: Wakeup code way too big, S3 disabled.\n"); + return; + } + + acpi_wakeup_address = (unsigned long)alloc_bootmem_low(PAGE_SIZE); + if (!acpi_wakeup_address) + printk(KERN_ERR "ACPI: Cannot allocate lowmem, S3 disabled.\n"); +} + +static int __init acpi_sleep_setup(char *str) +{ + while ((str != NULL) && (*str != '\0')) { + if (strncmp(str, "s3_bios", 7) == 0) + acpi_video_flags = 1; + if (strncmp(str, "s3_mode", 7) == 0) + acpi_video_flags |= 2; + str = strchr(str, ','); + if (str != NULL) + str += strspn(str, ", \t"); + } + return 1; +} + + +__setup("acpi_sleep=", acpi_sleep_setup); diff --git a/arch/i386/kernel/acpi/wakeup.S b/arch/i386/kernel/acpi/wakeup.S new file mode 100644 index 00000000000..39d32484f6f --- /dev/null +++ b/arch/i386/kernel/acpi/wakeup.S @@ -0,0 +1,318 @@ +.text +#include <linux/linkage.h> +#include <asm/segment.h> +#include <asm/page.h> + +# +# wakeup_code runs in real mode, and at unknown address (determined at run-time). +# Therefore it must only use relative jumps/calls. +# +# Do we need to deal with A20? It is okay: ACPI specs says A20 must be enabled +# +# If physical address of wakeup_code is 0x12345, BIOS should call us with +# cs = 0x1234, eip = 0x05 +# + +ALIGN + .align 4096 +ENTRY(wakeup_start) +wakeup_code: + wakeup_code_start = . + .code16 + + movw $0xb800, %ax + movw %ax,%fs + movw $0x0e00 + 'L', %fs:(0x10) + + cli + cld + + # setup data segment + movw %cs, %ax + movw %ax, %ds # Make ds:0 point to wakeup_start + movw %ax, %ss + mov $(wakeup_stack - wakeup_code), %sp # Private stack is needed for ASUS board + movw $0x0e00 + 'S', %fs:(0x12) + + pushl $0 # Kill any dangerous flags + popfl + + movl real_magic - wakeup_code, %eax + cmpl $0x12345678, %eax + jne bogus_real_magic + + testl $1, video_flags - wakeup_code + jz 1f + lcall $0xc000,$3 + movw %cs, %ax + movw %ax, %ds # Bios might have played with that + movw %ax, %ss +1: + + testl $2, video_flags - wakeup_code + jz 1f + mov video_mode - wakeup_code, %ax + call mode_set +1: + + # set up page table + movl $swapper_pg_dir-__PAGE_OFFSET, %eax + movl %eax, %cr3 + + testl $1, real_efer_save_restore - wakeup_code + jz 4f + # restore efer setting + movl real_save_efer_edx - wakeup_code, %edx + movl real_save_efer_eax - wakeup_code, %eax + mov $0xc0000080, %ecx + wrmsr +4: + # make sure %cr4 is set correctly (features, etc) + movl real_save_cr4 - wakeup_code, %eax + movl %eax, %cr4 + movw $0xb800, %ax + movw %ax,%fs + movw $0x0e00 + 'i', %fs:(0x12) + + # need a gdt + lgdt real_save_gdt - wakeup_code + + movl real_save_cr0 - wakeup_code, %eax + movl %eax, %cr0 + jmp 1f +1: + movw $0x0e00 + 'n', %fs:(0x14) + + movl real_magic - wakeup_code, %eax + cmpl $0x12345678, %eax + jne bogus_real_magic + + ljmpl $__KERNEL_CS,$wakeup_pmode_return + +real_save_gdt: .word 0 + .long 0 +real_save_cr0: .long 0 +real_save_cr3: .long 0 +real_save_cr4: .long 0 +real_magic: .long 0 +video_mode: .long 0 +video_flags: .long 0 +real_efer_save_restore: .long 0 +real_save_efer_edx: .long 0 +real_save_efer_eax: .long 0 + +bogus_real_magic: + movw $0x0e00 + 'B', %fs:(0x12) + jmp bogus_real_magic + +/* This code uses an extended set of video mode numbers. These include: + * Aliases for standard modes + * NORMAL_VGA (-1) + * EXTENDED_VGA (-2) + * ASK_VGA (-3) + * Video modes numbered by menu position -- NOT RECOMMENDED because of lack + * of compatibility when extending the table. These are between 0x00 and 0xff. + */ +#define VIDEO_FIRST_MENU 0x0000 + +/* Standard BIOS video modes (BIOS number + 0x0100) */ +#define VIDEO_FIRST_BIOS 0x0100 + +/* VESA BIOS video modes (VESA number + 0x0200) */ +#define VIDEO_FIRST_VESA 0x0200 + +/* Video7 special modes (BIOS number + 0x0900) */ +#define VIDEO_FIRST_V7 0x0900 + +# Setting of user mode (AX=mode ID) => CF=success +mode_set: + movw %ax, %bx +#if 0 + cmpb $0xff, %ah + jz setalias + + testb $VIDEO_RECALC>>8, %ah + jnz _setrec + + cmpb $VIDEO_FIRST_RESOLUTION>>8, %ah + jnc setres + + cmpb $VIDEO_FIRST_SPECIAL>>8, %ah + jz setspc + + cmpb $VIDEO_FIRST_V7>>8, %ah + jz setv7 +#endif + + cmpb $VIDEO_FIRST_VESA>>8, %ah + jnc check_vesa +#if 0 + orb %ah, %ah + jz setmenu +#endif + + decb %ah +# jz setbios Add bios modes later + +setbad: clc + ret + +check_vesa: + subb $VIDEO_FIRST_VESA>>8, %bh + orw $0x4000, %bx # Use linear frame buffer + movw $0x4f02, %ax # VESA BIOS mode set call + int $0x10 + cmpw $0x004f, %ax # AL=4f if implemented + jnz _setbad # AH=0 if OK + + stc + ret + +_setbad: jmp setbad + + .code32 + ALIGN + +.org 0x800 +wakeup_stack_begin: # Stack grows down + +.org 0xff0 # Just below end of page +wakeup_stack: +ENTRY(wakeup_end) + +.org 0x1000 + +wakeup_pmode_return: + movw $__KERNEL_DS, %ax + movw %ax, %ss + movw %ax, %ds + movw %ax, %es + movw %ax, %fs + movw %ax, %gs + movw $0x0e00 + 'u', 0xb8016 + + # reload the gdt, as we need the full 32 bit address + lgdt saved_gdt + lidt saved_idt + lldt saved_ldt + ljmp $(__KERNEL_CS),$1f +1: + movl %cr3, %eax + movl %eax, %cr3 + wbinvd + + # and restore the stack ... but you need gdt for this to work + movl saved_context_esp, %esp + + movl %cs:saved_magic, %eax + cmpl $0x12345678, %eax + jne bogus_magic + + # jump to place where we left off + movl saved_eip,%eax + jmp *%eax + +bogus_magic: + movw $0x0e00 + 'B', 0xb8018 + jmp bogus_magic + + +## +# acpi_copy_wakeup_routine +# +# Copy the above routine to low memory. +# +# Parameters: +# %eax: place to copy wakeup routine to +# +# Returned address is location of code in low memory (past data and stack) +# +ENTRY(acpi_copy_wakeup_routine) + + sgdt saved_gdt + sidt saved_idt + sldt saved_ldt + str saved_tss + + movl nx_enabled, %edx + movl %edx, real_efer_save_restore - wakeup_start (%eax) + testl $1, real_efer_save_restore - wakeup_start (%eax) + jz 2f + # save efer setting + pushl %eax + movl %eax, %ebx + mov $0xc0000080, %ecx + rdmsr + movl %edx, real_save_efer_edx - wakeup_start (%ebx) + movl %eax, real_save_efer_eax - wakeup_start (%ebx) + popl %eax +2: + + movl %cr3, %edx + movl %edx, real_save_cr3 - wakeup_start (%eax) + movl %cr4, %edx + movl %edx, real_save_cr4 - wakeup_start (%eax) + movl %cr0, %edx + movl %edx, real_save_cr0 - wakeup_start (%eax) + sgdt real_save_gdt - wakeup_start (%eax) + + movl saved_videomode, %edx + movl %edx, video_mode - wakeup_start (%eax) + movl acpi_video_flags, %edx + movl %edx, video_flags - wakeup_start (%eax) + movl $0x12345678, real_magic - wakeup_start (%eax) + movl $0x12345678, saved_magic + ret + +.data +ALIGN +ENTRY(saved_magic) .long 0 +ENTRY(saved_eip) .long 0 + +save_registers: + leal 4(%esp), %eax + movl %eax, saved_context_esp + movl %ebx, saved_context_ebx + movl %ebp, saved_context_ebp + movl %esi, saved_context_esi + movl %edi, saved_context_edi + pushfl ; popl saved_context_eflags + + movl $ret_point, saved_eip + ret + + +restore_registers: + movl saved_context_ebp, %ebp + movl saved_context_ebx, %ebx + movl saved_context_esi, %esi + movl saved_context_edi, %edi + pushl saved_context_eflags ; popfl + ret + +ENTRY(do_suspend_lowlevel) + call save_processor_state + call save_registers + pushl $3 + call acpi_enter_sleep_state + addl $4, %esp + ret + .p2align 4,,7 +ret_point: + call restore_registers + call restore_processor_state + ret + +ENTRY(do_suspend_lowlevel_s4bios) + call save_processor_state + call save_registers + call acpi_enter_sleep_state_s4bios + ret + +ALIGN +# saved registers +saved_gdt: .long 0,0 +saved_idt: .long 0,0 +saved_ldt: .long 0 +saved_tss: .long 0 + diff --git a/arch/i386/kernel/apic.c b/arch/i386/kernel/apic.c new file mode 100644 index 00000000000..35c1751ea0b --- /dev/null +++ b/arch/i386/kernel/apic.c @@ -0,0 +1,1278 @@ +/* + * Local APIC handling, local APIC timers + * + * (c) 1999, 2000 Ingo Molnar <mingo@redhat.com> + * + * Fixes + * Maciej W. Rozycki : Bits for genuine 82489DX APICs; + * thanks to Eric Gilmore + * and Rolf G. Tews + * for testing these extensively. + * Maciej W. Rozycki : Various updates and fixes. + * Mikael Pettersson : Power Management for UP-APIC. + * Pavel Machek and + * Mikael Pettersson : PM converted to driver model. + */ + +#include <linux/config.h> +#include <linux/init.h> + +#include <linux/mm.h> +#include <linux/irq.h> +#include <linux/delay.h> +#include <linux/bootmem.h> +#include <linux/smp_lock.h> +#include <linux/interrupt.h> +#include <linux/mc146818rtc.h> +#include <linux/kernel_stat.h> +#include <linux/sysdev.h> + +#include <asm/atomic.h> +#include <asm/smp.h> +#include <asm/mtrr.h> +#include <asm/mpspec.h> +#include <asm/desc.h> +#include <asm/arch_hooks.h> +#include <asm/hpet.h> + +#include <mach_apic.h> + +#include "io_ports.h" + +/* + * Debug level + */ +int apic_verbosity; + + +static void apic_pm_activate(void); + +/* + * 'what should we do if we get a hw irq event on an illegal vector'. + * each architecture has to answer this themselves. + */ +void ack_bad_irq(unsigned int irq) +{ + printk("unexpected IRQ trap at vector %02x\n", irq); + /* + * Currently unexpected vectors happen only on SMP and APIC. + * We _must_ ack these because every local APIC has only N + * irq slots per priority level, and a 'hanging, unacked' IRQ + * holds up an irq slot - in excessive cases (when multiple + * unexpected vectors occur) that might lock up the APIC + * completely. + */ + ack_APIC_irq(); +} + +void __init apic_intr_init(void) +{ +#ifdef CONFIG_SMP + smp_intr_init(); +#endif + /* self generated IPI for local APIC timer */ + set_intr_gate(LOCAL_TIMER_VECTOR, apic_timer_interrupt); + + /* IPI vectors for APIC spurious and error interrupts */ + set_intr_gate(SPURIOUS_APIC_VECTOR, spurious_interrupt); + set_intr_gate(ERROR_APIC_VECTOR, error_interrupt); + + /* thermal monitor LVT interrupt */ +#ifdef CONFIG_X86_MCE_P4THERMAL + set_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt); +#endif +} + +/* Using APIC to generate smp_local_timer_interrupt? */ +int using_apic_timer = 0; + +static DEFINE_PER_CPU(int, prof_multiplier) = 1; +static DEFINE_PER_CPU(int, prof_old_multiplier) = 1; +static DEFINE_PER_CPU(int, prof_counter) = 1; + +static int enabled_via_apicbase; + +void enable_NMI_through_LVT0 (void * dummy) +{ + unsigned int v, ver; + + ver = apic_read(APIC_LVR); + ver = GET_APIC_VERSION(ver); + v = APIC_DM_NMI; /* unmask and set to NMI */ + if (!APIC_INTEGRATED(ver)) /* 82489DX */ + v |= APIC_LVT_LEVEL_TRIGGER; + apic_write_around(APIC_LVT0, v); +} + +int get_physical_broadcast(void) +{ + unsigned int lvr, version; + lvr = apic_read(APIC_LVR); + version = GET_APIC_VERSION(lvr); + if (!APIC_INTEGRATED(version) || version >= 0x14) + return 0xff; + else + return 0xf; +} + +int get_maxlvt(void) +{ + unsigned int v, ver, maxlvt; + + v = apic_read(APIC_LVR); + ver = GET_APIC_VERSION(v); + /* 82489DXs do not report # of LVT entries. */ + maxlvt = APIC_INTEGRATED(ver) ? GET_APIC_MAXLVT(v) : 2; + return maxlvt; +} + +void clear_local_APIC(void) +{ + int maxlvt; + unsigned long v; + + maxlvt = get_maxlvt(); + + /* + * Masking an LVT entry on a P6 can trigger a local APIC error + * if the vector is zero. Mask LVTERR first to prevent this. + */ + if (maxlvt >= 3) { + v = ERROR_APIC_VECTOR; /* any non-zero vector will do */ + apic_write_around(APIC_LVTERR, v | APIC_LVT_MASKED); + } + /* + * Careful: we have to set masks only first to deassert + * any level-triggered sources. + */ + v = apic_read(APIC_LVTT); + apic_write_around(APIC_LVTT, v | APIC_LVT_MASKED); + v = apic_read(APIC_LVT0); + apic_write_around(APIC_LVT0, v | APIC_LVT_MASKED); + v = apic_read(APIC_LVT1); + apic_write_around(APIC_LVT1, v | APIC_LVT_MASKED); + if (maxlvt >= 4) { + v = apic_read(APIC_LVTPC); + apic_write_around(APIC_LVTPC, v | APIC_LVT_MASKED); + } + +/* lets not touch this if we didn't frob it */ +#ifdef CONFIG_X86_MCE_P4THERMAL + if (maxlvt >= 5) { + v = apic_read(APIC_LVTTHMR); + apic_write_around(APIC_LVTTHMR, v | APIC_LVT_MASKED); + } +#endif + /* + * Clean APIC state for other OSs: + */ + apic_write_around(APIC_LVTT, APIC_LVT_MASKED); + apic_write_around(APIC_LVT0, APIC_LVT_MASKED); + apic_write_around(APIC_LVT1, APIC_LVT_MASKED); + if (maxlvt >= 3) + apic_write_around(APIC_LVTERR, APIC_LVT_MASKED); + if (maxlvt >= 4) + apic_write_around(APIC_LVTPC, APIC_LVT_MASKED); + +#ifdef CONFIG_X86_MCE_P4THERMAL + if (maxlvt >= 5) + apic_write_around(APIC_LVTTHMR, APIC_LVT_MASKED); +#endif + v = GET_APIC_VERSION(apic_read(APIC_LVR)); + if (APIC_INTEGRATED(v)) { /* !82489DX */ + if (maxlvt > 3) /* Due to Pentium errata 3AP and 11AP. */ + apic_write(APIC_ESR, 0); + apic_read(APIC_ESR); + } +} + +void __init connect_bsp_APIC(void) +{ + if (pic_mode) { + /* + * Do not trust the local APIC being empty at bootup. + */ + clear_local_APIC(); + /* + * PIC mode, enable APIC mode in the IMCR, i.e. + * connect BSP's local APIC to INT and NMI lines. + */ + apic_printk(APIC_VERBOSE, "leaving PIC mode, " + "enabling APIC mode.\n"); + outb(0x70, 0x22); + outb(0x01, 0x23); + } + enable_apic_mode(); +} + +void disconnect_bsp_APIC(void) +{ + if (pic_mode) { + /* + * Put the board back into PIC mode (has an effect + * only on certain older boards). Note that APIC + * interrupts, including IPIs, won't work beyond + * this point! The only exception are INIT IPIs. + */ + apic_printk(APIC_VERBOSE, "disabling APIC mode, " + "entering PIC mode.\n"); + outb(0x70, 0x22); + outb(0x00, 0x23); + } +} + +void disable_local_APIC(void) +{ + unsigned long value; + + clear_local_APIC(); + + /* + * Disable APIC (implies clearing of registers + * for 82489DX!). + */ + value = apic_read(APIC_SPIV); + value &= ~APIC_SPIV_APIC_ENABLED; + apic_write_around(APIC_SPIV, value); + + if (enabled_via_apicbase) { + unsigned int l, h; + rdmsr(MSR_IA32_APICBASE, l, h); + l &= ~MSR_IA32_APICBASE_ENABLE; + wrmsr(MSR_IA32_APICBASE, l, h); + } +} + +/* + * This is to verify that we're looking at a real local APIC. + * Check these against your board if the CPUs aren't getting + * started for no apparent reason. + */ +int __init verify_local_APIC(void) +{ + unsigned int reg0, reg1; + + /* + * The version register is read-only in a real APIC. + */ + reg0 = apic_read(APIC_LVR); + apic_printk(APIC_DEBUG, "Getting VERSION: %x\n", reg0); + apic_write(APIC_LVR, reg0 ^ APIC_LVR_MASK); + reg1 = apic_read(APIC_LVR); + apic_printk(APIC_DEBUG, "Getting VERSION: %x\n", reg1); + + /* + * The two version reads above should print the same + * numbers. If the second one is different, then we + * poke at a non-APIC. + */ + if (reg1 != reg0) + return 0; + + /* + * Check if the version looks reasonably. + */ + reg1 = GET_APIC_VERSION(reg0); + if (reg1 == 0x00 || reg1 == 0xff) + return 0; + reg1 = get_maxlvt(); + if (reg1 < 0x02 || reg1 == 0xff) + return 0; + + /* + * The ID register is read/write in a real APIC. + */ + reg0 = apic_read(APIC_ID); + apic_printk(APIC_DEBUG, "Getting ID: %x\n", reg0); + + /* + * The next two are just to see if we have sane values. + * They're only really relevant if we're in Virtual Wire + * compatibility mode, but most boxes are anymore. + */ + reg0 = apic_read(APIC_LVT0); + apic_printk(APIC_DEBUG, "Getting LVT0: %x\n", reg0); + reg1 = apic_read(APIC_LVT1); + apic_printk(APIC_DEBUG, "Getting LVT1: %x\n", reg1); + + return 1; +} + +void __init sync_Arb_IDs(void) +{ + /* Unsupported on P4 - see Intel Dev. Manual Vol. 3, Ch. 8.6.1 */ + unsigned int ver = GET_APIC_VERSION(apic_read(APIC_LVR)); + if (ver >= 0x14) /* P4 or higher */ + return; + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + + apic_printk(APIC_DEBUG, "Synchronizing Arb IDs.\n"); + apic_write_around(APIC_ICR, APIC_DEST_ALLINC | APIC_INT_LEVELTRIG + | APIC_DM_INIT); +} + +extern void __error_in_apic_c (void); + +/* + * An initial setup of the virtual wire mode. + */ +void __init init_bsp_APIC(void) +{ + unsigned long value, ver; + + /* + * Don't do the setup now if we have a SMP BIOS as the + * through-I/O-APIC virtual wire mode might be active. + */ + if (smp_found_config || !cpu_has_apic) + return; + + value = apic_read(APIC_LVR); + ver = GET_APIC_VERSION(value); + + /* + * Do not trust the local APIC being empty at bootup. + */ + clear_local_APIC(); + + /* + * Enable APIC. + */ + value = apic_read(APIC_SPIV); + value &= ~APIC_VECTOR_MASK; + value |= APIC_SPIV_APIC_ENABLED; + + /* This bit is reserved on P4/Xeon and should be cleared */ + if ((boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) && (boot_cpu_data.x86 == 15)) + value &= ~APIC_SPIV_FOCUS_DISABLED; + else + value |= APIC_SPIV_FOCUS_DISABLED; + value |= SPURIOUS_APIC_VECTOR; + apic_write_around(APIC_SPIV, value); + + /* + * Set up the virtual wire mode. + */ + apic_write_around(APIC_LVT0, APIC_DM_EXTINT); + value = APIC_DM_NMI; + if (!APIC_INTEGRATED(ver)) /* 82489DX */ + value |= APIC_LVT_LEVEL_TRIGGER; + apic_write_around(APIC_LVT1, value); +} + +void __init setup_local_APIC (void) +{ + unsigned long oldvalue, value, ver, maxlvt; + + /* Pound the ESR really hard over the head with a big hammer - mbligh */ + if (esr_disable) { + apic_write(APIC_ESR, 0); + apic_write(APIC_ESR, 0); + apic_write(APIC_ESR, 0); + apic_write(APIC_ESR, 0); + } + + value = apic_read(APIC_LVR); + ver = GET_APIC_VERSION(value); + + if ((SPURIOUS_APIC_VECTOR & 0x0f) != 0x0f) + __error_in_apic_c(); + + /* + * Double-check whether this APIC is really registered. + */ + if (!apic_id_registered()) + BUG(); + + /* + * Intel recommends to set DFR, LDR and TPR before enabling + * an APIC. See e.g. "AP-388 82489DX User's Manual" (Intel + * document number 292116). So here it goes... + */ + init_apic_ldr(); + + /* + * Set Task Priority to 'accept all'. We never change this + * later on. + */ + value = apic_read(APIC_TASKPRI); + value &= ~APIC_TPRI_MASK; + apic_write_around(APIC_TASKPRI, value); + + /* + * Now that we are all set up, enable the APIC + */ + value = apic_read(APIC_SPIV); + value &= ~APIC_VECTOR_MASK; + /* + * Enable APIC + */ + value |= APIC_SPIV_APIC_ENABLED; + + /* + * Some unknown Intel IO/APIC (or APIC) errata is biting us with + * certain networking cards. If high frequency interrupts are + * happening on a particular IOAPIC pin, plus the IOAPIC routing + * entry is masked/unmasked at a high rate as well then sooner or + * later IOAPIC line gets 'stuck', no more interrupts are received + * from the device. If focus CPU is disabled then the hang goes + * away, oh well :-( + * + * [ This bug can be reproduced easily with a level-triggered + * PCI Ne2000 networking cards and PII/PIII processors, dual + * BX chipset. ] + */ + /* + * Actually disabling the focus CPU check just makes the hang less + * frequent as it makes the interrupt distributon model be more + * like LRU than MRU (the short-term load is more even across CPUs). + * See also the comment in end_level_ioapic_irq(). --macro + */ +#if 1 + /* Enable focus processor (bit==0) */ + value &= ~APIC_SPIV_FOCUS_DISABLED; +#else + /* Disable focus processor (bit==1) */ + value |= APIC_SPIV_FOCUS_DISABLED; +#endif + /* + * Set spurious IRQ vector + */ + value |= SPURIOUS_APIC_VECTOR; + apic_write_around(APIC_SPIV, value); + + /* + * Set up LVT0, LVT1: + * + * set up through-local-APIC on the BP's LINT0. This is not + * strictly necessery in pure symmetric-IO mode, but sometimes + * we delegate interrupts to the 8259A. + */ + /* + * TODO: set up through-local-APIC from through-I/O-APIC? --macro + */ + value = apic_read(APIC_LVT0) & APIC_LVT_MASKED; + if (!smp_processor_id() && (pic_mode || !value)) { + value = APIC_DM_EXTINT; + apic_printk(APIC_VERBOSE, "enabled ExtINT on CPU#%d\n", + smp_processor_id()); + } else { + value = APIC_DM_EXTINT | APIC_LVT_MASKED; + apic_printk(APIC_VERBOSE, "masked ExtINT on CPU#%d\n", + smp_processor_id()); + } + apic_write_around(APIC_LVT0, value); + + /* + * only the BP should see the LINT1 NMI signal, obviously. + */ + if (!smp_processor_id()) + value = APIC_DM_NMI; + else + value = APIC_DM_NMI | APIC_LVT_MASKED; + if (!APIC_INTEGRATED(ver)) /* 82489DX */ + value |= APIC_LVT_LEVEL_TRIGGER; + apic_write_around(APIC_LVT1, value); + + if (APIC_INTEGRATED(ver) && !esr_disable) { /* !82489DX */ + maxlvt = get_maxlvt(); + if (maxlvt > 3) /* Due to the Pentium erratum 3AP. */ + apic_write(APIC_ESR, 0); + oldvalue = apic_read(APIC_ESR); + + value = ERROR_APIC_VECTOR; // enables sending errors + apic_write_around(APIC_LVTERR, value); + /* + * spec says clear errors after enabling vector. + */ + if (maxlvt > 3) + apic_write(APIC_ESR, 0); + value = apic_read(APIC_ESR); + if (value != oldvalue) + apic_printk(APIC_VERBOSE, "ESR value before enabling " + "vector: 0x%08lx after: 0x%08lx\n", + oldvalue, value); + } else { + if (esr_disable) + /* + * Something untraceble is creating bad interrupts on + * secondary quads ... for the moment, just leave the + * ESR disabled - we can't do anything useful with the + * errors anyway - mbligh + */ + printk("Leaving ESR disabled.\n"); + else + printk("No ESR for 82489DX.\n"); + } + + if (nmi_watchdog == NMI_LOCAL_APIC) + setup_apic_nmi_watchdog(); + apic_pm_activate(); +} + +/* + * If Linux enabled the LAPIC against the BIOS default + * disable it down before re-entering the BIOS on shutdown. + * Otherwise the BIOS may get confused and not power-off. + */ +void lapic_shutdown(void) +{ + if (!cpu_has_apic || !enabled_via_apicbase) + return; + + local_irq_disable(); + disable_local_APIC(); + local_irq_enable(); +} + +#ifdef CONFIG_PM + +static struct { + int active; + /* r/w apic fields */ + unsigned int apic_id; + unsigned int apic_taskpri; + unsigned int apic_ldr; + unsigned int apic_dfr; + unsigned int apic_spiv; + unsigned int apic_lvtt; + unsigned int apic_lvtpc; + unsigned int apic_lvt0; + unsigned int apic_lvt1; + unsigned int apic_lvterr; + unsigned int apic_tmict; + unsigned int apic_tdcr; + unsigned int apic_thmr; +} apic_pm_state; + +static int lapic_suspend(struct sys_device *dev, u32 state) +{ + unsigned long flags; + + if (!apic_pm_state.active) + return 0; + + apic_pm_state.apic_id = apic_read(APIC_ID); + apic_pm_state.apic_taskpri = apic_read(APIC_TASKPRI); + apic_pm_state.apic_ldr = apic_read(APIC_LDR); + apic_pm_state.apic_dfr = apic_read(APIC_DFR); + apic_pm_state.apic_spiv = apic_read(APIC_SPIV); + apic_pm_state.apic_lvtt = apic_read(APIC_LVTT); + apic_pm_state.apic_lvtpc = apic_read(APIC_LVTPC); + apic_pm_state.apic_lvt0 = apic_read(APIC_LVT0); + apic_pm_state.apic_lvt1 = apic_read(APIC_LVT1); + apic_pm_state.apic_lvterr = apic_read(APIC_LVTERR); + apic_pm_state.apic_tmict = apic_read(APIC_TMICT); + apic_pm_state.apic_tdcr = apic_read(APIC_TDCR); + apic_pm_state.apic_thmr = apic_read(APIC_LVTTHMR); + + local_irq_save(flags); + disable_local_APIC(); + local_irq_restore(flags); + return 0; +} + +static int lapic_resume(struct sys_device *dev) +{ + unsigned int l, h; + unsigned long flags; + + if (!apic_pm_state.active) + return 0; + + local_irq_save(flags); + + /* + * Make sure the APICBASE points to the right address + * + * FIXME! This will be wrong if we ever support suspend on + * SMP! We'll need to do this as part of the CPU restore! + */ + rdmsr(MSR_IA32_APICBASE, l, h); + l &= ~MSR_IA32_APICBASE_BASE; + l |= MSR_IA32_APICBASE_ENABLE | mp_lapic_addr; + wrmsr(MSR_IA32_APICBASE, l, h); + + apic_write(APIC_LVTERR, ERROR_APIC_VECTOR | APIC_LVT_MASKED); + apic_write(APIC_ID, apic_pm_state.apic_id); + apic_write(APIC_DFR, apic_pm_state.apic_dfr); + apic_write(APIC_LDR, apic_pm_state.apic_ldr); + apic_write(APIC_TASKPRI, apic_pm_state.apic_taskpri); + apic_write(APIC_SPIV, apic_pm_state.apic_spiv); + apic_write(APIC_LVT0, apic_pm_state.apic_lvt0); + apic_write(APIC_LVT1, apic_pm_state.apic_lvt1); + apic_write(APIC_LVTTHMR, apic_pm_state.apic_thmr); + apic_write(APIC_LVTPC, apic_pm_state.apic_lvtpc); + apic_write(APIC_LVTT, apic_pm_state.apic_lvtt); + apic_write(APIC_TDCR, apic_pm_state.apic_tdcr); + apic_write(APIC_TMICT, apic_pm_state.apic_tmict); + apic_write(APIC_ESR, 0); + apic_read(APIC_ESR); + apic_write(APIC_LVTERR, apic_pm_state.apic_lvterr); + apic_write(APIC_ESR, 0); + apic_read(APIC_ESR); + local_irq_restore(flags); + return 0; +} + +/* + * This device has no shutdown method - fully functioning local APICs + * are needed on every CPU up until machine_halt/restart/poweroff. + */ + +static struct sysdev_class lapic_sysclass = { + set_kset_name("lapic"), + .resume = lapic_resume, + .suspend = lapic_suspend, +}; + +static struct sys_device device_lapic = { + .id = 0, + .cls = &lapic_sysclass, +}; + +static void __init apic_pm_activate(void) +{ + apic_pm_state.active = 1; +} + +static int __init init_lapic_sysfs(void) +{ + int error; + + if (!cpu_has_apic) + return 0; + /* XXX: remove suspend/resume procs if !apic_pm_state.active? */ + + error = sysdev_class_register(&lapic_sysclass); + if (!error) + error = sysdev_register(&device_lapic); + return error; +} +device_initcall(init_lapic_sysfs); + +#else /* CONFIG_PM */ + +static void apic_pm_activate(void) { } + +#endif /* CONFIG_PM */ + +/* + * Detect and enable local APICs on non-SMP boards. + * Original code written by Keir Fraser. + */ + +/* + * Knob to control our willingness to enable the local APIC. + */ +int enable_local_apic __initdata = 0; /* -1=force-disable, +1=force-enable */ + +static int __init lapic_disable(char *str) +{ + enable_local_apic = -1; + clear_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability); + return 0; +} +__setup("nolapic", lapic_disable); + +static int __init lapic_enable(char *str) +{ + enable_local_apic = 1; + return 0; +} +__setup("lapic", lapic_enable); + +static int __init apic_set_verbosity(char *str) +{ + if (strcmp("debug", str) == 0) + apic_verbosity = APIC_DEBUG; + else if (strcmp("verbose", str) == 0) + apic_verbosity = APIC_VERBOSE; + else + printk(KERN_WARNING "APIC Verbosity level %s not recognised" + " use apic=verbose or apic=debug", str); + + return 0; +} + +__setup("apic=", apic_set_verbosity); + +static int __init detect_init_APIC (void) +{ + u32 h, l, features; + extern void get_cpu_vendor(struct cpuinfo_x86*); + + /* Disabled by kernel option? */ + if (enable_local_apic < 0) + return -1; + + /* Workaround for us being called before identify_cpu(). */ + get_cpu_vendor(&boot_cpu_data); + + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + if ((boot_cpu_data.x86 == 6 && boot_cpu_data.x86_model > 1) || + (boot_cpu_data.x86 == 15)) + break; + goto no_apic; + case X86_VENDOR_INTEL: + if (boot_cpu_data.x86 == 6 || boot_cpu_data.x86 == 15 || + (boot_cpu_data.x86 == 5 && cpu_has_apic)) + break; + goto no_apic; + default: + goto no_apic; + } + + if (!cpu_has_apic) { + /* + * Over-ride BIOS and try to enable the local + * APIC only if "lapic" specified. + */ + if (enable_local_apic <= 0) { + printk("Local APIC disabled by BIOS -- " + "you can enable it with \"lapic\"\n"); + return -1; + } + /* + * Some BIOSes disable the local APIC in the + * APIC_BASE MSR. This can only be done in + * software for Intel P6 or later and AMD K7 + * (Model > 1) or later. + */ + rdmsr(MSR_IA32_APICBASE, l, h); + if (!(l & MSR_IA32_APICBASE_ENABLE)) { + printk("Local APIC disabled by BIOS -- reenabling.\n"); + l &= ~MSR_IA32_APICBASE_BASE; + l |= MSR_IA32_APICBASE_ENABLE | APIC_DEFAULT_PHYS_BASE; + wrmsr(MSR_IA32_APICBASE, l, h); + enabled_via_apicbase = 1; + } + } + /* + * The APIC feature bit should now be enabled + * in `cpuid' + */ + features = cpuid_edx(1); + if (!(features & (1 << X86_FEATURE_APIC))) { + printk("Could not enable APIC!\n"); + return -1; + } + set_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability); + mp_lapic_addr = APIC_DEFAULT_PHYS_BASE; + + /* The BIOS may have set up the APIC at some other address */ + rdmsr(MSR_IA32_APICBASE, l, h); + if (l & MSR_IA32_APICBASE_ENABLE) + mp_lapic_addr = l & MSR_IA32_APICBASE_BASE; + + if (nmi_watchdog != NMI_NONE) + nmi_watchdog = NMI_LOCAL_APIC; + + printk("Found and enabled local APIC!\n"); + + apic_pm_activate(); + + return 0; + +no_apic: + printk("No local APIC present or hardware disabled\n"); + return -1; +} + +void __init init_apic_mappings(void) +{ + unsigned long apic_phys; + + /* + * If no local APIC can be found then set up a fake all + * zeroes page to simulate the local APIC and another + * one for the IO-APIC. + */ + if (!smp_found_config && detect_init_APIC()) { + apic_phys = (unsigned long) alloc_bootmem_pages(PAGE_SIZE); + apic_phys = __pa(apic_phys); + } else + apic_phys = mp_lapic_addr; + + set_fixmap_nocache(FIX_APIC_BASE, apic_phys); + printk(KERN_DEBUG "mapped APIC to %08lx (%08lx)\n", APIC_BASE, + apic_phys); + + /* + * Fetch the APIC ID of the BSP in case we have a + * default configuration (or the MP table is broken). + */ + if (boot_cpu_physical_apicid == -1U) + boot_cpu_physical_apicid = GET_APIC_ID(apic_read(APIC_ID)); + +#ifdef CONFIG_X86_IO_APIC + { + unsigned long ioapic_phys, idx = FIX_IO_APIC_BASE_0; + int i; + + for (i = 0; i < nr_ioapics; i++) { + if (smp_found_config) { + ioapic_phys = mp_ioapics[i].mpc_apicaddr; + if (!ioapic_phys) { + printk(KERN_ERR + "WARNING: bogus zero IO-APIC " + "address found in MPTABLE, " + "disabling IO/APIC support!\n"); + smp_found_config = 0; + skip_ioapic_setup = 1; + goto fake_ioapic_page; + } + } else { +fake_ioapic_page: + ioapic_phys = (unsigned long) + alloc_bootmem_pages(PAGE_SIZE); + ioapic_phys = __pa(ioapic_phys); + } + set_fixmap_nocache(idx, ioapic_phys); + printk(KERN_DEBUG "mapped IOAPIC to %08lx (%08lx)\n", + __fix_to_virt(idx), ioapic_phys); + idx++; + } + } +#endif +} + +/* + * This part sets up the APIC 32 bit clock in LVTT1, with HZ interrupts + * per second. We assume that the caller has already set up the local + * APIC. + * + * The APIC timer is not exactly sync with the external timer chip, it + * closely follows bus clocks. + */ + +/* + * The timer chip is already set up at HZ interrupts per second here, + * but we do not accept timer interrupts yet. We only allow the BP + * to calibrate. + */ +static unsigned int __init get_8254_timer_count(void) +{ + extern spinlock_t i8253_lock; + unsigned long flags; + + unsigned int count; + + spin_lock_irqsave(&i8253_lock, flags); + + outb_p(0x00, PIT_MODE); + count = inb_p(PIT_CH0); + count |= inb_p(PIT_CH0) << 8; + + spin_unlock_irqrestore(&i8253_lock, flags); + + return count; +} + +/* next tick in 8254 can be caught by catching timer wraparound */ +static void __init wait_8254_wraparound(void) +{ + unsigned int curr_count, prev_count; + + curr_count = get_8254_timer_count(); + do { + prev_count = curr_count; + curr_count = get_8254_timer_count(); + + /* workaround for broken Mercury/Neptune */ + if (prev_count >= curr_count + 0x100) + curr_count = get_8254_timer_count(); + + } while (prev_count >= curr_count); +} + +/* + * Default initialization for 8254 timers. If we use other timers like HPET, + * we override this later + */ +void (*wait_timer_tick)(void) __initdata = wait_8254_wraparound; + +/* + * This function sets up the local APIC timer, with a timeout of + * 'clocks' APIC bus clock. During calibration we actually call + * this function twice on the boot CPU, once with a bogus timeout + * value, second time for real. The other (noncalibrating) CPUs + * call this function only once, with the real, calibrated value. + * + * We do reads before writes even if unnecessary, to get around the + * P5 APIC double write bug. + */ + +#define APIC_DIVISOR 16 + +static void __setup_APIC_LVTT(unsigned int clocks) +{ + unsigned int lvtt_value, tmp_value, ver; + + ver = GET_APIC_VERSION(apic_read(APIC_LVR)); + lvtt_value = APIC_LVT_TIMER_PERIODIC | LOCAL_TIMER_VECTOR; + if (!APIC_INTEGRATED(ver)) + lvtt_value |= SET_APIC_TIMER_BASE(APIC_TIMER_BASE_DIV); + apic_write_around(APIC_LVTT, lvtt_value); + + /* + * Divide PICLK by 16 + */ + tmp_value = apic_read(APIC_TDCR); + apic_write_around(APIC_TDCR, (tmp_value + & ~(APIC_TDR_DIV_1 | APIC_TDR_DIV_TMBASE)) + | APIC_TDR_DIV_16); + + apic_write_around(APIC_TMICT, clocks/APIC_DIVISOR); +} + +static void __init setup_APIC_timer(unsigned int clocks) +{ + unsigned long flags; + + local_irq_save(flags); + + /* + * Wait for IRQ0's slice: + */ + wait_timer_tick(); + + __setup_APIC_LVTT(clocks); + + local_irq_restore(flags); +} + +/* + * In this function we calibrate APIC bus clocks to the external + * timer. Unfortunately we cannot use jiffies and the timer irq + * to calibrate, since some later bootup code depends on getting + * the first irq? Ugh. + * + * We want to do the calibration only once since we + * want to have local timer irqs syncron. CPUs connected + * by the same APIC bus have the very same bus frequency. + * And we want to have irqs off anyways, no accidental + * APIC irq that way. + */ + +static int __init calibrate_APIC_clock(void) +{ + unsigned long long t1 = 0, t2 = 0; + long tt1, tt2; + long result; + int i; + const int LOOPS = HZ/10; + + apic_printk(APIC_VERBOSE, "calibrating APIC timer ...\n"); + + /* + * Put whatever arbitrary (but long enough) timeout + * value into the APIC clock, we just want to get the + * counter running for calibration. + */ + __setup_APIC_LVTT(1000000000); + + /* + * The timer chip counts down to zero. Let's wait + * for a wraparound to start exact measurement: + * (the current tick might have been already half done) + */ + + wait_timer_tick(); + + /* + * We wrapped around just now. Let's start: + */ + if (cpu_has_tsc) + rdtscll(t1); + tt1 = apic_read(APIC_TMCCT); + + /* + * Let's wait LOOPS wraprounds: + */ + for (i = 0; i < LOOPS; i++) + wait_timer_tick(); + + tt2 = apic_read(APIC_TMCCT); + if (cpu_has_tsc) + rdtscll(t2); + + /* + * The APIC bus clock counter is 32 bits only, it + * might have overflown, but note that we use signed + * longs, thus no extra care needed. + * + * underflown to be exact, as the timer counts down ;) + */ + + result = (tt1-tt2)*APIC_DIVISOR/LOOPS; + + if (cpu_has_tsc) + apic_printk(APIC_VERBOSE, "..... CPU clock speed is " + "%ld.%04ld MHz.\n", + ((long)(t2-t1)/LOOPS)/(1000000/HZ), + ((long)(t2-t1)/LOOPS)%(1000000/HZ)); + + apic_printk(APIC_VERBOSE, "..... host bus clock speed is " + "%ld.%04ld MHz.\n", + result/(1000000/HZ), + result%(1000000/HZ)); + + return result; +} + +static unsigned int calibration_result; + +void __init setup_boot_APIC_clock(void) +{ + apic_printk(APIC_VERBOSE, "Using local APIC timer interrupts.\n"); + using_apic_timer = 1; + + local_irq_disable(); + + calibration_result = calibrate_APIC_clock(); + /* + * Now set up the timer for real. + */ + setup_APIC_timer(calibration_result); + + local_irq_enable(); +} + +void __init setup_secondary_APIC_clock(void) +{ + setup_APIC_timer(calibration_result); +} + +void __init disable_APIC_timer(void) +{ + if (using_apic_timer) { + unsigned long v; + + v = apic_read(APIC_LVTT); + apic_write_around(APIC_LVTT, v | APIC_LVT_MASKED); + } +} + +void enable_APIC_timer(void) +{ + if (using_apic_timer) { + unsigned long v; + + v = apic_read(APIC_LVTT); + apic_write_around(APIC_LVTT, v & ~APIC_LVT_MASKED); + } +} + +/* + * the frequency of the profiling timer can be changed + * by writing a multiplier value into /proc/profile. + */ +int setup_profiling_timer(unsigned int multiplier) +{ + int i; + + /* + * Sanity check. [at least 500 APIC cycles should be + * between APIC interrupts as a rule of thumb, to avoid + * irqs flooding us] + */ + if ( (!multiplier) || (calibration_result/multiplier < 500)) + return -EINVAL; + + /* + * Set the new multiplier for each CPU. CPUs don't start using the + * new values until the next timer interrupt in which they do process + * accounting. At that time they also adjust their APIC timers + * accordingly. + */ + for (i = 0; i < NR_CPUS; ++i) + per_cpu(prof_multiplier, i) = multiplier; + + return 0; +} + +#undef APIC_DIVISOR + +/* + * Local timer interrupt handler. It does both profiling and + * process statistics/rescheduling. + * + * We do profiling in every local tick, statistics/rescheduling + * happen only every 'profiling multiplier' ticks. The default + * multiplier is 1 and it can be changed by writing the new multiplier + * value into /proc/profile. + */ + +inline void smp_local_timer_interrupt(struct pt_regs * regs) +{ + int cpu = smp_processor_id(); + + profile_tick(CPU_PROFILING, regs); + if (--per_cpu(prof_counter, cpu) <= 0) { + /* + * The multiplier may have changed since the last time we got + * to this point as a result of the user writing to + * /proc/profile. In this case we need to adjust the APIC + * timer accordingly. + * + * Interrupts are already masked off at this point. + */ + per_cpu(prof_counter, cpu) = per_cpu(prof_multiplier, cpu); + if (per_cpu(prof_counter, cpu) != + per_cpu(prof_old_multiplier, cpu)) { + __setup_APIC_LVTT( + calibration_result/ + per_cpu(prof_counter, cpu)); + per_cpu(prof_old_multiplier, cpu) = + per_cpu(prof_counter, cpu); + } + +#ifdef CONFIG_SMP + update_process_times(user_mode(regs)); +#endif + } + + /* + * We take the 'long' return path, and there every subsystem + * grabs the apropriate locks (kernel lock/ irq lock). + * + * we might want to decouple profiling from the 'long path', + * and do the profiling totally in assembly. + * + * Currently this isn't too much of an issue (performance wise), + * we can take more than 100K local irqs per second on a 100 MHz P5. + */ +} + +/* + * Local APIC timer interrupt. This is the most natural way for doing + * local interrupts, but local timer interrupts can be emulated by + * broadcast interrupts too. [in case the hw doesn't support APIC timers] + * + * [ if a single-CPU system runs an SMP kernel then we call the local + * interrupt as well. Thus we cannot inline the local irq ... ] + */ + +fastcall void smp_apic_timer_interrupt(struct pt_regs *regs) +{ + int cpu = smp_processor_id(); + + /* + * the NMI deadlock-detector uses this. + */ + per_cpu(irq_stat, cpu).apic_timer_irqs++; + + /* + * NOTE! We'd better ACK the irq immediately, + * because timer handling can be slow. + */ + ack_APIC_irq(); + /* + * update_process_times() expects us to have done irq_enter(). + * Besides, if we don't timer interrupts ignore the global + * interrupt lock, which is the WrongThing (tm) to do. + */ + irq_enter(); + smp_local_timer_interrupt(regs); + irq_exit(); +} + +/* + * This interrupt should _never_ happen with our APIC/SMP architecture + */ +fastcall void smp_spurious_interrupt(struct pt_regs *regs) +{ + unsigned long v; + + irq_enter(); + /* + * Check if this really is a spurious interrupt and ACK it + * if it is a vectored one. Just in case... + * Spurious interrupts should not be ACKed. + */ + v = apic_read(APIC_ISR + ((SPURIOUS_APIC_VECTOR & ~0x1f) >> 1)); + if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f))) + ack_APIC_irq(); + + /* see sw-dev-man vol 3, chapter 7.4.13.5 */ + printk(KERN_INFO "spurious APIC interrupt on CPU#%d, should never happen.\n", + smp_processor_id()); + irq_exit(); +} + +/* + * This interrupt should never happen with our APIC/SMP architecture + */ + +fastcall void smp_error_interrupt(struct pt_regs *regs) +{ + unsigned long v, v1; + + irq_enter(); + /* First tickle the hardware, only then report what went on. -- REW */ + v = apic_read(APIC_ESR); + apic_write(APIC_ESR, 0); + v1 = apic_read(APIC_ESR); + ack_APIC_irq(); + atomic_inc(&irq_err_count); + + /* Here is what the APIC error bits mean: + 0: Send CS error + 1: Receive CS error + 2: Send accept error + 3: Receive accept error + 4: Reserved + 5: Send illegal vector + 6: Received illegal vector + 7: Illegal register address + */ + printk (KERN_DEBUG "APIC error on CPU%d: %02lx(%02lx)\n", + smp_processor_id(), v , v1); + irq_exit(); +} + +/* + * This initializes the IO-APIC and APIC hardware if this is + * a UP kernel. + */ +int __init APIC_init_uniprocessor (void) +{ + if (enable_local_apic < 0) + clear_bit(X86_FEATURE_APIC, boot_cpu_data.x86_capability); + + if (!smp_found_config && !cpu_has_apic) + return -1; + + /* + * Complain if the BIOS pretends there is one. + */ + if (!cpu_has_apic && APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid])) { + printk(KERN_ERR "BIOS bug, local APIC #%d not detected!...\n", + boot_cpu_physical_apicid); + return -1; + } + + verify_local_APIC(); + + connect_bsp_APIC(); + + phys_cpu_present_map = physid_mask_of_physid(boot_cpu_physical_apicid); + + setup_local_APIC(); + + if (nmi_watchdog == NMI_LOCAL_APIC) + check_nmi_watchdog(); +#ifdef CONFIG_X86_IO_APIC + if (smp_found_config) + if (!skip_ioapic_setup && nr_ioapics) + setup_IO_APIC(); +#endif + setup_boot_APIC_clock(); + + return 0; +} diff --git a/arch/i386/kernel/apm.c b/arch/i386/kernel/apm.c new file mode 100644 index 00000000000..45641a87255 --- /dev/null +++ b/arch/i386/kernel/apm.c @@ -0,0 +1,2428 @@ +/* -*- linux-c -*- + * APM BIOS driver for Linux + * Copyright 1994-2001 Stephen Rothwell (sfr@canb.auug.org.au) + * + * Initial development of this driver was funded by NEC Australia P/L + * and NEC Corporation + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the + * Free Software Foundation; either version 2, or (at your option) any + * later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * October 1995, Rik Faith (faith@cs.unc.edu): + * Minor enhancements and updates (to the patch set) for 1.3.x + * Documentation + * January 1996, Rik Faith (faith@cs.unc.edu): + * Make /proc/apm easy to format (bump driver version) + * March 1996, Rik Faith (faith@cs.unc.edu): + * Prohibit APM BIOS calls unless apm_enabled. + * (Thanks to Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de>) + * April 1996, Stephen Rothwell (sfr@canb.auug.org.au) + * Version 1.0 and 1.1 + * May 1996, Version 1.2 + * Feb 1998, Version 1.3 + * Feb 1998, Version 1.4 + * Aug 1998, Version 1.5 + * Sep 1998, Version 1.6 + * Nov 1998, Version 1.7 + * Jan 1999, Version 1.8 + * Jan 1999, Version 1.9 + * Oct 1999, Version 1.10 + * Nov 1999, Version 1.11 + * Jan 2000, Version 1.12 + * Feb 2000, Version 1.13 + * Nov 2000, Version 1.14 + * Oct 2001, Version 1.15 + * Jan 2002, Version 1.16 + * Oct 2002, Version 1.16ac + * + * History: + * 0.6b: first version in official kernel, Linux 1.3.46 + * 0.7: changed /proc/apm format, Linux 1.3.58 + * 0.8: fixed gcc 2.7.[12] compilation problems, Linux 1.3.59 + * 0.9: only call bios if bios is present, Linux 1.3.72 + * 1.0: use fixed device number, consolidate /proc/apm into this file, + * Linux 1.3.85 + * 1.1: support user-space standby and suspend, power off after system + * halted, Linux 1.3.98 + * 1.2: When resetting RTC after resume, take care so that the time + * is only incorrect by 30-60mS (vs. 1S previously) (Gabor J. Toth + * <jtoth@princeton.edu>); improve interaction between + * screen-blanking and gpm (Stephen Rothwell); Linux 1.99.4 + * 1.2a:Simple change to stop mysterious bug reports with SMP also added + * levels to the printk calls. APM is not defined for SMP machines. + * The new replacment for it is, but Linux doesn't yet support this. + * Alan Cox Linux 2.1.55 + * 1.3: Set up a valid data descriptor 0x40 for buggy BIOS's + * 1.4: Upgraded to support APM 1.2. Integrated ThinkPad suspend patch by + * Dean Gaudet <dgaudet@arctic.org>. + * C. Scott Ananian <cananian@alumni.princeton.edu> Linux 2.1.87 + * 1.5: Fix segment register reloading (in case of bad segments saved + * across BIOS call). + * Stephen Rothwell + * 1.6: Cope with complier/assembler differences. + * Only try to turn off the first display device. + * Fix OOPS at power off with no APM BIOS by Jan Echternach + * <echter@informatik.uni-rostock.de> + * Stephen Rothwell + * 1.7: Modify driver's cached copy of the disabled/disengaged flags + * to reflect current state of APM BIOS. + * Chris Rankin <rankinc@bellsouth.net> + * Reset interrupt 0 timer to 100Hz after suspend + * Chad Miller <cmiller@surfsouth.com> + * Add CONFIG_APM_IGNORE_SUSPEND_BOUNCE + * Richard Gooch <rgooch@atnf.csiro.au> + * Allow boot time disabling of APM + * Make boot messages far less verbose by default + * Make asm safer + * Stephen Rothwell + * 1.8: Add CONFIG_APM_RTC_IS_GMT + * Richard Gooch <rgooch@atnf.csiro.au> + * change APM_NOINTS to CONFIG_APM_ALLOW_INTS + * remove dependency on CONFIG_PROC_FS + * Stephen Rothwell + * 1.9: Fix small typo. <laslo@wodip.opole.pl> + * Try to cope with BIOS's that need to have all display + * devices blanked and not just the first one. + * Ross Paterson <ross@soi.city.ac.uk> + * Fix segment limit setting it has always been wrong as + * the segments needed to have byte granularity. + * Mark a few things __init. + * Add hack to allow power off of SMP systems by popular request. + * Use CONFIG_SMP instead of __SMP__ + * Ignore BOUNCES for three seconds. + * Stephen Rothwell + * 1.10: Fix for Thinkpad return code. + * Merge 2.2 and 2.3 drivers. + * Remove APM dependencies in arch/i386/kernel/process.c + * Remove APM dependencies in drivers/char/sysrq.c + * Reset time across standby. + * Allow more inititialisation on SMP. + * Remove CONFIG_APM_POWER_OFF and make it boot time + * configurable (default on). + * Make debug only a boot time parameter (remove APM_DEBUG). + * Try to blank all devices on any error. + * 1.11: Remove APM dependencies in drivers/char/console.c + * Check nr_running to detect if we are idle (from + * Borislav Deianov <borislav@lix.polytechnique.fr>) + * Fix for bioses that don't zero the top part of the + * entrypoint offset (Mario Sitta <sitta@al.unipmn.it>) + * (reported by Panos Katsaloulis <teras@writeme.com>). + * Real mode power off patch (Walter Hofmann + * <Walter.Hofmann@physik.stud.uni-erlangen.de>). + * 1.12: Remove CONFIG_SMP as the compiler will optimize + * the code away anyway (smp_num_cpus == 1 in UP) + * noted by Artur Skawina <skawina@geocities.com>. + * Make power off under SMP work again. + * Fix thinko with initial engaging of BIOS. + * Make sure power off only happens on CPU 0 + * (Paul "Rusty" Russell <rusty@rustcorp.com.au>). + * Do error notification to user mode if BIOS calls fail. + * Move entrypoint offset fix to ...boot/setup.S + * where it belongs (Cosmos <gis88564@cis.nctu.edu.tw>). + * Remove smp-power-off. SMP users must now specify + * "apm=power-off" on the kernel command line. Suggested + * by Jim Avera <jima@hal.com>, modified by Alan Cox + * <alan@lxorguk.ukuu.org.uk>. + * Register the /proc/apm entry even on SMP so that + * scripts that check for it before doing power off + * work (Jim Avera <jima@hal.com>). + * 1.13: Changes for new pm_ interfaces (Andy Henroid + * <andy_henroid@yahoo.com>). + * Modularize the code. + * Fix the Thinkpad (again) :-( (CONFIG_APM_IGNORE_MULTIPLE_SUSPENDS + * is now the way life works). + * Fix thinko in suspend() (wrong return). + * Notify drivers on critical suspend. + * Make kapmd absorb more idle time (Pavel Machek <pavel@suse.cz> + * modified by sfr). + * Disable interrupts while we are suspended (Andy Henroid + * <andy_henroid@yahoo.com> fixed by sfr). + * Make power off work on SMP again (Tony Hoyle + * <tmh@magenta-logic.com> and <zlatko@iskon.hr>) modified by sfr. + * Remove CONFIG_APM_SUSPEND_BOUNCE. The bounce ignore + * interval is now configurable. + * 1.14: Make connection version persist across module unload/load. + * Enable and engage power management earlier. + * Disengage power management on module unload. + * Changed to use the sysrq-register hack for registering the + * power off function called by magic sysrq based upon discussions + * in irc://irc.openprojects.net/#kernelnewbies + * (Crutcher Dunnavant <crutcher+kernel@datastacks.com>). + * Make CONFIG_APM_REAL_MODE_POWER_OFF run time configurable. + * (Arjan van de Ven <arjanv@redhat.com>) modified by sfr. + * Work around byte swap bug in one of the Vaio's BIOS's + * (Marc Boucher <marc@mbsi.ca>). + * Exposed the disable flag to dmi so that we can handle known + * broken APM (Alan Cox <alan@redhat.com>). + * 1.14ac: If the BIOS says "I slowed the CPU down" then don't spin + * calling it - instead idle. (Alan Cox <alan@redhat.com>) + * If an APM idle fails log it and idle sensibly + * 1.15: Don't queue events to clients who open the device O_WRONLY. + * Don't expect replies from clients who open the device O_RDONLY. + * (Idea from Thomas Hood) + * Minor waitqueue cleanups. (John Fremlin <chief@bandits.org>) + * 1.16: Fix idle calling. (Andreas Steinmetz <ast@domdv.de> et al.) + * Notify listeners of standby or suspend events before notifying + * drivers. Return EBUSY to ioctl() if suspend is rejected. + * (Russell King <rmk@arm.linux.org.uk> and Thomas Hood) + * Ignore first resume after we generate our own resume event + * after a suspend (Thomas Hood) + * Daemonize now gets rid of our controlling terminal (sfr). + * CONFIG_APM_CPU_IDLE now just affects the default value of + * idle_threshold (sfr). + * Change name of kernel apm daemon (as it no longer idles) (sfr). + * 1.16ac: Fix up SMP support somewhat. You can now force SMP on and we + * make _all_ APM calls on the CPU#0. Fix unsafe sign bug. + * TODO: determine if its "boot CPU" or "CPU0" we want to lock to. + * + * APM 1.1 Reference: + * + * Intel Corporation, Microsoft Corporation. Advanced Power Management + * (APM) BIOS Interface Specification, Revision 1.1, September 1993. + * Intel Order Number 241704-001. Microsoft Part Number 781-110-X01. + * + * [This document is available free from Intel by calling 800.628.8686 (fax + * 916.356.6100) or 800.548.4725; or via anonymous ftp from + * ftp://ftp.intel.com/pub/IAL/software_specs/apmv11.doc. It is also + * available from Microsoft by calling 206.882.8080.] + * + * APM 1.2 Reference: + * Intel Corporation, Microsoft Corporation. Advanced Power Management + * (APM) BIOS Interface Specification, Revision 1.2, February 1996. + * + * [This document is available from Microsoft at: + * http://www.microsoft.com/hwdev/busbios/amp_12.htm] + */ + +#include <linux/config.h> +#include <linux/module.h> + +#include <linux/poll.h> +#include <linux/types.h> +#include <linux/stddef.h> +#include <linux/timer.h> +#include <linux/fcntl.h> +#include <linux/slab.h> +#include <linux/stat.h> +#include <linux/proc_fs.h> +#include <linux/miscdevice.h> +#include <linux/apm_bios.h> +#include <linux/init.h> +#include <linux/time.h> +#include <linux/sched.h> +#include <linux/pm.h> +#include <linux/device.h> +#include <linux/kernel.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/dmi.h> +#include <linux/suspend.h> + +#include <asm/system.h> +#include <asm/uaccess.h> +#include <asm/desc.h> + +#include "io_ports.h" + +extern spinlock_t i8253_lock; +extern unsigned long get_cmos_time(void); +extern void machine_real_restart(unsigned char *, int); + +#if defined(CONFIG_APM_DISPLAY_BLANK) && defined(CONFIG_VT) +extern int (*console_blank_hook)(int); +#endif + +/* + * The apm_bios device is one of the misc char devices. + * This is its minor number. + */ +#define APM_MINOR_DEV 134 + +/* + * See Documentation/Config.help for the configuration options. + * + * Various options can be changed at boot time as follows: + * (We allow underscores for compatibility with the modules code) + * apm=on/off enable/disable APM + * [no-]allow[-_]ints allow interrupts during BIOS calls + * [no-]broken[-_]psr BIOS has a broken GetPowerStatus call + * [no-]realmode[-_]power[-_]off switch to real mode before + * powering off + * [no-]debug log some debugging messages + * [no-]power[-_]off power off on shutdown + * [no-]smp Use apm even on an SMP box + * bounce[-_]interval=<n> number of ticks to ignore suspend + * bounces + * idle[-_]threshold=<n> System idle percentage above which to + * make APM BIOS idle calls. Set it to + * 100 to disable. + * idle[-_]period=<n> Period (in 1/100s of a second) over + * which the idle percentage is + * calculated. + */ + +/* KNOWN PROBLEM MACHINES: + * + * U: TI 4000M TravelMate: BIOS is *NOT* APM compliant + * [Confirmed by TI representative] + * ?: ACER 486DX4/75: uses dseg 0040, in violation of APM specification + * [Confirmed by BIOS disassembly] + * [This may work now ...] + * P: Toshiba 1950S: battery life information only gets updated after resume + * P: Midwest Micro Soundbook Elite DX2/66 monochrome: screen blanking + * broken in BIOS [Reported by Garst R. Reese <reese@isn.net>] + * ?: AcerNote-950: oops on reading /proc/apm - workaround is a WIP + * Neale Banks <neale@lowendale.com.au> December 2000 + * + * Legend: U = unusable with APM patches + * P = partially usable with APM patches + */ + +/* + * Define as 1 to make the driver always call the APM BIOS busy + * routine even if the clock was not reported as slowed by the + * idle routine. Otherwise, define as 0. + */ +#define ALWAYS_CALL_BUSY 1 + +/* + * Define to make the APM BIOS calls zero all data segment registers (so + * that an incorrect BIOS implementation will cause a kernel panic if it + * tries to write to arbitrary memory). + */ +#define APM_ZERO_SEGS + +#include "apm.h" + +/* + * Define to make all _set_limit calls use 64k limits. The APM 1.1 BIOS is + * supposed to provide limit information that it recognizes. Many machines + * do this correctly, but many others do not restrict themselves to their + * claimed limit. When this happens, they will cause a segmentation + * violation in the kernel at boot time. Most BIOS's, however, will + * respect a 64k limit, so we use that. If you want to be pedantic and + * hold your BIOS to its claims, then undefine this. + */ +#define APM_RELAX_SEGMENTS + +/* + * Define to re-initialize the interrupt 0 timer to 100 Hz after a suspend. + * This patched by Chad Miller <cmiller@surfsouth.com>, original code by + * David Chen <chen@ctpa04.mit.edu> + */ +#undef INIT_TIMER_AFTER_SUSPEND + +#ifdef INIT_TIMER_AFTER_SUSPEND +#include <linux/timex.h> +#include <asm/io.h> +#include <linux/delay.h> +#endif + +/* + * Need to poll the APM BIOS every second + */ +#define APM_CHECK_TIMEOUT (HZ) + +/* + * Ignore suspend events for this amount of time after a resume + */ +#define DEFAULT_BOUNCE_INTERVAL (3 * HZ) + +/* + * Maximum number of events stored + */ +#define APM_MAX_EVENTS 20 + +/* + * The per-file APM data + */ +struct apm_user { + int magic; + struct apm_user * next; + int suser: 1; + int writer: 1; + int reader: 1; + int suspend_wait: 1; + int suspend_result; + int suspends_pending; + int standbys_pending; + int suspends_read; + int standbys_read; + int event_head; + int event_tail; + apm_event_t events[APM_MAX_EVENTS]; +}; + +/* + * The magic number in apm_user + */ +#define APM_BIOS_MAGIC 0x4101 + +/* + * idle percentage above which bios idle calls are done + */ +#ifdef CONFIG_APM_CPU_IDLE +#define DEFAULT_IDLE_THRESHOLD 95 +#else +#define DEFAULT_IDLE_THRESHOLD 100 +#endif +#define DEFAULT_IDLE_PERIOD (100 / 3) + +/* + * Local variables + */ +static struct { + unsigned long offset; + unsigned short segment; +} apm_bios_entry; +static int clock_slowed; +static int idle_threshold = DEFAULT_IDLE_THRESHOLD; +static int idle_period = DEFAULT_IDLE_PERIOD; +static int set_pm_idle; +static int suspends_pending; +static int standbys_pending; +static int ignore_sys_suspend; +static int ignore_normal_resume; +static int bounce_interval = DEFAULT_BOUNCE_INTERVAL; + +#ifdef CONFIG_APM_RTC_IS_GMT +# define clock_cmos_diff 0 +# define got_clock_diff 1 +#else +static long clock_cmos_diff; +static int got_clock_diff; +#endif +static int debug; +static int smp; +static int apm_disabled = -1; +#ifdef CONFIG_SMP +static int power_off; +#else +static int power_off = 1; +#endif +#ifdef CONFIG_APM_REAL_MODE_POWER_OFF +static int realmode_power_off = 1; +#else +static int realmode_power_off; +#endif +static int exit_kapmd; +static int kapmd_running; +#ifdef CONFIG_APM_ALLOW_INTS +static int allow_ints = 1; +#else +static int allow_ints; +#endif +static int broken_psr; + +static DECLARE_WAIT_QUEUE_HEAD(apm_waitqueue); +static DECLARE_WAIT_QUEUE_HEAD(apm_suspend_waitqueue); +static struct apm_user * user_list; +static DEFINE_SPINLOCK(user_list_lock); +static struct desc_struct bad_bios_desc = { 0, 0x00409200 }; + +static char driver_version[] = "1.16ac"; /* no spaces */ + +/* + * APM event names taken from the APM 1.2 specification. These are + * the message codes that the BIOS uses to tell us about events + */ +static char * apm_event_name[] = { + "system standby", + "system suspend", + "normal resume", + "critical resume", + "low battery", + "power status change", + "update time", + "critical suspend", + "user standby", + "user suspend", + "system standby resume", + "capabilities change" +}; +#define NR_APM_EVENT_NAME \ + (sizeof(apm_event_name) / sizeof(apm_event_name[0])) + +typedef struct lookup_t { + int key; + char * msg; +} lookup_t; + +/* + * The BIOS returns a set of standard error codes in AX when the + * carry flag is set. + */ + +static const lookup_t error_table[] = { +/* N/A { APM_SUCCESS, "Operation succeeded" }, */ + { APM_DISABLED, "Power management disabled" }, + { APM_CONNECTED, "Real mode interface already connected" }, + { APM_NOT_CONNECTED, "Interface not connected" }, + { APM_16_CONNECTED, "16 bit interface already connected" }, +/* N/A { APM_16_UNSUPPORTED, "16 bit interface not supported" }, */ + { APM_32_CONNECTED, "32 bit interface already connected" }, + { APM_32_UNSUPPORTED, "32 bit interface not supported" }, + { APM_BAD_DEVICE, "Unrecognized device ID" }, + { APM_BAD_PARAM, "Parameter out of range" }, + { APM_NOT_ENGAGED, "Interface not engaged" }, + { APM_BAD_FUNCTION, "Function not supported" }, + { APM_RESUME_DISABLED, "Resume timer disabled" }, + { APM_BAD_STATE, "Unable to enter requested state" }, +/* N/A { APM_NO_EVENTS, "No events pending" }, */ + { APM_NO_ERROR, "BIOS did not set a return code" }, + { APM_NOT_PRESENT, "No APM present" } +}; +#define ERROR_COUNT (sizeof(error_table)/sizeof(lookup_t)) + +/** + * apm_error - display an APM error + * @str: information string + * @err: APM BIOS return code + * + * Write a meaningful log entry to the kernel log in the event of + * an APM error. + */ + +static void apm_error(char *str, int err) +{ + int i; + + for (i = 0; i < ERROR_COUNT; i++) + if (error_table[i].key == err) break; + if (i < ERROR_COUNT) + printk(KERN_NOTICE "apm: %s: %s\n", str, error_table[i].msg); + else + printk(KERN_NOTICE "apm: %s: unknown error code %#2.2x\n", + str, err); +} + +/* + * Lock APM functionality to physical CPU 0 + */ + +#ifdef CONFIG_SMP + +static cpumask_t apm_save_cpus(void) +{ + cpumask_t x = current->cpus_allowed; + /* Some bioses don't like being called from CPU != 0 */ + set_cpus_allowed(current, cpumask_of_cpu(0)); + BUG_ON(smp_processor_id() != 0); + return x; +} + +static inline void apm_restore_cpus(cpumask_t mask) +{ + set_cpus_allowed(current, mask); +} + +#else + +/* + * No CPU lockdown needed on a uniprocessor + */ + +#define apm_save_cpus() (current->cpus_allowed) +#define apm_restore_cpus(x) (void)(x) + +#endif + +/* + * These are the actual BIOS calls. Depending on APM_ZERO_SEGS and + * apm_info.allow_ints, we are being really paranoid here! Not only + * are interrupts disabled, but all the segment registers (except SS) + * are saved and zeroed this means that if the BIOS tries to reference + * any data without explicitly loading the segment registers, the kernel + * will fault immediately rather than have some unforeseen circumstances + * for the rest of the kernel. And it will be very obvious! :-) Doing + * this depends on CS referring to the same physical memory as DS so that + * DS can be zeroed before the call. Unfortunately, we can't do anything + * about the stack segment/pointer. Also, we tell the compiler that + * everything could change. + * + * Also, we KNOW that for the non error case of apm_bios_call, there + * is no useful data returned in the low order 8 bits of eax. + */ +#define APM_DO_CLI \ + if (apm_info.allow_ints) \ + local_irq_enable(); \ + else \ + local_irq_disable(); + +#ifdef APM_ZERO_SEGS +# define APM_DECL_SEGS \ + unsigned int saved_fs; unsigned int saved_gs; +# define APM_DO_SAVE_SEGS \ + savesegment(fs, saved_fs); savesegment(gs, saved_gs) +# define APM_DO_RESTORE_SEGS \ + loadsegment(fs, saved_fs); loadsegment(gs, saved_gs) +#else +# define APM_DECL_SEGS +# define APM_DO_SAVE_SEGS +# define APM_DO_RESTORE_SEGS +#endif + +/** + * apm_bios_call - Make an APM BIOS 32bit call + * @func: APM function to execute + * @ebx_in: EBX register for call entry + * @ecx_in: ECX register for call entry + * @eax: EAX register return + * @ebx: EBX register return + * @ecx: ECX register return + * @edx: EDX register return + * @esi: ESI register return + * + * Make an APM call using the 32bit protected mode interface. The + * caller is responsible for knowing if APM BIOS is configured and + * enabled. This call can disable interrupts for a long period of + * time on some laptops. The return value is in AH and the carry + * flag is loaded into AL. If there is an error, then the error + * code is returned in AH (bits 8-15 of eax) and this function + * returns non-zero. + */ + +static u8 apm_bios_call(u32 func, u32 ebx_in, u32 ecx_in, + u32 *eax, u32 *ebx, u32 *ecx, u32 *edx, u32 *esi) +{ + APM_DECL_SEGS + unsigned long flags; + cpumask_t cpus; + int cpu; + struct desc_struct save_desc_40; + + cpus = apm_save_cpus(); + + cpu = get_cpu(); + save_desc_40 = per_cpu(cpu_gdt_table, cpu)[0x40 / 8]; + per_cpu(cpu_gdt_table, cpu)[0x40 / 8] = bad_bios_desc; + + local_save_flags(flags); + APM_DO_CLI; + APM_DO_SAVE_SEGS; + apm_bios_call_asm(func, ebx_in, ecx_in, eax, ebx, ecx, edx, esi); + APM_DO_RESTORE_SEGS; + local_irq_restore(flags); + per_cpu(cpu_gdt_table, cpu)[0x40 / 8] = save_desc_40; + put_cpu(); + apm_restore_cpus(cpus); + + return *eax & 0xff; +} + +/** + * apm_bios_call_simple - make a simple APM BIOS 32bit call + * @func: APM function to invoke + * @ebx_in: EBX register value for BIOS call + * @ecx_in: ECX register value for BIOS call + * @eax: EAX register on return from the BIOS call + * + * Make a BIOS call that does only returns one value, or just status. + * If there is an error, then the error code is returned in AH + * (bits 8-15 of eax) and this function returns non-zero. This is + * used for simpler BIOS operations. This call may hold interrupts + * off for a long time on some laptops. + */ + +static u8 apm_bios_call_simple(u32 func, u32 ebx_in, u32 ecx_in, u32 *eax) +{ + u8 error; + APM_DECL_SEGS + unsigned long flags; + cpumask_t cpus; + int cpu; + struct desc_struct save_desc_40; + + + cpus = apm_save_cpus(); + + cpu = get_cpu(); + save_desc_40 = per_cpu(cpu_gdt_table, cpu)[0x40 / 8]; + per_cpu(cpu_gdt_table, cpu)[0x40 / 8] = bad_bios_desc; + + local_save_flags(flags); + APM_DO_CLI; + APM_DO_SAVE_SEGS; + error = apm_bios_call_simple_asm(func, ebx_in, ecx_in, eax); + APM_DO_RESTORE_SEGS; + local_irq_restore(flags); + __get_cpu_var(cpu_gdt_table)[0x40 / 8] = save_desc_40; + put_cpu(); + apm_restore_cpus(cpus); + return error; +} + +/** + * apm_driver_version - APM driver version + * @val: loaded with the APM version on return + * + * Retrieve the APM version supported by the BIOS. This is only + * supported for APM 1.1 or higher. An error indicates APM 1.0 is + * probably present. + * + * On entry val should point to a value indicating the APM driver + * version with the high byte being the major and the low byte the + * minor number both in BCD + * + * On return it will hold the BIOS revision supported in the + * same format. + */ + +static int apm_driver_version(u_short *val) +{ + u32 eax; + + if (apm_bios_call_simple(APM_FUNC_VERSION, 0, *val, &eax)) + return (eax >> 8) & 0xff; + *val = eax; + return APM_SUCCESS; +} + +/** + * apm_get_event - get an APM event from the BIOS + * @event: pointer to the event + * @info: point to the event information + * + * The APM BIOS provides a polled information for event + * reporting. The BIOS expects to be polled at least every second + * when events are pending. When a message is found the caller should + * poll until no more messages are present. However, this causes + * problems on some laptops where a suspend event notification is + * not cleared until it is acknowledged. + * + * Additional information is returned in the info pointer, providing + * that APM 1.2 is in use. If no messges are pending the value 0x80 + * is returned (No power management events pending). + */ + +static int apm_get_event(apm_event_t *event, apm_eventinfo_t *info) +{ + u32 eax; + u32 ebx; + u32 ecx; + u32 dummy; + + if (apm_bios_call(APM_FUNC_GET_EVENT, 0, 0, &eax, &ebx, &ecx, + &dummy, &dummy)) + return (eax >> 8) & 0xff; + *event = ebx; + if (apm_info.connection_version < 0x0102) + *info = ~0; /* indicate info not valid */ + else + *info = ecx; + return APM_SUCCESS; +} + +/** + * set_power_state - set the power management state + * @what: which items to transition + * @state: state to transition to + * + * Request an APM change of state for one or more system devices. The + * processor state must be transitioned last of all. what holds the + * class of device in the upper byte and the device number (0xFF for + * all) for the object to be transitioned. + * + * The state holds the state to transition to, which may in fact + * be an acceptance of a BIOS requested state change. + */ + +static int set_power_state(u_short what, u_short state) +{ + u32 eax; + + if (apm_bios_call_simple(APM_FUNC_SET_STATE, what, state, &eax)) + return (eax >> 8) & 0xff; + return APM_SUCCESS; +} + +/** + * set_system_power_state - set system wide power state + * @state: which state to enter + * + * Transition the entire system into a new APM power state. + */ + +static int set_system_power_state(u_short state) +{ + return set_power_state(APM_DEVICE_ALL, state); +} + +/** + * apm_do_idle - perform power saving + * + * This function notifies the BIOS that the processor is (in the view + * of the OS) idle. It returns -1 in the event that the BIOS refuses + * to handle the idle request. On a success the function returns 1 + * if the BIOS did clock slowing or 0 otherwise. + */ + +static int apm_do_idle(void) +{ + u32 eax; + + if (apm_bios_call_simple(APM_FUNC_IDLE, 0, 0, &eax)) { + static unsigned long t; + + /* This always fails on some SMP boards running UP kernels. + * Only report the failure the first 5 times. + */ + if (++t < 5) + { + printk(KERN_DEBUG "apm_do_idle failed (%d)\n", + (eax >> 8) & 0xff); + t = jiffies; + } + return -1; + } + clock_slowed = (apm_info.bios.flags & APM_IDLE_SLOWS_CLOCK) != 0; + return clock_slowed; +} + +/** + * apm_do_busy - inform the BIOS the CPU is busy + * + * Request that the BIOS brings the CPU back to full performance. + */ + +static void apm_do_busy(void) +{ + u32 dummy; + + if (clock_slowed || ALWAYS_CALL_BUSY) { + (void) apm_bios_call_simple(APM_FUNC_BUSY, 0, 0, &dummy); + clock_slowed = 0; + } +} + +/* + * If no process has really been interested in + * the CPU for some time, we want to call BIOS + * power management - we probably want + * to conserve power. + */ +#define IDLE_CALC_LIMIT (HZ * 100) +#define IDLE_LEAKY_MAX 16 + +static void (*original_pm_idle)(void); + +extern void default_idle(void); + +/** + * apm_cpu_idle - cpu idling for APM capable Linux + * + * This is the idling function the kernel executes when APM is available. It + * tries to do BIOS powermanagement based on the average system idle time. + * Furthermore it calls the system default idle routine. + */ + +static void apm_cpu_idle(void) +{ + static int use_apm_idle; /* = 0 */ + static unsigned int last_jiffies; /* = 0 */ + static unsigned int last_stime; /* = 0 */ + + int apm_idle_done = 0; + unsigned int jiffies_since_last_check = jiffies - last_jiffies; + unsigned int bucket; + +recalc: + if (jiffies_since_last_check > IDLE_CALC_LIMIT) { + use_apm_idle = 0; + last_jiffies = jiffies; + last_stime = current->stime; + } else if (jiffies_since_last_check > idle_period) { + unsigned int idle_percentage; + + idle_percentage = current->stime - last_stime; + idle_percentage *= 100; + idle_percentage /= jiffies_since_last_check; + use_apm_idle = (idle_percentage > idle_threshold); + if (apm_info.forbid_idle) + use_apm_idle = 0; + last_jiffies = jiffies; + last_stime = current->stime; + } + + bucket = IDLE_LEAKY_MAX; + + while (!need_resched()) { + if (use_apm_idle) { + unsigned int t; + + t = jiffies; + switch (apm_do_idle()) { + case 0: apm_idle_done = 1; + if (t != jiffies) { + if (bucket) { + bucket = IDLE_LEAKY_MAX; + continue; + } + } else if (bucket) { + bucket--; + continue; + } + break; + case 1: apm_idle_done = 1; + break; + default: /* BIOS refused */ + break; + } + } + if (original_pm_idle) + original_pm_idle(); + else + default_idle(); + jiffies_since_last_check = jiffies - last_jiffies; + if (jiffies_since_last_check > idle_period) + goto recalc; + } + + if (apm_idle_done) + apm_do_busy(); +} + +/** + * apm_power_off - ask the BIOS to power off + * + * Handle the power off sequence. This is the one piece of code we + * will execute even on SMP machines. In order to deal with BIOS + * bugs we support real mode APM BIOS power off calls. We also make + * the SMP call on CPU0 as some systems will only honour this call + * on their first cpu. + */ + +static void apm_power_off(void) +{ + unsigned char po_bios_call[] = { + 0xb8, 0x00, 0x10, /* movw $0x1000,ax */ + 0x8e, 0xd0, /* movw ax,ss */ + 0xbc, 0x00, 0xf0, /* movw $0xf000,sp */ + 0xb8, 0x07, 0x53, /* movw $0x5307,ax */ + 0xbb, 0x01, 0x00, /* movw $0x0001,bx */ + 0xb9, 0x03, 0x00, /* movw $0x0003,cx */ + 0xcd, 0x15 /* int $0x15 */ + }; + + /* + * This may be called on an SMP machine. + */ +#ifdef CONFIG_SMP + /* Some bioses don't like being called from CPU != 0 */ + set_cpus_allowed(current, cpumask_of_cpu(0)); + BUG_ON(smp_processor_id() != 0); +#endif + if (apm_info.realmode_power_off) + { + (void)apm_save_cpus(); + machine_real_restart(po_bios_call, sizeof(po_bios_call)); + } + else + (void) set_system_power_state(APM_STATE_OFF); +} + +#ifdef CONFIG_APM_DO_ENABLE + +/** + * apm_enable_power_management - enable BIOS APM power management + * @enable: enable yes/no + * + * Enable or disable the APM BIOS power services. + */ + +static int apm_enable_power_management(int enable) +{ + u32 eax; + + if ((enable == 0) && (apm_info.bios.flags & APM_BIOS_DISENGAGED)) + return APM_NOT_ENGAGED; + if (apm_bios_call_simple(APM_FUNC_ENABLE_PM, APM_DEVICE_BALL, + enable, &eax)) + return (eax >> 8) & 0xff; + if (enable) + apm_info.bios.flags &= ~APM_BIOS_DISABLED; + else + apm_info.bios.flags |= APM_BIOS_DISABLED; + return APM_SUCCESS; +} +#endif + +/** + * apm_get_power_status - get current power state + * @status: returned status + * @bat: battery info + * @life: estimated life + * + * Obtain the current power status from the APM BIOS. We return a + * status which gives the rough battery status, and current power + * source. The bat value returned give an estimate as a percentage + * of life and a status value for the battery. The estimated life + * if reported is a lifetime in secodnds/minutes at current powwer + * consumption. + */ + +static int apm_get_power_status(u_short *status, u_short *bat, u_short *life) +{ + u32 eax; + u32 ebx; + u32 ecx; + u32 edx; + u32 dummy; + + if (apm_info.get_power_status_broken) + return APM_32_UNSUPPORTED; + if (apm_bios_call(APM_FUNC_GET_STATUS, APM_DEVICE_ALL, 0, + &eax, &ebx, &ecx, &edx, &dummy)) + return (eax >> 8) & 0xff; + *status = ebx; + *bat = ecx; + if (apm_info.get_power_status_swabinminutes) { + *life = swab16((u16)edx); + *life |= 0x8000; + } else + *life = edx; + return APM_SUCCESS; +} + +#if 0 +static int apm_get_battery_status(u_short which, u_short *status, + u_short *bat, u_short *life, u_short *nbat) +{ + u32 eax; + u32 ebx; + u32 ecx; + u32 edx; + u32 esi; + + if (apm_info.connection_version < 0x0102) { + /* pretend we only have one battery. */ + if (which != 1) + return APM_BAD_DEVICE; + *nbat = 1; + return apm_get_power_status(status, bat, life); + } + + if (apm_bios_call(APM_FUNC_GET_STATUS, (0x8000 | (which)), 0, &eax, + &ebx, &ecx, &edx, &esi)) + return (eax >> 8) & 0xff; + *status = ebx; + *bat = ecx; + *life = edx; + *nbat = esi; + return APM_SUCCESS; +} +#endif + +/** + * apm_engage_power_management - enable PM on a device + * @device: identity of device + * @enable: on/off + * + * Activate or deactive power management on either a specific device + * or the entire system (%APM_DEVICE_ALL). + */ + +static int apm_engage_power_management(u_short device, int enable) +{ + u32 eax; + + if ((enable == 0) && (device == APM_DEVICE_ALL) + && (apm_info.bios.flags & APM_BIOS_DISABLED)) + return APM_DISABLED; + if (apm_bios_call_simple(APM_FUNC_ENGAGE_PM, device, enable, &eax)) + return (eax >> 8) & 0xff; + if (device == APM_DEVICE_ALL) { + if (enable) + apm_info.bios.flags &= ~APM_BIOS_DISENGAGED; + else + apm_info.bios.flags |= APM_BIOS_DISENGAGED; + } + return APM_SUCCESS; +} + +#if defined(CONFIG_APM_DISPLAY_BLANK) && defined(CONFIG_VT) + +/** + * apm_console_blank - blank the display + * @blank: on/off + * + * Attempt to blank the console, firstly by blanking just video device + * zero, and if that fails (some BIOSes don't support it) then it blanks + * all video devices. Typically the BIOS will do laptop backlight and + * monitor powerdown for us. + */ + +static int apm_console_blank(int blank) +{ + int error; + u_short state; + + state = blank ? APM_STATE_STANDBY : APM_STATE_READY; + /* Blank the first display device */ + error = set_power_state(0x100, state); + if ((error != APM_SUCCESS) && (error != APM_NO_ERROR)) { + /* try to blank them all instead */ + error = set_power_state(0x1ff, state); + if ((error != APM_SUCCESS) && (error != APM_NO_ERROR)) + /* try to blank device one instead */ + error = set_power_state(0x101, state); + } + if ((error == APM_SUCCESS) || (error == APM_NO_ERROR)) + return 1; + if (error == APM_NOT_ENGAGED) { + static int tried; + int eng_error; + if (tried++ == 0) { + eng_error = apm_engage_power_management(APM_DEVICE_ALL, 1); + if (eng_error) { + apm_error("set display", error); + apm_error("engage interface", eng_error); + return 0; + } else + return apm_console_blank(blank); + } + } + apm_error("set display", error); + return 0; +} +#endif + +static int queue_empty(struct apm_user *as) +{ + return as->event_head == as->event_tail; +} + +static apm_event_t get_queued_event(struct apm_user *as) +{ + as->event_tail = (as->event_tail + 1) % APM_MAX_EVENTS; + return as->events[as->event_tail]; +} + +static void queue_event(apm_event_t event, struct apm_user *sender) +{ + struct apm_user * as; + + spin_lock(&user_list_lock); + if (user_list == NULL) + goto out; + for (as = user_list; as != NULL; as = as->next) { + if ((as == sender) || (!as->reader)) + continue; + as->event_head = (as->event_head + 1) % APM_MAX_EVENTS; + if (as->event_head == as->event_tail) { + static int notified; + + if (notified++ == 0) + printk(KERN_ERR "apm: an event queue overflowed\n"); + as->event_tail = (as->event_tail + 1) % APM_MAX_EVENTS; + } + as->events[as->event_head] = event; + if ((!as->suser) || (!as->writer)) + continue; + switch (event) { + case APM_SYS_SUSPEND: + case APM_USER_SUSPEND: + as->suspends_pending++; + suspends_pending++; + break; + + case APM_SYS_STANDBY: + case APM_USER_STANDBY: + as->standbys_pending++; + standbys_pending++; + break; + } + } + wake_up_interruptible(&apm_waitqueue); +out: + spin_unlock(&user_list_lock); +} + +static void set_time(void) +{ + if (got_clock_diff) { /* Must know time zone in order to set clock */ + xtime.tv_sec = get_cmos_time() + clock_cmos_diff; + xtime.tv_nsec = 0; + } +} + +static void get_time_diff(void) +{ +#ifndef CONFIG_APM_RTC_IS_GMT + /* + * Estimate time zone so that set_time can update the clock + */ + clock_cmos_diff = -get_cmos_time(); + clock_cmos_diff += get_seconds(); + got_clock_diff = 1; +#endif +} + +static void reinit_timer(void) +{ +#ifdef INIT_TIMER_AFTER_SUSPEND + unsigned long flags; + extern spinlock_t i8253_lock; + + spin_lock_irqsave(&i8253_lock, flags); + /* set the clock to 100 Hz */ + outb_p(0x34, PIT_MODE); /* binary, mode 2, LSB/MSB, ch 0 */ + udelay(10); + outb_p(LATCH & 0xff, PIT_CH0); /* LSB */ + udelay(10); + outb(LATCH >> 8, PIT_CH0); /* MSB */ + udelay(10); + spin_unlock_irqrestore(&i8253_lock, flags); +#endif +} + +static int suspend(int vetoable) +{ + int err; + struct apm_user *as; + + if (pm_send_all(PM_SUSPEND, (void *)3)) { + /* Vetoed */ + if (vetoable) { + if (apm_info.connection_version > 0x100) + set_system_power_state(APM_STATE_REJECT); + err = -EBUSY; + ignore_sys_suspend = 0; + printk(KERN_WARNING "apm: suspend was vetoed.\n"); + goto out; + } + printk(KERN_CRIT "apm: suspend was vetoed, but suspending anyway.\n"); + } + + device_suspend(PMSG_SUSPEND); + local_irq_disable(); + device_power_down(PMSG_SUSPEND); + + /* serialize with the timer interrupt */ + write_seqlock(&xtime_lock); + + /* protect against access to timer chip registers */ + spin_lock(&i8253_lock); + + get_time_diff(); + /* + * Irq spinlock must be dropped around set_system_power_state. + * We'll undo any timer changes due to interrupts below. + */ + spin_unlock(&i8253_lock); + write_sequnlock(&xtime_lock); + local_irq_enable(); + + save_processor_state(); + err = set_system_power_state(APM_STATE_SUSPEND); + restore_processor_state(); + + local_irq_disable(); + write_seqlock(&xtime_lock); + spin_lock(&i8253_lock); + reinit_timer(); + set_time(); + ignore_normal_resume = 1; + + spin_unlock(&i8253_lock); + write_sequnlock(&xtime_lock); + + if (err == APM_NO_ERROR) + err = APM_SUCCESS; + if (err != APM_SUCCESS) + apm_error("suspend", err); + err = (err == APM_SUCCESS) ? 0 : -EIO; + device_power_up(); + local_irq_enable(); + device_resume(); + pm_send_all(PM_RESUME, (void *)0); + queue_event(APM_NORMAL_RESUME, NULL); + out: + spin_lock(&user_list_lock); + for (as = user_list; as != NULL; as = as->next) { + as->suspend_wait = 0; + as->suspend_result = err; + } + spin_unlock(&user_list_lock); + wake_up_interruptible(&apm_suspend_waitqueue); + return err; +} + +static void standby(void) +{ + int err; + + local_irq_disable(); + device_power_down(PMSG_SUSPEND); + /* serialize with the timer interrupt */ + write_seqlock(&xtime_lock); + /* If needed, notify drivers here */ + get_time_diff(); + write_sequnlock(&xtime_lock); + local_irq_enable(); + + err = set_system_power_state(APM_STATE_STANDBY); + if ((err != APM_SUCCESS) && (err != APM_NO_ERROR)) + apm_error("standby", err); + + local_irq_disable(); + device_power_up(); + local_irq_enable(); +} + +static apm_event_t get_event(void) +{ + int error; + apm_event_t event; + apm_eventinfo_t info; + + static int notified; + + /* we don't use the eventinfo */ + error = apm_get_event(&event, &info); + if (error == APM_SUCCESS) + return event; + + if ((error != APM_NO_EVENTS) && (notified++ == 0)) + apm_error("get_event", error); + + return 0; +} + +static void check_events(void) +{ + apm_event_t event; + static unsigned long last_resume; + static int ignore_bounce; + + while ((event = get_event()) != 0) { + if (debug) { + if (event <= NR_APM_EVENT_NAME) + printk(KERN_DEBUG "apm: received %s notify\n", + apm_event_name[event - 1]); + else + printk(KERN_DEBUG "apm: received unknown " + "event 0x%02x\n", event); + } + if (ignore_bounce + && ((jiffies - last_resume) > bounce_interval)) + ignore_bounce = 0; + + switch (event) { + case APM_SYS_STANDBY: + case APM_USER_STANDBY: + queue_event(event, NULL); + if (standbys_pending <= 0) + standby(); + break; + + case APM_USER_SUSPEND: +#ifdef CONFIG_APM_IGNORE_USER_SUSPEND + if (apm_info.connection_version > 0x100) + set_system_power_state(APM_STATE_REJECT); + break; +#endif + case APM_SYS_SUSPEND: + if (ignore_bounce) { + if (apm_info.connection_version > 0x100) + set_system_power_state(APM_STATE_REJECT); + break; + } + /* + * If we are already processing a SUSPEND, + * then further SUSPEND events from the BIOS + * will be ignored. We also return here to + * cope with the fact that the Thinkpads keep + * sending a SUSPEND event until something else + * happens! + */ + if (ignore_sys_suspend) + return; + ignore_sys_suspend = 1; + queue_event(event, NULL); + if (suspends_pending <= 0) + (void) suspend(1); + break; + + case APM_NORMAL_RESUME: + case APM_CRITICAL_RESUME: + case APM_STANDBY_RESUME: + ignore_sys_suspend = 0; + last_resume = jiffies; + ignore_bounce = 1; + if ((event != APM_NORMAL_RESUME) + || (ignore_normal_resume == 0)) { + write_seqlock_irq(&xtime_lock); + set_time(); + write_sequnlock_irq(&xtime_lock); + device_resume(); + pm_send_all(PM_RESUME, (void *)0); + queue_event(event, NULL); + } + ignore_normal_resume = 0; + break; + + case APM_CAPABILITY_CHANGE: + case APM_LOW_BATTERY: + case APM_POWER_STATUS_CHANGE: + queue_event(event, NULL); + /* If needed, notify drivers here */ + break; + + case APM_UPDATE_TIME: + write_seqlock_irq(&xtime_lock); + set_time(); + write_sequnlock_irq(&xtime_lock); + break; + + case APM_CRITICAL_SUSPEND: + /* + * We are not allowed to reject a critical suspend. + */ + (void) suspend(0); + break; + } + } +} + +static void apm_event_handler(void) +{ + static int pending_count = 4; + int err; + + if ((standbys_pending > 0) || (suspends_pending > 0)) { + if ((apm_info.connection_version > 0x100) && + (pending_count-- <= 0)) { + pending_count = 4; + if (debug) + printk(KERN_DEBUG "apm: setting state busy\n"); + err = set_system_power_state(APM_STATE_BUSY); + if (err) + apm_error("busy", err); + } + } else + pending_count = 4; + check_events(); +} + +/* + * This is the APM thread main loop. + */ + +static void apm_mainloop(void) +{ + DECLARE_WAITQUEUE(wait, current); + + add_wait_queue(&apm_waitqueue, &wait); + set_current_state(TASK_INTERRUPTIBLE); + for (;;) { + schedule_timeout(APM_CHECK_TIMEOUT); + if (exit_kapmd) + break; + /* + * Ok, check all events, check for idle (and mark us sleeping + * so as not to count towards the load average).. + */ + set_current_state(TASK_INTERRUPTIBLE); + apm_event_handler(); + } + remove_wait_queue(&apm_waitqueue, &wait); +} + +static int check_apm_user(struct apm_user *as, const char *func) +{ + if ((as == NULL) || (as->magic != APM_BIOS_MAGIC)) { + printk(KERN_ERR "apm: %s passed bad filp\n", func); + return 1; + } + return 0; +} + +static ssize_t do_read(struct file *fp, char __user *buf, size_t count, loff_t *ppos) +{ + struct apm_user * as; + int i; + apm_event_t event; + + as = fp->private_data; + if (check_apm_user(as, "read")) + return -EIO; + if ((int)count < sizeof(apm_event_t)) + return -EINVAL; + if ((queue_empty(as)) && (fp->f_flags & O_NONBLOCK)) + return -EAGAIN; + wait_event_interruptible(apm_waitqueue, !queue_empty(as)); + i = count; + while ((i >= sizeof(event)) && !queue_empty(as)) { + event = get_queued_event(as); + if (copy_to_user(buf, &event, sizeof(event))) { + if (i < count) + break; + return -EFAULT; + } + switch (event) { + case APM_SYS_SUSPEND: + case APM_USER_SUSPEND: + as->suspends_read++; + break; + + case APM_SYS_STANDBY: + case APM_USER_STANDBY: + as->standbys_read++; + break; + } + buf += sizeof(event); + i -= sizeof(event); + } + if (i < count) + return count - i; + if (signal_pending(current)) + return -ERESTARTSYS; + return 0; +} + +static unsigned int do_poll(struct file *fp, poll_table * wait) +{ + struct apm_user * as; + + as = fp->private_data; + if (check_apm_user(as, "poll")) + return 0; + poll_wait(fp, &apm_waitqueue, wait); + if (!queue_empty(as)) + return POLLIN | POLLRDNORM; + return 0; +} + +static int do_ioctl(struct inode * inode, struct file *filp, + u_int cmd, u_long arg) +{ + struct apm_user * as; + + as = filp->private_data; + if (check_apm_user(as, "ioctl")) + return -EIO; + if ((!as->suser) || (!as->writer)) + return -EPERM; + switch (cmd) { + case APM_IOC_STANDBY: + if (as->standbys_read > 0) { + as->standbys_read--; + as->standbys_pending--; + standbys_pending--; + } else + queue_event(APM_USER_STANDBY, as); + if (standbys_pending <= 0) + standby(); + break; + case APM_IOC_SUSPEND: + if (as->suspends_read > 0) { + as->suspends_read--; + as->suspends_pending--; + suspends_pending--; + } else + queue_event(APM_USER_SUSPEND, as); + if (suspends_pending <= 0) { + return suspend(1); + } else { + as->suspend_wait = 1; + wait_event_interruptible(apm_suspend_waitqueue, + as->suspend_wait == 0); + return as->suspend_result; + } + break; + default: + return -EINVAL; + } + return 0; +} + +static int do_release(struct inode * inode, struct file * filp) +{ + struct apm_user * as; + + as = filp->private_data; + if (check_apm_user(as, "release")) + return 0; + filp->private_data = NULL; + if (as->standbys_pending > 0) { + standbys_pending -= as->standbys_pending; + if (standbys_pending <= 0) + standby(); + } + if (as->suspends_pending > 0) { + suspends_pending -= as->suspends_pending; + if (suspends_pending <= 0) + (void) suspend(1); + } + spin_lock(&user_list_lock); + if (user_list == as) + user_list = as->next; + else { + struct apm_user * as1; + + for (as1 = user_list; + (as1 != NULL) && (as1->next != as); + as1 = as1->next) + ; + if (as1 == NULL) + printk(KERN_ERR "apm: filp not in user list\n"); + else + as1->next = as->next; + } + spin_unlock(&user_list_lock); + kfree(as); + return 0; +} + +static int do_open(struct inode * inode, struct file * filp) +{ + struct apm_user * as; + + as = (struct apm_user *)kmalloc(sizeof(*as), GFP_KERNEL); + if (as == NULL) { + printk(KERN_ERR "apm: cannot allocate struct of size %d bytes\n", + sizeof(*as)); + return -ENOMEM; + } + as->magic = APM_BIOS_MAGIC; + as->event_tail = as->event_head = 0; + as->suspends_pending = as->standbys_pending = 0; + as->suspends_read = as->standbys_read = 0; + /* + * XXX - this is a tiny bit broken, when we consider BSD + * process accounting. If the device is opened by root, we + * instantly flag that we used superuser privs. Who knows, + * we might close the device immediately without doing a + * privileged operation -- cevans + */ + as->suser = capable(CAP_SYS_ADMIN); + as->writer = (filp->f_mode & FMODE_WRITE) == FMODE_WRITE; + as->reader = (filp->f_mode & FMODE_READ) == FMODE_READ; + spin_lock(&user_list_lock); + as->next = user_list; + user_list = as; + spin_unlock(&user_list_lock); + filp->private_data = as; + return 0; +} + +static int apm_get_info(char *buf, char **start, off_t fpos, int length) +{ + char * p; + unsigned short bx; + unsigned short cx; + unsigned short dx; + int error; + unsigned short ac_line_status = 0xff; + unsigned short battery_status = 0xff; + unsigned short battery_flag = 0xff; + int percentage = -1; + int time_units = -1; + char *units = "?"; + + p = buf; + + if ((num_online_cpus() == 1) && + !(error = apm_get_power_status(&bx, &cx, &dx))) { + ac_line_status = (bx >> 8) & 0xff; + battery_status = bx & 0xff; + if ((cx & 0xff) != 0xff) + percentage = cx & 0xff; + + if (apm_info.connection_version > 0x100) { + battery_flag = (cx >> 8) & 0xff; + if (dx != 0xffff) { + units = (dx & 0x8000) ? "min" : "sec"; + time_units = dx & 0x7fff; + } + } + } + /* Arguments, with symbols from linux/apm_bios.h. Information is + from the Get Power Status (0x0a) call unless otherwise noted. + + 0) Linux driver version (this will change if format changes) + 1) APM BIOS Version. Usually 1.0, 1.1 or 1.2. + 2) APM flags from APM Installation Check (0x00): + bit 0: APM_16_BIT_SUPPORT + bit 1: APM_32_BIT_SUPPORT + bit 2: APM_IDLE_SLOWS_CLOCK + bit 3: APM_BIOS_DISABLED + bit 4: APM_BIOS_DISENGAGED + 3) AC line status + 0x00: Off-line + 0x01: On-line + 0x02: On backup power (BIOS >= 1.1 only) + 0xff: Unknown + 4) Battery status + 0x00: High + 0x01: Low + 0x02: Critical + 0x03: Charging + 0x04: Selected battery not present (BIOS >= 1.2 only) + 0xff: Unknown + 5) Battery flag + bit 0: High + bit 1: Low + bit 2: Critical + bit 3: Charging + bit 7: No system battery + 0xff: Unknown + 6) Remaining battery life (percentage of charge): + 0-100: valid + -1: Unknown + 7) Remaining battery life (time units): + Number of remaining minutes or seconds + -1: Unknown + 8) min = minutes; sec = seconds */ + + p += sprintf(p, "%s %d.%d 0x%02x 0x%02x 0x%02x 0x%02x %d%% %d %s\n", + driver_version, + (apm_info.bios.version >> 8) & 0xff, + apm_info.bios.version & 0xff, + apm_info.bios.flags, + ac_line_status, + battery_status, + battery_flag, + percentage, + time_units, + units); + + return p - buf; +} + +static int apm(void *unused) +{ + unsigned short bx; + unsigned short cx; + unsigned short dx; + int error; + char * power_stat; + char * bat_stat; + + kapmd_running = 1; + + daemonize("kapmd"); + + current->flags |= PF_NOFREEZE; + +#ifdef CONFIG_SMP + /* 2002/08/01 - WT + * This is to avoid random crashes at boot time during initialization + * on SMP systems in case of "apm=power-off" mode. Seen on ASUS A7M266D. + * Some bioses don't like being called from CPU != 0. + * Method suggested by Ingo Molnar. + */ + set_cpus_allowed(current, cpumask_of_cpu(0)); + BUG_ON(smp_processor_id() != 0); +#endif + + if (apm_info.connection_version == 0) { + apm_info.connection_version = apm_info.bios.version; + if (apm_info.connection_version > 0x100) { + /* + * We only support BIOSs up to version 1.2 + */ + if (apm_info.connection_version > 0x0102) + apm_info.connection_version = 0x0102; + error = apm_driver_version(&apm_info.connection_version); + if (error != APM_SUCCESS) { + apm_error("driver version", error); + /* Fall back to an APM 1.0 connection. */ + apm_info.connection_version = 0x100; + } + } + } + + if (debug) + printk(KERN_INFO "apm: Connection version %d.%d\n", + (apm_info.connection_version >> 8) & 0xff, + apm_info.connection_version & 0xff); + +#ifdef CONFIG_APM_DO_ENABLE + if (apm_info.bios.flags & APM_BIOS_DISABLED) { + /* + * This call causes my NEC UltraLite Versa 33/C to hang if it + * is booted with PM disabled but not in the docking station. + * Unfortunate ... + */ + error = apm_enable_power_management(1); + if (error) { + apm_error("enable power management", error); + return -1; + } + } +#endif + + if ((apm_info.bios.flags & APM_BIOS_DISENGAGED) + && (apm_info.connection_version > 0x0100)) { + error = apm_engage_power_management(APM_DEVICE_ALL, 1); + if (error) { + apm_error("engage power management", error); + return -1; + } + } + + if (debug && (num_online_cpus() == 1 || smp )) { + error = apm_get_power_status(&bx, &cx, &dx); + if (error) + printk(KERN_INFO "apm: power status not available\n"); + else { + switch ((bx >> 8) & 0xff) { + case 0: power_stat = "off line"; break; + case 1: power_stat = "on line"; break; + case 2: power_stat = "on backup power"; break; + default: power_stat = "unknown"; break; + } + switch (bx & 0xff) { + case 0: bat_stat = "high"; break; + case 1: bat_stat = "low"; break; + case 2: bat_stat = "critical"; break; + case 3: bat_stat = "charging"; break; + default: bat_stat = "unknown"; break; + } + printk(KERN_INFO + "apm: AC %s, battery status %s, battery life ", + power_stat, bat_stat); + if ((cx & 0xff) == 0xff) + printk("unknown\n"); + else + printk("%d%%\n", cx & 0xff); + if (apm_info.connection_version > 0x100) { + printk(KERN_INFO + "apm: battery flag 0x%02x, battery life ", + (cx >> 8) & 0xff); + if (dx == 0xffff) + printk("unknown\n"); + else + printk("%d %s\n", dx & 0x7fff, + (dx & 0x8000) ? + "minutes" : "seconds"); + } + } + } + + /* Install our power off handler.. */ + if (power_off) + pm_power_off = apm_power_off; + + if (num_online_cpus() == 1 || smp) { +#if defined(CONFIG_APM_DISPLAY_BLANK) && defined(CONFIG_VT) + console_blank_hook = apm_console_blank; +#endif + apm_mainloop(); +#if defined(CONFIG_APM_DISPLAY_BLANK) && defined(CONFIG_VT) + console_blank_hook = NULL; +#endif + } + kapmd_running = 0; + + return 0; +} + +#ifndef MODULE +static int __init apm_setup(char *str) +{ + int invert; + + while ((str != NULL) && (*str != '\0')) { + if (strncmp(str, "off", 3) == 0) + apm_disabled = 1; + if (strncmp(str, "on", 2) == 0) + apm_disabled = 0; + if ((strncmp(str, "bounce-interval=", 16) == 0) || + (strncmp(str, "bounce_interval=", 16) == 0)) + bounce_interval = simple_strtol(str + 16, NULL, 0); + if ((strncmp(str, "idle-threshold=", 15) == 0) || + (strncmp(str, "idle_threshold=", 15) == 0)) + idle_threshold = simple_strtol(str + 15, NULL, 0); + if ((strncmp(str, "idle-period=", 12) == 0) || + (strncmp(str, "idle_period=", 12) == 0)) + idle_period = simple_strtol(str + 12, NULL, 0); + invert = (strncmp(str, "no-", 3) == 0) || + (strncmp(str, "no_", 3) == 0); + if (invert) + str += 3; + if (strncmp(str, "debug", 5) == 0) + debug = !invert; + if ((strncmp(str, "power-off", 9) == 0) || + (strncmp(str, "power_off", 9) == 0)) + power_off = !invert; + if (strncmp(str, "smp", 3) == 0) + { + smp = !invert; + idle_threshold = 100; + } + if ((strncmp(str, "allow-ints", 10) == 0) || + (strncmp(str, "allow_ints", 10) == 0)) + apm_info.allow_ints = !invert; + if ((strncmp(str, "broken-psr", 10) == 0) || + (strncmp(str, "broken_psr", 10) == 0)) + apm_info.get_power_status_broken = !invert; + if ((strncmp(str, "realmode-power-off", 18) == 0) || + (strncmp(str, "realmode_power_off", 18) == 0)) + apm_info.realmode_power_off = !invert; + str = strchr(str, ','); + if (str != NULL) + str += strspn(str, ", \t"); + } + return 1; +} + +__setup("apm=", apm_setup); +#endif + +static struct file_operations apm_bios_fops = { + .owner = THIS_MODULE, + .read = do_read, + .poll = do_poll, + .ioctl = do_ioctl, + .open = do_open, + .release = do_release, +}; + +static struct miscdevice apm_device = { + APM_MINOR_DEV, + "apm_bios", + &apm_bios_fops +}; + + +/* Simple "print if true" callback */ +static int __init print_if_true(struct dmi_system_id *d) +{ + printk("%s\n", d->ident); + return 0; +} + +/* + * Some Bioses enable the PS/2 mouse (touchpad) at resume, even if it was + * disabled before the suspend. Linux used to get terribly confused by that. + */ +static int __init broken_ps2_resume(struct dmi_system_id *d) +{ + printk(KERN_INFO "%s machine detected. Mousepad Resume Bug workaround hopefully not needed.\n", d->ident); + return 0; +} + +/* Some bioses have a broken protected mode poweroff and need to use realmode */ +static int __init set_realmode_power_off(struct dmi_system_id *d) +{ + if (apm_info.realmode_power_off == 0) { + apm_info.realmode_power_off = 1; + printk(KERN_INFO "%s bios detected. Using realmode poweroff only.\n", d->ident); + } + return 0; +} + +/* Some laptops require interrupts to be enabled during APM calls */ +static int __init set_apm_ints(struct dmi_system_id *d) +{ + if (apm_info.allow_ints == 0) { + apm_info.allow_ints = 1; + printk(KERN_INFO "%s machine detected. Enabling interrupts during APM calls.\n", d->ident); + } + return 0; +} + +/* Some APM bioses corrupt memory or just plain do not work */ +static int __init apm_is_horked(struct dmi_system_id *d) +{ + if (apm_info.disabled == 0) { + apm_info.disabled = 1; + printk(KERN_INFO "%s machine detected. Disabling APM.\n", d->ident); + } + return 0; +} + +static int __init apm_is_horked_d850md(struct dmi_system_id *d) +{ + if (apm_info.disabled == 0) { + apm_info.disabled = 1; + printk(KERN_INFO "%s machine detected. Disabling APM.\n", d->ident); + printk(KERN_INFO "This bug is fixed in bios P15 which is available for \n"); + printk(KERN_INFO "download from support.intel.com \n"); + } + return 0; +} + +/* Some APM bioses hang on APM idle calls */ +static int __init apm_likes_to_melt(struct dmi_system_id *d) +{ + if (apm_info.forbid_idle == 0) { + apm_info.forbid_idle = 1; + printk(KERN_INFO "%s machine detected. Disabling APM idle calls.\n", d->ident); + } + return 0; +} + +/* + * Check for clue free BIOS implementations who use + * the following QA technique + * + * [ Write BIOS Code ]<------ + * | ^ + * < Does it Compile >----N-- + * |Y ^ + * < Does it Boot Win98 >-N-- + * |Y + * [Ship It] + * + * Phoenix A04 08/24/2000 is known bad (Dell Inspiron 5000e) + * Phoenix A07 09/29/2000 is known good (Dell Inspiron 5000) + */ +static int __init broken_apm_power(struct dmi_system_id *d) +{ + apm_info.get_power_status_broken = 1; + printk(KERN_WARNING "BIOS strings suggest APM bugs, disabling power status reporting.\n"); + return 0; +} + +/* + * This bios swaps the APM minute reporting bytes over (Many sony laptops + * have this problem). + */ +static int __init swab_apm_power_in_minutes(struct dmi_system_id *d) +{ + apm_info.get_power_status_swabinminutes = 1; + printk(KERN_WARNING "BIOS strings suggest APM reports battery life in minutes and wrong byte order.\n"); + return 0; +} + +static struct dmi_system_id __initdata apm_dmi_table[] = { + { + print_if_true, + KERN_WARNING "IBM T23 - BIOS 1.03b+ and controller firmware 1.02+ may be needed for Linux APM.", + { DMI_MATCH(DMI_SYS_VENDOR, "IBM"), + DMI_MATCH(DMI_BIOS_VERSION, "1AET38WW (1.01b)"), }, + }, + { /* Handle problems with APM on the C600 */ + broken_ps2_resume, "Dell Latitude C600", + { DMI_MATCH(DMI_SYS_VENDOR, "Dell"), + DMI_MATCH(DMI_PRODUCT_NAME, "Latitude C600"), }, + }, + { /* Allow interrupts during suspend on Dell Latitude laptops*/ + set_apm_ints, "Dell Latitude", + { DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "Latitude C510"), } + }, + { /* APM crashes */ + apm_is_horked, "Dell Inspiron 2500", + { DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "Inspiron 2500"), + DMI_MATCH(DMI_BIOS_VENDOR,"Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION,"A11"), }, + }, + { /* Allow interrupts during suspend on Dell Inspiron laptops*/ + set_apm_ints, "Dell Inspiron", { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "Inspiron 4000"), }, + }, + { /* Handle problems with APM on Inspiron 5000e */ + broken_apm_power, "Dell Inspiron 5000e", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "A04"), + DMI_MATCH(DMI_BIOS_DATE, "08/24/2000"), }, + }, + { /* Handle problems with APM on Inspiron 2500 */ + broken_apm_power, "Dell Inspiron 2500", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "A12"), + DMI_MATCH(DMI_BIOS_DATE, "02/04/2002"), }, + }, + { /* APM crashes */ + apm_is_horked, "Dell Dimension 4100", + { DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "XPS-Z"), + DMI_MATCH(DMI_BIOS_VENDOR,"Intel Corp."), + DMI_MATCH(DMI_BIOS_VERSION,"A11"), }, + }, + { /* Allow interrupts during suspend on Compaq Laptops*/ + set_apm_ints, "Compaq 12XL125", + { DMI_MATCH(DMI_SYS_VENDOR, "Compaq"), + DMI_MATCH(DMI_PRODUCT_NAME, "Compaq PC"), + DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION,"4.06"), }, + }, + { /* Allow interrupts during APM or the clock goes slow */ + set_apm_ints, "ASUSTeK", + { DMI_MATCH(DMI_SYS_VENDOR, "ASUSTeK Computer Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "L8400K series Notebook PC"), }, + }, + { /* APM blows on shutdown */ + apm_is_horked, "ABIT KX7-333[R]", + { DMI_MATCH(DMI_BOARD_VENDOR, "ABIT"), + DMI_MATCH(DMI_BOARD_NAME, "VT8367-8233A (KX7-333[R])"), }, + }, + { /* APM crashes */ + apm_is_horked, "Trigem Delhi3", + { DMI_MATCH(DMI_SYS_VENDOR, "TriGem Computer, Inc"), + DMI_MATCH(DMI_PRODUCT_NAME, "Delhi3"), }, + }, + { /* APM crashes */ + apm_is_horked, "Fujitsu-Siemens", + { DMI_MATCH(DMI_BIOS_VENDOR, "hoenix/FUJITSU SIEMENS"), + DMI_MATCH(DMI_BIOS_VERSION, "Version1.01"), }, + }, + { /* APM crashes */ + apm_is_horked_d850md, "Intel D850MD", + { DMI_MATCH(DMI_BIOS_VENDOR, "Intel Corp."), + DMI_MATCH(DMI_BIOS_VERSION, "MV85010A.86A.0016.P07.0201251536"), }, + }, + { /* APM crashes */ + apm_is_horked, "Intel D810EMO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Intel Corp."), + DMI_MATCH(DMI_BIOS_VERSION, "MO81010A.86A.0008.P04.0004170800"), }, + }, + { /* APM crashes */ + apm_is_horked, "Dell XPS-Z", + { DMI_MATCH(DMI_BIOS_VENDOR, "Intel Corp."), + DMI_MATCH(DMI_BIOS_VERSION, "A11"), + DMI_MATCH(DMI_PRODUCT_NAME, "XPS-Z"), }, + }, + { /* APM crashes */ + apm_is_horked, "Sharp PC-PJ/AX", + { DMI_MATCH(DMI_SYS_VENDOR, "SHARP"), + DMI_MATCH(DMI_PRODUCT_NAME, "PC-PJ/AX"), + DMI_MATCH(DMI_BIOS_VENDOR,"SystemSoft"), + DMI_MATCH(DMI_BIOS_VERSION,"Version R2.08"), }, + }, + { /* APM crashes */ + apm_is_horked, "Dell Inspiron 2500", + { DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "Inspiron 2500"), + DMI_MATCH(DMI_BIOS_VENDOR,"Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION,"A11"), }, + }, + { /* APM idle hangs */ + apm_likes_to_melt, "Jabil AMD", + { DMI_MATCH(DMI_BIOS_VENDOR, "American Megatrends Inc."), + DMI_MATCH(DMI_BIOS_VERSION, "0AASNP06"), }, + }, + { /* APM idle hangs */ + apm_likes_to_melt, "AMI Bios", + { DMI_MATCH(DMI_BIOS_VENDOR, "American Megatrends Inc."), + DMI_MATCH(DMI_BIOS_VERSION, "0AASNP05"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-N505X(DE) */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0206H"), + DMI_MATCH(DMI_BIOS_DATE, "08/23/99"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-N505VX */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "W2K06H0"), + DMI_MATCH(DMI_BIOS_DATE, "02/03/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-XG29 */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0117A0"), + DMI_MATCH(DMI_BIOS_DATE, "04/25/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z600NE */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0121Z1"), + DMI_MATCH(DMI_BIOS_DATE, "05/11/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z600NE */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "WME01Z1"), + DMI_MATCH(DMI_BIOS_DATE, "08/11/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z600LEK(DE) */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0206Z3"), + DMI_MATCH(DMI_BIOS_DATE, "12/25/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z505LS */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0203D0"), + DMI_MATCH(DMI_BIOS_DATE, "05/12/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z505LS */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0203Z3"), + DMI_MATCH(DMI_BIOS_DATE, "08/25/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-Z505LS (with updated BIOS) */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0209Z3"), + DMI_MATCH(DMI_BIOS_DATE, "05/12/01"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-F104K */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0204K2"), + DMI_MATCH(DMI_BIOS_DATE, "08/28/00"), }, + }, + + { /* Handle problems with APM on Sony Vaio PCG-C1VN/C1VE */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0208P1"), + DMI_MATCH(DMI_BIOS_DATE, "11/09/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-C1VE */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "R0204P1"), + DMI_MATCH(DMI_BIOS_DATE, "09/12/00"), }, + }, + { /* Handle problems with APM on Sony Vaio PCG-C1VE */ + swab_apm_power_in_minutes, "Sony VAIO", + { DMI_MATCH(DMI_BIOS_VENDOR, "Phoenix Technologies LTD"), + DMI_MATCH(DMI_BIOS_VERSION, "WXPO1Z3"), + DMI_MATCH(DMI_BIOS_DATE, "10/26/01"), }, + }, + { /* broken PM poweroff bios */ + set_realmode_power_off, "Award Software v4.60 PGMA", + { DMI_MATCH(DMI_BIOS_VENDOR, "Award Software International, Inc."), + DMI_MATCH(DMI_BIOS_VERSION, "4.60 PGMA"), + DMI_MATCH(DMI_BIOS_DATE, "134526184"), }, + }, + + /* Generic per vendor APM settings */ + + { /* Allow interrupts during suspend on IBM laptops */ + set_apm_ints, "IBM", + { DMI_MATCH(DMI_SYS_VENDOR, "IBM"), }, + }, + + { } +}; + +/* + * Just start the APM thread. We do NOT want to do APM BIOS + * calls from anything but the APM thread, if for no other reason + * than the fact that we don't trust the APM BIOS. This way, + * most common APM BIOS problems that lead to protection errors + * etc will have at least some level of being contained... + * + * In short, if something bad happens, at least we have a choice + * of just killing the apm thread.. + */ +static int __init apm_init(void) +{ + struct proc_dir_entry *apm_proc; + int ret; + int i; + + dmi_check_system(apm_dmi_table); + + if (apm_info.bios.version == 0) { + printk(KERN_INFO "apm: BIOS not found.\n"); + return -ENODEV; + } + printk(KERN_INFO + "apm: BIOS version %d.%d Flags 0x%02x (Driver version %s)\n", + ((apm_info.bios.version >> 8) & 0xff), + (apm_info.bios.version & 0xff), + apm_info.bios.flags, + driver_version); + if ((apm_info.bios.flags & APM_32_BIT_SUPPORT) == 0) { + printk(KERN_INFO "apm: no 32 bit BIOS support\n"); + return -ENODEV; + } + + if (allow_ints) + apm_info.allow_ints = 1; + if (broken_psr) + apm_info.get_power_status_broken = 1; + if (realmode_power_off) + apm_info.realmode_power_off = 1; + /* User can override, but default is to trust DMI */ + if (apm_disabled != -1) + apm_info.disabled = apm_disabled; + + /* + * Fix for the Compaq Contura 3/25c which reports BIOS version 0.1 + * but is reportedly a 1.0 BIOS. + */ + if (apm_info.bios.version == 0x001) + apm_info.bios.version = 0x100; + + /* BIOS < 1.2 doesn't set cseg_16_len */ + if (apm_info.bios.version < 0x102) + apm_info.bios.cseg_16_len = 0; /* 64k */ + + if (debug) { + printk(KERN_INFO "apm: entry %x:%lx cseg16 %x dseg %x", + apm_info.bios.cseg, apm_info.bios.offset, + apm_info.bios.cseg_16, apm_info.bios.dseg); + if (apm_info.bios.version > 0x100) + printk(" cseg len %x, dseg len %x", + apm_info.bios.cseg_len, + apm_info.bios.dseg_len); + if (apm_info.bios.version > 0x101) + printk(" cseg16 len %x", apm_info.bios.cseg_16_len); + printk("\n"); + } + + if (apm_info.disabled) { + printk(KERN_NOTICE "apm: disabled on user request.\n"); + return -ENODEV; + } + if ((num_online_cpus() > 1) && !power_off && !smp) { + printk(KERN_NOTICE "apm: disabled - APM is not SMP safe.\n"); + apm_info.disabled = 1; + return -ENODEV; + } + if (PM_IS_ACTIVE()) { + printk(KERN_NOTICE "apm: overridden by ACPI.\n"); + apm_info.disabled = 1; + return -ENODEV; + } + pm_active = 1; + + /* + * Set up a segment that references the real mode segment 0x40 + * that extends up to the end of page zero (that we have reserved). + * This is for buggy BIOS's that refer to (real mode) segment 0x40 + * even though they are called in protected mode. + */ + set_base(bad_bios_desc, __va((unsigned long)0x40 << 4)); + _set_limit((char *)&bad_bios_desc, 4095 - (0x40 << 4)); + + apm_bios_entry.offset = apm_info.bios.offset; + apm_bios_entry.segment = APM_CS; + + for (i = 0; i < NR_CPUS; i++) { + set_base(per_cpu(cpu_gdt_table, i)[APM_CS >> 3], + __va((unsigned long)apm_info.bios.cseg << 4)); + set_base(per_cpu(cpu_gdt_table, i)[APM_CS_16 >> 3], + __va((unsigned long)apm_info.bios.cseg_16 << 4)); + set_base(per_cpu(cpu_gdt_table, i)[APM_DS >> 3], + __va((unsigned long)apm_info.bios.dseg << 4)); +#ifndef APM_RELAX_SEGMENTS + if (apm_info.bios.version == 0x100) { +#endif + /* For ASUS motherboard, Award BIOS rev 110 (and others?) */ + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_CS >> 3], 64 * 1024 - 1); + /* For some unknown machine. */ + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_CS_16 >> 3], 64 * 1024 - 1); + /* For the DEC Hinote Ultra CT475 (and others?) */ + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_DS >> 3], 64 * 1024 - 1); +#ifndef APM_RELAX_SEGMENTS + } else { + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_CS >> 3], + (apm_info.bios.cseg_len - 1) & 0xffff); + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_CS_16 >> 3], + (apm_info.bios.cseg_16_len - 1) & 0xffff); + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_DS >> 3], + (apm_info.bios.dseg_len - 1) & 0xffff); + /* workaround for broken BIOSes */ + if (apm_info.bios.cseg_len <= apm_info.bios.offset) + _set_limit((char *)&per_cpu(cpu_gdt_table, i)[APM_CS >> 3], 64 * 1024 -1); + if (apm_info.bios.dseg_len <= 0x40) { /* 0x40 * 4kB == 64kB */ + /* for the BIOS that assumes granularity = 1 */ + per_cpu(cpu_gdt_table, i)[APM_DS >> 3].b |= 0x800000; + printk(KERN_NOTICE "apm: we set the granularity of dseg.\n"); + } + } +#endif + } + + apm_proc = create_proc_info_entry("apm", 0, NULL, apm_get_info); + if (apm_proc) + apm_proc->owner = THIS_MODULE; + + ret = kernel_thread(apm, NULL, CLONE_KERNEL | SIGCHLD); + if (ret < 0) { + printk(KERN_ERR "apm: disabled - Unable to start kernel thread.\n"); + return -ENOMEM; + } + + if (num_online_cpus() > 1 && !smp ) { + printk(KERN_NOTICE + "apm: disabled - APM is not SMP safe (power off active).\n"); + return 0; + } + + misc_register(&apm_device); + + if (HZ != 100) + idle_period = (idle_period * HZ) / 100; + if (idle_threshold < 100) { + original_pm_idle = pm_idle; + pm_idle = apm_cpu_idle; + set_pm_idle = 1; + } + + return 0; +} + +static void __exit apm_exit(void) +{ + int error; + + if (set_pm_idle) { + pm_idle = original_pm_idle; + /* + * We are about to unload the current idle thread pm callback + * (pm_idle), Wait for all processors to update cached/local + * copies of pm_idle before proceeding. + */ + cpu_idle_wait(); + } + if (((apm_info.bios.flags & APM_BIOS_DISENGAGED) == 0) + && (apm_info.connection_version > 0x0100)) { + error = apm_engage_power_management(APM_DEVICE_ALL, 0); + if (error) + apm_error("disengage power management", error); + } + misc_deregister(&apm_device); + remove_proc_entry("apm", NULL); + if (power_off) + pm_power_off = NULL; + exit_kapmd = 1; + while (kapmd_running) + schedule(); + pm_active = 0; +} + +module_init(apm_init); +module_exit(apm_exit); + +MODULE_AUTHOR("Stephen Rothwell"); +MODULE_DESCRIPTION("Advanced Power Management"); +MODULE_LICENSE("GPL"); +module_param(debug, bool, 0644); +MODULE_PARM_DESC(debug, "Enable debug mode"); +module_param(power_off, bool, 0444); +MODULE_PARM_DESC(power_off, "Enable power off"); +module_param(bounce_interval, int, 0444); +MODULE_PARM_DESC(bounce_interval, + "Set the number of ticks to ignore suspend bounces"); +module_param(allow_ints, bool, 0444); +MODULE_PARM_DESC(allow_ints, "Allow interrupts during BIOS calls"); +module_param(broken_psr, bool, 0444); +MODULE_PARM_DESC(broken_psr, "BIOS has a broken GetPowerStatus call"); +module_param(realmode_power_off, bool, 0444); +MODULE_PARM_DESC(realmode_power_off, + "Switch to real mode before powering off"); +module_param(idle_threshold, int, 0444); +MODULE_PARM_DESC(idle_threshold, + "System idle percentage above which to make APM BIOS idle calls"); +module_param(idle_period, int, 0444); +MODULE_PARM_DESC(idle_period, + "Period (in sec/100) over which to caculate the idle percentage"); +module_param(smp, bool, 0444); +MODULE_PARM_DESC(smp, + "Set this to enable APM use on an SMP platform. Use with caution on older systems"); +MODULE_ALIAS_MISCDEV(APM_MINOR_DEV); diff --git a/arch/i386/kernel/asm-offsets.c b/arch/i386/kernel/asm-offsets.c new file mode 100644 index 00000000000..36d66e2077d --- /dev/null +++ b/arch/i386/kernel/asm-offsets.c @@ -0,0 +1,72 @@ +/* + * Generate definitions needed by assembly language modules. + * This code generates raw asm output which is post-processed + * to extract and format the required data. + */ + +#include <linux/sched.h> +#include <linux/signal.h> +#include <linux/personality.h> +#include <linux/suspend.h> +#include <asm/ucontext.h> +#include "sigframe.h" +#include <asm/fixmap.h> +#include <asm/processor.h> +#include <asm/thread_info.h> + +#define DEFINE(sym, val) \ + asm volatile("\n->" #sym " %0 " #val : : "i" (val)) + +#define BLANK() asm volatile("\n->" : : ) + +#define OFFSET(sym, str, mem) \ + DEFINE(sym, offsetof(struct str, mem)); + +void foo(void) +{ + OFFSET(SIGCONTEXT_eax, sigcontext, eax); + OFFSET(SIGCONTEXT_ebx, sigcontext, ebx); + OFFSET(SIGCONTEXT_ecx, sigcontext, ecx); + OFFSET(SIGCONTEXT_edx, sigcontext, edx); + OFFSET(SIGCONTEXT_esi, sigcontext, esi); + OFFSET(SIGCONTEXT_edi, sigcontext, edi); + OFFSET(SIGCONTEXT_ebp, sigcontext, ebp); + OFFSET(SIGCONTEXT_esp, sigcontext, esp); + OFFSET(SIGCONTEXT_eip, sigcontext, eip); + BLANK(); + + OFFSET(CPUINFO_x86, cpuinfo_x86, x86); + OFFSET(CPUINFO_x86_vendor, cpuinfo_x86, x86_vendor); + OFFSET(CPUINFO_x86_model, cpuinfo_x86, x86_model); + OFFSET(CPUINFO_x86_mask, cpuinfo_x86, x86_mask); + OFFSET(CPUINFO_hard_math, cpuinfo_x86, hard_math); + OFFSET(CPUINFO_cpuid_level, cpuinfo_x86, cpuid_level); + OFFSET(CPUINFO_x86_capability, cpuinfo_x86, x86_capability); + OFFSET(CPUINFO_x86_vendor_id, cpuinfo_x86, x86_vendor_id); + BLANK(); + + OFFSET(TI_task, thread_info, task); + OFFSET(TI_exec_domain, thread_info, exec_domain); + OFFSET(TI_flags, thread_info, flags); + OFFSET(TI_status, thread_info, status); + OFFSET(TI_cpu, thread_info, cpu); + OFFSET(TI_preempt_count, thread_info, preempt_count); + OFFSET(TI_addr_limit, thread_info, addr_limit); + OFFSET(TI_restart_block, thread_info, restart_block); + BLANK(); + + OFFSET(EXEC_DOMAIN_handler, exec_domain, handler); + OFFSET(RT_SIGFRAME_sigcontext, rt_sigframe, uc.uc_mcontext); + BLANK(); + + OFFSET(pbe_address, pbe, address); + OFFSET(pbe_orig_address, pbe, orig_address); + OFFSET(pbe_next, pbe, next); + + /* Offset from the sysenter stack to tss.esp0 */ + DEFINE(TSS_sysenter_esp0, offsetof(struct tss_struct, esp0) - + sizeof(struct tss_struct)); + + DEFINE(PAGE_SIZE_asm, PAGE_SIZE); + DEFINE(VSYSCALL_BASE, __fix_to_virt(FIX_VSYSCALL)); +} diff --git a/arch/i386/kernel/bootflag.c b/arch/i386/kernel/bootflag.c new file mode 100644 index 00000000000..4c30ed01f4e --- /dev/null +++ b/arch/i386/kernel/bootflag.c @@ -0,0 +1,99 @@ +/* + * Implement 'Simple Boot Flag Specification 2.0' + */ + + +#include <linux/config.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/string.h> +#include <linux/slab.h> +#include <linux/spinlock.h> +#include <linux/acpi.h> +#include <asm/io.h> + +#include <linux/mc146818rtc.h> + + +#define SBF_RESERVED (0x78) +#define SBF_PNPOS (1<<0) +#define SBF_BOOTING (1<<1) +#define SBF_DIAG (1<<2) +#define SBF_PARITY (1<<7) + + +int sbf_port __initdata = -1; /* set via acpi_boot_init() */ + + +static int __init parity(u8 v) +{ + int x = 0; + int i; + + for(i=0;i<8;i++) + { + x^=(v&1); + v>>=1; + } + return x; +} + +static void __init sbf_write(u8 v) +{ + unsigned long flags; + if(sbf_port != -1) + { + v &= ~SBF_PARITY; + if(!parity(v)) + v|=SBF_PARITY; + + printk(KERN_INFO "Simple Boot Flag at 0x%x set to 0x%x\n", sbf_port, v); + + spin_lock_irqsave(&rtc_lock, flags); + CMOS_WRITE(v, sbf_port); + spin_unlock_irqrestore(&rtc_lock, flags); + } +} + +static u8 __init sbf_read(void) +{ + u8 v; + unsigned long flags; + if(sbf_port == -1) + return 0; + spin_lock_irqsave(&rtc_lock, flags); + v = CMOS_READ(sbf_port); + spin_unlock_irqrestore(&rtc_lock, flags); + return v; +} + +static int __init sbf_value_valid(u8 v) +{ + if(v&SBF_RESERVED) /* Reserved bits */ + return 0; + if(!parity(v)) + return 0; + return 1; +} + +static int __init sbf_init(void) +{ + u8 v; + if(sbf_port == -1) + return 0; + v = sbf_read(); + if(!sbf_value_valid(v)) + printk(KERN_WARNING "Simple Boot Flag value 0x%x read from CMOS RAM was invalid\n",v); + + v &= ~SBF_RESERVED; + v &= ~SBF_BOOTING; + v &= ~SBF_DIAG; +#if defined(CONFIG_ISAPNP) + v |= SBF_PNPOS; +#endif + sbf_write(v); + return 0; +} + +module_init(sbf_init); diff --git a/arch/i386/kernel/cpu/Makefile b/arch/i386/kernel/cpu/Makefile new file mode 100644 index 00000000000..010aecfffbc --- /dev/null +++ b/arch/i386/kernel/cpu/Makefile @@ -0,0 +1,19 @@ +# +# Makefile for x86-compatible CPU details and quirks +# + +obj-y := common.o proc.o + +obj-y += amd.o +obj-y += cyrix.o +obj-y += centaur.o +obj-y += transmeta.o +obj-y += intel.o intel_cacheinfo.o +obj-y += rise.o +obj-y += nexgen.o +obj-y += umc.o + +obj-$(CONFIG_X86_MCE) += mcheck/ + +obj-$(CONFIG_MTRR) += mtrr/ +obj-$(CONFIG_CPU_FREQ) += cpufreq/ diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c new file mode 100644 index 00000000000..ae94585d044 --- /dev/null +++ b/arch/i386/kernel/cpu/amd.c @@ -0,0 +1,249 @@ +#include <linux/init.h> +#include <linux/bitops.h> +#include <linux/mm.h> +#include <asm/io.h> +#include <asm/processor.h> + +#include "cpu.h" + +/* + * B step AMD K6 before B 9730xxxx have hardware bugs that can cause + * misexecution of code under Linux. Owners of such processors should + * contact AMD for precise details and a CPU swap. + * + * See http://www.multimania.com/poulot/k6bug.html + * http://www.amd.com/K6/k6docs/revgd.html + * + * The following test is erm.. interesting. AMD neglected to up + * the chip setting when fixing the bug but they also tweaked some + * performance at the same time.. + */ + +extern void vide(void); +__asm__(".align 4\nvide: ret"); + +static void __init init_amd(struct cpuinfo_x86 *c) +{ + u32 l, h; + int mbytes = num_physpages >> (20-PAGE_SHIFT); + int r; + + /* + * FIXME: We should handle the K5 here. Set up the write + * range and also turn on MSR 83 bits 4 and 31 (write alloc, + * no bus pipeline) + */ + + /* Bit 31 in normal CPUID used for nonstandard 3DNow ID; + 3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */ + clear_bit(0*32+31, c->x86_capability); + + r = get_model_name(c); + + switch(c->x86) + { + case 4: + /* + * General Systems BIOSen alias the cpu frequency registers + * of the Elan at 0x000df000. Unfortuantly, one of the Linux + * drivers subsequently pokes it, and changes the CPU speed. + * Workaround : Remove the unneeded alias. + */ +#define CBAR (0xfffc) /* Configuration Base Address (32-bit) */ +#define CBAR_ENB (0x80000000) +#define CBAR_KEY (0X000000CB) + if (c->x86_model==9 || c->x86_model == 10) { + if (inl (CBAR) & CBAR_ENB) + outl (0 | CBAR_KEY, CBAR); + } + break; + case 5: + if( c->x86_model < 6 ) + { + /* Based on AMD doc 20734R - June 2000 */ + if ( c->x86_model == 0 ) { + clear_bit(X86_FEATURE_APIC, c->x86_capability); + set_bit(X86_FEATURE_PGE, c->x86_capability); + } + break; + } + + if ( c->x86_model == 6 && c->x86_mask == 1 ) { + const int K6_BUG_LOOP = 1000000; + int n; + void (*f_vide)(void); + unsigned long d, d2; + + printk(KERN_INFO "AMD K6 stepping B detected - "); + + /* + * It looks like AMD fixed the 2.6.2 bug and improved indirect + * calls at the same time. + */ + + n = K6_BUG_LOOP; + f_vide = vide; + rdtscl(d); + while (n--) + f_vide(); + rdtscl(d2); + d = d2-d; + + /* Knock these two lines out if it debugs out ok */ + printk(KERN_INFO "AMD K6 stepping B detected - "); + /* -- cut here -- */ + if (d > 20*K6_BUG_LOOP) + printk("system stability may be impaired when more than 32 MB are used.\n"); + else + printk("probably OK (after B9730xxxx).\n"); + printk(KERN_INFO "Please see http://membres.lycos.fr/poulot/k6bug.html\n"); + } + + /* K6 with old style WHCR */ + if (c->x86_model < 8 || + (c->x86_model== 8 && c->x86_mask < 8)) { + /* We can only write allocate on the low 508Mb */ + if(mbytes>508) + mbytes=508; + + rdmsr(MSR_K6_WHCR, l, h); + if ((l&0x0000FFFF)==0) { + unsigned long flags; + l=(1<<0)|((mbytes/4)<<1); + local_irq_save(flags); + wbinvd(); + wrmsr(MSR_K6_WHCR, l, h); + local_irq_restore(flags); + printk(KERN_INFO "Enabling old style K6 write allocation for %d Mb\n", + mbytes); + } + break; + } + + if ((c->x86_model == 8 && c->x86_mask >7) || + c->x86_model == 9 || c->x86_model == 13) { + /* The more serious chips .. */ + + if(mbytes>4092) + mbytes=4092; + + rdmsr(MSR_K6_WHCR, l, h); + if ((l&0xFFFF0000)==0) { + unsigned long flags; + l=((mbytes>>2)<<22)|(1<<16); + local_irq_save(flags); + wbinvd(); + wrmsr(MSR_K6_WHCR, l, h); + local_irq_restore(flags); + printk(KERN_INFO "Enabling new style K6 write allocation for %d Mb\n", + mbytes); + } + + /* Set MTRR capability flag if appropriate */ + if (c->x86_model == 13 || c->x86_model == 9 || + (c->x86_model == 8 && c->x86_mask >= 8)) + set_bit(X86_FEATURE_K6_MTRR, c->x86_capability); + break; + } + break; + + case 6: /* An Athlon/Duron */ + + /* Bit 15 of Athlon specific MSR 15, needs to be 0 + * to enable SSE on Palomino/Morgan/Barton CPU's. + * If the BIOS didn't enable it already, enable it here. + */ + if (c->x86_model >= 6 && c->x86_model <= 10) { + if (!cpu_has(c, X86_FEATURE_XMM)) { + printk(KERN_INFO "Enabling disabled K7/SSE Support.\n"); + rdmsr(MSR_K7_HWCR, l, h); + l &= ~0x00008000; + wrmsr(MSR_K7_HWCR, l, h); + set_bit(X86_FEATURE_XMM, c->x86_capability); + } + } + + /* It's been determined by AMD that Athlons since model 8 stepping 1 + * are more robust with CLK_CTL set to 200xxxxx instead of 600xxxxx + * As per AMD technical note 27212 0.2 + */ + if ((c->x86_model == 8 && c->x86_mask>=1) || (c->x86_model > 8)) { + rdmsr(MSR_K7_CLK_CTL, l, h); + if ((l & 0xfff00000) != 0x20000000) { + printk ("CPU: CLK_CTL MSR was %x. Reprogramming to %x\n", l, + ((l & 0x000fffff)|0x20000000)); + wrmsr(MSR_K7_CLK_CTL, (l & 0x000fffff)|0x20000000, h); + } + } + break; + } + + switch (c->x86) { + case 15: + set_bit(X86_FEATURE_K8, c->x86_capability); + break; + case 6: + set_bit(X86_FEATURE_K7, c->x86_capability); + break; + } + + display_cacheinfo(c); + detect_ht(c); + +#ifdef CONFIG_X86_HT + /* AMD dual core looks like HT but isn't really. Hide it from the + scheduler. This works around problems with the domain scheduler. + Also probably gives slightly better scheduling and disables + SMT nice which is harmful on dual core. + TBD tune the domain scheduler for dual core. */ + if (cpu_has(c, X86_FEATURE_CMP_LEGACY)) + smp_num_siblings = 1; +#endif + + if (cpuid_eax(0x80000000) >= 0x80000008) { + c->x86_num_cores = (cpuid_ecx(0x80000008) & 0xff) + 1; + if (c->x86_num_cores & (c->x86_num_cores - 1)) + c->x86_num_cores = 1; + } +} + +static unsigned int amd_size_cache(struct cpuinfo_x86 * c, unsigned int size) +{ + /* AMD errata T13 (order #21922) */ + if ((c->x86 == 6)) { + if (c->x86_model == 3 && c->x86_mask == 0) /* Duron Rev A0 */ + size = 64; + if (c->x86_model == 4 && + (c->x86_mask==0 || c->x86_mask==1)) /* Tbird rev A1/A2 */ + size = 256; + } + return size; +} + +static struct cpu_dev amd_cpu_dev __initdata = { + .c_vendor = "AMD", + .c_ident = { "AuthenticAMD" }, + .c_models = { + { .vendor = X86_VENDOR_AMD, .family = 4, .model_names = + { + [3] = "486 DX/2", + [7] = "486 DX/2-WB", + [8] = "486 DX/4", + [9] = "486 DX/4-WB", + [14] = "Am5x86-WT", + [15] = "Am5x86-WB" + } + }, + }, + .c_init = init_amd, + .c_identify = generic_identify, + .c_size_cache = amd_size_cache, +}; + +int __init amd_init_cpu(void) +{ + cpu_devs[X86_VENDOR_AMD] = &amd_cpu_dev; + return 0; +} + +//early_arch_initcall(amd_init_cpu); diff --git a/arch/i386/kernel/cpu/centaur.c b/arch/i386/kernel/cpu/centaur.c new file mode 100644 index 00000000000..394814e5767 --- /dev/null +++ b/arch/i386/kernel/cpu/centaur.c @@ -0,0 +1,476 @@ +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/bitops.h> +#include <asm/processor.h> +#include <asm/msr.h> +#include <asm/e820.h> +#include "cpu.h" + +#ifdef CONFIG_X86_OOSTORE + +static u32 __init power2(u32 x) +{ + u32 s=1; + while(s<=x) + s<<=1; + return s>>=1; +} + + +/* + * Set up an actual MCR + */ + +static void __init centaur_mcr_insert(int reg, u32 base, u32 size, int key) +{ + u32 lo, hi; + + hi = base & ~0xFFF; + lo = ~(size-1); /* Size is a power of 2 so this makes a mask */ + lo &= ~0xFFF; /* Remove the ctrl value bits */ + lo |= key; /* Attribute we wish to set */ + wrmsr(reg+MSR_IDT_MCR0, lo, hi); + mtrr_centaur_report_mcr(reg, lo, hi); /* Tell the mtrr driver */ +} + +/* + * Figure what we can cover with MCR's + * + * Shortcut: We know you can't put 4Gig of RAM on a winchip + */ + +static u32 __init ramtop(void) /* 16388 */ +{ + int i; + u32 top = 0; + u32 clip = 0xFFFFFFFFUL; + + for (i = 0; i < e820.nr_map; i++) { + unsigned long start, end; + + if (e820.map[i].addr > 0xFFFFFFFFUL) + continue; + /* + * Don't MCR over reserved space. Ignore the ISA hole + * we frob around that catastrophy already + */ + + if (e820.map[i].type == E820_RESERVED) + { + if(e820.map[i].addr >= 0x100000UL && e820.map[i].addr < clip) + clip = e820.map[i].addr; + continue; + } + start = e820.map[i].addr; + end = e820.map[i].addr + e820.map[i].size; + if (start >= end) + continue; + if (end > top) + top = end; + } + /* Everything below 'top' should be RAM except for the ISA hole. + Because of the limited MCR's we want to map NV/ACPI into our + MCR range for gunk in RAM + + Clip might cause us to MCR insufficient RAM but that is an + acceptable failure mode and should only bite obscure boxes with + a VESA hole at 15Mb + + The second case Clip sometimes kicks in is when the EBDA is marked + as reserved. Again we fail safe with reasonable results + */ + + if(top>clip) + top=clip; + + return top; +} + +/* + * Compute a set of MCR's to give maximum coverage + */ + +static int __init centaur_mcr_compute(int nr, int key) +{ + u32 mem = ramtop(); + u32 root = power2(mem); + u32 base = root; + u32 top = root; + u32 floor = 0; + int ct = 0; + + while(ct<nr) + { + u32 fspace = 0; + + /* + * Find the largest block we will fill going upwards + */ + + u32 high = power2(mem-top); + + /* + * Find the largest block we will fill going downwards + */ + + u32 low = base/2; + + /* + * Don't fill below 1Mb going downwards as there + * is an ISA hole in the way. + */ + + if(base <= 1024*1024) + low = 0; + + /* + * See how much space we could cover by filling below + * the ISA hole + */ + + if(floor == 0) + fspace = 512*1024; + else if(floor ==512*1024) + fspace = 128*1024; + + /* And forget ROM space */ + + /* + * Now install the largest coverage we get + */ + + if(fspace > high && fspace > low) + { + centaur_mcr_insert(ct, floor, fspace, key); + floor += fspace; + } + else if(high > low) + { + centaur_mcr_insert(ct, top, high, key); + top += high; + } + else if(low > 0) + { + base -= low; + centaur_mcr_insert(ct, base, low, key); + } + else break; + ct++; + } + /* + * We loaded ct values. We now need to set the mask. The caller + * must do this bit. + */ + + return ct; +} + +static void __init centaur_create_optimal_mcr(void) +{ + int i; + /* + * Allocate up to 6 mcrs to mark as much of ram as possible + * as write combining and weak write ordered. + * + * To experiment with: Linux never uses stack operations for + * mmio spaces so we could globally enable stack operation wc + * + * Load the registers with type 31 - full write combining, all + * writes weakly ordered. + */ + int used = centaur_mcr_compute(6, 31); + + /* + * Wipe unused MCRs + */ + + for(i=used;i<8;i++) + wrmsr(MSR_IDT_MCR0+i, 0, 0); +} + +static void __init winchip2_create_optimal_mcr(void) +{ + u32 lo, hi; + int i; + + /* + * Allocate up to 6 mcrs to mark as much of ram as possible + * as write combining, weak store ordered. + * + * Load the registers with type 25 + * 8 - weak write ordering + * 16 - weak read ordering + * 1 - write combining + */ + + int used = centaur_mcr_compute(6, 25); + + /* + * Mark the registers we are using. + */ + + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + for(i=0;i<used;i++) + lo|=1<<(9+i); + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); + + /* + * Wipe unused MCRs + */ + + for(i=used;i<8;i++) + wrmsr(MSR_IDT_MCR0+i, 0, 0); +} + +/* + * Handle the MCR key on the Winchip 2. + */ + +static void __init winchip2_unprotect_mcr(void) +{ + u32 lo, hi; + u32 key; + + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + lo&=~0x1C0; /* blank bits 8-6 */ + key = (lo>>17) & 7; + lo |= key<<6; /* replace with unlock key */ + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); +} + +static void __init winchip2_protect_mcr(void) +{ + u32 lo, hi; + + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + lo&=~0x1C0; /* blank bits 8-6 */ + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); +} +#endif /* CONFIG_X86_OOSTORE */ + +#define ACE_PRESENT (1 << 6) +#define ACE_ENABLED (1 << 7) +#define ACE_FCR (1 << 28) /* MSR_VIA_FCR */ + +#define RNG_PRESENT (1 << 2) +#define RNG_ENABLED (1 << 3) +#define RNG_ENABLE (1 << 6) /* MSR_VIA_RNG */ + +static void __init init_c3(struct cpuinfo_x86 *c) +{ + u32 lo, hi; + + /* Test for Centaur Extended Feature Flags presence */ + if (cpuid_eax(0xC0000000) >= 0xC0000001) { + u32 tmp = cpuid_edx(0xC0000001); + + /* enable ACE unit, if present and disabled */ + if ((tmp & (ACE_PRESENT | ACE_ENABLED)) == ACE_PRESENT) { + rdmsr (MSR_VIA_FCR, lo, hi); + lo |= ACE_FCR; /* enable ACE unit */ + wrmsr (MSR_VIA_FCR, lo, hi); + printk(KERN_INFO "CPU: Enabled ACE h/w crypto\n"); + } + + /* enable RNG unit, if present and disabled */ + if ((tmp & (RNG_PRESENT | RNG_ENABLED)) == RNG_PRESENT) { + rdmsr (MSR_VIA_RNG, lo, hi); + lo |= RNG_ENABLE; /* enable RNG unit */ + wrmsr (MSR_VIA_RNG, lo, hi); + printk(KERN_INFO "CPU: Enabled h/w RNG\n"); + } + + /* store Centaur Extended Feature Flags as + * word 5 of the CPU capability bit array + */ + c->x86_capability[5] = cpuid_edx(0xC0000001); + } + + /* Cyrix III family needs CX8 & PGE explicity enabled. */ + if (c->x86_model >=6 && c->x86_model <= 9) { + rdmsr (MSR_VIA_FCR, lo, hi); + lo |= (1<<1 | 1<<7); + wrmsr (MSR_VIA_FCR, lo, hi); + set_bit(X86_FEATURE_CX8, c->x86_capability); + } + + /* Before Nehemiah, the C3's had 3dNOW! */ + if (c->x86_model >=6 && c->x86_model <9) + set_bit(X86_FEATURE_3DNOW, c->x86_capability); + + get_model_name(c); + display_cacheinfo(c); +} + +static void __init init_centaur(struct cpuinfo_x86 *c) +{ + enum { + ECX8=1<<1, + EIERRINT=1<<2, + DPM=1<<3, + DMCE=1<<4, + DSTPCLK=1<<5, + ELINEAR=1<<6, + DSMC=1<<7, + DTLOCK=1<<8, + EDCTLB=1<<8, + EMMX=1<<9, + DPDC=1<<11, + EBRPRED=1<<12, + DIC=1<<13, + DDC=1<<14, + DNA=1<<15, + ERETSTK=1<<16, + E2MMX=1<<19, + EAMD3D=1<<20, + }; + + char *name; + u32 fcr_set=0; + u32 fcr_clr=0; + u32 lo,hi,newlo; + u32 aa,bb,cc,dd; + + /* Bit 31 in normal CPUID used for nonstandard 3DNow ID; + 3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */ + clear_bit(0*32+31, c->x86_capability); + + switch (c->x86) { + + case 5: + switch(c->x86_model) { + case 4: + name="C6"; + fcr_set=ECX8|DSMC|EDCTLB|EMMX|ERETSTK; + fcr_clr=DPDC; + printk(KERN_NOTICE "Disabling bugged TSC.\n"); + clear_bit(X86_FEATURE_TSC, c->x86_capability); +#ifdef CONFIG_X86_OOSTORE + centaur_create_optimal_mcr(); + /* Enable + write combining on non-stack, non-string + write combining on string, all types + weak write ordering + + The C6 original lacks weak read order + + Note 0x120 is write only on Winchip 1 */ + + wrmsr(MSR_IDT_MCR_CTRL, 0x01F0001F, 0); +#endif + break; + case 8: + switch(c->x86_mask) { + default: + name="2"; + break; + case 7 ... 9: + name="2A"; + break; + case 10 ... 15: + name="2B"; + break; + } + fcr_set=ECX8|DSMC|DTLOCK|EMMX|EBRPRED|ERETSTK|E2MMX|EAMD3D; + fcr_clr=DPDC; +#ifdef CONFIG_X86_OOSTORE + winchip2_unprotect_mcr(); + winchip2_create_optimal_mcr(); + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + /* Enable + write combining on non-stack, non-string + write combining on string, all types + weak write ordering + */ + lo|=31; + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); + winchip2_protect_mcr(); +#endif + break; + case 9: + name="3"; + fcr_set=ECX8|DSMC|DTLOCK|EMMX|EBRPRED|ERETSTK|E2MMX|EAMD3D; + fcr_clr=DPDC; +#ifdef CONFIG_X86_OOSTORE + winchip2_unprotect_mcr(); + winchip2_create_optimal_mcr(); + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + /* Enable + write combining on non-stack, non-string + write combining on string, all types + weak write ordering + */ + lo|=31; + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); + winchip2_protect_mcr(); +#endif + break; + case 10: + name="4"; + /* no info on the WC4 yet */ + break; + default: + name="??"; + } + + rdmsr(MSR_IDT_FCR1, lo, hi); + newlo=(lo|fcr_set) & (~fcr_clr); + + if (newlo!=lo) { + printk(KERN_INFO "Centaur FCR was 0x%X now 0x%X\n", lo, newlo ); + wrmsr(MSR_IDT_FCR1, newlo, hi ); + } else { + printk(KERN_INFO "Centaur FCR is 0x%X\n",lo); + } + /* Emulate MTRRs using Centaur's MCR. */ + set_bit(X86_FEATURE_CENTAUR_MCR, c->x86_capability); + /* Report CX8 */ + set_bit(X86_FEATURE_CX8, c->x86_capability); + /* Set 3DNow! on Winchip 2 and above. */ + if (c->x86_model >=8) + set_bit(X86_FEATURE_3DNOW, c->x86_capability); + /* See if we can find out some more. */ + if ( cpuid_eax(0x80000000) >= 0x80000005 ) { + /* Yes, we can. */ + cpuid(0x80000005,&aa,&bb,&cc,&dd); + /* Add L1 data and code cache sizes. */ + c->x86_cache_size = (cc>>24)+(dd>>24); + } + sprintf( c->x86_model_id, "WinChip %s", name ); + break; + + case 6: + init_c3(c); + break; + } +} + +static unsigned int centaur_size_cache(struct cpuinfo_x86 * c, unsigned int size) +{ + /* VIA C3 CPUs (670-68F) need further shifting. */ + if ((c->x86 == 6) && ((c->x86_model == 7) || (c->x86_model == 8))) + size >>= 8; + + /* VIA also screwed up Nehemiah stepping 1, and made + it return '65KB' instead of '64KB' + - Note, it seems this may only be in engineering samples. */ + if ((c->x86==6) && (c->x86_model==9) && (c->x86_mask==1) && (size==65)) + size -=1; + + return size; +} + +static struct cpu_dev centaur_cpu_dev __initdata = { + .c_vendor = "Centaur", + .c_ident = { "CentaurHauls" }, + .c_init = init_centaur, + .c_size_cache = centaur_size_cache, +}; + +int __init centaur_init_cpu(void) +{ + cpu_devs[X86_VENDOR_CENTAUR] = ¢aur_cpu_dev; + return 0; +} + +//early_arch_initcall(centaur_init_cpu); diff --git a/arch/i386/kernel/cpu/changelog b/arch/i386/kernel/cpu/changelog new file mode 100644 index 00000000000..cef76b80a71 --- /dev/null +++ b/arch/i386/kernel/cpu/changelog @@ -0,0 +1,63 @@ +/* + * Enhanced CPU type detection by Mike Jagdis, Patrick St. Jean + * and Martin Mares, November 1997. + * + * Force Cyrix 6x86(MX) and M II processors to report MTRR capability + * and Cyrix "coma bug" recognition by + * Zoltán Böszörményi <zboszor@mail.externet.hu> February 1999. + * + * Force Centaur C6 processors to report MTRR capability. + * Bart Hartgers <bart@etpmod.phys.tue.nl>, May 1999. + * + * Intel Mobile Pentium II detection fix. Sean Gilley, June 1999. + * + * IDT Winchip tweaks, misc clean ups. + * Dave Jones <davej@suse.de>, August 1999 + * + * Better detection of Centaur/IDT WinChip models. + * Bart Hartgers <bart@etpmod.phys.tue.nl>, August 1999. + * + * Cleaned up cache-detection code + * Dave Jones <davej@suse.de>, October 1999 + * + * Added proper L2 cache detection for Coppermine + * Dragan Stancevic <visitor@valinux.com>, October 1999 + * + * Added the original array for capability flags but forgot to credit + * myself :) (~1998) Fixed/cleaned up some cpu_model_info and other stuff + * Jauder Ho <jauderho@carumba.com>, January 2000 + * + * Detection for Celeron coppermine, identify_cpu() overhauled, + * and a few other clean ups. + * Dave Jones <davej@suse.de>, April 2000 + * + * Pentium III FXSR, SSE support + * General FPU state handling cleanups + * Gareth Hughes <gareth@valinux.com>, May 2000 + * + * Added proper Cascades CPU and L2 cache detection for Cascades + * and 8-way type cache happy bunch from Intel:^) + * Dragan Stancevic <visitor@valinux.com>, May 2000 + * + * Forward port AMD Duron errata T13 from 2.2.17pre + * Dave Jones <davej@suse.de>, August 2000 + * + * Forward port lots of fixes/improvements from 2.2.18pre + * Cyrix III, Pentium IV support. + * Dave Jones <davej@suse.de>, October 2000 + * + * Massive cleanup of CPU detection and bug handling; + * Transmeta CPU detection, + * H. Peter Anvin <hpa@zytor.com>, November 2000 + * + * VIA C3 Support. + * Dave Jones <davej@suse.de>, March 2001 + * + * AMD Athlon/Duron/Thunderbird bluesmoke support. + * Dave Jones <davej@suse.de>, April 2001. + * + * CacheSize bug workaround updates for AMD, Intel & VIA Cyrix. + * Dave Jones <davej@suse.de>, September, October 2001. + * + */ + diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c new file mode 100644 index 00000000000..ebd5d8247fa --- /dev/null +++ b/arch/i386/kernel/cpu/common.c @@ -0,0 +1,634 @@ +#include <linux/init.h> +#include <linux/string.h> +#include <linux/delay.h> +#include <linux/smp.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <asm/semaphore.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/msr.h> +#include <asm/io.h> +#include <asm/mmu_context.h> +#ifdef CONFIG_X86_LOCAL_APIC +#include <asm/mpspec.h> +#include <asm/apic.h> +#include <mach_apic.h> +#endif + +#include "cpu.h" + +DEFINE_PER_CPU(struct desc_struct, cpu_gdt_table[GDT_ENTRIES]); +EXPORT_PER_CPU_SYMBOL(cpu_gdt_table); + +DEFINE_PER_CPU(unsigned char, cpu_16bit_stack[CPU_16BIT_STACK_SIZE]); +EXPORT_PER_CPU_SYMBOL(cpu_16bit_stack); + +static int cachesize_override __initdata = -1; +static int disable_x86_fxsr __initdata = 0; +static int disable_x86_serial_nr __initdata = 1; + +struct cpu_dev * cpu_devs[X86_VENDOR_NUM] = {}; + +extern void mcheck_init(struct cpuinfo_x86 *c); + +extern int disable_pse; + +static void default_init(struct cpuinfo_x86 * c) +{ + /* Not much we can do here... */ + /* Check if at least it has cpuid */ + if (c->cpuid_level == -1) { + /* No cpuid. It must be an ancient CPU */ + if (c->x86 == 4) + strcpy(c->x86_model_id, "486"); + else if (c->x86 == 3) + strcpy(c->x86_model_id, "386"); + } +} + +static struct cpu_dev default_cpu = { + .c_init = default_init, +}; +static struct cpu_dev * this_cpu = &default_cpu; + +static int __init cachesize_setup(char *str) +{ + get_option (&str, &cachesize_override); + return 1; +} +__setup("cachesize=", cachesize_setup); + +int __init get_model_name(struct cpuinfo_x86 *c) +{ + unsigned int *v; + char *p, *q; + + if (cpuid_eax(0x80000000) < 0x80000004) + return 0; + + v = (unsigned int *) c->x86_model_id; + cpuid(0x80000002, &v[0], &v[1], &v[2], &v[3]); + cpuid(0x80000003, &v[4], &v[5], &v[6], &v[7]); + cpuid(0x80000004, &v[8], &v[9], &v[10], &v[11]); + c->x86_model_id[48] = 0; + + /* Intel chips right-justify this string for some dumb reason; + undo that brain damage */ + p = q = &c->x86_model_id[0]; + while ( *p == ' ' ) + p++; + if ( p != q ) { + while ( *p ) + *q++ = *p++; + while ( q <= &c->x86_model_id[48] ) + *q++ = '\0'; /* Zero-pad the rest */ + } + + return 1; +} + + +void __init display_cacheinfo(struct cpuinfo_x86 *c) +{ + unsigned int n, dummy, ecx, edx, l2size; + + n = cpuid_eax(0x80000000); + + if (n >= 0x80000005) { + cpuid(0x80000005, &dummy, &dummy, &ecx, &edx); + printk(KERN_INFO "CPU: L1 I Cache: %dK (%d bytes/line), D cache %dK (%d bytes/line)\n", + edx>>24, edx&0xFF, ecx>>24, ecx&0xFF); + c->x86_cache_size=(ecx>>24)+(edx>>24); + } + + if (n < 0x80000006) /* Some chips just has a large L1. */ + return; + + ecx = cpuid_ecx(0x80000006); + l2size = ecx >> 16; + + /* do processor-specific cache resizing */ + if (this_cpu->c_size_cache) + l2size = this_cpu->c_size_cache(c,l2size); + + /* Allow user to override all this if necessary. */ + if (cachesize_override != -1) + l2size = cachesize_override; + + if ( l2size == 0 ) + return; /* Again, no L2 cache is possible */ + + c->x86_cache_size = l2size; + + printk(KERN_INFO "CPU: L2 Cache: %dK (%d bytes/line)\n", + l2size, ecx & 0xFF); +} + +/* Naming convention should be: <Name> [(<Codename>)] */ +/* This table only is used unless init_<vendor>() below doesn't set it; */ +/* in particular, if CPUID levels 0x80000002..4 are supported, this isn't used */ + +/* Look up CPU names by table lookup. */ +static char __init *table_lookup_model(struct cpuinfo_x86 *c) +{ + struct cpu_model_info *info; + + if ( c->x86_model >= 16 ) + return NULL; /* Range check */ + + if (!this_cpu) + return NULL; + + info = this_cpu->c_models; + + while (info && info->family) { + if (info->family == c->x86) + return info->model_names[c->x86_model]; + info++; + } + return NULL; /* Not found */ +} + + +void __init get_cpu_vendor(struct cpuinfo_x86 *c, int early) +{ + char *v = c->x86_vendor_id; + int i; + + for (i = 0; i < X86_VENDOR_NUM; i++) { + if (cpu_devs[i]) { + if (!strcmp(v,cpu_devs[i]->c_ident[0]) || + (cpu_devs[i]->c_ident[1] && + !strcmp(v,cpu_devs[i]->c_ident[1]))) { + c->x86_vendor = i; + if (!early) + this_cpu = cpu_devs[i]; + break; + } + } + } +} + + +static int __init x86_fxsr_setup(char * s) +{ + disable_x86_fxsr = 1; + return 1; +} +__setup("nofxsr", x86_fxsr_setup); + + +/* Standard macro to see if a specific flag is changeable */ +static inline int flag_is_changeable_p(u32 flag) +{ + u32 f1, f2; + + asm("pushfl\n\t" + "pushfl\n\t" + "popl %0\n\t" + "movl %0,%1\n\t" + "xorl %2,%0\n\t" + "pushl %0\n\t" + "popfl\n\t" + "pushfl\n\t" + "popl %0\n\t" + "popfl\n\t" + : "=&r" (f1), "=&r" (f2) + : "ir" (flag)); + + return ((f1^f2) & flag) != 0; +} + + +/* Probe for the CPUID instruction */ +static int __init have_cpuid_p(void) +{ + return flag_is_changeable_p(X86_EFLAGS_ID); +} + +/* Do minimum CPU detection early. + Fields really needed: vendor, cpuid_level, family, model, mask, cache alignment. + The others are not touched to avoid unwanted side effects. */ +static void __init early_cpu_detect(void) +{ + struct cpuinfo_x86 *c = &boot_cpu_data; + + c->x86_cache_alignment = 32; + + if (!have_cpuid_p()) + return; + + /* Get vendor name */ + cpuid(0x00000000, &c->cpuid_level, + (int *)&c->x86_vendor_id[0], + (int *)&c->x86_vendor_id[8], + (int *)&c->x86_vendor_id[4]); + + get_cpu_vendor(c, 1); + + c->x86 = 4; + if (c->cpuid_level >= 0x00000001) { + u32 junk, tfms, cap0, misc; + cpuid(0x00000001, &tfms, &misc, &junk, &cap0); + c->x86 = (tfms >> 8) & 15; + c->x86_model = (tfms >> 4) & 15; + if (c->x86 == 0xf) { + c->x86 += (tfms >> 20) & 0xff; + c->x86_model += ((tfms >> 16) & 0xF) << 4; + } + c->x86_mask = tfms & 15; + if (cap0 & (1<<19)) + c->x86_cache_alignment = ((misc >> 8) & 0xff) * 8; + } + + early_intel_workaround(c); +} + +void __init generic_identify(struct cpuinfo_x86 * c) +{ + u32 tfms, xlvl; + int junk; + + if (have_cpuid_p()) { + /* Get vendor name */ + cpuid(0x00000000, &c->cpuid_level, + (int *)&c->x86_vendor_id[0], + (int *)&c->x86_vendor_id[8], + (int *)&c->x86_vendor_id[4]); + + get_cpu_vendor(c, 0); + /* Initialize the standard set of capabilities */ + /* Note that the vendor-specific code below might override */ + + /* Intel-defined flags: level 0x00000001 */ + if ( c->cpuid_level >= 0x00000001 ) { + u32 capability, excap; + cpuid(0x00000001, &tfms, &junk, &excap, &capability); + c->x86_capability[0] = capability; + c->x86_capability[4] = excap; + c->x86 = (tfms >> 8) & 15; + c->x86_model = (tfms >> 4) & 15; + if (c->x86 == 0xf) { + c->x86 += (tfms >> 20) & 0xff; + c->x86_model += ((tfms >> 16) & 0xF) << 4; + } + c->x86_mask = tfms & 15; + } else { + /* Have CPUID level 0 only - unheard of */ + c->x86 = 4; + } + + /* AMD-defined flags: level 0x80000001 */ + xlvl = cpuid_eax(0x80000000); + if ( (xlvl & 0xffff0000) == 0x80000000 ) { + if ( xlvl >= 0x80000001 ) { + c->x86_capability[1] = cpuid_edx(0x80000001); + c->x86_capability[6] = cpuid_ecx(0x80000001); + } + if ( xlvl >= 0x80000004 ) + get_model_name(c); /* Default name */ + } + } +} + +static void __init squash_the_stupid_serial_number(struct cpuinfo_x86 *c) +{ + if (cpu_has(c, X86_FEATURE_PN) && disable_x86_serial_nr ) { + /* Disable processor serial number */ + unsigned long lo,hi; + rdmsr(MSR_IA32_BBL_CR_CTL,lo,hi); + lo |= 0x200000; + wrmsr(MSR_IA32_BBL_CR_CTL,lo,hi); + printk(KERN_NOTICE "CPU serial number disabled.\n"); + clear_bit(X86_FEATURE_PN, c->x86_capability); + + /* Disabling the serial number may affect the cpuid level */ + c->cpuid_level = cpuid_eax(0); + } +} + +static int __init x86_serial_nr_setup(char *s) +{ + disable_x86_serial_nr = 0; + return 1; +} +__setup("serialnumber", x86_serial_nr_setup); + + + +/* + * This does the hard work of actually picking apart the CPU stuff... + */ +void __init identify_cpu(struct cpuinfo_x86 *c) +{ + int i; + + c->loops_per_jiffy = loops_per_jiffy; + c->x86_cache_size = -1; + c->x86_vendor = X86_VENDOR_UNKNOWN; + c->cpuid_level = -1; /* CPUID not detected */ + c->x86_model = c->x86_mask = 0; /* So far unknown... */ + c->x86_vendor_id[0] = '\0'; /* Unset */ + c->x86_model_id[0] = '\0'; /* Unset */ + c->x86_num_cores = 1; + memset(&c->x86_capability, 0, sizeof c->x86_capability); + + if (!have_cpuid_p()) { + /* First of all, decide if this is a 486 or higher */ + /* It's a 486 if we can modify the AC flag */ + if ( flag_is_changeable_p(X86_EFLAGS_AC) ) + c->x86 = 4; + else + c->x86 = 3; + } + + generic_identify(c); + + printk(KERN_DEBUG "CPU: After generic identify, caps:"); + for (i = 0; i < NCAPINTS; i++) + printk(" %08lx", c->x86_capability[i]); + printk("\n"); + + if (this_cpu->c_identify) { + this_cpu->c_identify(c); + + printk(KERN_DEBUG "CPU: After vendor identify, caps:"); + for (i = 0; i < NCAPINTS; i++) + printk(" %08lx", c->x86_capability[i]); + printk("\n"); + } + + /* + * Vendor-specific initialization. In this section we + * canonicalize the feature flags, meaning if there are + * features a certain CPU supports which CPUID doesn't + * tell us, CPUID claiming incorrect flags, or other bugs, + * we handle them here. + * + * At the end of this section, c->x86_capability better + * indicate the features this CPU genuinely supports! + */ + if (this_cpu->c_init) + this_cpu->c_init(c); + + /* Disable the PN if appropriate */ + squash_the_stupid_serial_number(c); + + /* + * The vendor-specific functions might have changed features. Now + * we do "generic changes." + */ + + /* TSC disabled? */ + if ( tsc_disable ) + clear_bit(X86_FEATURE_TSC, c->x86_capability); + + /* FXSR disabled? */ + if (disable_x86_fxsr) { + clear_bit(X86_FEATURE_FXSR, c->x86_capability); + clear_bit(X86_FEATURE_XMM, c->x86_capability); + } + + if (disable_pse) + clear_bit(X86_FEATURE_PSE, c->x86_capability); + + /* If the model name is still unset, do table lookup. */ + if ( !c->x86_model_id[0] ) { + char *p; + p = table_lookup_model(c); + if ( p ) + strcpy(c->x86_model_id, p); + else + /* Last resort... */ + sprintf(c->x86_model_id, "%02x/%02x", + c->x86_vendor, c->x86_model); + } + + /* Now the feature flags better reflect actual CPU features! */ + + printk(KERN_DEBUG "CPU: After all inits, caps:"); + for (i = 0; i < NCAPINTS; i++) + printk(" %08lx", c->x86_capability[i]); + printk("\n"); + + /* + * On SMP, boot_cpu_data holds the common feature set between + * all CPUs; so make sure that we indicate which features are + * common between the CPUs. The first time this routine gets + * executed, c == &boot_cpu_data. + */ + if ( c != &boot_cpu_data ) { + /* AND the already accumulated flags with these */ + for ( i = 0 ; i < NCAPINTS ; i++ ) + boot_cpu_data.x86_capability[i] &= c->x86_capability[i]; + } + + /* Init Machine Check Exception if available. */ +#ifdef CONFIG_X86_MCE + mcheck_init(c); +#endif +} + +#ifdef CONFIG_X86_HT +void __init detect_ht(struct cpuinfo_x86 *c) +{ + u32 eax, ebx, ecx, edx; + int index_lsb, index_msb, tmp; + int cpu = smp_processor_id(); + + if (!cpu_has(c, X86_FEATURE_HT)) + return; + + cpuid(1, &eax, &ebx, &ecx, &edx); + smp_num_siblings = (ebx & 0xff0000) >> 16; + + if (smp_num_siblings == 1) { + printk(KERN_INFO "CPU: Hyper-Threading is disabled\n"); + } else if (smp_num_siblings > 1 ) { + index_lsb = 0; + index_msb = 31; + + if (smp_num_siblings > NR_CPUS) { + printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings); + smp_num_siblings = 1; + return; + } + tmp = smp_num_siblings; + while ((tmp & 1) == 0) { + tmp >>=1 ; + index_lsb++; + } + tmp = smp_num_siblings; + while ((tmp & 0x80000000 ) == 0) { + tmp <<=1 ; + index_msb--; + } + if (index_lsb != index_msb ) + index_msb++; + phys_proc_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb); + + printk(KERN_INFO "CPU: Physical Processor ID: %d\n", + phys_proc_id[cpu]); + } +} +#endif + +void __init print_cpu_info(struct cpuinfo_x86 *c) +{ + char *vendor = NULL; + + if (c->x86_vendor < X86_VENDOR_NUM) + vendor = this_cpu->c_vendor; + else if (c->cpuid_level >= 0) + vendor = c->x86_vendor_id; + + if (vendor && strncmp(c->x86_model_id, vendor, strlen(vendor))) + printk("%s ", vendor); + + if (!c->x86_model_id[0]) + printk("%d86", c->x86); + else + printk("%s", c->x86_model_id); + + if (c->x86_mask || c->cpuid_level >= 0) + printk(" stepping %02x\n", c->x86_mask); + else + printk("\n"); +} + +cpumask_t cpu_initialized __initdata = CPU_MASK_NONE; + +/* This is hacky. :) + * We're emulating future behavior. + * In the future, the cpu-specific init functions will be called implicitly + * via the magic of initcalls. + * They will insert themselves into the cpu_devs structure. + * Then, when cpu_init() is called, we can just iterate over that array. + */ + +extern int intel_cpu_init(void); +extern int cyrix_init_cpu(void); +extern int nsc_init_cpu(void); +extern int amd_init_cpu(void); +extern int centaur_init_cpu(void); +extern int transmeta_init_cpu(void); +extern int rise_init_cpu(void); +extern int nexgen_init_cpu(void); +extern int umc_init_cpu(void); + +void __init early_cpu_init(void) +{ + intel_cpu_init(); + cyrix_init_cpu(); + nsc_init_cpu(); + amd_init_cpu(); + centaur_init_cpu(); + transmeta_init_cpu(); + rise_init_cpu(); + nexgen_init_cpu(); + umc_init_cpu(); + early_cpu_detect(); + +#ifdef CONFIG_DEBUG_PAGEALLOC + /* pse is not compatible with on-the-fly unmapping, + * disable it even if the cpus claim to support it. + */ + clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability); + disable_pse = 1; +#endif +} +/* + * cpu_init() initializes state that is per-CPU. Some data is already + * initialized (naturally) in the bootstrap process, such as the GDT + * and IDT. We reload them nevertheless, this function acts as a + * 'CPU state barrier', nothing should get across. + */ +void __init cpu_init (void) +{ + int cpu = smp_processor_id(); + struct tss_struct * t = &per_cpu(init_tss, cpu); + struct thread_struct *thread = ¤t->thread; + __u32 stk16_off = (__u32)&per_cpu(cpu_16bit_stack, cpu); + + if (cpu_test_and_set(cpu, cpu_initialized)) { + printk(KERN_WARNING "CPU#%d already initialized!\n", cpu); + for (;;) local_irq_enable(); + } + printk(KERN_INFO "Initializing CPU#%d\n", cpu); + + if (cpu_has_vme || cpu_has_tsc || cpu_has_de) + clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); + if (tsc_disable && cpu_has_tsc) { + printk(KERN_NOTICE "Disabling TSC...\n"); + /**** FIX-HPA: DOES THIS REALLY BELONG HERE? ****/ + clear_bit(X86_FEATURE_TSC, boot_cpu_data.x86_capability); + set_in_cr4(X86_CR4_TSD); + } + + /* + * Initialize the per-CPU GDT with the boot GDT, + * and set up the GDT descriptor: + */ + memcpy(&per_cpu(cpu_gdt_table, cpu), cpu_gdt_table, + GDT_SIZE); + + /* Set up GDT entry for 16bit stack */ + *(__u64 *)&(per_cpu(cpu_gdt_table, cpu)[GDT_ENTRY_ESPFIX_SS]) |= + ((((__u64)stk16_off) << 16) & 0x000000ffffff0000ULL) | + ((((__u64)stk16_off) << 32) & 0xff00000000000000ULL) | + (CPU_16BIT_STACK_SIZE - 1); + + cpu_gdt_descr[cpu].size = GDT_SIZE - 1; + cpu_gdt_descr[cpu].address = + (unsigned long)&per_cpu(cpu_gdt_table, cpu); + + /* + * Set up the per-thread TLS descriptor cache: + */ + memcpy(thread->tls_array, &per_cpu(cpu_gdt_table, cpu), + GDT_ENTRY_TLS_ENTRIES * 8); + + __asm__ __volatile__("lgdt %0" : : "m" (cpu_gdt_descr[cpu])); + __asm__ __volatile__("lidt %0" : : "m" (idt_descr)); + + /* + * Delete NT + */ + __asm__("pushfl ; andl $0xffffbfff,(%esp) ; popfl"); + + /* + * Set up and load the per-CPU TSS and LDT + */ + atomic_inc(&init_mm.mm_count); + current->active_mm = &init_mm; + if (current->mm) + BUG(); + enter_lazy_tlb(&init_mm, current); + + load_esp0(t, thread); + set_tss_desc(cpu,t); + load_TR_desc(); + load_LDT(&init_mm.context); + + /* Set up doublefault TSS pointer in the GDT */ + __set_tss_desc(cpu, GDT_ENTRY_DOUBLEFAULT_TSS, &doublefault_tss); + + /* Clear %fs and %gs. */ + asm volatile ("xorl %eax, %eax; movl %eax, %fs; movl %eax, %gs"); + + /* Clear all 6 debug registers: */ + +#define CD(register) __asm__("movl %0,%%db" #register ::"r"(0) ); + + CD(0); CD(1); CD(2); CD(3); /* no db4 and db5 */; CD(6); CD(7); + +#undef CD + + /* + * Force FPU initialization: + */ + current_thread_info()->status = 0; + clear_used_math(); + mxcsr_feature_mask_init(); +} diff --git a/arch/i386/kernel/cpu/cpu.h b/arch/i386/kernel/cpu/cpu.h new file mode 100644 index 00000000000..5a1d4f163e8 --- /dev/null +++ b/arch/i386/kernel/cpu/cpu.h @@ -0,0 +1,30 @@ + +struct cpu_model_info { + int vendor; + int family; + char *model_names[16]; +}; + +/* attempt to consolidate cpu attributes */ +struct cpu_dev { + char * c_vendor; + + /* some have two possibilities for cpuid string */ + char * c_ident[2]; + + struct cpu_model_info c_models[4]; + + void (*c_init)(struct cpuinfo_x86 * c); + void (*c_identify)(struct cpuinfo_x86 * c); + unsigned int (*c_size_cache)(struct cpuinfo_x86 * c, unsigned int size); +}; + +extern struct cpu_dev * cpu_devs [X86_VENDOR_NUM]; + +extern int get_model_name(struct cpuinfo_x86 *c); +extern void display_cacheinfo(struct cpuinfo_x86 *c); + +extern void generic_identify(struct cpuinfo_x86 * c); + +extern void early_intel_workaround(struct cpuinfo_x86 *c); + diff --git a/arch/i386/kernel/cpu/cpufreq/Kconfig b/arch/i386/kernel/cpu/cpufreq/Kconfig new file mode 100644 index 00000000000..f25ffd74235 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/Kconfig @@ -0,0 +1,231 @@ +# +# CPU Frequency scaling +# + +menu "CPU Frequency scaling" + +source "drivers/cpufreq/Kconfig" + +if CPU_FREQ + +comment "CPUFreq processor drivers" + +config X86_ACPI_CPUFREQ + tristate "ACPI Processor P-States driver" + select CPU_FREQ_TABLE + depends on ACPI_PROCESSOR + help + This driver adds a CPUFreq driver which utilizes the ACPI + Processor Performance States. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config ELAN_CPUFREQ + tristate "AMD Elan" + select CPU_FREQ_TABLE + depends on X86_ELAN + ---help--- + This adds the CPUFreq driver for AMD Elan SC400 and SC410 + processors. + + You need to specify the processor maximum speed as boot + parameter: elanfreq=maxspeed (in kHz) or as module + parameter "max_freq". + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_POWERNOW_K6 + tristate "AMD Mobile K6-2/K6-3 PowerNow!" + select CPU_FREQ_TABLE + help + This adds the CPUFreq driver for mobile AMD K6-2+ and mobile + AMD K6-3+ processors. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_POWERNOW_K7 + tristate "AMD Mobile Athlon/Duron PowerNow!" + select CPU_FREQ_TABLE + help + This adds the CPUFreq driver for mobile AMD K7 mobile processors. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_POWERNOW_K7_ACPI + bool + depends on X86_POWERNOW_K7 && ACPI_PROCESSOR + depends on !(X86_POWERNOW_K7 = y && ACPI_PROCESSOR = m) + default y + +config X86_POWERNOW_K8 + tristate "AMD Opteron/Athlon64 PowerNow!" + select CPU_FREQ_TABLE + depends on EXPERIMENTAL + help + This adds the CPUFreq driver for mobile AMD Opteron/Athlon64 processors. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_POWERNOW_K8_ACPI + bool + depends on X86_POWERNOW_K8 && ACPI_PROCESSOR + depends on !(X86_POWERNOW_K8 = y && ACPI_PROCESSOR = m) + default y + +config X86_GX_SUSPMOD + tristate "Cyrix MediaGX/NatSemi Geode Suspend Modulation" + help + This add the CPUFreq driver for NatSemi Geode processors which + support suspend modulation. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_SPEEDSTEP_CENTRINO + tristate "Intel Enhanced SpeedStep" + select CPU_FREQ_TABLE + select X86_SPEEDSTEP_CENTRINO_TABLE if (!X86_SPEEDSTEP_CENTRINO_ACPI) + help + This adds the CPUFreq driver for Enhanced SpeedStep enabled + mobile CPUs. This means Intel Pentium M (Centrino) CPUs. However, + you also need to say Y to "Use ACPI tables to decode..." below + [which might imply enabling ACPI] if you want to use this driver + on non-Banias CPUs. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_SPEEDSTEP_CENTRINO_ACPI + bool "Use ACPI tables to decode valid frequency/voltage pairs" + depends on X86_SPEEDSTEP_CENTRINO && ACPI_PROCESSOR + depends on !(X86_SPEEDSTEP_CENTRINO = y && ACPI_PROCESSOR = m) + default y + help + Use primarily the information provided in the BIOS ACPI tables + to determine valid CPU frequency and voltage pairings. It is + required for the driver to work on non-Banias CPUs. + + If in doubt, say Y. + +config X86_SPEEDSTEP_CENTRINO_TABLE + bool "Built-in tables for Banias CPUs" + depends on X86_SPEEDSTEP_CENTRINO + default y + help + Use built-in tables for Banias CPUs if ACPI encoding + is not available. + + If in doubt, say N. + +config X86_SPEEDSTEP_ICH + tristate "Intel Speedstep on ICH-M chipsets (ioport interface)" + select CPU_FREQ_TABLE + help + This adds the CPUFreq driver for certain mobile Intel Pentium III + (Coppermine), all mobile Intel Pentium III-M (Tualatin) and all + mobile Intel Pentium 4 P4-M on systems which have an Intel ICH2, + ICH3 or ICH4 southbridge. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_SPEEDSTEP_SMI + tristate "Intel SpeedStep on 440BX/ZX/MX chipsets (SMI interface)" + select CPU_FREQ_TABLE + depends on EXPERIMENTAL + help + This adds the CPUFreq driver for certain mobile Intel Pentium III + (Coppermine), all mobile Intel Pentium III-M (Tualatin) + on systems which have an Intel 440BX/ZX/MX southbridge. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_P4_CLOCKMOD + tristate "Intel Pentium 4 clock modulation" + select CPU_FREQ_TABLE + help + This adds the CPUFreq driver for Intel Pentium 4 / XEON + processors. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_CPUFREQ_NFORCE2 + tristate "nVidia nForce2 FSB changing" + depends on EXPERIMENTAL + help + This adds the CPUFreq driver for FSB changing on nVidia nForce2 + platforms. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_LONGRUN + tristate "Transmeta LongRun" + help + This adds the CPUFreq driver for Transmeta Crusoe and Efficeon processors + which support LongRun. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +config X86_LONGHAUL + tristate "VIA Cyrix III Longhaul" + select CPU_FREQ_TABLE + help + This adds the CPUFreq driver for VIA Samuel/CyrixIII, + VIA Cyrix Samuel/C3, VIA Cyrix Ezra and VIA Cyrix Ezra-T + processors. + + For details, take a look at <file:Documentation/cpu-freq/>. + + If in doubt, say N. + +comment "shared options" + +config X86_ACPI_CPUFREQ_PROC_INTF + bool "/proc/acpi/processor/../performance interface (deprecated)" + depends on PROC_FS + depends on X86_ACPI_CPUFREQ || X86_SPEEDSTEP_CENTRINO_ACPI || X86_POWERNOW_K7_ACPI || X86_POWERNOW_K8_ACPI + help + This enables the deprecated /proc/acpi/processor/../performance + interface. While it is helpful for debugging, the generic, + cross-architecture cpufreq interfaces should be used. + + If in doubt, say N. + +config X86_SPEEDSTEP_LIB + tristate + default X86_SPEEDSTEP_ICH || X86_SPEEDSTEP_SMI || X86_P4_CLOCKMOD + +config X86_SPEEDSTEP_RELAXED_CAP_CHECK + bool "Relaxed speedstep capability checks" + depends on (X86_SPEEDSTEP_SMI || X86_SPEEDSTEP_ICH) + help + Don't perform all checks for a speedstep capable system which would + normally be done. Some ancient or strange systems, though speedstep + capable, don't always indicate that they are speedstep capable. This + option lets the probing code bypass some of those checks if the + parameter "relaxed_check=1" is passed to the module. + +endif # CPU_FREQ + +endmenu diff --git a/arch/i386/kernel/cpu/cpufreq/Makefile b/arch/i386/kernel/cpu/cpufreq/Makefile new file mode 100644 index 00000000000..a922e97aeed --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/Makefile @@ -0,0 +1,14 @@ +obj-$(CONFIG_X86_POWERNOW_K6) += powernow-k6.o +obj-$(CONFIG_X86_POWERNOW_K7) += powernow-k7.o +obj-$(CONFIG_X86_POWERNOW_K8) += powernow-k8.o +obj-$(CONFIG_X86_LONGHAUL) += longhaul.o +obj-$(CONFIG_ELAN_CPUFREQ) += elanfreq.o +obj-$(CONFIG_X86_LONGRUN) += longrun.o +obj-$(CONFIG_X86_GX_SUSPMOD) += gx-suspmod.o +obj-$(CONFIG_X86_SPEEDSTEP_ICH) += speedstep-ich.o +obj-$(CONFIG_X86_SPEEDSTEP_CENTRINO) += speedstep-centrino.o +obj-$(CONFIG_X86_SPEEDSTEP_LIB) += speedstep-lib.o +obj-$(CONFIG_X86_SPEEDSTEP_SMI) += speedstep-smi.o +obj-$(CONFIG_X86_ACPI_CPUFREQ) += acpi-cpufreq.o +obj-$(CONFIG_X86_P4_CLOCKMOD) += p4-clockmod.o +obj-$(CONFIG_X86_CPUFREQ_NFORCE2) += cpufreq-nforce2.o diff --git a/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c new file mode 100644 index 00000000000..963e17aa205 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c @@ -0,0 +1,537 @@ +/* + * acpi-cpufreq.c - ACPI Processor P-States Driver ($Revision: 1.3 $) + * + * Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com> + * Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com> + * Copyright (C) 2002 - 2004 Dominik Brodowski <linux@brodo.de> + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or (at + * your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA. + * + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + */ + +#include <linux/config.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/proc_fs.h> +#include <linux/seq_file.h> +#include <asm/io.h> +#include <asm/delay.h> +#include <asm/uaccess.h> + +#include <linux/acpi.h> +#include <acpi/processor.h> + +#include "speedstep-est-common.h" + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "acpi-cpufreq", msg) + +MODULE_AUTHOR("Paul Diefenbaugh, Dominik Brodowski"); +MODULE_DESCRIPTION("ACPI Processor P-States Driver"); +MODULE_LICENSE("GPL"); + + +struct cpufreq_acpi_io { + struct acpi_processor_performance acpi_data; + struct cpufreq_frequency_table *freq_table; + unsigned int resume; +}; + +static struct cpufreq_acpi_io *acpi_io_data[NR_CPUS]; + +static struct cpufreq_driver acpi_cpufreq_driver; + +static int +acpi_processor_write_port( + u16 port, + u8 bit_width, + u32 value) +{ + if (bit_width <= 8) { + outb(value, port); + } else if (bit_width <= 16) { + outw(value, port); + } else if (bit_width <= 32) { + outl(value, port); + } else { + return -ENODEV; + } + return 0; +} + +static int +acpi_processor_read_port( + u16 port, + u8 bit_width, + u32 *ret) +{ + *ret = 0; + if (bit_width <= 8) { + *ret = inb(port); + } else if (bit_width <= 16) { + *ret = inw(port); + } else if (bit_width <= 32) { + *ret = inl(port); + } else { + return -ENODEV; + } + return 0; +} + +static int +acpi_processor_set_performance ( + struct cpufreq_acpi_io *data, + unsigned int cpu, + int state) +{ + u16 port = 0; + u8 bit_width = 0; + int ret = 0; + u32 value = 0; + int i = 0; + struct cpufreq_freqs cpufreq_freqs; + cpumask_t saved_mask; + int retval; + + dprintk("acpi_processor_set_performance\n"); + + /* + * TBD: Use something other than set_cpus_allowed. + * As set_cpus_allowed is a bit racy, + * with any other set_cpus_allowed for this process. + */ + saved_mask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + if (smp_processor_id() != cpu) { + return (-EAGAIN); + } + + if (state == data->acpi_data.state) { + if (unlikely(data->resume)) { + dprintk("Called after resume, resetting to P%d\n", state); + data->resume = 0; + } else { + dprintk("Already at target state (P%d)\n", state); + retval = 0; + goto migrate_end; + } + } + + dprintk("Transitioning from P%d to P%d\n", + data->acpi_data.state, state); + + /* cpufreq frequency struct */ + cpufreq_freqs.cpu = cpu; + cpufreq_freqs.old = data->freq_table[data->acpi_data.state].frequency; + cpufreq_freqs.new = data->freq_table[state].frequency; + + /* notify cpufreq */ + cpufreq_notify_transition(&cpufreq_freqs, CPUFREQ_PRECHANGE); + + /* + * First we write the target state's 'control' value to the + * control_register. + */ + + port = data->acpi_data.control_register.address; + bit_width = data->acpi_data.control_register.bit_width; + value = (u32) data->acpi_data.states[state].control; + + dprintk("Writing 0x%08x to port 0x%04x\n", value, port); + + ret = acpi_processor_write_port(port, bit_width, value); + if (ret) { + dprintk("Invalid port width 0x%04x\n", bit_width); + retval = ret; + goto migrate_end; + } + + /* + * Then we read the 'status_register' and compare the value with the + * target state's 'status' to make sure the transition was successful. + * Note that we'll poll for up to 1ms (100 cycles of 10us) before + * giving up. + */ + + port = data->acpi_data.status_register.address; + bit_width = data->acpi_data.status_register.bit_width; + + dprintk("Looking for 0x%08x from port 0x%04x\n", + (u32) data->acpi_data.states[state].status, port); + + for (i=0; i<100; i++) { + ret = acpi_processor_read_port(port, bit_width, &value); + if (ret) { + dprintk("Invalid port width 0x%04x\n", bit_width); + retval = ret; + goto migrate_end; + } + if (value == (u32) data->acpi_data.states[state].status) + break; + udelay(10); + } + + /* notify cpufreq */ + cpufreq_notify_transition(&cpufreq_freqs, CPUFREQ_POSTCHANGE); + + if (value != (u32) data->acpi_data.states[state].status) { + unsigned int tmp = cpufreq_freqs.new; + cpufreq_freqs.new = cpufreq_freqs.old; + cpufreq_freqs.old = tmp; + cpufreq_notify_transition(&cpufreq_freqs, CPUFREQ_PRECHANGE); + cpufreq_notify_transition(&cpufreq_freqs, CPUFREQ_POSTCHANGE); + printk(KERN_WARNING "acpi-cpufreq: Transition failed\n"); + retval = -ENODEV; + goto migrate_end; + } + + dprintk("Transition successful after %d microseconds\n", i * 10); + + data->acpi_data.state = state; + + retval = 0; +migrate_end: + set_cpus_allowed(current, saved_mask); + return (retval); +} + + +static int +acpi_cpufreq_target ( + struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu]; + unsigned int next_state = 0; + unsigned int result = 0; + + dprintk("acpi_cpufreq_setpolicy\n"); + + result = cpufreq_frequency_table_target(policy, + data->freq_table, + target_freq, + relation, + &next_state); + if (result) + return (result); + + result = acpi_processor_set_performance (data, policy->cpu, next_state); + + return (result); +} + + +static int +acpi_cpufreq_verify ( + struct cpufreq_policy *policy) +{ + unsigned int result = 0; + struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu]; + + dprintk("acpi_cpufreq_verify\n"); + + result = cpufreq_frequency_table_verify(policy, + data->freq_table); + + return (result); +} + + +static unsigned long +acpi_cpufreq_guess_freq ( + struct cpufreq_acpi_io *data, + unsigned int cpu) +{ + if (cpu_khz) { + /* search the closest match to cpu_khz */ + unsigned int i; + unsigned long freq; + unsigned long freqn = data->acpi_data.states[0].core_frequency * 1000; + + for (i=0; i < (data->acpi_data.state_count - 1); i++) { + freq = freqn; + freqn = data->acpi_data.states[i+1].core_frequency * 1000; + if ((2 * cpu_khz) > (freqn + freq)) { + data->acpi_data.state = i; + return (freq); + } + } + data->acpi_data.state = data->acpi_data.state_count - 1; + return (freqn); + } else + /* assume CPU is at P0... */ + data->acpi_data.state = 0; + return data->acpi_data.states[0].core_frequency * 1000; + +} + + +/* + * acpi_processor_cpu_init_pdc_est - let BIOS know about the SMP capabilities + * of this driver + * @perf: processor-specific acpi_io_data struct + * @cpu: CPU being initialized + * + * To avoid issues with legacy OSes, some BIOSes require to be informed of + * the SMP capabilities of OS P-state driver. Here we set the bits in _PDC + * accordingly, for Enhanced Speedstep. Actual call to _PDC is done in + * driver/acpi/processor.c + */ +static void +acpi_processor_cpu_init_pdc_est( + struct acpi_processor_performance *perf, + unsigned int cpu, + struct acpi_object_list *obj_list + ) +{ + union acpi_object *obj; + u32 *buf; + struct cpuinfo_x86 *c = cpu_data + cpu; + dprintk("acpi_processor_cpu_init_pdc_est\n"); + + if (!cpu_has(c, X86_FEATURE_EST)) + return; + + /* Initialize pdc. It will be used later. */ + if (!obj_list) + return; + + if (!(obj_list->count && obj_list->pointer)) + return; + + obj = obj_list->pointer; + if ((obj->buffer.length == 12) && obj->buffer.pointer) { + buf = (u32 *)obj->buffer.pointer; + buf[0] = ACPI_PDC_REVISION_ID; + buf[1] = 1; + buf[2] = ACPI_PDC_EST_CAPABILITY_SMP; + perf->pdc = obj_list; + } + return; +} + + +/* CPU specific PDC initialization */ +static void +acpi_processor_cpu_init_pdc( + struct acpi_processor_performance *perf, + unsigned int cpu, + struct acpi_object_list *obj_list + ) +{ + struct cpuinfo_x86 *c = cpu_data + cpu; + dprintk("acpi_processor_cpu_init_pdc\n"); + perf->pdc = NULL; + if (cpu_has(c, X86_FEATURE_EST)) + acpi_processor_cpu_init_pdc_est(perf, cpu, obj_list); + return; +} + + +static int +acpi_cpufreq_cpu_init ( + struct cpufreq_policy *policy) +{ + unsigned int i; + unsigned int cpu = policy->cpu; + struct cpufreq_acpi_io *data; + unsigned int result = 0; + + union acpi_object arg0 = {ACPI_TYPE_BUFFER}; + u32 arg0_buf[3]; + struct acpi_object_list arg_list = {1, &arg0}; + + dprintk("acpi_cpufreq_cpu_init\n"); + /* setup arg_list for _PDC settings */ + arg0.buffer.length = 12; + arg0.buffer.pointer = (u8 *) arg0_buf; + + data = kmalloc(sizeof(struct cpufreq_acpi_io), GFP_KERNEL); + if (!data) + return (-ENOMEM); + memset(data, 0, sizeof(struct cpufreq_acpi_io)); + + acpi_io_data[cpu] = data; + + acpi_processor_cpu_init_pdc(&data->acpi_data, cpu, &arg_list); + result = acpi_processor_register_performance(&data->acpi_data, cpu); + data->acpi_data.pdc = NULL; + + if (result) + goto err_free; + + if (is_const_loops_cpu(cpu)) { + acpi_cpufreq_driver.flags |= CPUFREQ_CONST_LOOPS; + } + + /* capability check */ + if (data->acpi_data.state_count <= 1) { + dprintk("No P-States\n"); + result = -ENODEV; + goto err_unreg; + } + if ((data->acpi_data.control_register.space_id != ACPI_ADR_SPACE_SYSTEM_IO) || + (data->acpi_data.status_register.space_id != ACPI_ADR_SPACE_SYSTEM_IO)) { + dprintk("Unsupported address space [%d, %d]\n", + (u32) (data->acpi_data.control_register.space_id), + (u32) (data->acpi_data.status_register.space_id)); + result = -ENODEV; + goto err_unreg; + } + + /* alloc freq_table */ + data->freq_table = kmalloc(sizeof(struct cpufreq_frequency_table) * (data->acpi_data.state_count + 1), GFP_KERNEL); + if (!data->freq_table) { + result = -ENOMEM; + goto err_unreg; + } + + /* detect transition latency */ + policy->cpuinfo.transition_latency = 0; + for (i=0; i<data->acpi_data.state_count; i++) { + if ((data->acpi_data.states[i].transition_latency * 1000) > policy->cpuinfo.transition_latency) + policy->cpuinfo.transition_latency = data->acpi_data.states[i].transition_latency * 1000; + } + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + + /* The current speed is unknown and not detectable by ACPI... */ + policy->cur = acpi_cpufreq_guess_freq(data, policy->cpu); + + /* table init */ + for (i=0; i<=data->acpi_data.state_count; i++) + { + data->freq_table[i].index = i; + if (i<data->acpi_data.state_count) + data->freq_table[i].frequency = data->acpi_data.states[i].core_frequency * 1000; + else + data->freq_table[i].frequency = CPUFREQ_TABLE_END; + } + + result = cpufreq_frequency_table_cpuinfo(policy, data->freq_table); + if (result) { + goto err_freqfree; + } + + /* notify BIOS that we exist */ + acpi_processor_notify_smm(THIS_MODULE); + + printk(KERN_INFO "acpi-cpufreq: CPU%u - ACPI performance management activated.\n", + cpu); + for (i = 0; i < data->acpi_data.state_count; i++) + dprintk(" %cP%d: %d MHz, %d mW, %d uS\n", + (i == data->acpi_data.state?'*':' '), i, + (u32) data->acpi_data.states[i].core_frequency, + (u32) data->acpi_data.states[i].power, + (u32) data->acpi_data.states[i].transition_latency); + + cpufreq_frequency_table_get_attr(data->freq_table, policy->cpu); + return (result); + + err_freqfree: + kfree(data->freq_table); + err_unreg: + acpi_processor_unregister_performance(&data->acpi_data, cpu); + err_free: + kfree(data); + acpi_io_data[cpu] = NULL; + + return (result); +} + + +static int +acpi_cpufreq_cpu_exit ( + struct cpufreq_policy *policy) +{ + struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu]; + + + dprintk("acpi_cpufreq_cpu_exit\n"); + + if (data) { + cpufreq_frequency_table_put_attr(policy->cpu); + acpi_io_data[policy->cpu] = NULL; + acpi_processor_unregister_performance(&data->acpi_data, policy->cpu); + kfree(data); + } + + return (0); +} + +static int +acpi_cpufreq_resume ( + struct cpufreq_policy *policy) +{ + struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu]; + + + dprintk("acpi_cpufreq_resume\n"); + + data->resume = 1; + + return (0); +} + + +static struct freq_attr* acpi_cpufreq_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver acpi_cpufreq_driver = { + .verify = acpi_cpufreq_verify, + .target = acpi_cpufreq_target, + .init = acpi_cpufreq_cpu_init, + .exit = acpi_cpufreq_cpu_exit, + .resume = acpi_cpufreq_resume, + .name = "acpi-cpufreq", + .owner = THIS_MODULE, + .attr = acpi_cpufreq_attr, +}; + + +static int __init +acpi_cpufreq_init (void) +{ + int result = 0; + + dprintk("acpi_cpufreq_init\n"); + + result = cpufreq_register_driver(&acpi_cpufreq_driver); + + return (result); +} + + +static void __exit +acpi_cpufreq_exit (void) +{ + dprintk("acpi_cpufreq_exit\n"); + + cpufreq_unregister_driver(&acpi_cpufreq_driver); + + return; +} + + +late_initcall(acpi_cpufreq_init); +module_exit(acpi_cpufreq_exit); + +MODULE_ALIAS("acpi"); diff --git a/arch/i386/kernel/cpu/cpufreq/cpufreq-nforce2.c b/arch/i386/kernel/cpu/cpufreq/cpufreq-nforce2.c new file mode 100644 index 00000000000..04a40534520 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/cpufreq-nforce2.c @@ -0,0 +1,457 @@ +/* + * (C) 2004 Sebastian Witt <se.witt@gmx.net> + * + * Licensed under the terms of the GNU GPL License version 2. + * Based upon reverse engineered information + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/pci.h> +#include <linux/delay.h> + +#define NFORCE2_XTAL 25 +#define NFORCE2_BOOTFSB 0x48 +#define NFORCE2_PLLENABLE 0xa8 +#define NFORCE2_PLLREG 0xa4 +#define NFORCE2_PLLADR 0xa0 +#define NFORCE2_PLL(mul, div) (0x100000 | (mul << 8) | div) + +#define NFORCE2_MIN_FSB 50 +#define NFORCE2_SAFE_DISTANCE 50 + +/* Delay in ms between FSB changes */ +//#define NFORCE2_DELAY 10 + +/* nforce2_chipset: + * FSB is changed using the chipset + */ +static struct pci_dev *nforce2_chipset_dev; + +/* fid: + * multiplier * 10 + */ +static int fid = 0; + +/* min_fsb, max_fsb: + * minimum and maximum FSB (= FSB at boot time) + */ +static int min_fsb = 0; +static int max_fsb = 0; + +MODULE_AUTHOR("Sebastian Witt <se.witt@gmx.net>"); +MODULE_DESCRIPTION("nForce2 FSB changing cpufreq driver"); +MODULE_LICENSE("GPL"); + +module_param(fid, int, 0444); +module_param(min_fsb, int, 0444); + +MODULE_PARM_DESC(fid, "CPU multiplier to use (11.5 = 115)"); +MODULE_PARM_DESC(min_fsb, + "Minimum FSB to use, if not defined: current FSB - 50"); + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "cpufreq-nforce2", msg) + +/* + * nforce2_calc_fsb - calculate FSB + * @pll: PLL value + * + * Calculates FSB from PLL value + */ +static int nforce2_calc_fsb(int pll) +{ + unsigned char mul, div; + + mul = (pll >> 8) & 0xff; + div = pll & 0xff; + + if (div > 0) + return NFORCE2_XTAL * mul / div; + + return 0; +} + +/* + * nforce2_calc_pll - calculate PLL value + * @fsb: FSB + * + * Calculate PLL value for given FSB + */ +static int nforce2_calc_pll(unsigned int fsb) +{ + unsigned char xmul, xdiv; + unsigned char mul = 0, div = 0; + int tried = 0; + + /* Try to calculate multiplier and divider up to 4 times */ + while (((mul == 0) || (div == 0)) && (tried <= 3)) { + for (xdiv = 1; xdiv <= 0x80; xdiv++) + for (xmul = 1; xmul <= 0xfe; xmul++) + if (nforce2_calc_fsb(NFORCE2_PLL(xmul, xdiv)) == + fsb + tried) { + mul = xmul; + div = xdiv; + } + tried++; + } + + if ((mul == 0) || (div == 0)) + return -1; + + return NFORCE2_PLL(mul, div); +} + +/* + * nforce2_write_pll - write PLL value to chipset + * @pll: PLL value + * + * Writes new FSB PLL value to chipset + */ +static void nforce2_write_pll(int pll) +{ + int temp; + + /* Set the pll addr. to 0x00 */ + temp = 0x00; + pci_write_config_dword(nforce2_chipset_dev, NFORCE2_PLLADR, temp); + + /* Now write the value in all 64 registers */ + for (temp = 0; temp <= 0x3f; temp++) { + pci_write_config_dword(nforce2_chipset_dev, + NFORCE2_PLLREG, pll); + } + + return; +} + +/* + * nforce2_fsb_read - Read FSB + * + * Read FSB from chipset + * If bootfsb != 0, return FSB at boot-time + */ +static unsigned int nforce2_fsb_read(int bootfsb) +{ + struct pci_dev *nforce2_sub5; + u32 fsb, temp = 0; + + + /* Get chipset boot FSB from subdevice 5 (FSB at boot-time) */ + nforce2_sub5 = pci_get_subsys(PCI_VENDOR_ID_NVIDIA, + 0x01EF, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + + if (!nforce2_sub5) + return 0; + + pci_read_config_dword(nforce2_sub5, NFORCE2_BOOTFSB, &fsb); + fsb /= 1000000; + + /* Check if PLL register is already set */ + pci_read_config_byte(nforce2_chipset_dev, + NFORCE2_PLLENABLE, (u8 *)&temp); + + if(bootfsb || !temp) + return fsb; + + /* Use PLL register FSB value */ + pci_read_config_dword(nforce2_chipset_dev, + NFORCE2_PLLREG, &temp); + fsb = nforce2_calc_fsb(temp); + + return fsb; +} + +/* + * nforce2_set_fsb - set new FSB + * @fsb: New FSB + * + * Sets new FSB + */ +static int nforce2_set_fsb(unsigned int fsb) +{ + u32 pll, temp = 0; + unsigned int tfsb; + int diff; + + if ((fsb > max_fsb) || (fsb < NFORCE2_MIN_FSB)) { + printk(KERN_ERR "cpufreq: FSB %d is out of range!\n", fsb); + return -EINVAL; + } + + tfsb = nforce2_fsb_read(0); + if (!tfsb) { + printk(KERN_ERR "cpufreq: Error while reading the FSB\n"); + return -EINVAL; + } + + /* First write? Then set actual value */ + pci_read_config_byte(nforce2_chipset_dev, + NFORCE2_PLLENABLE, (u8 *)&temp); + if (!temp) { + pll = nforce2_calc_pll(tfsb); + + if (pll < 0) + return -EINVAL; + + nforce2_write_pll(pll); + } + + /* Enable write access */ + temp = 0x01; + pci_write_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8)temp); + + diff = tfsb - fsb; + + if (!diff) + return 0; + + while ((tfsb != fsb) && (tfsb <= max_fsb) && (tfsb >= min_fsb)) { + if (diff < 0) + tfsb++; + else + tfsb--; + + /* Calculate the PLL reg. value */ + if ((pll = nforce2_calc_pll(tfsb)) == -1) + return -EINVAL; + + nforce2_write_pll(pll); +#ifdef NFORCE2_DELAY + mdelay(NFORCE2_DELAY); +#endif + } + + temp = 0x40; + pci_write_config_byte(nforce2_chipset_dev, NFORCE2_PLLADR, (u8)temp); + + return 0; +} + +/** + * nforce2_get - get the CPU frequency + * @cpu: CPU number + * + * Returns the CPU frequency + */ +static unsigned int nforce2_get(unsigned int cpu) +{ + if (cpu) + return 0; + return nforce2_fsb_read(0) * fid * 100; +} + +/** + * nforce2_target - set a new CPUFreq policy + * @policy: new policy + * @target_freq: the target frequency + * @relation: how that frequency relates to achieved frequency (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H) + * + * Sets a new CPUFreq policy. + */ +static int nforce2_target(struct cpufreq_policy *policy, + unsigned int target_freq, unsigned int relation) +{ +// unsigned long flags; + struct cpufreq_freqs freqs; + unsigned int target_fsb; + + if ((target_freq > policy->max) || (target_freq < policy->min)) + return -EINVAL; + + target_fsb = target_freq / (fid * 100); + + freqs.old = nforce2_get(policy->cpu); + freqs.new = target_fsb * fid * 100; + freqs.cpu = 0; /* Only one CPU on nForce2 plattforms */ + + if (freqs.old == freqs.new) + return 0; + + dprintk(KERN_INFO "cpufreq: Old CPU frequency %d kHz, new %d kHz\n", + freqs.old, freqs.new); + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + /* Disable IRQs */ + //local_irq_save(flags); + + if (nforce2_set_fsb(target_fsb) < 0) + printk(KERN_ERR "cpufreq: Changing FSB to %d failed\n", + target_fsb); + else + dprintk(KERN_INFO "cpufreq: Changed FSB successfully to %d\n", + target_fsb); + + /* Enable IRQs */ + //local_irq_restore(flags); + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + return 0; +} + +/** + * nforce2_verify - verifies a new CPUFreq policy + * @policy: new policy + */ +static int nforce2_verify(struct cpufreq_policy *policy) +{ + unsigned int fsb_pol_max; + + fsb_pol_max = policy->max / (fid * 100); + + if (policy->min < (fsb_pol_max * fid * 100)) + policy->max = (fsb_pol_max + 1) * fid * 100; + + cpufreq_verify_within_limits(policy, + policy->cpuinfo.min_freq, + policy->cpuinfo.max_freq); + return 0; +} + +static int nforce2_cpu_init(struct cpufreq_policy *policy) +{ + unsigned int fsb; + unsigned int rfid; + + /* capability check */ + if (policy->cpu != 0) + return -ENODEV; + + /* Get current FSB */ + fsb = nforce2_fsb_read(0); + + if (!fsb) + return -EIO; + + /* FIX: Get FID from CPU */ + if (!fid) { + if (!cpu_khz) { + printk(KERN_WARNING + "cpufreq: cpu_khz not set, can't calculate multiplier!\n"); + return -ENODEV; + } + + fid = cpu_khz / (fsb * 100); + rfid = fid % 5; + + if (rfid) { + if (rfid > 2) + fid += 5 - rfid; + else + fid -= rfid; + } + } + + printk(KERN_INFO "cpufreq: FSB currently at %i MHz, FID %d.%d\n", fsb, + fid / 10, fid % 10); + + /* Set maximum FSB to FSB at boot time */ + max_fsb = nforce2_fsb_read(1); + + if(!max_fsb) + return -EIO; + + if (!min_fsb) + min_fsb = max_fsb - NFORCE2_SAFE_DISTANCE; + + if (min_fsb < NFORCE2_MIN_FSB) + min_fsb = NFORCE2_MIN_FSB; + + /* cpuinfo and default policy values */ + policy->cpuinfo.min_freq = min_fsb * fid * 100; + policy->cpuinfo.max_freq = max_fsb * fid * 100; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = nforce2_get(policy->cpu); + policy->min = policy->cpuinfo.min_freq; + policy->max = policy->cpuinfo.max_freq; + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + + return 0; +} + +static int nforce2_cpu_exit(struct cpufreq_policy *policy) +{ + return 0; +} + +static struct cpufreq_driver nforce2_driver = { + .name = "nforce2", + .verify = nforce2_verify, + .target = nforce2_target, + .get = nforce2_get, + .init = nforce2_cpu_init, + .exit = nforce2_cpu_exit, + .owner = THIS_MODULE, +}; + +/** + * nforce2_detect_chipset - detect the Southbridge which contains FSB PLL logic + * + * Detects nForce2 A2 and C1 stepping + * + */ +static unsigned int nforce2_detect_chipset(void) +{ + u8 revision; + + nforce2_chipset_dev = pci_get_subsys(PCI_VENDOR_ID_NVIDIA, + PCI_DEVICE_ID_NVIDIA_NFORCE2, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + + if (nforce2_chipset_dev == NULL) + return -ENODEV; + + pci_read_config_byte(nforce2_chipset_dev, PCI_REVISION_ID, &revision); + + printk(KERN_INFO "cpufreq: Detected nForce2 chipset revision %X\n", + revision); + printk(KERN_INFO + "cpufreq: FSB changing is maybe unstable and can lead to crashes and data loss.\n"); + + return 0; +} + +/** + * nforce2_init - initializes the nForce2 CPUFreq driver + * + * Initializes the nForce2 FSB support. Returns -ENODEV on unsupported + * devices, -EINVAL on problems during initiatization, and zero on + * success. + */ +static int __init nforce2_init(void) +{ + /* TODO: do we need to detect the processor? */ + + /* detect chipset */ + if (nforce2_detect_chipset()) { + printk(KERN_ERR "cpufreq: No nForce2 chipset.\n"); + return -ENODEV; + } + + return cpufreq_register_driver(&nforce2_driver); +} + +/** + * nforce2_exit - unregisters cpufreq module + * + * Unregisters nForce2 FSB change support. + */ +static void __exit nforce2_exit(void) +{ + cpufreq_unregister_driver(&nforce2_driver); +} + +module_init(nforce2_init); +module_exit(nforce2_exit); + diff --git a/arch/i386/kernel/cpu/cpufreq/elanfreq.c b/arch/i386/kernel/cpu/cpufreq/elanfreq.c new file mode 100644 index 00000000000..3f7caa4ae6d --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/elanfreq.c @@ -0,0 +1,312 @@ +/* + * elanfreq: cpufreq driver for the AMD ELAN family + * + * (c) Copyright 2002 Robert Schwebel <r.schwebel@pengutronix.de> + * + * Parts of this code are (c) Sven Geggus <sven@geggus.net> + * + * All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * 2002-02-13: - initial revision for 2.4.18-pre9 by Robert Schwebel + * + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> + +#include <linux/slab.h> +#include <linux/delay.h> +#include <linux/cpufreq.h> + +#include <asm/msr.h> +#include <asm/timex.h> +#include <asm/io.h> + +#define REG_CSCIR 0x22 /* Chip Setup and Control Index Register */ +#define REG_CSCDR 0x23 /* Chip Setup and Control Data Register */ + +/* Module parameter */ +static int max_freq; + +struct s_elan_multiplier { + int clock; /* frequency in kHz */ + int val40h; /* PMU Force Mode register */ + int val80h; /* CPU Clock Speed Register */ +}; + +/* + * It is important that the frequencies + * are listed in ascending order here! + */ +struct s_elan_multiplier elan_multiplier[] = { + {1000, 0x02, 0x18}, + {2000, 0x02, 0x10}, + {4000, 0x02, 0x08}, + {8000, 0x00, 0x00}, + {16000, 0x00, 0x02}, + {33000, 0x00, 0x04}, + {66000, 0x01, 0x04}, + {99000, 0x01, 0x05} +}; + +static struct cpufreq_frequency_table elanfreq_table[] = { + {0, 1000}, + {1, 2000}, + {2, 4000}, + {3, 8000}, + {4, 16000}, + {5, 33000}, + {6, 66000}, + {7, 99000}, + {0, CPUFREQ_TABLE_END}, +}; + + +/** + * elanfreq_get_cpu_frequency: determine current cpu speed + * + * Finds out at which frequency the CPU of the Elan SOC runs + * at the moment. Frequencies from 1 to 33 MHz are generated + * the normal way, 66 and 99 MHz are called "Hyperspeed Mode" + * and have the rest of the chip running with 33 MHz. + */ + +static unsigned int elanfreq_get_cpu_frequency(unsigned int cpu) +{ + u8 clockspeed_reg; /* Clock Speed Register */ + + local_irq_disable(); + outb_p(0x80,REG_CSCIR); + clockspeed_reg = inb_p(REG_CSCDR); + local_irq_enable(); + + if ((clockspeed_reg & 0xE0) == 0xE0) { return 0; } + + /* Are we in CPU clock multiplied mode (66/99 MHz)? */ + if ((clockspeed_reg & 0xE0) == 0xC0) { + if ((clockspeed_reg & 0x01) == 0) { + return 66000; + } else { + return 99000; + } + } + + /* 33 MHz is not 32 MHz... */ + if ((clockspeed_reg & 0xE0)==0xA0) + return 33000; + + return ((1<<((clockspeed_reg & 0xE0) >> 5)) * 1000); +} + + +/** + * elanfreq_set_cpu_frequency: Change the CPU core frequency + * @cpu: cpu number + * @freq: frequency in kHz + * + * This function takes a frequency value and changes the CPU frequency + * according to this. Note that the frequency has to be checked by + * elanfreq_validatespeed() for correctness! + * + * There is no return value. + */ + +static void elanfreq_set_cpu_state (unsigned int state) { + + struct cpufreq_freqs freqs; + + freqs.old = elanfreq_get_cpu_frequency(0); + freqs.new = elan_multiplier[state].clock; + freqs.cpu = 0; /* elanfreq.c is UP only driver */ + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + printk(KERN_INFO "elanfreq: attempting to set frequency to %i kHz\n",elan_multiplier[state].clock); + + + /* + * Access to the Elan's internal registers is indexed via + * 0x22: Chip Setup & Control Register Index Register (CSCI) + * 0x23: Chip Setup & Control Register Data Register (CSCD) + * + */ + + /* + * 0x40 is the Power Management Unit's Force Mode Register. + * Bit 6 enables Hyperspeed Mode (66/100 MHz core frequency) + */ + + local_irq_disable(); + outb_p(0x40,REG_CSCIR); /* Disable hyperspeed mode */ + outb_p(0x00,REG_CSCDR); + local_irq_enable(); /* wait till internal pipelines and */ + udelay(1000); /* buffers have cleaned up */ + + local_irq_disable(); + + /* now, set the CPU clock speed register (0x80) */ + outb_p(0x80,REG_CSCIR); + outb_p(elan_multiplier[state].val80h,REG_CSCDR); + + /* now, the hyperspeed bit in PMU Force Mode Register (0x40) */ + outb_p(0x40,REG_CSCIR); + outb_p(elan_multiplier[state].val40h,REG_CSCDR); + udelay(10000); + local_irq_enable(); + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); +}; + + +/** + * elanfreq_validatespeed: test if frequency range is valid + * @policy: the policy to validate + * + * This function checks if a given frequency range in kHz is valid + * for the hardware supported by the driver. + */ + +static int elanfreq_verify (struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, &elanfreq_table[0]); +} + +static int elanfreq_target (struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = 0; + + if (cpufreq_frequency_table_target(policy, &elanfreq_table[0], target_freq, relation, &newstate)) + return -EINVAL; + + elanfreq_set_cpu_state(newstate); + + return 0; +} + + +/* + * Module init and exit code + */ + +static int elanfreq_cpu_init(struct cpufreq_policy *policy) +{ + struct cpuinfo_x86 *c = cpu_data; + unsigned int i; + int result; + + /* capability check */ + if ((c->x86_vendor != X86_VENDOR_AMD) || + (c->x86 != 4) || (c->x86_model!=10)) + return -ENODEV; + + /* max freq */ + if (!max_freq) + max_freq = elanfreq_get_cpu_frequency(0); + + /* table init */ + for (i=0; (elanfreq_table[i].frequency != CPUFREQ_TABLE_END); i++) { + if (elanfreq_table[i].frequency > max_freq) + elanfreq_table[i].frequency = CPUFREQ_ENTRY_INVALID; + } + + /* cpuinfo and default policy values */ + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = elanfreq_get_cpu_frequency(0); + + result = cpufreq_frequency_table_cpuinfo(policy, elanfreq_table); + if (result) + return (result); + + cpufreq_frequency_table_get_attr(elanfreq_table, policy->cpu); + + return 0; +} + + +static int elanfreq_cpu_exit(struct cpufreq_policy *policy) +{ + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + + +#ifndef MODULE +/** + * elanfreq_setup - elanfreq command line parameter parsing + * + * elanfreq command line parameter. Use: + * elanfreq=66000 + * to set the maximum CPU frequency to 66 MHz. Note that in + * case you do not give this boot parameter, the maximum + * frequency will fall back to _current_ CPU frequency which + * might be lower. If you build this as a module, use the + * max_freq module parameter instead. + */ +static int __init elanfreq_setup(char *str) +{ + max_freq = simple_strtoul(str, &str, 0); + printk(KERN_WARNING "You're using the deprecated elanfreq command line option. Use elanfreq.max_freq instead, please!\n"); + return 1; +} +__setup("elanfreq=", elanfreq_setup); +#endif + + +static struct freq_attr* elanfreq_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + + +static struct cpufreq_driver elanfreq_driver = { + .get = elanfreq_get_cpu_frequency, + .verify = elanfreq_verify, + .target = elanfreq_target, + .init = elanfreq_cpu_init, + .exit = elanfreq_cpu_exit, + .name = "elanfreq", + .owner = THIS_MODULE, + .attr = elanfreq_attr, +}; + + +static int __init elanfreq_init(void) +{ + struct cpuinfo_x86 *c = cpu_data; + + /* Test if we have the right hardware */ + if ((c->x86_vendor != X86_VENDOR_AMD) || + (c->x86 != 4) || (c->x86_model!=10)) + { + printk(KERN_INFO "elanfreq: error: no Elan processor found!\n"); + return -ENODEV; + } + + return cpufreq_register_driver(&elanfreq_driver); +} + + +static void __exit elanfreq_exit(void) +{ + cpufreq_unregister_driver(&elanfreq_driver); +} + + +module_param (max_freq, int, 0444); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Robert Schwebel <r.schwebel@pengutronix.de>, Sven Geggus <sven@geggus.net>"); +MODULE_DESCRIPTION("cpufreq driver for AMD's Elan CPUs"); + +module_init(elanfreq_init); +module_exit(elanfreq_exit); + diff --git a/arch/i386/kernel/cpu/cpufreq/gx-suspmod.c b/arch/i386/kernel/cpu/cpufreq/gx-suspmod.c new file mode 100644 index 00000000000..1a49adb1f4a --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/gx-suspmod.c @@ -0,0 +1,502 @@ +/* + * Cyrix MediaGX and NatSemi Geode Suspend Modulation + * (C) 2002 Zwane Mwaikambo <zwane@commfireservices.com> + * (C) 2002 Hiroshi Miura <miura@da-cha.org> + * All Rights Reserved + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * version 2 as published by the Free Software Foundation + * + * The author(s) of this software shall not be held liable for damages + * of any nature resulting due to the use of this software. This + * software is provided AS-IS with no warranties. + * + * Theoritical note: + * + * (see Geode(tm) CS5530 manual (rev.4.1) page.56) + * + * CPU frequency control on NatSemi Geode GX1/GXLV processor and CS55x0 + * are based on Suspend Moduration. + * + * Suspend Modulation works by asserting and de-asserting the SUSP# pin + * to CPU(GX1/GXLV) for configurable durations. When asserting SUSP# + * the CPU enters an idle state. GX1 stops its core clock when SUSP# is + * asserted then power consumption is reduced. + * + * Suspend Modulation's OFF/ON duration are configurable + * with 'Suspend Modulation OFF Count Register' + * and 'Suspend Modulation ON Count Register'. + * These registers are 8bit counters that represent the number of + * 32us intervals which the SUSP# pin is asserted(ON)/de-asserted(OFF) + * to the processor. + * + * These counters define a ratio which is the effective frequency + * of operation of the system. + * + * OFF Count + * F_eff = Fgx * ---------------------- + * OFF Count + ON Count + * + * 0 <= On Count, Off Count <= 255 + * + * From these limits, we can get register values + * + * off_duration + on_duration <= MAX_DURATION + * on_duration = off_duration * (stock_freq - freq) / freq + * + * off_duration = (freq * DURATION) / stock_freq + * on_duration = DURATION - off_duration + * + * + *--------------------------------------------------------------------------- + * + * ChangeLog: + * Dec. 12, 2003 Hiroshi Miura <miura@da-cha.org> + * - fix on/off register mistake + * - fix cpu_khz calc when it stops cpu modulation. + * + * Dec. 11, 2002 Hiroshi Miura <miura@da-cha.org> + * - rewrite for Cyrix MediaGX Cx5510/5520 and + * NatSemi Geode Cs5530(A). + * + * Jul. ??, 2002 Zwane Mwaikambo <zwane@commfireservices.com> + * - cs5530_mod patch for 2.4.19-rc1. + * + *--------------------------------------------------------------------------- + * + * Todo + * Test on machines with 5510, 5530, 5530A + */ + +/************************************************************************ + * Suspend Modulation - Definitions * + ************************************************************************/ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/smp.h> +#include <linux/cpufreq.h> +#include <linux/pci.h> +#include <asm/processor.h> +#include <asm/errno.h> + +/* PCI config registers, all at F0 */ +#define PCI_PMER1 0x80 /* power management enable register 1 */ +#define PCI_PMER2 0x81 /* power management enable register 2 */ +#define PCI_PMER3 0x82 /* power management enable register 3 */ +#define PCI_IRQTC 0x8c /* irq speedup timer counter register:typical 2 to 4ms */ +#define PCI_VIDTC 0x8d /* video speedup timer counter register: typical 50 to 100ms */ +#define PCI_MODOFF 0x94 /* suspend modulation OFF counter register, 1 = 32us */ +#define PCI_MODON 0x95 /* suspend modulation ON counter register */ +#define PCI_SUSCFG 0x96 /* suspend configuration register */ + +/* PMER1 bits */ +#define GPM (1<<0) /* global power management */ +#define GIT (1<<1) /* globally enable PM device idle timers */ +#define GTR (1<<2) /* globally enable IO traps */ +#define IRQ_SPDUP (1<<3) /* disable clock throttle during interrupt handling */ +#define VID_SPDUP (1<<4) /* disable clock throttle during vga video handling */ + +/* SUSCFG bits */ +#define SUSMOD (1<<0) /* enable/disable suspend modulation */ +/* the belows support only with cs5530 (after rev.1.2)/cs5530A */ +#define SMISPDUP (1<<1) /* select how SMI re-enable suspend modulation: */ + /* IRQTC timer or read SMI speedup disable reg.(F1BAR[08-09h]) */ +#define SUSCFG (1<<2) /* enable powering down a GXLV processor. "Special 3Volt Suspend" mode */ +/* the belows support only with cs5530A */ +#define PWRSVE_ISA (1<<3) /* stop ISA clock */ +#define PWRSVE (1<<4) /* active idle */ + +struct gxfreq_params { + u8 on_duration; + u8 off_duration; + u8 pci_suscfg; + u8 pci_pmer1; + u8 pci_pmer2; + u8 pci_rev; + struct pci_dev *cs55x0; +}; + +static struct gxfreq_params *gx_params; +static int stock_freq; + +/* PCI bus clock - defaults to 30.000 if cpu_khz is not available */ +static int pci_busclk = 0; +module_param (pci_busclk, int, 0444); + +/* maximum duration for which the cpu may be suspended + * (32us * MAX_DURATION). If no parameter is given, this defaults + * to 255. + * Note that this leads to a maximum of 8 ms(!) where the CPU clock + * is suspended -- processing power is just 0.39% of what it used to be, + * though. 781.25 kHz(!) for a 200 MHz processor -- wow. */ +static int max_duration = 255; +module_param (max_duration, int, 0444); + +/* For the default policy, we want at least some processing power + * - let's say 5%. (min = maxfreq / POLICY_MIN_DIV) + */ +#define POLICY_MIN_DIV 20 + + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "gx-suspmod", msg) + +/** + * we can detect a core multipiler from dir0_lsb + * from GX1 datasheet p.56, + * MULT[3:0]: + * 0000 = SYSCLK multiplied by 4 (test only) + * 0001 = SYSCLK multiplied by 10 + * 0010 = SYSCLK multiplied by 4 + * 0011 = SYSCLK multiplied by 6 + * 0100 = SYSCLK multiplied by 9 + * 0101 = SYSCLK multiplied by 5 + * 0110 = SYSCLK multiplied by 7 + * 0111 = SYSCLK multiplied by 8 + * of 33.3MHz + **/ +static int gx_freq_mult[16] = { + 4, 10, 4, 6, 9, 5, 7, 8, + 0, 0, 0, 0, 0, 0, 0, 0 +}; + + +/**************************************************************** + * Low Level chipset interface * + ****************************************************************/ +static struct pci_device_id gx_chipset_tbl[] __initdata = { + { PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5530_LEGACY, PCI_ANY_ID, PCI_ANY_ID }, + { PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5520, PCI_ANY_ID, PCI_ANY_ID }, + { PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5510, PCI_ANY_ID, PCI_ANY_ID }, + { 0, }, +}; + +/** + * gx_detect_chipset: + * + **/ +static __init struct pci_dev *gx_detect_chipset(void) +{ + struct pci_dev *gx_pci = NULL; + + /* check if CPU is a MediaGX or a Geode. */ + if ((current_cpu_data.x86_vendor != X86_VENDOR_NSC) && + (current_cpu_data.x86_vendor != X86_VENDOR_CYRIX)) { + dprintk("error: no MediaGX/Geode processor found!\n"); + return NULL; + } + + /* detect which companion chip is used */ + while ((gx_pci = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, gx_pci)) != NULL) { + if ((pci_match_device (gx_chipset_tbl, gx_pci)) != NULL) { + return gx_pci; + } + } + + dprintk("error: no supported chipset found!\n"); + return NULL; +} + +/** + * gx_get_cpuspeed: + * + * Finds out at which efficient frequency the Cyrix MediaGX/NatSemi Geode CPU runs. + */ +static unsigned int gx_get_cpuspeed(unsigned int cpu) +{ + if ((gx_params->pci_suscfg & SUSMOD) == 0) + return stock_freq; + + return (stock_freq * gx_params->off_duration) + / (gx_params->on_duration + gx_params->off_duration); +} + +/** + * gx_validate_speed: + * determine current cpu speed + * +**/ + +static unsigned int gx_validate_speed(unsigned int khz, u8 *on_duration, u8 *off_duration) +{ + unsigned int i; + u8 tmp_on, tmp_off; + int old_tmp_freq = stock_freq; + int tmp_freq; + + *off_duration=1; + *on_duration=0; + + for (i=max_duration; i>0; i--) { + tmp_off = ((khz * i) / stock_freq) & 0xff; + tmp_on = i - tmp_off; + tmp_freq = (stock_freq * tmp_off) / i; + /* if this relation is closer to khz, use this. If it's equal, + * prefer it, too - lower latency */ + if (abs(tmp_freq - khz) <= abs(old_tmp_freq - khz)) { + *on_duration = tmp_on; + *off_duration = tmp_off; + old_tmp_freq = tmp_freq; + } + } + + return old_tmp_freq; +} + + +/** + * gx_set_cpuspeed: + * set cpu speed in khz. + **/ + +static void gx_set_cpuspeed(unsigned int khz) +{ + u8 suscfg, pmer1; + unsigned int new_khz; + unsigned long flags; + struct cpufreq_freqs freqs; + + + freqs.cpu = 0; + freqs.old = gx_get_cpuspeed(0); + + new_khz = gx_validate_speed(khz, &gx_params->on_duration, &gx_params->off_duration); + + freqs.new = new_khz; + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + local_irq_save(flags); + + if (new_khz != stock_freq) { /* if new khz == 100% of CPU speed, it is special case */ + switch (gx_params->cs55x0->device) { + case PCI_DEVICE_ID_CYRIX_5530_LEGACY: + pmer1 = gx_params->pci_pmer1 | IRQ_SPDUP | VID_SPDUP; + /* FIXME: need to test other values -- Zwane,Miura */ + pci_write_config_byte(gx_params->cs55x0, PCI_IRQTC, 4); /* typical 2 to 4ms */ + pci_write_config_byte(gx_params->cs55x0, PCI_VIDTC, 100);/* typical 50 to 100ms */ + pci_write_config_byte(gx_params->cs55x0, PCI_PMER1, pmer1); + + if (gx_params->pci_rev < 0x10) { /* CS5530(rev 1.2, 1.3) */ + suscfg = gx_params->pci_suscfg | SUSMOD; + } else { /* CS5530A,B.. */ + suscfg = gx_params->pci_suscfg | SUSMOD | PWRSVE; + } + break; + case PCI_DEVICE_ID_CYRIX_5520: + case PCI_DEVICE_ID_CYRIX_5510: + suscfg = gx_params->pci_suscfg | SUSMOD; + break; + default: + local_irq_restore(flags); + dprintk("fatal: try to set unknown chipset.\n"); + return; + } + } else { + suscfg = gx_params->pci_suscfg & ~(SUSMOD); + gx_params->off_duration = 0; + gx_params->on_duration = 0; + dprintk("suspend modulation disabled: cpu runs 100 percent speed.\n"); + } + + pci_write_config_byte(gx_params->cs55x0, PCI_MODOFF, gx_params->off_duration); + pci_write_config_byte(gx_params->cs55x0, PCI_MODON, gx_params->on_duration); + + pci_write_config_byte(gx_params->cs55x0, PCI_SUSCFG, suscfg); + pci_read_config_byte(gx_params->cs55x0, PCI_SUSCFG, &suscfg); + + local_irq_restore(flags); + + gx_params->pci_suscfg = suscfg; + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + dprintk("suspend modulation w/ duration of ON:%d us, OFF:%d us\n", + gx_params->on_duration * 32, gx_params->off_duration * 32); + dprintk("suspend modulation w/ clock speed: %d kHz.\n", freqs.new); +} + +/**************************************************************** + * High level functions * + ****************************************************************/ + +/* + * cpufreq_gx_verify: test if frequency range is valid + * + * This function checks if a given frequency range in kHz is valid + * for the hardware supported by the driver. + */ + +static int cpufreq_gx_verify(struct cpufreq_policy *policy) +{ + unsigned int tmp_freq = 0; + u8 tmp1, tmp2; + + if (!stock_freq || !policy) + return -EINVAL; + + policy->cpu = 0; + cpufreq_verify_within_limits(policy, (stock_freq / max_duration), stock_freq); + + /* it needs to be assured that at least one supported frequency is + * within policy->min and policy->max. If it is not, policy->max + * needs to be increased until one freuqency is supported. + * policy->min may not be decreased, though. This way we guarantee a + * specific processing capacity. + */ + tmp_freq = gx_validate_speed(policy->min, &tmp1, &tmp2); + if (tmp_freq < policy->min) + tmp_freq += stock_freq / max_duration; + policy->min = tmp_freq; + if (policy->min > policy->max) + policy->max = tmp_freq; + tmp_freq = gx_validate_speed(policy->max, &tmp1, &tmp2); + if (tmp_freq > policy->max) + tmp_freq -= stock_freq / max_duration; + policy->max = tmp_freq; + if (policy->max < policy->min) + policy->max = policy->min; + cpufreq_verify_within_limits(policy, (stock_freq / max_duration), stock_freq); + + return 0; +} + +/* + * cpufreq_gx_target: + * + */ +static int cpufreq_gx_target(struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + u8 tmp1, tmp2; + unsigned int tmp_freq; + + if (!stock_freq || !policy) + return -EINVAL; + + policy->cpu = 0; + + tmp_freq = gx_validate_speed(target_freq, &tmp1, &tmp2); + while (tmp_freq < policy->min) { + tmp_freq += stock_freq / max_duration; + tmp_freq = gx_validate_speed(tmp_freq, &tmp1, &tmp2); + } + while (tmp_freq > policy->max) { + tmp_freq -= stock_freq / max_duration; + tmp_freq = gx_validate_speed(tmp_freq, &tmp1, &tmp2); + } + + gx_set_cpuspeed(tmp_freq); + + return 0; +} + +static int cpufreq_gx_cpu_init(struct cpufreq_policy *policy) +{ + unsigned int maxfreq, curfreq; + + if (!policy || policy->cpu != 0) + return -ENODEV; + + /* determine maximum frequency */ + if (pci_busclk) { + maxfreq = pci_busclk * gx_freq_mult[getCx86(CX86_DIR1) & 0x0f]; + } else if (cpu_khz) { + maxfreq = cpu_khz; + } else { + maxfreq = 30000 * gx_freq_mult[getCx86(CX86_DIR1) & 0x0f]; + } + stock_freq = maxfreq; + curfreq = gx_get_cpuspeed(0); + + dprintk("cpu max frequency is %d.\n", maxfreq); + dprintk("cpu current frequency is %dkHz.\n",curfreq); + + /* setup basic struct for cpufreq API */ + policy->cpu = 0; + + if (max_duration < POLICY_MIN_DIV) + policy->min = maxfreq / max_duration; + else + policy->min = maxfreq / POLICY_MIN_DIV; + policy->max = maxfreq; + policy->cur = curfreq; + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.min_freq = maxfreq / max_duration; + policy->cpuinfo.max_freq = maxfreq; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + + return 0; +} + +/* + * cpufreq_gx_init: + * MediaGX/Geode GX initialize cpufreq driver + */ +static struct cpufreq_driver gx_suspmod_driver = { + .get = gx_get_cpuspeed, + .verify = cpufreq_gx_verify, + .target = cpufreq_gx_target, + .init = cpufreq_gx_cpu_init, + .name = "gx-suspmod", + .owner = THIS_MODULE, +}; + +static int __init cpufreq_gx_init(void) +{ + int ret; + struct gxfreq_params *params; + struct pci_dev *gx_pci; + u32 class_rev; + + /* Test if we have the right hardware */ + if ((gx_pci = gx_detect_chipset()) == NULL) + return -ENODEV; + + /* check whether module parameters are sane */ + if (max_duration > 0xff) + max_duration = 0xff; + + dprintk("geode suspend modulation available.\n"); + + params = kmalloc(sizeof(struct gxfreq_params), GFP_KERNEL); + if (params == NULL) + return -ENOMEM; + memset(params, 0, sizeof(struct gxfreq_params)); + + params->cs55x0 = gx_pci; + gx_params = params; + + /* keep cs55x0 configurations */ + pci_read_config_byte(params->cs55x0, PCI_SUSCFG, &(params->pci_suscfg)); + pci_read_config_byte(params->cs55x0, PCI_PMER1, &(params->pci_pmer1)); + pci_read_config_byte(params->cs55x0, PCI_PMER2, &(params->pci_pmer2)); + pci_read_config_byte(params->cs55x0, PCI_MODON, &(params->on_duration)); + pci_read_config_byte(params->cs55x0, PCI_MODOFF, &(params->off_duration)); + pci_read_config_dword(params->cs55x0, PCI_CLASS_REVISION, &class_rev); + params->pci_rev = class_rev && 0xff; + + if ((ret = cpufreq_register_driver(&gx_suspmod_driver))) { + kfree(params); + return ret; /* register error! */ + } + + return 0; +} + +static void __exit cpufreq_gx_exit(void) +{ + cpufreq_unregister_driver(&gx_suspmod_driver); + pci_dev_put(gx_params->cs55x0); + kfree(gx_params); +} + +MODULE_AUTHOR ("Hiroshi Miura <miura@da-cha.org>"); +MODULE_DESCRIPTION ("Cpufreq driver for Cyrix MediaGX and NatSemi Geode"); +MODULE_LICENSE ("GPL"); + +module_init(cpufreq_gx_init); +module_exit(cpufreq_gx_exit); + diff --git a/arch/i386/kernel/cpu/cpufreq/longhaul.c b/arch/i386/kernel/cpu/cpufreq/longhaul.c new file mode 100644 index 00000000000..ab0f9f5aac1 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/longhaul.c @@ -0,0 +1,658 @@ +/* + * (C) 2001-2004 Dave Jones. <davej@codemonkey.org.uk> + * (C) 2002 Padraig Brady. <padraig@antefacto.com> + * + * Licensed under the terms of the GNU GPL License version 2. + * Based upon datasheets & sample CPUs kindly provided by VIA. + * + * VIA have currently 3 different versions of Longhaul. + * Version 1 (Longhaul) uses the BCR2 MSR at 0x1147. + * It is present only in Samuel 1 (C5A), Samuel 2 (C5B) stepping 0. + * Version 2 of longhaul is the same as v1, but adds voltage scaling. + * Present in Samuel 2 (steppings 1-7 only) (C5B), and Ezra (C5C) + * voltage scaling support has currently been disabled in this driver + * until we have code that gets it right. + * Version 3 of longhaul got renamed to Powersaver and redesigned + * to use the POWERSAVER MSR at 0x110a. + * It is present in Ezra-T (C5M), Nehemiah (C5X) and above. + * It's pretty much the same feature wise to longhaul v2, though + * there is provision for scaling FSB too, but this doesn't work + * too well in practice so we don't even try to use this. + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/slab.h> +#include <linux/string.h> + +#include <asm/msr.h> +#include <asm/timex.h> +#include <asm/io.h> + +#include "longhaul.h" + +#define PFX "longhaul: " + +#define TYPE_LONGHAUL_V1 1 +#define TYPE_LONGHAUL_V2 2 +#define TYPE_POWERSAVER 3 + +#define CPU_SAMUEL 1 +#define CPU_SAMUEL2 2 +#define CPU_EZRA 3 +#define CPU_EZRA_T 4 +#define CPU_NEHEMIAH 5 + +static int cpu_model; +static unsigned int numscales=16, numvscales; +static unsigned int fsb; +static int minvid, maxvid; +static unsigned int minmult, maxmult; +static int can_scale_voltage; +static int vrmrev; + +/* Module parameters */ +static int dont_scale_voltage; + + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "longhaul", msg) + + +#define __hlt() __asm__ __volatile__("hlt": : :"memory") + +/* Clock ratios multiplied by 10 */ +static int clock_ratio[32]; +static int eblcr_table[32]; +static int voltage_table[32]; +static unsigned int highest_speed, lowest_speed; /* kHz */ +static int longhaul_version; +static struct cpufreq_frequency_table *longhaul_table; + +#ifdef CONFIG_CPU_FREQ_DEBUG +static char speedbuffer[8]; + +static char *print_speed(int speed) +{ + if (speed > 1000) { + if (speed%1000 == 0) + sprintf (speedbuffer, "%dGHz", speed/1000); + else + sprintf (speedbuffer, "%d.%dGHz", speed/1000, (speed%1000)/100); + } else + sprintf (speedbuffer, "%dMHz", speed); + + return speedbuffer; +} +#endif + + +static unsigned int calc_speed(int mult) +{ + int khz; + khz = (mult/10)*fsb; + if (mult%10) + khz += fsb/2; + khz *= 1000; + return khz; +} + + +static int longhaul_get_cpu_mult(void) +{ + unsigned long invalue=0,lo, hi; + + rdmsr (MSR_IA32_EBL_CR_POWERON, lo, hi); + invalue = (lo & (1<<22|1<<23|1<<24|1<<25)) >>22; + if (longhaul_version==TYPE_LONGHAUL_V2 || longhaul_version==TYPE_POWERSAVER) { + if (lo & (1<<27)) + invalue+=16; + } + return eblcr_table[invalue]; +} + + +static void do_powersaver(union msr_longhaul *longhaul, + unsigned int clock_ratio_index) +{ + int version; + + switch (cpu_model) { + case CPU_EZRA_T: + version = 3; + break; + case CPU_NEHEMIAH: + version = 0xf; + break; + default: + return; + } + + rdmsrl(MSR_VIA_LONGHAUL, longhaul->val); + longhaul->bits.SoftBusRatio = clock_ratio_index & 0xf; + longhaul->bits.SoftBusRatio4 = (clock_ratio_index & 0x10) >> 4; + longhaul->bits.EnableSoftBusRatio = 1; + longhaul->bits.RevisionKey = 0; + local_irq_disable(); + wrmsrl(MSR_VIA_LONGHAUL, longhaul->val); + local_irq_enable(); + __hlt(); + + rdmsrl(MSR_VIA_LONGHAUL, longhaul->val); + longhaul->bits.EnableSoftBusRatio = 0; + longhaul->bits.RevisionKey = version; + local_irq_disable(); + wrmsrl(MSR_VIA_LONGHAUL, longhaul->val); + local_irq_enable(); +} + +/** + * longhaul_set_cpu_frequency() + * @clock_ratio_index : bitpattern of the new multiplier. + * + * Sets a new clock ratio. + */ + +static void longhaul_setstate(unsigned int clock_ratio_index) +{ + int speed, mult; + struct cpufreq_freqs freqs; + union msr_longhaul longhaul; + union msr_bcr2 bcr2; + static unsigned int old_ratio=-1; + + if (old_ratio == clock_ratio_index) + return; + old_ratio = clock_ratio_index; + + mult = clock_ratio[clock_ratio_index]; + if (mult == -1) + return; + + speed = calc_speed(mult); + if ((speed > highest_speed) || (speed < lowest_speed)) + return; + + freqs.old = calc_speed(longhaul_get_cpu_mult()); + freqs.new = speed; + freqs.cpu = 0; /* longhaul.c is UP only driver */ + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + dprintk ("Setting to FSB:%dMHz Mult:%d.%dx (%s)\n", + fsb, mult/10, mult%10, print_speed(speed/1000)); + + switch (longhaul_version) { + + /* + * Longhaul v1. (Samuel[C5A] and Samuel2 stepping 0[C5B]) + * Software controlled multipliers only. + * + * *NB* Until we get voltage scaling working v1 & v2 are the same code. + * Longhaul v2 appears in Samuel2 Steppings 1->7 [C5b] and Ezra [C5C] + */ + case TYPE_LONGHAUL_V1: + case TYPE_LONGHAUL_V2: + rdmsrl (MSR_VIA_BCR2, bcr2.val); + /* Enable software clock multiplier */ + bcr2.bits.ESOFTBF = 1; + bcr2.bits.CLOCKMUL = clock_ratio_index; + local_irq_disable(); + wrmsrl (MSR_VIA_BCR2, bcr2.val); + local_irq_enable(); + + __hlt(); + + /* Disable software clock multiplier */ + rdmsrl (MSR_VIA_BCR2, bcr2.val); + bcr2.bits.ESOFTBF = 0; + local_irq_disable(); + wrmsrl (MSR_VIA_BCR2, bcr2.val); + local_irq_enable(); + break; + + /* + * Longhaul v3 (aka Powersaver). (Ezra-T [C5M] & Nehemiah [C5N]) + * We can scale voltage with this too, but that's currently + * disabled until we come up with a decent 'match freq to voltage' + * algorithm. + * When we add voltage scaling, we will also need to do the + * voltage/freq setting in order depending on the direction + * of scaling (like we do in powernow-k7.c) + * Nehemiah can do FSB scaling too, but this has never been proven + * to work in practice. + */ + case TYPE_POWERSAVER: + do_powersaver(&longhaul, clock_ratio_index); + break; + } + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); +} + +/* + * Centaur decided to make life a little more tricky. + * Only longhaul v1 is allowed to read EBLCR BSEL[0:1]. + * Samuel2 and above have to try and guess what the FSB is. + * We do this by assuming we booted at maximum multiplier, and interpolate + * between that value multiplied by possible FSBs and cpu_mhz which + * was calculated at boot time. Really ugly, but no other way to do this. + */ + +#define ROUNDING 0xf + +static int _guess(int guess) +{ + int target; + + target = ((maxmult/10)*guess); + if (maxmult%10 != 0) + target += (guess/2); + target += ROUNDING/2; + target &= ~ROUNDING; + return target; +} + + +static int guess_fsb(void) +{ + int speed = (cpu_khz/1000); + int i; + int speeds[3] = { 66, 100, 133 }; + + speed += ROUNDING/2; + speed &= ~ROUNDING; + + for (i=0; i<3; i++) { + if (_guess(speeds[i]) == speed) + return speeds[i]; + } + return 0; +} + + +static int __init longhaul_get_ranges(void) +{ + unsigned long invalue; + unsigned int multipliers[32]= { + 50,30,40,100,55,35,45,95,90,70,80,60,120,75,85,65, + -1,110,120,-1,135,115,125,105,130,150,160,140,-1,155,-1,145 }; + unsigned int j, k = 0; + union msr_longhaul longhaul; + unsigned long lo, hi; + unsigned int eblcr_fsb_table_v1[] = { 66, 133, 100, -1 }; + unsigned int eblcr_fsb_table_v2[] = { 133, 100, -1, 66 }; + + switch (longhaul_version) { + case TYPE_LONGHAUL_V1: + case TYPE_LONGHAUL_V2: + /* Ugh, Longhaul v1 didn't have the min/max MSRs. + Assume min=3.0x & max = whatever we booted at. */ + minmult = 30; + maxmult = longhaul_get_cpu_mult(); + rdmsr (MSR_IA32_EBL_CR_POWERON, lo, hi); + invalue = (lo & (1<<18|1<<19)) >>18; + if (cpu_model==CPU_SAMUEL || cpu_model==CPU_SAMUEL2) + fsb = eblcr_fsb_table_v1[invalue]; + else + fsb = guess_fsb(); + break; + + case TYPE_POWERSAVER: + /* Ezra-T */ + if (cpu_model==CPU_EZRA_T) { + rdmsrl (MSR_VIA_LONGHAUL, longhaul.val); + invalue = longhaul.bits.MaxMHzBR; + if (longhaul.bits.MaxMHzBR4) + invalue += 16; + maxmult=multipliers[invalue]; + + invalue = longhaul.bits.MinMHzBR; + if (longhaul.bits.MinMHzBR4 == 1) + minmult = 30; + else + minmult = multipliers[invalue]; + fsb = eblcr_fsb_table_v2[longhaul.bits.MaxMHzFSB]; + break; + } + + /* Nehemiah */ + if (cpu_model==CPU_NEHEMIAH) { + rdmsrl (MSR_VIA_LONGHAUL, longhaul.val); + + /* + * TODO: This code works, but raises a lot of questions. + * - Some Nehemiah's seem to have broken Min/MaxMHzBR's. + * We get around this by using a hardcoded multiplier of 4.0x + * for the minimimum speed, and the speed we booted up at for the max. + * This is done in longhaul_get_cpu_mult() by reading the EBLCR register. + * - According to some VIA documentation EBLCR is only + * in pre-Nehemiah C3s. How this still works is a mystery. + * We're possibly using something undocumented and unsupported, + * But it works, so we don't grumble. + */ + minmult=40; + maxmult=longhaul_get_cpu_mult(); + + /* Starting with the 1.2GHz parts, theres a 200MHz bus. */ + if ((cpu_khz/1000) > 1200) + fsb = 200; + else + fsb = eblcr_fsb_table_v2[longhaul.bits.MaxMHzFSB]; + break; + } + } + + dprintk ("MinMult:%d.%dx MaxMult:%d.%dx\n", + minmult/10, minmult%10, maxmult/10, maxmult%10); + + if (fsb == -1) { + printk (KERN_INFO PFX "Invalid (reserved) FSB!\n"); + return -EINVAL; + } + + highest_speed = calc_speed(maxmult); + lowest_speed = calc_speed(minmult); + dprintk ("FSB:%dMHz Lowest speed: %s Highest speed:%s\n", fsb, + print_speed(lowest_speed/1000), + print_speed(highest_speed/1000)); + + if (lowest_speed == highest_speed) { + printk (KERN_INFO PFX "highestspeed == lowest, aborting.\n"); + return -EINVAL; + } + if (lowest_speed > highest_speed) { + printk (KERN_INFO PFX "nonsense! lowest (%d > %d) !\n", + lowest_speed, highest_speed); + return -EINVAL; + } + + longhaul_table = kmalloc((numscales + 1) * sizeof(struct cpufreq_frequency_table), GFP_KERNEL); + if(!longhaul_table) + return -ENOMEM; + + for (j=0; j < numscales; j++) { + unsigned int ratio; + ratio = clock_ratio[j]; + if (ratio == -1) + continue; + if (ratio > maxmult || ratio < minmult) + continue; + longhaul_table[k].frequency = calc_speed(ratio); + longhaul_table[k].index = j; + k++; + } + + longhaul_table[k].frequency = CPUFREQ_TABLE_END; + if (!k) { + kfree (longhaul_table); + return -EINVAL; + } + + return 0; +} + + +static void __init longhaul_setup_voltagescaling(void) +{ + union msr_longhaul longhaul; + + rdmsrl (MSR_VIA_LONGHAUL, longhaul.val); + + if (!(longhaul.bits.RevisionID & 1)) + return; + + minvid = longhaul.bits.MinimumVID; + maxvid = longhaul.bits.MaximumVID; + vrmrev = longhaul.bits.VRMRev; + + if (minvid == 0 || maxvid == 0) { + printk (KERN_INFO PFX "Bogus values Min:%d.%03d Max:%d.%03d. " + "Voltage scaling disabled.\n", + minvid/1000, minvid%1000, maxvid/1000, maxvid%1000); + return; + } + + if (minvid == maxvid) { + printk (KERN_INFO PFX "Claims to support voltage scaling but min & max are " + "both %d.%03d. Voltage scaling disabled\n", + maxvid/1000, maxvid%1000); + return; + } + + if (vrmrev==0) { + dprintk ("VRM 8.5 \n"); + memcpy (voltage_table, vrm85scales, sizeof(voltage_table)); + numvscales = (voltage_table[maxvid]-voltage_table[minvid])/25; + } else { + dprintk ("Mobile VRM \n"); + memcpy (voltage_table, mobilevrmscales, sizeof(voltage_table)); + numvscales = (voltage_table[maxvid]-voltage_table[minvid])/5; + } + + /* Current voltage isn't readable at first, so we need to + set it to a known value. The spec says to use maxvid */ + longhaul.bits.RevisionKey = longhaul.bits.RevisionID; /* FIXME: This is bad. */ + longhaul.bits.EnableSoftVID = 1; + longhaul.bits.SoftVID = maxvid; + wrmsrl (MSR_VIA_LONGHAUL, longhaul.val); + + minvid = voltage_table[minvid]; + maxvid = voltage_table[maxvid]; + + dprintk ("Min VID=%d.%03d Max VID=%d.%03d, %d possible voltage scales\n", + maxvid/1000, maxvid%1000, minvid/1000, minvid%1000, numvscales); + + can_scale_voltage = 1; +} + + +static int longhaul_verify(struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, longhaul_table); +} + + +static int longhaul_target(struct cpufreq_policy *policy, + unsigned int target_freq, unsigned int relation) +{ + unsigned int table_index = 0; + unsigned int new_clock_ratio = 0; + + if (cpufreq_frequency_table_target(policy, longhaul_table, target_freq, relation, &table_index)) + return -EINVAL; + + new_clock_ratio = longhaul_table[table_index].index & 0xFF; + + longhaul_setstate(new_clock_ratio); + + return 0; +} + + +static unsigned int longhaul_get(unsigned int cpu) +{ + if (cpu) + return 0; + return calc_speed(longhaul_get_cpu_mult()); +} + + +static int __init longhaul_cpu_init(struct cpufreq_policy *policy) +{ + struct cpuinfo_x86 *c = cpu_data; + char *cpuname=NULL; + int ret; + + switch (c->x86_model) { + case 6: + cpu_model = CPU_SAMUEL; + cpuname = "C3 'Samuel' [C5A]"; + longhaul_version = TYPE_LONGHAUL_V1; + memcpy (clock_ratio, samuel1_clock_ratio, sizeof(samuel1_clock_ratio)); + memcpy (eblcr_table, samuel1_eblcr, sizeof(samuel1_eblcr)); + break; + + case 7: + longhaul_version = TYPE_LONGHAUL_V1; + switch (c->x86_mask) { + case 0: + cpu_model = CPU_SAMUEL2; + cpuname = "C3 'Samuel 2' [C5B]"; + /* Note, this is not a typo, early Samuel2's had Samuel1 ratios. */ + memcpy (clock_ratio, samuel1_clock_ratio, sizeof(samuel1_clock_ratio)); + memcpy (eblcr_table, samuel2_eblcr, sizeof(samuel2_eblcr)); + break; + case 1 ... 15: + if (c->x86_mask < 8) { + cpu_model = CPU_SAMUEL2; + cpuname = "C3 'Samuel 2' [C5B]"; + } else { + cpu_model = CPU_EZRA; + cpuname = "C3 'Ezra' [C5C]"; + } + memcpy (clock_ratio, ezra_clock_ratio, sizeof(ezra_clock_ratio)); + memcpy (eblcr_table, ezra_eblcr, sizeof(ezra_eblcr)); + break; + } + break; + + case 8: + cpu_model = CPU_EZRA_T; + cpuname = "C3 'Ezra-T' [C5M]"; + longhaul_version = TYPE_POWERSAVER; + numscales=32; + memcpy (clock_ratio, ezrat_clock_ratio, sizeof(ezrat_clock_ratio)); + memcpy (eblcr_table, ezrat_eblcr, sizeof(ezrat_eblcr)); + break; + + case 9: + cpu_model = CPU_NEHEMIAH; + longhaul_version = TYPE_POWERSAVER; + numscales=32; + switch (c->x86_mask) { + case 0 ... 1: + cpuname = "C3 'Nehemiah A' [C5N]"; + memcpy (clock_ratio, nehemiah_a_clock_ratio, sizeof(nehemiah_a_clock_ratio)); + memcpy (eblcr_table, nehemiah_a_eblcr, sizeof(nehemiah_a_eblcr)); + break; + case 2 ... 4: + cpuname = "C3 'Nehemiah B' [C5N]"; + memcpy (clock_ratio, nehemiah_b_clock_ratio, sizeof(nehemiah_b_clock_ratio)); + memcpy (eblcr_table, nehemiah_b_eblcr, sizeof(nehemiah_b_eblcr)); + break; + case 5 ... 15: + cpuname = "C3 'Nehemiah C' [C5N]"; + memcpy (clock_ratio, nehemiah_c_clock_ratio, sizeof(nehemiah_c_clock_ratio)); + memcpy (eblcr_table, nehemiah_c_eblcr, sizeof(nehemiah_c_eblcr)); + break; + } + break; + + default: + cpuname = "Unknown"; + break; + } + + printk (KERN_INFO PFX "VIA %s CPU detected. ", cpuname); + switch (longhaul_version) { + case TYPE_LONGHAUL_V1: + case TYPE_LONGHAUL_V2: + printk ("Longhaul v%d supported.\n", longhaul_version); + break; + case TYPE_POWERSAVER: + printk ("Powersaver supported.\n"); + break; + }; + + ret = longhaul_get_ranges(); + if (ret != 0) + return ret; + + if ((longhaul_version==TYPE_LONGHAUL_V2 || longhaul_version==TYPE_POWERSAVER) && + (dont_scale_voltage==0)) + longhaul_setup_voltagescaling(); + + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = calc_speed(longhaul_get_cpu_mult()); + + ret = cpufreq_frequency_table_cpuinfo(policy, longhaul_table); + if (ret) + return ret; + + cpufreq_frequency_table_get_attr(longhaul_table, policy->cpu); + + return 0; +} + +static int __devexit longhaul_cpu_exit(struct cpufreq_policy *policy) +{ + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + +static struct freq_attr* longhaul_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver longhaul_driver = { + .verify = longhaul_verify, + .target = longhaul_target, + .get = longhaul_get, + .init = longhaul_cpu_init, + .exit = __devexit_p(longhaul_cpu_exit), + .name = "longhaul", + .owner = THIS_MODULE, + .attr = longhaul_attr, +}; + + +static int __init longhaul_init(void) +{ + struct cpuinfo_x86 *c = cpu_data; + + if (c->x86_vendor != X86_VENDOR_CENTAUR || c->x86 != 6) + return -ENODEV; + + switch (c->x86_model) { + case 6 ... 9: + return cpufreq_register_driver(&longhaul_driver); + default: + printk (KERN_INFO PFX "Unknown VIA CPU. Contact davej@codemonkey.org.uk\n"); + } + + return -ENODEV; +} + + +static void __exit longhaul_exit(void) +{ + int i=0; + + for (i=0; i < numscales; i++) { + if (clock_ratio[i] == maxmult) { + longhaul_setstate(i); + break; + } + } + + cpufreq_unregister_driver(&longhaul_driver); + kfree(longhaul_table); +} + +module_param (dont_scale_voltage, int, 0644); +MODULE_PARM_DESC(dont_scale_voltage, "Don't scale voltage of processor"); + +MODULE_AUTHOR ("Dave Jones <davej@codemonkey.org.uk>"); +MODULE_DESCRIPTION ("Longhaul driver for VIA Cyrix processors."); +MODULE_LICENSE ("GPL"); + +module_init(longhaul_init); +module_exit(longhaul_exit); + diff --git a/arch/i386/kernel/cpu/cpufreq/longhaul.h b/arch/i386/kernel/cpu/cpufreq/longhaul.h new file mode 100644 index 00000000000..2a495c162ec --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/longhaul.h @@ -0,0 +1,466 @@ +/* + * longhaul.h + * (C) 2003 Dave Jones. + * + * Licensed under the terms of the GNU GPL License version 2. + * + * VIA-specific information + */ + +union msr_bcr2 { + struct { + unsigned Reseved:19, // 18:0 + ESOFTBF:1, // 19 + Reserved2:3, // 22:20 + CLOCKMUL:4, // 26:23 + Reserved3:5; // 31:27 + } bits; + unsigned long val; +}; + +union msr_longhaul { + struct { + unsigned RevisionID:4, // 3:0 + RevisionKey:4, // 7:4 + EnableSoftBusRatio:1, // 8 + EnableSoftVID:1, // 9 + EnableSoftBSEL:1, // 10 + Reserved:3, // 11:13 + SoftBusRatio4:1, // 14 + VRMRev:1, // 15 + SoftBusRatio:4, // 19:16 + SoftVID:5, // 24:20 + Reserved2:3, // 27:25 + SoftBSEL:2, // 29:28 + Reserved3:2, // 31:30 + MaxMHzBR:4, // 35:32 + MaximumVID:5, // 40:36 + MaxMHzFSB:2, // 42:41 + MaxMHzBR4:1, // 43 + Reserved4:4, // 47:44 + MinMHzBR:4, // 51:48 + MinimumVID:5, // 56:52 + MinMHzFSB:2, // 58:57 + MinMHzBR4:1, // 59 + Reserved5:4; // 63:60 + } bits; + unsigned long long val; +}; + +/* + * Clock ratio tables. Div/Mod by 10 to get ratio. + * The eblcr ones specify the ratio read from the CPU. + * The clock_ratio ones specify what to write to the CPU. + */ + +/* + * VIA C3 Samuel 1 & Samuel 2 (stepping 0) + */ +static int __initdata samuel1_clock_ratio[16] = { + -1, /* 0000 -> RESERVED */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + -1, /* 0011 -> RESERVED */ + -1, /* 0100 -> RESERVED */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + -1, /* 1110 -> RESERVED */ + -1, /* 1111 -> RESERVED */ +}; + +static int __initdata samuel1_eblcr[16] = { + 50, /* 0000 -> RESERVED */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + -1, /* 0011 -> RESERVED */ + 55, /* 0100 -> 5.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + -1, /* 0111 -> RESERVED */ + -1, /* 1000 -> RESERVED */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + -1, /* 1100 -> RESERVED */ + 75, /* 1101 -> 7.5x */ + -1, /* 1110 -> RESERVED */ + 65, /* 1111 -> 6.5x */ +}; + +/* + * VIA C3 Samuel2 Stepping 1->15 + */ +static int __initdata samuel2_eblcr[16] = { + 50, /* 0000 -> 5.0x */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 110, /* 0111 -> 11.0x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 130, /* 1110 -> 13.0x */ + 65, /* 1111 -> 6.5x */ +}; + +/* + * VIA C3 Ezra + */ +static int __initdata ezra_clock_ratio[16] = { + 100, /* 0000 -> 10.0x */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + 90, /* 0011 -> 9.0x */ + 95, /* 0100 -> 9.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 120, /* 1111 -> 12.0x */ +}; + +static int __initdata ezra_eblcr[16] = { + 50, /* 0000 -> 5.0x */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 95, /* 0111 -> 9.5x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 65, /* 1111 -> 6.5x */ +}; + +/* + * VIA C3 (Ezra-T) [C5M]. + */ +static int __initdata ezrat_clock_ratio[32] = { + 100, /* 0000 -> 10.0x */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + 90, /* 0011 -> 9.0x */ + 95, /* 0100 -> 9.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 120, /* 1111 -> 12.0x */ + + -1, /* 0000 -> RESERVED (10.0x) */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + -1, /* 0011 -> RESERVED (9.0x)*/ + 105, /* 0100 -> 10.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 135, /* 0111 -> 13.5x */ + 140, /* 1000 -> 14.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 130, /* 1011 -> 13.0x */ + 145, /* 1100 -> 14.5x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + -1, /* 1111 -> RESERVED (12.0x) */ +}; + +static int __initdata ezrat_eblcr[32] = { + 50, /* 0000 -> 5.0x */ + 30, /* 0001 -> 3.0x */ + 40, /* 0010 -> 4.0x */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + 35, /* 0101 -> 3.5x */ + 45, /* 0110 -> 4.5x */ + 95, /* 0111 -> 9.5x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 65, /* 1111 -> 6.5x */ + + -1, /* 0000 -> RESERVED (9.0x) */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + -1, /* 0011 -> RESERVED (10.0x)*/ + 135, /* 0100 -> 13.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 105, /* 0111 -> 10.5x */ + 130, /* 1000 -> 13.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 140, /* 1011 -> 14.0x */ + -1, /* 1100 -> RESERVED (12.0x) */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 145, /* 1111 -> 14.5x */ +}; + +/* + * VIA C3 Nehemiah */ + +static int __initdata nehemiah_a_clock_ratio[32] = { + 100, /* 0000 -> 10.0x */ + 160, /* 0001 -> 16.0x */ + -1, /* 0010 -> RESERVED */ + 90, /* 0011 -> 9.0x */ + 95, /* 0100 -> 9.5x */ + -1, /* 0101 -> RESERVED */ + -1, /* 0110 -> RESERVED */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 120, /* 1111 -> 12.0x */ + 100, /* 0000 -> 10.0x */ + -1, /* 0001 -> RESERVED */ + 120, /* 0010 -> 12.0x */ + 90, /* 0011 -> 9.0x */ + 105, /* 0100 -> 10.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 135, /* 0111 -> 13.5x */ + 140, /* 1000 -> 14.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 130, /* 1011 -> 13.0x */ + 145, /* 1100 -> 14.5x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 120, /* 1111 -> 12.0x */ +}; + +static int __initdata nehemiah_b_clock_ratio[32] = { + 100, /* 0000 -> 10.0x */ + 160, /* 0001 -> 16.0x */ + -1, /* 0010 -> RESERVED */ + 90, /* 0011 -> 9.0x */ + 95, /* 0100 -> 9.5x */ + -1, /* 0101 -> RESERVED */ + -1, /* 0110 -> RESERVED */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 120, /* 1111 -> 12.0x */ + 100, /* 0000 -> 10.0x */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + 90, /* 0011 -> 9.0x */ + 105, /* 0100 -> 10.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 135, /* 0111 -> 13.5x */ + 140, /* 1000 -> 14.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 130, /* 1011 -> 13.0x */ + 145, /* 1100 -> 14.5x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 120, /* 1111 -> 12.0x */ +}; + +static int __initdata nehemiah_c_clock_ratio[32] = { + 100, /* 0000 -> 10.0x */ + 160, /* 0001 -> 16.0x */ + 40, /* 0010 -> RESERVED */ + 90, /* 0011 -> 9.0x */ + 95, /* 0100 -> 9.5x */ + -1, /* 0101 -> RESERVED */ + 45, /* 0110 -> RESERVED */ + 55, /* 0111 -> 5.5x */ + 60, /* 1000 -> 6.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 50, /* 1011 -> 5.0x */ + 65, /* 1100 -> 6.5x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 120, /* 1111 -> 12.0x */ + 100, /* 0000 -> 10.0x */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + 90, /* 0011 -> 9.0x */ + 105, /* 0100 -> 10.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 135, /* 0111 -> 13.5x */ + 140, /* 1000 -> 14.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 130, /* 1011 -> 13.0x */ + 145, /* 1100 -> 14.5x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 120, /* 1111 -> 12.0x */ +}; + +static int __initdata nehemiah_a_eblcr[32] = { + 50, /* 0000 -> 5.0x */ + 160, /* 0001 -> 16.0x */ + -1, /* 0010 -> RESERVED */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + -1, /* 0101 -> RESERVED */ + -1, /* 0110 -> RESERVED */ + 95, /* 0111 -> 9.5x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 65, /* 1111 -> 6.5x */ + 90, /* 0000 -> 9.0x */ + -1, /* 0001 -> RESERVED */ + 120, /* 0010 -> 12.0x */ + 100, /* 0011 -> 10.0x */ + 135, /* 0100 -> 13.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 105, /* 0111 -> 10.5x */ + 130, /* 1000 -> 13.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 140, /* 1011 -> 14.0x */ + 120, /* 1100 -> 12.0x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 145 /* 1111 -> 14.5x */ + /* end of table */ +}; +static int __initdata nehemiah_b_eblcr[32] = { + 50, /* 0000 -> 5.0x */ + 160, /* 0001 -> 16.0x */ + -1, /* 0010 -> RESERVED */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + -1, /* 0101 -> RESERVED */ + -1, /* 0110 -> RESERVED */ + 95, /* 0111 -> 9.5x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 65, /* 1111 -> 6.5x */ + 90, /* 0000 -> 9.0x */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + 100, /* 0011 -> 10.0x */ + 135, /* 0100 -> 13.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 105, /* 0111 -> 10.5x */ + 130, /* 1000 -> 13.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 140, /* 1011 -> 14.0x */ + 120, /* 1100 -> 12.0x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 145 /* 1111 -> 14.5x */ + /* end of table */ +}; +static int __initdata nehemiah_c_eblcr[32] = { + 50, /* 0000 -> 5.0x */ + 160, /* 0001 -> 16.0x */ + 40, /* 0010 -> RESERVED */ + 100, /* 0011 -> 10.0x */ + 55, /* 0100 -> 5.5x */ + -1, /* 0101 -> RESERVED */ + 45, /* 0110 -> RESERVED */ + 95, /* 0111 -> 9.5x */ + 90, /* 1000 -> 9.0x */ + 70, /* 1001 -> 7.0x */ + 80, /* 1010 -> 8.0x */ + 60, /* 1011 -> 6.0x */ + 120, /* 1100 -> 12.0x */ + 75, /* 1101 -> 7.5x */ + 85, /* 1110 -> 8.5x */ + 65, /* 1111 -> 6.5x */ + 90, /* 0000 -> 9.0x */ + 110, /* 0001 -> 11.0x */ + 120, /* 0010 -> 12.0x */ + 100, /* 0011 -> 10.0x */ + 135, /* 0100 -> 13.5x */ + 115, /* 0101 -> 11.5x */ + 125, /* 0110 -> 12.5x */ + 105, /* 0111 -> 10.5x */ + 130, /* 1000 -> 13.0x */ + 150, /* 1001 -> 15.0x */ + 160, /* 1010 -> 16.0x */ + 140, /* 1011 -> 14.0x */ + 120, /* 1100 -> 12.0x */ + 155, /* 1101 -> 15.5x */ + -1, /* 1110 -> RESERVED (13.0x) */ + 145 /* 1111 -> 14.5x */ + /* end of table */ +}; + +/* + * Voltage scales. Div/Mod by 1000 to get actual voltage. + * Which scale to use depends on the VRM type in use. + */ +static int __initdata vrm85scales[32] = { + 1250, 1200, 1150, 1100, 1050, 1800, 1750, 1700, + 1650, 1600, 1550, 1500, 1450, 1400, 1350, 1300, + 1275, 1225, 1175, 1125, 1075, 1825, 1775, 1725, + 1675, 1625, 1575, 1525, 1475, 1425, 1375, 1325, +}; + +static int __initdata mobilevrmscales[32] = { + 2000, 1950, 1900, 1850, 1800, 1750, 1700, 1650, + 1600, 1550, 1500, 1450, 1500, 1350, 1300, -1, + 1275, 1250, 1225, 1200, 1175, 1150, 1125, 1100, + 1075, 1050, 1025, 1000, 975, 950, 925, -1, +}; + diff --git a/arch/i386/kernel/cpu/cpufreq/longrun.c b/arch/i386/kernel/cpu/cpufreq/longrun.c new file mode 100644 index 00000000000..e3868de4dc2 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/longrun.c @@ -0,0 +1,326 @@ +/* + * (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de> + * + * Licensed under the terms of the GNU GPL License version 2. + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/cpufreq.h> + +#include <asm/msr.h> +#include <asm/processor.h> +#include <asm/timex.h> + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "longrun", msg) + +static struct cpufreq_driver longrun_driver; + +/** + * longrun_{low,high}_freq is needed for the conversion of cpufreq kHz + * values into per cent values. In TMTA microcode, the following is valid: + * performance_pctg = (current_freq - low_freq)/(high_freq - low_freq) + */ +static unsigned int longrun_low_freq, longrun_high_freq; + + +/** + * longrun_get_policy - get the current LongRun policy + * @policy: struct cpufreq_policy where current policy is written into + * + * Reads the current LongRun policy by access to MSR_TMTA_LONGRUN_FLAGS + * and MSR_TMTA_LONGRUN_CTRL + */ +static void __init longrun_get_policy(struct cpufreq_policy *policy) +{ + u32 msr_lo, msr_hi; + + rdmsr(MSR_TMTA_LONGRUN_FLAGS, msr_lo, msr_hi); + dprintk("longrun flags are %x - %x\n", msr_lo, msr_hi); + if (msr_lo & 0x01) + policy->policy = CPUFREQ_POLICY_PERFORMANCE; + else + policy->policy = CPUFREQ_POLICY_POWERSAVE; + + rdmsr(MSR_TMTA_LONGRUN_CTRL, msr_lo, msr_hi); + dprintk("longrun ctrl is %x - %x\n", msr_lo, msr_hi); + msr_lo &= 0x0000007F; + msr_hi &= 0x0000007F; + + if ( longrun_high_freq <= longrun_low_freq ) { + /* Assume degenerate Longrun table */ + policy->min = policy->max = longrun_high_freq; + } else { + policy->min = longrun_low_freq + msr_lo * + ((longrun_high_freq - longrun_low_freq) / 100); + policy->max = longrun_low_freq + msr_hi * + ((longrun_high_freq - longrun_low_freq) / 100); + } + policy->cpu = 0; +} + + +/** + * longrun_set_policy - sets a new CPUFreq policy + * @policy: new policy + * + * Sets a new CPUFreq policy on LongRun-capable processors. This function + * has to be called with cpufreq_driver locked. + */ +static int longrun_set_policy(struct cpufreq_policy *policy) +{ + u32 msr_lo, msr_hi; + u32 pctg_lo, pctg_hi; + + if (!policy) + return -EINVAL; + + if ( longrun_high_freq <= longrun_low_freq ) { + /* Assume degenerate Longrun table */ + pctg_lo = pctg_hi = 100; + } else { + pctg_lo = (policy->min - longrun_low_freq) / + ((longrun_high_freq - longrun_low_freq) / 100); + pctg_hi = (policy->max - longrun_low_freq) / + ((longrun_high_freq - longrun_low_freq) / 100); + } + + if (pctg_hi > 100) + pctg_hi = 100; + if (pctg_lo > pctg_hi) + pctg_lo = pctg_hi; + + /* performance or economy mode */ + rdmsr(MSR_TMTA_LONGRUN_FLAGS, msr_lo, msr_hi); + msr_lo &= 0xFFFFFFFE; + switch (policy->policy) { + case CPUFREQ_POLICY_PERFORMANCE: + msr_lo |= 0x00000001; + break; + case CPUFREQ_POLICY_POWERSAVE: + break; + } + wrmsr(MSR_TMTA_LONGRUN_FLAGS, msr_lo, msr_hi); + + /* lower and upper boundary */ + rdmsr(MSR_TMTA_LONGRUN_CTRL, msr_lo, msr_hi); + msr_lo &= 0xFFFFFF80; + msr_hi &= 0xFFFFFF80; + msr_lo |= pctg_lo; + msr_hi |= pctg_hi; + wrmsr(MSR_TMTA_LONGRUN_CTRL, msr_lo, msr_hi); + + return 0; +} + + +/** + * longrun_verify_poliy - verifies a new CPUFreq policy + * @policy: the policy to verify + * + * Validates a new CPUFreq policy. This function has to be called with + * cpufreq_driver locked. + */ +static int longrun_verify_policy(struct cpufreq_policy *policy) +{ + if (!policy) + return -EINVAL; + + policy->cpu = 0; + cpufreq_verify_within_limits(policy, + policy->cpuinfo.min_freq, + policy->cpuinfo.max_freq); + + if ((policy->policy != CPUFREQ_POLICY_POWERSAVE) && + (policy->policy != CPUFREQ_POLICY_PERFORMANCE)) + return -EINVAL; + + return 0; +} + +static unsigned int longrun_get(unsigned int cpu) +{ + u32 eax, ebx, ecx, edx; + + if (cpu) + return 0; + + cpuid(0x80860007, &eax, &ebx, &ecx, &edx); + dprintk("cpuid eax is %u\n", eax); + + return (eax * 1000); +} + +/** + * longrun_determine_freqs - determines the lowest and highest possible core frequency + * @low_freq: an int to put the lowest frequency into + * @high_freq: an int to put the highest frequency into + * + * Determines the lowest and highest possible core frequencies on this CPU. + * This is necessary to calculate the performance percentage according to + * TMTA rules: + * performance_pctg = (target_freq - low_freq)/(high_freq - low_freq) + */ +static unsigned int __init longrun_determine_freqs(unsigned int *low_freq, + unsigned int *high_freq) +{ + u32 msr_lo, msr_hi; + u32 save_lo, save_hi; + u32 eax, ebx, ecx, edx; + u32 try_hi; + struct cpuinfo_x86 *c = cpu_data; + + if (!low_freq || !high_freq) + return -EINVAL; + + if (cpu_has(c, X86_FEATURE_LRTI)) { + /* if the LongRun Table Interface is present, the + * detection is a bit easier: + * For minimum frequency, read out the maximum + * level (msr_hi), write that into "currently + * selected level", and read out the frequency. + * For maximum frequency, read out level zero. + */ + /* minimum */ + rdmsr(MSR_TMTA_LRTI_READOUT, msr_lo, msr_hi); + wrmsr(MSR_TMTA_LRTI_READOUT, msr_hi, msr_hi); + rdmsr(MSR_TMTA_LRTI_VOLT_MHZ, msr_lo, msr_hi); + *low_freq = msr_lo * 1000; /* to kHz */ + + /* maximum */ + wrmsr(MSR_TMTA_LRTI_READOUT, 0, msr_hi); + rdmsr(MSR_TMTA_LRTI_VOLT_MHZ, msr_lo, msr_hi); + *high_freq = msr_lo * 1000; /* to kHz */ + + dprintk("longrun table interface told %u - %u kHz\n", *low_freq, *high_freq); + + if (*low_freq > *high_freq) + *low_freq = *high_freq; + return 0; + } + + /* set the upper border to the value determined during TSC init */ + *high_freq = (cpu_khz / 1000); + *high_freq = *high_freq * 1000; + dprintk("high frequency is %u kHz\n", *high_freq); + + /* get current borders */ + rdmsr(MSR_TMTA_LONGRUN_CTRL, msr_lo, msr_hi); + save_lo = msr_lo & 0x0000007F; + save_hi = msr_hi & 0x0000007F; + + /* if current perf_pctg is larger than 90%, we need to decrease the + * upper limit to make the calculation more accurate. + */ + cpuid(0x80860007, &eax, &ebx, &ecx, &edx); + /* try decreasing in 10% steps, some processors react only + * on some barrier values */ + for (try_hi = 80; try_hi > 0 && ecx > 90; try_hi -=10) { + /* set to 0 to try_hi perf_pctg */ + msr_lo &= 0xFFFFFF80; + msr_hi &= 0xFFFFFF80; + msr_lo |= 0; + msr_hi |= try_hi; + wrmsr(MSR_TMTA_LONGRUN_CTRL, msr_lo, msr_hi); + + /* read out current core MHz and current perf_pctg */ + cpuid(0x80860007, &eax, &ebx, &ecx, &edx); + + /* restore values */ + wrmsr(MSR_TMTA_LONGRUN_CTRL, save_lo, save_hi); + } + dprintk("percentage is %u %%, freq is %u MHz\n", ecx, eax); + + /* performance_pctg = (current_freq - low_freq)/(high_freq - low_freq) + * eqals + * low_freq * ( 1 - perf_pctg) = (cur_freq - high_freq * perf_pctg) + * + * high_freq * perf_pctg is stored tempoarily into "ebx". + */ + ebx = (((cpu_khz / 1000) * ecx) / 100); /* to MHz */ + + if ((ecx > 95) || (ecx == 0) || (eax < ebx)) + return -EIO; + + edx = (eax - ebx) / (100 - ecx); + *low_freq = edx * 1000; /* back to kHz */ + + dprintk("low frequency is %u kHz\n", *low_freq); + + if (*low_freq > *high_freq) + *low_freq = *high_freq; + + return 0; +} + + +static int __init longrun_cpu_init(struct cpufreq_policy *policy) +{ + int result = 0; + + /* capability check */ + if (policy->cpu != 0) + return -ENODEV; + + /* detect low and high frequency */ + result = longrun_determine_freqs(&longrun_low_freq, &longrun_high_freq); + if (result) + return result; + + /* cpuinfo and default policy values */ + policy->cpuinfo.min_freq = longrun_low_freq; + policy->cpuinfo.max_freq = longrun_high_freq; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + longrun_get_policy(policy); + + return 0; +} + + +static struct cpufreq_driver longrun_driver = { + .flags = CPUFREQ_CONST_LOOPS, + .verify = longrun_verify_policy, + .setpolicy = longrun_set_policy, + .get = longrun_get, + .init = longrun_cpu_init, + .name = "longrun", + .owner = THIS_MODULE, +}; + + +/** + * longrun_init - initializes the Transmeta Crusoe LongRun CPUFreq driver + * + * Initializes the LongRun support. + */ +static int __init longrun_init(void) +{ + struct cpuinfo_x86 *c = cpu_data; + + if (c->x86_vendor != X86_VENDOR_TRANSMETA || + !cpu_has(c, X86_FEATURE_LONGRUN)) + return -ENODEV; + + return cpufreq_register_driver(&longrun_driver); +} + + +/** + * longrun_exit - unregisters LongRun support + */ +static void __exit longrun_exit(void) +{ + cpufreq_unregister_driver(&longrun_driver); +} + + +MODULE_AUTHOR ("Dominik Brodowski <linux@brodo.de>"); +MODULE_DESCRIPTION ("LongRun driver for Transmeta Crusoe and Efficeon processors."); +MODULE_LICENSE ("GPL"); + +module_init(longrun_init); +module_exit(longrun_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/p4-clockmod.c b/arch/i386/kernel/cpu/cpufreq/p4-clockmod.c new file mode 100644 index 00000000000..aa622d52c6e --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/p4-clockmod.c @@ -0,0 +1,337 @@ +/* + * Pentium 4/Xeon CPU on demand clock modulation/speed scaling + * (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de> + * (C) 2002 Zwane Mwaikambo <zwane@commfireservices.com> + * (C) 2002 Arjan van de Ven <arjanv@redhat.com> + * (C) 2002 Tora T. Engstad + * All Rights Reserved + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * The author(s) of this software shall not be held liable for damages + * of any nature resulting due to the use of this software. This + * software is provided AS-IS with no warranties. + * + * Date Errata Description + * 20020525 N44, O17 12.5% or 25% DC causes lockup + * + */ + +#include <linux/config.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/smp.h> +#include <linux/cpufreq.h> +#include <linux/slab.h> +#include <linux/cpumask.h> + +#include <asm/processor.h> +#include <asm/msr.h> +#include <asm/timex.h> + +#include "speedstep-lib.h" + +#define PFX "p4-clockmod: " +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "p4-clockmod", msg) + +/* + * Duty Cycle (3bits), note DC_DISABLE is not specified in + * intel docs i just use it to mean disable + */ +enum { + DC_RESV, DC_DFLT, DC_25PT, DC_38PT, DC_50PT, + DC_64PT, DC_75PT, DC_88PT, DC_DISABLE +}; + +#define DC_ENTRIES 8 + + +static int has_N44_O17_errata[NR_CPUS]; +static unsigned int stock_freq; +static struct cpufreq_driver p4clockmod_driver; +static unsigned int cpufreq_p4_get(unsigned int cpu); + +static int cpufreq_p4_setdc(unsigned int cpu, unsigned int newstate) +{ + u32 l, h; + + if (!cpu_online(cpu) || (newstate > DC_DISABLE) || (newstate == DC_RESV)) + return -EINVAL; + + rdmsr(MSR_IA32_THERM_STATUS, l, h); + + if (l & 0x01) + dprintk("CPU#%d currently thermal throttled\n", cpu); + + if (has_N44_O17_errata[cpu] && (newstate == DC_25PT || newstate == DC_DFLT)) + newstate = DC_38PT; + + rdmsr(MSR_IA32_THERM_CONTROL, l, h); + if (newstate == DC_DISABLE) { + dprintk("CPU#%d disabling modulation\n", cpu); + wrmsr(MSR_IA32_THERM_CONTROL, l & ~(1<<4), h); + } else { + dprintk("CPU#%d setting duty cycle to %d%%\n", + cpu, ((125 * newstate) / 10)); + /* bits 63 - 5 : reserved + * bit 4 : enable/disable + * bits 3-1 : duty cycle + * bit 0 : reserved + */ + l = (l & ~14); + l = l | (1<<4) | ((newstate & 0x7)<<1); + wrmsr(MSR_IA32_THERM_CONTROL, l, h); + } + + return 0; +} + + +static struct cpufreq_frequency_table p4clockmod_table[] = { + {DC_RESV, CPUFREQ_ENTRY_INVALID}, + {DC_DFLT, 0}, + {DC_25PT, 0}, + {DC_38PT, 0}, + {DC_50PT, 0}, + {DC_64PT, 0}, + {DC_75PT, 0}, + {DC_88PT, 0}, + {DC_DISABLE, 0}, + {DC_RESV, CPUFREQ_TABLE_END}, +}; + + +static int cpufreq_p4_target(struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = DC_RESV; + struct cpufreq_freqs freqs; + cpumask_t cpus_allowed; + int i; + + if (cpufreq_frequency_table_target(policy, &p4clockmod_table[0], target_freq, relation, &newstate)) + return -EINVAL; + + freqs.old = cpufreq_p4_get(policy->cpu); + freqs.new = stock_freq * p4clockmod_table[newstate].index / 8; + + if (freqs.new == freqs.old) + return 0; + + /* notifiers */ + for_each_cpu_mask(i, policy->cpus) { + freqs.cpu = i; + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + } + + /* run on each logical CPU, see section 13.15.3 of IA32 Intel Architecture Software + * Developer's Manual, Volume 3 + */ + cpus_allowed = current->cpus_allowed; + + for_each_cpu_mask(i, policy->cpus) { + cpumask_t this_cpu = cpumask_of_cpu(i); + + set_cpus_allowed(current, this_cpu); + BUG_ON(smp_processor_id() != i); + + cpufreq_p4_setdc(i, p4clockmod_table[newstate].index); + } + set_cpus_allowed(current, cpus_allowed); + + /* notifiers */ + for_each_cpu_mask(i, policy->cpus) { + freqs.cpu = i; + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + } + + return 0; +} + + +static int cpufreq_p4_verify(struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, &p4clockmod_table[0]); +} + + +static unsigned int cpufreq_p4_get_frequency(struct cpuinfo_x86 *c) +{ + if ((c->x86 == 0x06) && (c->x86_model == 0x09)) { + /* Pentium M (Banias) */ + printk(KERN_WARNING PFX "Warning: Pentium M detected. " + "The speedstep_centrino module offers voltage scaling" + " in addition of frequency scaling. You should use " + "that instead of p4-clockmod, if possible.\n"); + return speedstep_get_processor_frequency(SPEEDSTEP_PROCESSOR_PM); + } + + if ((c->x86 == 0x06) && (c->x86_model == 0x0D)) { + /* Pentium M (Dothan) */ + printk(KERN_WARNING PFX "Warning: Pentium M detected. " + "The speedstep_centrino module offers voltage scaling" + " in addition of frequency scaling. You should use " + "that instead of p4-clockmod, if possible.\n"); + /* on P-4s, the TSC runs with constant frequency independent whether + * throttling is active or not. */ + p4clockmod_driver.flags |= CPUFREQ_CONST_LOOPS; + return speedstep_get_processor_frequency(SPEEDSTEP_PROCESSOR_PM); + } + + if (c->x86 != 0xF) { + printk(KERN_WARNING PFX "Unknown p4-clockmod-capable CPU. Please send an e-mail to <linux@brodo.de>\n"); + return 0; + } + + /* on P-4s, the TSC runs with constant frequency independent whether + * throttling is active or not. */ + p4clockmod_driver.flags |= CPUFREQ_CONST_LOOPS; + + if (speedstep_detect_processor() == SPEEDSTEP_PROCESSOR_P4M) { + printk(KERN_WARNING PFX "Warning: Pentium 4-M detected. " + "The speedstep-ich or acpi cpufreq modules offer " + "voltage scaling in addition of frequency scaling. " + "You should use either one instead of p4-clockmod, " + "if possible.\n"); + return speedstep_get_processor_frequency(SPEEDSTEP_PROCESSOR_P4M); + } + + return speedstep_get_processor_frequency(SPEEDSTEP_PROCESSOR_P4D); +} + + + +static int cpufreq_p4_cpu_init(struct cpufreq_policy *policy) +{ + struct cpuinfo_x86 *c = &cpu_data[policy->cpu]; + int cpuid = 0; + unsigned int i; + +#ifdef CONFIG_SMP + policy->cpus = cpu_sibling_map[policy->cpu]; +#endif + + /* Errata workaround */ + cpuid = (c->x86 << 8) | (c->x86_model << 4) | c->x86_mask; + switch (cpuid) { + case 0x0f07: + case 0x0f0a: + case 0x0f11: + case 0x0f12: + has_N44_O17_errata[policy->cpu] = 1; + dprintk("has errata -- disabling low frequencies\n"); + } + + /* get max frequency */ + stock_freq = cpufreq_p4_get_frequency(c); + if (!stock_freq) + return -EINVAL; + + /* table init */ + for (i=1; (p4clockmod_table[i].frequency != CPUFREQ_TABLE_END); i++) { + if ((i<2) && (has_N44_O17_errata[policy->cpu])) + p4clockmod_table[i].frequency = CPUFREQ_ENTRY_INVALID; + else + p4clockmod_table[i].frequency = (stock_freq * i)/8; + } + cpufreq_frequency_table_get_attr(p4clockmod_table, policy->cpu); + + /* cpuinfo and default policy values */ + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = 1000000; /* assumed */ + policy->cur = stock_freq; + + return cpufreq_frequency_table_cpuinfo(policy, &p4clockmod_table[0]); +} + + +static int cpufreq_p4_cpu_exit(struct cpufreq_policy *policy) +{ + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + +static unsigned int cpufreq_p4_get(unsigned int cpu) +{ + cpumask_t cpus_allowed; + u32 l, h; + + cpus_allowed = current->cpus_allowed; + + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + BUG_ON(smp_processor_id() != cpu); + + rdmsr(MSR_IA32_THERM_CONTROL, l, h); + + set_cpus_allowed(current, cpus_allowed); + + if (l & 0x10) { + l = l >> 1; + l &= 0x7; + } else + l = DC_DISABLE; + + if (l != DC_DISABLE) + return (stock_freq * l / 8); + + return stock_freq; +} + +static struct freq_attr* p4clockmod_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver p4clockmod_driver = { + .verify = cpufreq_p4_verify, + .target = cpufreq_p4_target, + .init = cpufreq_p4_cpu_init, + .exit = cpufreq_p4_cpu_exit, + .get = cpufreq_p4_get, + .name = "p4-clockmod", + .owner = THIS_MODULE, + .attr = p4clockmod_attr, +}; + + +static int __init cpufreq_p4_init(void) +{ + struct cpuinfo_x86 *c = cpu_data; + int ret; + + /* + * THERM_CONTROL is architectural for IA32 now, so + * we can rely on the capability checks + */ + if (c->x86_vendor != X86_VENDOR_INTEL) + return -ENODEV; + + if (!test_bit(X86_FEATURE_ACPI, c->x86_capability) || + !test_bit(X86_FEATURE_ACC, c->x86_capability)) + return -ENODEV; + + ret = cpufreq_register_driver(&p4clockmod_driver); + if (!ret) + printk(KERN_INFO PFX "P4/Xeon(TM) CPU On-Demand Clock Modulation available\n"); + + return (ret); +} + + +static void __exit cpufreq_p4_exit(void) +{ + cpufreq_unregister_driver(&p4clockmod_driver); +} + + +MODULE_AUTHOR ("Zwane Mwaikambo <zwane@commfireservices.com>"); +MODULE_DESCRIPTION ("cpufreq driver for Pentium(TM) 4/Xeon(TM)"); +MODULE_LICENSE ("GPL"); + +late_initcall(cpufreq_p4_init); +module_exit(cpufreq_p4_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/powernow-k6.c b/arch/i386/kernel/cpu/cpufreq/powernow-k6.c new file mode 100644 index 00000000000..222f8cfe3c5 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/powernow-k6.c @@ -0,0 +1,256 @@ +/* + * This file was based upon code in Powertweak Linux (http://powertweak.sf.net) + * (C) 2000-2003 Dave Jones, Arjan van de Ven, Janne Pänkälä, Dominik Brodowski. + * + * Licensed under the terms of the GNU GPL License version 2. + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/ioport.h> +#include <linux/slab.h> + +#include <asm/msr.h> +#include <asm/timex.h> +#include <asm/io.h> + + +#define POWERNOW_IOPORT 0xfff0 /* it doesn't matter where, as long + as it is unused */ + +static unsigned int busfreq; /* FSB, in 10 kHz */ +static unsigned int max_multiplier; + + +/* Clock ratio multiplied by 10 - see table 27 in AMD#23446 */ +static struct cpufreq_frequency_table clock_ratio[] = { + {45, /* 000 -> 4.5x */ 0}, + {50, /* 001 -> 5.0x */ 0}, + {40, /* 010 -> 4.0x */ 0}, + {55, /* 011 -> 5.5x */ 0}, + {20, /* 100 -> 2.0x */ 0}, + {30, /* 101 -> 3.0x */ 0}, + {60, /* 110 -> 6.0x */ 0}, + {35, /* 111 -> 3.5x */ 0}, + {0, CPUFREQ_TABLE_END} +}; + + +/** + * powernow_k6_get_cpu_multiplier - returns the current FSB multiplier + * + * Returns the current setting of the frequency multiplier. Core clock + * speed is frequency of the Front-Side Bus multiplied with this value. + */ +static int powernow_k6_get_cpu_multiplier(void) +{ + u64 invalue = 0; + u32 msrval; + + msrval = POWERNOW_IOPORT + 0x1; + wrmsr(MSR_K6_EPMR, msrval, 0); /* enable the PowerNow port */ + invalue=inl(POWERNOW_IOPORT + 0x8); + msrval = POWERNOW_IOPORT + 0x0; + wrmsr(MSR_K6_EPMR, msrval, 0); /* disable it again */ + + return clock_ratio[(invalue >> 5)&7].index; +} + + +/** + * powernow_k6_set_state - set the PowerNow! multiplier + * @best_i: clock_ratio[best_i] is the target multiplier + * + * Tries to change the PowerNow! multiplier + */ +static void powernow_k6_set_state (unsigned int best_i) +{ + unsigned long outvalue=0, invalue=0; + unsigned long msrval; + struct cpufreq_freqs freqs; + + if (clock_ratio[best_i].index > max_multiplier) { + printk(KERN_ERR "cpufreq: invalid target frequency\n"); + return; + } + + freqs.old = busfreq * powernow_k6_get_cpu_multiplier(); + freqs.new = busfreq * clock_ratio[best_i].index; + freqs.cpu = 0; /* powernow-k6.c is UP only driver */ + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + /* we now need to transform best_i to the BVC format, see AMD#23446 */ + + outvalue = (1<<12) | (1<<10) | (1<<9) | (best_i<<5); + + msrval = POWERNOW_IOPORT + 0x1; + wrmsr(MSR_K6_EPMR, msrval, 0); /* enable the PowerNow port */ + invalue=inl(POWERNOW_IOPORT + 0x8); + invalue = invalue & 0xf; + outvalue = outvalue | invalue; + outl(outvalue ,(POWERNOW_IOPORT + 0x8)); + msrval = POWERNOW_IOPORT + 0x0; + wrmsr(MSR_K6_EPMR, msrval, 0); /* disable it again */ + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + return; +} + + +/** + * powernow_k6_verify - verifies a new CPUfreq policy + * @policy: new policy + * + * Policy must be within lowest and highest possible CPU Frequency, + * and at least one possible state must be within min and max. + */ +static int powernow_k6_verify(struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, &clock_ratio[0]); +} + + +/** + * powernow_k6_setpolicy - sets a new CPUFreq policy + * @policy: new policy + * @target_freq: the target frequency + * @relation: how that frequency relates to achieved frequency (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H) + * + * sets a new CPUFreq policy + */ +static int powernow_k6_target (struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = 0; + + if (cpufreq_frequency_table_target(policy, &clock_ratio[0], target_freq, relation, &newstate)) + return -EINVAL; + + powernow_k6_set_state(newstate); + + return 0; +} + + +static int powernow_k6_cpu_init(struct cpufreq_policy *policy) +{ + unsigned int i; + int result; + + if (policy->cpu != 0) + return -ENODEV; + + /* get frequencies */ + max_multiplier = powernow_k6_get_cpu_multiplier(); + busfreq = cpu_khz / max_multiplier; + + /* table init */ + for (i=0; (clock_ratio[i].frequency != CPUFREQ_TABLE_END); i++) { + if (clock_ratio[i].index > max_multiplier) + clock_ratio[i].frequency = CPUFREQ_ENTRY_INVALID; + else + clock_ratio[i].frequency = busfreq * clock_ratio[i].index; + } + + /* cpuinfo and default policy values */ + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = busfreq * max_multiplier; + + result = cpufreq_frequency_table_cpuinfo(policy, clock_ratio); + if (result) + return (result); + + cpufreq_frequency_table_get_attr(clock_ratio, policy->cpu); + + return 0; +} + + +static int powernow_k6_cpu_exit(struct cpufreq_policy *policy) +{ + unsigned int i; + for (i=0; i<8; i++) { + if (i==max_multiplier) + powernow_k6_set_state(i); + } + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + +static unsigned int powernow_k6_get(unsigned int cpu) +{ + return busfreq * powernow_k6_get_cpu_multiplier(); +} + +static struct freq_attr* powernow_k6_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver powernow_k6_driver = { + .verify = powernow_k6_verify, + .target = powernow_k6_target, + .init = powernow_k6_cpu_init, + .exit = powernow_k6_cpu_exit, + .get = powernow_k6_get, + .name = "powernow-k6", + .owner = THIS_MODULE, + .attr = powernow_k6_attr, +}; + + +/** + * powernow_k6_init - initializes the k6 PowerNow! CPUFreq driver + * + * Initializes the K6 PowerNow! support. Returns -ENODEV on unsupported + * devices, -EINVAL or -ENOMEM on problems during initiatization, and zero + * on success. + */ +static int __init powernow_k6_init(void) +{ + struct cpuinfo_x86 *c = cpu_data; + + if ((c->x86_vendor != X86_VENDOR_AMD) || (c->x86 != 5) || + ((c->x86_model != 12) && (c->x86_model != 13))) + return -ENODEV; + + if (!request_region(POWERNOW_IOPORT, 16, "PowerNow!")) { + printk("cpufreq: PowerNow IOPORT region already used.\n"); + return -EIO; + } + + if (cpufreq_register_driver(&powernow_k6_driver)) { + release_region (POWERNOW_IOPORT, 16); + return -EINVAL; + } + + return 0; +} + + +/** + * powernow_k6_exit - unregisters AMD K6-2+/3+ PowerNow! support + * + * Unregisters AMD K6-2+ / K6-3+ PowerNow! support. + */ +static void __exit powernow_k6_exit(void) +{ + cpufreq_unregister_driver(&powernow_k6_driver); + release_region (POWERNOW_IOPORT, 16); +} + + +MODULE_AUTHOR ("Arjan van de Ven <arjanv@redhat.com>, Dave Jones <davej@codemonkey.org.uk>, Dominik Brodowski <linux@brodo.de>"); +MODULE_DESCRIPTION ("PowerNow! driver for AMD K6-2+ / K6-3+ processors."); +MODULE_LICENSE ("GPL"); + +module_init(powernow_k6_init); +module_exit(powernow_k6_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/powernow-k7.c b/arch/i386/kernel/cpu/cpufreq/powernow-k7.c new file mode 100644 index 00000000000..913f652623d --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/powernow-k7.c @@ -0,0 +1,690 @@ +/* + * AMD K7 Powernow driver. + * (C) 2003 Dave Jones <davej@codemonkey.org.uk> on behalf of SuSE Labs. + * (C) 2003-2004 Dave Jones <davej@redhat.com> + * + * Licensed under the terms of the GNU GPL License version 2. + * Based upon datasheets & sample CPUs kindly provided by AMD. + * + * Errata 5: Processor may fail to execute a FID/VID change in presence of interrupt. + * - We cli/sti on stepping A0 CPUs around the FID/VID transition. + * Errata 15: Processors with half frequency multipliers may hang upon wakeup from disconnect. + * - We disable half multipliers if ACPI is used on A0 stepping CPUs. + */ + +#include <linux/config.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/slab.h> +#include <linux/string.h> +#include <linux/dmi.h> + +#include <asm/msr.h> +#include <asm/timex.h> +#include <asm/io.h> +#include <asm/system.h> + +#ifdef CONFIG_X86_POWERNOW_K7_ACPI +#include <linux/acpi.h> +#include <acpi/processor.h> +#endif + +#include "powernow-k7.h" + +#define PFX "powernow: " + + +struct psb_s { + u8 signature[10]; + u8 tableversion; + u8 flags; + u16 settlingtime; + u8 reserved1; + u8 numpst; +}; + +struct pst_s { + u32 cpuid; + u8 fsbspeed; + u8 maxfid; + u8 startvid; + u8 numpstates; +}; + +#ifdef CONFIG_X86_POWERNOW_K7_ACPI +union powernow_acpi_control_t { + struct { + unsigned long fid:5, + vid:5, + sgtc:20, + res1:2; + } bits; + unsigned long val; +}; +#endif + +#ifdef CONFIG_CPU_FREQ_DEBUG +/* divide by 1000 to get VCore voltage in V. */ +static int mobile_vid_table[32] = { + 2000, 1950, 1900, 1850, 1800, 1750, 1700, 1650, + 1600, 1550, 1500, 1450, 1400, 1350, 1300, 0, + 1275, 1250, 1225, 1200, 1175, 1150, 1125, 1100, + 1075, 1050, 1025, 1000, 975, 950, 925, 0, +}; +#endif + +/* divide by 10 to get FID. */ +static int fid_codes[32] = { + 110, 115, 120, 125, 50, 55, 60, 65, + 70, 75, 80, 85, 90, 95, 100, 105, + 30, 190, 40, 200, 130, 135, 140, 210, + 150, 225, 160, 165, 170, 180, -1, -1, +}; + +/* This parameter is used in order to force ACPI instead of legacy method for + * configuration purpose. + */ + +static int acpi_force; + +static struct cpufreq_frequency_table *powernow_table; + +static unsigned int can_scale_bus; +static unsigned int can_scale_vid; +static unsigned int minimum_speed=-1; +static unsigned int maximum_speed; +static unsigned int number_scales; +static unsigned int fsb; +static unsigned int latency; +static char have_a0; + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "powernow-k7", msg) + +static int check_fsb(unsigned int fsbspeed) +{ + int delta; + unsigned int f = fsb / 1000; + + delta = (fsbspeed > f) ? fsbspeed - f : f - fsbspeed; + return (delta < 5); +} + +static int check_powernow(void) +{ + struct cpuinfo_x86 *c = cpu_data; + unsigned int maxei, eax, ebx, ecx, edx; + + if ((c->x86_vendor != X86_VENDOR_AMD) || (c->x86 !=6)) { +#ifdef MODULE + printk (KERN_INFO PFX "This module only works with AMD K7 CPUs\n"); +#endif + return 0; + } + + /* Get maximum capabilities */ + maxei = cpuid_eax (0x80000000); + if (maxei < 0x80000007) { /* Any powernow info ? */ +#ifdef MODULE + printk (KERN_INFO PFX "No powernow capabilities detected\n"); +#endif + return 0; + } + + if ((c->x86_model == 6) && (c->x86_mask == 0)) { + printk (KERN_INFO PFX "K7 660[A0] core detected, enabling errata workarounds\n"); + have_a0 = 1; + } + + cpuid(0x80000007, &eax, &ebx, &ecx, &edx); + + /* Check we can actually do something before we say anything.*/ + if (!(edx & (1 << 1 | 1 << 2))) + return 0; + + printk (KERN_INFO PFX "PowerNOW! Technology present. Can scale: "); + + if (edx & 1 << 1) { + printk ("frequency"); + can_scale_bus=1; + } + + if ((edx & (1 << 1 | 1 << 2)) == 0x6) + printk (" and "); + + if (edx & 1 << 2) { + printk ("voltage"); + can_scale_vid=1; + } + + printk (".\n"); + return 1; +} + + +static int get_ranges (unsigned char *pst) +{ + unsigned int j; + unsigned int speed; + u8 fid, vid; + + powernow_table = kmalloc((sizeof(struct cpufreq_frequency_table) * (number_scales + 1)), GFP_KERNEL); + if (!powernow_table) + return -ENOMEM; + memset(powernow_table, 0, (sizeof(struct cpufreq_frequency_table) * (number_scales + 1))); + + for (j=0 ; j < number_scales; j++) { + fid = *pst++; + + powernow_table[j].frequency = (fsb * fid_codes[fid]) / 10; + powernow_table[j].index = fid; /* lower 8 bits */ + + speed = powernow_table[j].frequency; + + if ((fid_codes[fid] % 10)==5) { +#ifdef CONFIG_X86_POWERNOW_K7_ACPI + if (have_a0 == 1) + powernow_table[j].frequency = CPUFREQ_ENTRY_INVALID; +#endif + } + + if (speed < minimum_speed) + minimum_speed = speed; + if (speed > maximum_speed) + maximum_speed = speed; + + vid = *pst++; + powernow_table[j].index |= (vid << 8); /* upper 8 bits */ + + dprintk (" FID: 0x%x (%d.%dx [%dMHz]) " + "VID: 0x%x (%d.%03dV)\n", fid, fid_codes[fid] / 10, + fid_codes[fid] % 10, speed/1000, vid, + mobile_vid_table[vid]/1000, + mobile_vid_table[vid]%1000); + } + powernow_table[number_scales].frequency = CPUFREQ_TABLE_END; + powernow_table[number_scales].index = 0; + + return 0; +} + + +static void change_FID(int fid) +{ + union msr_fidvidctl fidvidctl; + + rdmsrl (MSR_K7_FID_VID_CTL, fidvidctl.val); + if (fidvidctl.bits.FID != fid) { + fidvidctl.bits.SGTC = latency; + fidvidctl.bits.FID = fid; + fidvidctl.bits.VIDC = 0; + fidvidctl.bits.FIDC = 1; + wrmsrl (MSR_K7_FID_VID_CTL, fidvidctl.val); + } +} + + +static void change_VID(int vid) +{ + union msr_fidvidctl fidvidctl; + + rdmsrl (MSR_K7_FID_VID_CTL, fidvidctl.val); + if (fidvidctl.bits.VID != vid) { + fidvidctl.bits.SGTC = latency; + fidvidctl.bits.VID = vid; + fidvidctl.bits.FIDC = 0; + fidvidctl.bits.VIDC = 1; + wrmsrl (MSR_K7_FID_VID_CTL, fidvidctl.val); + } +} + + +static void change_speed (unsigned int index) +{ + u8 fid, vid; + struct cpufreq_freqs freqs; + union msr_fidvidstatus fidvidstatus; + int cfid; + + /* fid are the lower 8 bits of the index we stored into + * the cpufreq frequency table in powernow_decode_bios, + * vid are the upper 8 bits. + */ + + fid = powernow_table[index].index & 0xFF; + vid = (powernow_table[index].index & 0xFF00) >> 8; + + freqs.cpu = 0; + + rdmsrl (MSR_K7_FID_VID_STATUS, fidvidstatus.val); + cfid = fidvidstatus.bits.CFID; + freqs.old = fsb * fid_codes[cfid] / 10; + + freqs.new = powernow_table[index].frequency; + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + /* Now do the magic poking into the MSRs. */ + + if (have_a0 == 1) /* A0 errata 5 */ + local_irq_disable(); + + if (freqs.old > freqs.new) { + /* Going down, so change FID first */ + change_FID(fid); + change_VID(vid); + } else { + /* Going up, so change VID first */ + change_VID(vid); + change_FID(fid); + } + + + if (have_a0 == 1) + local_irq_enable(); + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); +} + + +#ifdef CONFIG_X86_POWERNOW_K7_ACPI + +static struct acpi_processor_performance *acpi_processor_perf; + +static int powernow_acpi_init(void) +{ + int i; + int retval = 0; + union powernow_acpi_control_t pc; + + if (acpi_processor_perf != NULL && powernow_table != NULL) { + retval = -EINVAL; + goto err0; + } + + acpi_processor_perf = kmalloc(sizeof(struct acpi_processor_performance), + GFP_KERNEL); + + if (!acpi_processor_perf) { + retval = -ENOMEM; + goto err0; + } + + memset(acpi_processor_perf, 0, sizeof(struct acpi_processor_performance)); + + if (acpi_processor_register_performance(acpi_processor_perf, 0)) { + retval = -EIO; + goto err1; + } + + if (acpi_processor_perf->control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) { + retval = -ENODEV; + goto err2; + } + + if (acpi_processor_perf->status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) { + retval = -ENODEV; + goto err2; + } + + number_scales = acpi_processor_perf->state_count; + + if (number_scales < 2) { + retval = -ENODEV; + goto err2; + } + + powernow_table = kmalloc((number_scales + 1) * (sizeof(struct cpufreq_frequency_table)), GFP_KERNEL); + if (!powernow_table) { + retval = -ENOMEM; + goto err2; + } + + memset(powernow_table, 0, ((number_scales + 1) * sizeof(struct cpufreq_frequency_table))); + + pc.val = (unsigned long) acpi_processor_perf->states[0].control; + for (i = 0; i < number_scales; i++) { + u8 fid, vid; + unsigned int speed; + + pc.val = (unsigned long) acpi_processor_perf->states[i].control; + dprintk ("acpi: P%d: %d MHz %d mW %d uS control %08x SGTC %d\n", + i, + (u32) acpi_processor_perf->states[i].core_frequency, + (u32) acpi_processor_perf->states[i].power, + (u32) acpi_processor_perf->states[i].transition_latency, + (u32) acpi_processor_perf->states[i].control, + pc.bits.sgtc); + + vid = pc.bits.vid; + fid = pc.bits.fid; + + powernow_table[i].frequency = fsb * fid_codes[fid] / 10; + powernow_table[i].index = fid; /* lower 8 bits */ + powernow_table[i].index |= (vid << 8); /* upper 8 bits */ + + speed = powernow_table[i].frequency; + + if ((fid_codes[fid] % 10)==5) { + if (have_a0 == 1) + powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID; + } + + dprintk (" FID: 0x%x (%d.%dx [%dMHz]) " + "VID: 0x%x (%d.%03dV)\n", fid, fid_codes[fid] / 10, + fid_codes[fid] % 10, speed/1000, vid, + mobile_vid_table[vid]/1000, + mobile_vid_table[vid]%1000); + + if (latency < pc.bits.sgtc) + latency = pc.bits.sgtc; + + if (speed < minimum_speed) + minimum_speed = speed; + if (speed > maximum_speed) + maximum_speed = speed; + } + + powernow_table[i].frequency = CPUFREQ_TABLE_END; + powernow_table[i].index = 0; + + /* notify BIOS that we exist */ + acpi_processor_notify_smm(THIS_MODULE); + + return 0; + +err2: + acpi_processor_unregister_performance(acpi_processor_perf, 0); +err1: + kfree(acpi_processor_perf); +err0: + printk(KERN_WARNING PFX "ACPI perflib can not be used in this platform\n"); + acpi_processor_perf = NULL; + return retval; +} +#else +static int powernow_acpi_init(void) +{ + printk(KERN_INFO PFX "no support for ACPI processor found." + " Please recompile your kernel with ACPI processor\n"); + return -EINVAL; +} +#endif + +static int powernow_decode_bios (int maxfid, int startvid) +{ + struct psb_s *psb; + struct pst_s *pst; + unsigned int i, j; + unsigned char *p; + unsigned int etuple; + unsigned int ret; + + etuple = cpuid_eax(0x80000001); + + for (i=0xC0000; i < 0xffff0 ; i+=16) { + + p = phys_to_virt(i); + + if (memcmp(p, "AMDK7PNOW!", 10) == 0){ + dprintk ("Found PSB header at %p\n", p); + psb = (struct psb_s *) p; + dprintk ("Table version: 0x%x\n", psb->tableversion); + if (psb->tableversion != 0x12) { + printk (KERN_INFO PFX "Sorry, only v1.2 tables supported right now\n"); + return -ENODEV; + } + + dprintk ("Flags: 0x%x\n", psb->flags); + if ((psb->flags & 1)==0) { + dprintk ("Mobile voltage regulator\n"); + } else { + dprintk ("Desktop voltage regulator\n"); + } + + latency = psb->settlingtime; + if (latency < 100) { + printk (KERN_INFO PFX "BIOS set settling time to %d microseconds." + "Should be at least 100. Correcting.\n", latency); + latency = 100; + } + dprintk ("Settling Time: %d microseconds.\n", psb->settlingtime); + dprintk ("Has %d PST tables. (Only dumping ones relevant to this CPU).\n", psb->numpst); + + p += sizeof (struct psb_s); + + pst = (struct pst_s *) p; + + for (i = 0 ; i <psb->numpst; i++) { + pst = (struct pst_s *) p; + number_scales = pst->numpstates; + + if ((etuple == pst->cpuid) && check_fsb(pst->fsbspeed) && + (maxfid==pst->maxfid) && (startvid==pst->startvid)) + { + dprintk ("PST:%d (@%p)\n", i, pst); + dprintk (" cpuid: 0x%x fsb: %d maxFID: 0x%x startvid: 0x%x\n", + pst->cpuid, pst->fsbspeed, pst->maxfid, pst->startvid); + + ret = get_ranges ((char *) pst + sizeof (struct pst_s)); + return ret; + + } else { + p = (char *) pst + sizeof (struct pst_s); + for (j=0 ; j < number_scales; j++) + p+=2; + } + } + printk (KERN_INFO PFX "No PST tables match this cpuid (0x%x)\n", etuple); + printk (KERN_INFO PFX "This is indicative of a broken BIOS.\n"); + + return -EINVAL; + } + p++; + } + + return -ENODEV; +} + + +static int powernow_target (struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate; + + if (cpufreq_frequency_table_target(policy, powernow_table, target_freq, relation, &newstate)) + return -EINVAL; + + change_speed(newstate); + + return 0; +} + + +static int powernow_verify (struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, powernow_table); +} + +/* + * We use the fact that the bus frequency is somehow + * a multiple of 100000/3 khz, then we compute sgtc according + * to this multiple. + * That way, we match more how AMD thinks all of that work. + * We will then get the same kind of behaviour already tested under + * the "well-known" other OS. + */ +static int __init fixup_sgtc(void) +{ + unsigned int sgtc; + unsigned int m; + + m = fsb / 3333; + if ((m % 10) >= 5) + m += 5; + + m /= 10; + + sgtc = 100 * m * latency; + sgtc = sgtc / 3; + if (sgtc > 0xfffff) { + printk(KERN_WARNING PFX "SGTC too large %d\n", sgtc); + sgtc = 0xfffff; + } + return sgtc; +} + +static unsigned int powernow_get(unsigned int cpu) +{ + union msr_fidvidstatus fidvidstatus; + unsigned int cfid; + + if (cpu) + return 0; + rdmsrl (MSR_K7_FID_VID_STATUS, fidvidstatus.val); + cfid = fidvidstatus.bits.CFID; + + return (fsb * fid_codes[cfid] / 10); +} + + +static int __init acer_cpufreq_pst(struct dmi_system_id *d) +{ + printk(KERN_WARNING "%s laptop with broken PST tables in BIOS detected.\n", d->ident); + printk(KERN_WARNING "You need to downgrade to 3A21 (09/09/2002), or try a newer BIOS than 3A71 (01/20/2003)\n"); + printk(KERN_WARNING "cpufreq scaling has been disabled as a result of this.\n"); + return 0; +} + +/* + * Some Athlon laptops have really fucked PST tables. + * A BIOS update is all that can save them. + * Mention this, and disable cpufreq. + */ +static struct dmi_system_id __initdata powernow_dmi_table[] = { + { + .callback = acer_cpufreq_pst, + .ident = "Acer Aspire", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Insyde Software"), + DMI_MATCH(DMI_BIOS_VERSION, "3A71"), + }, + }, + { } +}; + +static int __init powernow_cpu_init (struct cpufreq_policy *policy) +{ + union msr_fidvidstatus fidvidstatus; + int result; + + if (policy->cpu != 0) + return -ENODEV; + + rdmsrl (MSR_K7_FID_VID_STATUS, fidvidstatus.val); + + /* A K7 with powernow technology is set to max frequency by BIOS */ + fsb = (10 * cpu_khz) / fid_codes[fidvidstatus.bits.MFID]; + if (!fsb) { + printk(KERN_WARNING PFX "can not determine bus frequency\n"); + return -EINVAL; + } + dprintk("FSB: %3d.%03d MHz\n", fsb/1000, fsb%1000); + + if (dmi_check_system(powernow_dmi_table) || acpi_force) { + printk (KERN_INFO PFX "PSB/PST known to be broken. Trying ACPI instead\n"); + result = powernow_acpi_init(); + } else { + result = powernow_decode_bios(fidvidstatus.bits.MFID, fidvidstatus.bits.SVID); + if (result) { + printk (KERN_INFO PFX "Trying ACPI perflib\n"); + maximum_speed = 0; + minimum_speed = -1; + latency = 0; + result = powernow_acpi_init(); + if (result) { + printk (KERN_INFO PFX "ACPI and legacy methods failed\n"); + printk (KERN_INFO PFX "See http://www.codemonkey.org.uk/projects/cpufreq/powernow-k7.shtml\n"); + } + } else { + /* SGTC use the bus clock as timer */ + latency = fixup_sgtc(); + printk(KERN_INFO PFX "SGTC: %d\n", latency); + } + } + + if (result) + return result; + + printk (KERN_INFO PFX "Minimum speed %d MHz. Maximum speed %d MHz.\n", + minimum_speed/1000, maximum_speed/1000); + + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + + policy->cpuinfo.transition_latency = cpufreq_scale(2000000UL, fsb, latency); + + policy->cur = powernow_get(0); + + cpufreq_frequency_table_get_attr(powernow_table, policy->cpu); + + return cpufreq_frequency_table_cpuinfo(policy, powernow_table); +} + +static int powernow_cpu_exit (struct cpufreq_policy *policy) { + cpufreq_frequency_table_put_attr(policy->cpu); + +#ifdef CONFIG_X86_POWERNOW_K7_ACPI + if (acpi_processor_perf) { + acpi_processor_unregister_performance(acpi_processor_perf, 0); + kfree(acpi_processor_perf); + } +#endif + + if (powernow_table) + kfree(powernow_table); + + return 0; +} + +static struct freq_attr* powernow_table_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver powernow_driver = { + .verify = powernow_verify, + .target = powernow_target, + .get = powernow_get, + .init = powernow_cpu_init, + .exit = powernow_cpu_exit, + .name = "powernow-k7", + .owner = THIS_MODULE, + .attr = powernow_table_attr, +}; + +static int __init powernow_init (void) +{ + if (check_powernow()==0) + return -ENODEV; + return cpufreq_register_driver(&powernow_driver); +} + + +static void __exit powernow_exit (void) +{ + cpufreq_unregister_driver(&powernow_driver); +} + +module_param(acpi_force, int, 0444); +MODULE_PARM_DESC(acpi_force, "Force ACPI to be used."); + +MODULE_AUTHOR ("Dave Jones <davej@codemonkey.org.uk>"); +MODULE_DESCRIPTION ("Powernow driver for AMD K7 processors."); +MODULE_LICENSE ("GPL"); + +late_initcall(powernow_init); +module_exit(powernow_exit); + diff --git a/arch/i386/kernel/cpu/cpufreq/powernow-k7.h b/arch/i386/kernel/cpu/cpufreq/powernow-k7.h new file mode 100644 index 00000000000..f8a63b3664e --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/powernow-k7.h @@ -0,0 +1,44 @@ +/* + * $Id: powernow-k7.h,v 1.2 2003/02/10 18:26:01 davej Exp $ + * (C) 2003 Dave Jones. + * + * Licensed under the terms of the GNU GPL License version 2. + * + * AMD-specific information + * + */ + +union msr_fidvidctl { + struct { + unsigned FID:5, // 4:0 + reserved1:3, // 7:5 + VID:5, // 12:8 + reserved2:3, // 15:13 + FIDC:1, // 16 + VIDC:1, // 17 + reserved3:2, // 19:18 + FIDCHGRATIO:1, // 20 + reserved4:11, // 31-21 + SGTC:20, // 32:51 + reserved5:12; // 63:52 + } bits; + unsigned long long val; +}; + +union msr_fidvidstatus { + struct { + unsigned CFID:5, // 4:0 + reserved1:3, // 7:5 + SFID:5, // 12:8 + reserved2:3, // 15:13 + MFID:5, // 20:16 + reserved3:11, // 31:21 + CVID:5, // 36:32 + reserved4:3, // 39:37 + SVID:5, // 44:40 + reserved5:3, // 47:45 + MVID:5, // 52:48 + reserved6:11; // 63:53 + } bits; + unsigned long long val; +}; diff --git a/arch/i386/kernel/cpu/cpufreq/powernow-k8.c b/arch/i386/kernel/cpu/cpufreq/powernow-k8.c new file mode 100644 index 00000000000..a65ff7e32e5 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/powernow-k8.c @@ -0,0 +1,1135 @@ +/* + * (c) 2003, 2004 Advanced Micro Devices, Inc. + * Your use of this code is subject to the terms and conditions of the + * GNU general public license version 2. See "COPYING" or + * http://www.gnu.org/licenses/gpl.html + * + * Support : paul.devriendt@amd.com + * + * Based on the powernow-k7.c module written by Dave Jones. + * (C) 2003 Dave Jones <davej@codemonkey.org.uk> on behalf of SuSE Labs + * (C) 2004 Dominik Brodowski <linux@brodo.de> + * (C) 2004 Pavel Machek <pavel@suse.cz> + * Licensed under the terms of the GNU GPL License version 2. + * Based upon datasheets & sample CPUs kindly provided by AMD. + * + * Valuable input gratefully received from Dave Jones, Pavel Machek, + * Dominik Brodowski, and others. + * Processor information obtained from Chapter 9 (Power and Thermal Management) + * of the "BIOS and Kernel Developer's Guide for the AMD Athlon 64 and AMD + * Opteron Processors" available for download from www.amd.com + * + * Tables for specific CPUs can be infrerred from + * http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf + */ + +#include <linux/kernel.h> +#include <linux/smp.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/slab.h> +#include <linux/string.h> + +#include <asm/msr.h> +#include <asm/io.h> +#include <asm/delay.h> + +#ifdef CONFIG_X86_POWERNOW_K8_ACPI +#include <linux/acpi.h> +#include <acpi/processor.h> +#endif + +#define PFX "powernow-k8: " +#define BFX PFX "BIOS error: " +#define VERSION "version 1.00.09e" +#include "powernow-k8.h" + +/* serialize freq changes */ +static DECLARE_MUTEX(fidvid_sem); + +static struct powernow_k8_data *powernow_data[NR_CPUS]; + +/* Return a frequency in MHz, given an input fid */ +static u32 find_freq_from_fid(u32 fid) +{ + return 800 + (fid * 100); +} + +/* Return a frequency in KHz, given an input fid */ +static u32 find_khz_freq_from_fid(u32 fid) +{ + return 1000 * find_freq_from_fid(fid); +} + +/* Return a voltage in miliVolts, given an input vid */ +static u32 find_millivolts_from_vid(struct powernow_k8_data *data, u32 vid) +{ + return 1550-vid*25; +} + +/* Return the vco fid for an input fid + * + * Each "low" fid has corresponding "high" fid, and you can get to "low" fids + * only from corresponding high fids. This returns "high" fid corresponding to + * "low" one. + */ +static u32 convert_fid_to_vco_fid(u32 fid) +{ + if (fid < HI_FID_TABLE_BOTTOM) { + return 8 + (2 * fid); + } else { + return fid; + } +} + +/* + * Return 1 if the pending bit is set. Unless we just instructed the processor + * to transition to a new state, seeing this bit set is really bad news. + */ +static int pending_bit_stuck(void) +{ + u32 lo, hi; + + rdmsr(MSR_FIDVID_STATUS, lo, hi); + return lo & MSR_S_LO_CHANGE_PENDING ? 1 : 0; +} + +/* + * Update the global current fid / vid values from the status msr. + * Returns 1 on error. + */ +static int query_current_values_with_pending_wait(struct powernow_k8_data *data) +{ + u32 lo, hi; + u32 i = 0; + + lo = MSR_S_LO_CHANGE_PENDING; + while (lo & MSR_S_LO_CHANGE_PENDING) { + if (i++ > 0x1000000) { + printk(KERN_ERR PFX "detected change pending stuck\n"); + return 1; + } + rdmsr(MSR_FIDVID_STATUS, lo, hi); + } + + data->currvid = hi & MSR_S_HI_CURRENT_VID; + data->currfid = lo & MSR_S_LO_CURRENT_FID; + + return 0; +} + +/* the isochronous relief time */ +static void count_off_irt(struct powernow_k8_data *data) +{ + udelay((1 << data->irt) * 10); + return; +} + +/* the voltage stabalization time */ +static void count_off_vst(struct powernow_k8_data *data) +{ + udelay(data->vstable * VST_UNITS_20US); + return; +} + +/* need to init the control msr to a safe value (for each cpu) */ +static void fidvid_msr_init(void) +{ + u32 lo, hi; + u8 fid, vid; + + rdmsr(MSR_FIDVID_STATUS, lo, hi); + vid = hi & MSR_S_HI_CURRENT_VID; + fid = lo & MSR_S_LO_CURRENT_FID; + lo = fid | (vid << MSR_C_LO_VID_SHIFT); + hi = MSR_C_HI_STP_GNT_BENIGN; + dprintk("cpu%d, init lo 0x%x, hi 0x%x\n", smp_processor_id(), lo, hi); + wrmsr(MSR_FIDVID_CTL, lo, hi); +} + + +/* write the new fid value along with the other control fields to the msr */ +static int write_new_fid(struct powernow_k8_data *data, u32 fid) +{ + u32 lo; + u32 savevid = data->currvid; + + if ((fid & INVALID_FID_MASK) || (data->currvid & INVALID_VID_MASK)) { + printk(KERN_ERR PFX "internal error - overflow on fid write\n"); + return 1; + } + + lo = fid | (data->currvid << MSR_C_LO_VID_SHIFT) | MSR_C_LO_INIT_FID_VID; + + dprintk("writing fid 0x%x, lo 0x%x, hi 0x%x\n", + fid, lo, data->plllock * PLL_LOCK_CONVERSION); + + wrmsr(MSR_FIDVID_CTL, lo, data->plllock * PLL_LOCK_CONVERSION); + + if (query_current_values_with_pending_wait(data)) + return 1; + + count_off_irt(data); + + if (savevid != data->currvid) { + printk(KERN_ERR PFX "vid change on fid trans, old 0x%x, new 0x%x\n", + savevid, data->currvid); + return 1; + } + + if (fid != data->currfid) { + printk(KERN_ERR PFX "fid trans failed, fid 0x%x, curr 0x%x\n", fid, + data->currfid); + return 1; + } + + return 0; +} + +/* Write a new vid to the hardware */ +static int write_new_vid(struct powernow_k8_data *data, u32 vid) +{ + u32 lo; + u32 savefid = data->currfid; + + if ((data->currfid & INVALID_FID_MASK) || (vid & INVALID_VID_MASK)) { + printk(KERN_ERR PFX "internal error - overflow on vid write\n"); + return 1; + } + + lo = data->currfid | (vid << MSR_C_LO_VID_SHIFT) | MSR_C_LO_INIT_FID_VID; + + dprintk("writing vid 0x%x, lo 0x%x, hi 0x%x\n", + vid, lo, STOP_GRANT_5NS); + + wrmsr(MSR_FIDVID_CTL, lo, STOP_GRANT_5NS); + + if (query_current_values_with_pending_wait(data)) + return 1; + + if (savefid != data->currfid) { + printk(KERN_ERR PFX "fid changed on vid trans, old 0x%x new 0x%x\n", + savefid, data->currfid); + return 1; + } + + if (vid != data->currvid) { + printk(KERN_ERR PFX "vid trans failed, vid 0x%x, curr 0x%x\n", vid, + data->currvid); + return 1; + } + + return 0; +} + +/* + * Reduce the vid by the max of step or reqvid. + * Decreasing vid codes represent increasing voltages: + * vid of 0 is 1.550V, vid of 0x1e is 0.800V, vid of 0x1f is off. + */ +static int decrease_vid_code_by_step(struct powernow_k8_data *data, u32 reqvid, u32 step) +{ + if ((data->currvid - reqvid) > step) + reqvid = data->currvid - step; + + if (write_new_vid(data, reqvid)) + return 1; + + count_off_vst(data); + + return 0; +} + +/* Change the fid and vid, by the 3 phases. */ +static int transition_fid_vid(struct powernow_k8_data *data, u32 reqfid, u32 reqvid) +{ + if (core_voltage_pre_transition(data, reqvid)) + return 1; + + if (core_frequency_transition(data, reqfid)) + return 1; + + if (core_voltage_post_transition(data, reqvid)) + return 1; + + if (query_current_values_with_pending_wait(data)) + return 1; + + if ((reqfid != data->currfid) || (reqvid != data->currvid)) { + printk(KERN_ERR PFX "failed (cpu%d): req 0x%x 0x%x, curr 0x%x 0x%x\n", + smp_processor_id(), + reqfid, reqvid, data->currfid, data->currvid); + return 1; + } + + dprintk("transitioned (cpu%d): new fid 0x%x, vid 0x%x\n", + smp_processor_id(), data->currfid, data->currvid); + + return 0; +} + +/* Phase 1 - core voltage transition ... setup voltage */ +static int core_voltage_pre_transition(struct powernow_k8_data *data, u32 reqvid) +{ + u32 rvosteps = data->rvo; + u32 savefid = data->currfid; + + dprintk("ph1 (cpu%d): start, currfid 0x%x, currvid 0x%x, reqvid 0x%x, rvo 0x%x\n", + smp_processor_id(), + data->currfid, data->currvid, reqvid, data->rvo); + + while (data->currvid > reqvid) { + dprintk("ph1: curr 0x%x, req vid 0x%x\n", + data->currvid, reqvid); + if (decrease_vid_code_by_step(data, reqvid, data->vidmvs)) + return 1; + } + + while ((rvosteps > 0) && ((data->rvo + data->currvid) > reqvid)) { + if (data->currvid == 0) { + rvosteps = 0; + } else { + dprintk("ph1: changing vid for rvo, req 0x%x\n", + data->currvid - 1); + if (decrease_vid_code_by_step(data, data->currvid - 1, 1)) + return 1; + rvosteps--; + } + } + + if (query_current_values_with_pending_wait(data)) + return 1; + + if (savefid != data->currfid) { + printk(KERN_ERR PFX "ph1 err, currfid changed 0x%x\n", data->currfid); + return 1; + } + + dprintk("ph1 complete, currfid 0x%x, currvid 0x%x\n", + data->currfid, data->currvid); + + return 0; +} + +/* Phase 2 - core frequency transition */ +static int core_frequency_transition(struct powernow_k8_data *data, u32 reqfid) +{ + u32 vcoreqfid, vcocurrfid, vcofiddiff, savevid = data->currvid; + + if ((reqfid < HI_FID_TABLE_BOTTOM) && (data->currfid < HI_FID_TABLE_BOTTOM)) { + printk(KERN_ERR PFX "ph2: illegal lo-lo transition 0x%x 0x%x\n", + reqfid, data->currfid); + return 1; + } + + if (data->currfid == reqfid) { + printk(KERN_ERR PFX "ph2 null fid transition 0x%x\n", data->currfid); + return 0; + } + + dprintk("ph2 (cpu%d): starting, currfid 0x%x, currvid 0x%x, reqfid 0x%x\n", + smp_processor_id(), + data->currfid, data->currvid, reqfid); + + vcoreqfid = convert_fid_to_vco_fid(reqfid); + vcocurrfid = convert_fid_to_vco_fid(data->currfid); + vcofiddiff = vcocurrfid > vcoreqfid ? vcocurrfid - vcoreqfid + : vcoreqfid - vcocurrfid; + + while (vcofiddiff > 2) { + if (reqfid > data->currfid) { + if (data->currfid > LO_FID_TABLE_TOP) { + if (write_new_fid(data, data->currfid + 2)) { + return 1; + } + } else { + if (write_new_fid + (data, 2 + convert_fid_to_vco_fid(data->currfid))) { + return 1; + } + } + } else { + if (write_new_fid(data, data->currfid - 2)) + return 1; + } + + vcocurrfid = convert_fid_to_vco_fid(data->currfid); + vcofiddiff = vcocurrfid > vcoreqfid ? vcocurrfid - vcoreqfid + : vcoreqfid - vcocurrfid; + } + + if (write_new_fid(data, reqfid)) + return 1; + + if (query_current_values_with_pending_wait(data)) + return 1; + + if (data->currfid != reqfid) { + printk(KERN_ERR PFX + "ph2: mismatch, failed fid transition, curr 0x%x, req 0x%x\n", + data->currfid, reqfid); + return 1; + } + + if (savevid != data->currvid) { + printk(KERN_ERR PFX "ph2: vid changed, save 0x%x, curr 0x%x\n", + savevid, data->currvid); + return 1; + } + + dprintk("ph2 complete, currfid 0x%x, currvid 0x%x\n", + data->currfid, data->currvid); + + return 0; +} + +/* Phase 3 - core voltage transition flow ... jump to the final vid. */ +static int core_voltage_post_transition(struct powernow_k8_data *data, u32 reqvid) +{ + u32 savefid = data->currfid; + u32 savereqvid = reqvid; + + dprintk("ph3 (cpu%d): starting, currfid 0x%x, currvid 0x%x\n", + smp_processor_id(), + data->currfid, data->currvid); + + if (reqvid != data->currvid) { + if (write_new_vid(data, reqvid)) + return 1; + + if (savefid != data->currfid) { + printk(KERN_ERR PFX + "ph3: bad fid change, save 0x%x, curr 0x%x\n", + savefid, data->currfid); + return 1; + } + + if (data->currvid != reqvid) { + printk(KERN_ERR PFX + "ph3: failed vid transition\n, req 0x%x, curr 0x%x", + reqvid, data->currvid); + return 1; + } + } + + if (query_current_values_with_pending_wait(data)) + return 1; + + if (savereqvid != data->currvid) { + dprintk("ph3 failed, currvid 0x%x\n", data->currvid); + return 1; + } + + if (savefid != data->currfid) { + dprintk("ph3 failed, currfid changed 0x%x\n", + data->currfid); + return 1; + } + + dprintk("ph3 complete, currfid 0x%x, currvid 0x%x\n", + data->currfid, data->currvid); + + return 0; +} + +static int check_supported_cpu(unsigned int cpu) +{ + cpumask_t oldmask = CPU_MASK_ALL; + u32 eax, ebx, ecx, edx; + unsigned int rc = 0; + + oldmask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + schedule(); + + if (smp_processor_id() != cpu) { + printk(KERN_ERR "limiting to cpu %u failed\n", cpu); + goto out; + } + + if (current_cpu_data.x86_vendor != X86_VENDOR_AMD) + goto out; + + eax = cpuid_eax(CPUID_PROCESSOR_SIGNATURE); + if (((eax & CPUID_USE_XFAM_XMOD) != CPUID_USE_XFAM_XMOD) || + ((eax & CPUID_XFAM) != CPUID_XFAM_K8) || + ((eax & CPUID_XMOD) > CPUID_XMOD_REV_E)) { + printk(KERN_INFO PFX "Processor cpuid %x not supported\n", eax); + goto out; + } + + eax = cpuid_eax(CPUID_GET_MAX_CAPABILITIES); + if (eax < CPUID_FREQ_VOLT_CAPABILITIES) { + printk(KERN_INFO PFX + "No frequency change capabilities detected\n"); + goto out; + } + + cpuid(CPUID_FREQ_VOLT_CAPABILITIES, &eax, &ebx, &ecx, &edx); + if ((edx & P_STATE_TRANSITION_CAPABLE) != P_STATE_TRANSITION_CAPABLE) { + printk(KERN_INFO PFX "Power state transitions not supported\n"); + goto out; + } + + rc = 1; + +out: + set_cpus_allowed(current, oldmask); + schedule(); + return rc; + +} + +static int check_pst_table(struct powernow_k8_data *data, struct pst_s *pst, u8 maxvid) +{ + unsigned int j; + u8 lastfid = 0xff; + + for (j = 0; j < data->numps; j++) { + if (pst[j].vid > LEAST_VID) { + printk(KERN_ERR PFX "vid %d invalid : 0x%x\n", j, pst[j].vid); + return -EINVAL; + } + if (pst[j].vid < data->rvo) { /* vid + rvo >= 0 */ + printk(KERN_ERR BFX "0 vid exceeded with pstate %d\n", j); + return -ENODEV; + } + if (pst[j].vid < maxvid + data->rvo) { /* vid + rvo >= maxvid */ + printk(KERN_ERR BFX "maxvid exceeded with pstate %d\n", j); + return -ENODEV; + } + if ((pst[j].fid > MAX_FID) + || (pst[j].fid & 1) + || (j && (pst[j].fid < HI_FID_TABLE_BOTTOM))) { + /* Only first fid is allowed to be in "low" range */ + printk(KERN_ERR PFX "two low fids - %d : 0x%x\n", j, pst[j].fid); + return -EINVAL; + } + if (pst[j].fid < lastfid) + lastfid = pst[j].fid; + } + if (lastfid & 1) { + printk(KERN_ERR PFX "lastfid invalid\n"); + return -EINVAL; + } + if (lastfid > LO_FID_TABLE_TOP) + printk(KERN_INFO PFX "first fid not from lo freq table\n"); + + return 0; +} + +static void print_basics(struct powernow_k8_data *data) +{ + int j; + for (j = 0; j < data->numps; j++) { + if (data->powernow_table[j].frequency != CPUFREQ_ENTRY_INVALID) + printk(KERN_INFO PFX " %d : fid 0x%x (%d MHz), vid 0x%x (%d mV)\n", j, + data->powernow_table[j].index & 0xff, + data->powernow_table[j].frequency/1000, + data->powernow_table[j].index >> 8, + find_millivolts_from_vid(data, data->powernow_table[j].index >> 8)); + } + if (data->batps) + printk(KERN_INFO PFX "Only %d pstates on battery\n", data->batps); +} + +static int fill_powernow_table(struct powernow_k8_data *data, struct pst_s *pst, u8 maxvid) +{ + struct cpufreq_frequency_table *powernow_table; + unsigned int j; + + if (data->batps) { /* use ACPI support to get full speed on mains power */ + printk(KERN_WARNING PFX "Only %d pstates usable (use ACPI driver for full range\n", data->batps); + data->numps = data->batps; + } + + for ( j=1; j<data->numps; j++ ) { + if (pst[j-1].fid >= pst[j].fid) { + printk(KERN_ERR PFX "PST out of sequence\n"); + return -EINVAL; + } + } + + if (data->numps < 2) { + printk(KERN_ERR PFX "no p states to transition\n"); + return -ENODEV; + } + + if (check_pst_table(data, pst, maxvid)) + return -EINVAL; + + powernow_table = kmalloc((sizeof(struct cpufreq_frequency_table) + * (data->numps + 1)), GFP_KERNEL); + if (!powernow_table) { + printk(KERN_ERR PFX "powernow_table memory alloc failure\n"); + return -ENOMEM; + } + + for (j = 0; j < data->numps; j++) { + powernow_table[j].index = pst[j].fid; /* lower 8 bits */ + powernow_table[j].index |= (pst[j].vid << 8); /* upper 8 bits */ + powernow_table[j].frequency = find_khz_freq_from_fid(pst[j].fid); + } + powernow_table[data->numps].frequency = CPUFREQ_TABLE_END; + powernow_table[data->numps].index = 0; + + if (query_current_values_with_pending_wait(data)) { + kfree(powernow_table); + return -EIO; + } + + dprintk("cfid 0x%x, cvid 0x%x\n", data->currfid, data->currvid); + data->powernow_table = powernow_table; + print_basics(data); + + for (j = 0; j < data->numps; j++) + if ((pst[j].fid==data->currfid) && (pst[j].vid==data->currvid)) + return 0; + + dprintk("currfid/vid do not match PST, ignoring\n"); + return 0; +} + +/* Find and validate the PSB/PST table in BIOS. */ +static int find_psb_table(struct powernow_k8_data *data) +{ + struct psb_s *psb; + unsigned int i; + u32 mvs; + u8 maxvid; + u32 cpst = 0; + u32 thiscpuid; + + for (i = 0xc0000; i < 0xffff0; i += 0x10) { + /* Scan BIOS looking for the signature. */ + /* It can not be at ffff0 - it is too big. */ + + psb = phys_to_virt(i); + if (memcmp(psb, PSB_ID_STRING, PSB_ID_STRING_LEN) != 0) + continue; + + dprintk("found PSB header at 0x%p\n", psb); + + dprintk("table vers: 0x%x\n", psb->tableversion); + if (psb->tableversion != PSB_VERSION_1_4) { + printk(KERN_INFO BFX "PSB table is not v1.4\n"); + return -ENODEV; + } + + dprintk("flags: 0x%x\n", psb->flags1); + if (psb->flags1) { + printk(KERN_ERR BFX "unknown flags\n"); + return -ENODEV; + } + + data->vstable = psb->vstable; + dprintk("voltage stabilization time: %d(*20us)\n", data->vstable); + + dprintk("flags2: 0x%x\n", psb->flags2); + data->rvo = psb->flags2 & 3; + data->irt = ((psb->flags2) >> 2) & 3; + mvs = ((psb->flags2) >> 4) & 3; + data->vidmvs = 1 << mvs; + data->batps = ((psb->flags2) >> 6) & 3; + + dprintk("ramp voltage offset: %d\n", data->rvo); + dprintk("isochronous relief time: %d\n", data->irt); + dprintk("maximum voltage step: %d - 0x%x\n", mvs, data->vidmvs); + + dprintk("numpst: 0x%x\n", psb->num_tables); + cpst = psb->num_tables; + if ((psb->cpuid == 0x00000fc0) || (psb->cpuid == 0x00000fe0) ){ + thiscpuid = cpuid_eax(CPUID_PROCESSOR_SIGNATURE); + if ((thiscpuid == 0x00000fc0) || (thiscpuid == 0x00000fe0) ) { + cpst = 1; + } + } + if (cpst != 1) { + printk(KERN_ERR BFX "numpst must be 1\n"); + return -ENODEV; + } + + data->plllock = psb->plllocktime; + dprintk("plllocktime: 0x%x (units 1us)\n", psb->plllocktime); + dprintk("maxfid: 0x%x\n", psb->maxfid); + dprintk("maxvid: 0x%x\n", psb->maxvid); + maxvid = psb->maxvid; + + data->numps = psb->numps; + dprintk("numpstates: 0x%x\n", data->numps); + return fill_powernow_table(data, (struct pst_s *)(psb+1), maxvid); + } + /* + * If you see this message, complain to BIOS manufacturer. If + * he tells you "we do not support Linux" or some similar + * nonsense, remember that Windows 2000 uses the same legacy + * mechanism that the old Linux PSB driver uses. Tell them it + * is broken with Windows 2000. + * + * The reference to the AMD documentation is chapter 9 in the + * BIOS and Kernel Developer's Guide, which is available on + * www.amd.com + */ + printk(KERN_ERR PFX "BIOS error - no PSB\n"); + return -ENODEV; +} + +#ifdef CONFIG_X86_POWERNOW_K8_ACPI +static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) +{ + if (!data->acpi_data.state_count) + return; + + data->irt = (data->acpi_data.states[index].control >> IRT_SHIFT) & IRT_MASK; + data->rvo = (data->acpi_data.states[index].control >> RVO_SHIFT) & RVO_MASK; + data->plllock = (data->acpi_data.states[index].control >> PLL_L_SHIFT) & PLL_L_MASK; + data->vidmvs = 1 << ((data->acpi_data.states[index].control >> MVS_SHIFT) & MVS_MASK); + data->vstable = (data->acpi_data.states[index].control >> VST_SHIFT) & VST_MASK; +} + +static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) +{ + int i; + int cntlofreq = 0; + struct cpufreq_frequency_table *powernow_table; + + if (acpi_processor_register_performance(&data->acpi_data, data->cpu)) { + dprintk("register performance failed\n"); + return -EIO; + } + + /* verify the data contained in the ACPI structures */ + if (data->acpi_data.state_count <= 1) { + dprintk("No ACPI P-States\n"); + goto err_out; + } + + if ((data->acpi_data.control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) || + (data->acpi_data.status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) { + dprintk("Invalid control/status registers (%x - %x)\n", + data->acpi_data.control_register.space_id, + data->acpi_data.status_register.space_id); + goto err_out; + } + + /* fill in data->powernow_table */ + powernow_table = kmalloc((sizeof(struct cpufreq_frequency_table) + * (data->acpi_data.state_count + 1)), GFP_KERNEL); + if (!powernow_table) { + dprintk("powernow_table memory alloc failure\n"); + goto err_out; + } + + for (i = 0; i < data->acpi_data.state_count; i++) { + u32 fid = data->acpi_data.states[i].control & FID_MASK; + u32 vid = (data->acpi_data.states[i].control >> VID_SHIFT) & VID_MASK; + + dprintk(" %d : fid 0x%x, vid 0x%x\n", i, fid, vid); + + powernow_table[i].index = fid; /* lower 8 bits */ + powernow_table[i].index |= (vid << 8); /* upper 8 bits */ + powernow_table[i].frequency = find_khz_freq_from_fid(fid); + + /* verify frequency is OK */ + if ((powernow_table[i].frequency > (MAX_FREQ * 1000)) || + (powernow_table[i].frequency < (MIN_FREQ * 1000))) { + dprintk("invalid freq %u kHz, ignoring\n", powernow_table[i].frequency); + powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID; + continue; + } + + /* verify voltage is OK - BIOSs are using "off" to indicate invalid */ + if (vid == 0x1f) { + dprintk("invalid vid %u, ignoring\n", vid); + powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID; + continue; + } + + if (fid < HI_FID_TABLE_BOTTOM) { + if (cntlofreq) { + /* if both entries are the same, ignore this + * one... + */ + if ((powernow_table[i].frequency != powernow_table[cntlofreq].frequency) || + (powernow_table[i].index != powernow_table[cntlofreq].index)) { + printk(KERN_ERR PFX "Too many lo freq table entries\n"); + goto err_out_mem; + } + + dprintk("double low frequency table entry, ignoring it.\n"); + powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID; + continue; + } else + cntlofreq = i; + } + + if (powernow_table[i].frequency != (data->acpi_data.states[i].core_frequency * 1000)) { + printk(KERN_INFO PFX "invalid freq entries %u kHz vs. %u kHz\n", + powernow_table[i].frequency, + (unsigned int) (data->acpi_data.states[i].core_frequency * 1000)); + powernow_table[i].frequency = CPUFREQ_ENTRY_INVALID; + continue; + } + } + + powernow_table[data->acpi_data.state_count].frequency = CPUFREQ_TABLE_END; + powernow_table[data->acpi_data.state_count].index = 0; + data->powernow_table = powernow_table; + + /* fill in data */ + data->numps = data->acpi_data.state_count; + print_basics(data); + powernow_k8_acpi_pst_values(data, 0); + + /* notify BIOS that we exist */ + acpi_processor_notify_smm(THIS_MODULE); + + return 0; + +err_out_mem: + kfree(powernow_table); + +err_out: + acpi_processor_unregister_performance(&data->acpi_data, data->cpu); + + /* data->acpi_data.state_count informs us at ->exit() whether ACPI was used */ + data->acpi_data.state_count = 0; + + return -ENODEV; +} + +static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) +{ + if (data->acpi_data.state_count) + acpi_processor_unregister_performance(&data->acpi_data, data->cpu); +} + +#else +static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data) { return -ENODEV; } +static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data) { return; } +static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index) { return; } +#endif /* CONFIG_X86_POWERNOW_K8_ACPI */ + +/* Take a frequency, and issue the fid/vid transition command */ +static int transition_frequency(struct powernow_k8_data *data, unsigned int index) +{ + u32 fid; + u32 vid; + int res; + struct cpufreq_freqs freqs; + + dprintk("cpu %d transition to index %u\n", smp_processor_id(), index); + + /* fid are the lower 8 bits of the index we stored into + * the cpufreq frequency table in find_psb_table, vid are + * the upper 8 bits. + */ + + fid = data->powernow_table[index].index & 0xFF; + vid = (data->powernow_table[index].index & 0xFF00) >> 8; + + dprintk("table matched fid 0x%x, giving vid 0x%x\n", fid, vid); + + if (query_current_values_with_pending_wait(data)) + return 1; + + if ((data->currvid == vid) && (data->currfid == fid)) { + dprintk("target matches current values (fid 0x%x, vid 0x%x)\n", + fid, vid); + return 0; + } + + if ((fid < HI_FID_TABLE_BOTTOM) && (data->currfid < HI_FID_TABLE_BOTTOM)) { + printk("ignoring illegal change in lo freq table-%x to 0x%x\n", + data->currfid, fid); + return 1; + } + + dprintk("cpu %d, changing to fid 0x%x, vid 0x%x\n", + smp_processor_id(), fid, vid); + + freqs.cpu = data->cpu; + + freqs.old = find_khz_freq_from_fid(data->currfid); + freqs.new = find_khz_freq_from_fid(fid); + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + down(&fidvid_sem); + res = transition_fid_vid(data, fid, vid); + up(&fidvid_sem); + + freqs.new = find_khz_freq_from_fid(data->currfid); + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + return res; +} + +/* Driver entry point to switch to the target frequency */ +static int powernowk8_target(struct cpufreq_policy *pol, unsigned targfreq, unsigned relation) +{ + cpumask_t oldmask = CPU_MASK_ALL; + struct powernow_k8_data *data = powernow_data[pol->cpu]; + u32 checkfid = data->currfid; + u32 checkvid = data->currvid; + unsigned int newstate; + int ret = -EIO; + + /* only run on specific CPU from here on */ + oldmask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(pol->cpu)); + schedule(); + + if (smp_processor_id() != pol->cpu) { + printk(KERN_ERR "limiting to cpu %u failed\n", pol->cpu); + goto err_out; + } + + if (pending_bit_stuck()) { + printk(KERN_ERR PFX "failing targ, change pending bit set\n"); + goto err_out; + } + + dprintk("targ: cpu %d, %d kHz, min %d, max %d, relation %d\n", + pol->cpu, targfreq, pol->min, pol->max, relation); + + if (query_current_values_with_pending_wait(data)) { + ret = -EIO; + goto err_out; + } + + dprintk("targ: curr fid 0x%x, vid 0x%x\n", + data->currfid, data->currvid); + + if ((checkvid != data->currvid) || (checkfid != data->currfid)) { + printk(KERN_ERR PFX + "error - out of sync, fid 0x%x 0x%x, vid 0x%x 0x%x\n", + checkfid, data->currfid, checkvid, data->currvid); + } + + if (cpufreq_frequency_table_target(pol, data->powernow_table, targfreq, relation, &newstate)) + goto err_out; + + powernow_k8_acpi_pst_values(data, newstate); + + if (transition_frequency(data, newstate)) { + printk(KERN_ERR PFX "transition frequency failed\n"); + ret = 1; + goto err_out; + } + + pol->cur = find_khz_freq_from_fid(data->currfid); + ret = 0; + +err_out: + set_cpus_allowed(current, oldmask); + schedule(); + + return ret; +} + +/* Driver entry point to verify the policy and range of frequencies */ +static int powernowk8_verify(struct cpufreq_policy *pol) +{ + struct powernow_k8_data *data = powernow_data[pol->cpu]; + + return cpufreq_frequency_table_verify(pol, data->powernow_table); +} + +/* per CPU init entry point to the driver */ +static int __init powernowk8_cpu_init(struct cpufreq_policy *pol) +{ + struct powernow_k8_data *data; + cpumask_t oldmask = CPU_MASK_ALL; + int rc; + + if (!check_supported_cpu(pol->cpu)) + return -ENODEV; + + data = kmalloc(sizeof(struct powernow_k8_data), GFP_KERNEL); + if (!data) { + printk(KERN_ERR PFX "unable to alloc powernow_k8_data"); + return -ENOMEM; + } + memset(data,0,sizeof(struct powernow_k8_data)); + + data->cpu = pol->cpu; + + if (powernow_k8_cpu_init_acpi(data)) { + /* + * Use the PSB BIOS structure. This is only availabe on + * an UP version, and is deprecated by AMD. + */ + + if ((num_online_cpus() != 1) || (num_possible_cpus() != 1)) { + printk(KERN_INFO PFX "MP systems not supported by PSB BIOS structure\n"); + kfree(data); + return -ENODEV; + } + if (pol->cpu != 0) { + printk(KERN_ERR PFX "init not cpu 0\n"); + kfree(data); + return -ENODEV; + } + rc = find_psb_table(data); + if (rc) { + kfree(data); + return -ENODEV; + } + } + + /* only run on specific CPU from here on */ + oldmask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(pol->cpu)); + schedule(); + + if (smp_processor_id() != pol->cpu) { + printk(KERN_ERR "limiting to cpu %u failed\n", pol->cpu); + goto err_out; + } + + if (pending_bit_stuck()) { + printk(KERN_ERR PFX "failing init, change pending bit set\n"); + goto err_out; + } + + if (query_current_values_with_pending_wait(data)) + goto err_out; + + fidvid_msr_init(); + + /* run on any CPU again */ + set_cpus_allowed(current, oldmask); + schedule(); + + pol->governor = CPUFREQ_DEFAULT_GOVERNOR; + + /* Take a crude guess here. + * That guess was in microseconds, so multiply with 1000 */ + pol->cpuinfo.transition_latency = (((data->rvo + 8) * data->vstable * VST_UNITS_20US) + + (3 * (1 << data->irt) * 10)) * 1000; + + pol->cur = find_khz_freq_from_fid(data->currfid); + dprintk("policy current frequency %d kHz\n", pol->cur); + + /* min/max the cpu is capable of */ + if (cpufreq_frequency_table_cpuinfo(pol, data->powernow_table)) { + printk(KERN_ERR PFX "invalid powernow_table\n"); + powernow_k8_cpu_exit_acpi(data); + kfree(data->powernow_table); + kfree(data); + return -EINVAL; + } + + cpufreq_frequency_table_get_attr(data->powernow_table, pol->cpu); + + printk("cpu_init done, current fid 0x%x, vid 0x%x\n", + data->currfid, data->currvid); + + powernow_data[pol->cpu] = data; + + return 0; + +err_out: + set_cpus_allowed(current, oldmask); + schedule(); + powernow_k8_cpu_exit_acpi(data); + + kfree(data); + return -ENODEV; +} + +static int __devexit powernowk8_cpu_exit (struct cpufreq_policy *pol) +{ + struct powernow_k8_data *data = powernow_data[pol->cpu]; + + if (!data) + return -EINVAL; + + powernow_k8_cpu_exit_acpi(data); + + cpufreq_frequency_table_put_attr(pol->cpu); + + kfree(data->powernow_table); + kfree(data); + + return 0; +} + +static unsigned int powernowk8_get (unsigned int cpu) +{ + struct powernow_k8_data *data = powernow_data[cpu]; + cpumask_t oldmask = current->cpus_allowed; + unsigned int khz = 0; + + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + if (smp_processor_id() != cpu) { + printk(KERN_ERR PFX "limiting to CPU %d failed in powernowk8_get\n", cpu); + set_cpus_allowed(current, oldmask); + return 0; + } + preempt_disable(); + + if (query_current_values_with_pending_wait(data)) + goto out; + + khz = find_khz_freq_from_fid(data->currfid); + + out: + preempt_enable_no_resched(); + set_cpus_allowed(current, oldmask); + + return khz; +} + +static struct freq_attr* powernow_k8_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver cpufreq_amd64_driver = { + .verify = powernowk8_verify, + .target = powernowk8_target, + .init = powernowk8_cpu_init, + .exit = __devexit_p(powernowk8_cpu_exit), + .get = powernowk8_get, + .name = "powernow-k8", + .owner = THIS_MODULE, + .attr = powernow_k8_attr, +}; + +/* driver entry point for init */ +static int __init powernowk8_init(void) +{ + unsigned int i, supported_cpus = 0; + + for (i=0; i<NR_CPUS; i++) { + if (!cpu_online(i)) + continue; + if (check_supported_cpu(i)) + supported_cpus++; + } + + if (supported_cpus == num_online_cpus()) { + printk(KERN_INFO PFX "Found %d AMD Athlon 64 / Opteron processors (" VERSION ")\n", + supported_cpus); + return cpufreq_register_driver(&cpufreq_amd64_driver); + } + + return -ENODEV; +} + +/* driver entry point for term */ +static void __exit powernowk8_exit(void) +{ + dprintk("exit\n"); + + cpufreq_unregister_driver(&cpufreq_amd64_driver); +} + +MODULE_AUTHOR("Paul Devriendt <paul.devriendt@amd.com>"); +MODULE_DESCRIPTION("AMD Athlon 64 and Opteron processor frequency driver."); +MODULE_LICENSE("GPL"); + +late_initcall(powernowk8_init); +module_exit(powernowk8_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/powernow-k8.h b/arch/i386/kernel/cpu/cpufreq/powernow-k8.h new file mode 100644 index 00000000000..63ebc8470f5 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/powernow-k8.h @@ -0,0 +1,176 @@ +/* + * (c) 2003, 2004 Advanced Micro Devices, Inc. + * Your use of this code is subject to the terms and conditions of the + * GNU general public license version 2. See "COPYING" or + * http://www.gnu.org/licenses/gpl.html + */ + +struct powernow_k8_data { + unsigned int cpu; + + u32 numps; /* number of p-states */ + u32 batps; /* number of p-states supported on battery */ + + /* these values are constant when the PSB is used to determine + * vid/fid pairings, but are modified during the ->target() call + * when ACPI is used */ + u32 rvo; /* ramp voltage offset */ + u32 irt; /* isochronous relief time */ + u32 vidmvs; /* usable value calculated from mvs */ + u32 vstable; /* voltage stabilization time, units 20 us */ + u32 plllock; /* pll lock time, units 1 us */ + + /* keep track of the current fid / vid */ + u32 currvid, currfid; + + /* the powernow_table includes all frequency and vid/fid pairings: + * fid are the lower 8 bits of the index, vid are the upper 8 bits. + * frequency is in kHz */ + struct cpufreq_frequency_table *powernow_table; + +#ifdef CONFIG_X86_POWERNOW_K8_ACPI + /* the acpi table needs to be kept. it's only available if ACPI was + * used to determine valid frequency/vid/fid states */ + struct acpi_processor_performance acpi_data; +#endif +}; + + +/* processor's cpuid instruction support */ +#define CPUID_PROCESSOR_SIGNATURE 1 /* function 1 */ +#define CPUID_XFAM 0x0ff00000 /* extended family */ +#define CPUID_XFAM_K8 0 +#define CPUID_XMOD 0x000f0000 /* extended model */ +#define CPUID_XMOD_REV_E 0x00020000 +#define CPUID_USE_XFAM_XMOD 0x00000f00 +#define CPUID_GET_MAX_CAPABILITIES 0x80000000 +#define CPUID_FREQ_VOLT_CAPABILITIES 0x80000007 +#define P_STATE_TRANSITION_CAPABLE 6 + +/* Model Specific Registers for p-state transitions. MSRs are 64-bit. For */ +/* writes (wrmsr - opcode 0f 30), the register number is placed in ecx, and */ +/* the value to write is placed in edx:eax. For reads (rdmsr - opcode 0f 32), */ +/* the register number is placed in ecx, and the data is returned in edx:eax. */ + +#define MSR_FIDVID_CTL 0xc0010041 +#define MSR_FIDVID_STATUS 0xc0010042 + +/* Field definitions within the FID VID Low Control MSR : */ +#define MSR_C_LO_INIT_FID_VID 0x00010000 +#define MSR_C_LO_NEW_VID 0x00001f00 +#define MSR_C_LO_NEW_FID 0x0000002f +#define MSR_C_LO_VID_SHIFT 8 + +/* Field definitions within the FID VID High Control MSR : */ +#define MSR_C_HI_STP_GNT_TO 0x000fffff + +/* Field definitions within the FID VID Low Status MSR : */ +#define MSR_S_LO_CHANGE_PENDING 0x80000000 /* cleared when completed */ +#define MSR_S_LO_MAX_RAMP_VID 0x1f000000 +#define MSR_S_LO_MAX_FID 0x003f0000 +#define MSR_S_LO_START_FID 0x00003f00 +#define MSR_S_LO_CURRENT_FID 0x0000003f + +/* Field definitions within the FID VID High Status MSR : */ +#define MSR_S_HI_MAX_WORKING_VID 0x001f0000 +#define MSR_S_HI_START_VID 0x00001f00 +#define MSR_S_HI_CURRENT_VID 0x0000001f +#define MSR_C_HI_STP_GNT_BENIGN 0x00000001 + +/* + * There are restrictions frequencies have to follow: + * - only 1 entry in the low fid table ( <=1.4GHz ) + * - lowest entry in the high fid table must be >= 2 * the entry in the + * low fid table + * - lowest entry in the high fid table must be a <= 200MHz + 2 * the entry + * in the low fid table + * - the parts can only step at 200 MHz intervals, so 1.9 GHz is never valid + * - lowest frequency must be >= interprocessor hypertransport link speed + * (only applies to MP systems obviously) + */ + +/* fids (frequency identifiers) are arranged in 2 tables - lo and hi */ +#define LO_FID_TABLE_TOP 6 /* fid values marking the boundary */ +#define HI_FID_TABLE_BOTTOM 8 /* between the low and high tables */ + +#define LO_VCOFREQ_TABLE_TOP 1400 /* corresponding vco frequency values */ +#define HI_VCOFREQ_TABLE_BOTTOM 1600 + +#define MIN_FREQ_RESOLUTION 200 /* fids jump by 2 matching freq jumps by 200 */ + +#define MAX_FID 0x2a /* Spec only gives FID values as far as 5 GHz */ +#define LEAST_VID 0x1e /* Lowest (numerically highest) useful vid value */ + +#define MIN_FREQ 800 /* Min and max freqs, per spec */ +#define MAX_FREQ 5000 + +#define INVALID_FID_MASK 0xffffffc1 /* not a valid fid if these bits are set */ +#define INVALID_VID_MASK 0xffffffe0 /* not a valid vid if these bits are set */ + +#define STOP_GRANT_5NS 1 /* min poss memory access latency for voltage change */ + +#define PLL_LOCK_CONVERSION (1000/5) /* ms to ns, then divide by clock period */ + +#define MAXIMUM_VID_STEPS 1 /* Current cpus only allow a single step of 25mV */ +#define VST_UNITS_20US 20 /* Voltage Stabalization Time is in units of 20us */ + +/* + * Most values of interest are enocoded in a single field of the _PSS + * entries: the "control" value. + */ + +#define IRT_SHIFT 30 +#define RVO_SHIFT 28 +#define PLL_L_SHIFT 20 +#define MVS_SHIFT 18 +#define VST_SHIFT 11 +#define VID_SHIFT 6 +#define IRT_MASK 3 +#define RVO_MASK 3 +#define PLL_L_MASK 0x7f +#define MVS_MASK 3 +#define VST_MASK 0x7f +#define VID_MASK 0x1f +#define FID_MASK 0x3f + + +/* + * Version 1.4 of the PSB table. This table is constructed by BIOS and is + * to tell the OS's power management driver which VIDs and FIDs are + * supported by this particular processor. + * If the data in the PSB / PST is wrong, then this driver will program the + * wrong values into hardware, which is very likely to lead to a crash. + */ + +#define PSB_ID_STRING "AMDK7PNOW!" +#define PSB_ID_STRING_LEN 10 + +#define PSB_VERSION_1_4 0x14 + +struct psb_s { + u8 signature[10]; + u8 tableversion; + u8 flags1; + u16 vstable; + u8 flags2; + u8 num_tables; + u32 cpuid; + u8 plllocktime; + u8 maxfid; + u8 maxvid; + u8 numps; +}; + +/* Pairs of fid/vid values are appended to the version 1.4 PSB table. */ +struct pst_s { + u8 fid; + u8 vid; +}; + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "powernow-k8", msg) + +static int core_voltage_pre_transition(struct powernow_k8_data *data, u32 reqvid); +static int core_voltage_post_transition(struct powernow_k8_data *data, u32 reqvid); +static int core_frequency_transition(struct powernow_k8_data *data, u32 reqfid); + +static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data, unsigned int index); diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-centrino.c b/arch/i386/kernel/cpu/cpufreq/speedstep-centrino.c new file mode 100644 index 00000000000..07d5612dc00 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-centrino.c @@ -0,0 +1,715 @@ +/* + * cpufreq driver for Enhanced SpeedStep, as found in Intel's Pentium + * M (part of the Centrino chipset). + * + * Despite the "SpeedStep" in the name, this is almost entirely unlike + * traditional SpeedStep. + * + * Modelled on speedstep.c + * + * Copyright (C) 2003 Jeremy Fitzhardinge <jeremy@goop.org> + * + * WARNING WARNING WARNING + * + * This driver manipulates the PERF_CTL MSR, which is only somewhat + * documented. While it seems to work on my laptop, it has not been + * tested anywhere else, and it may not work for you, do strange + * things or simply crash. + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/config.h> +#include <linux/delay.h> +#include <linux/compiler.h> + +#ifdef CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI +#include <linux/acpi.h> +#include <acpi/processor.h> +#endif + +#include <asm/msr.h> +#include <asm/processor.h> +#include <asm/cpufeature.h> + +#include "speedstep-est-common.h" + +#define PFX "speedstep-centrino: " +#define MAINTAINER "Jeremy Fitzhardinge <jeremy@goop.org>" + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "speedstep-centrino", msg) + + +struct cpu_id +{ + __u8 x86; /* CPU family */ + __u8 x86_model; /* model */ + __u8 x86_mask; /* stepping */ +}; + +enum { + CPU_BANIAS, + CPU_DOTHAN_A1, + CPU_DOTHAN_A2, + CPU_DOTHAN_B0, +}; + +static const struct cpu_id cpu_ids[] = { + [CPU_BANIAS] = { 6, 9, 5 }, + [CPU_DOTHAN_A1] = { 6, 13, 1 }, + [CPU_DOTHAN_A2] = { 6, 13, 2 }, + [CPU_DOTHAN_B0] = { 6, 13, 6 }, +}; +#define N_IDS (sizeof(cpu_ids)/sizeof(cpu_ids[0])) + +struct cpu_model +{ + const struct cpu_id *cpu_id; + const char *model_name; + unsigned max_freq; /* max clock in kHz */ + + struct cpufreq_frequency_table *op_points; /* clock/voltage pairs */ +}; +static int centrino_verify_cpu_id(const struct cpuinfo_x86 *c, const struct cpu_id *x); + +/* Operating points for current CPU */ +static struct cpu_model *centrino_model[NR_CPUS]; +static const struct cpu_id *centrino_cpu[NR_CPUS]; + +static struct cpufreq_driver centrino_driver; + +#ifdef CONFIG_X86_SPEEDSTEP_CENTRINO_TABLE + +/* Computes the correct form for IA32_PERF_CTL MSR for a particular + frequency/voltage operating point; frequency in MHz, volts in mV. + This is stored as "index" in the structure. */ +#define OP(mhz, mv) \ + { \ + .frequency = (mhz) * 1000, \ + .index = (((mhz)/100) << 8) | ((mv - 700) / 16) \ + } + +/* + * These voltage tables were derived from the Intel Pentium M + * datasheet, document 25261202.pdf, Table 5. I have verified they + * are consistent with my IBM ThinkPad X31, which has a 1.3GHz Pentium + * M. + */ + +/* Ultra Low Voltage Intel Pentium M processor 900MHz (Banias) */ +static struct cpufreq_frequency_table banias_900[] = +{ + OP(600, 844), + OP(800, 988), + OP(900, 1004), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Ultra Low Voltage Intel Pentium M processor 1000MHz (Banias) */ +static struct cpufreq_frequency_table banias_1000[] = +{ + OP(600, 844), + OP(800, 972), + OP(900, 988), + OP(1000, 1004), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Low Voltage Intel Pentium M processor 1.10GHz (Banias) */ +static struct cpufreq_frequency_table banias_1100[] = +{ + OP( 600, 956), + OP( 800, 1020), + OP( 900, 1100), + OP(1000, 1164), + OP(1100, 1180), + { .frequency = CPUFREQ_TABLE_END } +}; + + +/* Low Voltage Intel Pentium M processor 1.20GHz (Banias) */ +static struct cpufreq_frequency_table banias_1200[] = +{ + OP( 600, 956), + OP( 800, 1004), + OP( 900, 1020), + OP(1000, 1100), + OP(1100, 1164), + OP(1200, 1180), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Intel Pentium M processor 1.30GHz (Banias) */ +static struct cpufreq_frequency_table banias_1300[] = +{ + OP( 600, 956), + OP( 800, 1260), + OP(1000, 1292), + OP(1200, 1356), + OP(1300, 1388), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Intel Pentium M processor 1.40GHz (Banias) */ +static struct cpufreq_frequency_table banias_1400[] = +{ + OP( 600, 956), + OP( 800, 1180), + OP(1000, 1308), + OP(1200, 1436), + OP(1400, 1484), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Intel Pentium M processor 1.50GHz (Banias) */ +static struct cpufreq_frequency_table banias_1500[] = +{ + OP( 600, 956), + OP( 800, 1116), + OP(1000, 1228), + OP(1200, 1356), + OP(1400, 1452), + OP(1500, 1484), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Intel Pentium M processor 1.60GHz (Banias) */ +static struct cpufreq_frequency_table banias_1600[] = +{ + OP( 600, 956), + OP( 800, 1036), + OP(1000, 1164), + OP(1200, 1276), + OP(1400, 1420), + OP(1600, 1484), + { .frequency = CPUFREQ_TABLE_END } +}; + +/* Intel Pentium M processor 1.70GHz (Banias) */ +static struct cpufreq_frequency_table banias_1700[] = +{ + OP( 600, 956), + OP( 800, 1004), + OP(1000, 1116), + OP(1200, 1228), + OP(1400, 1308), + OP(1700, 1484), + { .frequency = CPUFREQ_TABLE_END } +}; +#undef OP + +#define _BANIAS(cpuid, max, name) \ +{ .cpu_id = cpuid, \ + .model_name = "Intel(R) Pentium(R) M processor " name "MHz", \ + .max_freq = (max)*1000, \ + .op_points = banias_##max, \ +} +#define BANIAS(max) _BANIAS(&cpu_ids[CPU_BANIAS], max, #max) + +/* CPU models, their operating frequency range, and freq/voltage + operating points */ +static struct cpu_model models[] = +{ + _BANIAS(&cpu_ids[CPU_BANIAS], 900, " 900"), + BANIAS(1000), + BANIAS(1100), + BANIAS(1200), + BANIAS(1300), + BANIAS(1400), + BANIAS(1500), + BANIAS(1600), + BANIAS(1700), + + /* NULL model_name is a wildcard */ + { &cpu_ids[CPU_DOTHAN_A1], NULL, 0, NULL }, + { &cpu_ids[CPU_DOTHAN_A2], NULL, 0, NULL }, + { &cpu_ids[CPU_DOTHAN_B0], NULL, 0, NULL }, + + { NULL, } +}; +#undef _BANIAS +#undef BANIAS + +static int centrino_cpu_init_table(struct cpufreq_policy *policy) +{ + struct cpuinfo_x86 *cpu = &cpu_data[policy->cpu]; + struct cpu_model *model; + + for(model = models; model->cpu_id != NULL; model++) + if (centrino_verify_cpu_id(cpu, model->cpu_id) && + (model->model_name == NULL || + strcmp(cpu->x86_model_id, model->model_name) == 0)) + break; + + if (model->cpu_id == NULL) { + /* No match at all */ + dprintk(KERN_INFO PFX "no support for CPU model \"%s\": " + "send /proc/cpuinfo to " MAINTAINER "\n", + cpu->x86_model_id); + return -ENOENT; + } + + if (model->op_points == NULL) { + /* Matched a non-match */ + dprintk(KERN_INFO PFX "no table support for CPU model \"%s\": \n", + cpu->x86_model_id); +#ifndef CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI + dprintk(KERN_INFO PFX "try compiling with CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI enabled\n"); +#endif + return -ENOENT; + } + + centrino_model[policy->cpu] = model; + + dprintk("found \"%s\": max frequency: %dkHz\n", + model->model_name, model->max_freq); + + return 0; +} + +#else +static inline int centrino_cpu_init_table(struct cpufreq_policy *policy) { return -ENODEV; } +#endif /* CONFIG_X86_SPEEDSTEP_CENTRINO_TABLE */ + +static int centrino_verify_cpu_id(const struct cpuinfo_x86 *c, const struct cpu_id *x) +{ + if ((c->x86 == x->x86) && + (c->x86_model == x->x86_model) && + (c->x86_mask == x->x86_mask)) + return 1; + return 0; +} + +/* To be called only after centrino_model is initialized */ +static unsigned extract_clock(unsigned msr, unsigned int cpu, int failsafe) +{ + int i; + + /* + * Extract clock in kHz from PERF_CTL value + * for centrino, as some DSDTs are buggy. + * Ideally, this can be done using the acpi_data structure. + */ + if ((centrino_cpu[cpu] == &cpu_ids[CPU_BANIAS]) || + (centrino_cpu[cpu] == &cpu_ids[CPU_DOTHAN_A1]) || + (centrino_cpu[cpu] == &cpu_ids[CPU_DOTHAN_B0])) { + msr = (msr >> 8) & 0xff; + return msr * 100000; + } + + if ((!centrino_model[cpu]) || (!centrino_model[cpu]->op_points)) + return 0; + + msr &= 0xffff; + for (i=0;centrino_model[cpu]->op_points[i].frequency != CPUFREQ_TABLE_END; i++) { + if (msr == centrino_model[cpu]->op_points[i].index) + return centrino_model[cpu]->op_points[i].frequency; + } + if (failsafe) + return centrino_model[cpu]->op_points[i-1].frequency; + else + return 0; +} + +/* Return the current CPU frequency in kHz */ +static unsigned int get_cur_freq(unsigned int cpu) +{ + unsigned l, h; + unsigned clock_freq; + cpumask_t saved_mask; + + saved_mask = current->cpus_allowed; + set_cpus_allowed(current, cpumask_of_cpu(cpu)); + if (smp_processor_id() != cpu) + return 0; + + rdmsr(MSR_IA32_PERF_STATUS, l, h); + clock_freq = extract_clock(l, cpu, 0); + + if (unlikely(clock_freq == 0)) { + /* + * On some CPUs, we can see transient MSR values (which are + * not present in _PSS), while CPU is doing some automatic + * P-state transition (like TM2). Get the last freq set + * in PERF_CTL. + */ + rdmsr(MSR_IA32_PERF_CTL, l, h); + clock_freq = extract_clock(l, cpu, 1); + } + + set_cpus_allowed(current, saved_mask); + return clock_freq; +} + + +#ifdef CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI + +static struct acpi_processor_performance p; + +/* + * centrino_cpu_init_acpi - register with ACPI P-States library + * + * Register with the ACPI P-States library (part of drivers/acpi/processor.c) + * in order to determine correct frequency and voltage pairings by reading + * the _PSS of the ACPI DSDT or SSDT tables. + */ +static int centrino_cpu_init_acpi(struct cpufreq_policy *policy) +{ + union acpi_object arg0 = {ACPI_TYPE_BUFFER}; + u32 arg0_buf[3]; + struct acpi_object_list arg_list = {1, &arg0}; + unsigned long cur_freq; + int result = 0, i; + unsigned int cpu = policy->cpu; + + /* _PDC settings */ + arg0.buffer.length = 12; + arg0.buffer.pointer = (u8 *) arg0_buf; + arg0_buf[0] = ACPI_PDC_REVISION_ID; + arg0_buf[1] = 1; + arg0_buf[2] = ACPI_PDC_EST_CAPABILITY_SMP | ACPI_PDC_EST_CAPABILITY_MSR; + + p.pdc = &arg_list; + + /* register with ACPI core */ + if (acpi_processor_register_performance(&p, cpu)) { + dprintk(KERN_INFO PFX "obtaining ACPI data failed\n"); + return -EIO; + } + + /* verify the acpi_data */ + if (p.state_count <= 1) { + dprintk("No P-States\n"); + result = -ENODEV; + goto err_unreg; + } + + if ((p.control_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE) || + (p.status_register.space_id != ACPI_ADR_SPACE_FIXED_HARDWARE)) { + dprintk("Invalid control/status registers (%x - %x)\n", + p.control_register.space_id, p.status_register.space_id); + result = -EIO; + goto err_unreg; + } + + for (i=0; i<p.state_count; i++) { + if (p.states[i].control != p.states[i].status) { + dprintk("Different control (%x) and status values (%x)\n", + p.states[i].control, p.states[i].status); + result = -EINVAL; + goto err_unreg; + } + + if (!p.states[i].core_frequency) { + dprintk("Zero core frequency for state %u\n", i); + result = -EINVAL; + goto err_unreg; + } + + if (p.states[i].core_frequency > p.states[0].core_frequency) { + dprintk("P%u has larger frequency (%u) than P0 (%u), skipping\n", i, + p.states[i].core_frequency, p.states[0].core_frequency); + p.states[i].core_frequency = 0; + continue; + } + } + + centrino_model[cpu] = kmalloc(sizeof(struct cpu_model), GFP_KERNEL); + if (!centrino_model[cpu]) { + result = -ENOMEM; + goto err_unreg; + } + memset(centrino_model[cpu], 0, sizeof(struct cpu_model)); + + centrino_model[cpu]->model_name=NULL; + centrino_model[cpu]->max_freq = p.states[0].core_frequency * 1000; + centrino_model[cpu]->op_points = kmalloc(sizeof(struct cpufreq_frequency_table) * + (p.state_count + 1), GFP_KERNEL); + if (!centrino_model[cpu]->op_points) { + result = -ENOMEM; + goto err_kfree; + } + + for (i=0; i<p.state_count; i++) { + centrino_model[cpu]->op_points[i].index = p.states[i].control; + centrino_model[cpu]->op_points[i].frequency = p.states[i].core_frequency * 1000; + dprintk("adding state %i with frequency %u and control value %04x\n", + i, centrino_model[cpu]->op_points[i].frequency, centrino_model[cpu]->op_points[i].index); + } + centrino_model[cpu]->op_points[p.state_count].frequency = CPUFREQ_TABLE_END; + + cur_freq = get_cur_freq(cpu); + + for (i=0; i<p.state_count; i++) { + if (!p.states[i].core_frequency) { + dprintk("skipping state %u\n", i); + centrino_model[cpu]->op_points[i].frequency = CPUFREQ_ENTRY_INVALID; + continue; + } + + if (extract_clock(centrino_model[cpu]->op_points[i].index, cpu, 0) != + (centrino_model[cpu]->op_points[i].frequency)) { + dprintk("Invalid encoded frequency (%u vs. %u)\n", + extract_clock(centrino_model[cpu]->op_points[i].index, cpu, 0), + centrino_model[cpu]->op_points[i].frequency); + result = -EINVAL; + goto err_kfree_all; + } + + if (cur_freq == centrino_model[cpu]->op_points[i].frequency) + p.state = i; + } + + /* notify BIOS that we exist */ + acpi_processor_notify_smm(THIS_MODULE); + + return 0; + + err_kfree_all: + kfree(centrino_model[cpu]->op_points); + err_kfree: + kfree(centrino_model[cpu]); + err_unreg: + acpi_processor_unregister_performance(&p, cpu); + dprintk(KERN_INFO PFX "invalid ACPI data\n"); + return (result); +} +#else +static inline int centrino_cpu_init_acpi(struct cpufreq_policy *policy) { return -ENODEV; } +#endif + +static int centrino_cpu_init(struct cpufreq_policy *policy) +{ + struct cpuinfo_x86 *cpu = &cpu_data[policy->cpu]; + unsigned freq; + unsigned l, h; + int ret; + int i; + + /* Only Intel makes Enhanced Speedstep-capable CPUs */ + if (cpu->x86_vendor != X86_VENDOR_INTEL || !cpu_has(cpu, X86_FEATURE_EST)) + return -ENODEV; + + for (i = 0; i < N_IDS; i++) + if (centrino_verify_cpu_id(cpu, &cpu_ids[i])) + break; + + if (i != N_IDS) + centrino_cpu[policy->cpu] = &cpu_ids[i]; + + if (is_const_loops_cpu(policy->cpu)) { + centrino_driver.flags |= CPUFREQ_CONST_LOOPS; + } + + if (centrino_cpu_init_acpi(policy)) { + if (policy->cpu != 0) + return -ENODEV; + + if (!centrino_cpu[policy->cpu]) { + dprintk(KERN_INFO PFX "found unsupported CPU with " + "Enhanced SpeedStep: send /proc/cpuinfo to " + MAINTAINER "\n"); + return -ENODEV; + } + + if (centrino_cpu_init_table(policy)) { + return -ENODEV; + } + } + + /* Check to see if Enhanced SpeedStep is enabled, and try to + enable it if not. */ + rdmsr(MSR_IA32_MISC_ENABLE, l, h); + + if (!(l & (1<<16))) { + l |= (1<<16); + dprintk("trying to enable Enhanced SpeedStep (%x)\n", l); + wrmsr(MSR_IA32_MISC_ENABLE, l, h); + + /* check to see if it stuck */ + rdmsr(MSR_IA32_MISC_ENABLE, l, h); + if (!(l & (1<<16))) { + printk(KERN_INFO PFX "couldn't enable Enhanced SpeedStep\n"); + return -ENODEV; + } + } + + freq = get_cur_freq(policy->cpu); + + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = 10000; /* 10uS transition latency */ + policy->cur = freq; + + dprintk("centrino_cpu_init: cur=%dkHz\n", policy->cur); + + ret = cpufreq_frequency_table_cpuinfo(policy, centrino_model[policy->cpu]->op_points); + if (ret) + return (ret); + + cpufreq_frequency_table_get_attr(centrino_model[policy->cpu]->op_points, policy->cpu); + + return 0; +} + +static int centrino_cpu_exit(struct cpufreq_policy *policy) +{ + unsigned int cpu = policy->cpu; + + if (!centrino_model[cpu]) + return -ENODEV; + + cpufreq_frequency_table_put_attr(cpu); + +#ifdef CONFIG_X86_SPEEDSTEP_CENTRINO_ACPI + if (!centrino_model[cpu]->model_name) { + dprintk("unregistering and freeing ACPI data\n"); + acpi_processor_unregister_performance(&p, cpu); + kfree(centrino_model[cpu]->op_points); + kfree(centrino_model[cpu]); + } +#endif + + centrino_model[cpu] = NULL; + + return 0; +} + +/** + * centrino_verify - verifies a new CPUFreq policy + * @policy: new policy + * + * Limit must be within this model's frequency range at least one + * border included. + */ +static int centrino_verify (struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, centrino_model[policy->cpu]->op_points); +} + +/** + * centrino_setpolicy - set a new CPUFreq policy + * @policy: new policy + * @target_freq: the target frequency + * @relation: how that frequency relates to achieved frequency (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H) + * + * Sets a new CPUFreq policy. + */ +static int centrino_target (struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = 0; + unsigned int msr, oldmsr, h, cpu = policy->cpu; + struct cpufreq_freqs freqs; + cpumask_t saved_mask; + int retval; + + if (centrino_model[cpu] == NULL) + return -ENODEV; + + /* + * Support for SMP systems. + * Make sure we are running on the CPU that wants to change frequency + */ + saved_mask = current->cpus_allowed; + set_cpus_allowed(current, policy->cpus); + if (!cpu_isset(smp_processor_id(), policy->cpus)) { + dprintk("couldn't limit to CPUs in this domain\n"); + return(-EAGAIN); + } + + if (cpufreq_frequency_table_target(policy, centrino_model[cpu]->op_points, target_freq, + relation, &newstate)) { + retval = -EINVAL; + goto migrate_end; + } + + msr = centrino_model[cpu]->op_points[newstate].index; + rdmsr(MSR_IA32_PERF_CTL, oldmsr, h); + + if (msr == (oldmsr & 0xffff)) { + retval = 0; + dprintk("no change needed - msr was and needs to be %x\n", oldmsr); + goto migrate_end; + } + + freqs.cpu = cpu; + freqs.old = extract_clock(oldmsr, cpu, 0); + freqs.new = extract_clock(msr, cpu, 0); + + dprintk("target=%dkHz old=%d new=%d msr=%04x\n", + target_freq, freqs.old, freqs.new, msr); + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + + /* all but 16 LSB are "reserved", so treat them with + care */ + oldmsr &= ~0xffff; + msr &= 0xffff; + oldmsr |= msr; + + wrmsr(MSR_IA32_PERF_CTL, oldmsr, h); + + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + retval = 0; +migrate_end: + set_cpus_allowed(current, saved_mask); + return (retval); +} + +static struct freq_attr* centrino_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver centrino_driver = { + .name = "centrino", /* should be speedstep-centrino, + but there's a 16 char limit */ + .init = centrino_cpu_init, + .exit = centrino_cpu_exit, + .verify = centrino_verify, + .target = centrino_target, + .get = get_cur_freq, + .attr = centrino_attr, + .owner = THIS_MODULE, +}; + + +/** + * centrino_init - initializes the Enhanced SpeedStep CPUFreq driver + * + * Initializes the Enhanced SpeedStep support. Returns -ENODEV on + * unsupported devices, -ENOENT if there's no voltage table for this + * particular CPU model, -EINVAL on problems during initiatization, + * and zero on success. + * + * This is quite picky. Not only does the CPU have to advertise the + * "est" flag in the cpuid capability flags, we look for a specific + * CPU model and stepping, and we need to have the exact model name in + * our voltage tables. That is, be paranoid about not releasing + * someone's valuable magic smoke. + */ +static int __init centrino_init(void) +{ + struct cpuinfo_x86 *cpu = cpu_data; + + if (!cpu_has(cpu, X86_FEATURE_EST)) + return -ENODEV; + + return cpufreq_register_driver(¢rino_driver); +} + +static void __exit centrino_exit(void) +{ + cpufreq_unregister_driver(¢rino_driver); +} + +MODULE_AUTHOR ("Jeremy Fitzhardinge <jeremy@goop.org>"); +MODULE_DESCRIPTION ("Enhanced SpeedStep driver for Intel Pentium M processors."); +MODULE_LICENSE ("GPL"); + +late_initcall(centrino_init); +module_exit(centrino_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-est-common.h b/arch/i386/kernel/cpu/cpufreq/speedstep-est-common.h new file mode 100644 index 00000000000..5ce995c9d86 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-est-common.h @@ -0,0 +1,25 @@ +/* + * Routines common for drivers handling Enhanced Speedstep Technology + * Copyright (C) 2004 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com> + * + * Licensed under the terms of the GNU GPL License version 2 -- see + * COPYING for details. + */ + +static inline int is_const_loops_cpu(unsigned int cpu) +{ + struct cpuinfo_x86 *c = cpu_data + cpu; + + if (c->x86_vendor != X86_VENDOR_INTEL || !cpu_has(c, X86_FEATURE_EST)) + return 0; + + /* + * on P-4s, the TSC runs with constant frequency independent of cpu freq + * when we use EST + */ + if (c->x86 == 0xf) + return 1; + + return 0; +} + diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-ich.c b/arch/i386/kernel/cpu/cpufreq/speedstep-ich.c new file mode 100644 index 00000000000..5b7d18a06af --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-ich.c @@ -0,0 +1,424 @@ +/* + * (C) 2001 Dave Jones, Arjan van de ven. + * (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de> + * + * Licensed under the terms of the GNU GPL License version 2. + * Based upon reverse engineered information, and on Intel documentation + * for chipsets ICH2-M and ICH3-M. + * + * Many thanks to Ducrot Bruno for finding and fixing the last + * "missing link" for ICH2-M/ICH3-M support, and to Thomas Winkler + * for extensive testing. + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + + +/********************************************************************* + * SPEEDSTEP - DEFINITIONS * + *********************************************************************/ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/pci.h> +#include <linux/slab.h> + +#include "speedstep-lib.h" + + +/* speedstep_chipset: + * It is necessary to know which chipset is used. As accesses to + * this device occur at various places in this module, we need a + * static struct pci_dev * pointing to that device. + */ +static struct pci_dev *speedstep_chipset_dev; + + +/* speedstep_processor + */ +static unsigned int speedstep_processor = 0; + + +/* + * There are only two frequency states for each processor. Values + * are in kHz for the time being. + */ +static struct cpufreq_frequency_table speedstep_freqs[] = { + {SPEEDSTEP_HIGH, 0}, + {SPEEDSTEP_LOW, 0}, + {0, CPUFREQ_TABLE_END}, +}; + + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "speedstep-ich", msg) + + +/** + * speedstep_set_state - set the SpeedStep state + * @state: new processor frequency state (SPEEDSTEP_LOW or SPEEDSTEP_HIGH) + * + * Tries to change the SpeedStep state. + */ +static void speedstep_set_state (unsigned int state) +{ + u32 pmbase; + u8 pm2_blk; + u8 value; + unsigned long flags; + + if (!speedstep_chipset_dev || (state > 0x1)) + return; + + /* get PMBASE */ + pci_read_config_dword(speedstep_chipset_dev, 0x40, &pmbase); + if (!(pmbase & 0x01)) { + printk(KERN_ERR "speedstep-ich: could not find speedstep register\n"); + return; + } + + pmbase &= 0xFFFFFFFE; + if (!pmbase) { + printk(KERN_ERR "speedstep-ich: could not find speedstep register\n"); + return; + } + + /* Disable IRQs */ + local_irq_save(flags); + + /* read state */ + value = inb(pmbase + 0x50); + + dprintk("read at pmbase 0x%x + 0x50 returned 0x%x\n", pmbase, value); + + /* write new state */ + value &= 0xFE; + value |= state; + + dprintk("writing 0x%x to pmbase 0x%x + 0x50\n", value, pmbase); + + /* Disable bus master arbitration */ + pm2_blk = inb(pmbase + 0x20); + pm2_blk |= 0x01; + outb(pm2_blk, (pmbase + 0x20)); + + /* Actual transition */ + outb(value, (pmbase + 0x50)); + + /* Restore bus master arbitration */ + pm2_blk &= 0xfe; + outb(pm2_blk, (pmbase + 0x20)); + + /* check if transition was successful */ + value = inb(pmbase + 0x50); + + /* Enable IRQs */ + local_irq_restore(flags); + + dprintk("read at pmbase 0x%x + 0x50 returned 0x%x\n", pmbase, value); + + if (state == (value & 0x1)) { + dprintk("change to %u MHz succeeded\n", (speedstep_get_processor_frequency(speedstep_processor) / 1000)); + } else { + printk (KERN_ERR "cpufreq: change failed - I/O error\n"); + } + + return; +} + + +/** + * speedstep_activate - activate SpeedStep control in the chipset + * + * Tries to activate the SpeedStep status and control registers. + * Returns -EINVAL on an unsupported chipset, and zero on success. + */ +static int speedstep_activate (void) +{ + u16 value = 0; + + if (!speedstep_chipset_dev) + return -EINVAL; + + pci_read_config_word(speedstep_chipset_dev, 0x00A0, &value); + if (!(value & 0x08)) { + value |= 0x08; + dprintk("activating SpeedStep (TM) registers\n"); + pci_write_config_word(speedstep_chipset_dev, 0x00A0, value); + } + + return 0; +} + + +/** + * speedstep_detect_chipset - detect the Southbridge which contains SpeedStep logic + * + * Detects ICH2-M, ICH3-M and ICH4-M so far. The pci_dev points to + * the LPC bridge / PM module which contains all power-management + * functions. Returns the SPEEDSTEP_CHIPSET_-number for the detected + * chipset, or zero on failure. + */ +static unsigned int speedstep_detect_chipset (void) +{ + speedstep_chipset_dev = pci_get_subsys(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82801DB_12, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + if (speedstep_chipset_dev) + return 4; /* 4-M */ + + speedstep_chipset_dev = pci_get_subsys(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82801CA_12, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + if (speedstep_chipset_dev) + return 3; /* 3-M */ + + + speedstep_chipset_dev = pci_get_subsys(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82801BA_10, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + if (speedstep_chipset_dev) { + /* speedstep.c causes lockups on Dell Inspirons 8000 and + * 8100 which use a pretty old revision of the 82815 + * host brige. Abort on these systems. + */ + static struct pci_dev *hostbridge; + u8 rev = 0; + + hostbridge = pci_get_subsys(PCI_VENDOR_ID_INTEL, + PCI_DEVICE_ID_INTEL_82815_MC, + PCI_ANY_ID, + PCI_ANY_ID, + NULL); + + if (!hostbridge) + return 2; /* 2-M */ + + pci_read_config_byte(hostbridge, PCI_REVISION_ID, &rev); + if (rev < 5) { + dprintk("hostbridge does not support speedstep\n"); + speedstep_chipset_dev = NULL; + pci_dev_put(hostbridge); + return 0; + } + + pci_dev_put(hostbridge); + return 2; /* 2-M */ + } + + return 0; +} + +static unsigned int _speedstep_get(cpumask_t cpus) +{ + unsigned int speed; + cpumask_t cpus_allowed; + + cpus_allowed = current->cpus_allowed; + set_cpus_allowed(current, cpus); + speed = speedstep_get_processor_frequency(speedstep_processor); + set_cpus_allowed(current, cpus_allowed); + dprintk("detected %u kHz as current frequency\n", speed); + return speed; +} + +static unsigned int speedstep_get(unsigned int cpu) +{ + return _speedstep_get(cpumask_of_cpu(cpu)); +} + +/** + * speedstep_target - set a new CPUFreq policy + * @policy: new policy + * @target_freq: the target frequency + * @relation: how that frequency relates to achieved frequency (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H) + * + * Sets a new CPUFreq policy. + */ +static int speedstep_target (struct cpufreq_policy *policy, + unsigned int target_freq, + unsigned int relation) +{ + unsigned int newstate = 0; + struct cpufreq_freqs freqs; + cpumask_t cpus_allowed; + int i; + + if (cpufreq_frequency_table_target(policy, &speedstep_freqs[0], target_freq, relation, &newstate)) + return -EINVAL; + + freqs.old = _speedstep_get(policy->cpus); + freqs.new = speedstep_freqs[newstate].frequency; + freqs.cpu = policy->cpu; + + dprintk("transiting from %u to %u kHz\n", freqs.old, freqs.new); + + /* no transition necessary */ + if (freqs.old == freqs.new) + return 0; + + cpus_allowed = current->cpus_allowed; + + for_each_cpu_mask(i, policy->cpus) { + freqs.cpu = i; + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + } + + /* switch to physical CPU where state is to be changed */ + set_cpus_allowed(current, policy->cpus); + + speedstep_set_state(newstate); + + /* allow to be run on all CPUs */ + set_cpus_allowed(current, cpus_allowed); + + for_each_cpu_mask(i, policy->cpus) { + freqs.cpu = i; + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + } + + return 0; +} + + +/** + * speedstep_verify - verifies a new CPUFreq policy + * @policy: new policy + * + * Limit must be within speedstep_low_freq and speedstep_high_freq, with + * at least one border included. + */ +static int speedstep_verify (struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, &speedstep_freqs[0]); +} + + +static int speedstep_cpu_init(struct cpufreq_policy *policy) +{ + int result = 0; + unsigned int speed; + cpumask_t cpus_allowed; + + /* only run on CPU to be set, or on its sibling */ +#ifdef CONFIG_SMP + policy->cpus = cpu_sibling_map[policy->cpu]; +#endif + + cpus_allowed = current->cpus_allowed; + set_cpus_allowed(current, policy->cpus); + + /* detect low and high frequency */ + result = speedstep_get_freqs(speedstep_processor, + &speedstep_freqs[SPEEDSTEP_LOW].frequency, + &speedstep_freqs[SPEEDSTEP_HIGH].frequency, + &speedstep_set_state); + set_cpus_allowed(current, cpus_allowed); + if (result) + return result; + + /* get current speed setting */ + speed = _speedstep_get(policy->cpus); + if (!speed) + return -EIO; + + dprintk("currently at %s speed setting - %i MHz\n", + (speed == speedstep_freqs[SPEEDSTEP_LOW].frequency) ? "low" : "high", + (speed / 1000)); + + /* cpuinfo and default policy values */ + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = speed; + + result = cpufreq_frequency_table_cpuinfo(policy, speedstep_freqs); + if (result) + return (result); + + cpufreq_frequency_table_get_attr(speedstep_freqs, policy->cpu); + + return 0; +} + + +static int speedstep_cpu_exit(struct cpufreq_policy *policy) +{ + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + +static struct freq_attr* speedstep_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + + +static struct cpufreq_driver speedstep_driver = { + .name = "speedstep-ich", + .verify = speedstep_verify, + .target = speedstep_target, + .init = speedstep_cpu_init, + .exit = speedstep_cpu_exit, + .get = speedstep_get, + .owner = THIS_MODULE, + .attr = speedstep_attr, +}; + + +/** + * speedstep_init - initializes the SpeedStep CPUFreq driver + * + * Initializes the SpeedStep support. Returns -ENODEV on unsupported + * devices, -EINVAL on problems during initiatization, and zero on + * success. + */ +static int __init speedstep_init(void) +{ + /* detect processor */ + speedstep_processor = speedstep_detect_processor(); + if (!speedstep_processor) { + dprintk("Intel(R) SpeedStep(TM) capable processor not found\n"); + return -ENODEV; + } + + /* detect chipset */ + if (!speedstep_detect_chipset()) { + dprintk("Intel(R) SpeedStep(TM) for this chipset not (yet) available.\n"); + return -ENODEV; + } + + /* activate speedstep support */ + if (speedstep_activate()) { + pci_dev_put(speedstep_chipset_dev); + return -EINVAL; + } + + return cpufreq_register_driver(&speedstep_driver); +} + + +/** + * speedstep_exit - unregisters SpeedStep support + * + * Unregisters SpeedStep support. + */ +static void __exit speedstep_exit(void) +{ + pci_dev_put(speedstep_chipset_dev); + cpufreq_unregister_driver(&speedstep_driver); +} + + +MODULE_AUTHOR ("Dave Jones <davej@codemonkey.org.uk>, Dominik Brodowski <linux@brodo.de>"); +MODULE_DESCRIPTION ("Speedstep driver for Intel mobile processors on chipsets with ICH-M southbridges."); +MODULE_LICENSE ("GPL"); + +module_init(speedstep_init); +module_exit(speedstep_exit); diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-lib.c b/arch/i386/kernel/cpu/cpufreq/speedstep-lib.c new file mode 100644 index 00000000000..8ba430a9c3a --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-lib.c @@ -0,0 +1,385 @@ +/* + * (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de> + * + * Licensed under the terms of the GNU GPL License version 2. + * + * Library for common functions for Intel SpeedStep v.1 and v.2 support + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/pci.h> +#include <linux/slab.h> + +#include <asm/msr.h> +#include "speedstep-lib.h" + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "speedstep-lib", msg) + +#ifdef CONFIG_X86_SPEEDSTEP_RELAXED_CAP_CHECK +static int relaxed_check = 0; +#else +#define relaxed_check 0 +#endif + +/********************************************************************* + * GET PROCESSOR CORE SPEED IN KHZ * + *********************************************************************/ + +static unsigned int pentium3_get_frequency (unsigned int processor) +{ + /* See table 14 of p3_ds.pdf and table 22 of 29834003.pdf */ + struct { + unsigned int ratio; /* Frequency Multiplier (x10) */ + u8 bitmap; /* power on configuration bits + [27, 25:22] (in MSR 0x2a) */ + } msr_decode_mult [] = { + { 30, 0x01 }, + { 35, 0x05 }, + { 40, 0x02 }, + { 45, 0x06 }, + { 50, 0x00 }, + { 55, 0x04 }, + { 60, 0x0b }, + { 65, 0x0f }, + { 70, 0x09 }, + { 75, 0x0d }, + { 80, 0x0a }, + { 85, 0x26 }, + { 90, 0x20 }, + { 100, 0x2b }, + { 0, 0xff } /* error or unknown value */ + }; + + /* PIII(-M) FSB settings: see table b1-b of 24547206.pdf */ + struct { + unsigned int value; /* Front Side Bus speed in MHz */ + u8 bitmap; /* power on configuration bits [18: 19] + (in MSR 0x2a) */ + } msr_decode_fsb [] = { + { 66, 0x0 }, + { 100, 0x2 }, + { 133, 0x1 }, + { 0, 0xff} + }; + + u32 msr_lo, msr_tmp; + int i = 0, j = 0; + + /* read MSR 0x2a - we only need the low 32 bits */ + rdmsr(MSR_IA32_EBL_CR_POWERON, msr_lo, msr_tmp); + dprintk("P3 - MSR_IA32_EBL_CR_POWERON: 0x%x 0x%x\n", msr_lo, msr_tmp); + msr_tmp = msr_lo; + + /* decode the FSB */ + msr_tmp &= 0x00c0000; + msr_tmp >>= 18; + while (msr_tmp != msr_decode_fsb[i].bitmap) { + if (msr_decode_fsb[i].bitmap == 0xff) + return 0; + i++; + } + + /* decode the multiplier */ + if (processor == SPEEDSTEP_PROCESSOR_PIII_C_EARLY) { + dprintk("workaround for early PIIIs\n"); + msr_lo &= 0x03c00000; + } else + msr_lo &= 0x0bc00000; + msr_lo >>= 22; + while (msr_lo != msr_decode_mult[j].bitmap) { + if (msr_decode_mult[j].bitmap == 0xff) + return 0; + j++; + } + + dprintk("speed is %u\n", (msr_decode_mult[j].ratio * msr_decode_fsb[i].value * 100)); + + return (msr_decode_mult[j].ratio * msr_decode_fsb[i].value * 100); +} + + +static unsigned int pentiumM_get_frequency(void) +{ + u32 msr_lo, msr_tmp; + + rdmsr(MSR_IA32_EBL_CR_POWERON, msr_lo, msr_tmp); + dprintk("PM - MSR_IA32_EBL_CR_POWERON: 0x%x 0x%x\n", msr_lo, msr_tmp); + + /* see table B-2 of 24547212.pdf */ + if (msr_lo & 0x00040000) { + printk(KERN_DEBUG "speedstep-lib: PM - invalid FSB: 0x%x 0x%x\n", msr_lo, msr_tmp); + return 0; + } + + msr_tmp = (msr_lo >> 22) & 0x1f; + dprintk("bits 22-26 are 0x%x, speed is %u\n", msr_tmp, (msr_tmp * 100 * 1000)); + + return (msr_tmp * 100 * 1000); +} + + +static unsigned int pentium4_get_frequency(void) +{ + struct cpuinfo_x86 *c = &boot_cpu_data; + u32 msr_lo, msr_hi, mult; + unsigned int fsb = 0; + + rdmsr(0x2c, msr_lo, msr_hi); + + dprintk("P4 - MSR_EBC_FREQUENCY_ID: 0x%x 0x%x\n", msr_lo, msr_hi); + + /* decode the FSB: see IA-32 Intel (C) Architecture Software + * Developer's Manual, Volume 3: System Prgramming Guide, + * revision #12 in Table B-1: MSRs in the Pentium 4 and + * Intel Xeon Processors, on page B-4 and B-5. + */ + if (c->x86_model < 2) + fsb = 100 * 1000; + else { + u8 fsb_code = (msr_lo >> 16) & 0x7; + switch (fsb_code) { + case 0: + fsb = 100 * 1000; + break; + case 1: + fsb = 13333 * 10; + break; + case 2: + fsb = 200 * 1000; + break; + } + } + + if (!fsb) + printk(KERN_DEBUG "speedstep-lib: couldn't detect FSB speed. Please send an e-mail to <linux@brodo.de>\n"); + + /* Multiplier. */ + if (c->x86_model < 2) + mult = msr_lo >> 27; + else + mult = msr_lo >> 24; + + dprintk("P4 - FSB %u kHz; Multiplier %u; Speed %u kHz\n", fsb, mult, (fsb * mult)); + + return (fsb * mult); +} + + +unsigned int speedstep_get_processor_frequency(unsigned int processor) +{ + switch (processor) { + case SPEEDSTEP_PROCESSOR_PM: + return pentiumM_get_frequency(); + case SPEEDSTEP_PROCESSOR_P4D: + case SPEEDSTEP_PROCESSOR_P4M: + return pentium4_get_frequency(); + case SPEEDSTEP_PROCESSOR_PIII_T: + case SPEEDSTEP_PROCESSOR_PIII_C: + case SPEEDSTEP_PROCESSOR_PIII_C_EARLY: + return pentium3_get_frequency(processor); + default: + return 0; + }; + return 0; +} +EXPORT_SYMBOL_GPL(speedstep_get_processor_frequency); + + +/********************************************************************* + * DETECT SPEEDSTEP-CAPABLE PROCESSOR * + *********************************************************************/ + +unsigned int speedstep_detect_processor (void) +{ + struct cpuinfo_x86 *c = cpu_data; + u32 ebx, msr_lo, msr_hi; + + dprintk("x86: %x, model: %x\n", c->x86, c->x86_model); + + if ((c->x86_vendor != X86_VENDOR_INTEL) || + ((c->x86 != 6) && (c->x86 != 0xF))) + return 0; + + if (c->x86 == 0xF) { + /* Intel Mobile Pentium 4-M + * or Intel Mobile Pentium 4 with 533 MHz FSB */ + if (c->x86_model != 2) + return 0; + + ebx = cpuid_ebx(0x00000001); + ebx &= 0x000000FF; + + dprintk("ebx value is %x, x86_mask is %x\n", ebx, c->x86_mask); + + switch (c->x86_mask) { + case 4: + /* + * B-stepping [M-P4-M] + * sample has ebx = 0x0f, production has 0x0e. + */ + if ((ebx == 0x0e) || (ebx == 0x0f)) + return SPEEDSTEP_PROCESSOR_P4M; + break; + case 7: + /* + * C-stepping [M-P4-M] + * needs to have ebx=0x0e, else it's a celeron: + * cf. 25130917.pdf / page 7, footnote 5 even + * though 25072120.pdf / page 7 doesn't say + * samples are only of B-stepping... + */ + if (ebx == 0x0e) + return SPEEDSTEP_PROCESSOR_P4M; + break; + case 9: + /* + * D-stepping [M-P4-M or M-P4/533] + * + * this is totally strange: CPUID 0x0F29 is + * used by M-P4-M, M-P4/533 and(!) Celeron CPUs. + * The latter need to be sorted out as they don't + * support speedstep. + * Celerons with CPUID 0x0F29 may have either + * ebx=0x8 or 0xf -- 25130917.pdf doesn't say anything + * specific. + * M-P4-Ms may have either ebx=0xe or 0xf [see above] + * M-P4/533 have either ebx=0xe or 0xf. [25317607.pdf] + * also, M-P4M HTs have ebx=0x8, too + * For now, they are distinguished by the model_id string + */ + if ((ebx == 0x0e) || (strstr(c->x86_model_id,"Mobile Intel(R) Pentium(R) 4") != NULL)) + return SPEEDSTEP_PROCESSOR_P4M; + break; + default: + break; + } + return 0; + } + + switch (c->x86_model) { + case 0x0B: /* Intel PIII [Tualatin] */ + /* cpuid_ebx(1) is 0x04 for desktop PIII, + 0x06 for mobile PIII-M */ + ebx = cpuid_ebx(0x00000001); + dprintk("ebx is %x\n", ebx); + + ebx &= 0x000000FF; + + if (ebx != 0x06) + return 0; + + /* So far all PIII-M processors support SpeedStep. See + * Intel's 24540640.pdf of June 2003 + */ + + return SPEEDSTEP_PROCESSOR_PIII_T; + + case 0x08: /* Intel PIII [Coppermine] */ + + /* all mobile PIII Coppermines have FSB 100 MHz + * ==> sort out a few desktop PIIIs. */ + rdmsr(MSR_IA32_EBL_CR_POWERON, msr_lo, msr_hi); + dprintk("Coppermine: MSR_IA32_EBL_CR_POWERON is 0x%x, 0x%x\n", msr_lo, msr_hi); + msr_lo &= 0x00c0000; + if (msr_lo != 0x0080000) + return 0; + + /* + * If the processor is a mobile version, + * platform ID has bit 50 set + * it has SpeedStep technology if either + * bit 56 or 57 is set + */ + rdmsr(MSR_IA32_PLATFORM_ID, msr_lo, msr_hi); + dprintk("Coppermine: MSR_IA32_PLATFORM ID is 0x%x, 0x%x\n", msr_lo, msr_hi); + if ((msr_hi & (1<<18)) && (relaxed_check ? 1 : (msr_hi & (3<<24)))) { + if (c->x86_mask == 0x01) { + dprintk("early PIII version\n"); + return SPEEDSTEP_PROCESSOR_PIII_C_EARLY; + } else + return SPEEDSTEP_PROCESSOR_PIII_C; + } + + default: + return 0; + } +} +EXPORT_SYMBOL_GPL(speedstep_detect_processor); + + +/********************************************************************* + * DETECT SPEEDSTEP SPEEDS * + *********************************************************************/ + +unsigned int speedstep_get_freqs(unsigned int processor, + unsigned int *low_speed, + unsigned int *high_speed, + void (*set_state) (unsigned int state)) +{ + unsigned int prev_speed; + unsigned int ret = 0; + unsigned long flags; + + if ((!processor) || (!low_speed) || (!high_speed) || (!set_state)) + return -EINVAL; + + dprintk("trying to determine both speeds\n"); + + /* get current speed */ + prev_speed = speedstep_get_processor_frequency(processor); + if (!prev_speed) + return -EIO; + + dprintk("previous seped is %u\n", prev_speed); + + local_irq_save(flags); + + /* switch to low state */ + set_state(SPEEDSTEP_LOW); + *low_speed = speedstep_get_processor_frequency(processor); + if (!*low_speed) { + ret = -EIO; + goto out; + } + + dprintk("low seped is %u\n", *low_speed); + + /* switch to high state */ + set_state(SPEEDSTEP_HIGH); + *high_speed = speedstep_get_processor_frequency(processor); + if (!*high_speed) { + ret = -EIO; + goto out; + } + + dprintk("high seped is %u\n", *high_speed); + + if (*low_speed == *high_speed) { + ret = -ENODEV; + goto out; + } + + /* switch to previous state, if necessary */ + if (*high_speed != prev_speed) + set_state(SPEEDSTEP_LOW); + + out: + local_irq_restore(flags); + return (ret); +} +EXPORT_SYMBOL_GPL(speedstep_get_freqs); + +#ifdef CONFIG_X86_SPEEDSTEP_RELAXED_CAP_CHECK +module_param(relaxed_check, int, 0444); +MODULE_PARM_DESC(relaxed_check, "Don't do all checks for speedstep capability."); +#endif + +MODULE_AUTHOR ("Dominik Brodowski <linux@brodo.de>"); +MODULE_DESCRIPTION ("Library for Intel SpeedStep 1 or 2 cpufreq drivers."); +MODULE_LICENSE ("GPL"); diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-lib.h b/arch/i386/kernel/cpu/cpufreq/speedstep-lib.h new file mode 100644 index 00000000000..261a2c9b7f6 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-lib.h @@ -0,0 +1,47 @@ +/* + * (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de> + * + * Licensed under the terms of the GNU GPL License version 2. + * + * Library for common functions for Intel SpeedStep v.1 and v.2 support + * + * BIG FAT DISCLAIMER: Work in progress code. Possibly *dangerous* + */ + + + +/* processors */ + +#define SPEEDSTEP_PROCESSOR_PIII_C_EARLY 0x00000001 /* Coppermine core */ +#define SPEEDSTEP_PROCESSOR_PIII_C 0x00000002 /* Coppermine core */ +#define SPEEDSTEP_PROCESSOR_PIII_T 0x00000003 /* Tualatin core */ +#define SPEEDSTEP_PROCESSOR_P4M 0x00000004 /* P4-M */ + +/* the following processors are not speedstep-capable and are not auto-detected + * in speedstep_detect_processor(). However, their speed can be detected using + * the speedstep_get_processor_frequency() call. */ +#define SPEEDSTEP_PROCESSOR_PM 0xFFFFFF03 /* Pentium M */ +#define SPEEDSTEP_PROCESSOR_P4D 0xFFFFFF04 /* desktop P4 */ + +/* speedstep states -- only two of them */ + +#define SPEEDSTEP_HIGH 0x00000000 +#define SPEEDSTEP_LOW 0x00000001 + + +/* detect a speedstep-capable processor */ +extern unsigned int speedstep_detect_processor (void); + +/* detect the current speed (in khz) of the processor */ +extern unsigned int speedstep_get_processor_frequency(unsigned int processor); + + +/* detect the low and high speeds of the processor. The callback + * set_state"'s first argument is either SPEEDSTEP_HIGH or + * SPEEDSTEP_LOW; the second argument is zero so that no + * cpufreq_notify_transition calls are initiated. + */ +extern unsigned int speedstep_get_freqs(unsigned int processor, + unsigned int *low_speed, + unsigned int *high_speed, + void (*set_state) (unsigned int state)); diff --git a/arch/i386/kernel/cpu/cpufreq/speedstep-smi.c b/arch/i386/kernel/cpu/cpufreq/speedstep-smi.c new file mode 100644 index 00000000000..79440b3f087 --- /dev/null +++ b/arch/i386/kernel/cpu/cpufreq/speedstep-smi.c @@ -0,0 +1,424 @@ +/* + * Intel SpeedStep SMI driver. + * + * (C) 2003 Hiroshi Miura <miura@da-cha.org> + * + * Licensed under the terms of the GNU GPL License version 2. + * + */ + + +/********************************************************************* + * SPEEDSTEP - DEFINITIONS * + *********************************************************************/ + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/moduleparam.h> +#include <linux/init.h> +#include <linux/cpufreq.h> +#include <linux/pci.h> +#include <linux/slab.h> +#include <linux/delay.h> +#include <asm/ist.h> + +#include "speedstep-lib.h" + +/* speedstep system management interface port/command. + * + * These parameters are got from IST-SMI BIOS call. + * If user gives it, these are used. + * + */ +static int smi_port = 0; +static int smi_cmd = 0; +static unsigned int smi_sig = 0; + +/* info about the processor */ +static unsigned int speedstep_processor = 0; + +/* + * There are only two frequency states for each processor. Values + * are in kHz for the time being. + */ +static struct cpufreq_frequency_table speedstep_freqs[] = { + {SPEEDSTEP_HIGH, 0}, + {SPEEDSTEP_LOW, 0}, + {0, CPUFREQ_TABLE_END}, +}; + +#define GET_SPEEDSTEP_OWNER 0 +#define GET_SPEEDSTEP_STATE 1 +#define SET_SPEEDSTEP_STATE 2 +#define GET_SPEEDSTEP_FREQS 4 + +/* how often shall the SMI call be tried if it failed, e.g. because + * of DMA activity going on? */ +#define SMI_TRIES 5 + +#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "speedstep-smi", msg) + +/** + * speedstep_smi_ownership + */ +static int speedstep_smi_ownership (void) +{ + u32 command, result, magic; + u32 function = GET_SPEEDSTEP_OWNER; + unsigned char magic_data[] = "Copyright (c) 1999 Intel Corporation"; + + command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff); + magic = virt_to_phys(magic_data); + + dprintk("trying to obtain ownership with command %x at port %x\n", command, smi_port); + + __asm__ __volatile__( + "out %%al, (%%dx)\n" + : "=D" (result) + : "a" (command), "b" (function), "c" (0), "d" (smi_port), "D" (0), "S" (magic) + ); + + dprintk("result is %x\n", result); + + return result; +} + +/** + * speedstep_smi_get_freqs - get SpeedStep preferred & current freq. + * @low: the low frequency value is placed here + * @high: the high frequency value is placed here + * + * Only available on later SpeedStep-enabled systems, returns false results or + * even hangs [cf. bugme.osdl.org # 1422] on earlier systems. Empirical testing + * shows that the latter occurs if !(ist_info.event & 0xFFFF). + */ +static int speedstep_smi_get_freqs (unsigned int *low, unsigned int *high) +{ + u32 command, result = 0, edi, high_mhz, low_mhz; + u32 state=0; + u32 function = GET_SPEEDSTEP_FREQS; + + if (!(ist_info.event & 0xFFFF)) { + dprintk("bug #1422 -- can't read freqs from BIOS\n", result); + return -ENODEV; + } + + command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff); + + dprintk("trying to determine frequencies with command %x at port %x\n", command, smi_port); + + __asm__ __volatile__("movl $0, %%edi\n" + "out %%al, (%%dx)\n" + : "=a" (result), "=b" (high_mhz), "=c" (low_mhz), "=d" (state), "=D" (edi) + : "a" (command), "b" (function), "c" (state), "d" (smi_port), "S" (0) + ); + + dprintk("result %x, low_freq %u, high_freq %u\n", result, low_mhz, high_mhz); + + /* abort if results are obviously incorrect... */ + if ((high_mhz + low_mhz) < 600) + return -EINVAL; + + *high = high_mhz * 1000; + *low = low_mhz * 1000; + + return result; +} + +/** + * speedstep_get_state - set the SpeedStep state + * @state: processor frequency state (SPEEDSTEP_LOW or SPEEDSTEP_HIGH) + * + */ +static int speedstep_get_state (void) +{ + u32 function=GET_SPEEDSTEP_STATE; + u32 result, state, edi, command; + + command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff); + + dprintk("trying to determine current setting with command %x at port %x\n", command, smi_port); + + __asm__ __volatile__("movl $0, %%edi\n" + "out %%al, (%%dx)\n" + : "=a" (result), "=b" (state), "=D" (edi) + : "a" (command), "b" (function), "c" (0), "d" (smi_port), "S" (0) + ); + + dprintk("state is %x, result is %x\n", state, result); + + return (state & 1); +} + + +/** + * speedstep_set_state - set the SpeedStep state + * @state: new processor frequency state (SPEEDSTEP_LOW or SPEEDSTEP_HIGH) + * + */ +static void speedstep_set_state (unsigned int state) +{ + unsigned int result = 0, command, new_state; + unsigned long flags; + unsigned int function=SET_SPEEDSTEP_STATE; + unsigned int retry = 0; + + if (state > 0x1) + return; + + /* Disable IRQs */ + local_irq_save(flags); + + command = (smi_sig & 0xffffff00) | (smi_cmd & 0xff); + + dprintk("trying to set frequency to state %u with command %x at port %x\n", state, command, smi_port); + + do { + if (retry) { + dprintk("retry %u, previous result %u, waiting...\n", retry, result); + mdelay(retry * 50); + } + retry++; + __asm__ __volatile__( + "movl $0, %%edi\n" + "out %%al, (%%dx)\n" + : "=b" (new_state), "=D" (result) + : "a" (command), "b" (function), "c" (state), "d" (smi_port), "S" (0) + ); + } while ((new_state != state) && (retry <= SMI_TRIES)); + + /* enable IRQs */ + local_irq_restore(flags); + + if (new_state == state) { + dprintk("change to %u MHz succeeded after %u tries with result %u\n", (speedstep_freqs[new_state].frequency / 1000), retry, result); + } else { + printk(KERN_ERR "cpufreq: change failed with new_state %u and result %u\n", new_state, result); + } + + return; +} + + +/** + * speedstep_target - set a new CPUFreq policy + * @policy: new policy + * @target_freq: new freq + * @relation: + * + * Sets a new CPUFreq policy/freq. + */ +static int speedstep_target (struct cpufreq_policy *policy, + unsigned int target_freq, unsigned int relation) +{ + unsigned int newstate = 0; + struct cpufreq_freqs freqs; + + if (cpufreq_frequency_table_target(policy, &speedstep_freqs[0], target_freq, relation, &newstate)) + return -EINVAL; + + freqs.old = speedstep_freqs[speedstep_get_state()].frequency; + freqs.new = speedstep_freqs[newstate].frequency; + freqs.cpu = 0; /* speedstep.c is UP only driver */ + + if (freqs.old == freqs.new) + return 0; + + cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE); + speedstep_set_state(newstate); + cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE); + + return 0; +} + + +/** + * speedstep_verify - verifies a new CPUFreq policy + * @policy: new policy + * + * Limit must be within speedstep_low_freq and speedstep_high_freq, with + * at least one border included. + */ +static int speedstep_verify (struct cpufreq_policy *policy) +{ + return cpufreq_frequency_table_verify(policy, &speedstep_freqs[0]); +} + + +static int speedstep_cpu_init(struct cpufreq_policy *policy) +{ + int result; + unsigned int speed,state; + + /* capability check */ + if (policy->cpu != 0) + return -ENODEV; + + result = speedstep_smi_ownership(); + if (result) { + dprintk("fails in aquiring ownership of a SMI interface.\n"); + return -EINVAL; + } + + /* detect low and high frequency */ + result = speedstep_smi_get_freqs(&speedstep_freqs[SPEEDSTEP_LOW].frequency, + &speedstep_freqs[SPEEDSTEP_HIGH].frequency); + if (result) { + /* fall back to speedstep_lib.c dection mechanism: try both states out */ + dprintk("could not detect low and high frequencies by SMI call.\n"); + result = speedstep_get_freqs(speedstep_processor, + &speedstep_freqs[SPEEDSTEP_LOW].frequency, + &speedstep_freqs[SPEEDSTEP_HIGH].frequency, + &speedstep_set_state); + + if (result) { + dprintk("could not detect two different speeds -- aborting.\n"); + return result; + } else + dprintk("workaround worked.\n"); + } + + /* get current speed setting */ + state = speedstep_get_state(); + speed = speedstep_freqs[state].frequency; + + dprintk("currently at %s speed setting - %i MHz\n", + (speed == speedstep_freqs[SPEEDSTEP_LOW].frequency) ? "low" : "high", + (speed / 1000)); + + /* cpuinfo and default policy values */ + policy->governor = CPUFREQ_DEFAULT_GOVERNOR; + policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL; + policy->cur = speed; + + result = cpufreq_frequency_table_cpuinfo(policy, speedstep_freqs); + if (result) + return (result); + + cpufreq_frequency_table_get_attr(speedstep_freqs, policy->cpu); + + return 0; +} + +static int speedstep_cpu_exit(struct cpufreq_policy *policy) +{ + cpufreq_frequency_table_put_attr(policy->cpu); + return 0; +} + +static unsigned int speedstep_get(unsigned int cpu) +{ + if (cpu) + return -ENODEV; + return speedstep_get_processor_frequency(speedstep_processor); +} + + +static int speedstep_resume(struct cpufreq_policy *policy) +{ + int result = speedstep_smi_ownership(); + + if (result) + dprintk("fails in re-aquiring ownership of a SMI interface.\n"); + + return result; +} + +static struct freq_attr* speedstep_attr[] = { + &cpufreq_freq_attr_scaling_available_freqs, + NULL, +}; + +static struct cpufreq_driver speedstep_driver = { + .name = "speedstep-smi", + .verify = speedstep_verify, + .target = speedstep_target, + .init = speedstep_cpu_init, + .exit = speedstep_cpu_exit, + .get = speedstep_get, + .resume = speedstep_resume, + .owner = THIS_MODULE, + .attr = speedstep_attr, +}; + +/** + * speedstep_init - initializes the SpeedStep CPUFreq driver + * + * Initializes the SpeedStep support. Returns -ENODEV on unsupported + * BIOS, -EINVAL on problems during initiatization, and zero on + * success. + */ +static int __init speedstep_init(void) +{ + speedstep_processor = speedstep_detect_processor(); + + switch (speedstep_processor) { + case SPEEDSTEP_PROCESSOR_PIII_T: + case SPEEDSTEP_PROCESSOR_PIII_C: + case SPEEDSTEP_PROCESSOR_PIII_C_EARLY: + break; + default: + speedstep_processor = 0; + } + + if (!speedstep_processor) { + dprintk ("No supported Intel CPU detected.\n"); + return -ENODEV; + } + + dprintk("signature:0x%.8lx, command:0x%.8lx, event:0x%.8lx, perf_level:0x%.8lx.\n", + ist_info.signature, ist_info.command, ist_info.event, ist_info.perf_level); + + + /* Error if no IST-SMI BIOS or no PARM + sig= 'ISGE' aka 'Intel Speedstep Gate E' */ + if ((ist_info.signature != 0x47534943) && ( + (smi_port == 0) || (smi_cmd == 0))) + return -ENODEV; + + if (smi_sig == 1) + smi_sig = 0x47534943; + else + smi_sig = ist_info.signature; + + /* setup smi_port from MODLULE_PARM or BIOS */ + if ((smi_port > 0xff) || (smi_port < 0)) { + return -EINVAL; + } else if (smi_port == 0) { + smi_port = ist_info.command & 0xff; + } + + if ((smi_cmd > 0xff) || (smi_cmd < 0)) { + return -EINVAL; + } else if (smi_cmd == 0) { + smi_cmd = (ist_info.command >> 16) & 0xff; + } + + return cpufreq_register_driver(&speedstep_driver); +} + + +/** + * speedstep_exit - unregisters SpeedStep support + * + * Unregisters SpeedStep support. + */ +static void __exit speedstep_exit(void) +{ + cpufreq_unregister_driver(&speedstep_driver); +} + +module_param(smi_port, int, 0444); +module_param(smi_cmd, int, 0444); +module_param(smi_sig, uint, 0444); + +MODULE_PARM_DESC(smi_port, "Override the BIOS-given IST port with this value -- Intel's default setting is 0xb2"); +MODULE_PARM_DESC(smi_cmd, "Override the BIOS-given IST command with this value -- Intel's default setting is 0x82"); +MODULE_PARM_DESC(smi_sig, "Set to 1 to fake the IST signature when using the SMI interface."); + +MODULE_AUTHOR ("Hiroshi Miura"); +MODULE_DESCRIPTION ("Speedstep driver for IST applet SMI interface."); +MODULE_LICENSE ("GPL"); + +module_init(speedstep_init); +module_exit(speedstep_exit); diff --git a/arch/i386/kernel/cpu/cyrix.c b/arch/i386/kernel/cpu/cyrix.c new file mode 100644 index 00000000000..ba4b01138c8 --- /dev/null +++ b/arch/i386/kernel/cpu/cyrix.c @@ -0,0 +1,439 @@ +#include <linux/init.h> +#include <linux/bitops.h> +#include <linux/delay.h> +#include <linux/pci.h> +#include <asm/dma.h> +#include <asm/io.h> +#include <asm/processor.h> +#include <asm/timer.h> + +#include "cpu.h" + +/* + * Read NSC/Cyrix DEVID registers (DIR) to get more detailed info. about the CPU + */ +static void __init do_cyrix_devid(unsigned char *dir0, unsigned char *dir1) +{ + unsigned char ccr2, ccr3; + unsigned long flags; + + /* we test for DEVID by checking whether CCR3 is writable */ + local_irq_save(flags); + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, ccr3 ^ 0x80); + getCx86(0xc0); /* dummy to change bus */ + + if (getCx86(CX86_CCR3) == ccr3) { /* no DEVID regs. */ + ccr2 = getCx86(CX86_CCR2); + setCx86(CX86_CCR2, ccr2 ^ 0x04); + getCx86(0xc0); /* dummy */ + + if (getCx86(CX86_CCR2) == ccr2) /* old Cx486SLC/DLC */ + *dir0 = 0xfd; + else { /* Cx486S A step */ + setCx86(CX86_CCR2, ccr2); + *dir0 = 0xfe; + } + } + else { + setCx86(CX86_CCR3, ccr3); /* restore CCR3 */ + + /* read DIR0 and DIR1 CPU registers */ + *dir0 = getCx86(CX86_DIR0); + *dir1 = getCx86(CX86_DIR1); + } + local_irq_restore(flags); +} + +/* + * Cx86_dir0_msb is a HACK needed by check_cx686_cpuid/slop in bugs.h in + * order to identify the Cyrix CPU model after we're out of setup.c + * + * Actually since bugs.h doesn't even reference this perhaps someone should + * fix the documentation ??? + */ +static unsigned char Cx86_dir0_msb __initdata = 0; + +static char Cx86_model[][9] __initdata = { + "Cx486", "Cx486", "5x86 ", "6x86", "MediaGX ", "6x86MX ", + "M II ", "Unknown" +}; +static char Cx486_name[][5] __initdata = { + "SLC", "DLC", "SLC2", "DLC2", "SRx", "DRx", + "SRx2", "DRx2" +}; +static char Cx486S_name[][4] __initdata = { + "S", "S2", "Se", "S2e" +}; +static char Cx486D_name[][4] __initdata = { + "DX", "DX2", "?", "?", "?", "DX4" +}; +static char Cx86_cb[] __initdata = "?.5x Core/Bus Clock"; +static char cyrix_model_mult1[] __initdata = "12??43"; +static char cyrix_model_mult2[] __initdata = "12233445"; + +/* + * Reset the slow-loop (SLOP) bit on the 686(L) which is set by some old + * BIOSes for compatibility with DOS games. This makes the udelay loop + * work correctly, and improves performance. + * + * FIXME: our newer udelay uses the tsc. We don't need to frob with SLOP + */ + +extern void calibrate_delay(void) __init; + +static void __init check_cx686_slop(struct cpuinfo_x86 *c) +{ + unsigned long flags; + + if (Cx86_dir0_msb == 3) { + unsigned char ccr3, ccr5; + + local_irq_save(flags); + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + ccr5 = getCx86(CX86_CCR5); + if (ccr5 & 2) + setCx86(CX86_CCR5, ccr5 & 0xfd); /* reset SLOP */ + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ + local_irq_restore(flags); + + if (ccr5 & 2) { /* possible wrong calibration done */ + printk(KERN_INFO "Recalibrating delay loop with SLOP bit reset\n"); + calibrate_delay(); + c->loops_per_jiffy = loops_per_jiffy; + } + } +} + + +static void __init set_cx86_reorder(void) +{ + u8 ccr3; + + printk(KERN_INFO "Enable Memory access reorder on Cyrix/NSC processor.\n"); + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + + /* Load/Store Serialize to mem access disable (=reorder it) */ + setCx86(CX86_PCR0, getCx86(CX86_PCR0) & ~0x80); + /* set load/store serialize from 1GB to 4GB */ + ccr3 |= 0xe0; + setCx86(CX86_CCR3, ccr3); +} + +static void __init set_cx86_memwb(void) +{ + u32 cr0; + + printk(KERN_INFO "Enable Memory-Write-back mode on Cyrix/NSC processor.\n"); + + /* CCR2 bit 2: unlock NW bit */ + setCx86(CX86_CCR2, getCx86(CX86_CCR2) & ~0x04); + /* set 'Not Write-through' */ + cr0 = 0x20000000; + __asm__("movl %%cr0,%%eax\n\t" + "orl %0,%%eax\n\t" + "movl %%eax,%%cr0\n" + : : "r" (cr0) + :"ax"); + /* CCR2 bit 2: lock NW bit and set WT1 */ + setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x14 ); +} + +static void __init set_cx86_inc(void) +{ + unsigned char ccr3; + + printk(KERN_INFO "Enable Incrementor on Cyrix/NSC processor.\n"); + + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + /* PCR1 -- Performance Control */ + /* Incrementor on, whatever that is */ + setCx86(CX86_PCR1, getCx86(CX86_PCR1) | 0x02); + /* PCR0 -- Performance Control */ + /* Incrementor Margin 10 */ + setCx86(CX86_PCR0, getCx86(CX86_PCR0) | 0x04); + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ +} + +/* + * Configure later MediaGX and/or Geode processor. + */ + +static void __init geode_configure(void) +{ + unsigned long flags; + u8 ccr3, ccr4; + local_irq_save(flags); + + /* Suspend on halt power saving and enable #SUSP pin */ + setCx86(CX86_CCR2, getCx86(CX86_CCR2) | 0x88); + + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* Enable */ + + ccr4 = getCx86(CX86_CCR4); + ccr4 |= 0x38; /* FPU fast, DTE cache, Mem bypass */ + + setCx86(CX86_CCR3, ccr3); + + set_cx86_memwb(); + set_cx86_reorder(); + set_cx86_inc(); + + local_irq_restore(flags); +} + + +#ifdef CONFIG_PCI +static struct pci_device_id cyrix_55x0[] = { + { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5510) }, + { PCI_DEVICE(PCI_VENDOR_ID_CYRIX, PCI_DEVICE_ID_CYRIX_5520) }, + { }, +}; +#endif + +static void __init init_cyrix(struct cpuinfo_x86 *c) +{ + unsigned char dir0, dir0_msn, dir0_lsn, dir1 = 0; + char *buf = c->x86_model_id; + const char *p = NULL; + + /* Bit 31 in normal CPUID used for nonstandard 3DNow ID; + 3DNow is IDd by bit 31 in extended CPUID (1*32+31) anyway */ + clear_bit(0*32+31, c->x86_capability); + + /* Cyrix used bit 24 in extended (AMD) CPUID for Cyrix MMX extensions */ + if ( test_bit(1*32+24, c->x86_capability) ) { + clear_bit(1*32+24, c->x86_capability); + set_bit(X86_FEATURE_CXMMX, c->x86_capability); + } + + do_cyrix_devid(&dir0, &dir1); + + check_cx686_slop(c); + + Cx86_dir0_msb = dir0_msn = dir0 >> 4; /* identifies CPU "family" */ + dir0_lsn = dir0 & 0xf; /* model or clock multiplier */ + + /* common case step number/rev -- exceptions handled below */ + c->x86_model = (dir1 >> 4) + 1; + c->x86_mask = dir1 & 0xf; + + /* Now cook; the original recipe is by Channing Corn, from Cyrix. + * We do the same thing for each generation: we work out + * the model, multiplier and stepping. Black magic included, + * to make the silicon step/rev numbers match the printed ones. + */ + + switch (dir0_msn) { + unsigned char tmp; + + case 0: /* Cx486SLC/DLC/SRx/DRx */ + p = Cx486_name[dir0_lsn & 7]; + break; + + case 1: /* Cx486S/DX/DX2/DX4 */ + p = (dir0_lsn & 8) ? Cx486D_name[dir0_lsn & 5] + : Cx486S_name[dir0_lsn & 3]; + break; + + case 2: /* 5x86 */ + Cx86_cb[2] = cyrix_model_mult1[dir0_lsn & 5]; + p = Cx86_cb+2; + break; + + case 3: /* 6x86/6x86L */ + Cx86_cb[1] = ' '; + Cx86_cb[2] = cyrix_model_mult1[dir0_lsn & 5]; + if (dir1 > 0x21) { /* 686L */ + Cx86_cb[0] = 'L'; + p = Cx86_cb; + (c->x86_model)++; + } else /* 686 */ + p = Cx86_cb+1; + /* Emulate MTRRs using Cyrix's ARRs. */ + set_bit(X86_FEATURE_CYRIX_ARR, c->x86_capability); + /* 6x86's contain this bug */ + c->coma_bug = 1; + break; + + case 4: /* MediaGX/GXm or Geode GXM/GXLV/GX1 */ +#ifdef CONFIG_PCI + /* It isn't really a PCI quirk directly, but the cure is the + same. The MediaGX has deep magic SMM stuff that handles the + SB emulation. It thows away the fifo on disable_dma() which + is wrong and ruins the audio. + + Bug2: VSA1 has a wrap bug so that using maximum sized DMA + causes bad things. According to NatSemi VSA2 has another + bug to do with 'hlt'. I've not seen any boards using VSA2 + and X doesn't seem to support it either so who cares 8). + VSA1 we work around however. + */ + + printk(KERN_INFO "Working around Cyrix MediaGX virtual DMA bugs.\n"); + isa_dma_bridge_buggy = 2; +#endif + c->x86_cache_size=16; /* Yep 16K integrated cache thats it */ + + /* + * The 5510/5520 companion chips have a funky PIT. + */ + if (pci_dev_present(cyrix_55x0)) + pit_latch_buggy = 1; + + /* GXm supports extended cpuid levels 'ala' AMD */ + if (c->cpuid_level == 2) { + /* Enable cxMMX extensions (GX1 Datasheet 54) */ + setCx86(CX86_CCR7, getCx86(CX86_CCR7)|1); + + /* GXlv/GXm/GX1 */ + if((dir1 >= 0x50 && dir1 <= 0x54) || dir1 >= 0x63) + geode_configure(); + get_model_name(c); /* get CPU marketing name */ + return; + } + else { /* MediaGX */ + Cx86_cb[2] = (dir0_lsn & 1) ? '3' : '4'; + p = Cx86_cb+2; + c->x86_model = (dir1 & 0x20) ? 1 : 2; + } + break; + + case 5: /* 6x86MX/M II */ + if (dir1 > 7) + { + dir0_msn++; /* M II */ + /* Enable MMX extensions (App note 108) */ + setCx86(CX86_CCR7, getCx86(CX86_CCR7)|1); + } + else + { + c->coma_bug = 1; /* 6x86MX, it has the bug. */ + } + tmp = (!(dir0_lsn & 7) || dir0_lsn & 1) ? 2 : 0; + Cx86_cb[tmp] = cyrix_model_mult2[dir0_lsn & 7]; + p = Cx86_cb+tmp; + if (((dir1 & 0x0f) > 4) || ((dir1 & 0xf0) == 0x20)) + (c->x86_model)++; + /* Emulate MTRRs using Cyrix's ARRs. */ + set_bit(X86_FEATURE_CYRIX_ARR, c->x86_capability); + break; + + case 0xf: /* Cyrix 486 without DEVID registers */ + switch (dir0_lsn) { + case 0xd: /* either a 486SLC or DLC w/o DEVID */ + dir0_msn = 0; + p = Cx486_name[(c->hard_math) ? 1 : 0]; + break; + + case 0xe: /* a 486S A step */ + dir0_msn = 0; + p = Cx486S_name[0]; + break; + } + break; + + default: /* unknown (shouldn't happen, we know everyone ;-) */ + dir0_msn = 7; + break; + } + strcpy(buf, Cx86_model[dir0_msn & 7]); + if (p) strcat(buf, p); + return; +} + +/* + * Cyrix CPUs without cpuid or with cpuid not yet enabled can be detected + * by the fact that they preserve the flags across the division of 5/2. + * PII and PPro exhibit this behavior too, but they have cpuid available. + */ + +/* + * Perform the Cyrix 5/2 test. A Cyrix won't change + * the flags, while other 486 chips will. + */ +static inline int test_cyrix_52div(void) +{ + unsigned int test; + + __asm__ __volatile__( + "sahf\n\t" /* clear flags (%eax = 0x0005) */ + "div %b2\n\t" /* divide 5 by 2 */ + "lahf" /* store flags into %ah */ + : "=a" (test) + : "0" (5), "q" (2) + : "cc"); + + /* AH is 0x02 on Cyrix after the divide.. */ + return (unsigned char) (test >> 8) == 0x02; +} + +static void cyrix_identify(struct cpuinfo_x86 * c) +{ + /* Detect Cyrix with disabled CPUID */ + if ( c->x86 == 4 && test_cyrix_52div() ) { + unsigned char dir0, dir1; + + strcpy(c->x86_vendor_id, "CyrixInstead"); + c->x86_vendor = X86_VENDOR_CYRIX; + + /* Actually enable cpuid on the older cyrix */ + + /* Retrieve CPU revisions */ + + do_cyrix_devid(&dir0, &dir1); + + dir0>>=4; + + /* Check it is an affected model */ + + if (dir0 == 5 || dir0 == 3) + { + unsigned char ccr3, ccr4; + unsigned long flags; + printk(KERN_INFO "Enabling CPUID on Cyrix processor.\n"); + local_irq_save(flags); + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + ccr4 = getCx86(CX86_CCR4); + setCx86(CX86_CCR4, ccr4 | 0x80); /* enable cpuid */ + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ + local_irq_restore(flags); + } + } + generic_identify(c); +} + +static struct cpu_dev cyrix_cpu_dev __initdata = { + .c_vendor = "Cyrix", + .c_ident = { "CyrixInstead" }, + .c_init = init_cyrix, + .c_identify = cyrix_identify, +}; + +int __init cyrix_init_cpu(void) +{ + cpu_devs[X86_VENDOR_CYRIX] = &cyrix_cpu_dev; + return 0; +} + +//early_arch_initcall(cyrix_init_cpu); + +static struct cpu_dev nsc_cpu_dev __initdata = { + .c_vendor = "NSC", + .c_ident = { "Geode by NSC" }, + .c_init = init_cyrix, + .c_identify = generic_identify, +}; + +int __init nsc_init_cpu(void) +{ + cpu_devs[X86_VENDOR_NSC] = &nsc_cpu_dev; + return 0; +} + +//early_arch_initcall(nsc_init_cpu); diff --git a/arch/i386/kernel/cpu/intel.c b/arch/i386/kernel/cpu/intel.c new file mode 100644 index 00000000000..b8d847b850d --- /dev/null +++ b/arch/i386/kernel/cpu/intel.c @@ -0,0 +1,248 @@ +#include <linux/config.h> +#include <linux/init.h> +#include <linux/kernel.h> + +#include <linux/string.h> +#include <linux/bitops.h> +#include <linux/smp.h> +#include <linux/thread_info.h> + +#include <asm/processor.h> +#include <asm/msr.h> +#include <asm/uaccess.h> + +#include "cpu.h" + +#ifdef CONFIG_X86_LOCAL_APIC +#include <asm/mpspec.h> +#include <asm/apic.h> +#include <mach_apic.h> +#endif + +extern int trap_init_f00f_bug(void); + +#ifdef CONFIG_X86_INTEL_USERCOPY +/* + * Alignment at which movsl is preferred for bulk memory copies. + */ +struct movsl_mask movsl_mask; +#endif + +void __init early_intel_workaround(struct cpuinfo_x86 *c) +{ + if (c->x86_vendor != X86_VENDOR_INTEL) + return; + /* Netburst reports 64 bytes clflush size, but does IO in 128 bytes */ + if (c->x86 == 15 && c->x86_cache_alignment == 64) + c->x86_cache_alignment = 128; +} + +/* + * Early probe support logic for ppro memory erratum #50 + * + * This is called before we do cpu ident work + */ + +int __init ppro_with_ram_bug(void) +{ + /* Uses data from early_cpu_detect now */ + if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL && + boot_cpu_data.x86 == 6 && + boot_cpu_data.x86_model == 1 && + boot_cpu_data.x86_mask < 8) { + printk(KERN_INFO "Pentium Pro with Errata#50 detected. Taking evasive action.\n"); + return 1; + } + return 0; +} + + +/* + * P4 Xeon errata 037 workaround. + * Hardware prefetcher may cause stale data to be loaded into the cache. + */ +static void __init Intel_errata_workarounds(struct cpuinfo_x86 *c) +{ + unsigned long lo, hi; + + if ((c->x86 == 15) && (c->x86_model == 1) && (c->x86_mask == 1)) { + rdmsr (MSR_IA32_MISC_ENABLE, lo, hi); + if ((lo & (1<<9)) == 0) { + printk (KERN_INFO "CPU: C0 stepping P4 Xeon detected.\n"); + printk (KERN_INFO "CPU: Disabling hardware prefetching (Errata 037)\n"); + lo |= (1<<9); /* Disable hw prefetching */ + wrmsr (MSR_IA32_MISC_ENABLE, lo, hi); + } + } +} + + +static void __init init_intel(struct cpuinfo_x86 *c) +{ + unsigned int l2 = 0; + char *p = NULL; + +#ifdef CONFIG_X86_F00F_BUG + /* + * All current models of Pentium and Pentium with MMX technology CPUs + * have the F0 0F bug, which lets nonprivileged users lock up the system. + * Note that the workaround only should be initialized once... + */ + c->f00f_bug = 0; + if ( c->x86 == 5 ) { + static int f00f_workaround_enabled = 0; + + c->f00f_bug = 1; + if ( !f00f_workaround_enabled ) { + trap_init_f00f_bug(); + printk(KERN_NOTICE "Intel Pentium with F0 0F bug - workaround enabled.\n"); + f00f_workaround_enabled = 1; + } + } +#endif + + select_idle_routine(c); + l2 = init_intel_cacheinfo(c); + + /* SEP CPUID bug: Pentium Pro reports SEP but doesn't have it until model 3 mask 3 */ + if ((c->x86<<8 | c->x86_model<<4 | c->x86_mask) < 0x633) + clear_bit(X86_FEATURE_SEP, c->x86_capability); + + /* Names for the Pentium II/Celeron processors + detectable only by also checking the cache size. + Dixon is NOT a Celeron. */ + if (c->x86 == 6) { + switch (c->x86_model) { + case 5: + if (c->x86_mask == 0) { + if (l2 == 0) + p = "Celeron (Covington)"; + else if (l2 == 256) + p = "Mobile Pentium II (Dixon)"; + } + break; + + case 6: + if (l2 == 128) + p = "Celeron (Mendocino)"; + else if (c->x86_mask == 0 || c->x86_mask == 5) + p = "Celeron-A"; + break; + + case 8: + if (l2 == 128) + p = "Celeron (Coppermine)"; + break; + } + } + + if ( p ) + strcpy(c->x86_model_id, p); + + detect_ht(c); + + /* Work around errata */ + Intel_errata_workarounds(c); + +#ifdef CONFIG_X86_INTEL_USERCOPY + /* + * Set up the preferred alignment for movsl bulk memory moves + */ + switch (c->x86) { + case 4: /* 486: untested */ + break; + case 5: /* Old Pentia: untested */ + break; + case 6: /* PII/PIII only like movsl with 8-byte alignment */ + movsl_mask.mask = 7; + break; + case 15: /* P4 is OK down to 8-byte alignment */ + movsl_mask.mask = 7; + break; + } +#endif + + if (c->x86 == 15) + set_bit(X86_FEATURE_P4, c->x86_capability); + if (c->x86 == 6) + set_bit(X86_FEATURE_P3, c->x86_capability); +} + + +static unsigned int intel_size_cache(struct cpuinfo_x86 * c, unsigned int size) +{ + /* Intel PIII Tualatin. This comes in two flavours. + * One has 256kb of cache, the other 512. We have no way + * to determine which, so we use a boottime override + * for the 512kb model, and assume 256 otherwise. + */ + if ((c->x86 == 6) && (c->x86_model == 11) && (size == 0)) + size = 256; + return size; +} + +static struct cpu_dev intel_cpu_dev __initdata = { + .c_vendor = "Intel", + .c_ident = { "GenuineIntel" }, + .c_models = { + { .vendor = X86_VENDOR_INTEL, .family = 4, .model_names = + { + [0] = "486 DX-25/33", + [1] = "486 DX-50", + [2] = "486 SX", + [3] = "486 DX/2", + [4] = "486 SL", + [5] = "486 SX/2", + [7] = "486 DX/2-WB", + [8] = "486 DX/4", + [9] = "486 DX/4-WB" + } + }, + { .vendor = X86_VENDOR_INTEL, .family = 5, .model_names = + { + [0] = "Pentium 60/66 A-step", + [1] = "Pentium 60/66", + [2] = "Pentium 75 - 200", + [3] = "OverDrive PODP5V83", + [4] = "Pentium MMX", + [7] = "Mobile Pentium 75 - 200", + [8] = "Mobile Pentium MMX" + } + }, + { .vendor = X86_VENDOR_INTEL, .family = 6, .model_names = + { + [0] = "Pentium Pro A-step", + [1] = "Pentium Pro", + [3] = "Pentium II (Klamath)", + [4] = "Pentium II (Deschutes)", + [5] = "Pentium II (Deschutes)", + [6] = "Mobile Pentium II", + [7] = "Pentium III (Katmai)", + [8] = "Pentium III (Coppermine)", + [10] = "Pentium III (Cascades)", + [11] = "Pentium III (Tualatin)", + } + }, + { .vendor = X86_VENDOR_INTEL, .family = 15, .model_names = + { + [0] = "Pentium 4 (Unknown)", + [1] = "Pentium 4 (Willamette)", + [2] = "Pentium 4 (Northwood)", + [4] = "Pentium 4 (Foster)", + [5] = "Pentium 4 (Foster)", + } + }, + }, + .c_init = init_intel, + .c_identify = generic_identify, + .c_size_cache = intel_size_cache, +}; + +__init int intel_cpu_init(void) +{ + cpu_devs[X86_VENDOR_INTEL] = &intel_cpu_dev; + return 0; +} + +// arch_initcall(intel_cpu_init); + diff --git a/arch/i386/kernel/cpu/intel_cacheinfo.c b/arch/i386/kernel/cpu/intel_cacheinfo.c new file mode 100644 index 00000000000..aeb5b4ef8c8 --- /dev/null +++ b/arch/i386/kernel/cpu/intel_cacheinfo.c @@ -0,0 +1,598 @@ +/* + * Routines to indentify caches on Intel CPU. + * + * Changes: + * Venkatesh Pallipadi : Adding cache identification through cpuid(4) + */ + +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/device.h> +#include <linux/compiler.h> +#include <linux/cpu.h> + +#include <asm/processor.h> +#include <asm/smp.h> + +#define LVL_1_INST 1 +#define LVL_1_DATA 2 +#define LVL_2 3 +#define LVL_3 4 +#define LVL_TRACE 5 + +struct _cache_table +{ + unsigned char descriptor; + char cache_type; + short size; +}; + +/* all the cache descriptor types we care about (no TLB or trace cache entries) */ +static struct _cache_table cache_table[] __initdata = +{ + { 0x06, LVL_1_INST, 8 }, /* 4-way set assoc, 32 byte line size */ + { 0x08, LVL_1_INST, 16 }, /* 4-way set assoc, 32 byte line size */ + { 0x0a, LVL_1_DATA, 8 }, /* 2 way set assoc, 32 byte line size */ + { 0x0c, LVL_1_DATA, 16 }, /* 4-way set assoc, 32 byte line size */ + { 0x22, LVL_3, 512 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x23, LVL_3, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x25, LVL_3, 2048 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x29, LVL_3, 4096 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x2c, LVL_1_DATA, 32 }, /* 8-way set assoc, 64 byte line size */ + { 0x30, LVL_1_INST, 32 }, /* 8-way set assoc, 64 byte line size */ + { 0x39, LVL_2, 128 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x3b, LVL_2, 128 }, /* 2-way set assoc, sectored cache, 64 byte line size */ + { 0x3c, LVL_2, 256 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x41, LVL_2, 128 }, /* 4-way set assoc, 32 byte line size */ + { 0x42, LVL_2, 256 }, /* 4-way set assoc, 32 byte line size */ + { 0x43, LVL_2, 512 }, /* 4-way set assoc, 32 byte line size */ + { 0x44, LVL_2, 1024 }, /* 4-way set assoc, 32 byte line size */ + { 0x45, LVL_2, 2048 }, /* 4-way set assoc, 32 byte line size */ + { 0x60, LVL_1_DATA, 16 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x66, LVL_1_DATA, 8 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x67, LVL_1_DATA, 16 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x68, LVL_1_DATA, 32 }, /* 4-way set assoc, sectored cache, 64 byte line size */ + { 0x70, LVL_TRACE, 12 }, /* 8-way set assoc */ + { 0x71, LVL_TRACE, 16 }, /* 8-way set assoc */ + { 0x72, LVL_TRACE, 32 }, /* 8-way set assoc */ + { 0x78, LVL_2, 1024 }, /* 4-way set assoc, 64 byte line size */ + { 0x79, LVL_2, 128 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7a, LVL_2, 256 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7b, LVL_2, 512 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7c, LVL_2, 1024 }, /* 8-way set assoc, sectored cache, 64 byte line size */ + { 0x7d, LVL_2, 2048 }, /* 8-way set assoc, 64 byte line size */ + { 0x7f, LVL_2, 512 }, /* 2-way set assoc, 64 byte line size */ + { 0x82, LVL_2, 256 }, /* 8-way set assoc, 32 byte line size */ + { 0x83, LVL_2, 512 }, /* 8-way set assoc, 32 byte line size */ + { 0x84, LVL_2, 1024 }, /* 8-way set assoc, 32 byte line size */ + { 0x85, LVL_2, 2048 }, /* 8-way set assoc, 32 byte line size */ + { 0x86, LVL_2, 512 }, /* 4-way set assoc, 64 byte line size */ + { 0x87, LVL_2, 1024 }, /* 8-way set assoc, 64 byte line size */ + { 0x00, 0, 0} +}; + + +enum _cache_type +{ + CACHE_TYPE_NULL = 0, + CACHE_TYPE_DATA = 1, + CACHE_TYPE_INST = 2, + CACHE_TYPE_UNIFIED = 3 +}; + +union _cpuid4_leaf_eax { + struct { + enum _cache_type type:5; + unsigned int level:3; + unsigned int is_self_initializing:1; + unsigned int is_fully_associative:1; + unsigned int reserved:4; + unsigned int num_threads_sharing:12; + unsigned int num_cores_on_die:6; + } split; + u32 full; +}; + +union _cpuid4_leaf_ebx { + struct { + unsigned int coherency_line_size:12; + unsigned int physical_line_partition:10; + unsigned int ways_of_associativity:10; + } split; + u32 full; +}; + +union _cpuid4_leaf_ecx { + struct { + unsigned int number_of_sets:32; + } split; + u32 full; +}; + +struct _cpuid4_info { + union _cpuid4_leaf_eax eax; + union _cpuid4_leaf_ebx ebx; + union _cpuid4_leaf_ecx ecx; + unsigned long size; + cpumask_t shared_cpu_map; +}; + +#define MAX_CACHE_LEAVES 4 +static unsigned short __devinitdata num_cache_leaves; + +static int __devinit cpuid4_cache_lookup(int index, struct _cpuid4_info *this_leaf) +{ + unsigned int eax, ebx, ecx, edx; + union _cpuid4_leaf_eax cache_eax; + + cpuid_count(4, index, &eax, &ebx, &ecx, &edx); + cache_eax.full = eax; + if (cache_eax.split.type == CACHE_TYPE_NULL) + return -1; + + this_leaf->eax.full = eax; + this_leaf->ebx.full = ebx; + this_leaf->ecx.full = ecx; + this_leaf->size = (this_leaf->ecx.split.number_of_sets + 1) * + (this_leaf->ebx.split.coherency_line_size + 1) * + (this_leaf->ebx.split.physical_line_partition + 1) * + (this_leaf->ebx.split.ways_of_associativity + 1); + return 0; +} + +static int __init find_num_cache_leaves(void) +{ + unsigned int eax, ebx, ecx, edx; + union _cpuid4_leaf_eax cache_eax; + int i; + int retval; + + retval = MAX_CACHE_LEAVES; + /* Do cpuid(4) loop to find out num_cache_leaves */ + for (i = 0; i < MAX_CACHE_LEAVES; i++) { + cpuid_count(4, i, &eax, &ebx, &ecx, &edx); + cache_eax.full = eax; + if (cache_eax.split.type == CACHE_TYPE_NULL) { + retval = i; + break; + } + } + return retval; +} + +unsigned int __init init_intel_cacheinfo(struct cpuinfo_x86 *c) +{ + unsigned int trace = 0, l1i = 0, l1d = 0, l2 = 0, l3 = 0; /* Cache sizes */ + unsigned int new_l1d = 0, new_l1i = 0; /* Cache sizes from cpuid(4) */ + unsigned int new_l2 = 0, new_l3 = 0, i; /* Cache sizes from cpuid(4) */ + + if (c->cpuid_level > 4) { + static int is_initialized; + + if (is_initialized == 0) { + /* Init num_cache_leaves from boot CPU */ + num_cache_leaves = find_num_cache_leaves(); + is_initialized++; + } + + /* + * Whenever possible use cpuid(4), deterministic cache + * parameters cpuid leaf to find the cache details + */ + for (i = 0; i < num_cache_leaves; i++) { + struct _cpuid4_info this_leaf; + + int retval; + + retval = cpuid4_cache_lookup(i, &this_leaf); + if (retval >= 0) { + switch(this_leaf.eax.split.level) { + case 1: + if (this_leaf.eax.split.type == + CACHE_TYPE_DATA) + new_l1d = this_leaf.size/1024; + else if (this_leaf.eax.split.type == + CACHE_TYPE_INST) + new_l1i = this_leaf.size/1024; + break; + case 2: + new_l2 = this_leaf.size/1024; + break; + case 3: + new_l3 = this_leaf.size/1024; + break; + default: + break; + } + } + } + } + if (c->cpuid_level > 1) { + /* supports eax=2 call */ + int i, j, n; + int regs[4]; + unsigned char *dp = (unsigned char *)regs; + + /* Number of times to iterate */ + n = cpuid_eax(2) & 0xFF; + + for ( i = 0 ; i < n ; i++ ) { + cpuid(2, ®s[0], ®s[1], ®s[2], ®s[3]); + + /* If bit 31 is set, this is an unknown format */ + for ( j = 0 ; j < 3 ; j++ ) { + if ( regs[j] < 0 ) regs[j] = 0; + } + + /* Byte 0 is level count, not a descriptor */ + for ( j = 1 ; j < 16 ; j++ ) { + unsigned char des = dp[j]; + unsigned char k = 0; + + /* look up this descriptor in the table */ + while (cache_table[k].descriptor != 0) + { + if (cache_table[k].descriptor == des) { + switch (cache_table[k].cache_type) { + case LVL_1_INST: + l1i += cache_table[k].size; + break; + case LVL_1_DATA: + l1d += cache_table[k].size; + break; + case LVL_2: + l2 += cache_table[k].size; + break; + case LVL_3: + l3 += cache_table[k].size; + break; + case LVL_TRACE: + trace += cache_table[k].size; + break; + } + + break; + } + + k++; + } + } + } + + if (new_l1d) + l1d = new_l1d; + + if (new_l1i) + l1i = new_l1i; + + if (new_l2) + l2 = new_l2; + + if (new_l3) + l3 = new_l3; + + if ( trace ) + printk (KERN_INFO "CPU: Trace cache: %dK uops", trace); + else if ( l1i ) + printk (KERN_INFO "CPU: L1 I cache: %dK", l1i); + if ( l1d ) + printk(", L1 D cache: %dK\n", l1d); + else + printk("\n"); + if ( l2 ) + printk(KERN_INFO "CPU: L2 cache: %dK\n", l2); + if ( l3 ) + printk(KERN_INFO "CPU: L3 cache: %dK\n", l3); + + /* + * This assumes the L3 cache is shared; it typically lives in + * the northbridge. The L1 caches are included by the L2 + * cache, and so should not be included for the purpose of + * SMP switching weights. + */ + c->x86_cache_size = l2 ? l2 : (l1i+l1d); + } + + return l2; +} + +/* pointer to _cpuid4_info array (for each cache leaf) */ +static struct _cpuid4_info *cpuid4_info[NR_CPUS]; +#define CPUID4_INFO_IDX(x,y) (&((cpuid4_info[x])[y])) + +#ifdef CONFIG_SMP +static void __devinit cache_shared_cpu_map_setup(unsigned int cpu, int index) +{ + struct _cpuid4_info *this_leaf; + unsigned long num_threads_sharing; + + this_leaf = CPUID4_INFO_IDX(cpu, index); + num_threads_sharing = 1 + this_leaf->eax.split.num_threads_sharing; + + if (num_threads_sharing == 1) + cpu_set(cpu, this_leaf->shared_cpu_map); +#ifdef CONFIG_X86_HT + else if (num_threads_sharing == smp_num_siblings) + this_leaf->shared_cpu_map = cpu_sibling_map[cpu]; +#endif + else + printk(KERN_INFO "Number of CPUs sharing cache didn't match " + "any known set of CPUs\n"); +} +#else +static void __init cache_shared_cpu_map_setup(unsigned int cpu, int index) {} +#endif + +static void free_cache_attributes(unsigned int cpu) +{ + kfree(cpuid4_info[cpu]); + cpuid4_info[cpu] = NULL; +} + +static int __devinit detect_cache_attributes(unsigned int cpu) +{ + struct _cpuid4_info *this_leaf; + unsigned long j; + int retval; + + if (num_cache_leaves == 0) + return -ENOENT; + + cpuid4_info[cpu] = kmalloc( + sizeof(struct _cpuid4_info) * num_cache_leaves, GFP_KERNEL); + if (unlikely(cpuid4_info[cpu] == NULL)) + return -ENOMEM; + memset(cpuid4_info[cpu], 0, + sizeof(struct _cpuid4_info) * num_cache_leaves); + + /* Do cpuid and store the results */ + for (j = 0; j < num_cache_leaves; j++) { + this_leaf = CPUID4_INFO_IDX(cpu, j); + retval = cpuid4_cache_lookup(j, this_leaf); + if (unlikely(retval < 0)) + goto err_out; + cache_shared_cpu_map_setup(cpu, j); + } + return 0; + +err_out: + free_cache_attributes(cpu); + return -ENOMEM; +} + +#ifdef CONFIG_SYSFS + +#include <linux/kobject.h> +#include <linux/sysfs.h> + +extern struct sysdev_class cpu_sysdev_class; /* from drivers/base/cpu.c */ + +/* pointer to kobject for cpuX/cache */ +static struct kobject * cache_kobject[NR_CPUS]; + +struct _index_kobject { + struct kobject kobj; + unsigned int cpu; + unsigned short index; +}; + +/* pointer to array of kobjects for cpuX/cache/indexY */ +static struct _index_kobject *index_kobject[NR_CPUS]; +#define INDEX_KOBJECT_PTR(x,y) (&((index_kobject[x])[y])) + +#define show_one_plus(file_name, object, val) \ +static ssize_t show_##file_name \ + (struct _cpuid4_info *this_leaf, char *buf) \ +{ \ + return sprintf (buf, "%lu\n", (unsigned long)this_leaf->object + val); \ +} + +show_one_plus(level, eax.split.level, 0); +show_one_plus(coherency_line_size, ebx.split.coherency_line_size, 1); +show_one_plus(physical_line_partition, ebx.split.physical_line_partition, 1); +show_one_plus(ways_of_associativity, ebx.split.ways_of_associativity, 1); +show_one_plus(number_of_sets, ecx.split.number_of_sets, 1); + +static ssize_t show_size(struct _cpuid4_info *this_leaf, char *buf) +{ + return sprintf (buf, "%luK\n", this_leaf->size / 1024); +} + +static ssize_t show_shared_cpu_map(struct _cpuid4_info *this_leaf, char *buf) +{ + char mask_str[NR_CPUS]; + cpumask_scnprintf(mask_str, NR_CPUS, this_leaf->shared_cpu_map); + return sprintf(buf, "%s\n", mask_str); +} + +static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) { + switch(this_leaf->eax.split.type) { + case CACHE_TYPE_DATA: + return sprintf(buf, "Data\n"); + break; + case CACHE_TYPE_INST: + return sprintf(buf, "Instruction\n"); + break; + case CACHE_TYPE_UNIFIED: + return sprintf(buf, "Unified\n"); + break; + default: + return sprintf(buf, "Unknown\n"); + break; + } +} + +struct _cache_attr { + struct attribute attr; + ssize_t (*show)(struct _cpuid4_info *, char *); + ssize_t (*store)(struct _cpuid4_info *, const char *, size_t count); +}; + +#define define_one_ro(_name) \ +static struct _cache_attr _name = \ + __ATTR(_name, 0444, show_##_name, NULL) + +define_one_ro(level); +define_one_ro(type); +define_one_ro(coherency_line_size); +define_one_ro(physical_line_partition); +define_one_ro(ways_of_associativity); +define_one_ro(number_of_sets); +define_one_ro(size); +define_one_ro(shared_cpu_map); + +static struct attribute * default_attrs[] = { + &type.attr, + &level.attr, + &coherency_line_size.attr, + &physical_line_partition.attr, + &ways_of_associativity.attr, + &number_of_sets.attr, + &size.attr, + &shared_cpu_map.attr, + NULL +}; + +#define to_object(k) container_of(k, struct _index_kobject, kobj) +#define to_attr(a) container_of(a, struct _cache_attr, attr) + +static ssize_t show(struct kobject * kobj, struct attribute * attr, char * buf) +{ + struct _cache_attr *fattr = to_attr(attr); + struct _index_kobject *this_leaf = to_object(kobj); + ssize_t ret; + + ret = fattr->show ? + fattr->show(CPUID4_INFO_IDX(this_leaf->cpu, this_leaf->index), + buf) : + 0; + return ret; +} + +static ssize_t store(struct kobject * kobj, struct attribute * attr, + const char * buf, size_t count) +{ + return 0; +} + +static struct sysfs_ops sysfs_ops = { + .show = show, + .store = store, +}; + +static struct kobj_type ktype_cache = { + .sysfs_ops = &sysfs_ops, + .default_attrs = default_attrs, +}; + +static struct kobj_type ktype_percpu_entry = { + .sysfs_ops = &sysfs_ops, +}; + +static void cpuid4_cache_sysfs_exit(unsigned int cpu) +{ + kfree(cache_kobject[cpu]); + kfree(index_kobject[cpu]); + cache_kobject[cpu] = NULL; + index_kobject[cpu] = NULL; + free_cache_attributes(cpu); +} + +static int __devinit cpuid4_cache_sysfs_init(unsigned int cpu) +{ + + if (num_cache_leaves == 0) + return -ENOENT; + + detect_cache_attributes(cpu); + if (cpuid4_info[cpu] == NULL) + return -ENOENT; + + /* Allocate all required memory */ + cache_kobject[cpu] = kmalloc(sizeof(struct kobject), GFP_KERNEL); + if (unlikely(cache_kobject[cpu] == NULL)) + goto err_out; + memset(cache_kobject[cpu], 0, sizeof(struct kobject)); + + index_kobject[cpu] = kmalloc( + sizeof(struct _index_kobject ) * num_cache_leaves, GFP_KERNEL); + if (unlikely(index_kobject[cpu] == NULL)) + goto err_out; + memset(index_kobject[cpu], 0, + sizeof(struct _index_kobject) * num_cache_leaves); + + return 0; + +err_out: + cpuid4_cache_sysfs_exit(cpu); + return -ENOMEM; +} + +/* Add/Remove cache interface for CPU device */ +static int __devinit cache_add_dev(struct sys_device * sys_dev) +{ + unsigned int cpu = sys_dev->id; + unsigned long i, j; + struct _index_kobject *this_object; + int retval = 0; + + retval = cpuid4_cache_sysfs_init(cpu); + if (unlikely(retval < 0)) + return retval; + + cache_kobject[cpu]->parent = &sys_dev->kobj; + kobject_set_name(cache_kobject[cpu], "%s", "cache"); + cache_kobject[cpu]->ktype = &ktype_percpu_entry; + retval = kobject_register(cache_kobject[cpu]); + + for (i = 0; i < num_cache_leaves; i++) { + this_object = INDEX_KOBJECT_PTR(cpu,i); + this_object->cpu = cpu; + this_object->index = i; + this_object->kobj.parent = cache_kobject[cpu]; + kobject_set_name(&(this_object->kobj), "index%1lu", i); + this_object->kobj.ktype = &ktype_cache; + retval = kobject_register(&(this_object->kobj)); + if (unlikely(retval)) { + for (j = 0; j < i; j++) { + kobject_unregister( + &(INDEX_KOBJECT_PTR(cpu,j)->kobj)); + } + kobject_unregister(cache_kobject[cpu]); + cpuid4_cache_sysfs_exit(cpu); + break; + } + } + return retval; +} + +static int __devexit cache_remove_dev(struct sys_device * sys_dev) +{ + unsigned int cpu = sys_dev->id; + unsigned long i; + + for (i = 0; i < num_cache_leaves; i++) + kobject_unregister(&(INDEX_KOBJECT_PTR(cpu,i)->kobj)); + kobject_unregister(cache_kobject[cpu]); + cpuid4_cache_sysfs_exit(cpu); + return 0; +} + +static struct sysdev_driver cache_sysdev_driver = { + .add = cache_add_dev, + .remove = __devexit_p(cache_remove_dev), +}; + +/* Register/Unregister the cpu_cache driver */ +static int __devinit cache_register_driver(void) +{ + if (num_cache_leaves == 0) + return 0; + + return sysdev_driver_register(&cpu_sysdev_class,&cache_sysdev_driver); +} + +device_initcall(cache_register_driver); + +#endif + diff --git a/arch/i386/kernel/cpu/mcheck/Makefile b/arch/i386/kernel/cpu/mcheck/Makefile new file mode 100644 index 00000000000..30808f3d671 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/Makefile @@ -0,0 +1,2 @@ +obj-y = mce.o k7.o p4.o p5.o p6.o winchip.o +obj-$(CONFIG_X86_MCE_NONFATAL) += non-fatal.o diff --git a/arch/i386/kernel/cpu/mcheck/k7.c b/arch/i386/kernel/cpu/mcheck/k7.c new file mode 100644 index 00000000000..8df52e86c4d --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/k7.c @@ -0,0 +1,97 @@ +/* + * Athlon/Hammer specific Machine Check Exception Reporting + * (C) Copyright 2002 Dave Jones <davej@codemonkey.org.uk> + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/config.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/smp.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> + +#include "mce.h" + +/* Machine Check Handler For AMD Athlon/Duron */ +static fastcall void k7_machine_check(struct pt_regs * regs, long error_code) +{ + int recover=1; + u32 alow, ahigh, high, low; + u32 mcgstl, mcgsth; + int i; + + rdmsr (MSR_IA32_MCG_STATUS, mcgstl, mcgsth); + if (mcgstl & (1<<0)) /* Recoverable ? */ + recover=0; + + printk (KERN_EMERG "CPU %d: Machine Check Exception: %08x%08x\n", + smp_processor_id(), mcgsth, mcgstl); + + for (i=1; i<nr_mce_banks; i++) { + rdmsr (MSR_IA32_MC0_STATUS+i*4,low, high); + if (high&(1<<31)) { + if (high & (1<<29)) + recover |= 1; + if (high & (1<<25)) + recover |= 2; + printk (KERN_EMERG "Bank %d: %08x%08x", i, high, low); + high &= ~(1<<31); + if (high & (1<<27)) { + rdmsr (MSR_IA32_MC0_MISC+i*4, alow, ahigh); + printk ("[%08x%08x]", ahigh, alow); + } + if (high & (1<<26)) { + rdmsr (MSR_IA32_MC0_ADDR+i*4, alow, ahigh); + printk (" at %08x%08x", ahigh, alow); + } + printk ("\n"); + /* Clear it */ + wrmsr (MSR_IA32_MC0_STATUS+i*4, 0UL, 0UL); + /* Serialize */ + wmb(); + add_taint(TAINT_MACHINE_CHECK); + } + } + + if (recover&2) + panic ("CPU context corrupt"); + if (recover&1) + panic ("Unable to continue"); + printk (KERN_EMERG "Attempting to continue.\n"); + mcgstl &= ~(1<<2); + wrmsr (MSR_IA32_MCG_STATUS,mcgstl, mcgsth); +} + + +/* AMD K7 machine check is Intel like */ +void __init amd_mcheck_init(struct cpuinfo_x86 *c) +{ + u32 l, h; + int i; + + machine_check_vector = k7_machine_check; + wmb(); + + printk (KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr (MSR_IA32_MCG_CAP, l, h); + if (l & (1<<8)) /* Control register present ? */ + wrmsr (MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); + nr_mce_banks = l & 0xff; + + /* Clear status for MC index 0 separately, we don't touch CTL, + * as some Athlons cause spurious MCEs when its enabled. */ + wrmsr (MSR_IA32_MC0_STATUS, 0x0, 0x0); + for (i=1; i<nr_mce_banks; i++) { + wrmsr (MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); + wrmsr (MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); + } + + set_in_cr4 (X86_CR4_MCE); + printk (KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", + smp_processor_id()); +} diff --git a/arch/i386/kernel/cpu/mcheck/mce.c b/arch/i386/kernel/cpu/mcheck/mce.c new file mode 100644 index 00000000000..bf6d1aefafc --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/mce.c @@ -0,0 +1,77 @@ +/* + * mce.c - x86 Machine Check Exception Reporting + * (c) 2002 Alan Cox <alan@redhat.com>, Dave Jones <davej@codemonkey.org.uk> + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/config.h> +#include <linux/module.h> +#include <linux/smp.h> +#include <linux/thread_info.h> + +#include <asm/processor.h> +#include <asm/system.h> + +#include "mce.h" + +int mce_disabled __initdata = 0; +int nr_mce_banks; + +EXPORT_SYMBOL_GPL(nr_mce_banks); /* non-fatal.o */ + +/* Handle unconfigured int18 (should never happen) */ +static fastcall void unexpected_machine_check(struct pt_regs * regs, long error_code) +{ + printk(KERN_ERR "CPU#%d: Unexpected int18 (Machine Check).\n", smp_processor_id()); +} + +/* Call the installed machine check handler for this CPU setup. */ +void fastcall (*machine_check_vector)(struct pt_regs *, long error_code) = unexpected_machine_check; + +/* This has to be run for each processor */ +void __init mcheck_init(struct cpuinfo_x86 *c) +{ + if (mce_disabled==1) + return; + + switch (c->x86_vendor) { + case X86_VENDOR_AMD: + if (c->x86==6 || c->x86==15) + amd_mcheck_init(c); + break; + + case X86_VENDOR_INTEL: + if (c->x86==5) + intel_p5_mcheck_init(c); + if (c->x86==6) + intel_p6_mcheck_init(c); + if (c->x86==15) + intel_p4_mcheck_init(c); + break; + + case X86_VENDOR_CENTAUR: + if (c->x86==5) + winchip_mcheck_init(c); + break; + + default: + break; + } +} + +static int __init mcheck_disable(char *str) +{ + mce_disabled = 1; + return 0; +} + +static int __init mcheck_enable(char *str) +{ + mce_disabled = -1; + return 0; +} + +__setup("nomce", mcheck_disable); +__setup("mce", mcheck_enable); diff --git a/arch/i386/kernel/cpu/mcheck/mce.h b/arch/i386/kernel/cpu/mcheck/mce.h new file mode 100644 index 00000000000..dc2416dfef1 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/mce.h @@ -0,0 +1,14 @@ +#include <linux/init.h> + +void amd_mcheck_init(struct cpuinfo_x86 *c); +void intel_p4_mcheck_init(struct cpuinfo_x86 *c); +void intel_p5_mcheck_init(struct cpuinfo_x86 *c); +void intel_p6_mcheck_init(struct cpuinfo_x86 *c); +void winchip_mcheck_init(struct cpuinfo_x86 *c); + +/* Call the installed machine check handler for this CPU setup. */ +extern fastcall void (*machine_check_vector)(struct pt_regs *, long error_code); + +extern int mce_disabled __initdata; +extern int nr_mce_banks; + diff --git a/arch/i386/kernel/cpu/mcheck/non-fatal.c b/arch/i386/kernel/cpu/mcheck/non-fatal.c new file mode 100644 index 00000000000..7864ddfccf0 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/non-fatal.c @@ -0,0 +1,93 @@ +/* + * Non Fatal Machine Check Exception Reporting + * + * (C) Copyright 2002 Dave Jones. <davej@codemonkey.org.uk> + * + * This file contains routines to check for non-fatal MCEs every 15s + * + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/jiffies.h> +#include <linux/config.h> +#include <linux/irq.h> +#include <linux/workqueue.h> +#include <linux/interrupt.h> +#include <linux/smp.h> +#include <linux/module.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> + +#include "mce.h" + +static int firstbank; + +#define MCE_RATE 15*HZ /* timer rate is 15s */ + +static void mce_checkregs (void *info) +{ + u32 low, high; + int i; + + for (i=firstbank; i<nr_mce_banks; i++) { + rdmsr (MSR_IA32_MC0_STATUS+i*4, low, high); + + if (high & (1<<31)) { + printk(KERN_INFO "MCE: The hardware reports a non " + "fatal, correctable incident occurred on " + "CPU %d.\n", + smp_processor_id()); + printk (KERN_INFO "Bank %d: %08x%08x\n", i, high, low); + + /* Scrub the error so we don't pick it up in MCE_RATE seconds time. */ + wrmsr (MSR_IA32_MC0_STATUS+i*4, 0UL, 0UL); + + /* Serialize */ + wmb(); + add_taint(TAINT_MACHINE_CHECK); + } + } +} + +static void mce_work_fn(void *data); +static DECLARE_WORK(mce_work, mce_work_fn, NULL); + +static void mce_work_fn(void *data) +{ + on_each_cpu(mce_checkregs, NULL, 1, 1); + schedule_delayed_work(&mce_work, MCE_RATE); +} + +static int __init init_nonfatal_mce_checker(void) +{ + struct cpuinfo_x86 *c = &boot_cpu_data; + + /* Check for MCE support */ + if (!cpu_has(c, X86_FEATURE_MCE)) + return -ENODEV; + + /* Check for PPro style MCA */ + if (!cpu_has(c, X86_FEATURE_MCA)) + return -ENODEV; + + /* Some Athlons misbehave when we frob bank 0 */ + if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD && + boot_cpu_data.x86 == 6) + firstbank = 1; + else + firstbank = 0; + + /* + * Check for non-fatal errors every MCE_RATE s + */ + schedule_delayed_work(&mce_work, MCE_RATE); + printk(KERN_INFO "Machine check exception polling timer started.\n"); + return 0; +} +module_init(init_nonfatal_mce_checker); + +MODULE_LICENSE("GPL"); diff --git a/arch/i386/kernel/cpu/mcheck/p4.c b/arch/i386/kernel/cpu/mcheck/p4.c new file mode 100644 index 00000000000..8b16ceb929b --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/p4.c @@ -0,0 +1,271 @@ +/* + * P4 specific Machine Check Exception Reporting + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/config.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/smp.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> +#include <asm/apic.h> + +#include "mce.h" + +/* as supported by the P4/Xeon family */ +struct intel_mce_extended_msrs { + u32 eax; + u32 ebx; + u32 ecx; + u32 edx; + u32 esi; + u32 edi; + u32 ebp; + u32 esp; + u32 eflags; + u32 eip; + /* u32 *reserved[]; */ +}; + +static int mce_num_extended_msrs = 0; + + +#ifdef CONFIG_X86_MCE_P4THERMAL +static void unexpected_thermal_interrupt(struct pt_regs *regs) +{ + printk(KERN_ERR "CPU%d: Unexpected LVT TMR interrupt!\n", + smp_processor_id()); + add_taint(TAINT_MACHINE_CHECK); +} + +/* P4/Xeon Thermal transition interrupt handler */ +static void intel_thermal_interrupt(struct pt_regs *regs) +{ + u32 l, h; + unsigned int cpu = smp_processor_id(); + static unsigned long next[NR_CPUS]; + + ack_APIC_irq(); + + if (time_after(next[cpu], jiffies)) + return; + + next[cpu] = jiffies + HZ*5; + rdmsr(MSR_IA32_THERM_STATUS, l, h); + if (l & 0x1) { + printk(KERN_EMERG "CPU%d: Temperature above threshold\n", cpu); + printk(KERN_EMERG "CPU%d: Running in modulated clock mode\n", + cpu); + add_taint(TAINT_MACHINE_CHECK); + } else { + printk(KERN_INFO "CPU%d: Temperature/speed normal\n", cpu); + } +} + +/* Thermal interrupt handler for this CPU setup */ +static void (*vendor_thermal_interrupt)(struct pt_regs *regs) = unexpected_thermal_interrupt; + +fastcall void smp_thermal_interrupt(struct pt_regs *regs) +{ + irq_enter(); + vendor_thermal_interrupt(regs); + irq_exit(); +} + +/* P4/Xeon Thermal regulation detect and init */ +static void __init intel_init_thermal(struct cpuinfo_x86 *c) +{ + u32 l, h; + unsigned int cpu = smp_processor_id(); + + /* Thermal monitoring */ + if (!cpu_has(c, X86_FEATURE_ACPI)) + return; /* -ENODEV */ + + /* Clock modulation */ + if (!cpu_has(c, X86_FEATURE_ACC)) + return; /* -ENODEV */ + + /* first check if its enabled already, in which case there might + * be some SMM goo which handles it, so we can't even put a handler + * since it might be delivered via SMI already -zwanem. + */ + rdmsr (MSR_IA32_MISC_ENABLE, l, h); + h = apic_read(APIC_LVTTHMR); + if ((l & (1<<3)) && (h & APIC_DM_SMI)) { + printk(KERN_DEBUG "CPU%d: Thermal monitoring handled by SMI\n", + cpu); + return; /* -EBUSY */ + } + + /* check whether a vector already exists, temporarily masked? */ + if (h & APIC_VECTOR_MASK) { + printk(KERN_DEBUG "CPU%d: Thermal LVT vector (%#x) already " + "installed\n", + cpu, (h & APIC_VECTOR_MASK)); + return; /* -EBUSY */ + } + + /* The temperature transition interrupt handler setup */ + h = THERMAL_APIC_VECTOR; /* our delivery vector */ + h |= (APIC_DM_FIXED | APIC_LVT_MASKED); /* we'll mask till we're ready */ + apic_write_around(APIC_LVTTHMR, h); + + rdmsr (MSR_IA32_THERM_INTERRUPT, l, h); + wrmsr (MSR_IA32_THERM_INTERRUPT, l | 0x03 , h); + + /* ok we're good to go... */ + vendor_thermal_interrupt = intel_thermal_interrupt; + + rdmsr (MSR_IA32_MISC_ENABLE, l, h); + wrmsr (MSR_IA32_MISC_ENABLE, l | (1<<3), h); + + l = apic_read (APIC_LVTTHMR); + apic_write_around (APIC_LVTTHMR, l & ~APIC_LVT_MASKED); + printk (KERN_INFO "CPU%d: Thermal monitoring enabled\n", cpu); + return; +} +#endif /* CONFIG_X86_MCE_P4THERMAL */ + + +/* P4/Xeon Extended MCE MSR retrieval, return 0 if unsupported */ +static inline int intel_get_extended_msrs(struct intel_mce_extended_msrs *r) +{ + u32 h; + + if (mce_num_extended_msrs == 0) + goto done; + + rdmsr (MSR_IA32_MCG_EAX, r->eax, h); + rdmsr (MSR_IA32_MCG_EBX, r->ebx, h); + rdmsr (MSR_IA32_MCG_ECX, r->ecx, h); + rdmsr (MSR_IA32_MCG_EDX, r->edx, h); + rdmsr (MSR_IA32_MCG_ESI, r->esi, h); + rdmsr (MSR_IA32_MCG_EDI, r->edi, h); + rdmsr (MSR_IA32_MCG_EBP, r->ebp, h); + rdmsr (MSR_IA32_MCG_ESP, r->esp, h); + rdmsr (MSR_IA32_MCG_EFLAGS, r->eflags, h); + rdmsr (MSR_IA32_MCG_EIP, r->eip, h); + + /* can we rely on kmalloc to do a dynamic + * allocation for the reserved registers? + */ +done: + return mce_num_extended_msrs; +} + +static fastcall void intel_machine_check(struct pt_regs * regs, long error_code) +{ + int recover=1; + u32 alow, ahigh, high, low; + u32 mcgstl, mcgsth; + int i; + struct intel_mce_extended_msrs dbg; + + rdmsr (MSR_IA32_MCG_STATUS, mcgstl, mcgsth); + if (mcgstl & (1<<0)) /* Recoverable ? */ + recover=0; + + printk (KERN_EMERG "CPU %d: Machine Check Exception: %08x%08x\n", + smp_processor_id(), mcgsth, mcgstl); + + if (intel_get_extended_msrs(&dbg)) { + printk (KERN_DEBUG "CPU %d: EIP: %08x EFLAGS: %08x\n", + smp_processor_id(), dbg.eip, dbg.eflags); + printk (KERN_DEBUG "\teax: %08x ebx: %08x ecx: %08x edx: %08x\n", + dbg.eax, dbg.ebx, dbg.ecx, dbg.edx); + printk (KERN_DEBUG "\tesi: %08x edi: %08x ebp: %08x esp: %08x\n", + dbg.esi, dbg.edi, dbg.ebp, dbg.esp); + } + + for (i=0; i<nr_mce_banks; i++) { + rdmsr (MSR_IA32_MC0_STATUS+i*4,low, high); + if (high & (1<<31)) { + if (high & (1<<29)) + recover |= 1; + if (high & (1<<25)) + recover |= 2; + printk (KERN_EMERG "Bank %d: %08x%08x", i, high, low); + high &= ~(1<<31); + if (high & (1<<27)) { + rdmsr (MSR_IA32_MC0_MISC+i*4, alow, ahigh); + printk ("[%08x%08x]", ahigh, alow); + } + if (high & (1<<26)) { + rdmsr (MSR_IA32_MC0_ADDR+i*4, alow, ahigh); + printk (" at %08x%08x", ahigh, alow); + } + printk ("\n"); + } + } + + if (recover & 2) + panic ("CPU context corrupt"); + if (recover & 1) + panic ("Unable to continue"); + + printk(KERN_EMERG "Attempting to continue.\n"); + /* + * Do not clear the MSR_IA32_MCi_STATUS if the error is not + * recoverable/continuable.This will allow BIOS to look at the MSRs + * for errors if the OS could not log the error. + */ + for (i=0; i<nr_mce_banks; i++) { + u32 msr; + msr = MSR_IA32_MC0_STATUS+i*4; + rdmsr (msr, low, high); + if (high&(1<<31)) { + /* Clear it */ + wrmsr(msr, 0UL, 0UL); + /* Serialize */ + wmb(); + add_taint(TAINT_MACHINE_CHECK); + } + } + mcgstl &= ~(1<<2); + wrmsr (MSR_IA32_MCG_STATUS,mcgstl, mcgsth); +} + + +void __init intel_p4_mcheck_init(struct cpuinfo_x86 *c) +{ + u32 l, h; + int i; + + machine_check_vector = intel_machine_check; + wmb(); + + printk (KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr (MSR_IA32_MCG_CAP, l, h); + if (l & (1<<8)) /* Control register present ? */ + wrmsr (MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); + nr_mce_banks = l & 0xff; + + for (i=0; i<nr_mce_banks; i++) { + wrmsr (MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); + wrmsr (MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); + } + + set_in_cr4 (X86_CR4_MCE); + printk (KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", + smp_processor_id()); + + /* Check for P4/Xeon extended MCE MSRs */ + rdmsr (MSR_IA32_MCG_CAP, l, h); + if (l & (1<<9)) {/* MCG_EXT_P */ + mce_num_extended_msrs = (l >> 16) & 0xff; + printk (KERN_INFO "CPU%d: Intel P4/Xeon Extended MCE MSRs (%d)" + " available\n", + smp_processor_id(), mce_num_extended_msrs); + +#ifdef CONFIG_X86_MCE_P4THERMAL + /* Check for P4/Xeon Thermal monitor */ + intel_init_thermal(c); +#endif + } +} diff --git a/arch/i386/kernel/cpu/mcheck/p5.c b/arch/i386/kernel/cpu/mcheck/p5.c new file mode 100644 index 00000000000..c45a1b485c8 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/p5.c @@ -0,0 +1,54 @@ +/* + * P5 specific Machine Check Exception Reporting + * (C) Copyright 2002 Alan Cox <alan@redhat.com> + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/smp.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> + +#include "mce.h" + +/* Machine check handler for Pentium class Intel */ +static fastcall void pentium_machine_check(struct pt_regs * regs, long error_code) +{ + u32 loaddr, hi, lotype; + rdmsr(MSR_IA32_P5_MC_ADDR, loaddr, hi); + rdmsr(MSR_IA32_P5_MC_TYPE, lotype, hi); + printk(KERN_EMERG "CPU#%d: Machine Check Exception: 0x%8X (type 0x%8X).\n", smp_processor_id(), loaddr, lotype); + if(lotype&(1<<5)) + printk(KERN_EMERG "CPU#%d: Possible thermal failure (CPU on fire ?).\n", smp_processor_id()); + add_taint(TAINT_MACHINE_CHECK); +} + +/* Set up machine check reporting for processors with Intel style MCE */ +void __init intel_p5_mcheck_init(struct cpuinfo_x86 *c) +{ + u32 l, h; + + /*Check for MCE support */ + if( !cpu_has(c, X86_FEATURE_MCE) ) + return; + + /* Default P5 to off as its often misconnected */ + if(mce_disabled != -1) + return; + machine_check_vector = pentium_machine_check; + wmb(); + + /* Read registers before enabling */ + rdmsr(MSR_IA32_P5_MC_ADDR, l, h); + rdmsr(MSR_IA32_P5_MC_TYPE, l, h); + printk(KERN_INFO "Intel old style machine check architecture supported.\n"); + + /* Enable MCE */ + set_in_cr4(X86_CR4_MCE); + printk(KERN_INFO "Intel old style machine check reporting enabled on CPU#%d.\n", smp_processor_id()); +} diff --git a/arch/i386/kernel/cpu/mcheck/p6.c b/arch/i386/kernel/cpu/mcheck/p6.c new file mode 100644 index 00000000000..46640f8c249 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/p6.c @@ -0,0 +1,115 @@ +/* + * P6 specific Machine Check Exception Reporting + * (C) Copyright 2002 Alan Cox <alan@redhat.com> + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/smp.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> + +#include "mce.h" + +/* Machine Check Handler For PII/PIII */ +static fastcall void intel_machine_check(struct pt_regs * regs, long error_code) +{ + int recover=1; + u32 alow, ahigh, high, low; + u32 mcgstl, mcgsth; + int i; + + rdmsr (MSR_IA32_MCG_STATUS, mcgstl, mcgsth); + if (mcgstl & (1<<0)) /* Recoverable ? */ + recover=0; + + printk (KERN_EMERG "CPU %d: Machine Check Exception: %08x%08x\n", + smp_processor_id(), mcgsth, mcgstl); + + for (i=0; i<nr_mce_banks; i++) { + rdmsr (MSR_IA32_MC0_STATUS+i*4,low, high); + if (high & (1<<31)) { + if (high & (1<<29)) + recover |= 1; + if (high & (1<<25)) + recover |= 2; + printk (KERN_EMERG "Bank %d: %08x%08x", i, high, low); + high &= ~(1<<31); + if (high & (1<<27)) { + rdmsr (MSR_IA32_MC0_MISC+i*4, alow, ahigh); + printk ("[%08x%08x]", ahigh, alow); + } + if (high & (1<<26)) { + rdmsr (MSR_IA32_MC0_ADDR+i*4, alow, ahigh); + printk (" at %08x%08x", ahigh, alow); + } + printk ("\n"); + } + } + + if (recover & 2) + panic ("CPU context corrupt"); + if (recover & 1) + panic ("Unable to continue"); + + printk (KERN_EMERG "Attempting to continue.\n"); + /* + * Do not clear the MSR_IA32_MCi_STATUS if the error is not + * recoverable/continuable.This will allow BIOS to look at the MSRs + * for errors if the OS could not log the error. + */ + for (i=0; i<nr_mce_banks; i++) { + unsigned int msr; + msr = MSR_IA32_MC0_STATUS+i*4; + rdmsr (msr,low, high); + if (high & (1<<31)) { + /* Clear it */ + wrmsr (msr, 0UL, 0UL); + /* Serialize */ + wmb(); + add_taint(TAINT_MACHINE_CHECK); + } + } + mcgstl &= ~(1<<2); + wrmsr (MSR_IA32_MCG_STATUS,mcgstl, mcgsth); +} + +/* Set up machine check reporting for processors with Intel style MCE */ +void __init intel_p6_mcheck_init(struct cpuinfo_x86 *c) +{ + u32 l, h; + int i; + + /* Check for MCE support */ + if (!cpu_has(c, X86_FEATURE_MCE)) + return; + + /* Check for PPro style MCA */ + if (!cpu_has(c, X86_FEATURE_MCA)) + return; + + /* Ok machine check is available */ + machine_check_vector = intel_machine_check; + wmb(); + + printk (KERN_INFO "Intel machine check architecture supported.\n"); + rdmsr (MSR_IA32_MCG_CAP, l, h); + if (l & (1<<8)) /* Control register present ? */ + wrmsr(MSR_IA32_MCG_CTL, 0xffffffff, 0xffffffff); + nr_mce_banks = l & 0xff; + + /* Don't enable bank 0 on intel P6 cores, it goes bang quickly. */ + for (i=1; i<nr_mce_banks; i++) { + wrmsr (MSR_IA32_MC0_CTL+4*i, 0xffffffff, 0xffffffff); + wrmsr (MSR_IA32_MC0_STATUS+4*i, 0x0, 0x0); + } + + set_in_cr4 (X86_CR4_MCE); + printk (KERN_INFO "Intel machine check reporting enabled on CPU#%d.\n", + smp_processor_id()); +} diff --git a/arch/i386/kernel/cpu/mcheck/winchip.c b/arch/i386/kernel/cpu/mcheck/winchip.c new file mode 100644 index 00000000000..753fa7acb98 --- /dev/null +++ b/arch/i386/kernel/cpu/mcheck/winchip.c @@ -0,0 +1,37 @@ +/* + * IDT Winchip specific Machine Check Exception Reporting + * (C) Copyright 2002 Alan Cox <alan@redhat.com> + */ + +#include <linux/init.h> +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/irq.h> +#include <linux/interrupt.h> + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/msr.h> + +#include "mce.h" + +/* Machine check handler for WinChip C6 */ +static fastcall void winchip_machine_check(struct pt_regs * regs, long error_code) +{ + printk(KERN_EMERG "CPU0: Machine Check Exception.\n"); + add_taint(TAINT_MACHINE_CHECK); +} + +/* Set up machine check reporting on the Winchip C6 series */ +void __init winchip_mcheck_init(struct cpuinfo_x86 *c) +{ + u32 lo, hi; + machine_check_vector = winchip_machine_check; + wmb(); + rdmsr(MSR_IDT_FCR1, lo, hi); + lo|= (1<<2); /* Enable EIERRINT (int 18 MCE) */ + lo&= ~(1<<4); /* Enable MCE */ + wrmsr(MSR_IDT_FCR1, lo, hi); + set_in_cr4(X86_CR4_MCE); + printk(KERN_INFO "Winchip machine check reporting enabled on CPU#0.\n"); +} diff --git a/arch/i386/kernel/cpu/mtrr/Makefile b/arch/i386/kernel/cpu/mtrr/Makefile new file mode 100644 index 00000000000..a25b701ab84 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/Makefile @@ -0,0 +1,5 @@ +obj-y := main.o if.o generic.o state.o +obj-y += amd.o +obj-y += cyrix.o +obj-y += centaur.o + diff --git a/arch/i386/kernel/cpu/mtrr/amd.c b/arch/i386/kernel/cpu/mtrr/amd.c new file mode 100644 index 00000000000..1a1e04b6fd0 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/amd.c @@ -0,0 +1,121 @@ +#include <linux/init.h> +#include <linux/mm.h> +#include <asm/mtrr.h> +#include <asm/msr.h> + +#include "mtrr.h" + +static void +amd_get_mtrr(unsigned int reg, unsigned long *base, + unsigned int *size, mtrr_type * type) +{ + unsigned long low, high; + + rdmsr(MSR_K6_UWCCR, low, high); + /* Upper dword is region 1, lower is region 0 */ + if (reg == 1) + low = high; + /* The base masks off on the right alignment */ + *base = (low & 0xFFFE0000) >> PAGE_SHIFT; + *type = 0; + if (low & 1) + *type = MTRR_TYPE_UNCACHABLE; + if (low & 2) + *type = MTRR_TYPE_WRCOMB; + if (!(low & 3)) { + *size = 0; + return; + } + /* + * This needs a little explaining. The size is stored as an + * inverted mask of bits of 128K granularity 15 bits long offset + * 2 bits + * + * So to get a size we do invert the mask and add 1 to the lowest + * mask bit (4 as its 2 bits in). This gives us a size we then shift + * to turn into 128K blocks + * + * eg 111 1111 1111 1100 is 512K + * + * invert 000 0000 0000 0011 + * +1 000 0000 0000 0100 + * *128K ... + */ + low = (~low) & 0x1FFFC; + *size = (low + 4) << (15 - PAGE_SHIFT); + return; +} + +static void amd_set_mtrr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type) +/* [SUMMARY] Set variable MTRR register on the local CPU. + <reg> The register to set. + <base> The base address of the region. + <size> The size of the region. If this is 0 the region is disabled. + <type> The type of the region. + <do_safe> If TRUE, do the change safely. If FALSE, safety measures should + be done externally. + [RETURNS] Nothing. +*/ +{ + u32 regs[2]; + + /* + * Low is MTRR0 , High MTRR 1 + */ + rdmsr(MSR_K6_UWCCR, regs[0], regs[1]); + /* + * Blank to disable + */ + if (size == 0) + regs[reg] = 0; + else + /* Set the register to the base, the type (off by one) and an + inverted bitmask of the size The size is the only odd + bit. We are fed say 512K We invert this and we get 111 1111 + 1111 1011 but if you subtract one and invert you get the + desired 111 1111 1111 1100 mask + + But ~(x - 1) == ~x + 1 == -x. Two's complement rocks! */ + regs[reg] = (-size >> (15 - PAGE_SHIFT) & 0x0001FFFC) + | (base << PAGE_SHIFT) | (type + 1); + + /* + * The writeback rule is quite specific. See the manual. Its + * disable local interrupts, write back the cache, set the mtrr + */ + wbinvd(); + wrmsr(MSR_K6_UWCCR, regs[0], regs[1]); +} + +static int amd_validate_add_page(unsigned long base, unsigned long size, unsigned int type) +{ + /* Apply the K6 block alignment and size rules + In order + o Uncached or gathering only + o 128K or bigger block + o Power of 2 block + o base suitably aligned to the power + */ + if (type > MTRR_TYPE_WRCOMB || size < (1 << (17 - PAGE_SHIFT)) + || (size & ~(size - 1)) - size || (base & (size - 1))) + return -EINVAL; + return 0; +} + +static struct mtrr_ops amd_mtrr_ops = { + .vendor = X86_VENDOR_AMD, + .set = amd_set_mtrr, + .get = amd_get_mtrr, + .get_free_region = generic_get_free_region, + .validate_add_page = amd_validate_add_page, + .have_wrcomb = positive_have_wrcomb, +}; + +int __init amd_init_mtrr(void) +{ + set_mtrr_ops(&amd_mtrr_ops); + return 0; +} + +//arch_initcall(amd_mtrr_init); diff --git a/arch/i386/kernel/cpu/mtrr/centaur.c b/arch/i386/kernel/cpu/mtrr/centaur.c new file mode 100644 index 00000000000..33f00ac314e --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/centaur.c @@ -0,0 +1,223 @@ +#include <linux/init.h> +#include <linux/mm.h> +#include <asm/mtrr.h> +#include <asm/msr.h> +#include "mtrr.h" + +static struct { + unsigned long high; + unsigned long low; +} centaur_mcr[8]; + +static u8 centaur_mcr_reserved; +static u8 centaur_mcr_type; /* 0 for winchip, 1 for winchip2 */ + +/* + * Report boot time MCR setups + */ + +static int +centaur_get_free_region(unsigned long base, unsigned long size) +/* [SUMMARY] Get a free MTRR. + <base> The starting (base) address of the region. + <size> The size (in bytes) of the region. + [RETURNS] The index of the region on success, else -1 on error. +*/ +{ + int i, max; + mtrr_type ltype; + unsigned long lbase; + unsigned int lsize; + + max = num_var_ranges; + for (i = 0; i < max; ++i) { + if (centaur_mcr_reserved & (1 << i)) + continue; + mtrr_if->get(i, &lbase, &lsize, <ype); + if (lsize == 0) + return i; + } + return -ENOSPC; +} + +void +mtrr_centaur_report_mcr(int mcr, u32 lo, u32 hi) +{ + centaur_mcr[mcr].low = lo; + centaur_mcr[mcr].high = hi; +} + +static void +centaur_get_mcr(unsigned int reg, unsigned long *base, + unsigned int *size, mtrr_type * type) +{ + *base = centaur_mcr[reg].high >> PAGE_SHIFT; + *size = -(centaur_mcr[reg].low & 0xfffff000) >> PAGE_SHIFT; + *type = MTRR_TYPE_WRCOMB; /* If it is there, it is write-combining */ + if (centaur_mcr_type == 1 && ((centaur_mcr[reg].low & 31) & 2)) + *type = MTRR_TYPE_UNCACHABLE; + if (centaur_mcr_type == 1 && (centaur_mcr[reg].low & 31) == 25) + *type = MTRR_TYPE_WRBACK; + if (centaur_mcr_type == 0 && (centaur_mcr[reg].low & 31) == 31) + *type = MTRR_TYPE_WRBACK; + +} + +static void centaur_set_mcr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type) +{ + unsigned long low, high; + + if (size == 0) { + /* Disable */ + high = low = 0; + } else { + high = base << PAGE_SHIFT; + if (centaur_mcr_type == 0) + low = -size << PAGE_SHIFT | 0x1f; /* only support write-combining... */ + else { + if (type == MTRR_TYPE_UNCACHABLE) + low = -size << PAGE_SHIFT | 0x02; /* NC */ + else + low = -size << PAGE_SHIFT | 0x09; /* WWO,WC */ + } + } + centaur_mcr[reg].high = high; + centaur_mcr[reg].low = low; + wrmsr(MSR_IDT_MCR0 + reg, low, high); +} + +#if 0 +/* + * Initialise the later (saner) Winchip MCR variant. In this version + * the BIOS can pass us the registers it has used (but not their values) + * and the control register is read/write + */ + +static void __init +centaur_mcr1_init(void) +{ + unsigned i; + u32 lo, hi; + + /* Unfortunately, MCR's are read-only, so there is no way to + * find out what the bios might have done. + */ + + rdmsr(MSR_IDT_MCR_CTRL, lo, hi); + if (((lo >> 17) & 7) == 1) { /* Type 1 Winchip2 MCR */ + lo &= ~0x1C0; /* clear key */ + lo |= 0x040; /* set key to 1 */ + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); /* unlock MCR */ + } + + centaur_mcr_type = 1; + + /* + * Clear any unconfigured MCR's. + */ + + for (i = 0; i < 8; ++i) { + if (centaur_mcr[i].high == 0 && centaur_mcr[i].low == 0) { + if (!(lo & (1 << (9 + i)))) + wrmsr(MSR_IDT_MCR0 + i, 0, 0); + else + /* + * If the BIOS set up an MCR we cannot see it + * but we don't wish to obliterate it + */ + centaur_mcr_reserved |= (1 << i); + } + } + /* + * Throw the main write-combining switch... + * However if OOSTORE is enabled then people have already done far + * cleverer things and we should behave. + */ + + lo |= 15; /* Write combine enables */ + wrmsr(MSR_IDT_MCR_CTRL, lo, hi); +} + +/* + * Initialise the original winchip with read only MCR registers + * no used bitmask for the BIOS to pass on and write only control + */ + +static void __init +centaur_mcr0_init(void) +{ + unsigned i; + + /* Unfortunately, MCR's are read-only, so there is no way to + * find out what the bios might have done. + */ + + /* Clear any unconfigured MCR's. + * This way we are sure that the centaur_mcr array contains the actual + * values. The disadvantage is that any BIOS tweaks are thus undone. + * + */ + for (i = 0; i < 8; ++i) { + if (centaur_mcr[i].high == 0 && centaur_mcr[i].low == 0) + wrmsr(MSR_IDT_MCR0 + i, 0, 0); + } + + wrmsr(MSR_IDT_MCR_CTRL, 0x01F0001F, 0); /* Write only */ +} + +/* + * Initialise Winchip series MCR registers + */ + +static void __init +centaur_mcr_init(void) +{ + struct set_mtrr_context ctxt; + + set_mtrr_prepare_save(&ctxt); + set_mtrr_cache_disable(&ctxt); + + if (boot_cpu_data.x86_model == 4) + centaur_mcr0_init(); + else if (boot_cpu_data.x86_model == 8 || boot_cpu_data.x86_model == 9) + centaur_mcr1_init(); + + set_mtrr_done(&ctxt); +} +#endif + +static int centaur_validate_add_page(unsigned long base, + unsigned long size, unsigned int type) +{ + /* + * FIXME: Winchip2 supports uncached + */ + if (type != MTRR_TYPE_WRCOMB && + (centaur_mcr_type == 0 || type != MTRR_TYPE_UNCACHABLE)) { + printk(KERN_WARNING + "mtrr: only write-combining%s supported\n", + centaur_mcr_type ? " and uncacheable are" + : " is"); + return -EINVAL; + } + return 0; +} + +static struct mtrr_ops centaur_mtrr_ops = { + .vendor = X86_VENDOR_CENTAUR, +// .init = centaur_mcr_init, + .set = centaur_set_mcr, + .get = centaur_get_mcr, + .get_free_region = centaur_get_free_region, + .validate_add_page = centaur_validate_add_page, + .have_wrcomb = positive_have_wrcomb, +}; + +int __init centaur_init_mtrr(void) +{ + set_mtrr_ops(¢aur_mtrr_ops); + return 0; +} + +//arch_initcall(centaur_init_mtrr); diff --git a/arch/i386/kernel/cpu/mtrr/changelog b/arch/i386/kernel/cpu/mtrr/changelog new file mode 100644 index 00000000000..af136853595 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/changelog @@ -0,0 +1,229 @@ + ChangeLog + + Prehistory Martin Tischhäuser <martin@ikcbarka.fzk.de> + Initial register-setting code (from proform-1.0). + 19971216 Richard Gooch <rgooch@atnf.csiro.au> + Original version for /proc/mtrr interface, SMP-safe. + v1.0 + 19971217 Richard Gooch <rgooch@atnf.csiro.au> + Bug fix for ioctls()'s. + Added sample code in Documentation/mtrr.txt + v1.1 + 19971218 Richard Gooch <rgooch@atnf.csiro.au> + Disallow overlapping regions. + 19971219 Jens Maurer <jmaurer@menuett.rhein-main.de> + Register-setting fixups. + v1.2 + 19971222 Richard Gooch <rgooch@atnf.csiro.au> + Fixups for kernel 2.1.75. + v1.3 + 19971229 David Wragg <dpw@doc.ic.ac.uk> + Register-setting fixups and conformity with Intel conventions. + 19971229 Richard Gooch <rgooch@atnf.csiro.au> + Cosmetic changes and wrote this ChangeLog ;-) + 19980106 Richard Gooch <rgooch@atnf.csiro.au> + Fixups for kernel 2.1.78. + v1.4 + 19980119 David Wragg <dpw@doc.ic.ac.uk> + Included passive-release enable code (elsewhere in PCI setup). + v1.5 + 19980131 Richard Gooch <rgooch@atnf.csiro.au> + Replaced global kernel lock with private spinlock. + v1.6 + 19980201 Richard Gooch <rgooch@atnf.csiro.au> + Added wait for other CPUs to complete changes. + v1.7 + 19980202 Richard Gooch <rgooch@atnf.csiro.au> + Bug fix in definition of <set_mtrr> for UP. + v1.8 + 19980319 Richard Gooch <rgooch@atnf.csiro.au> + Fixups for kernel 2.1.90. + 19980323 Richard Gooch <rgooch@atnf.csiro.au> + Move SMP BIOS fixup before secondary CPUs call <calibrate_delay> + v1.9 + 19980325 Richard Gooch <rgooch@atnf.csiro.au> + Fixed test for overlapping regions: confused by adjacent regions + 19980326 Richard Gooch <rgooch@atnf.csiro.au> + Added wbinvd in <set_mtrr_prepare>. + 19980401 Richard Gooch <rgooch@atnf.csiro.au> + Bug fix for non-SMP compilation. + 19980418 David Wragg <dpw@doc.ic.ac.uk> + Fixed-MTRR synchronisation for SMP and use atomic operations + instead of spinlocks. + 19980418 Richard Gooch <rgooch@atnf.csiro.au> + Differentiate different MTRR register classes for BIOS fixup. + v1.10 + 19980419 David Wragg <dpw@doc.ic.ac.uk> + Bug fix in variable MTRR synchronisation. + v1.11 + 19980419 Richard Gooch <rgooch@atnf.csiro.au> + Fixups for kernel 2.1.97. + v1.12 + 19980421 Richard Gooch <rgooch@atnf.csiro.au> + Safer synchronisation across CPUs when changing MTRRs. + v1.13 + 19980423 Richard Gooch <rgooch@atnf.csiro.au> + Bugfix for SMP systems without MTRR support. + v1.14 + 19980427 Richard Gooch <rgooch@atnf.csiro.au> + Trap calls to <mtrr_add> and <mtrr_del> on non-MTRR machines. + v1.15 + 19980427 Richard Gooch <rgooch@atnf.csiro.au> + Use atomic bitops for setting SMP change mask. + v1.16 + 19980428 Richard Gooch <rgooch@atnf.csiro.au> + Removed spurious diagnostic message. + v1.17 + 19980429 Richard Gooch <rgooch@atnf.csiro.au> + Moved register-setting macros into this file. + Moved setup code from init/main.c to i386-specific areas. + v1.18 + 19980502 Richard Gooch <rgooch@atnf.csiro.au> + Moved MTRR detection outside conditionals in <mtrr_init>. + v1.19 + 19980502 Richard Gooch <rgooch@atnf.csiro.au> + Documentation improvement: mention Pentium II and AGP. + v1.20 + 19980521 Richard Gooch <rgooch@atnf.csiro.au> + Only manipulate interrupt enable flag on local CPU. + Allow enclosed uncachable regions. + v1.21 + 19980611 Richard Gooch <rgooch@atnf.csiro.au> + Always define <main_lock>. + v1.22 + 19980901 Richard Gooch <rgooch@atnf.csiro.au> + Removed module support in order to tidy up code. + Added sanity check for <mtrr_add>/<mtrr_del> before <mtrr_init>. + Created addition queue for prior to SMP commence. + v1.23 + 19980902 Richard Gooch <rgooch@atnf.csiro.au> + Ported patch to kernel 2.1.120-pre3. + v1.24 + 19980910 Richard Gooch <rgooch@atnf.csiro.au> + Removed sanity checks and addition queue: Linus prefers an OOPS. + v1.25 + 19981001 Richard Gooch <rgooch@atnf.csiro.au> + Fixed harmless compiler warning in include/asm-i386/mtrr.h + Fixed version numbering and history for v1.23 -> v1.24. + v1.26 + 19990118 Richard Gooch <rgooch@atnf.csiro.au> + Added devfs support. + v1.27 + 19990123 Richard Gooch <rgooch@atnf.csiro.au> + Changed locking to spin with reschedule. + Made use of new <smp_call_function>. + v1.28 + 19990201 Zoltán Böszörményi <zboszor@mail.externet.hu> + Extended the driver to be able to use Cyrix style ARRs. + 19990204 Richard Gooch <rgooch@atnf.csiro.au> + Restructured Cyrix support. + v1.29 + 19990204 Zoltán Böszörményi <zboszor@mail.externet.hu> + Refined ARR support: enable MAPEN in set_mtrr_prepare() + and disable MAPEN in set_mtrr_done(). + 19990205 Richard Gooch <rgooch@atnf.csiro.au> + Minor cleanups. + v1.30 + 19990208 Zoltán Böszörményi <zboszor@mail.externet.hu> + Protect plain 6x86s (and other processors without the + Page Global Enable feature) against accessing CR4 in + set_mtrr_prepare() and set_mtrr_done(). + 19990210 Richard Gooch <rgooch@atnf.csiro.au> + Turned <set_mtrr_up> and <get_mtrr> into function pointers. + v1.31 + 19990212 Zoltán Böszörményi <zboszor@mail.externet.hu> + Major rewrite of cyrix_arr_init(): do not touch ARRs, + leave them as the BIOS have set them up. + Enable usage of all 8 ARRs. + Avoid multiplications by 3 everywhere and other + code clean ups/speed ups. + 19990213 Zoltán Böszörményi <zboszor@mail.externet.hu> + Set up other Cyrix processors identical to the boot cpu. + Since Cyrix don't support Intel APIC, this is l'art pour l'art. + Weigh ARRs by size: + If size <= 32M is given, set up ARR# we were given. + If size > 32M is given, set up ARR7 only if it is free, + fail otherwise. + 19990214 Zoltán Böszörményi <zboszor@mail.externet.hu> + Also check for size >= 256K if we are to set up ARR7, + mtrr_add() returns the value it gets from set_mtrr() + 19990218 Zoltán Böszörményi <zboszor@mail.externet.hu> + Remove Cyrix "coma bug" workaround from here. + Moved to linux/arch/i386/kernel/setup.c and + linux/include/asm-i386/bugs.h + 19990228 Richard Gooch <rgooch@atnf.csiro.au> + Added MTRRIOC_KILL_ENTRY ioctl(2) + Trap for counter underflow in <mtrr_file_del>. + Trap for 4 MiB aligned regions for PPro, stepping <= 7. + 19990301 Richard Gooch <rgooch@atnf.csiro.au> + Created <get_free_region> hook. + 19990305 Richard Gooch <rgooch@atnf.csiro.au> + Temporarily disable AMD support now MTRR capability flag is set. + v1.32 + 19990308 Zoltán Böszörményi <zboszor@mail.externet.hu> + Adjust my changes (19990212-19990218) to Richard Gooch's + latest changes. (19990228-19990305) + v1.33 + 19990309 Richard Gooch <rgooch@atnf.csiro.au> + Fixed typo in <printk> message. + 19990310 Richard Gooch <rgooch@atnf.csiro.au> + Support K6-II/III based on Alan Cox's <alan@redhat.com> patches. + v1.34 + 19990511 Bart Hartgers <bart@etpmod.phys.tue.nl> + Support Centaur C6 MCR's. + 19990512 Richard Gooch <rgooch@atnf.csiro.au> + Minor cleanups. + v1.35 + 19990707 Zoltán Böszörményi <zboszor@mail.externet.hu> + Check whether ARR3 is protected in cyrix_get_free_region() + and mtrr_del(). The code won't attempt to delete or change it + from now on if the BIOS protected ARR3. It silently skips ARR3 + in cyrix_get_free_region() or returns with an error code from + mtrr_del(). + 19990711 Zoltán Böszörményi <zboszor@mail.externet.hu> + Reset some bits in the CCRs in cyrix_arr_init() to disable SMM + if ARR3 isn't protected. This is needed because if SMM is active + and ARR3 isn't protected then deleting and setting ARR3 again + may lock up the processor. With SMM entirely disabled, it does + not happen. + 19990812 Zoltán Böszörményi <zboszor@mail.externet.hu> + Rearrange switch() statements so the driver accomodates to + the fact that the AMD Athlon handles its MTRRs the same way + as Intel does. + 19990814 Zoltán Böszörményi <zboszor@mail.externet.hu> + Double check for Intel in mtrr_add()'s big switch() because + that revision check is only valid for Intel CPUs. + 19990819 Alan Cox <alan@redhat.com> + Tested Zoltan's changes on a pre production Athlon - 100% + success. + 19991008 Manfred Spraul <manfreds@colorfullife.com> + replaced spin_lock_reschedule() with a normal semaphore. + v1.36 + 20000221 Richard Gooch <rgooch@atnf.csiro.au> + Compile fix if procfs and devfs not enabled. + Formatting changes. + v1.37 + 20001109 H. Peter Anvin <hpa@zytor.com> + Use the new centralized CPU feature detects. + + v1.38 + 20010309 Dave Jones <davej@suse.de> + Add support for Cyrix III. + + v1.39 + 20010312 Dave Jones <davej@suse.de> + Ugh, I broke AMD support. + Reworked fix by Troels Walsted Hansen <troels@thule.no> + + v1.40 + 20010327 Dave Jones <davej@suse.de> + Adapted Cyrix III support to include VIA C3. + + v2.0 + 20020306 Patrick Mochel <mochel@osdl.org> + Split mtrr.c -> mtrr/*.c + Converted to Linux Kernel Coding Style + Fixed several minor nits in form + Moved some SMP-only functions out, so they can be used + for power management in the future. + TODO: Fix user interface cruft. diff --git a/arch/i386/kernel/cpu/mtrr/cyrix.c b/arch/i386/kernel/cpu/mtrr/cyrix.c new file mode 100644 index 00000000000..933b0dd62f4 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/cyrix.c @@ -0,0 +1,364 @@ +#include <linux/init.h> +#include <linux/mm.h> +#include <asm/mtrr.h> +#include <asm/msr.h> +#include <asm/io.h> +#include "mtrr.h" + +int arr3_protected; + +static void +cyrix_get_arr(unsigned int reg, unsigned long *base, + unsigned int *size, mtrr_type * type) +{ + unsigned long flags; + unsigned char arr, ccr3, rcr, shift; + + arr = CX86_ARR_BASE + (reg << 1) + reg; /* avoid multiplication by 3 */ + + /* Save flags and disable interrupts */ + local_irq_save(flags); + + ccr3 = getCx86(CX86_CCR3); + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); /* enable MAPEN */ + ((unsigned char *) base)[3] = getCx86(arr); + ((unsigned char *) base)[2] = getCx86(arr + 1); + ((unsigned char *) base)[1] = getCx86(arr + 2); + rcr = getCx86(CX86_RCR_BASE + reg); + setCx86(CX86_CCR3, ccr3); /* disable MAPEN */ + + /* Enable interrupts if it was enabled previously */ + local_irq_restore(flags); + shift = ((unsigned char *) base)[1] & 0x0f; + *base >>= PAGE_SHIFT; + + /* Power of two, at least 4K on ARR0-ARR6, 256K on ARR7 + * Note: shift==0xf means 4G, this is unsupported. + */ + if (shift) + *size = (reg < 7 ? 0x1UL : 0x40UL) << (shift - 1); + else + *size = 0; + + /* Bit 0 is Cache Enable on ARR7, Cache Disable on ARR0-ARR6 */ + if (reg < 7) { + switch (rcr) { + case 1: + *type = MTRR_TYPE_UNCACHABLE; + break; + case 8: + *type = MTRR_TYPE_WRBACK; + break; + case 9: + *type = MTRR_TYPE_WRCOMB; + break; + case 24: + default: + *type = MTRR_TYPE_WRTHROUGH; + break; + } + } else { + switch (rcr) { + case 0: + *type = MTRR_TYPE_UNCACHABLE; + break; + case 8: + *type = MTRR_TYPE_WRCOMB; + break; + case 9: + *type = MTRR_TYPE_WRBACK; + break; + case 25: + default: + *type = MTRR_TYPE_WRTHROUGH; + break; + } + } +} + +static int +cyrix_get_free_region(unsigned long base, unsigned long size) +/* [SUMMARY] Get a free ARR. + <base> The starting (base) address of the region. + <size> The size (in bytes) of the region. + [RETURNS] The index of the region on success, else -1 on error. +*/ +{ + int i; + mtrr_type ltype; + unsigned long lbase; + unsigned int lsize; + + /* If we are to set up a region >32M then look at ARR7 immediately */ + if (size > 0x2000) { + cyrix_get_arr(7, &lbase, &lsize, <ype); + if (lsize == 0) + return 7; + /* Else try ARR0-ARR6 first */ + } else { + for (i = 0; i < 7; i++) { + cyrix_get_arr(i, &lbase, &lsize, <ype); + if ((i == 3) && arr3_protected) + continue; + if (lsize == 0) + return i; + } + /* ARR0-ARR6 isn't free, try ARR7 but its size must be at least 256K */ + cyrix_get_arr(i, &lbase, &lsize, <ype); + if ((lsize == 0) && (size >= 0x40)) + return i; + } + return -ENOSPC; +} + +static u32 cr4 = 0; +static u32 ccr3; + +static void prepare_set(void) +{ + u32 cr0; + + /* Save value of CR4 and clear Page Global Enable (bit 7) */ + if ( cpu_has_pge ) { + cr4 = read_cr4(); + write_cr4(cr4 & (unsigned char) ~(1 << 7)); + } + + /* Disable and flush caches. Note that wbinvd flushes the TLBs as + a side-effect */ + cr0 = read_cr0() | 0x40000000; + wbinvd(); + write_cr0(cr0); + wbinvd(); + + /* Cyrix ARRs - everything else were excluded at the top */ + ccr3 = getCx86(CX86_CCR3); + + /* Cyrix ARRs - everything else were excluded at the top */ + setCx86(CX86_CCR3, (ccr3 & 0x0f) | 0x10); + +} + +static void post_set(void) +{ + /* Flush caches and TLBs */ + wbinvd(); + + /* Cyrix ARRs - everything else was excluded at the top */ + setCx86(CX86_CCR3, ccr3); + + /* Enable caches */ + write_cr0(read_cr0() & 0xbfffffff); + + /* Restore value of CR4 */ + if ( cpu_has_pge ) + write_cr4(cr4); +} + +static void cyrix_set_arr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type) +{ + unsigned char arr, arr_type, arr_size; + + arr = CX86_ARR_BASE + (reg << 1) + reg; /* avoid multiplication by 3 */ + + /* count down from 32M (ARR0-ARR6) or from 2G (ARR7) */ + if (reg >= 7) + size >>= 6; + + size &= 0x7fff; /* make sure arr_size <= 14 */ + for (arr_size = 0; size; arr_size++, size >>= 1) ; + + if (reg < 7) { + switch (type) { + case MTRR_TYPE_UNCACHABLE: + arr_type = 1; + break; + case MTRR_TYPE_WRCOMB: + arr_type = 9; + break; + case MTRR_TYPE_WRTHROUGH: + arr_type = 24; + break; + default: + arr_type = 8; + break; + } + } else { + switch (type) { + case MTRR_TYPE_UNCACHABLE: + arr_type = 0; + break; + case MTRR_TYPE_WRCOMB: + arr_type = 8; + break; + case MTRR_TYPE_WRTHROUGH: + arr_type = 25; + break; + default: + arr_type = 9; + break; + } + } + + prepare_set(); + + base <<= PAGE_SHIFT; + setCx86(arr, ((unsigned char *) &base)[3]); + setCx86(arr + 1, ((unsigned char *) &base)[2]); + setCx86(arr + 2, (((unsigned char *) &base)[1]) | arr_size); + setCx86(CX86_RCR_BASE + reg, arr_type); + + post_set(); +} + +typedef struct { + unsigned long base; + unsigned int size; + mtrr_type type; +} arr_state_t; + +static arr_state_t arr_state[8] __initdata = { + {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL}, + {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL}, {0UL, 0UL, 0UL} +}; + +static unsigned char ccr_state[7] __initdata = { 0, 0, 0, 0, 0, 0, 0 }; + +static void cyrix_set_all(void) +{ + int i; + + prepare_set(); + + /* the CCRs are not contiguous */ + for (i = 0; i < 4; i++) + setCx86(CX86_CCR0 + i, ccr_state[i]); + for (; i < 7; i++) + setCx86(CX86_CCR4 + i, ccr_state[i]); + for (i = 0; i < 8; i++) + cyrix_set_arr(i, arr_state[i].base, + arr_state[i].size, arr_state[i].type); + + post_set(); +} + +#if 0 +/* + * On Cyrix 6x86(MX) and M II the ARR3 is special: it has connection + * with the SMM (System Management Mode) mode. So we need the following: + * Check whether SMI_LOCK (CCR3 bit 0) is set + * if it is set, write a warning message: ARR3 cannot be changed! + * (it cannot be changed until the next processor reset) + * if it is reset, then we can change it, set all the needed bits: + * - disable access to SMM memory through ARR3 range (CCR1 bit 7 reset) + * - disable access to SMM memory (CCR1 bit 2 reset) + * - disable SMM mode (CCR1 bit 1 reset) + * - disable write protection of ARR3 (CCR6 bit 1 reset) + * - (maybe) disable ARR3 + * Just to be sure, we enable ARR usage by the processor (CCR5 bit 5 set) + */ +static void __init +cyrix_arr_init(void) +{ + struct set_mtrr_context ctxt; + unsigned char ccr[7]; + int ccrc[7] = { 0, 0, 0, 0, 0, 0, 0 }; +#ifdef CONFIG_SMP + int i; +#endif + + /* flush cache and enable MAPEN */ + set_mtrr_prepare_save(&ctxt); + set_mtrr_cache_disable(&ctxt); + + /* Save all CCRs locally */ + ccr[0] = getCx86(CX86_CCR0); + ccr[1] = getCx86(CX86_CCR1); + ccr[2] = getCx86(CX86_CCR2); + ccr[3] = ctxt.ccr3; + ccr[4] = getCx86(CX86_CCR4); + ccr[5] = getCx86(CX86_CCR5); + ccr[6] = getCx86(CX86_CCR6); + + if (ccr[3] & 1) { + ccrc[3] = 1; + arr3_protected = 1; + } else { + /* Disable SMM mode (bit 1), access to SMM memory (bit 2) and + * access to SMM memory through ARR3 (bit 7). + */ + if (ccr[1] & 0x80) { + ccr[1] &= 0x7f; + ccrc[1] |= 0x80; + } + if (ccr[1] & 0x04) { + ccr[1] &= 0xfb; + ccrc[1] |= 0x04; + } + if (ccr[1] & 0x02) { + ccr[1] &= 0xfd; + ccrc[1] |= 0x02; + } + arr3_protected = 0; + if (ccr[6] & 0x02) { + ccr[6] &= 0xfd; + ccrc[6] = 1; /* Disable write protection of ARR3 */ + setCx86(CX86_CCR6, ccr[6]); + } + /* Disable ARR3. This is safe now that we disabled SMM. */ + /* cyrix_set_arr_up (3, 0, 0, 0, FALSE); */ + } + /* If we changed CCR1 in memory, change it in the processor, too. */ + if (ccrc[1]) + setCx86(CX86_CCR1, ccr[1]); + + /* Enable ARR usage by the processor */ + if (!(ccr[5] & 0x20)) { + ccr[5] |= 0x20; + ccrc[5] = 1; + setCx86(CX86_CCR5, ccr[5]); + } +#ifdef CONFIG_SMP + for (i = 0; i < 7; i++) + ccr_state[i] = ccr[i]; + for (i = 0; i < 8; i++) + cyrix_get_arr(i, + &arr_state[i].base, &arr_state[i].size, + &arr_state[i].type); +#endif + + set_mtrr_done(&ctxt); /* flush cache and disable MAPEN */ + + if (ccrc[5]) + printk(KERN_INFO "mtrr: ARR usage was not enabled, enabled manually\n"); + if (ccrc[3]) + printk(KERN_INFO "mtrr: ARR3 cannot be changed\n"); +/* + if ( ccrc[1] & 0x80) printk ("mtrr: SMM memory access through ARR3 disabled\n"); + if ( ccrc[1] & 0x04) printk ("mtrr: SMM memory access disabled\n"); + if ( ccrc[1] & 0x02) printk ("mtrr: SMM mode disabled\n"); +*/ + if (ccrc[6]) + printk(KERN_INFO "mtrr: ARR3 was write protected, unprotected\n"); +} +#endif + +static struct mtrr_ops cyrix_mtrr_ops = { + .vendor = X86_VENDOR_CYRIX, +// .init = cyrix_arr_init, + .set_all = cyrix_set_all, + .set = cyrix_set_arr, + .get = cyrix_get_arr, + .get_free_region = cyrix_get_free_region, + .validate_add_page = generic_validate_add_page, + .have_wrcomb = positive_have_wrcomb, +}; + +int __init cyrix_init_mtrr(void) +{ + set_mtrr_ops(&cyrix_mtrr_ops); + return 0; +} + +//arch_initcall(cyrix_init_mtrr); diff --git a/arch/i386/kernel/cpu/mtrr/generic.c b/arch/i386/kernel/cpu/mtrr/generic.c new file mode 100644 index 00000000000..a4cce454d09 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/generic.c @@ -0,0 +1,417 @@ +/* This only handles 32bit MTRR on 32bit hosts. This is strictly wrong + because MTRRs can span upto 40 bits (36bits on most modern x86) */ +#include <linux/init.h> +#include <linux/slab.h> +#include <linux/mm.h> +#include <asm/io.h> +#include <asm/mtrr.h> +#include <asm/msr.h> +#include <asm/system.h> +#include <asm/cpufeature.h> +#include <asm/tlbflush.h> +#include "mtrr.h" + +struct mtrr_state { + struct mtrr_var_range *var_ranges; + mtrr_type fixed_ranges[NUM_FIXED_RANGES]; + unsigned char enabled; + mtrr_type def_type; +}; + +static unsigned long smp_changes_mask; +static struct mtrr_state mtrr_state = {}; + +/* Get the MSR pair relating to a var range */ +static void __init +get_mtrr_var_range(unsigned int index, struct mtrr_var_range *vr) +{ + rdmsr(MTRRphysBase_MSR(index), vr->base_lo, vr->base_hi); + rdmsr(MTRRphysMask_MSR(index), vr->mask_lo, vr->mask_hi); +} + +static void __init +get_fixed_ranges(mtrr_type * frs) +{ + unsigned int *p = (unsigned int *) frs; + int i; + + rdmsr(MTRRfix64K_00000_MSR, p[0], p[1]); + + for (i = 0; i < 2; i++) + rdmsr(MTRRfix16K_80000_MSR + i, p[2 + i * 2], p[3 + i * 2]); + for (i = 0; i < 8; i++) + rdmsr(MTRRfix4K_C0000_MSR + i, p[6 + i * 2], p[7 + i * 2]); +} + +/* Grab all of the MTRR state for this CPU into *state */ +void __init get_mtrr_state(void) +{ + unsigned int i; + struct mtrr_var_range *vrs; + unsigned lo, dummy; + + if (!mtrr_state.var_ranges) { + mtrr_state.var_ranges = kmalloc(num_var_ranges * sizeof (struct mtrr_var_range), + GFP_KERNEL); + if (!mtrr_state.var_ranges) + return; + } + vrs = mtrr_state.var_ranges; + + for (i = 0; i < num_var_ranges; i++) + get_mtrr_var_range(i, &vrs[i]); + get_fixed_ranges(mtrr_state.fixed_ranges); + + rdmsr(MTRRdefType_MSR, lo, dummy); + mtrr_state.def_type = (lo & 0xff); + mtrr_state.enabled = (lo & 0xc00) >> 10; +} + +/* Free resources associated with a struct mtrr_state */ +void __init finalize_mtrr_state(void) +{ + if (mtrr_state.var_ranges) + kfree(mtrr_state.var_ranges); + mtrr_state.var_ranges = NULL; +} + +/* Some BIOS's are fucked and don't set all MTRRs the same! */ +void __init mtrr_state_warn(void) +{ + unsigned long mask = smp_changes_mask; + + if (!mask) + return; + if (mask & MTRR_CHANGE_MASK_FIXED) + printk(KERN_WARNING "mtrr: your CPUs had inconsistent fixed MTRR settings\n"); + if (mask & MTRR_CHANGE_MASK_VARIABLE) + printk(KERN_WARNING "mtrr: your CPUs had inconsistent variable MTRR settings\n"); + if (mask & MTRR_CHANGE_MASK_DEFTYPE) + printk(KERN_WARNING "mtrr: your CPUs had inconsistent MTRRdefType settings\n"); + printk(KERN_INFO "mtrr: probably your BIOS does not setup all CPUs.\n"); + printk(KERN_INFO "mtrr: corrected configuration.\n"); +} + +/* Doesn't attempt to pass an error out to MTRR users + because it's quite complicated in some cases and probably not + worth it because the best error handling is to ignore it. */ +void mtrr_wrmsr(unsigned msr, unsigned a, unsigned b) +{ + if (wrmsr_safe(msr, a, b) < 0) + printk(KERN_ERR + "MTRR: CPU %u: Writing MSR %x to %x:%x failed\n", + smp_processor_id(), msr, a, b); +} + +int generic_get_free_region(unsigned long base, unsigned long size) +/* [SUMMARY] Get a free MTRR. + <base> The starting (base) address of the region. + <size> The size (in bytes) of the region. + [RETURNS] The index of the region on success, else -1 on error. +*/ +{ + int i, max; + mtrr_type ltype; + unsigned long lbase; + unsigned lsize; + + max = num_var_ranges; + for (i = 0; i < max; ++i) { + mtrr_if->get(i, &lbase, &lsize, <ype); + if (lsize == 0) + return i; + } + return -ENOSPC; +} + +void generic_get_mtrr(unsigned int reg, unsigned long *base, + unsigned int *size, mtrr_type * type) +{ + unsigned int mask_lo, mask_hi, base_lo, base_hi; + + rdmsr(MTRRphysMask_MSR(reg), mask_lo, mask_hi); + if ((mask_lo & 0x800) == 0) { + /* Invalid (i.e. free) range */ + *base = 0; + *size = 0; + *type = 0; + return; + } + + rdmsr(MTRRphysBase_MSR(reg), base_lo, base_hi); + + /* Work out the shifted address mask. */ + mask_lo = size_or_mask | mask_hi << (32 - PAGE_SHIFT) + | mask_lo >> PAGE_SHIFT; + + /* This works correctly if size is a power of two, i.e. a + contiguous range. */ + *size = -mask_lo; + *base = base_hi << (32 - PAGE_SHIFT) | base_lo >> PAGE_SHIFT; + *type = base_lo & 0xff; +} + +static int set_fixed_ranges(mtrr_type * frs) +{ + unsigned int *p = (unsigned int *) frs; + int changed = FALSE; + int i; + unsigned int lo, hi; + + rdmsr(MTRRfix64K_00000_MSR, lo, hi); + if (p[0] != lo || p[1] != hi) { + mtrr_wrmsr(MTRRfix64K_00000_MSR, p[0], p[1]); + changed = TRUE; + } + + for (i = 0; i < 2; i++) { + rdmsr(MTRRfix16K_80000_MSR + i, lo, hi); + if (p[2 + i * 2] != lo || p[3 + i * 2] != hi) { + mtrr_wrmsr(MTRRfix16K_80000_MSR + i, p[2 + i * 2], + p[3 + i * 2]); + changed = TRUE; + } + } + + for (i = 0; i < 8; i++) { + rdmsr(MTRRfix4K_C0000_MSR + i, lo, hi); + if (p[6 + i * 2] != lo || p[7 + i * 2] != hi) { + mtrr_wrmsr(MTRRfix4K_C0000_MSR + i, p[6 + i * 2], + p[7 + i * 2]); + changed = TRUE; + } + } + return changed; +} + +/* Set the MSR pair relating to a var range. Returns TRUE if + changes are made */ +static int set_mtrr_var_ranges(unsigned int index, struct mtrr_var_range *vr) +{ + unsigned int lo, hi; + int changed = FALSE; + + rdmsr(MTRRphysBase_MSR(index), lo, hi); + if ((vr->base_lo & 0xfffff0ffUL) != (lo & 0xfffff0ffUL) + || (vr->base_hi & 0xfUL) != (hi & 0xfUL)) { + mtrr_wrmsr(MTRRphysBase_MSR(index), vr->base_lo, vr->base_hi); + changed = TRUE; + } + + rdmsr(MTRRphysMask_MSR(index), lo, hi); + + if ((vr->mask_lo & 0xfffff800UL) != (lo & 0xfffff800UL) + || (vr->mask_hi & 0xfUL) != (hi & 0xfUL)) { + mtrr_wrmsr(MTRRphysMask_MSR(index), vr->mask_lo, vr->mask_hi); + changed = TRUE; + } + return changed; +} + +static unsigned long set_mtrr_state(u32 deftype_lo, u32 deftype_hi) +/* [SUMMARY] Set the MTRR state for this CPU. + <state> The MTRR state information to read. + <ctxt> Some relevant CPU context. + [NOTE] The CPU must already be in a safe state for MTRR changes. + [RETURNS] 0 if no changes made, else a mask indication what was changed. +*/ +{ + unsigned int i; + unsigned long change_mask = 0; + + for (i = 0; i < num_var_ranges; i++) + if (set_mtrr_var_ranges(i, &mtrr_state.var_ranges[i])) + change_mask |= MTRR_CHANGE_MASK_VARIABLE; + + if (set_fixed_ranges(mtrr_state.fixed_ranges)) + change_mask |= MTRR_CHANGE_MASK_FIXED; + + /* Set_mtrr_restore restores the old value of MTRRdefType, + so to set it we fiddle with the saved value */ + if ((deftype_lo & 0xff) != mtrr_state.def_type + || ((deftype_lo & 0xc00) >> 10) != mtrr_state.enabled) { + deftype_lo |= (mtrr_state.def_type | mtrr_state.enabled << 10); + change_mask |= MTRR_CHANGE_MASK_DEFTYPE; + } + + return change_mask; +} + + +static unsigned long cr4 = 0; +static u32 deftype_lo, deftype_hi; +static DEFINE_SPINLOCK(set_atomicity_lock); + +/* + * Since we are disabling the cache don't allow any interrupts - they + * would run extremely slow and would only increase the pain. The caller must + * ensure that local interrupts are disabled and are reenabled after post_set() + * has been called. + */ + +static void prepare_set(void) +{ + unsigned long cr0; + + /* Note that this is not ideal, since the cache is only flushed/disabled + for this CPU while the MTRRs are changed, but changing this requires + more invasive changes to the way the kernel boots */ + + spin_lock(&set_atomicity_lock); + + /* Enter the no-fill (CD=1, NW=0) cache mode and flush caches. */ + cr0 = read_cr0() | 0x40000000; /* set CD flag */ + write_cr0(cr0); + wbinvd(); + + /* Save value of CR4 and clear Page Global Enable (bit 7) */ + if ( cpu_has_pge ) { + cr4 = read_cr4(); + write_cr4(cr4 & ~X86_CR4_PGE); + } + + /* Flush all TLBs via a mov %cr3, %reg; mov %reg, %cr3 */ + __flush_tlb(); + + /* Save MTRR state */ + rdmsr(MTRRdefType_MSR, deftype_lo, deftype_hi); + + /* Disable MTRRs, and set the default type to uncached */ + mtrr_wrmsr(MTRRdefType_MSR, deftype_lo & 0xf300UL, deftype_hi); +} + +static void post_set(void) +{ + /* Flush TLBs (no need to flush caches - they are disabled) */ + __flush_tlb(); + + /* Intel (P6) standard MTRRs */ + mtrr_wrmsr(MTRRdefType_MSR, deftype_lo, deftype_hi); + + /* Enable caches */ + write_cr0(read_cr0() & 0xbfffffff); + + /* Restore value of CR4 */ + if ( cpu_has_pge ) + write_cr4(cr4); + spin_unlock(&set_atomicity_lock); +} + +static void generic_set_all(void) +{ + unsigned long mask, count; + unsigned long flags; + + local_irq_save(flags); + prepare_set(); + + /* Actually set the state */ + mask = set_mtrr_state(deftype_lo,deftype_hi); + + post_set(); + local_irq_restore(flags); + + /* Use the atomic bitops to update the global mask */ + for (count = 0; count < sizeof mask * 8; ++count) { + if (mask & 0x01) + set_bit(count, &smp_changes_mask); + mask >>= 1; + } + +} + +static void generic_set_mtrr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type) +/* [SUMMARY] Set variable MTRR register on the local CPU. + <reg> The register to set. + <base> The base address of the region. + <size> The size of the region. If this is 0 the region is disabled. + <type> The type of the region. + <do_safe> If TRUE, do the change safely. If FALSE, safety measures should + be done externally. + [RETURNS] Nothing. +*/ +{ + unsigned long flags; + + local_irq_save(flags); + prepare_set(); + + if (size == 0) { + /* The invalid bit is kept in the mask, so we simply clear the + relevant mask register to disable a range. */ + mtrr_wrmsr(MTRRphysMask_MSR(reg), 0, 0); + } else { + mtrr_wrmsr(MTRRphysBase_MSR(reg), base << PAGE_SHIFT | type, + (base & size_and_mask) >> (32 - PAGE_SHIFT)); + mtrr_wrmsr(MTRRphysMask_MSR(reg), -size << PAGE_SHIFT | 0x800, + (-size & size_and_mask) >> (32 - PAGE_SHIFT)); + } + + post_set(); + local_irq_restore(flags); +} + +int generic_validate_add_page(unsigned long base, unsigned long size, unsigned int type) +{ + unsigned long lbase, last; + + /* For Intel PPro stepping <= 7, must be 4 MiB aligned + and not touch 0x70000000->0x7003FFFF */ + if (is_cpu(INTEL) && boot_cpu_data.x86 == 6 && + boot_cpu_data.x86_model == 1 && + boot_cpu_data.x86_mask <= 7) { + if (base & ((1 << (22 - PAGE_SHIFT)) - 1)) { + printk(KERN_WARNING "mtrr: base(0x%lx000) is not 4 MiB aligned\n", base); + return -EINVAL; + } + if (!(base + size < 0x70000000 || base > 0x7003FFFF) && + (type == MTRR_TYPE_WRCOMB + || type == MTRR_TYPE_WRBACK)) { + printk(KERN_WARNING "mtrr: writable mtrr between 0x70000000 and 0x7003FFFF may hang the CPU.\n"); + return -EINVAL; + } + } + + if (base + size < 0x100) { + printk(KERN_WARNING "mtrr: cannot set region below 1 MiB (0x%lx000,0x%lx000)\n", + base, size); + return -EINVAL; + } + /* Check upper bits of base and last are equal and lower bits are 0 + for base and 1 for last */ + last = base + size - 1; + for (lbase = base; !(lbase & 1) && (last & 1); + lbase = lbase >> 1, last = last >> 1) ; + if (lbase != last) { + printk(KERN_WARNING "mtrr: base(0x%lx000) is not aligned on a size(0x%lx000) boundary\n", + base, size); + return -EINVAL; + } + return 0; +} + + +static int generic_have_wrcomb(void) +{ + unsigned long config, dummy; + rdmsr(MTRRcap_MSR, config, dummy); + return (config & (1 << 10)); +} + +int positive_have_wrcomb(void) +{ + return 1; +} + +/* generic structure... + */ +struct mtrr_ops generic_mtrr_ops = { + .use_intel_if = 1, + .set_all = generic_set_all, + .get = generic_get_mtrr, + .get_free_region = generic_get_free_region, + .set = generic_set_mtrr, + .validate_add_page = generic_validate_add_page, + .have_wrcomb = generic_have_wrcomb, +}; diff --git a/arch/i386/kernel/cpu/mtrr/if.c b/arch/i386/kernel/cpu/mtrr/if.c new file mode 100644 index 00000000000..1923e0aed26 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/if.c @@ -0,0 +1,374 @@ +#include <linux/init.h> +#include <linux/proc_fs.h> +#include <linux/ctype.h> +#include <linux/module.h> +#include <linux/seq_file.h> +#include <asm/uaccess.h> + +#define LINE_SIZE 80 + +#include <asm/mtrr.h> +#include "mtrr.h" + +/* RED-PEN: this is accessed without any locking */ +extern unsigned int *usage_table; + + +#define FILE_FCOUNT(f) (((struct seq_file *)((f)->private_data))->private) + +static char *mtrr_strings[MTRR_NUM_TYPES] = +{ + "uncachable", /* 0 */ + "write-combining", /* 1 */ + "?", /* 2 */ + "?", /* 3 */ + "write-through", /* 4 */ + "write-protect", /* 5 */ + "write-back", /* 6 */ +}; + +char *mtrr_attrib_to_str(int x) +{ + return (x <= 6) ? mtrr_strings[x] : "?"; +} + +#ifdef CONFIG_PROC_FS + +static int +mtrr_file_add(unsigned long base, unsigned long size, + unsigned int type, char increment, struct file *file, int page) +{ + int reg, max; + unsigned int *fcount = FILE_FCOUNT(file); + + max = num_var_ranges; + if (fcount == NULL) { + fcount = kmalloc(max * sizeof *fcount, GFP_KERNEL); + if (!fcount) + return -ENOMEM; + memset(fcount, 0, max * sizeof *fcount); + FILE_FCOUNT(file) = fcount; + } + if (!page) { + if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1))) + return -EINVAL; + base >>= PAGE_SHIFT; + size >>= PAGE_SHIFT; + } + reg = mtrr_add_page(base, size, type, 1); + if (reg >= 0) + ++fcount[reg]; + return reg; +} + +static int +mtrr_file_del(unsigned long base, unsigned long size, + struct file *file, int page) +{ + int reg; + unsigned int *fcount = FILE_FCOUNT(file); + + if (!page) { + if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1))) + return -EINVAL; + base >>= PAGE_SHIFT; + size >>= PAGE_SHIFT; + } + reg = mtrr_del_page(-1, base, size); + if (reg < 0) + return reg; + if (fcount == NULL) + return reg; + if (fcount[reg] < 1) + return -EINVAL; + --fcount[reg]; + return reg; +} + +/* RED-PEN: seq_file can seek now. this is ignored. */ +static ssize_t +mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos) +/* Format of control line: + "base=%Lx size=%Lx type=%s" OR: + "disable=%d" +*/ +{ + int i, err; + unsigned long reg; + unsigned long long base, size; + char *ptr; + char line[LINE_SIZE]; + size_t linelen; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (!len) + return -EINVAL; + memset(line, 0, LINE_SIZE); + if (len > LINE_SIZE) + len = LINE_SIZE; + if (copy_from_user(line, buf, len - 1)) + return -EFAULT; + linelen = strlen(line); + ptr = line + linelen - 1; + if (linelen && *ptr == '\n') + *ptr = '\0'; + if (!strncmp(line, "disable=", 8)) { + reg = simple_strtoul(line + 8, &ptr, 0); + err = mtrr_del_page(reg, 0, 0); + if (err < 0) + return err; + return len; + } + if (strncmp(line, "base=", 5)) + return -EINVAL; + base = simple_strtoull(line + 5, &ptr, 0); + for (; isspace(*ptr); ++ptr) ; + if (strncmp(ptr, "size=", 5)) + return -EINVAL; + size = simple_strtoull(ptr + 5, &ptr, 0); + if ((base & 0xfff) || (size & 0xfff)) + return -EINVAL; + for (; isspace(*ptr); ++ptr) ; + if (strncmp(ptr, "type=", 5)) + return -EINVAL; + ptr += 5; + for (; isspace(*ptr); ++ptr) ; + for (i = 0; i < MTRR_NUM_TYPES; ++i) { + if (strcmp(ptr, mtrr_strings[i])) + continue; + base >>= PAGE_SHIFT; + size >>= PAGE_SHIFT; + err = + mtrr_add_page((unsigned long) base, (unsigned long) size, i, + 1); + if (err < 0) + return err; + return len; + } + return -EINVAL; +} + +static int +mtrr_ioctl(struct inode *inode, struct file *file, + unsigned int cmd, unsigned long __arg) +{ + int err; + mtrr_type type; + struct mtrr_sentry sentry; + struct mtrr_gentry gentry; + void __user *arg = (void __user *) __arg; + + switch (cmd) { + default: + return -ENOTTY; + case MTRRIOC_ADD_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = + mtrr_file_add(sentry.base, sentry.size, sentry.type, 1, + file, 0); + if (err < 0) + return err; + break; + case MTRRIOC_SET_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_add(sentry.base, sentry.size, sentry.type, 0); + if (err < 0) + return err; + break; + case MTRRIOC_DEL_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_file_del(sentry.base, sentry.size, file, 0); + if (err < 0) + return err; + break; + case MTRRIOC_KILL_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_del(-1, sentry.base, sentry.size); + if (err < 0) + return err; + break; + case MTRRIOC_GET_ENTRY: + if (copy_from_user(&gentry, arg, sizeof gentry)) + return -EFAULT; + if (gentry.regnum >= num_var_ranges) + return -EINVAL; + mtrr_if->get(gentry.regnum, &gentry.base, &gentry.size, &type); + + /* Hide entries that go above 4GB */ + if (gentry.base + gentry.size > 0x100000 + || gentry.size == 0x100000) + gentry.base = gentry.size = gentry.type = 0; + else { + gentry.base <<= PAGE_SHIFT; + gentry.size <<= PAGE_SHIFT; + gentry.type = type; + } + + if (copy_to_user(arg, &gentry, sizeof gentry)) + return -EFAULT; + break; + case MTRRIOC_ADD_PAGE_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = + mtrr_file_add(sentry.base, sentry.size, sentry.type, 1, + file, 1); + if (err < 0) + return err; + break; + case MTRRIOC_SET_PAGE_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_add_page(sentry.base, sentry.size, sentry.type, 0); + if (err < 0) + return err; + break; + case MTRRIOC_DEL_PAGE_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_file_del(sentry.base, sentry.size, file, 1); + if (err < 0) + return err; + break; + case MTRRIOC_KILL_PAGE_ENTRY: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (copy_from_user(&sentry, arg, sizeof sentry)) + return -EFAULT; + err = mtrr_del_page(-1, sentry.base, sentry.size); + if (err < 0) + return err; + break; + case MTRRIOC_GET_PAGE_ENTRY: + if (copy_from_user(&gentry, arg, sizeof gentry)) + return -EFAULT; + if (gentry.regnum >= num_var_ranges) + return -EINVAL; + mtrr_if->get(gentry.regnum, &gentry.base, &gentry.size, &type); + gentry.type = type; + + if (copy_to_user(arg, &gentry, sizeof gentry)) + return -EFAULT; + break; + } + return 0; +} + +static int +mtrr_close(struct inode *ino, struct file *file) +{ + int i, max; + unsigned int *fcount = FILE_FCOUNT(file); + + if (fcount != NULL) { + max = num_var_ranges; + for (i = 0; i < max; ++i) { + while (fcount[i] > 0) { + mtrr_del(i, 0, 0); + --fcount[i]; + } + } + kfree(fcount); + FILE_FCOUNT(file) = NULL; + } + return single_release(ino, file); +} + +static int mtrr_seq_show(struct seq_file *seq, void *offset); + +static int mtrr_open(struct inode *inode, struct file *file) +{ + if (!mtrr_if) + return -EIO; + if (!mtrr_if->get) + return -ENXIO; + return single_open(file, mtrr_seq_show, NULL); +} + +static struct file_operations mtrr_fops = { + .owner = THIS_MODULE, + .open = mtrr_open, + .read = seq_read, + .llseek = seq_lseek, + .write = mtrr_write, + .ioctl = mtrr_ioctl, + .release = mtrr_close, +}; + + +static struct proc_dir_entry *proc_root_mtrr; + + +static int mtrr_seq_show(struct seq_file *seq, void *offset) +{ + char factor; + int i, max, len; + mtrr_type type; + unsigned long base; + unsigned int size; + + len = 0; + max = num_var_ranges; + for (i = 0; i < max; i++) { + mtrr_if->get(i, &base, &size, &type); + if (size == 0) + usage_table[i] = 0; + else { + if (size < (0x100000 >> PAGE_SHIFT)) { + /* less than 1MB */ + factor = 'K'; + size <<= PAGE_SHIFT - 10; + } else { + factor = 'M'; + size >>= 20 - PAGE_SHIFT; + } + /* RED-PEN: base can be > 32bit */ + len += seq_printf(seq, + "reg%02i: base=0x%05lx000 (%4liMB), size=%4i%cB: %s, count=%d\n", + i, base, base >> (20 - PAGE_SHIFT), size, factor, + mtrr_attrib_to_str(type), usage_table[i]); + } + } + return 0; +} + +static int __init mtrr_if_init(void) +{ + struct cpuinfo_x86 *c = &boot_cpu_data; + + if ((!cpu_has(c, X86_FEATURE_MTRR)) && + (!cpu_has(c, X86_FEATURE_K6_MTRR)) && + (!cpu_has(c, X86_FEATURE_CYRIX_ARR)) && + (!cpu_has(c, X86_FEATURE_CENTAUR_MCR))) + return -ENODEV; + + proc_root_mtrr = + create_proc_entry("mtrr", S_IWUSR | S_IRUGO, &proc_root); + if (proc_root_mtrr) { + proc_root_mtrr->owner = THIS_MODULE; + proc_root_mtrr->proc_fops = &mtrr_fops; + } + return 0; +} + +arch_initcall(mtrr_if_init); +#endif /* CONFIG_PROC_FS */ diff --git a/arch/i386/kernel/cpu/mtrr/main.c b/arch/i386/kernel/cpu/mtrr/main.c new file mode 100644 index 00000000000..8f67b490a7f --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/main.c @@ -0,0 +1,693 @@ +/* Generic MTRR (Memory Type Range Register) driver. + + Copyright (C) 1997-2000 Richard Gooch + Copyright (c) 2002 Patrick Mochel + + This library is free software; you can redistribute it and/or + modify it under the terms of the GNU Library General Public + License as published by the Free Software Foundation; either + version 2 of the License, or (at your option) any later version. + + This library is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + Library General Public License for more details. + + You should have received a copy of the GNU Library General Public + License along with this library; if not, write to the Free + Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + + Richard Gooch may be reached by email at rgooch@atnf.csiro.au + The postal address is: + Richard Gooch, c/o ATNF, P. O. Box 76, Epping, N.S.W., 2121, Australia. + + Source: "Pentium Pro Family Developer's Manual, Volume 3: + Operating System Writer's Guide" (Intel document number 242692), + section 11.11.7 + + This was cleaned and made readable by Patrick Mochel <mochel@osdl.org> + on 6-7 March 2002. + Source: Intel Architecture Software Developers Manual, Volume 3: + System Programming Guide; Section 9.11. (1997 edition - PPro). +*/ + +#include <linux/module.h> +#include <linux/init.h> +#include <linux/pci.h> +#include <linux/smp.h> +#include <linux/cpu.h> + +#include <asm/mtrr.h> + +#include <asm/uaccess.h> +#include <asm/processor.h> +#include <asm/msr.h> +#include "mtrr.h" + +#define MTRR_VERSION "2.0 (20020519)" + +u32 num_var_ranges = 0; + +unsigned int *usage_table; +static DECLARE_MUTEX(main_lock); + +u32 size_or_mask, size_and_mask; + +static struct mtrr_ops * mtrr_ops[X86_VENDOR_NUM] = {}; + +struct mtrr_ops * mtrr_if = NULL; + +static void set_mtrr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type); + +extern int arr3_protected; + +void set_mtrr_ops(struct mtrr_ops * ops) +{ + if (ops->vendor && ops->vendor < X86_VENDOR_NUM) + mtrr_ops[ops->vendor] = ops; +} + +/* Returns non-zero if we have the write-combining memory type */ +static int have_wrcomb(void) +{ + struct pci_dev *dev; + + if ((dev = pci_get_class(PCI_CLASS_BRIDGE_HOST << 8, NULL)) != NULL) { + /* ServerWorks LE chipsets have problems with write-combining + Don't allow it and leave room for other chipsets to be tagged */ + if (dev->vendor == PCI_VENDOR_ID_SERVERWORKS && + dev->device == PCI_DEVICE_ID_SERVERWORKS_LE) { + printk(KERN_INFO "mtrr: Serverworks LE detected. Write-combining disabled.\n"); + pci_dev_put(dev); + return 0; + } + /* Intel 450NX errata # 23. Non ascending cachline evictions to + write combining memory may resulting in data corruption */ + if (dev->vendor == PCI_VENDOR_ID_INTEL && + dev->device == PCI_DEVICE_ID_INTEL_82451NX) { + printk(KERN_INFO "mtrr: Intel 450NX MMC detected. Write-combining disabled.\n"); + pci_dev_put(dev); + return 0; + } + pci_dev_put(dev); + } + return (mtrr_if->have_wrcomb ? mtrr_if->have_wrcomb() : 0); +} + +/* This function returns the number of variable MTRRs */ +static void __init set_num_var_ranges(void) +{ + unsigned long config = 0, dummy; + + if (use_intel()) { + rdmsr(MTRRcap_MSR, config, dummy); + } else if (is_cpu(AMD)) + config = 2; + else if (is_cpu(CYRIX) || is_cpu(CENTAUR)) + config = 8; + num_var_ranges = config & 0xff; +} + +static void __init init_table(void) +{ + int i, max; + + max = num_var_ranges; + if ((usage_table = kmalloc(max * sizeof *usage_table, GFP_KERNEL)) + == NULL) { + printk(KERN_ERR "mtrr: could not allocate\n"); + return; + } + for (i = 0; i < max; i++) + usage_table[i] = 1; +} + +struct set_mtrr_data { + atomic_t count; + atomic_t gate; + unsigned long smp_base; + unsigned long smp_size; + unsigned int smp_reg; + mtrr_type smp_type; +}; + +#ifdef CONFIG_SMP + +static void ipi_handler(void *info) +/* [SUMMARY] Synchronisation handler. Executed by "other" CPUs. + [RETURNS] Nothing. +*/ +{ + struct set_mtrr_data *data = info; + unsigned long flags; + + local_irq_save(flags); + + atomic_dec(&data->count); + while(!atomic_read(&data->gate)) + cpu_relax(); + + /* The master has cleared me to execute */ + if (data->smp_reg != ~0U) + mtrr_if->set(data->smp_reg, data->smp_base, + data->smp_size, data->smp_type); + else + mtrr_if->set_all(); + + atomic_dec(&data->count); + while(atomic_read(&data->gate)) + cpu_relax(); + + atomic_dec(&data->count); + local_irq_restore(flags); +} + +#endif + +/** + * set_mtrr - update mtrrs on all processors + * @reg: mtrr in question + * @base: mtrr base + * @size: mtrr size + * @type: mtrr type + * + * This is kinda tricky, but fortunately, Intel spelled it out for us cleanly: + * + * 1. Send IPI to do the following: + * 2. Disable Interrupts + * 3. Wait for all procs to do so + * 4. Enter no-fill cache mode + * 5. Flush caches + * 6. Clear PGE bit + * 7. Flush all TLBs + * 8. Disable all range registers + * 9. Update the MTRRs + * 10. Enable all range registers + * 11. Flush all TLBs and caches again + * 12. Enter normal cache mode and reenable caching + * 13. Set PGE + * 14. Wait for buddies to catch up + * 15. Enable interrupts. + * + * What does that mean for us? Well, first we set data.count to the number + * of CPUs. As each CPU disables interrupts, it'll decrement it once. We wait + * until it hits 0 and proceed. We set the data.gate flag and reset data.count. + * Meanwhile, they are waiting for that flag to be set. Once it's set, each + * CPU goes through the transition of updating MTRRs. The CPU vendors may each do it + * differently, so we call mtrr_if->set() callback and let them take care of it. + * When they're done, they again decrement data->count and wait for data.gate to + * be reset. + * When we finish, we wait for data.count to hit 0 and toggle the data.gate flag. + * Everyone then enables interrupts and we all continue on. + * + * Note that the mechanism is the same for UP systems, too; all the SMP stuff + * becomes nops. + */ +static void set_mtrr(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type) +{ + struct set_mtrr_data data; + unsigned long flags; + + data.smp_reg = reg; + data.smp_base = base; + data.smp_size = size; + data.smp_type = type; + atomic_set(&data.count, num_booting_cpus() - 1); + atomic_set(&data.gate,0); + + /* Start the ball rolling on other CPUs */ + if (smp_call_function(ipi_handler, &data, 1, 0) != 0) + panic("mtrr: timed out waiting for other CPUs\n"); + + local_irq_save(flags); + + while(atomic_read(&data.count)) + cpu_relax(); + + /* ok, reset count and toggle gate */ + atomic_set(&data.count, num_booting_cpus() - 1); + atomic_set(&data.gate,1); + + /* do our MTRR business */ + + /* HACK! + * We use this same function to initialize the mtrrs on boot. + * The state of the boot cpu's mtrrs has been saved, and we want + * to replicate across all the APs. + * If we're doing that @reg is set to something special... + */ + if (reg != ~0U) + mtrr_if->set(reg,base,size,type); + + /* wait for the others */ + while(atomic_read(&data.count)) + cpu_relax(); + + atomic_set(&data.count, num_booting_cpus() - 1); + atomic_set(&data.gate,0); + + /* + * Wait here for everyone to have seen the gate change + * So we're the last ones to touch 'data' + */ + while(atomic_read(&data.count)) + cpu_relax(); + + local_irq_restore(flags); +} + +/** + * mtrr_add_page - Add a memory type region + * @base: Physical base address of region in pages (4 KB) + * @size: Physical size of region in pages (4 KB) + * @type: Type of MTRR desired + * @increment: If this is true do usage counting on the region + * + * Memory type region registers control the caching on newer Intel and + * non Intel processors. This function allows drivers to request an + * MTRR is added. The details and hardware specifics of each processor's + * implementation are hidden from the caller, but nevertheless the + * caller should expect to need to provide a power of two size on an + * equivalent power of two boundary. + * + * If the region cannot be added either because all regions are in use + * or the CPU cannot support it a negative value is returned. On success + * the register number for this entry is returned, but should be treated + * as a cookie only. + * + * On a multiprocessor machine the changes are made to all processors. + * This is required on x86 by the Intel processors. + * + * The available types are + * + * %MTRR_TYPE_UNCACHABLE - No caching + * + * %MTRR_TYPE_WRBACK - Write data back in bursts whenever + * + * %MTRR_TYPE_WRCOMB - Write data back soon but allow bursts + * + * %MTRR_TYPE_WRTHROUGH - Cache reads but not writes + * + * BUGS: Needs a quiet flag for the cases where drivers do not mind + * failures and do not wish system log messages to be sent. + */ + +int mtrr_add_page(unsigned long base, unsigned long size, + unsigned int type, char increment) +{ + int i; + mtrr_type ltype; + unsigned long lbase; + unsigned int lsize; + int error; + + if (!mtrr_if) + return -ENXIO; + + if ((error = mtrr_if->validate_add_page(base,size,type))) + return error; + + if (type >= MTRR_NUM_TYPES) { + printk(KERN_WARNING "mtrr: type: %u invalid\n", type); + return -EINVAL; + } + + /* If the type is WC, check that this processor supports it */ + if ((type == MTRR_TYPE_WRCOMB) && !have_wrcomb()) { + printk(KERN_WARNING + "mtrr: your processor doesn't support write-combining\n"); + return -ENOSYS; + } + + if (base & size_or_mask || size & size_or_mask) { + printk(KERN_WARNING "mtrr: base or size exceeds the MTRR width\n"); + return -EINVAL; + } + + error = -EINVAL; + + /* Search for existing MTRR */ + down(&main_lock); + for (i = 0; i < num_var_ranges; ++i) { + mtrr_if->get(i, &lbase, &lsize, <ype); + if (base >= lbase + lsize) + continue; + if ((base < lbase) && (base + size <= lbase)) + continue; + /* At this point we know there is some kind of overlap/enclosure */ + if ((base < lbase) || (base + size > lbase + lsize)) { + printk(KERN_WARNING + "mtrr: 0x%lx000,0x%lx000 overlaps existing" + " 0x%lx000,0x%x000\n", base, size, lbase, + lsize); + goto out; + } + /* New region is enclosed by an existing region */ + if (ltype != type) { + if (type == MTRR_TYPE_UNCACHABLE) + continue; + printk (KERN_WARNING "mtrr: type mismatch for %lx000,%lx000 old: %s new: %s\n", + base, size, mtrr_attrib_to_str(ltype), + mtrr_attrib_to_str(type)); + goto out; + } + if (increment) + ++usage_table[i]; + error = i; + goto out; + } + /* Search for an empty MTRR */ + i = mtrr_if->get_free_region(base, size); + if (i >= 0) { + set_mtrr(i, base, size, type); + usage_table[i] = 1; + } else + printk(KERN_INFO "mtrr: no more MTRRs available\n"); + error = i; + out: + up(&main_lock); + return error; +} + +/** + * mtrr_add - Add a memory type region + * @base: Physical base address of region + * @size: Physical size of region + * @type: Type of MTRR desired + * @increment: If this is true do usage counting on the region + * + * Memory type region registers control the caching on newer Intel and + * non Intel processors. This function allows drivers to request an + * MTRR is added. The details and hardware specifics of each processor's + * implementation are hidden from the caller, but nevertheless the + * caller should expect to need to provide a power of two size on an + * equivalent power of two boundary. + * + * If the region cannot be added either because all regions are in use + * or the CPU cannot support it a negative value is returned. On success + * the register number for this entry is returned, but should be treated + * as a cookie only. + * + * On a multiprocessor machine the changes are made to all processors. + * This is required on x86 by the Intel processors. + * + * The available types are + * + * %MTRR_TYPE_UNCACHABLE - No caching + * + * %MTRR_TYPE_WRBACK - Write data back in bursts whenever + * + * %MTRR_TYPE_WRCOMB - Write data back soon but allow bursts + * + * %MTRR_TYPE_WRTHROUGH - Cache reads but not writes + * + * BUGS: Needs a quiet flag for the cases where drivers do not mind + * failures and do not wish system log messages to be sent. + */ + +int +mtrr_add(unsigned long base, unsigned long size, unsigned int type, + char increment) +{ + if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1))) { + printk(KERN_WARNING "mtrr: size and base must be multiples of 4 kiB\n"); + printk(KERN_DEBUG "mtrr: size: 0x%lx base: 0x%lx\n", size, base); + return -EINVAL; + } + return mtrr_add_page(base >> PAGE_SHIFT, size >> PAGE_SHIFT, type, + increment); +} + +/** + * mtrr_del_page - delete a memory type region + * @reg: Register returned by mtrr_add + * @base: Physical base address + * @size: Size of region + * + * If register is supplied then base and size are ignored. This is + * how drivers should call it. + * + * Releases an MTRR region. If the usage count drops to zero the + * register is freed and the region returns to default state. + * On success the register is returned, on failure a negative error + * code. + */ + +int mtrr_del_page(int reg, unsigned long base, unsigned long size) +{ + int i, max; + mtrr_type ltype; + unsigned long lbase; + unsigned int lsize; + int error = -EINVAL; + + if (!mtrr_if) + return -ENXIO; + + max = num_var_ranges; + down(&main_lock); + if (reg < 0) { + /* Search for existing MTRR */ + for (i = 0; i < max; ++i) { + mtrr_if->get(i, &lbase, &lsize, <ype); + if (lbase == base && lsize == size) { + reg = i; + break; + } + } + if (reg < 0) { + printk(KERN_DEBUG "mtrr: no MTRR for %lx000,%lx000 found\n", base, + size); + goto out; + } + } + if (reg >= max) { + printk(KERN_WARNING "mtrr: register: %d too big\n", reg); + goto out; + } + if (is_cpu(CYRIX) && !use_intel()) { + if ((reg == 3) && arr3_protected) { + printk(KERN_WARNING "mtrr: ARR3 cannot be changed\n"); + goto out; + } + } + mtrr_if->get(reg, &lbase, &lsize, <ype); + if (lsize < 1) { + printk(KERN_WARNING "mtrr: MTRR %d not used\n", reg); + goto out; + } + if (usage_table[reg] < 1) { + printk(KERN_WARNING "mtrr: reg: %d has count=0\n", reg); + goto out; + } + if (--usage_table[reg] < 1) + set_mtrr(reg, 0, 0, 0); + error = reg; + out: + up(&main_lock); + return error; +} +/** + * mtrr_del - delete a memory type region + * @reg: Register returned by mtrr_add + * @base: Physical base address + * @size: Size of region + * + * If register is supplied then base and size are ignored. This is + * how drivers should call it. + * + * Releases an MTRR region. If the usage count drops to zero the + * register is freed and the region returns to default state. + * On success the register is returned, on failure a negative error + * code. + */ + +int +mtrr_del(int reg, unsigned long base, unsigned long size) +{ + if ((base & (PAGE_SIZE - 1)) || (size & (PAGE_SIZE - 1))) { + printk(KERN_INFO "mtrr: size and base must be multiples of 4 kiB\n"); + printk(KERN_DEBUG "mtrr: size: 0x%lx base: 0x%lx\n", size, base); + return -EINVAL; + } + return mtrr_del_page(reg, base >> PAGE_SHIFT, size >> PAGE_SHIFT); +} + +EXPORT_SYMBOL(mtrr_add); +EXPORT_SYMBOL(mtrr_del); + +/* HACK ALERT! + * These should be called implicitly, but we can't yet until all the initcall + * stuff is done... + */ +extern void amd_init_mtrr(void); +extern void cyrix_init_mtrr(void); +extern void centaur_init_mtrr(void); + +static void __init init_ifs(void) +{ + amd_init_mtrr(); + cyrix_init_mtrr(); + centaur_init_mtrr(); +} + +static void __init init_other_cpus(void) +{ + if (use_intel()) + get_mtrr_state(); + + /* bring up the other processors */ + set_mtrr(~0U,0,0,0); + + if (use_intel()) { + finalize_mtrr_state(); + mtrr_state_warn(); + } +} + + +struct mtrr_value { + mtrr_type ltype; + unsigned long lbase; + unsigned int lsize; +}; + +static struct mtrr_value * mtrr_state; + +static int mtrr_save(struct sys_device * sysdev, u32 state) +{ + int i; + int size = num_var_ranges * sizeof(struct mtrr_value); + + mtrr_state = kmalloc(size,GFP_ATOMIC); + if (mtrr_state) + memset(mtrr_state,0,size); + else + return -ENOMEM; + + for (i = 0; i < num_var_ranges; i++) { + mtrr_if->get(i, + &mtrr_state[i].lbase, + &mtrr_state[i].lsize, + &mtrr_state[i].ltype); + } + return 0; +} + +static int mtrr_restore(struct sys_device * sysdev) +{ + int i; + + for (i = 0; i < num_var_ranges; i++) { + if (mtrr_state[i].lsize) + set_mtrr(i, + mtrr_state[i].lbase, + mtrr_state[i].lsize, + mtrr_state[i].ltype); + } + kfree(mtrr_state); + return 0; +} + + + +static struct sysdev_driver mtrr_sysdev_driver = { + .suspend = mtrr_save, + .resume = mtrr_restore, +}; + + +/** + * mtrr_init - initialize mtrrs on the boot CPU + * + * This needs to be called early; before any of the other CPUs are + * initialized (i.e. before smp_init()). + * + */ +static int __init mtrr_init(void) +{ + init_ifs(); + + if (cpu_has_mtrr) { + mtrr_if = &generic_mtrr_ops; + size_or_mask = 0xff000000; /* 36 bits */ + size_and_mask = 0x00f00000; + + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + /* The original Athlon docs said that + total addressable memory is 44 bits wide. + It was not really clear whether its MTRRs + follow this or not. (Read: 44 or 36 bits). + However, "x86-64_overview.pdf" explicitly + states that "previous implementations support + 36 bit MTRRs" and also provides a way to + query the width (in bits) of the physical + addressable memory on the Hammer family. + */ + if (boot_cpu_data.x86 == 15 + && (cpuid_eax(0x80000000) >= 0x80000008)) { + u32 phys_addr; + phys_addr = cpuid_eax(0x80000008) & 0xff; + size_or_mask = + ~((1 << (phys_addr - PAGE_SHIFT)) - 1); + size_and_mask = ~size_or_mask & 0xfff00000; + } + /* Athlon MTRRs use an Intel-compatible interface for + * getting and setting */ + break; + case X86_VENDOR_CENTAUR: + if (boot_cpu_data.x86 == 6) { + /* VIA Cyrix family have Intel style MTRRs, but don't support PAE */ + size_or_mask = 0xfff00000; /* 32 bits */ + size_and_mask = 0; + } + break; + + default: + break; + } + } else { + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + if (cpu_has_k6_mtrr) { + /* Pre-Athlon (K6) AMD CPU MTRRs */ + mtrr_if = mtrr_ops[X86_VENDOR_AMD]; + size_or_mask = 0xfff00000; /* 32 bits */ + size_and_mask = 0; + } + break; + case X86_VENDOR_CENTAUR: + if (cpu_has_centaur_mcr) { + mtrr_if = mtrr_ops[X86_VENDOR_CENTAUR]; + size_or_mask = 0xfff00000; /* 32 bits */ + size_and_mask = 0; + } + break; + case X86_VENDOR_CYRIX: + if (cpu_has_cyrix_arr) { + mtrr_if = mtrr_ops[X86_VENDOR_CYRIX]; + size_or_mask = 0xfff00000; /* 32 bits */ + size_and_mask = 0; + } + break; + default: + break; + } + } + printk(KERN_INFO "mtrr: v%s\n",MTRR_VERSION); + + if (mtrr_if) { + set_num_var_ranges(); + init_table(); + init_other_cpus(); + + return sysdev_driver_register(&cpu_sysdev_class, + &mtrr_sysdev_driver); + } + return -ENXIO; +} + +subsys_initcall(mtrr_init); diff --git a/arch/i386/kernel/cpu/mtrr/mtrr.h b/arch/i386/kernel/cpu/mtrr/mtrr.h new file mode 100644 index 00000000000..de135124559 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/mtrr.h @@ -0,0 +1,98 @@ +/* + * local mtrr defines. + */ + +#ifndef TRUE +#define TRUE 1 +#define FALSE 0 +#endif + +#define MTRRcap_MSR 0x0fe +#define MTRRdefType_MSR 0x2ff + +#define MTRRphysBase_MSR(reg) (0x200 + 2 * (reg)) +#define MTRRphysMask_MSR(reg) (0x200 + 2 * (reg) + 1) + +#define NUM_FIXED_RANGES 88 +#define MTRRfix64K_00000_MSR 0x250 +#define MTRRfix16K_80000_MSR 0x258 +#define MTRRfix16K_A0000_MSR 0x259 +#define MTRRfix4K_C0000_MSR 0x268 +#define MTRRfix4K_C8000_MSR 0x269 +#define MTRRfix4K_D0000_MSR 0x26a +#define MTRRfix4K_D8000_MSR 0x26b +#define MTRRfix4K_E0000_MSR 0x26c +#define MTRRfix4K_E8000_MSR 0x26d +#define MTRRfix4K_F0000_MSR 0x26e +#define MTRRfix4K_F8000_MSR 0x26f + +#define MTRR_CHANGE_MASK_FIXED 0x01 +#define MTRR_CHANGE_MASK_VARIABLE 0x02 +#define MTRR_CHANGE_MASK_DEFTYPE 0x04 + +/* In the Intel processor's MTRR interface, the MTRR type is always held in + an 8 bit field: */ +typedef u8 mtrr_type; + +struct mtrr_ops { + u32 vendor; + u32 use_intel_if; +// void (*init)(void); + void (*set)(unsigned int reg, unsigned long base, + unsigned long size, mtrr_type type); + void (*set_all)(void); + + void (*get)(unsigned int reg, unsigned long *base, + unsigned int *size, mtrr_type * type); + int (*get_free_region) (unsigned long base, unsigned long size); + + int (*validate_add_page)(unsigned long base, unsigned long size, + unsigned int type); + int (*have_wrcomb)(void); +}; + +extern int generic_get_free_region(unsigned long base, unsigned long size); +extern int generic_validate_add_page(unsigned long base, unsigned long size, + unsigned int type); + +extern struct mtrr_ops generic_mtrr_ops; + +extern int positive_have_wrcomb(void); + +/* library functions for processor-specific routines */ +struct set_mtrr_context { + unsigned long flags; + unsigned long deftype_lo; + unsigned long deftype_hi; + unsigned long cr4val; + unsigned long ccr3; +}; + +struct mtrr_var_range { + unsigned long base_lo; + unsigned long base_hi; + unsigned long mask_lo; + unsigned long mask_hi; +}; + +void set_mtrr_done(struct set_mtrr_context *ctxt); +void set_mtrr_cache_disable(struct set_mtrr_context *ctxt); +void set_mtrr_prepare_save(struct set_mtrr_context *ctxt); + +void get_mtrr_state(void); + +extern void set_mtrr_ops(struct mtrr_ops * ops); + +extern u32 size_or_mask, size_and_mask; +extern struct mtrr_ops * mtrr_if; + +#define is_cpu(vnd) (mtrr_if && mtrr_if->vendor == X86_VENDOR_##vnd) +#define use_intel() (mtrr_if && mtrr_if->use_intel_if == 1) + +extern unsigned int num_var_ranges; + +void finalize_mtrr_state(void); +void mtrr_state_warn(void); +char *mtrr_attrib_to_str(int x); +void mtrr_wrmsr(unsigned, unsigned, unsigned); + diff --git a/arch/i386/kernel/cpu/mtrr/state.c b/arch/i386/kernel/cpu/mtrr/state.c new file mode 100644 index 00000000000..f62ecd15811 --- /dev/null +++ b/arch/i386/kernel/cpu/mtrr/state.c @@ -0,0 +1,78 @@ +#include <linux/mm.h> +#include <linux/init.h> +#include <asm/io.h> +#include <asm/mtrr.h> +#include <asm/msr.h> +#include "mtrr.h" + + +/* Put the processor into a state where MTRRs can be safely set */ +void set_mtrr_prepare_save(struct set_mtrr_context *ctxt) +{ + unsigned int cr0; + + /* Disable interrupts locally */ + local_irq_save(ctxt->flags); + + if (use_intel() || is_cpu(CYRIX)) { + + /* Save value of CR4 and clear Page Global Enable (bit 7) */ + if ( cpu_has_pge ) { + ctxt->cr4val = read_cr4(); + write_cr4(ctxt->cr4val & (unsigned char) ~(1 << 7)); + } + + /* Disable and flush caches. Note that wbinvd flushes the TLBs as + a side-effect */ + cr0 = read_cr0() | 0x40000000; + wbinvd(); + write_cr0(cr0); + wbinvd(); + + if (use_intel()) + /* Save MTRR state */ + rdmsr(MTRRdefType_MSR, ctxt->deftype_lo, ctxt->deftype_hi); + else + /* Cyrix ARRs - everything else were excluded at the top */ + ctxt->ccr3 = getCx86(CX86_CCR3); + } +} + +void set_mtrr_cache_disable(struct set_mtrr_context *ctxt) +{ + if (use_intel()) + /* Disable MTRRs, and set the default type to uncached */ + mtrr_wrmsr(MTRRdefType_MSR, ctxt->deftype_lo & 0xf300UL, + ctxt->deftype_hi); + else if (is_cpu(CYRIX)) + /* Cyrix ARRs - everything else were excluded at the top */ + setCx86(CX86_CCR3, (ctxt->ccr3 & 0x0f) | 0x10); +} + +/* Restore the processor after a set_mtrr_prepare */ +void set_mtrr_done(struct set_mtrr_context *ctxt) +{ + if (use_intel() || is_cpu(CYRIX)) { + + /* Flush caches and TLBs */ + wbinvd(); + + /* Restore MTRRdefType */ + if (use_intel()) + /* Intel (P6) standard MTRRs */ + mtrr_wrmsr(MTRRdefType_MSR, ctxt->deftype_lo, ctxt->deftype_hi); + else + /* Cyrix ARRs - everything else was excluded at the top */ + setCx86(CX86_CCR3, ctxt->ccr3); + + /* Enable caches */ + write_cr0(read_cr0() & 0xbfffffff); + + /* Restore value of CR4 */ + if ( cpu_has_pge ) + write_cr4(ctxt->cr4val); + } + /* Re-enable interrupts locally (if enabled previously) */ + local_irq_restore(ctxt->flags); +} + diff --git a/arch/i386/kernel/cpu/nexgen.c b/arch/i386/kernel/cpu/nexgen.c new file mode 100644 index 00000000000..30898a260a5 --- /dev/null +++ b/arch/i386/kernel/cpu/nexgen.c @@ -0,0 +1,63 @@ +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/string.h> +#include <asm/processor.h> + +#include "cpu.h" + +/* + * Detect a NexGen CPU running without BIOS hypercode new enough + * to have CPUID. (Thanks to Herbert Oppmann) + */ + +static int __init deep_magic_nexgen_probe(void) +{ + int ret; + + __asm__ __volatile__ ( + " movw $0x5555, %%ax\n" + " xorw %%dx,%%dx\n" + " movw $2, %%cx\n" + " divw %%cx\n" + " movl $0, %%eax\n" + " jnz 1f\n" + " movl $1, %%eax\n" + "1:\n" + : "=a" (ret) : : "cx", "dx" ); + return ret; +} + +static void __init init_nexgen(struct cpuinfo_x86 * c) +{ + c->x86_cache_size = 256; /* A few had 1 MB... */ +} + +static void __init nexgen_identify(struct cpuinfo_x86 * c) +{ + /* Detect NexGen with old hypercode */ + if ( deep_magic_nexgen_probe() ) { + strcpy(c->x86_vendor_id, "NexGenDriven"); + } + generic_identify(c); +} + +static struct cpu_dev nexgen_cpu_dev __initdata = { + .c_vendor = "Nexgen", + .c_ident = { "NexGenDriven" }, + .c_models = { + { .vendor = X86_VENDOR_NEXGEN, + .family = 5, + .model_names = { [1] = "Nx586" } + }, + }, + .c_init = init_nexgen, + .c_identify = nexgen_identify, +}; + +int __init nexgen_init_cpu(void) +{ + cpu_devs[X86_VENDOR_NEXGEN] = &nexgen_cpu_dev; + return 0; +} + +//early_arch_initcall(nexgen_init_cpu); diff --git a/arch/i386/kernel/cpu/proc.c b/arch/i386/kernel/cpu/proc.c new file mode 100644 index 00000000000..c8d83fdc237 --- /dev/null +++ b/arch/i386/kernel/cpu/proc.c @@ -0,0 +1,149 @@ +#include <linux/smp.h> +#include <linux/timex.h> +#include <linux/string.h> +#include <asm/semaphore.h> +#include <linux/seq_file.h> + +/* + * Get CPU information for use by the procfs. + */ +static int show_cpuinfo(struct seq_file *m, void *v) +{ + /* + * These flag bits must match the definitions in <asm/cpufeature.h>. + * NULL means this bit is undefined or reserved; either way it doesn't + * have meaning as far as Linux is concerned. Note that it's important + * to realize there is a difference between this table and CPUID -- if + * applications want to get the raw CPUID data, they should access + * /dev/cpu/<cpu_nr>/cpuid instead. + */ + static char *x86_cap_flags[] = { + /* Intel-defined */ + "fpu", "vme", "de", "pse", "tsc", "msr", "pae", "mce", + "cx8", "apic", NULL, "sep", "mtrr", "pge", "mca", "cmov", + "pat", "pse36", "pn", "clflush", NULL, "dts", "acpi", "mmx", + "fxsr", "sse", "sse2", "ss", "ht", "tm", "ia64", "pbe", + + /* AMD-defined */ + "pni", NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, "syscall", NULL, NULL, NULL, NULL, + NULL, NULL, NULL, "mp", "nx", NULL, "mmxext", NULL, + NULL, "fxsr_opt", NULL, NULL, NULL, "lm", "3dnowext", "3dnow", + + /* Transmeta-defined */ + "recovery", "longrun", NULL, "lrti", NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + + /* Other (Linux-defined) */ + "cxmmx", "k6_mtrr", "cyrix_arr", "centaur_mcr", + NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + + /* Intel-defined (#2) */ + "pni", NULL, NULL, "monitor", "ds_cpl", NULL, NULL, "est", + "tm2", NULL, "cid", NULL, NULL, "cx16", "xtpr", NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + + /* VIA/Cyrix/Centaur-defined */ + NULL, NULL, "rng", "rng_en", NULL, NULL, "ace", "ace_en", + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + + /* AMD-defined (#2) */ + "lahf_lm", "cmp_legacy", NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, + }; + struct cpuinfo_x86 *c = v; + int i, n = c - cpu_data; + int fpu_exception; + +#ifdef CONFIG_SMP + if (!cpu_online(n)) + return 0; +#endif + seq_printf(m, "processor\t: %d\n" + "vendor_id\t: %s\n" + "cpu family\t: %d\n" + "model\t\t: %d\n" + "model name\t: %s\n", + n, + c->x86_vendor_id[0] ? c->x86_vendor_id : "unknown", + c->x86, + c->x86_model, + c->x86_model_id[0] ? c->x86_model_id : "unknown"); + + if (c->x86_mask || c->cpuid_level >= 0) + seq_printf(m, "stepping\t: %d\n", c->x86_mask); + else + seq_printf(m, "stepping\t: unknown\n"); + + if ( cpu_has(c, X86_FEATURE_TSC) ) { + seq_printf(m, "cpu MHz\t\t: %lu.%03lu\n", + cpu_khz / 1000, (cpu_khz % 1000)); + } + + /* Cache size */ + if (c->x86_cache_size >= 0) + seq_printf(m, "cache size\t: %d KB\n", c->x86_cache_size); +#ifdef CONFIG_X86_HT + seq_printf(m, "physical id\t: %d\n", phys_proc_id[n]); + seq_printf(m, "siblings\t: %d\n", c->x86_num_cores * smp_num_siblings); +#endif + + /* We use exception 16 if we have hardware math and we've either seen it or the CPU claims it is internal */ + fpu_exception = c->hard_math && (ignore_fpu_irq || cpu_has_fpu); + seq_printf(m, "fdiv_bug\t: %s\n" + "hlt_bug\t\t: %s\n" + "f00f_bug\t: %s\n" + "coma_bug\t: %s\n" + "fpu\t\t: %s\n" + "fpu_exception\t: %s\n" + "cpuid level\t: %d\n" + "wp\t\t: %s\n" + "flags\t\t:", + c->fdiv_bug ? "yes" : "no", + c->hlt_works_ok ? "no" : "yes", + c->f00f_bug ? "yes" : "no", + c->coma_bug ? "yes" : "no", + c->hard_math ? "yes" : "no", + fpu_exception ? "yes" : "no", + c->cpuid_level, + c->wp_works_ok ? "yes" : "no"); + + for ( i = 0 ; i < 32*NCAPINTS ; i++ ) + if ( test_bit(i, c->x86_capability) && + x86_cap_flags[i] != NULL ) + seq_printf(m, " %s", x86_cap_flags[i]); + + seq_printf(m, "\nbogomips\t: %lu.%02lu\n\n", + c->loops_per_jiffy/(500000/HZ), + (c->loops_per_jiffy/(5000/HZ)) % 100); + return 0; +} + +static void *c_start(struct seq_file *m, loff_t *pos) +{ + return *pos < NR_CPUS ? cpu_data + *pos : NULL; +} +static void *c_next(struct seq_file *m, void *v, loff_t *pos) +{ + ++*pos; + return c_start(m, pos); +} +static void c_stop(struct seq_file *m, void *v) +{ +} +struct seq_operations cpuinfo_op = { + .start = c_start, + .next = c_next, + .stop = c_stop, + .show = show_cpuinfo, +}; diff --git a/arch/i386/kernel/cpu/rise.c b/arch/i386/kernel/cpu/rise.c new file mode 100644 index 00000000000..8602425628c --- /dev/null +++ b/arch/i386/kernel/cpu/rise.c @@ -0,0 +1,53 @@ +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/bitops.h> +#include <asm/processor.h> + +#include "cpu.h" + +static void __init init_rise(struct cpuinfo_x86 *c) +{ + printk("CPU: Rise iDragon"); + if (c->x86_model > 2) + printk(" II"); + printk("\n"); + + /* Unhide possibly hidden capability flags + The mp6 iDragon family don't have MSRs. + We switch on extra features with this cpuid weirdness: */ + __asm__ ( + "movl $0x6363452a, %%eax\n\t" + "movl $0x3231206c, %%ecx\n\t" + "movl $0x2a32313a, %%edx\n\t" + "cpuid\n\t" + "movl $0x63634523, %%eax\n\t" + "movl $0x32315f6c, %%ecx\n\t" + "movl $0x2333313a, %%edx\n\t" + "cpuid\n\t" : : : "eax", "ebx", "ecx", "edx" + ); + set_bit(X86_FEATURE_CX8, c->x86_capability); +} + +static struct cpu_dev rise_cpu_dev __initdata = { + .c_vendor = "Rise", + .c_ident = { "RiseRiseRise" }, + .c_models = { + { .vendor = X86_VENDOR_RISE, .family = 5, .model_names = + { + [0] = "iDragon", + [2] = "iDragon", + [8] = "iDragon II", + [9] = "iDragon II" + } + }, + }, + .c_init = init_rise, +}; + +int __init rise_init_cpu(void) +{ + cpu_devs[X86_VENDOR_RISE] = &rise_cpu_dev; + return 0; +} + +//early_arch_initcall(rise_init_cpu); diff --git a/arch/i386/kernel/cpu/transmeta.c b/arch/i386/kernel/cpu/transmeta.c new file mode 100644 index 00000000000..f57e5ee9494 --- /dev/null +++ b/arch/i386/kernel/cpu/transmeta.c @@ -0,0 +1,107 @@ +#include <linux/kernel.h> +#include <linux/init.h> +#include <asm/processor.h> +#include <asm/msr.h> +#include "cpu.h" + +static void __init init_transmeta(struct cpuinfo_x86 *c) +{ + unsigned int cap_mask, uk, max, dummy; + unsigned int cms_rev1, cms_rev2; + unsigned int cpu_rev, cpu_freq, cpu_flags, new_cpu_rev; + char cpu_info[65]; + + get_model_name(c); /* Same as AMD/Cyrix */ + display_cacheinfo(c); + + /* Print CMS and CPU revision */ + max = cpuid_eax(0x80860000); + cpu_rev = 0; + if ( max >= 0x80860001 ) { + cpuid(0x80860001, &dummy, &cpu_rev, &cpu_freq, &cpu_flags); + if (cpu_rev != 0x02000000) { + printk(KERN_INFO "CPU: Processor revision %u.%u.%u.%u, %u MHz\n", + (cpu_rev >> 24) & 0xff, + (cpu_rev >> 16) & 0xff, + (cpu_rev >> 8) & 0xff, + cpu_rev & 0xff, + cpu_freq); + } + } + if ( max >= 0x80860002 ) { + cpuid(0x80860002, &new_cpu_rev, &cms_rev1, &cms_rev2, &dummy); + if (cpu_rev == 0x02000000) { + printk(KERN_INFO "CPU: Processor revision %08X, %u MHz\n", + new_cpu_rev, cpu_freq); + } + printk(KERN_INFO "CPU: Code Morphing Software revision %u.%u.%u-%u-%u\n", + (cms_rev1 >> 24) & 0xff, + (cms_rev1 >> 16) & 0xff, + (cms_rev1 >> 8) & 0xff, + cms_rev1 & 0xff, + cms_rev2); + } + if ( max >= 0x80860006 ) { + cpuid(0x80860003, + (void *)&cpu_info[0], + (void *)&cpu_info[4], + (void *)&cpu_info[8], + (void *)&cpu_info[12]); + cpuid(0x80860004, + (void *)&cpu_info[16], + (void *)&cpu_info[20], + (void *)&cpu_info[24], + (void *)&cpu_info[28]); + cpuid(0x80860005, + (void *)&cpu_info[32], + (void *)&cpu_info[36], + (void *)&cpu_info[40], + (void *)&cpu_info[44]); + cpuid(0x80860006, + (void *)&cpu_info[48], + (void *)&cpu_info[52], + (void *)&cpu_info[56], + (void *)&cpu_info[60]); + cpu_info[64] = '\0'; + printk(KERN_INFO "CPU: %s\n", cpu_info); + } + + /* Unhide possibly hidden capability flags */ + rdmsr(0x80860004, cap_mask, uk); + wrmsr(0x80860004, ~0, uk); + c->x86_capability[0] = cpuid_edx(0x00000001); + wrmsr(0x80860004, cap_mask, uk); + + /* If we can run i686 user-space code, call us an i686 */ +#define USER686 (X86_FEATURE_TSC|X86_FEATURE_CX8|X86_FEATURE_CMOV) + if ( c->x86 == 5 && (c->x86_capability[0] & USER686) == USER686 ) + c->x86 = 6; +} + +static void transmeta_identify(struct cpuinfo_x86 * c) +{ + u32 xlvl; + generic_identify(c); + + /* Transmeta-defined flags: level 0x80860001 */ + xlvl = cpuid_eax(0x80860000); + if ( (xlvl & 0xffff0000) == 0x80860000 ) { + if ( xlvl >= 0x80860001 ) + c->x86_capability[2] = cpuid_edx(0x80860001); + } +} + +static struct cpu_dev transmeta_cpu_dev __initdata = { + .c_vendor = "Transmeta", + .c_ident = { "GenuineTMx86", "TransmetaCPU" }, + .c_init = init_transmeta, + .c_identify = transmeta_identify, +}; + +int __init transmeta_init_cpu(void) +{ + cpu_devs[X86_VENDOR_TRANSMETA] = &transmeta_cpu_dev; + return 0; +} + +//early_arch_initcall(transmeta_init_cpu); diff --git a/arch/i386/kernel/cpu/umc.c b/arch/i386/kernel/cpu/umc.c new file mode 100644 index 00000000000..264fcad559d --- /dev/null +++ b/arch/i386/kernel/cpu/umc.c @@ -0,0 +1,33 @@ +#include <linux/kernel.h> +#include <linux/init.h> +#include <asm/processor.h> +#include "cpu.h" + +/* UMC chips appear to be only either 386 or 486, so no special init takes place. + */ +static void __init init_umc(struct cpuinfo_x86 * c) +{ + +} + +static struct cpu_dev umc_cpu_dev __initdata = { + .c_vendor = "UMC", + .c_ident = { "UMC UMC UMC" }, + .c_models = { + { .vendor = X86_VENDOR_UMC, .family = 4, .model_names = + { + [1] = "U5D", + [2] = "U5S", + } + }, + }, + .c_init = init_umc, +}; + +int __init umc_init_cpu(void) +{ + cpu_devs[X86_VENDOR_UMC] = &umc_cpu_dev; + return 0; +} + +//early_arch_initcall(umc_init_cpu); diff --git a/arch/i386/kernel/cpuid.c b/arch/i386/kernel/cpuid.c new file mode 100644 index 00000000000..2e2756345bb --- /dev/null +++ b/arch/i386/kernel/cpuid.c @@ -0,0 +1,246 @@ +/* ----------------------------------------------------------------------- * + * + * Copyright 2000 H. Peter Anvin - All Rights Reserved + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge MA 02139, + * USA; either version 2 of the License, or (at your option) any later + * version; incorporated herein by reference. + * + * ----------------------------------------------------------------------- */ + +/* + * cpuid.c + * + * x86 CPUID access device + * + * This device is accessed by lseek() to the appropriate CPUID level + * and then read in chunks of 16 bytes. A larger size means multiple + * reads of consecutive levels. + * + * This driver uses /dev/cpu/%d/cpuid where %d is the minor number, and on + * an SMP box will direct the access to CPU %d. + */ + +#include <linux/module.h> +#include <linux/config.h> + +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/fcntl.h> +#include <linux/init.h> +#include <linux/poll.h> +#include <linux/smp.h> +#include <linux/major.h> +#include <linux/fs.h> +#include <linux/smp_lock.h> +#include <linux/fs.h> +#include <linux/device.h> +#include <linux/cpu.h> +#include <linux/notifier.h> + +#include <asm/processor.h> +#include <asm/msr.h> +#include <asm/uaccess.h> +#include <asm/system.h> + +static struct class_simple *cpuid_class; + +#ifdef CONFIG_SMP + +struct cpuid_command { + int cpu; + u32 reg; + u32 *data; +}; + +static void cpuid_smp_cpuid(void *cmd_block) +{ + struct cpuid_command *cmd = (struct cpuid_command *)cmd_block; + + if (cmd->cpu == smp_processor_id()) + cpuid(cmd->reg, &cmd->data[0], &cmd->data[1], &cmd->data[2], + &cmd->data[3]); +} + +static inline void do_cpuid(int cpu, u32 reg, u32 * data) +{ + struct cpuid_command cmd; + + preempt_disable(); + if (cpu == smp_processor_id()) { + cpuid(reg, &data[0], &data[1], &data[2], &data[3]); + } else { + cmd.cpu = cpu; + cmd.reg = reg; + cmd.data = data; + + smp_call_function(cpuid_smp_cpuid, &cmd, 1, 1); + } + preempt_enable(); +} +#else /* ! CONFIG_SMP */ + +static inline void do_cpuid(int cpu, u32 reg, u32 * data) +{ + cpuid(reg, &data[0], &data[1], &data[2], &data[3]); +} + +#endif /* ! CONFIG_SMP */ + +static loff_t cpuid_seek(struct file *file, loff_t offset, int orig) +{ + loff_t ret; + + lock_kernel(); + + switch (orig) { + case 0: + file->f_pos = offset; + ret = file->f_pos; + break; + case 1: + file->f_pos += offset; + ret = file->f_pos; + break; + default: + ret = -EINVAL; + } + + unlock_kernel(); + return ret; +} + +static ssize_t cpuid_read(struct file *file, char __user *buf, + size_t count, loff_t * ppos) +{ + char __user *tmp = buf; + u32 data[4]; + size_t rv; + u32 reg = *ppos; + int cpu = iminor(file->f_dentry->d_inode); + + if (count % 16) + return -EINVAL; /* Invalid chunk size */ + + for (rv = 0; count; count -= 16) { + do_cpuid(cpu, reg, data); + if (copy_to_user(tmp, &data, 16)) + return -EFAULT; + tmp += 16; + *ppos = reg++; + } + + return tmp - buf; +} + +static int cpuid_open(struct inode *inode, struct file *file) +{ + unsigned int cpu = iminor(file->f_dentry->d_inode); + struct cpuinfo_x86 *c = &(cpu_data)[cpu]; + + if (cpu >= NR_CPUS || !cpu_online(cpu)) + return -ENXIO; /* No such CPU */ + if (c->cpuid_level < 0) + return -EIO; /* CPUID not supported */ + + return 0; +} + +/* + * File operations we support + */ +static struct file_operations cpuid_fops = { + .owner = THIS_MODULE, + .llseek = cpuid_seek, + .read = cpuid_read, + .open = cpuid_open, +}; + +static int cpuid_class_simple_device_add(int i) +{ + int err = 0; + struct class_device *class_err; + + class_err = class_simple_device_add(cpuid_class, MKDEV(CPUID_MAJOR, i), NULL, "cpu%d",i); + if (IS_ERR(class_err)) + err = PTR_ERR(class_err); + return err; +} + +static int __devinit cpuid_class_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) +{ + unsigned int cpu = (unsigned long)hcpu; + + switch (action) { + case CPU_ONLINE: + cpuid_class_simple_device_add(cpu); + break; + case CPU_DEAD: + class_simple_device_remove(MKDEV(CPUID_MAJOR, cpu)); + break; + } + return NOTIFY_OK; +} + +static struct notifier_block cpuid_class_cpu_notifier = +{ + .notifier_call = cpuid_class_cpu_callback, +}; + +static int __init cpuid_init(void) +{ + int i, err = 0; + i = 0; + + if (register_chrdev(CPUID_MAJOR, "cpu/cpuid", &cpuid_fops)) { + printk(KERN_ERR "cpuid: unable to get major %d for cpuid\n", + CPUID_MAJOR); + err = -EBUSY; + goto out; + } + cpuid_class = class_simple_create(THIS_MODULE, "cpuid"); + if (IS_ERR(cpuid_class)) { + err = PTR_ERR(cpuid_class); + goto out_chrdev; + } + for_each_online_cpu(i) { + err = cpuid_class_simple_device_add(i); + if (err != 0) + goto out_class; + } + register_cpu_notifier(&cpuid_class_cpu_notifier); + + err = 0; + goto out; + +out_class: + i = 0; + for_each_online_cpu(i) { + class_simple_device_remove(MKDEV(CPUID_MAJOR, i)); + } + class_simple_destroy(cpuid_class); +out_chrdev: + unregister_chrdev(CPUID_MAJOR, "cpu/cpuid"); +out: + return err; +} + +static void __exit cpuid_exit(void) +{ + int cpu = 0; + + for_each_online_cpu(cpu) + class_simple_device_remove(MKDEV(CPUID_MAJOR, cpu)); + class_simple_destroy(cpuid_class); + unregister_chrdev(CPUID_MAJOR, "cpu/cpuid"); + unregister_cpu_notifier(&cpuid_class_cpu_notifier); +} + +module_init(cpuid_init); +module_exit(cpuid_exit); + +MODULE_AUTHOR("H. Peter Anvin <hpa@zytor.com>"); +MODULE_DESCRIPTION("x86 generic CPUID driver"); +MODULE_LICENSE("GPL"); diff --git a/arch/i386/kernel/dmi_scan.c b/arch/i386/kernel/dmi_scan.c new file mode 100644 index 00000000000..6ed7e28f306 --- /dev/null +++ b/arch/i386/kernel/dmi_scan.c @@ -0,0 +1,487 @@ +#include <linux/types.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/init.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/acpi.h> +#include <asm/io.h> +#include <linux/pm.h> +#include <asm/system.h> +#include <linux/dmi.h> +#include <linux/bootmem.h> + + +struct dmi_header +{ + u8 type; + u8 length; + u16 handle; +}; + +#undef DMI_DEBUG + +#ifdef DMI_DEBUG +#define dmi_printk(x) printk x +#else +#define dmi_printk(x) +#endif + +static char * __init dmi_string(struct dmi_header *dm, u8 s) +{ + u8 *bp=(u8 *)dm; + bp+=dm->length; + if(!s) + return ""; + s--; + while(s>0 && *bp) + { + bp+=strlen(bp); + bp++; + s--; + } + return bp; +} + +/* + * We have to be cautious here. We have seen BIOSes with DMI pointers + * pointing to completely the wrong place for example + */ + +static int __init dmi_table(u32 base, int len, int num, void (*decode)(struct dmi_header *)) +{ + u8 *buf; + struct dmi_header *dm; + u8 *data; + int i=0; + + buf = bt_ioremap(base, len); + if(buf==NULL) + return -1; + + data = buf; + + /* + * Stop when we see all the items the table claimed to have + * OR we run off the end of the table (also happens) + */ + + while(i<num && data-buf+sizeof(struct dmi_header)<=len) + { + dm=(struct dmi_header *)data; + /* + * We want to know the total length (formated area and strings) + * before decoding to make sure we won't run off the table in + * dmi_decode or dmi_string + */ + data+=dm->length; + while(data-buf<len-1 && (data[0] || data[1])) + data++; + if(data-buf<len-1) + decode(dm); + data+=2; + i++; + } + bt_iounmap(buf, len); + return 0; +} + + +inline static int __init dmi_checksum(u8 *buf) +{ + u8 sum=0; + int a; + + for(a=0; a<15; a++) + sum+=buf[a]; + return (sum==0); +} + +static int __init dmi_iterate(void (*decode)(struct dmi_header *)) +{ + u8 buf[15]; + char __iomem *p, *q; + + /* + * no iounmap() for that ioremap(); it would be a no-op, but it's + * so early in setup that sucker gets confused into doing what + * it shouldn't if we actually call it. + */ + p = ioremap(0xF0000, 0x10000); + if (p == NULL) + return -1; + for (q = p; q < p + 0x10000; q += 16) { + memcpy_fromio(buf, q, 15); + if(memcmp(buf, "_DMI_", 5)==0 && dmi_checksum(buf)) + { + u16 num=buf[13]<<8|buf[12]; + u16 len=buf[7]<<8|buf[6]; + u32 base=buf[11]<<24|buf[10]<<16|buf[9]<<8|buf[8]; + + /* + * DMI version 0.0 means that the real version is taken from + * the SMBIOS version, which we don't know at this point. + */ + if(buf[14]!=0) + printk(KERN_INFO "DMI %d.%d present.\n", + buf[14]>>4, buf[14]&0x0F); + else + printk(KERN_INFO "DMI present.\n"); + dmi_printk((KERN_INFO "%d structures occupying %d bytes.\n", + num, len)); + dmi_printk((KERN_INFO "DMI table at 0x%08X.\n", + base)); + if(dmi_table(base,len, num, decode)==0) + return 0; + } + } + return -1; +} + +static char *dmi_ident[DMI_STRING_MAX]; + +/* + * Save a DMI string + */ + +static void __init dmi_save_ident(struct dmi_header *dm, int slot, int string) +{ + char *d = (char*)dm; + char *p = dmi_string(dm, d[string]); + if(p==NULL || *p == 0) + return; + if (dmi_ident[slot]) + return; + dmi_ident[slot] = alloc_bootmem(strlen(p)+1); + if(dmi_ident[slot]) + strcpy(dmi_ident[slot], p); + else + printk(KERN_ERR "dmi_save_ident: out of memory.\n"); +} + +/* + * Ugly compatibility crap. + */ +#define dmi_blacklist dmi_system_id +#define NO_MATCH { DMI_NONE, NULL} +#define MATCH DMI_MATCH + +/* + * Toshiba keyboard likes to repeat keys when they are not repeated. + */ + +static __init int broken_toshiba_keyboard(struct dmi_blacklist *d) +{ + printk(KERN_WARNING "Toshiba with broken keyboard detected. If your keyboard sometimes generates 3 keypresses instead of one, see http://davyd.ucc.asn.au/projects/toshiba/README\n"); + return 0; +} + + +#ifdef CONFIG_ACPI_SLEEP +static __init int reset_videomode_after_s3(struct dmi_blacklist *d) +{ + /* See acpi_wakeup.S */ + extern long acpi_video_flags; + acpi_video_flags |= 2; + return 0; +} +#endif + + +#ifdef CONFIG_ACPI_BOOT +extern int acpi_force; + +static __init __attribute__((unused)) int dmi_disable_acpi(struct dmi_blacklist *d) +{ + if (!acpi_force) { + printk(KERN_NOTICE "%s detected: acpi off\n",d->ident); + disable_acpi(); + } else { + printk(KERN_NOTICE + "Warning: DMI blacklist says broken, but acpi forced\n"); + } + return 0; +} + +/* + * Limit ACPI to CPU enumeration for HT + */ +static __init __attribute__((unused)) int force_acpi_ht(struct dmi_blacklist *d) +{ + if (!acpi_force) { + printk(KERN_NOTICE "%s detected: force use of acpi=ht\n", d->ident); + disable_acpi(); + acpi_ht = 1; + } else { + printk(KERN_NOTICE + "Warning: acpi=force overrules DMI blacklist: acpi=ht\n"); + } + return 0; +} +#endif + +#ifdef CONFIG_ACPI_PCI +static __init int disable_acpi_irq(struct dmi_blacklist *d) +{ + if (!acpi_force) { + printk(KERN_NOTICE "%s detected: force use of acpi=noirq\n", + d->ident); + acpi_noirq_set(); + } + return 0; +} +static __init int disable_acpi_pci(struct dmi_blacklist *d) +{ + if (!acpi_force) { + printk(KERN_NOTICE "%s detected: force use of pci=noacpi\n", + d->ident); + acpi_disable_pci(); + } + return 0; +} +#endif + +/* + * Process the DMI blacklists + */ + + +/* + * This will be expanded over time to force things like the APM + * interrupt mask settings according to the laptop + */ + +static __initdata struct dmi_blacklist dmi_blacklist[]={ + + { broken_toshiba_keyboard, "Toshiba Satellite 4030cdt", { /* Keyboard generates spurious repeats */ + MATCH(DMI_PRODUCT_NAME, "S4030CDT/4.3"), + NO_MATCH, NO_MATCH, NO_MATCH + } }, +#ifdef CONFIG_ACPI_SLEEP + { reset_videomode_after_s3, "Toshiba Satellite 4030cdt", { /* Reset video mode after returning from ACPI S3 sleep */ + MATCH(DMI_PRODUCT_NAME, "S4030CDT/4.3"), + NO_MATCH, NO_MATCH, NO_MATCH + } }, +#endif + +#ifdef CONFIG_ACPI_BOOT + /* + * If your system is blacklisted here, but you find that acpi=force + * works for you, please contact acpi-devel@sourceforge.net + */ + + /* + * Boxes that need ACPI disabled + */ + + { dmi_disable_acpi, "IBM Thinkpad", { + MATCH(DMI_BOARD_VENDOR, "IBM"), + MATCH(DMI_BOARD_NAME, "2629H1G"), + NO_MATCH, NO_MATCH }}, + + /* + * Boxes that need acpi=ht + */ + + { force_acpi_ht, "FSC Primergy T850", { + MATCH(DMI_SYS_VENDOR, "FUJITSU SIEMENS"), + MATCH(DMI_PRODUCT_NAME, "PRIMERGY T850"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "DELL GX240", { + MATCH(DMI_BOARD_VENDOR, "Dell Computer Corporation"), + MATCH(DMI_BOARD_NAME, "OptiPlex GX240"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "HP VISUALIZE NT Workstation", { + MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"), + MATCH(DMI_PRODUCT_NAME, "HP VISUALIZE NT Workstation"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "Compaq Workstation W8000", { + MATCH(DMI_SYS_VENDOR, "Compaq"), + MATCH(DMI_PRODUCT_NAME, "Workstation W8000"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "ASUS P4B266", { + MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."), + MATCH(DMI_BOARD_NAME, "P4B266"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "ASUS P2B-DS", { + MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."), + MATCH(DMI_BOARD_NAME, "P2B-DS"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "ASUS CUR-DLS", { + MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."), + MATCH(DMI_BOARD_NAME, "CUR-DLS"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "ABIT i440BX-W83977", { + MATCH(DMI_BOARD_VENDOR, "ABIT <http://www.abit.com>"), + MATCH(DMI_BOARD_NAME, "i440BX-W83977 (BP6)"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "IBM Bladecenter", { + MATCH(DMI_BOARD_VENDOR, "IBM"), + MATCH(DMI_BOARD_NAME, "IBM eServer BladeCenter HS20"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "IBM eServer xSeries 360", { + MATCH(DMI_BOARD_VENDOR, "IBM"), + MATCH(DMI_BOARD_NAME, "eServer xSeries 360"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "IBM eserver xSeries 330", { + MATCH(DMI_BOARD_VENDOR, "IBM"), + MATCH(DMI_BOARD_NAME, "eserver xSeries 330"), + NO_MATCH, NO_MATCH }}, + + { force_acpi_ht, "IBM eserver xSeries 440", { + MATCH(DMI_BOARD_VENDOR, "IBM"), + MATCH(DMI_PRODUCT_NAME, "eserver xSeries 440"), + NO_MATCH, NO_MATCH }}, + +#endif // CONFIG_ACPI_BOOT + +#ifdef CONFIG_ACPI_PCI + /* + * Boxes that need ACPI PCI IRQ routing disabled + */ + + { disable_acpi_irq, "ASUS A7V", { + MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC"), + MATCH(DMI_BOARD_NAME, "<A7V>"), + /* newer BIOS, Revision 1011, does work */ + MATCH(DMI_BIOS_VERSION, "ASUS A7V ACPI BIOS Revision 1007"), + NO_MATCH }}, + + /* + * Boxes that need ACPI PCI IRQ routing and PCI scan disabled + */ + { disable_acpi_pci, "ASUS PR-DLS", { /* _BBN 0 bug */ + MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."), + MATCH(DMI_BOARD_NAME, "PR-DLS"), + MATCH(DMI_BIOS_VERSION, "ASUS PR-DLS ACPI BIOS Revision 1010"), + MATCH(DMI_BIOS_DATE, "03/21/2003") }}, + + { disable_acpi_pci, "Acer TravelMate 36x Laptop", { + MATCH(DMI_SYS_VENDOR, "Acer"), + MATCH(DMI_PRODUCT_NAME, "TravelMate 360"), + NO_MATCH, NO_MATCH + } }, + +#endif + + { NULL, } +}; + +/* + * Process a DMI table entry. Right now all we care about are the BIOS + * and machine entries. For 2.5 we should pull the smbus controller info + * out of here. + */ + +static void __init dmi_decode(struct dmi_header *dm) +{ +#ifdef DMI_DEBUG + u8 *data = (u8 *)dm; +#endif + + switch(dm->type) + { + case 0: + dmi_printk(("BIOS Vendor: %s\n", + dmi_string(dm, data[4]))); + dmi_save_ident(dm, DMI_BIOS_VENDOR, 4); + dmi_printk(("BIOS Version: %s\n", + dmi_string(dm, data[5]))); + dmi_save_ident(dm, DMI_BIOS_VERSION, 5); + dmi_printk(("BIOS Release: %s\n", + dmi_string(dm, data[8]))); + dmi_save_ident(dm, DMI_BIOS_DATE, 8); + break; + case 1: + dmi_printk(("System Vendor: %s\n", + dmi_string(dm, data[4]))); + dmi_save_ident(dm, DMI_SYS_VENDOR, 4); + dmi_printk(("Product Name: %s\n", + dmi_string(dm, data[5]))); + dmi_save_ident(dm, DMI_PRODUCT_NAME, 5); + dmi_printk(("Version: %s\n", + dmi_string(dm, data[6]))); + dmi_save_ident(dm, DMI_PRODUCT_VERSION, 6); + dmi_printk(("Serial Number: %s\n", + dmi_string(dm, data[7]))); + break; + case 2: + dmi_printk(("Board Vendor: %s\n", + dmi_string(dm, data[4]))); + dmi_save_ident(dm, DMI_BOARD_VENDOR, 4); + dmi_printk(("Board Name: %s\n", + dmi_string(dm, data[5]))); + dmi_save_ident(dm, DMI_BOARD_NAME, 5); + dmi_printk(("Board Version: %s\n", + dmi_string(dm, data[6]))); + dmi_save_ident(dm, DMI_BOARD_VERSION, 6); + break; + } +} + +void __init dmi_scan_machine(void) +{ + int err = dmi_iterate(dmi_decode); + if(err == 0) + dmi_check_system(dmi_blacklist); + else + printk(KERN_INFO "DMI not present.\n"); +} + + +/** + * dmi_check_system - check system DMI data + * @list: array of dmi_system_id structures to match against + * + * Walk the blacklist table running matching functions until someone + * returns non zero or we hit the end. Callback function is called for + * each successfull match. Returns the number of matches. + */ +int dmi_check_system(struct dmi_system_id *list) +{ + int i, count = 0; + struct dmi_system_id *d = list; + + while (d->ident) { + for (i = 0; i < ARRAY_SIZE(d->matches); i++) { + int s = d->matches[i].slot; + if (s == DMI_NONE) + continue; + if (dmi_ident[s] && strstr(dmi_ident[s], d->matches[i].substr)) + continue; + /* No match */ + goto fail; + } + if (d->callback && d->callback(d)) + break; + count++; +fail: d++; + } + + return count; +} + +EXPORT_SYMBOL(dmi_check_system); + +/** + * dmi_get_system_info - return DMI data value + * @field: data index (see enum dmi_filed) + * + * Returns one DMI data value, can be used to perform + * complex DMI data checks. + */ +char * dmi_get_system_info(int field) +{ + return dmi_ident[field]; +} + diff --git a/arch/i386/kernel/doublefault.c b/arch/i386/kernel/doublefault.c new file mode 100644 index 00000000000..789af3e9fb1 --- /dev/null +++ b/arch/i386/kernel/doublefault.c @@ -0,0 +1,65 @@ +#include <linux/mm.h> +#include <linux/sched.h> +#include <linux/init.h> +#include <linux/init_task.h> +#include <linux/fs.h> + +#include <asm/uaccess.h> +#include <asm/pgtable.h> +#include <asm/processor.h> +#include <asm/desc.h> + +#define DOUBLEFAULT_STACKSIZE (1024) +static unsigned long doublefault_stack[DOUBLEFAULT_STACKSIZE]; +#define STACK_START (unsigned long)(doublefault_stack+DOUBLEFAULT_STACKSIZE) + +#define ptr_ok(x) ((x) > PAGE_OFFSET && (x) < PAGE_OFFSET + 0x1000000) + +static void doublefault_fn(void) +{ + struct Xgt_desc_struct gdt_desc = {0, 0}; + unsigned long gdt, tss; + + __asm__ __volatile__("sgdt %0": "=m" (gdt_desc): :"memory"); + gdt = gdt_desc.address; + + printk("double fault, gdt at %08lx [%d bytes]\n", gdt, gdt_desc.size); + + if (ptr_ok(gdt)) { + gdt += GDT_ENTRY_TSS << 3; + tss = *(u16 *)(gdt+2); + tss += *(u8 *)(gdt+4) << 16; + tss += *(u8 *)(gdt+7) << 24; + printk("double fault, tss at %08lx\n", tss); + + if (ptr_ok(tss)) { + struct tss_struct *t = (struct tss_struct *)tss; + + printk("eip = %08lx, esp = %08lx\n", t->eip, t->esp); + + printk("eax = %08lx, ebx = %08lx, ecx = %08lx, edx = %08lx\n", + t->eax, t->ebx, t->ecx, t->edx); + printk("esi = %08lx, edi = %08lx\n", + t->esi, t->edi); + } + } + + for (;;) /* nothing */; +} + +struct tss_struct doublefault_tss __cacheline_aligned = { + .esp0 = STACK_START, + .ss0 = __KERNEL_DS, + .ldt = 0, + .io_bitmap_base = INVALID_IO_BITMAP_OFFSET, + + .eip = (unsigned long) doublefault_fn, + .eflags = X86_EFLAGS_SF | 0x2, /* 0x2 bit is always set */ + .esp = STACK_START, + .es = __USER_DS, + .cs = __KERNEL_CS, + .ss = __KERNEL_DS, + .ds = __USER_DS, + + .__cr3 = __pa(swapper_pg_dir) +}; diff --git a/arch/i386/kernel/early_printk.c b/arch/i386/kernel/early_printk.c new file mode 100644 index 00000000000..92f812ba275 --- /dev/null +++ b/arch/i386/kernel/early_printk.c @@ -0,0 +1,2 @@ + +#include "../../x86_64/kernel/early_printk.c" diff --git a/arch/i386/kernel/efi.c b/arch/i386/kernel/efi.c new file mode 100644 index 00000000000..9e5e0d8bd36 --- /dev/null +++ b/arch/i386/kernel/efi.c @@ -0,0 +1,635 @@ +/* + * Extensible Firmware Interface + * + * Based on Extensible Firmware Interface Specification version 1.0 + * + * Copyright (C) 1999 VA Linux Systems + * Copyright (C) 1999 Walt Drummond <drummond@valinux.com> + * Copyright (C) 1999-2002 Hewlett-Packard Co. + * David Mosberger-Tang <davidm@hpl.hp.com> + * Stephane Eranian <eranian@hpl.hp.com> + * + * All EFI Runtime Services are not implemented yet as EFI only + * supports physical mode addressing on SoftSDV. This is to be fixed + * in a future version. --drummond 1999-07-20 + * + * Implemented EFI runtime services and virtual mode calls. --davidm + * + * Goutham Rao: <goutham.rao@intel.com> + * Skip non-WB memory and ignore empty memory ranges. + */ + +#include <linux/config.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/mm.h> +#include <linux/types.h> +#include <linux/time.h> +#include <linux/spinlock.h> +#include <linux/bootmem.h> +#include <linux/ioport.h> +#include <linux/module.h> +#include <linux/efi.h> + +#include <asm/setup.h> +#include <asm/io.h> +#include <asm/page.h> +#include <asm/pgtable.h> +#include <asm/processor.h> +#include <asm/desc.h> +#include <asm/tlbflush.h> + +#define EFI_DEBUG 0 +#define PFX "EFI: " + +extern efi_status_t asmlinkage efi_call_phys(void *, ...); + +struct efi efi; +EXPORT_SYMBOL(efi); +static struct efi efi_phys __initdata; +struct efi_memory_map memmap __initdata; + +/* + * We require an early boot_ioremap mapping mechanism initially + */ +extern void * boot_ioremap(unsigned long, unsigned long); + +/* + * To make EFI call EFI runtime service in physical addressing mode we need + * prelog/epilog before/after the invocation to disable interrupt, to + * claim EFI runtime service handler exclusively and to duplicate a memory in + * low memory space say 0 - 3G. + */ + +static unsigned long efi_rt_eflags; +static DEFINE_SPINLOCK(efi_rt_lock); +static pgd_t efi_bak_pg_dir_pointer[2]; + +static void efi_call_phys_prelog(void) +{ + unsigned long cr4; + unsigned long temp; + + spin_lock(&efi_rt_lock); + local_irq_save(efi_rt_eflags); + + /* + * If I don't have PSE, I should just duplicate two entries in page + * directory. If I have PSE, I just need to duplicate one entry in + * page directory. + */ + __asm__ __volatile__("movl %%cr4, %0":"=r"(cr4)); + + if (cr4 & X86_CR4_PSE) { + efi_bak_pg_dir_pointer[0].pgd = + swapper_pg_dir[pgd_index(0)].pgd; + swapper_pg_dir[0].pgd = + swapper_pg_dir[pgd_index(PAGE_OFFSET)].pgd; + } else { + efi_bak_pg_dir_pointer[0].pgd = + swapper_pg_dir[pgd_index(0)].pgd; + efi_bak_pg_dir_pointer[1].pgd = + swapper_pg_dir[pgd_index(0x400000)].pgd; + swapper_pg_dir[pgd_index(0)].pgd = + swapper_pg_dir[pgd_index(PAGE_OFFSET)].pgd; + temp = PAGE_OFFSET + 0x400000; + swapper_pg_dir[pgd_index(0x400000)].pgd = + swapper_pg_dir[pgd_index(temp)].pgd; + } + + /* + * After the lock is released, the original page table is restored. + */ + local_flush_tlb(); + + cpu_gdt_descr[0].address = __pa(cpu_gdt_descr[0].address); + __asm__ __volatile__("lgdt %0":"=m" + (*(struct Xgt_desc_struct *) __pa(&cpu_gdt_descr[0]))); +} + +static void efi_call_phys_epilog(void) +{ + unsigned long cr4; + + cpu_gdt_descr[0].address = + (unsigned long) __va(cpu_gdt_descr[0].address); + __asm__ __volatile__("lgdt %0":"=m"(cpu_gdt_descr)); + __asm__ __volatile__("movl %%cr4, %0":"=r"(cr4)); + + if (cr4 & X86_CR4_PSE) { + swapper_pg_dir[pgd_index(0)].pgd = + efi_bak_pg_dir_pointer[0].pgd; + } else { + swapper_pg_dir[pgd_index(0)].pgd = + efi_bak_pg_dir_pointer[0].pgd; + swapper_pg_dir[pgd_index(0x400000)].pgd = + efi_bak_pg_dir_pointer[1].pgd; + } + + /* + * After the lock is released, the original page table is restored. + */ + local_flush_tlb(); + + local_irq_restore(efi_rt_eflags); + spin_unlock(&efi_rt_lock); +} + +static efi_status_t +phys_efi_set_virtual_address_map(unsigned long memory_map_size, + unsigned long descriptor_size, + u32 descriptor_version, + efi_memory_desc_t *virtual_map) +{ + efi_status_t status; + + efi_call_phys_prelog(); + status = efi_call_phys(efi_phys.set_virtual_address_map, + memory_map_size, descriptor_size, + descriptor_version, virtual_map); + efi_call_phys_epilog(); + return status; +} + +static efi_status_t +phys_efi_get_time(efi_time_t *tm, efi_time_cap_t *tc) +{ + efi_status_t status; + + efi_call_phys_prelog(); + status = efi_call_phys(efi_phys.get_time, tm, tc); + efi_call_phys_epilog(); + return status; +} + +inline int efi_set_rtc_mmss(unsigned long nowtime) +{ + int real_seconds, real_minutes; + efi_status_t status; + efi_time_t eft; + efi_time_cap_t cap; + + spin_lock(&efi_rt_lock); + status = efi.get_time(&eft, &cap); + spin_unlock(&efi_rt_lock); + if (status != EFI_SUCCESS) + panic("Ooops, efitime: can't read time!\n"); + real_seconds = nowtime % 60; + real_minutes = nowtime / 60; + + if (((abs(real_minutes - eft.minute) + 15)/30) & 1) + real_minutes += 30; + real_minutes %= 60; + + eft.minute = real_minutes; + eft.second = real_seconds; + + if (status != EFI_SUCCESS) { + printk("Ooops: efitime: can't read time!\n"); + return -1; + } + return 0; +} +/* + * This should only be used during kernel init and before runtime + * services have been remapped, therefore, we'll need to call in physical + * mode. Note, this call isn't used later, so mark it __init. + */ +inline unsigned long __init efi_get_time(void) +{ + efi_status_t status; + efi_time_t eft; + efi_time_cap_t cap; + + status = phys_efi_get_time(&eft, &cap); + if (status != EFI_SUCCESS) + printk("Oops: efitime: can't read time status: 0x%lx\n",status); + + return mktime(eft.year, eft.month, eft.day, eft.hour, + eft.minute, eft.second); +} + +int is_available_memory(efi_memory_desc_t * md) +{ + if (!(md->attribute & EFI_MEMORY_WB)) + return 0; + + switch (md->type) { + case EFI_LOADER_CODE: + case EFI_LOADER_DATA: + case EFI_BOOT_SERVICES_CODE: + case EFI_BOOT_SERVICES_DATA: + case EFI_CONVENTIONAL_MEMORY: + return 1; + } + return 0; +} + +/* + * We need to map the EFI memory map again after paging_init(). + */ +void __init efi_map_memmap(void) +{ + memmap.map = NULL; + + memmap.map = (efi_memory_desc_t *) + bt_ioremap((unsigned long) memmap.phys_map, + (memmap.nr_map * sizeof(efi_memory_desc_t))); + + if (memmap.map == NULL) + printk(KERN_ERR PFX "Could not remap the EFI memmap!\n"); +} + +#if EFI_DEBUG +static void __init print_efi_memmap(void) +{ + efi_memory_desc_t *md; + int i; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + printk(KERN_INFO "mem%02u: type=%u, attr=0x%llx, " + "range=[0x%016llx-0x%016llx) (%lluMB)\n", + i, md->type, md->attribute, md->phys_addr, + md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT), + (md->num_pages >> (20 - EFI_PAGE_SHIFT))); + } +} +#endif /* EFI_DEBUG */ + +/* + * Walks the EFI memory map and calls CALLBACK once for each EFI + * memory descriptor that has memory that is available for kernel use. + */ +void efi_memmap_walk(efi_freemem_callback_t callback, void *arg) +{ + int prev_valid = 0; + struct range { + unsigned long start; + unsigned long end; + } prev, curr; + efi_memory_desc_t *md; + unsigned long start, end; + int i; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + + if ((md->num_pages == 0) || (!is_available_memory(md))) + continue; + + curr.start = md->phys_addr; + curr.end = curr.start + (md->num_pages << EFI_PAGE_SHIFT); + + if (!prev_valid) { + prev = curr; + prev_valid = 1; + } else { + if (curr.start < prev.start) + printk(KERN_INFO PFX "Unordered memory map\n"); + if (prev.end == curr.start) + prev.end = curr.end; + else { + start = + (unsigned long) (PAGE_ALIGN(prev.start)); + end = (unsigned long) (prev.end & PAGE_MASK); + if ((end > start) + && (*callback) (start, end, arg) < 0) + return; + prev = curr; + } + } + } + if (prev_valid) { + start = (unsigned long) PAGE_ALIGN(prev.start); + end = (unsigned long) (prev.end & PAGE_MASK); + if (end > start) + (*callback) (start, end, arg); + } +} + +void __init efi_init(void) +{ + efi_config_table_t *config_tables; + efi_runtime_services_t *runtime; + efi_char16_t *c16; + char vendor[100] = "unknown"; + unsigned long num_config_tables; + int i = 0; + + memset(&efi, 0, sizeof(efi) ); + memset(&efi_phys, 0, sizeof(efi_phys)); + + efi_phys.systab = EFI_SYSTAB; + memmap.phys_map = EFI_MEMMAP; + memmap.nr_map = EFI_MEMMAP_SIZE/EFI_MEMDESC_SIZE; + memmap.desc_version = EFI_MEMDESC_VERSION; + + efi.systab = (efi_system_table_t *) + boot_ioremap((unsigned long) efi_phys.systab, + sizeof(efi_system_table_t)); + /* + * Verify the EFI Table + */ + if (efi.systab == NULL) + printk(KERN_ERR PFX "Woah! Couldn't map the EFI system table.\n"); + if (efi.systab->hdr.signature != EFI_SYSTEM_TABLE_SIGNATURE) + printk(KERN_ERR PFX "Woah! EFI system table signature incorrect\n"); + if ((efi.systab->hdr.revision ^ EFI_SYSTEM_TABLE_REVISION) >> 16 != 0) + printk(KERN_ERR PFX + "Warning: EFI system table major version mismatch: " + "got %d.%02d, expected %d.%02d\n", + efi.systab->hdr.revision >> 16, + efi.systab->hdr.revision & 0xffff, + EFI_SYSTEM_TABLE_REVISION >> 16, + EFI_SYSTEM_TABLE_REVISION & 0xffff); + /* + * Grab some details from the system table + */ + num_config_tables = efi.systab->nr_tables; + config_tables = (efi_config_table_t *)efi.systab->tables; + runtime = efi.systab->runtime; + + /* + * Show what we know for posterity + */ + c16 = (efi_char16_t *) boot_ioremap(efi.systab->fw_vendor, 2); + if (c16) { + for (i = 0; i < sizeof(vendor) && *c16; ++i) + vendor[i] = *c16++; + vendor[i] = '\0'; + } else + printk(KERN_ERR PFX "Could not map the firmware vendor!\n"); + + printk(KERN_INFO PFX "EFI v%u.%.02u by %s \n", + efi.systab->hdr.revision >> 16, + efi.systab->hdr.revision & 0xffff, vendor); + + /* + * Let's see what config tables the firmware passed to us. + */ + config_tables = (efi_config_table_t *) + boot_ioremap((unsigned long) config_tables, + num_config_tables * sizeof(efi_config_table_t)); + + if (config_tables == NULL) + printk(KERN_ERR PFX "Could not map EFI Configuration Table!\n"); + + for (i = 0; i < num_config_tables; i++) { + if (efi_guidcmp(config_tables[i].guid, MPS_TABLE_GUID) == 0) { + efi.mps = (void *)config_tables[i].table; + printk(KERN_INFO " MPS=0x%lx ", config_tables[i].table); + } else + if (efi_guidcmp(config_tables[i].guid, ACPI_20_TABLE_GUID) == 0) { + efi.acpi20 = __va(config_tables[i].table); + printk(KERN_INFO " ACPI 2.0=0x%lx ", config_tables[i].table); + } else + if (efi_guidcmp(config_tables[i].guid, ACPI_TABLE_GUID) == 0) { + efi.acpi = __va(config_tables[i].table); + printk(KERN_INFO " ACPI=0x%lx ", config_tables[i].table); + } else + if (efi_guidcmp(config_tables[i].guid, SMBIOS_TABLE_GUID) == 0) { + efi.smbios = (void *) config_tables[i].table; + printk(KERN_INFO " SMBIOS=0x%lx ", config_tables[i].table); + } else + if (efi_guidcmp(config_tables[i].guid, HCDP_TABLE_GUID) == 0) { + efi.hcdp = (void *)config_tables[i].table; + printk(KERN_INFO " HCDP=0x%lx ", config_tables[i].table); + } else + if (efi_guidcmp(config_tables[i].guid, UGA_IO_PROTOCOL_GUID) == 0) { + efi.uga = (void *)config_tables[i].table; + printk(KERN_INFO " UGA=0x%lx ", config_tables[i].table); + } + } + printk("\n"); + + /* + * Check out the runtime services table. We need to map + * the runtime services table so that we can grab the physical + * address of several of the EFI runtime functions, needed to + * set the firmware into virtual mode. + */ + + runtime = (efi_runtime_services_t *) boot_ioremap((unsigned long) + runtime, + sizeof(efi_runtime_services_t)); + if (runtime != NULL) { + /* + * We will only need *early* access to the following + * two EFI runtime services before set_virtual_address_map + * is invoked. + */ + efi_phys.get_time = (efi_get_time_t *) runtime->get_time; + efi_phys.set_virtual_address_map = + (efi_set_virtual_address_map_t *) + runtime->set_virtual_address_map; + } else + printk(KERN_ERR PFX "Could not map the runtime service table!\n"); + + /* Map the EFI memory map for use until paging_init() */ + + memmap.map = (efi_memory_desc_t *) + boot_ioremap((unsigned long) EFI_MEMMAP, EFI_MEMMAP_SIZE); + + if (memmap.map == NULL) + printk(KERN_ERR PFX "Could not map the EFI memory map!\n"); + + if (EFI_MEMDESC_SIZE != sizeof(efi_memory_desc_t)) { + printk(KERN_WARNING PFX "Warning! Kernel-defined memdesc doesn't " + "match the one from EFI!\n"); + } +#if EFI_DEBUG + print_efi_memmap(); +#endif +} + +/* + * This function will switch the EFI runtime services to virtual mode. + * Essentially, look through the EFI memmap and map every region that + * has the runtime attribute bit set in its memory descriptor and update + * that memory descriptor with the virtual address obtained from ioremap(). + * This enables the runtime services to be called without having to + * thunk back into physical mode for every invocation. + */ + +void __init efi_enter_virtual_mode(void) +{ + efi_memory_desc_t *md; + efi_status_t status; + int i; + + efi.systab = NULL; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + + if (md->attribute & EFI_MEMORY_RUNTIME) { + md->virt_addr = + (unsigned long)ioremap(md->phys_addr, + md->num_pages << EFI_PAGE_SHIFT); + if (!(unsigned long)md->virt_addr) { + printk(KERN_ERR PFX "ioremap of 0x%lX failed\n", + (unsigned long)md->phys_addr); + } + + if (((unsigned long)md->phys_addr <= + (unsigned long)efi_phys.systab) && + ((unsigned long)efi_phys.systab < + md->phys_addr + + ((unsigned long)md->num_pages << + EFI_PAGE_SHIFT))) { + unsigned long addr; + + addr = md->virt_addr - md->phys_addr + + (unsigned long)efi_phys.systab; + efi.systab = (efi_system_table_t *)addr; + } + } + } + + if (!efi.systab) + BUG(); + + status = phys_efi_set_virtual_address_map( + sizeof(efi_memory_desc_t) * memmap.nr_map, + sizeof(efi_memory_desc_t), + memmap.desc_version, + memmap.phys_map); + + if (status != EFI_SUCCESS) { + printk (KERN_ALERT "You are screwed! " + "Unable to switch EFI into virtual mode " + "(status=%lx)\n", status); + panic("EFI call to SetVirtualAddressMap() failed!"); + } + + /* + * Now that EFI is in virtual mode, update the function + * pointers in the runtime service table to the new virtual addresses. + */ + + efi.get_time = (efi_get_time_t *) efi.systab->runtime->get_time; + efi.set_time = (efi_set_time_t *) efi.systab->runtime->set_time; + efi.get_wakeup_time = (efi_get_wakeup_time_t *) + efi.systab->runtime->get_wakeup_time; + efi.set_wakeup_time = (efi_set_wakeup_time_t *) + efi.systab->runtime->set_wakeup_time; + efi.get_variable = (efi_get_variable_t *) + efi.systab->runtime->get_variable; + efi.get_next_variable = (efi_get_next_variable_t *) + efi.systab->runtime->get_next_variable; + efi.set_variable = (efi_set_variable_t *) + efi.systab->runtime->set_variable; + efi.get_next_high_mono_count = (efi_get_next_high_mono_count_t *) + efi.systab->runtime->get_next_high_mono_count; + efi.reset_system = (efi_reset_system_t *) + efi.systab->runtime->reset_system; +} + +void __init +efi_initialize_iomem_resources(struct resource *code_resource, + struct resource *data_resource) +{ + struct resource *res; + efi_memory_desc_t *md; + int i; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + + if ((md->phys_addr + (md->num_pages << EFI_PAGE_SHIFT)) > + 0x100000000ULL) + continue; + res = alloc_bootmem_low(sizeof(struct resource)); + switch (md->type) { + case EFI_RESERVED_TYPE: + res->name = "Reserved Memory"; + break; + case EFI_LOADER_CODE: + res->name = "Loader Code"; + break; + case EFI_LOADER_DATA: + res->name = "Loader Data"; + break; + case EFI_BOOT_SERVICES_DATA: + res->name = "BootServices Data"; + break; + case EFI_BOOT_SERVICES_CODE: + res->name = "BootServices Code"; + break; + case EFI_RUNTIME_SERVICES_CODE: + res->name = "Runtime Service Code"; + break; + case EFI_RUNTIME_SERVICES_DATA: + res->name = "Runtime Service Data"; + break; + case EFI_CONVENTIONAL_MEMORY: + res->name = "Conventional Memory"; + break; + case EFI_UNUSABLE_MEMORY: + res->name = "Unusable Memory"; + break; + case EFI_ACPI_RECLAIM_MEMORY: + res->name = "ACPI Reclaim"; + break; + case EFI_ACPI_MEMORY_NVS: + res->name = "ACPI NVS"; + break; + case EFI_MEMORY_MAPPED_IO: + res->name = "Memory Mapped IO"; + break; + case EFI_MEMORY_MAPPED_IO_PORT_SPACE: + res->name = "Memory Mapped IO Port Space"; + break; + default: + res->name = "Reserved"; + break; + } + res->start = md->phys_addr; + res->end = res->start + ((md->num_pages << EFI_PAGE_SHIFT) - 1); + res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; + if (request_resource(&iomem_resource, res) < 0) + printk(KERN_ERR PFX "Failed to allocate res %s : 0x%lx-0x%lx\n", + res->name, res->start, res->end); + /* + * We don't know which region contains kernel data so we try + * it repeatedly and let the resource manager test it. + */ + if (md->type == EFI_CONVENTIONAL_MEMORY) { + request_resource(res, code_resource); + request_resource(res, data_resource); + } + } +} + +/* + * Convenience functions to obtain memory types and attributes + */ + +u32 efi_mem_type(unsigned long phys_addr) +{ + efi_memory_desc_t *md; + int i; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + if ((md->phys_addr <= phys_addr) && (phys_addr < + (md->phys_addr + (md-> num_pages << EFI_PAGE_SHIFT)) )) + return md->type; + } + return 0; +} + +u64 efi_mem_attributes(unsigned long phys_addr) +{ + efi_memory_desc_t *md; + int i; + + for (i = 0; i < memmap.nr_map; i++) { + md = &memmap.map[i]; + if ((md->phys_addr <= phys_addr) && (phys_addr < + (md->phys_addr + (md-> num_pages << EFI_PAGE_SHIFT)) )) + return md->attribute; + } + return 0; +} diff --git a/arch/i386/kernel/efi_stub.S b/arch/i386/kernel/efi_stub.S new file mode 100644 index 00000000000..08c0312d9b6 --- /dev/null +++ b/arch/i386/kernel/efi_stub.S @@ -0,0 +1,124 @@ +/* + * EFI call stub for IA32. + * + * This stub allows us to make EFI calls in physical mode with interrupts + * turned off. + */ + +#include <linux/config.h> +#include <linux/linkage.h> +#include <asm/page.h> +#include <asm/pgtable.h> + +/* + * efi_call_phys(void *, ...) is a function with variable parameters. + * All the callers of this function assure that all the parameters are 4-bytes. + */ + +/* + * In gcc calling convention, EBX, ESP, EBP, ESI and EDI are all callee save. + * So we'd better save all of them at the beginning of this function and restore + * at the end no matter how many we use, because we can not assure EFI runtime + * service functions will comply with gcc calling convention, too. + */ + +.text +ENTRY(efi_call_phys) + /* + * 0. The function can only be called in Linux kernel. So CS has been + * set to 0x0010, DS and SS have been set to 0x0018. In EFI, I found + * the values of these registers are the same. And, the corresponding + * GDT entries are identical. So I will do nothing about segment reg + * and GDT, but change GDT base register in prelog and epilog. + */ + + /* + * 1. Now I am running with EIP = <physical address> + PAGE_OFFSET. + * But to make it smoothly switch from virtual mode to flat mode. + * The mapping of lower virtual memory has been created in prelog and + * epilog. + */ + movl $1f, %edx + subl $__PAGE_OFFSET, %edx + jmp *%edx +1: + + /* + * 2. Now on the top of stack is the return + * address in the caller of efi_call_phys(), then parameter 1, + * parameter 2, ..., param n. To make things easy, we save the return + * address of efi_call_phys in a global variable. + */ + popl %edx + movl %edx, saved_return_addr + /* get the function pointer into ECX*/ + popl %ecx + movl %ecx, efi_rt_function_ptr + movl $2f, %edx + subl $__PAGE_OFFSET, %edx + pushl %edx + + /* + * 3. Clear PG bit in %CR0. + */ + movl %cr0, %edx + andl $0x7fffffff, %edx + movl %edx, %cr0 + jmp 1f +1: + + /* + * 4. Adjust stack pointer. + */ + subl $__PAGE_OFFSET, %esp + + /* + * 5. Call the physical function. + */ + jmp *%ecx + +2: + /* + * 6. After EFI runtime service returns, control will return to + * following instruction. We'd better readjust stack pointer first. + */ + addl $__PAGE_OFFSET, %esp + + /* + * 7. Restore PG bit + */ + movl %cr0, %edx + orl $0x80000000, %edx + movl %edx, %cr0 + jmp 1f +1: + /* + * 8. Now restore the virtual mode from flat mode by + * adding EIP with PAGE_OFFSET. + */ + movl $1f, %edx + jmp *%edx +1: + + /* + * 9. Balance the stack. And because EAX contain the return value, + * we'd better not clobber it. + */ + leal efi_rt_function_ptr, %edx + movl (%edx), %ecx + pushl %ecx + + /* + * 10. Push the saved return address onto the stack and return. + */ + leal saved_return_addr, %edx + movl (%edx), %ecx + pushl %ecx + ret +.previous + +.data +saved_return_addr: + .long 0 +efi_rt_function_ptr: + .long 0 diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S new file mode 100644 index 00000000000..1e45ff292bc --- /dev/null +++ b/arch/i386/kernel/entry.S @@ -0,0 +1,950 @@ +/* + * linux/arch/i386/entry.S + * + * Copyright (C) 1991, 1992 Linus Torvalds + */ + +/* + * entry.S contains the system-call and fault low-level handling routines. + * This also contains the timer-interrupt handler, as well as all interrupts + * and faults that can result in a task-switch. + * + * NOTE: This code handles signal-recognition, which happens every time + * after a timer-interrupt and after each system call. + * + * I changed all the .align's to 4 (16 byte alignment), as that's faster + * on a 486. + * + * Stack layout in 'ret_from_system_call': + * ptrace needs to have all regs on the stack. + * if the order here is changed, it needs to be + * updated in fork.c:copy_process, signal.c:do_signal, + * ptrace.c and ptrace.h + * + * 0(%esp) - %ebx + * 4(%esp) - %ecx + * 8(%esp) - %edx + * C(%esp) - %esi + * 10(%esp) - %edi + * 14(%esp) - %ebp + * 18(%esp) - %eax + * 1C(%esp) - %ds + * 20(%esp) - %es + * 24(%esp) - orig_eax + * 28(%esp) - %eip + * 2C(%esp) - %cs + * 30(%esp) - %eflags + * 34(%esp) - %oldesp + * 38(%esp) - %oldss + * + * "current" is in register %ebx during any slow entries. + */ + +#include <linux/config.h> +#include <linux/linkage.h> +#include <asm/thread_info.h> +#include <asm/errno.h> +#include <asm/segment.h> +#include <asm/smp.h> +#include <asm/page.h> +#include <asm/desc.h> +#include "irq_vectors.h" + +#define nr_syscalls ((syscall_table_size)/4) + +EBX = 0x00 +ECX = 0x04 +EDX = 0x08 +ESI = 0x0C +EDI = 0x10 +EBP = 0x14 +EAX = 0x18 +DS = 0x1C +ES = 0x20 +ORIG_EAX = 0x24 +EIP = 0x28 +CS = 0x2C +EFLAGS = 0x30 +OLDESP = 0x34 +OLDSS = 0x38 + +CF_MASK = 0x00000001 +TF_MASK = 0x00000100 +IF_MASK = 0x00000200 +DF_MASK = 0x00000400 +NT_MASK = 0x00004000 +VM_MASK = 0x00020000 + +#ifdef CONFIG_PREEMPT +#define preempt_stop cli +#else +#define preempt_stop +#define resume_kernel restore_nocheck +#endif + +#define SAVE_ALL \ + cld; \ + pushl %es; \ + pushl %ds; \ + pushl %eax; \ + pushl %ebp; \ + pushl %edi; \ + pushl %esi; \ + pushl %edx; \ + pushl %ecx; \ + pushl %ebx; \ + movl $(__USER_DS), %edx; \ + movl %edx, %ds; \ + movl %edx, %es; + +#define RESTORE_INT_REGS \ + popl %ebx; \ + popl %ecx; \ + popl %edx; \ + popl %esi; \ + popl %edi; \ + popl %ebp; \ + popl %eax + +#define RESTORE_REGS \ + RESTORE_INT_REGS; \ +1: popl %ds; \ +2: popl %es; \ +.section .fixup,"ax"; \ +3: movl $0,(%esp); \ + jmp 1b; \ +4: movl $0,(%esp); \ + jmp 2b; \ +.previous; \ +.section __ex_table,"a";\ + .align 4; \ + .long 1b,3b; \ + .long 2b,4b; \ +.previous + + +ENTRY(ret_from_fork) + pushl %eax + call schedule_tail + GET_THREAD_INFO(%ebp) + popl %eax + jmp syscall_exit + +/* + * Return to user mode is not as complex as all this looks, + * but we want the default path for a system call return to + * go as quickly as possible which is why some of this is + * less clear than it otherwise should be. + */ + + # userspace resumption stub bypassing syscall exit tracing + ALIGN +ret_from_exception: + preempt_stop +ret_from_intr: + GET_THREAD_INFO(%ebp) + movl EFLAGS(%esp), %eax # mix EFLAGS and CS + movb CS(%esp), %al + testl $(VM_MASK | 3), %eax + jz resume_kernel +ENTRY(resume_userspace) + cli # make sure we don't miss an interrupt + # setting need_resched or sigpending + # between sampling and the iret + movl TI_flags(%ebp), %ecx + andl $_TIF_WORK_MASK, %ecx # is there any work to be done on + # int/exception return? + jne work_pending + jmp restore_all + +#ifdef CONFIG_PREEMPT +ENTRY(resume_kernel) + cli + cmpl $0,TI_preempt_count(%ebp) # non-zero preempt_count ? + jnz restore_nocheck +need_resched: + movl TI_flags(%ebp), %ecx # need_resched set ? + testb $_TIF_NEED_RESCHED, %cl + jz restore_all + testl $IF_MASK,EFLAGS(%esp) # interrupts off (exception path) ? + jz restore_all + call preempt_schedule_irq + jmp need_resched +#endif + +/* SYSENTER_RETURN points to after the "sysenter" instruction in + the vsyscall page. See vsyscall-sysentry.S, which defines the symbol. */ + + # sysenter call handler stub +ENTRY(sysenter_entry) + movl TSS_sysenter_esp0(%esp),%esp +sysenter_past_esp: + sti + pushl $(__USER_DS) + pushl %ebp + pushfl + pushl $(__USER_CS) + pushl $SYSENTER_RETURN + +/* + * Load the potential sixth argument from user stack. + * Careful about security. + */ + cmpl $__PAGE_OFFSET-3,%ebp + jae syscall_fault +1: movl (%ebp),%ebp +.section __ex_table,"a" + .align 4 + .long 1b,syscall_fault +.previous + + pushl %eax + SAVE_ALL + GET_THREAD_INFO(%ebp) + + /* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */ + testw $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),TI_flags(%ebp) + jnz syscall_trace_entry + cmpl $(nr_syscalls), %eax + jae syscall_badsys + call *sys_call_table(,%eax,4) + movl %eax,EAX(%esp) + cli + movl TI_flags(%ebp), %ecx + testw $_TIF_ALLWORK_MASK, %cx + jne syscall_exit_work +/* if something modifies registers it must also disable sysexit */ + movl EIP(%esp), %edx + movl OLDESP(%esp), %ecx + xorl %ebp,%ebp + sti + sysexit + + + # system call handler stub +ENTRY(system_call) + pushl %eax # save orig_eax + SAVE_ALL + GET_THREAD_INFO(%ebp) + # system call tracing in operation + /* Note, _TIF_SECCOMP is bit number 8, and so it needs testw and not testb */ + testw $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP),TI_flags(%ebp) + jnz syscall_trace_entry + cmpl $(nr_syscalls), %eax + jae syscall_badsys +syscall_call: + call *sys_call_table(,%eax,4) + movl %eax,EAX(%esp) # store the return value +syscall_exit: + cli # make sure we don't miss an interrupt + # setting need_resched or sigpending + # between sampling and the iret + movl TI_flags(%ebp), %ecx + testw $_TIF_ALLWORK_MASK, %cx # current->work + jne syscall_exit_work + +restore_all: + movl EFLAGS(%esp), %eax # mix EFLAGS, SS and CS + movb OLDSS(%esp), %ah + movb CS(%esp), %al + andl $(VM_MASK | (4 << 8) | 3), %eax + cmpl $((4 << 8) | 3), %eax + je ldt_ss # returning to user-space with LDT SS +restore_nocheck: + RESTORE_REGS + addl $4, %esp +1: iret +.section .fixup,"ax" +iret_exc: + sti + movl $__USER_DS, %edx + movl %edx, %ds + movl %edx, %es + movl $11,%eax + call do_exit +.previous +.section __ex_table,"a" + .align 4 + .long 1b,iret_exc +.previous + +ldt_ss: + larl OLDSS(%esp), %eax + jnz restore_nocheck + testl $0x00400000, %eax # returning to 32bit stack? + jnz restore_nocheck # allright, normal return + /* If returning to userspace with 16bit stack, + * try to fix the higher word of ESP, as the CPU + * won't restore it. + * This is an "official" bug of all the x86-compatible + * CPUs, which we can try to work around to make + * dosemu and wine happy. */ + subl $8, %esp # reserve space for switch16 pointer + cli + movl %esp, %eax + /* Set up the 16bit stack frame with switch32 pointer on top, + * and a switch16 pointer on top of the current frame. */ + call setup_x86_bogus_stack + RESTORE_REGS + lss 20+4(%esp), %esp # switch to 16bit stack +1: iret +.section __ex_table,"a" + .align 4 + .long 1b,iret_exc +.previous + + # perform work that needs to be done immediately before resumption + ALIGN +work_pending: + testb $_TIF_NEED_RESCHED, %cl + jz work_notifysig +work_resched: + call schedule + cli # make sure we don't miss an interrupt + # setting need_resched or sigpending + # between sampling and the iret + movl TI_flags(%ebp), %ecx + andl $_TIF_WORK_MASK, %ecx # is there any work to be done other + # than syscall tracing? + jz restore_all + testb $_TIF_NEED_RESCHED, %cl + jnz work_resched + +work_notifysig: # deal with pending signals and + # notify-resume requests + testl $VM_MASK, EFLAGS(%esp) + movl %esp, %eax + jne work_notifysig_v86 # returning to kernel-space or + # vm86-space + xorl %edx, %edx + call do_notify_resume + jmp restore_all + + ALIGN +work_notifysig_v86: + pushl %ecx # save ti_flags for do_notify_resume + call save_v86_state # %eax contains pt_regs pointer + popl %ecx + movl %eax, %esp + xorl %edx, %edx + call do_notify_resume + jmp restore_all + + # perform syscall exit tracing + ALIGN +syscall_trace_entry: + movl $-ENOSYS,EAX(%esp) + movl %esp, %eax + xorl %edx,%edx + call do_syscall_trace + movl ORIG_EAX(%esp), %eax + cmpl $(nr_syscalls), %eax + jnae syscall_call + jmp syscall_exit + + # perform syscall exit tracing + ALIGN +syscall_exit_work: + testb $(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP), %cl + jz work_pending + sti # could let do_syscall_trace() call + # schedule() instead + movl %esp, %eax + movl $1, %edx + call do_syscall_trace + jmp resume_userspace + + ALIGN +syscall_fault: + pushl %eax # save orig_eax + SAVE_ALL + GET_THREAD_INFO(%ebp) + movl $-EFAULT,EAX(%esp) + jmp resume_userspace + + ALIGN +syscall_badsys: + movl $-ENOSYS,EAX(%esp) + jmp resume_userspace + +#define FIXUP_ESPFIX_STACK \ + movl %esp, %eax; \ + /* switch to 32bit stack using the pointer on top of 16bit stack */ \ + lss %ss:CPU_16BIT_STACK_SIZE-8, %esp; \ + /* copy data from 16bit stack to 32bit stack */ \ + call fixup_x86_bogus_stack; \ + /* put ESP to the proper location */ \ + movl %eax, %esp; +#define UNWIND_ESPFIX_STACK \ + pushl %eax; \ + movl %ss, %eax; \ + /* see if on 16bit stack */ \ + cmpw $__ESPFIX_SS, %ax; \ + jne 28f; \ + movl $__KERNEL_DS, %edx; \ + movl %edx, %ds; \ + movl %edx, %es; \ + /* switch to 32bit stack */ \ + FIXUP_ESPFIX_STACK \ +28: popl %eax; + +/* + * Build the entry stubs and pointer table with + * some assembler magic. + */ +.data +ENTRY(interrupt) +.text + +vector=0 +ENTRY(irq_entries_start) +.rept NR_IRQS + ALIGN +1: pushl $vector-256 + jmp common_interrupt +.data + .long 1b +.text +vector=vector+1 +.endr + + ALIGN +common_interrupt: + SAVE_ALL + movl %esp,%eax + call do_IRQ + jmp ret_from_intr + +#define BUILD_INTERRUPT(name, nr) \ +ENTRY(name) \ + pushl $nr-256; \ + SAVE_ALL \ + movl %esp,%eax; \ + call smp_/**/name; \ + jmp ret_from_intr; + +/* The include is where all of the SMP etc. interrupts come from */ +#include "entry_arch.h" + +ENTRY(divide_error) + pushl $0 # no error code + pushl $do_divide_error + ALIGN +error_code: + pushl %ds + pushl %eax + xorl %eax, %eax + pushl %ebp + pushl %edi + pushl %esi + pushl %edx + decl %eax # eax = -1 + pushl %ecx + pushl %ebx + cld + pushl %es + UNWIND_ESPFIX_STACK + popl %ecx + movl ES(%esp), %edi # get the function address + movl ORIG_EAX(%esp), %edx # get the error code + movl %eax, ORIG_EAX(%esp) + movl %ecx, ES(%esp) + movl $(__USER_DS), %ecx + movl %ecx, %ds + movl %ecx, %es + movl %esp,%eax # pt_regs pointer + call *%edi + jmp ret_from_exception + +ENTRY(coprocessor_error) + pushl $0 + pushl $do_coprocessor_error + jmp error_code + +ENTRY(simd_coprocessor_error) + pushl $0 + pushl $do_simd_coprocessor_error + jmp error_code + +ENTRY(device_not_available) + pushl $-1 # mark this as an int + SAVE_ALL + movl %cr0, %eax + testl $0x4, %eax # EM (math emulation bit) + jne device_not_available_emulate + preempt_stop + call math_state_restore + jmp ret_from_exception +device_not_available_emulate: + pushl $0 # temporary storage for ORIG_EIP + call math_emulate + addl $4, %esp + jmp ret_from_exception + +/* + * Debug traps and NMI can happen at the one SYSENTER instruction + * that sets up the real kernel stack. Check here, since we can't + * allow the wrong stack to be used. + * + * "TSS_sysenter_esp0+12" is because the NMI/debug handler will have + * already pushed 3 words if it hits on the sysenter instruction: + * eflags, cs and eip. + * + * We just load the right stack, and push the three (known) values + * by hand onto the new stack - while updating the return eip past + * the instruction that would have done it for sysenter. + */ +#define FIX_STACK(offset, ok, label) \ + cmpw $__KERNEL_CS,4(%esp); \ + jne ok; \ +label: \ + movl TSS_sysenter_esp0+offset(%esp),%esp; \ + pushfl; \ + pushl $__KERNEL_CS; \ + pushl $sysenter_past_esp + +ENTRY(debug) + cmpl $sysenter_entry,(%esp) + jne debug_stack_correct + FIX_STACK(12, debug_stack_correct, debug_esp_fix_insn) +debug_stack_correct: + pushl $-1 # mark this as an int + SAVE_ALL + xorl %edx,%edx # error code 0 + movl %esp,%eax # pt_regs pointer + call do_debug + testl %eax,%eax + jnz restore_all + jmp ret_from_exception + +/* + * NMI is doubly nasty. It can happen _while_ we're handling + * a debug fault, and the debug fault hasn't yet been able to + * clear up the stack. So we first check whether we got an + * NMI on the sysenter entry path, but after that we need to + * check whether we got an NMI on the debug path where the debug + * fault happened on the sysenter path. + */ +ENTRY(nmi) + pushl %eax + movl %ss, %eax + cmpw $__ESPFIX_SS, %ax + popl %eax + je nmi_16bit_stack + cmpl $sysenter_entry,(%esp) + je nmi_stack_fixup + pushl %eax + movl %esp,%eax + /* Do not access memory above the end of our stack page, + * it might not exist. + */ + andl $(THREAD_SIZE-1),%eax + cmpl $(THREAD_SIZE-20),%eax + popl %eax + jae nmi_stack_correct + cmpl $sysenter_entry,12(%esp) + je nmi_debug_stack_check +nmi_stack_correct: + pushl %eax + SAVE_ALL + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_nmi + jmp restore_all + +nmi_stack_fixup: + FIX_STACK(12,nmi_stack_correct, 1) + jmp nmi_stack_correct +nmi_debug_stack_check: + cmpw $__KERNEL_CS,16(%esp) + jne nmi_stack_correct + cmpl $debug - 1,(%esp) + jle nmi_stack_correct + cmpl $debug_esp_fix_insn,(%esp) + jle nmi_debug_stack_fixup +nmi_debug_stack_fixup: + FIX_STACK(24,nmi_stack_correct, 1) + jmp nmi_stack_correct + +nmi_16bit_stack: + /* create the pointer to lss back */ + pushl %ss + pushl %esp + movzwl %sp, %esp + addw $4, (%esp) + /* copy the iret frame of 12 bytes */ + .rept 3 + pushl 16(%esp) + .endr + pushl %eax + SAVE_ALL + FIXUP_ESPFIX_STACK # %eax == %esp + xorl %edx,%edx # zero error code + call do_nmi + RESTORE_REGS + lss 12+4(%esp), %esp # back to 16bit stack +1: iret +.section __ex_table,"a" + .align 4 + .long 1b,iret_exc +.previous + +ENTRY(int3) + pushl $-1 # mark this as an int + SAVE_ALL + xorl %edx,%edx # zero error code + movl %esp,%eax # pt_regs pointer + call do_int3 + testl %eax,%eax + jnz restore_all + jmp ret_from_exception + +ENTRY(overflow) + pushl $0 + pushl $do_overflow + jmp error_code + +ENTRY(bounds) + pushl $0 + pushl $do_bounds + jmp error_code + +ENTRY(invalid_op) + pushl $0 + pushl $do_invalid_op + jmp error_code + +ENTRY(coprocessor_segment_overrun) + pushl $0 + pushl $do_coprocessor_segment_overrun + jmp error_code + +ENTRY(invalid_TSS) + pushl $do_invalid_TSS + jmp error_code + +ENTRY(segment_not_present) + pushl $do_segment_not_present + jmp error_code + +ENTRY(stack_segment) + pushl $do_stack_segment + jmp error_code + +ENTRY(general_protection) + pushl $do_general_protection + jmp error_code + +ENTRY(alignment_check) + pushl $do_alignment_check + jmp error_code + +ENTRY(page_fault) + pushl $do_page_fault + jmp error_code + +#ifdef CONFIG_X86_MCE +ENTRY(machine_check) + pushl $0 + pushl machine_check_vector + jmp error_code +#endif + +ENTRY(spurious_interrupt_bug) + pushl $0 + pushl $do_spurious_interrupt_bug + jmp error_code + +.data +ENTRY(sys_call_table) + .long sys_restart_syscall /* 0 - old "setup()" system call, used for restarting */ + .long sys_exit + .long sys_fork + .long sys_read + .long sys_write + .long sys_open /* 5 */ + .long sys_close + .long sys_waitpid + .long sys_creat + .long sys_link + .long sys_unlink /* 10 */ + .long sys_execve + .long sys_chdir + .long sys_time + .long sys_mknod + .long sys_chmod /* 15 */ + .long sys_lchown16 + .long sys_ni_syscall /* old break syscall holder */ + .long sys_stat + .long sys_lseek + .long sys_getpid /* 20 */ + .long sys_mount + .long sys_oldumount + .long sys_setuid16 + .long sys_getuid16 + .long sys_stime /* 25 */ + .long sys_ptrace + .long sys_alarm + .long sys_fstat + .long sys_pause + .long sys_utime /* 30 */ + .long sys_ni_syscall /* old stty syscall holder */ + .long sys_ni_syscall /* old gtty syscall holder */ + .long sys_access + .long sys_nice + .long sys_ni_syscall /* 35 - old ftime syscall holder */ + .long sys_sync + .long sys_kill + .long sys_rename + .long sys_mkdir + .long sys_rmdir /* 40 */ + .long sys_dup + .long sys_pipe + .long sys_times + .long sys_ni_syscall /* old prof syscall holder */ + .long sys_brk /* 45 */ + .long sys_setgid16 + .long sys_getgid16 + .long sys_signal + .long sys_geteuid16 + .long sys_getegid16 /* 50 */ + .long sys_acct + .long sys_umount /* recycled never used phys() */ + .long sys_ni_syscall /* old lock syscall holder */ + .long sys_ioctl + .long sys_fcntl /* 55 */ + .long sys_ni_syscall /* old mpx syscall holder */ + .long sys_setpgid + .long sys_ni_syscall /* old ulimit syscall holder */ + .long sys_olduname + .long sys_umask /* 60 */ + .long sys_chroot + .long sys_ustat + .long sys_dup2 + .long sys_getppid + .long sys_getpgrp /* 65 */ + .long sys_setsid + .long sys_sigaction + .long sys_sgetmask + .long sys_ssetmask + .long sys_setreuid16 /* 70 */ + .long sys_setregid16 + .long sys_sigsuspend + .long sys_sigpending + .long sys_sethostname + .long sys_setrlimit /* 75 */ + .long sys_old_getrlimit + .long sys_getrusage + .long sys_gettimeofday + .long sys_settimeofday + .long sys_getgroups16 /* 80 */ + .long sys_setgroups16 + .long old_select + .long sys_symlink + .long sys_lstat + .long sys_readlink /* 85 */ + .long sys_uselib + .long sys_swapon + .long sys_reboot + .long old_readdir + .long old_mmap /* 90 */ + .long sys_munmap + .long sys_truncate + .long sys_ftruncate + .long sys_fchmod + .long sys_fchown16 /* 95 */ + .long sys_getpriority + .long sys_setpriority + .long sys_ni_syscall /* old profil syscall holder */ + .long sys_statfs + .long sys_fstatfs /* 100 */ + .long sys_ioperm + .long sys_socketcall + .long sys_syslog + .long sys_setitimer + .long sys_getitimer /* 105 */ + .long sys_newstat + .long sys_newlstat + .long sys_newfstat + .long sys_uname + .long sys_iopl /* 110 */ + .long sys_vhangup + .long sys_ni_syscall /* old "idle" system call */ + .long sys_vm86old + .long sys_wait4 + .long sys_swapoff /* 115 */ + .long sys_sysinfo + .long sys_ipc + .long sys_fsync + .long sys_sigreturn + .long sys_clone /* 120 */ + .long sys_setdomainname + .long sys_newuname + .long sys_modify_ldt + .long sys_adjtimex + .long sys_mprotect /* 125 */ + .long sys_sigprocmask + .long sys_ni_syscall /* old "create_module" */ + .long sys_init_module + .long sys_delete_module + .long sys_ni_syscall /* 130: old "get_kernel_syms" */ + .long sys_quotactl + .long sys_getpgid + .long sys_fchdir + .long sys_bdflush + .long sys_sysfs /* 135 */ + .long sys_personality + .long sys_ni_syscall /* reserved for afs_syscall */ + .long sys_setfsuid16 + .long sys_setfsgid16 + .long sys_llseek /* 140 */ + .long sys_getdents + .long sys_select + .long sys_flock + .long sys_msync + .long sys_readv /* 145 */ + .long sys_writev + .long sys_getsid + .long sys_fdatasync + .long sys_sysctl + .long sys_mlock /* 150 */ + .long sys_munlock + .long sys_mlockall + .long sys_munlockall + .long sys_sched_setparam + .long sys_sched_getparam /* 155 */ + .long sys_sched_setscheduler + .long sys_sched_getscheduler + .long sys_sched_yield + .long sys_sched_get_priority_max + .long sys_sched_get_priority_min /* 160 */ + .long sys_sched_rr_get_interval + .long sys_nanosleep + .long sys_mremap + .long sys_setresuid16 + .long sys_getresuid16 /* 165 */ + .long sys_vm86 + .long sys_ni_syscall /* Old sys_query_module */ + .long sys_poll + .long sys_nfsservctl + .long sys_setresgid16 /* 170 */ + .long sys_getresgid16 + .long sys_prctl + .long sys_rt_sigreturn + .long sys_rt_sigaction + .long sys_rt_sigprocmask /* 175 */ + .long sys_rt_sigpending + .long sys_rt_sigtimedwait + .long sys_rt_sigqueueinfo + .long sys_rt_sigsuspend + .long sys_pread64 /* 180 */ + .long sys_pwrite64 + .long sys_chown16 + .long sys_getcwd + .long sys_capget + .long sys_capset /* 185 */ + .long sys_sigaltstack + .long sys_sendfile + .long sys_ni_syscall /* reserved for streams1 */ + .long sys_ni_syscall /* reserved for streams2 */ + .long sys_vfork /* 190 */ + .long sys_getrlimit + .long sys_mmap2 + .long sys_truncate64 + .long sys_ftruncate64 + .long sys_stat64 /* 195 */ + .long sys_lstat64 + .long sys_fstat64 + .long sys_lchown + .long sys_getuid + .long sys_getgid /* 200 */ + .long sys_geteuid + .long sys_getegid + .long sys_setreuid + .long sys_setregid + .long sys_getgroups /* 205 */ + .long sys_setgroups + .long sys_fchown + .long sys_setresuid + .long sys_getresuid + .long sys_setresgid /* 210 */ + .long sys_getresgid + .long sys_chown + .long sys_setuid + .long sys_setgid + .long sys_setfsuid /* 215 */ + .long sys_setfsgid + .long sys_pivot_root + .long sys_mincore + .long sys_madvise + .long sys_getdents64 /* 220 */ + .long sys_fcntl64 + .long sys_ni_syscall /* reserved for TUX */ + .long sys_ni_syscall + .long sys_gettid + .long sys_readahead /* 225 */ + .long sys_setxattr + .long sys_lsetxattr + .long sys_fsetxattr + .long sys_getxattr + .long sys_lgetxattr /* 230 */ + .long sys_fgetxattr + .long sys_listxattr + .long sys_llistxattr + .long sys_flistxattr + .long sys_removexattr /* 235 */ + .long sys_lremovexattr + .long sys_fremovexattr + .long sys_tkill + .long sys_sendfile64 + .long sys_futex /* 240 */ + .long sys_sched_setaffinity + .long sys_sched_getaffinity + .long sys_set_thread_area + .long sys_get_thread_area + .long sys_io_setup /* 245 */ + .long sys_io_destroy + .long sys_io_getevents + .long sys_io_submit + .long sys_io_cancel + .long sys_fadvise64 /* 250 */ + .long sys_ni_syscall + .long sys_exit_group + .long sys_lookup_dcookie + .long sys_epoll_create + .long sys_epoll_ctl /* 255 */ + .long sys_epoll_wait + .long sys_remap_file_pages + .long sys_set_tid_address + .long sys_timer_create + .long sys_timer_settime /* 260 */ + .long sys_timer_gettime + .long sys_timer_getoverrun + .long sys_timer_delete + .long sys_clock_settime + .long sys_clock_gettime /* 265 */ + .long sys_clock_getres + .long sys_clock_nanosleep + .long sys_statfs64 + .long sys_fstatfs64 + .long sys_tgkill /* 270 */ + .long sys_utimes + .long sys_fadvise64_64 + .long sys_ni_syscall /* sys_vserver */ + .long sys_mbind + .long sys_get_mempolicy + .long sys_set_mempolicy + .long sys_mq_open + .long sys_mq_unlink + .long sys_mq_timedsend + .long sys_mq_timedreceive /* 280 */ + .long sys_mq_notify + .long sys_mq_getsetattr + .long sys_ni_syscall /* reserved for kexec */ + .long sys_waitid + .long sys_ni_syscall /* 285 */ /* available */ + .long sys_add_key + .long sys_request_key + .long sys_keyctl + +syscall_table_size=(.-sys_call_table) diff --git a/arch/i386/kernel/head.S b/arch/i386/kernel/head.S new file mode 100644 index 00000000000..d273fd74619 --- /dev/null +++ b/arch/i386/kernel/head.S @@ -0,0 +1,521 @@ +/* + * linux/arch/i386/kernel/head.S -- the 32-bit startup code. + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * Enhanced CPU detection and feature setting code by Mike Jagdis + * and Martin Mares, November 1997. + */ + +.text +#include <linux/config.h> +#include <linux/threads.h> +#include <linux/linkage.h> +#include <asm/segment.h> +#include <asm/page.h> +#include <asm/pgtable.h> +#include <asm/desc.h> +#include <asm/cache.h> +#include <asm/thread_info.h> +#include <asm/asm_offsets.h> +#include <asm/setup.h> + +/* + * References to members of the new_cpu_data structure. + */ + +#define X86 new_cpu_data+CPUINFO_x86 +#define X86_VENDOR new_cpu_data+CPUINFO_x86_vendor +#define X86_MODEL new_cpu_data+CPUINFO_x86_model +#define X86_MASK new_cpu_data+CPUINFO_x86_mask +#define X86_HARD_MATH new_cpu_data+CPUINFO_hard_math +#define X86_CPUID new_cpu_data+CPUINFO_cpuid_level +#define X86_CAPABILITY new_cpu_data+CPUINFO_x86_capability +#define X86_VENDOR_ID new_cpu_data+CPUINFO_x86_vendor_id + +/* + * This is how much memory *in addition to the memory covered up to + * and including _end* we need mapped initially. We need one bit for + * each possible page, but only in low memory, which means + * 2^32/4096/8 = 128K worst case (4G/4G split.) + * + * Modulo rounding, each megabyte assigned here requires a kilobyte of + * memory, which is currently unreclaimed. + * + * This should be a multiple of a page. + */ +#define INIT_MAP_BEYOND_END (128*1024) + + +/* + * 32-bit kernel entrypoint; only used by the boot CPU. On entry, + * %esi points to the real-mode code as a 32-bit pointer. + * CS and DS must be 4 GB flat segments, but we don't depend on + * any particular GDT layout, because we load our own as soon as we + * can. + */ +ENTRY(startup_32) + +/* + * Set segments to known values. + */ + cld + lgdt boot_gdt_descr - __PAGE_OFFSET + movl $(__BOOT_DS),%eax + movl %eax,%ds + movl %eax,%es + movl %eax,%fs + movl %eax,%gs + +/* + * Clear BSS first so that there are no surprises... + * No need to cld as DF is already clear from cld above... + */ + xorl %eax,%eax + movl $__bss_start - __PAGE_OFFSET,%edi + movl $__bss_stop - __PAGE_OFFSET,%ecx + subl %edi,%ecx + shrl $2,%ecx + rep ; stosl + +/* + * Initialize page tables. This creates a PDE and a set of page + * tables, which are located immediately beyond _end. The variable + * init_pg_tables_end is set up to point to the first "safe" location. + * Mappings are created both at virtual address 0 (identity mapping) + * and PAGE_OFFSET for up to _end+sizeof(page tables)+INIT_MAP_BEYOND_END. + * + * Warning: don't use %esi or the stack in this code. However, %esp + * can be used as a GPR if you really need it... + */ +page_pde_offset = (__PAGE_OFFSET >> 20); + + movl $(pg0 - __PAGE_OFFSET), %edi + movl $(swapper_pg_dir - __PAGE_OFFSET), %edx + movl $0x007, %eax /* 0x007 = PRESENT+RW+USER */ +10: + leal 0x007(%edi),%ecx /* Create PDE entry */ + movl %ecx,(%edx) /* Store identity PDE entry */ + movl %ecx,page_pde_offset(%edx) /* Store kernel PDE entry */ + addl $4,%edx + movl $1024, %ecx +11: + stosl + addl $0x1000,%eax + loop 11b + /* End condition: we must map up to and including INIT_MAP_BEYOND_END */ + /* bytes beyond the end of our own page tables; the +0x007 is the attribute bits */ + leal (INIT_MAP_BEYOND_END+0x007)(%edi),%ebp + cmpl %ebp,%eax + jb 10b + movl %edi,(init_pg_tables_end - __PAGE_OFFSET) + +#ifdef CONFIG_SMP + xorl %ebx,%ebx /* This is the boot CPU (BSP) */ + jmp 3f + +/* + * Non-boot CPU entry point; entered from trampoline.S + * We can't lgdt here, because lgdt itself uses a data segment, but + * we know the trampoline has already loaded the boot_gdt_table GDT + * for us. + */ +ENTRY(startup_32_smp) + cld + movl $(__BOOT_DS),%eax + movl %eax,%ds + movl %eax,%es + movl %eax,%fs + movl %eax,%gs + +/* + * New page tables may be in 4Mbyte page mode and may + * be using the global pages. + * + * NOTE! If we are on a 486 we may have no cr4 at all! + * So we do not try to touch it unless we really have + * some bits in it to set. This won't work if the BSP + * implements cr4 but this AP does not -- very unlikely + * but be warned! The same applies to the pse feature + * if not equally supported. --macro + * + * NOTE! We have to correct for the fact that we're + * not yet offset PAGE_OFFSET.. + */ +#define cr4_bits mmu_cr4_features-__PAGE_OFFSET + movl cr4_bits,%edx + andl %edx,%edx + jz 6f + movl %cr4,%eax # Turn on paging options (PSE,PAE,..) + orl %edx,%eax + movl %eax,%cr4 + + btl $5, %eax # check if PAE is enabled + jnc 6f + + /* Check if extended functions are implemented */ + movl $0x80000000, %eax + cpuid + cmpl $0x80000000, %eax + jbe 6f + mov $0x80000001, %eax + cpuid + /* Execute Disable bit supported? */ + btl $20, %edx + jnc 6f + + /* Setup EFER (Extended Feature Enable Register) */ + movl $0xc0000080, %ecx + rdmsr + + btsl $11, %eax + /* Make changes effective */ + wrmsr + +6: + /* This is a secondary processor (AP) */ + xorl %ebx,%ebx + incl %ebx + +3: +#endif /* CONFIG_SMP */ + +/* + * Enable paging + */ + movl $swapper_pg_dir-__PAGE_OFFSET,%eax + movl %eax,%cr3 /* set the page table pointer.. */ + movl %cr0,%eax + orl $0x80000000,%eax + movl %eax,%cr0 /* ..and set paging (PG) bit */ + ljmp $__BOOT_CS,$1f /* Clear prefetch and normalize %eip */ +1: + /* Set up the stack pointer */ + lss stack_start,%esp + +/* + * Initialize eflags. Some BIOS's leave bits like NT set. This would + * confuse the debugger if this code is traced. + * XXX - best to initialize before switching to protected mode. + */ + pushl $0 + popfl + +#ifdef CONFIG_SMP + andl %ebx,%ebx + jz 1f /* Initial CPU cleans BSS */ + jmp checkCPUtype +1: +#endif /* CONFIG_SMP */ + +/* + * start system 32-bit setup. We need to re-do some of the things done + * in 16-bit mode for the "real" operations. + */ + call setup_idt + +/* + * Copy bootup parameters out of the way. + * Note: %esi still has the pointer to the real-mode data. + */ + movl $boot_params,%edi + movl $(PARAM_SIZE/4),%ecx + cld + rep + movsl + movl boot_params+NEW_CL_POINTER,%esi + andl %esi,%esi + jnz 2f # New command line protocol + cmpw $(OLD_CL_MAGIC),OLD_CL_MAGIC_ADDR + jne 1f + movzwl OLD_CL_OFFSET,%esi + addl $(OLD_CL_BASE_ADDR),%esi +2: + movl $saved_command_line,%edi + movl $(COMMAND_LINE_SIZE/4),%ecx + rep + movsl +1: +checkCPUtype: + + movl $-1,X86_CPUID # -1 for no CPUID initially + +/* check if it is 486 or 386. */ +/* + * XXX - this does a lot of unnecessary setup. Alignment checks don't + * apply at our cpl of 0 and the stack ought to be aligned already, and + * we don't need to preserve eflags. + */ + + movb $3,X86 # at least 386 + pushfl # push EFLAGS + popl %eax # get EFLAGS + movl %eax,%ecx # save original EFLAGS + xorl $0x240000,%eax # flip AC and ID bits in EFLAGS + pushl %eax # copy to EFLAGS + popfl # set EFLAGS + pushfl # get new EFLAGS + popl %eax # put it in eax + xorl %ecx,%eax # change in flags + pushl %ecx # restore original EFLAGS + popfl + testl $0x40000,%eax # check if AC bit changed + je is386 + + movb $4,X86 # at least 486 + testl $0x200000,%eax # check if ID bit changed + je is486 + + /* get vendor info */ + xorl %eax,%eax # call CPUID with 0 -> return vendor ID + cpuid + movl %eax,X86_CPUID # save CPUID level + movl %ebx,X86_VENDOR_ID # lo 4 chars + movl %edx,X86_VENDOR_ID+4 # next 4 chars + movl %ecx,X86_VENDOR_ID+8 # last 4 chars + + orl %eax,%eax # do we have processor info as well? + je is486 + + movl $1,%eax # Use the CPUID instruction to get CPU type + cpuid + movb %al,%cl # save reg for future use + andb $0x0f,%ah # mask processor family + movb %ah,X86 + andb $0xf0,%al # mask model + shrb $4,%al + movb %al,X86_MODEL + andb $0x0f,%cl # mask mask revision + movb %cl,X86_MASK + movl %edx,X86_CAPABILITY + +is486: movl $0x50022,%ecx # set AM, WP, NE and MP + jmp 2f + +is386: movl $2,%ecx # set MP +2: movl %cr0,%eax + andl $0x80000011,%eax # Save PG,PE,ET + orl %ecx,%eax + movl %eax,%cr0 + + call check_x87 + incb ready + lgdt cpu_gdt_descr + lidt idt_descr + ljmp $(__KERNEL_CS),$1f +1: movl $(__KERNEL_DS),%eax # reload all the segment registers + movl %eax,%ss # after changing gdt. + + movl $(__USER_DS),%eax # DS/ES contains default USER segment + movl %eax,%ds + movl %eax,%es + + xorl %eax,%eax # Clear FS/GS and LDT + movl %eax,%fs + movl %eax,%gs + lldt %ax + cld # gcc2 wants the direction flag cleared at all times +#ifdef CONFIG_SMP + movb ready, %cl + cmpb $1,%cl + je 1f # the first CPU calls start_kernel + # all other CPUs call initialize_secondary + call initialize_secondary + jmp L6 +1: +#endif /* CONFIG_SMP */ + call start_kernel +L6: + jmp L6 # main should never return here, but + # just in case, we know what happens. + +/* + * We depend on ET to be correct. This checks for 287/387. + */ +check_x87: + movb $0,X86_HARD_MATH + clts + fninit + fstsw %ax + cmpb $0,%al + je 1f + movl %cr0,%eax /* no coprocessor: have to set bits */ + xorl $4,%eax /* set EM */ + movl %eax,%cr0 + ret + ALIGN +1: movb $1,X86_HARD_MATH + .byte 0xDB,0xE4 /* fsetpm for 287, ignored by 387 */ + ret + +/* + * setup_idt + * + * sets up a idt with 256 entries pointing to + * ignore_int, interrupt gates. It doesn't actually load + * idt - that can be done only after paging has been enabled + * and the kernel moved to PAGE_OFFSET. Interrupts + * are enabled elsewhere, when we can be relatively + * sure everything is ok. + * + * Warning: %esi is live across this function. + */ +setup_idt: + lea ignore_int,%edx + movl $(__KERNEL_CS << 16),%eax + movw %dx,%ax /* selector = 0x0010 = cs */ + movw $0x8E00,%dx /* interrupt gate - dpl=0, present */ + + lea idt_table,%edi + mov $256,%ecx +rp_sidt: + movl %eax,(%edi) + movl %edx,4(%edi) + addl $8,%edi + dec %ecx + jne rp_sidt + ret + +/* This is the default interrupt "handler" :-) */ + ALIGN +ignore_int: + cld + pushl %eax + pushl %ecx + pushl %edx + pushl %es + pushl %ds + movl $(__KERNEL_DS),%eax + movl %eax,%ds + movl %eax,%es + pushl 16(%esp) + pushl 24(%esp) + pushl 32(%esp) + pushl 40(%esp) + pushl $int_msg + call printk + addl $(5*4),%esp + popl %ds + popl %es + popl %edx + popl %ecx + popl %eax + iret + +/* + * Real beginning of normal "text" segment + */ +ENTRY(stext) +ENTRY(_stext) + +/* + * BSS section + */ +.section ".bss.page_aligned","w" +ENTRY(swapper_pg_dir) + .fill 1024,4,0 +ENTRY(empty_zero_page) + .fill 4096,1,0 + +/* + * This starts the data section. + */ +.data + +ENTRY(stack_start) + .long init_thread_union+THREAD_SIZE + .long __BOOT_DS + +ready: .byte 0 + +int_msg: + .asciz "Unknown interrupt or fault at EIP %p %p %p\n" + +/* + * The IDT and GDT 'descriptors' are a strange 48-bit object + * only used by the lidt and lgdt instructions. They are not + * like usual segment descriptors - they consist of a 16-bit + * segment size, and 32-bit linear address value: + */ + +.globl boot_gdt_descr +.globl idt_descr +.globl cpu_gdt_descr + + ALIGN +# early boot GDT descriptor (must use 1:1 address mapping) + .word 0 # 32 bit align gdt_desc.address +boot_gdt_descr: + .word __BOOT_DS+7 + .long boot_gdt_table - __PAGE_OFFSET + + .word 0 # 32-bit align idt_desc.address +idt_descr: + .word IDT_ENTRIES*8-1 # idt contains 256 entries + .long idt_table + +# boot GDT descriptor (later on used by CPU#0): + .word 0 # 32 bit align gdt_desc.address +cpu_gdt_descr: + .word GDT_ENTRIES*8-1 + .long cpu_gdt_table + + .fill NR_CPUS-1,8,0 # space for the other GDT descriptors + +/* + * The boot_gdt_table must mirror the equivalent in setup.S and is + * used only for booting. + */ + .align L1_CACHE_BYTES +ENTRY(boot_gdt_table) + .fill GDT_ENTRY_BOOT_CS,8,0 + .quad 0x00cf9a000000ffff /* kernel 4GB code at 0x00000000 */ + .quad 0x00cf92000000ffff /* kernel 4GB data at 0x00000000 */ + +/* + * The Global Descriptor Table contains 28 quadwords, per-CPU. + */ + .align PAGE_SIZE_asm +ENTRY(cpu_gdt_table) + .quad 0x0000000000000000 /* NULL descriptor */ + .quad 0x0000000000000000 /* 0x0b reserved */ + .quad 0x0000000000000000 /* 0x13 reserved */ + .quad 0x0000000000000000 /* 0x1b reserved */ + .quad 0x0000000000000000 /* 0x20 unused */ + .quad 0x0000000000000000 /* 0x28 unused */ + .quad 0x0000000000000000 /* 0x33 TLS entry 1 */ + .quad 0x0000000000000000 /* 0x3b TLS entry 2 */ + .quad 0x0000000000000000 /* 0x43 TLS entry 3 */ + .quad 0x0000000000000000 /* 0x4b reserved */ + .quad 0x0000000000000000 /* 0x53 reserved */ + .quad 0x0000000000000000 /* 0x5b reserved */ + + .quad 0x00cf9a000000ffff /* 0x60 kernel 4GB code at 0x00000000 */ + .quad 0x00cf92000000ffff /* 0x68 kernel 4GB data at 0x00000000 */ + .quad 0x00cffa000000ffff /* 0x73 user 4GB code at 0x00000000 */ + .quad 0x00cff2000000ffff /* 0x7b user 4GB data at 0x00000000 */ + + .quad 0x0000000000000000 /* 0x80 TSS descriptor */ + .quad 0x0000000000000000 /* 0x88 LDT descriptor */ + + /* Segments used for calling PnP BIOS */ + .quad 0x00c09a0000000000 /* 0x90 32-bit code */ + .quad 0x00809a0000000000 /* 0x98 16-bit code */ + .quad 0x0080920000000000 /* 0xa0 16-bit data */ + .quad 0x0080920000000000 /* 0xa8 16-bit data */ + .quad 0x0080920000000000 /* 0xb0 16-bit data */ + /* + * The APM segments have byte granularity and their bases + * and limits are set at run time. + */ + .quad 0x00409a0000000000 /* 0xb8 APM CS code */ + .quad 0x00009a0000000000 /* 0xc0 APM CS 16 code (16 bit) */ + .quad 0x0040920000000000 /* 0xc8 APM DS data */ + + .quad 0x0000920000000000 /* 0xd0 - ESPFIX 16-bit SS */ + .quad 0x0000000000000000 /* 0xd8 - unused */ + .quad 0x0000000000000000 /* 0xe0 - unused */ + .quad 0x0000000000000000 /* 0xe8 - unused */ + .quad 0x0000000000000000 /* 0xf0 - unused */ + .quad 0x0000000000000000 /* 0xf8 - GDT entry 31: double-fault TSS */ + diff --git a/arch/i386/kernel/i386_ksyms.c b/arch/i386/kernel/i386_ksyms.c new file mode 100644 index 00000000000..14ec354bec9 --- /dev/null +++ b/arch/i386/kernel/i386_ksyms.c @@ -0,0 +1,195 @@ +#include <linux/config.h> +#include <linux/module.h> +#include <linux/smp.h> +#include <linux/user.h> +#include <linux/elfcore.h> +#include <linux/mca.h> +#include <linux/sched.h> +#include <linux/in6.h> +#include <linux/interrupt.h> +#include <linux/smp_lock.h> +#include <linux/pm.h> +#include <linux/pci.h> +#include <linux/apm_bios.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/tty.h> +#include <linux/highmem.h> +#include <linux/time.h> + +#include <asm/semaphore.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/uaccess.h> +#include <asm/checksum.h> +#include <asm/io.h> +#include <asm/delay.h> +#include <asm/irq.h> +#include <asm/mmx.h> +#include <asm/desc.h> +#include <asm/pgtable.h> +#include <asm/tlbflush.h> +#include <asm/nmi.h> +#include <asm/ist.h> +#include <asm/kdebug.h> + +extern void dump_thread(struct pt_regs *, struct user *); +extern spinlock_t rtc_lock; + +/* This is definitely a GPL-only symbol */ +EXPORT_SYMBOL_GPL(cpu_gdt_table); + +#if defined(CONFIG_APM_MODULE) +extern void machine_real_restart(unsigned char *, int); +EXPORT_SYMBOL(machine_real_restart); +extern void default_idle(void); +EXPORT_SYMBOL(default_idle); +#endif + +#ifdef CONFIG_SMP +extern void FASTCALL( __write_lock_failed(rwlock_t *rw)); +extern void FASTCALL( __read_lock_failed(rwlock_t *rw)); +#endif + +#if defined(CONFIG_BLK_DEV_IDE) || defined(CONFIG_BLK_DEV_HD) || defined(CONFIG_BLK_DEV_IDE_MODULE) || defined(CONFIG_BLK_DEV_HD_MODULE) +extern struct drive_info_struct drive_info; +EXPORT_SYMBOL(drive_info); +#endif + +extern unsigned long cpu_khz; +extern unsigned long get_cmos_time(void); + +/* platform dependent support */ +EXPORT_SYMBOL(boot_cpu_data); +#ifdef CONFIG_DISCONTIGMEM +EXPORT_SYMBOL(node_data); +EXPORT_SYMBOL(physnode_map); +#endif +#ifdef CONFIG_X86_NUMAQ +EXPORT_SYMBOL(xquad_portio); +#endif +EXPORT_SYMBOL(dump_thread); +EXPORT_SYMBOL(dump_fpu); +EXPORT_SYMBOL_GPL(kernel_fpu_begin); +EXPORT_SYMBOL(__ioremap); +EXPORT_SYMBOL(ioremap_nocache); +EXPORT_SYMBOL(iounmap); +EXPORT_SYMBOL(kernel_thread); +EXPORT_SYMBOL(pm_idle); +EXPORT_SYMBOL(pm_power_off); +EXPORT_SYMBOL(get_cmos_time); +EXPORT_SYMBOL(cpu_khz); +EXPORT_SYMBOL(apm_info); + +EXPORT_SYMBOL(__down_failed); +EXPORT_SYMBOL(__down_failed_interruptible); +EXPORT_SYMBOL(__down_failed_trylock); +EXPORT_SYMBOL(__up_wakeup); +/* Networking helper routines. */ +EXPORT_SYMBOL(csum_partial_copy_generic); +/* Delay loops */ +EXPORT_SYMBOL(__ndelay); +EXPORT_SYMBOL(__udelay); +EXPORT_SYMBOL(__delay); +EXPORT_SYMBOL(__const_udelay); + +EXPORT_SYMBOL(__get_user_1); +EXPORT_SYMBOL(__get_user_2); +EXPORT_SYMBOL(__get_user_4); + +EXPORT_SYMBOL(__put_user_1); +EXPORT_SYMBOL(__put_user_2); +EXPORT_SYMBOL(__put_user_4); +EXPORT_SYMBOL(__put_user_8); + +EXPORT_SYMBOL(strpbrk); +EXPORT_SYMBOL(strstr); + +EXPORT_SYMBOL(strncpy_from_user); +EXPORT_SYMBOL(__strncpy_from_user); +EXPORT_SYMBOL(clear_user); +EXPORT_SYMBOL(__clear_user); +EXPORT_SYMBOL(__copy_from_user_ll); +EXPORT_SYMBOL(__copy_to_user_ll); +EXPORT_SYMBOL(strnlen_user); + +EXPORT_SYMBOL(dma_alloc_coherent); +EXPORT_SYMBOL(dma_free_coherent); + +#ifdef CONFIG_PCI +EXPORT_SYMBOL(pci_mem_start); +#endif + +#ifdef CONFIG_PCI_BIOS +EXPORT_SYMBOL(pcibios_set_irq_routing); +EXPORT_SYMBOL(pcibios_get_irq_routing_table); +#endif + +#ifdef CONFIG_X86_USE_3DNOW +EXPORT_SYMBOL(_mmx_memcpy); +EXPORT_SYMBOL(mmx_clear_page); +EXPORT_SYMBOL(mmx_copy_page); +#endif + +#ifdef CONFIG_X86_HT +EXPORT_SYMBOL(smp_num_siblings); +EXPORT_SYMBOL(cpu_sibling_map); +#endif + +#ifdef CONFIG_SMP +EXPORT_SYMBOL(cpu_data); +EXPORT_SYMBOL(cpu_online_map); +EXPORT_SYMBOL(cpu_callout_map); +EXPORT_SYMBOL(__write_lock_failed); +EXPORT_SYMBOL(__read_lock_failed); + +/* Global SMP stuff */ +EXPORT_SYMBOL(smp_call_function); + +/* TLB flushing */ +EXPORT_SYMBOL(flush_tlb_page); +#endif + +#ifdef CONFIG_X86_IO_APIC +EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); +#endif + +#ifdef CONFIG_MCA +EXPORT_SYMBOL(machine_id); +#endif + +#ifdef CONFIG_VT +EXPORT_SYMBOL(screen_info); +#endif + +EXPORT_SYMBOL(get_wchan); + +EXPORT_SYMBOL(rtc_lock); + +EXPORT_SYMBOL_GPL(set_nmi_callback); +EXPORT_SYMBOL_GPL(unset_nmi_callback); + +#undef memcmp +extern int memcmp(const void *,const void *,__kernel_size_t); +EXPORT_SYMBOL(memcmp); + +EXPORT_SYMBOL(register_die_notifier); +#ifdef CONFIG_HAVE_DEC_LOCK +EXPORT_SYMBOL(_atomic_dec_and_lock); +#endif + +EXPORT_SYMBOL(__PAGE_KERNEL); + +#ifdef CONFIG_HIGHMEM +EXPORT_SYMBOL(kmap); +EXPORT_SYMBOL(kunmap); +EXPORT_SYMBOL(kmap_atomic); +EXPORT_SYMBOL(kunmap_atomic); +EXPORT_SYMBOL(kmap_atomic_to_page); +#endif + +#if defined(CONFIG_X86_SPEEDSTEP_SMI) || defined(CONFIG_X86_SPEEDSTEP_SMI_MODULE) +EXPORT_SYMBOL(ist_info); +#endif + +EXPORT_SYMBOL(csum_partial); diff --git a/arch/i386/kernel/i387.c b/arch/i386/kernel/i387.c new file mode 100644 index 00000000000..c55e037f08f --- /dev/null +++ b/arch/i386/kernel/i387.c @@ -0,0 +1,555 @@ +/* + * linux/arch/i386/kernel/i387.c + * + * Copyright (C) 1994 Linus Torvalds + * + * Pentium III FXSR, SSE support + * General FPU state handling cleanups + * Gareth Hughes <gareth@valinux.com>, May 2000 + */ + +#include <linux/config.h> +#include <linux/sched.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/math_emu.h> +#include <asm/sigcontext.h> +#include <asm/user.h> +#include <asm/ptrace.h> +#include <asm/uaccess.h> + +#ifdef CONFIG_MATH_EMULATION +#define HAVE_HWFP (boot_cpu_data.hard_math) +#else +#define HAVE_HWFP 1 +#endif + +static unsigned long mxcsr_feature_mask = 0xffffffff; + +void mxcsr_feature_mask_init(void) +{ + unsigned long mask = 0; + clts(); + if (cpu_has_fxsr) { + memset(¤t->thread.i387.fxsave, 0, sizeof(struct i387_fxsave_struct)); + asm volatile("fxsave %0" : : "m" (current->thread.i387.fxsave)); + mask = current->thread.i387.fxsave.mxcsr_mask; + if (mask == 0) mask = 0x0000ffbf; + } + mxcsr_feature_mask &= mask; + stts(); +} + +/* + * The _current_ task is using the FPU for the first time + * so initialize it and set the mxcsr to its default + * value at reset if we support XMM instructions and then + * remeber the current task has used the FPU. + */ +void init_fpu(struct task_struct *tsk) +{ + if (cpu_has_fxsr) { + memset(&tsk->thread.i387.fxsave, 0, sizeof(struct i387_fxsave_struct)); + tsk->thread.i387.fxsave.cwd = 0x37f; + if (cpu_has_xmm) + tsk->thread.i387.fxsave.mxcsr = 0x1f80; + } else { + memset(&tsk->thread.i387.fsave, 0, sizeof(struct i387_fsave_struct)); + tsk->thread.i387.fsave.cwd = 0xffff037fu; + tsk->thread.i387.fsave.swd = 0xffff0000u; + tsk->thread.i387.fsave.twd = 0xffffffffu; + tsk->thread.i387.fsave.fos = 0xffff0000u; + } + /* only the device not available exception or ptrace can call init_fpu */ + set_stopped_child_used_math(tsk); +} + +/* + * FPU lazy state save handling. + */ + +void kernel_fpu_begin(void) +{ + struct thread_info *thread = current_thread_info(); + + preempt_disable(); + if (thread->status & TS_USEDFPU) { + __save_init_fpu(thread->task); + return; + } + clts(); +} + +void restore_fpu( struct task_struct *tsk ) +{ + if ( cpu_has_fxsr ) { + asm volatile( "fxrstor %0" + : : "m" (tsk->thread.i387.fxsave) ); + } else { + asm volatile( "frstor %0" + : : "m" (tsk->thread.i387.fsave) ); + } +} + +/* + * FPU tag word conversions. + */ + +static inline unsigned short twd_i387_to_fxsr( unsigned short twd ) +{ + unsigned int tmp; /* to avoid 16 bit prefixes in the code */ + + /* Transform each pair of bits into 01 (valid) or 00 (empty) */ + tmp = ~twd; + tmp = (tmp | (tmp>>1)) & 0x5555; /* 0V0V0V0V0V0V0V0V */ + /* and move the valid bits to the lower byte. */ + tmp = (tmp | (tmp >> 1)) & 0x3333; /* 00VV00VV00VV00VV */ + tmp = (tmp | (tmp >> 2)) & 0x0f0f; /* 0000VVVV0000VVVV */ + tmp = (tmp | (tmp >> 4)) & 0x00ff; /* 00000000VVVVVVVV */ + return tmp; +} + +static inline unsigned long twd_fxsr_to_i387( struct i387_fxsave_struct *fxsave ) +{ + struct _fpxreg *st = NULL; + unsigned long tos = (fxsave->swd >> 11) & 7; + unsigned long twd = (unsigned long) fxsave->twd; + unsigned long tag; + unsigned long ret = 0xffff0000u; + int i; + +#define FPREG_ADDR(f, n) ((void *)&(f)->st_space + (n) * 16); + + for ( i = 0 ; i < 8 ; i++ ) { + if ( twd & 0x1 ) { + st = FPREG_ADDR( fxsave, (i - tos) & 7 ); + + switch ( st->exponent & 0x7fff ) { + case 0x7fff: + tag = 2; /* Special */ + break; + case 0x0000: + if ( !st->significand[0] && + !st->significand[1] && + !st->significand[2] && + !st->significand[3] ) { + tag = 1; /* Zero */ + } else { + tag = 2; /* Special */ + } + break; + default: + if ( st->significand[3] & 0x8000 ) { + tag = 0; /* Valid */ + } else { + tag = 2; /* Special */ + } + break; + } + } else { + tag = 3; /* Empty */ + } + ret |= (tag << (2 * i)); + twd = twd >> 1; + } + return ret; +} + +/* + * FPU state interaction. + */ + +unsigned short get_fpu_cwd( struct task_struct *tsk ) +{ + if ( cpu_has_fxsr ) { + return tsk->thread.i387.fxsave.cwd; + } else { + return (unsigned short)tsk->thread.i387.fsave.cwd; + } +} + +unsigned short get_fpu_swd( struct task_struct *tsk ) +{ + if ( cpu_has_fxsr ) { + return tsk->thread.i387.fxsave.swd; + } else { + return (unsigned short)tsk->thread.i387.fsave.swd; + } +} + +#if 0 +unsigned short get_fpu_twd( struct task_struct *tsk ) +{ + if ( cpu_has_fxsr ) { + return tsk->thread.i387.fxsave.twd; + } else { + return (unsigned short)tsk->thread.i387.fsave.twd; + } +} +#endif /* 0 */ + +unsigned short get_fpu_mxcsr( struct task_struct *tsk ) +{ + if ( cpu_has_xmm ) { + return tsk->thread.i387.fxsave.mxcsr; + } else { + return 0x1f80; + } +} + +#if 0 + +void set_fpu_cwd( struct task_struct *tsk, unsigned short cwd ) +{ + if ( cpu_has_fxsr ) { + tsk->thread.i387.fxsave.cwd = cwd; + } else { + tsk->thread.i387.fsave.cwd = ((long)cwd | 0xffff0000u); + } +} + +void set_fpu_swd( struct task_struct *tsk, unsigned short swd ) +{ + if ( cpu_has_fxsr ) { + tsk->thread.i387.fxsave.swd = swd; + } else { + tsk->thread.i387.fsave.swd = ((long)swd | 0xffff0000u); + } +} + +void set_fpu_twd( struct task_struct *tsk, unsigned short twd ) +{ + if ( cpu_has_fxsr ) { + tsk->thread.i387.fxsave.twd = twd_i387_to_fxsr(twd); + } else { + tsk->thread.i387.fsave.twd = ((long)twd | 0xffff0000u); + } +} + +#endif /* 0 */ + +/* + * FXSR floating point environment conversions. + */ + +static int convert_fxsr_to_user( struct _fpstate __user *buf, + struct i387_fxsave_struct *fxsave ) +{ + unsigned long env[7]; + struct _fpreg __user *to; + struct _fpxreg *from; + int i; + + env[0] = (unsigned long)fxsave->cwd | 0xffff0000ul; + env[1] = (unsigned long)fxsave->swd | 0xffff0000ul; + env[2] = twd_fxsr_to_i387(fxsave); + env[3] = fxsave->fip; + env[4] = fxsave->fcs | ((unsigned long)fxsave->fop << 16); + env[5] = fxsave->foo; + env[6] = fxsave->fos; + + if ( __copy_to_user( buf, env, 7 * sizeof(unsigned long) ) ) + return 1; + + to = &buf->_st[0]; + from = (struct _fpxreg *) &fxsave->st_space[0]; + for ( i = 0 ; i < 8 ; i++, to++, from++ ) { + unsigned long __user *t = (unsigned long __user *)to; + unsigned long *f = (unsigned long *)from; + + if (__put_user(*f, t) || + __put_user(*(f + 1), t + 1) || + __put_user(from->exponent, &to->exponent)) + return 1; + } + return 0; +} + +static int convert_fxsr_from_user( struct i387_fxsave_struct *fxsave, + struct _fpstate __user *buf ) +{ + unsigned long env[7]; + struct _fpxreg *to; + struct _fpreg __user *from; + int i; + + if ( __copy_from_user( env, buf, 7 * sizeof(long) ) ) + return 1; + + fxsave->cwd = (unsigned short)(env[0] & 0xffff); + fxsave->swd = (unsigned short)(env[1] & 0xffff); + fxsave->twd = twd_i387_to_fxsr((unsigned short)(env[2] & 0xffff)); + fxsave->fip = env[3]; + fxsave->fop = (unsigned short)((env[4] & 0xffff0000ul) >> 16); + fxsave->fcs = (env[4] & 0xffff); + fxsave->foo = env[5]; + fxsave->fos = env[6]; + + to = (struct _fpxreg *) &fxsave->st_space[0]; + from = &buf->_st[0]; + for ( i = 0 ; i < 8 ; i++, to++, from++ ) { + unsigned long *t = (unsigned long *)to; + unsigned long __user *f = (unsigned long __user *)from; + + if (__get_user(*t, f) || + __get_user(*(t + 1), f + 1) || + __get_user(to->exponent, &from->exponent)) + return 1; + } + return 0; +} + +/* + * Signal frame handlers. + */ + +static inline int save_i387_fsave( struct _fpstate __user *buf ) +{ + struct task_struct *tsk = current; + + unlazy_fpu( tsk ); + tsk->thread.i387.fsave.status = tsk->thread.i387.fsave.swd; + if ( __copy_to_user( buf, &tsk->thread.i387.fsave, + sizeof(struct i387_fsave_struct) ) ) + return -1; + return 1; +} + +static int save_i387_fxsave( struct _fpstate __user *buf ) +{ + struct task_struct *tsk = current; + int err = 0; + + unlazy_fpu( tsk ); + + if ( convert_fxsr_to_user( buf, &tsk->thread.i387.fxsave ) ) + return -1; + + err |= __put_user( tsk->thread.i387.fxsave.swd, &buf->status ); + err |= __put_user( X86_FXSR_MAGIC, &buf->magic ); + if ( err ) + return -1; + + if ( __copy_to_user( &buf->_fxsr_env[0], &tsk->thread.i387.fxsave, + sizeof(struct i387_fxsave_struct) ) ) + return -1; + return 1; +} + +int save_i387( struct _fpstate __user *buf ) +{ + if ( !used_math() ) + return 0; + + /* This will cause a "finit" to be triggered by the next + * attempted FPU operation by the 'current' process. + */ + clear_used_math(); + + if ( HAVE_HWFP ) { + if ( cpu_has_fxsr ) { + return save_i387_fxsave( buf ); + } else { + return save_i387_fsave( buf ); + } + } else { + return save_i387_soft( ¤t->thread.i387.soft, buf ); + } +} + +static inline int restore_i387_fsave( struct _fpstate __user *buf ) +{ + struct task_struct *tsk = current; + clear_fpu( tsk ); + return __copy_from_user( &tsk->thread.i387.fsave, buf, + sizeof(struct i387_fsave_struct) ); +} + +static int restore_i387_fxsave( struct _fpstate __user *buf ) +{ + int err; + struct task_struct *tsk = current; + clear_fpu( tsk ); + err = __copy_from_user( &tsk->thread.i387.fxsave, &buf->_fxsr_env[0], + sizeof(struct i387_fxsave_struct) ); + /* mxcsr reserved bits must be masked to zero for security reasons */ + tsk->thread.i387.fxsave.mxcsr &= mxcsr_feature_mask; + return err ? 1 : convert_fxsr_from_user( &tsk->thread.i387.fxsave, buf ); +} + +int restore_i387( struct _fpstate __user *buf ) +{ + int err; + + if ( HAVE_HWFP ) { + if ( cpu_has_fxsr ) { + err = restore_i387_fxsave( buf ); + } else { + err = restore_i387_fsave( buf ); + } + } else { + err = restore_i387_soft( ¤t->thread.i387.soft, buf ); + } + set_used_math(); + return err; +} + +/* + * ptrace request handlers. + */ + +static inline int get_fpregs_fsave( struct user_i387_struct __user *buf, + struct task_struct *tsk ) +{ + return __copy_to_user( buf, &tsk->thread.i387.fsave, + sizeof(struct user_i387_struct) ); +} + +static inline int get_fpregs_fxsave( struct user_i387_struct __user *buf, + struct task_struct *tsk ) +{ + return convert_fxsr_to_user( (struct _fpstate __user *)buf, + &tsk->thread.i387.fxsave ); +} + +int get_fpregs( struct user_i387_struct __user *buf, struct task_struct *tsk ) +{ + if ( HAVE_HWFP ) { + if ( cpu_has_fxsr ) { + return get_fpregs_fxsave( buf, tsk ); + } else { + return get_fpregs_fsave( buf, tsk ); + } + } else { + return save_i387_soft( &tsk->thread.i387.soft, + (struct _fpstate __user *)buf ); + } +} + +static inline int set_fpregs_fsave( struct task_struct *tsk, + struct user_i387_struct __user *buf ) +{ + return __copy_from_user( &tsk->thread.i387.fsave, buf, + sizeof(struct user_i387_struct) ); +} + +static inline int set_fpregs_fxsave( struct task_struct *tsk, + struct user_i387_struct __user *buf ) +{ + return convert_fxsr_from_user( &tsk->thread.i387.fxsave, + (struct _fpstate __user *)buf ); +} + +int set_fpregs( struct task_struct *tsk, struct user_i387_struct __user *buf ) +{ + if ( HAVE_HWFP ) { + if ( cpu_has_fxsr ) { + return set_fpregs_fxsave( tsk, buf ); + } else { + return set_fpregs_fsave( tsk, buf ); + } + } else { + return restore_i387_soft( &tsk->thread.i387.soft, + (struct _fpstate __user *)buf ); + } +} + +int get_fpxregs( struct user_fxsr_struct __user *buf, struct task_struct *tsk ) +{ + if ( cpu_has_fxsr ) { + if (__copy_to_user( buf, &tsk->thread.i387.fxsave, + sizeof(struct user_fxsr_struct) )) + return -EFAULT; + return 0; + } else { + return -EIO; + } +} + +int set_fpxregs( struct task_struct *tsk, struct user_fxsr_struct __user *buf ) +{ + int ret = 0; + + if ( cpu_has_fxsr ) { + if (__copy_from_user( &tsk->thread.i387.fxsave, buf, + sizeof(struct user_fxsr_struct) )) + ret = -EFAULT; + /* mxcsr reserved bits must be masked to zero for security reasons */ + tsk->thread.i387.fxsave.mxcsr &= mxcsr_feature_mask; + } else { + ret = -EIO; + } + return ret; +} + +/* + * FPU state for core dumps. + */ + +static inline void copy_fpu_fsave( struct task_struct *tsk, + struct user_i387_struct *fpu ) +{ + memcpy( fpu, &tsk->thread.i387.fsave, + sizeof(struct user_i387_struct) ); +} + +static inline void copy_fpu_fxsave( struct task_struct *tsk, + struct user_i387_struct *fpu ) +{ + unsigned short *to; + unsigned short *from; + int i; + + memcpy( fpu, &tsk->thread.i387.fxsave, 7 * sizeof(long) ); + + to = (unsigned short *)&fpu->st_space[0]; + from = (unsigned short *)&tsk->thread.i387.fxsave.st_space[0]; + for ( i = 0 ; i < 8 ; i++, to += 5, from += 8 ) { + memcpy( to, from, 5 * sizeof(unsigned short) ); + } +} + +int dump_fpu( struct pt_regs *regs, struct user_i387_struct *fpu ) +{ + int fpvalid; + struct task_struct *tsk = current; + + fpvalid = !!used_math(); + if ( fpvalid ) { + unlazy_fpu( tsk ); + if ( cpu_has_fxsr ) { + copy_fpu_fxsave( tsk, fpu ); + } else { + copy_fpu_fsave( tsk, fpu ); + } + } + + return fpvalid; +} + +int dump_task_fpu(struct task_struct *tsk, struct user_i387_struct *fpu) +{ + int fpvalid = !!tsk_used_math(tsk); + + if (fpvalid) { + if (tsk == current) + unlazy_fpu(tsk); + if (cpu_has_fxsr) + copy_fpu_fxsave(tsk, fpu); + else + copy_fpu_fsave(tsk, fpu); + } + return fpvalid; +} + +int dump_task_extended_fpu(struct task_struct *tsk, struct user_fxsr_struct *fpu) +{ + int fpvalid = tsk_used_math(tsk) && cpu_has_fxsr; + + if (fpvalid) { + if (tsk == current) + unlazy_fpu(tsk); + memcpy(fpu, &tsk->thread.i387.fxsave, sizeof(*fpu)); + } + return fpvalid; +} diff --git a/arch/i386/kernel/i8259.c b/arch/i386/kernel/i8259.c new file mode 100644 index 00000000000..560bef1afb3 --- /dev/null +++ b/arch/i386/kernel/i8259.c @@ -0,0 +1,429 @@ +#include <linux/config.h> +#include <linux/errno.h> +#include <linux/signal.h> +#include <linux/sched.h> +#include <linux/ioport.h> +#include <linux/interrupt.h> +#include <linux/slab.h> +#include <linux/random.h> +#include <linux/smp_lock.h> +#include <linux/init.h> +#include <linux/kernel_stat.h> +#include <linux/sysdev.h> +#include <linux/bitops.h> + +#include <asm/8253pit.h> +#include <asm/atomic.h> +#include <asm/system.h> +#include <asm/io.h> +#include <asm/irq.h> +#include <asm/timer.h> +#include <asm/pgtable.h> +#include <asm/delay.h> +#include <asm/desc.h> +#include <asm/apic.h> +#include <asm/arch_hooks.h> +#include <asm/i8259.h> + +#include <linux/irq.h> + +#include <io_ports.h> + +/* + * This is the 'legacy' 8259A Programmable Interrupt Controller, + * present in the majority of PC/AT boxes. + * plus some generic x86 specific things if generic specifics makes + * any sense at all. + * this file should become arch/i386/kernel/irq.c when the old irq.c + * moves to arch independent land + */ + +DEFINE_SPINLOCK(i8259A_lock); + +static void end_8259A_irq (unsigned int irq) +{ + if (!(irq_desc[irq].status & (IRQ_DISABLED|IRQ_INPROGRESS)) && + irq_desc[irq].action) + enable_8259A_irq(irq); +} + +#define shutdown_8259A_irq disable_8259A_irq + +static void mask_and_ack_8259A(unsigned int); + +unsigned int startup_8259A_irq(unsigned int irq) +{ + enable_8259A_irq(irq); + return 0; /* never anything pending */ +} + +static struct hw_interrupt_type i8259A_irq_type = { + .typename = "XT-PIC", + .startup = startup_8259A_irq, + .shutdown = shutdown_8259A_irq, + .enable = enable_8259A_irq, + .disable = disable_8259A_irq, + .ack = mask_and_ack_8259A, + .end = end_8259A_irq, +}; + +/* + * 8259A PIC functions to handle ISA devices: + */ + +/* + * This contains the irq mask for both 8259A irq controllers, + */ +unsigned int cached_irq_mask = 0xffff; + +/* + * Not all IRQs can be routed through the IO-APIC, eg. on certain (older) + * boards the timer interrupt is not really connected to any IO-APIC pin, + * it's fed to the master 8259A's IR0 line only. + * + * Any '1' bit in this mask means the IRQ is routed through the IO-APIC. + * this 'mixed mode' IRQ handling costs nothing because it's only used + * at IRQ setup time. + */ +unsigned long io_apic_irqs; + +void disable_8259A_irq(unsigned int irq) +{ + unsigned int mask = 1 << irq; + unsigned long flags; + + spin_lock_irqsave(&i8259A_lock, flags); + cached_irq_mask |= mask; + if (irq & 8) + outb(cached_slave_mask, PIC_SLAVE_IMR); + else + outb(cached_master_mask, PIC_MASTER_IMR); + spin_unlock_irqrestore(&i8259A_lock, flags); +} + +void enable_8259A_irq(unsigned int irq) +{ + unsigned int mask = ~(1 << irq); + unsigned long flags; + + spin_lock_irqsave(&i8259A_lock, flags); + cached_irq_mask &= mask; + if (irq & 8) + outb(cached_slave_mask, PIC_SLAVE_IMR); + else + outb(cached_master_mask, PIC_MASTER_IMR); + spin_unlock_irqrestore(&i8259A_lock, flags); +} + +int i8259A_irq_pending(unsigned int irq) +{ + unsigned int mask = 1<<irq; + unsigned long flags; + int ret; + + spin_lock_irqsave(&i8259A_lock, flags); + if (irq < 8) + ret = inb(PIC_MASTER_CMD) & mask; + else + ret = inb(PIC_SLAVE_CMD) & (mask >> 8); + spin_unlock_irqrestore(&i8259A_lock, flags); + + return ret; +} + +void make_8259A_irq(unsigned int irq) +{ + disable_irq_nosync(irq); + io_apic_irqs &= ~(1<<irq); + irq_desc[irq].handler = &i8259A_irq_type; + enable_irq(irq); +} + +/* + * This function assumes to be called rarely. Switching between + * 8259A registers is slow. + * This has to be protected by the irq controller spinlock + * before being called. + */ +static inline int i8259A_irq_real(unsigned int irq) +{ + int value; + int irqmask = 1<<irq; + + if (irq < 8) { + outb(0x0B,PIC_MASTER_CMD); /* ISR register */ + value = inb(PIC_MASTER_CMD) & irqmask; + outb(0x0A,PIC_MASTER_CMD); /* back to the IRR register */ + return value; + } + outb(0x0B,PIC_SLAVE_CMD); /* ISR register */ + value = inb(PIC_SLAVE_CMD) & (irqmask >> 8); + outb(0x0A,PIC_SLAVE_CMD); /* back to the IRR register */ + return value; +} + +/* + * Careful! The 8259A is a fragile beast, it pretty + * much _has_ to be done exactly like this (mask it + * first, _then_ send the EOI, and the order of EOI + * to the two 8259s is important! + */ +static void mask_and_ack_8259A(unsigned int irq) +{ + unsigned int irqmask = 1 << irq; + unsigned long flags; + + spin_lock_irqsave(&i8259A_lock, flags); + /* + * Lightweight spurious IRQ detection. We do not want + * to overdo spurious IRQ handling - it's usually a sign + * of hardware problems, so we only do the checks we can + * do without slowing down good hardware unnecesserily. + * + * Note that IRQ7 and IRQ15 (the two spurious IRQs + * usually resulting from the 8259A-1|2 PICs) occur + * even if the IRQ is masked in the 8259A. Thus we + * can check spurious 8259A IRQs without doing the + * quite slow i8259A_irq_real() call for every IRQ. + * This does not cover 100% of spurious interrupts, + * but should be enough to warn the user that there + * is something bad going on ... + */ + if (cached_irq_mask & irqmask) + goto spurious_8259A_irq; + cached_irq_mask |= irqmask; + +handle_real_irq: + if (irq & 8) { + inb(PIC_SLAVE_IMR); /* DUMMY - (do we need this?) */ + outb(cached_slave_mask, PIC_SLAVE_IMR); + outb(0x60+(irq&7),PIC_SLAVE_CMD);/* 'Specific EOI' to slave */ + outb(0x60+PIC_CASCADE_IR,PIC_MASTER_CMD); /* 'Specific EOI' to master-IRQ2 */ + } else { + inb(PIC_MASTER_IMR); /* DUMMY - (do we need this?) */ + outb(cached_master_mask, PIC_MASTER_IMR); + outb(0x60+irq,PIC_MASTER_CMD); /* 'Specific EOI to master */ + } + spin_unlock_irqrestore(&i8259A_lock, flags); + return; + +spurious_8259A_irq: + /* + * this is the slow path - should happen rarely. + */ + if (i8259A_irq_real(irq)) + /* + * oops, the IRQ _is_ in service according to the + * 8259A - not spurious, go handle it. + */ + goto handle_real_irq; + + { + static int spurious_irq_mask; + /* + * At this point we can be sure the IRQ is spurious, + * lets ACK and report it. [once per IRQ] + */ + if (!(spurious_irq_mask & irqmask)) { + printk(KERN_DEBUG "spurious 8259A interrupt: IRQ%d.\n", irq); + spurious_irq_mask |= irqmask; + } + atomic_inc(&irq_err_count); + /* + * Theoretically we do not have to handle this IRQ, + * but in Linux this does not cause problems and is + * simpler for us. + */ + goto handle_real_irq; + } +} + +static char irq_trigger[2]; +/** + * ELCR registers (0x4d0, 0x4d1) control edge/level of IRQ + */ +static void restore_ELCR(char *trigger) +{ + outb(trigger[0], 0x4d0); + outb(trigger[1], 0x4d1); +} + +static void save_ELCR(char *trigger) +{ + /* IRQ 0,1,2,8,13 are marked as reserved */ + trigger[0] = inb(0x4d0) & 0xF8; + trigger[1] = inb(0x4d1) & 0xDE; +} + +static int i8259A_resume(struct sys_device *dev) +{ + init_8259A(0); + restore_ELCR(irq_trigger); + return 0; +} + +static int i8259A_suspend(struct sys_device *dev, u32 state) +{ + save_ELCR(irq_trigger); + return 0; +} + +static struct sysdev_class i8259_sysdev_class = { + set_kset_name("i8259"), + .suspend = i8259A_suspend, + .resume = i8259A_resume, +}; + +static struct sys_device device_i8259A = { + .id = 0, + .cls = &i8259_sysdev_class, +}; + +static int __init i8259A_init_sysfs(void) +{ + int error = sysdev_class_register(&i8259_sysdev_class); + if (!error) + error = sysdev_register(&device_i8259A); + return error; +} + +device_initcall(i8259A_init_sysfs); + +void init_8259A(int auto_eoi) +{ + unsigned long flags; + + spin_lock_irqsave(&i8259A_lock, flags); + + outb(0xff, PIC_MASTER_IMR); /* mask all of 8259A-1 */ + outb(0xff, PIC_SLAVE_IMR); /* mask all of 8259A-2 */ + + /* + * outb_p - this has to work on a wide range of PC hardware. + */ + outb_p(0x11, PIC_MASTER_CMD); /* ICW1: select 8259A-1 init */ + outb_p(0x20 + 0, PIC_MASTER_IMR); /* ICW2: 8259A-1 IR0-7 mapped to 0x20-0x27 */ + outb_p(1U << PIC_CASCADE_IR, PIC_MASTER_IMR); /* 8259A-1 (the master) has a slave on IR2 */ + if (auto_eoi) /* master does Auto EOI */ + outb_p(MASTER_ICW4_DEFAULT | PIC_ICW4_AEOI, PIC_MASTER_IMR); + else /* master expects normal EOI */ + outb_p(MASTER_ICW4_DEFAULT, PIC_MASTER_IMR); + + outb_p(0x11, PIC_SLAVE_CMD); /* ICW1: select 8259A-2 init */ + outb_p(0x20 + 8, PIC_SLAVE_IMR); /* ICW2: 8259A-2 IR0-7 mapped to 0x28-0x2f */ + outb_p(PIC_CASCADE_IR, PIC_SLAVE_IMR); /* 8259A-2 is a slave on master's IR2 */ + outb_p(SLAVE_ICW4_DEFAULT, PIC_SLAVE_IMR); /* (slave's support for AEOI in flat mode is to be investigated) */ + if (auto_eoi) + /* + * in AEOI mode we just have to mask the interrupt + * when acking. + */ + i8259A_irq_type.ack = disable_8259A_irq; + else + i8259A_irq_type.ack = mask_and_ack_8259A; + + udelay(100); /* wait for 8259A to initialize */ + + outb(cached_master_mask, PIC_MASTER_IMR); /* restore master IRQ mask */ + outb(cached_slave_mask, PIC_SLAVE_IMR); /* restore slave IRQ mask */ + + spin_unlock_irqrestore(&i8259A_lock, flags); +} + +/* + * Note that on a 486, we don't want to do a SIGFPE on an irq13 + * as the irq is unreliable, and exception 16 works correctly + * (ie as explained in the intel literature). On a 386, you + * can't use exception 16 due to bad IBM design, so we have to + * rely on the less exact irq13. + * + * Careful.. Not only is IRQ13 unreliable, but it is also + * leads to races. IBM designers who came up with it should + * be shot. + */ + + +static irqreturn_t math_error_irq(int cpl, void *dev_id, struct pt_regs *regs) +{ + extern void math_error(void __user *); + outb(0,0xF0); + if (ignore_fpu_irq || !boot_cpu_data.hard_math) + return IRQ_NONE; + math_error((void __user *)regs->eip); + return IRQ_HANDLED; +} + +/* + * New motherboards sometimes make IRQ 13 be a PCI interrupt, + * so allow interrupt sharing. + */ +static struct irqaction fpu_irq = { math_error_irq, 0, CPU_MASK_NONE, "fpu", NULL, NULL }; + +void __init init_ISA_irqs (void) +{ + int i; + +#ifdef CONFIG_X86_LOCAL_APIC + init_bsp_APIC(); +#endif + init_8259A(0); + + for (i = 0; i < NR_IRQS; i++) { + irq_desc[i].status = IRQ_DISABLED; + irq_desc[i].action = NULL; + irq_desc[i].depth = 1; + + if (i < 16) { + /* + * 16 old-style INTA-cycle interrupts: + */ + irq_desc[i].handler = &i8259A_irq_type; + } else { + /* + * 'high' PCI IRQs filled in on demand + */ + irq_desc[i].handler = &no_irq_type; + } + } +} + +void __init init_IRQ(void) +{ + int i; + + /* all the set up before the call gates are initialised */ + pre_intr_init_hook(); + + /* + * Cover the whole vector space, no vector can escape + * us. (some of these will be overridden and become + * 'special' SMP interrupts) + */ + for (i = 0; i < (NR_VECTORS - FIRST_EXTERNAL_VECTOR); i++) { + int vector = FIRST_EXTERNAL_VECTOR + i; + if (i >= NR_IRQS) + break; + if (vector != SYSCALL_VECTOR) + set_intr_gate(vector, interrupt[i]); + } + + /* setup after call gates are initialised (usually add in + * the architecture specific gates) + */ + intr_init_hook(); + + /* + * Set the clock to HZ Hz, we already have a valid + * vector now: + */ + setup_pit_timer(); + + /* + * External FPU? Set up irq13 if so, for + * original braindamaged IBM FERR coupling. + */ + if (boot_cpu_data.hard_math && !cpu_has_fpu) + setup_irq(FPU_IRQ, &fpu_irq); + + irq_ctx_init(smp_processor_id()); +} diff --git a/arch/i386/kernel/init_task.c b/arch/i386/kernel/init_task.c new file mode 100644 index 00000000000..9caa8e8db80 --- /dev/null +++ b/arch/i386/kernel/init_task.c @@ -0,0 +1,46 @@ +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/sched.h> +#include <linux/init.h> +#include <linux/init_task.h> +#include <linux/fs.h> +#include <linux/mqueue.h> + +#include <asm/uaccess.h> +#include <asm/pgtable.h> +#include <asm/desc.h> + +static struct fs_struct init_fs = INIT_FS; +static struct files_struct init_files = INIT_FILES; +static struct signal_struct init_signals = INIT_SIGNALS(init_signals); +static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand); +struct mm_struct init_mm = INIT_MM(init_mm); + +EXPORT_SYMBOL(init_mm); + +/* + * Initial thread structure. + * + * We need to make sure that this is THREAD_SIZE aligned due to the + * way process stacks are handled. This is done by having a special + * "init_task" linker map entry.. + */ +union thread_union init_thread_union + __attribute__((__section__(".data.init_task"))) = + { INIT_THREAD_INFO(init_task) }; + +/* + * Initial task structure. + * + * All other task structs will be allocated on slabs in fork.c + */ +struct task_struct init_task = INIT_TASK(init_task); + +EXPORT_SYMBOL(init_task); + +/* + * per-CPU TSS segments. Threads are completely 'soft' on Linux, + * no more per-task TSS's. + */ +DEFINE_PER_CPU(struct tss_struct, init_tss) ____cacheline_maxaligned_in_smp = INIT_TSS; + diff --git a/arch/i386/kernel/io_apic.c b/arch/i386/kernel/io_apic.c new file mode 100644 index 00000000000..9c1350e811d --- /dev/null +++ b/arch/i386/kernel/io_apic.c @@ -0,0 +1,2545 @@ +/* + * Intel IO-APIC support for multi-Pentium hosts. + * + * Copyright (C) 1997, 1998, 1999, 2000 Ingo Molnar, Hajnalka Szabo + * + * Many thanks to Stig Venaas for trying out countless experimental + * patches and reporting/debugging problems patiently! + * + * (c) 1999, Multiple IO-APIC support, developed by + * Ken-ichi Yaku <yaku@css1.kbnes.nec.co.jp> and + * Hidemi Kishimoto <kisimoto@css1.kbnes.nec.co.jp>, + * further tested and cleaned up by Zach Brown <zab@redhat.com> + * and Ingo Molnar <mingo@redhat.com> + * + * Fixes + * Maciej W. Rozycki : Bits for genuine 82489DX APICs; + * thanks to Eric Gilmore + * and Rolf G. Tews + * for testing these extensively + * Paul Diefenbaugh : Added full ACPI support + */ + +#include <linux/mm.h> +#include <linux/irq.h> +#include <linux/interrupt.h> +#include <linux/init.h> +#include <linux/delay.h> +#include <linux/sched.h> +#include <linux/config.h> +#include <linux/smp_lock.h> +#include <linux/mc146818rtc.h> +#include <linux/compiler.h> +#include <linux/acpi.h> + +#include <linux/sysdev.h> +#include <asm/io.h> +#include <asm/smp.h> +#include <asm/desc.h> +#include <asm/timer.h> + +#include <mach_apic.h> + +#include "io_ports.h" + +int (*ioapic_renumber_irq)(int ioapic, int irq); +atomic_t irq_mis_count; + +static DEFINE_SPINLOCK(ioapic_lock); + +/* + * Is the SiS APIC rmw bug present ? + * -1 = don't know, 0 = no, 1 = yes + */ +int sis_apic_bug = -1; + +/* + * # of IRQ routing registers + */ +int nr_ioapic_registers[MAX_IO_APICS]; + +/* + * Rough estimation of how many shared IRQs there are, can + * be changed anytime. + */ +#define MAX_PLUS_SHARED_IRQS NR_IRQS +#define PIN_MAP_SIZE (MAX_PLUS_SHARED_IRQS + NR_IRQS) + +/* + * This is performance-critical, we want to do it O(1) + * + * the indexing order of this array favors 1:1 mappings + * between pins and IRQs. + */ + +static struct irq_pin_list { + int apic, pin, next; +} irq_2_pin[PIN_MAP_SIZE]; + +int vector_irq[NR_VECTORS] = { [0 ... NR_VECTORS - 1] = -1}; +#ifdef CONFIG_PCI_MSI +#define vector_to_irq(vector) \ + (platform_legacy_irq(vector) ? vector : vector_irq[vector]) +#else +#define vector_to_irq(vector) (vector) +#endif + +/* + * The common case is 1:1 IRQ<->pin mappings. Sometimes there are + * shared ISA-space IRQs, so we have to support them. We are super + * fast in the common case, and fast for shared ISA-space IRQs. + */ +static void add_pin_to_irq(unsigned int irq, int apic, int pin) +{ + static int first_free_entry = NR_IRQS; + struct irq_pin_list *entry = irq_2_pin + irq; + + while (entry->next) + entry = irq_2_pin + entry->next; + + if (entry->pin != -1) { + entry->next = first_free_entry; + entry = irq_2_pin + entry->next; + if (++first_free_entry >= PIN_MAP_SIZE) + panic("io_apic.c: whoops"); + } + entry->apic = apic; + entry->pin = pin; +} + +/* + * Reroute an IRQ to a different pin. + */ +static void __init replace_pin_at_irq(unsigned int irq, + int oldapic, int oldpin, + int newapic, int newpin) +{ + struct irq_pin_list *entry = irq_2_pin + irq; + + while (1) { + if (entry->apic == oldapic && entry->pin == oldpin) { + entry->apic = newapic; + entry->pin = newpin; + } + if (!entry->next) + break; + entry = irq_2_pin + entry->next; + } +} + +static void __modify_IO_APIC_irq (unsigned int irq, unsigned long enable, unsigned long disable) +{ + struct irq_pin_list *entry = irq_2_pin + irq; + unsigned int pin, reg; + + for (;;) { + pin = entry->pin; + if (pin == -1) + break; + reg = io_apic_read(entry->apic, 0x10 + pin*2); + reg &= ~disable; + reg |= enable; + io_apic_modify(entry->apic, 0x10 + pin*2, reg); + if (!entry->next) + break; + entry = irq_2_pin + entry->next; + } +} + +/* mask = 1 */ +static void __mask_IO_APIC_irq (unsigned int irq) +{ + __modify_IO_APIC_irq(irq, 0x00010000, 0); +} + +/* mask = 0 */ +static void __unmask_IO_APIC_irq (unsigned int irq) +{ + __modify_IO_APIC_irq(irq, 0, 0x00010000); +} + +/* mask = 1, trigger = 0 */ +static void __mask_and_edge_IO_APIC_irq (unsigned int irq) +{ + __modify_IO_APIC_irq(irq, 0x00010000, 0x00008000); +} + +/* mask = 0, trigger = 1 */ +static void __unmask_and_level_IO_APIC_irq (unsigned int irq) +{ + __modify_IO_APIC_irq(irq, 0x00008000, 0x00010000); +} + +static void mask_IO_APIC_irq (unsigned int irq) +{ + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + __mask_IO_APIC_irq(irq); + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +static void unmask_IO_APIC_irq (unsigned int irq) +{ + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + __unmask_IO_APIC_irq(irq); + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +static void clear_IO_APIC_pin(unsigned int apic, unsigned int pin) +{ + struct IO_APIC_route_entry entry; + unsigned long flags; + + /* Check delivery_mode to be sure we're not clearing an SMI pin */ + spin_lock_irqsave(&ioapic_lock, flags); + *(((int*)&entry) + 0) = io_apic_read(apic, 0x10 + 2 * pin); + *(((int*)&entry) + 1) = io_apic_read(apic, 0x11 + 2 * pin); + spin_unlock_irqrestore(&ioapic_lock, flags); + if (entry.delivery_mode == dest_SMI) + return; + + /* + * Disable it in the IO-APIC irq-routing table: + */ + memset(&entry, 0, sizeof(entry)); + entry.mask = 1; + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(apic, 0x10 + 2 * pin, *(((int *)&entry) + 0)); + io_apic_write(apic, 0x11 + 2 * pin, *(((int *)&entry) + 1)); + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +static void clear_IO_APIC (void) +{ + int apic, pin; + + for (apic = 0; apic < nr_ioapics; apic++) + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) + clear_IO_APIC_pin(apic, pin); +} + +static void set_ioapic_affinity_irq(unsigned int irq, cpumask_t cpumask) +{ + unsigned long flags; + int pin; + struct irq_pin_list *entry = irq_2_pin + irq; + unsigned int apicid_value; + + apicid_value = cpu_mask_to_apicid(cpumask); + /* Prepare to do the io_apic_write */ + apicid_value = apicid_value << 24; + spin_lock_irqsave(&ioapic_lock, flags); + for (;;) { + pin = entry->pin; + if (pin == -1) + break; + io_apic_write(entry->apic, 0x10 + 1 + pin*2, apicid_value); + if (!entry->next) + break; + entry = irq_2_pin + entry->next; + } + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +#if defined(CONFIG_IRQBALANCE) +# include <asm/processor.h> /* kernel_thread() */ +# include <linux/kernel_stat.h> /* kstat */ +# include <linux/slab.h> /* kmalloc() */ +# include <linux/timer.h> /* time_after() */ + +# ifdef CONFIG_BALANCED_IRQ_DEBUG +# define TDprintk(x...) do { printk("<%ld:%s:%d>: ", jiffies, __FILE__, __LINE__); printk(x); } while (0) +# define Dprintk(x...) do { TDprintk(x); } while (0) +# else +# define TDprintk(x...) +# define Dprintk(x...) +# endif + +cpumask_t __cacheline_aligned pending_irq_balance_cpumask[NR_IRQS]; + +#define IRQBALANCE_CHECK_ARCH -999 +static int irqbalance_disabled = IRQBALANCE_CHECK_ARCH; +static int physical_balance = 0; + +static struct irq_cpu_info { + unsigned long * last_irq; + unsigned long * irq_delta; + unsigned long irq; +} irq_cpu_data[NR_CPUS]; + +#define CPU_IRQ(cpu) (irq_cpu_data[cpu].irq) +#define LAST_CPU_IRQ(cpu,irq) (irq_cpu_data[cpu].last_irq[irq]) +#define IRQ_DELTA(cpu,irq) (irq_cpu_data[cpu].irq_delta[irq]) + +#define IDLE_ENOUGH(cpu,now) \ + (idle_cpu(cpu) && ((now) - per_cpu(irq_stat, (cpu)).idle_timestamp > 1)) + +#define IRQ_ALLOWED(cpu, allowed_mask) cpu_isset(cpu, allowed_mask) + +#define CPU_TO_PACKAGEINDEX(i) (first_cpu(cpu_sibling_map[i])) + +#define MAX_BALANCED_IRQ_INTERVAL (5*HZ) +#define MIN_BALANCED_IRQ_INTERVAL (HZ/2) +#define BALANCED_IRQ_MORE_DELTA (HZ/10) +#define BALANCED_IRQ_LESS_DELTA (HZ) + +static long balanced_irq_interval = MAX_BALANCED_IRQ_INTERVAL; + +static unsigned long move(int curr_cpu, cpumask_t allowed_mask, + unsigned long now, int direction) +{ + int search_idle = 1; + int cpu = curr_cpu; + + goto inside; + + do { + if (unlikely(cpu == curr_cpu)) + search_idle = 0; +inside: + if (direction == 1) { + cpu++; + if (cpu >= NR_CPUS) + cpu = 0; + } else { + cpu--; + if (cpu == -1) + cpu = NR_CPUS-1; + } + } while (!cpu_online(cpu) || !IRQ_ALLOWED(cpu,allowed_mask) || + (search_idle && !IDLE_ENOUGH(cpu,now))); + + return cpu; +} + +static inline void balance_irq(int cpu, int irq) +{ + unsigned long now = jiffies; + cpumask_t allowed_mask; + unsigned int new_cpu; + + if (irqbalance_disabled) + return; + + cpus_and(allowed_mask, cpu_online_map, irq_affinity[irq]); + new_cpu = move(cpu, allowed_mask, now, 1); + if (cpu != new_cpu) { + irq_desc_t *desc = irq_desc + irq; + unsigned long flags; + + spin_lock_irqsave(&desc->lock, flags); + pending_irq_balance_cpumask[irq] = cpumask_of_cpu(new_cpu); + spin_unlock_irqrestore(&desc->lock, flags); + } +} + +static inline void rotate_irqs_among_cpus(unsigned long useful_load_threshold) +{ + int i, j; + Dprintk("Rotating IRQs among CPUs.\n"); + for (i = 0; i < NR_CPUS; i++) { + for (j = 0; cpu_online(i) && (j < NR_IRQS); j++) { + if (!irq_desc[j].action) + continue; + /* Is it a significant load ? */ + if (IRQ_DELTA(CPU_TO_PACKAGEINDEX(i),j) < + useful_load_threshold) + continue; + balance_irq(i, j); + } + } + balanced_irq_interval = max((long)MIN_BALANCED_IRQ_INTERVAL, + balanced_irq_interval - BALANCED_IRQ_LESS_DELTA); + return; +} + +static void do_irq_balance(void) +{ + int i, j; + unsigned long max_cpu_irq = 0, min_cpu_irq = (~0); + unsigned long move_this_load = 0; + int max_loaded = 0, min_loaded = 0; + int load; + unsigned long useful_load_threshold = balanced_irq_interval + 10; + int selected_irq; + int tmp_loaded, first_attempt = 1; + unsigned long tmp_cpu_irq; + unsigned long imbalance = 0; + cpumask_t allowed_mask, target_cpu_mask, tmp; + + for (i = 0; i < NR_CPUS; i++) { + int package_index; + CPU_IRQ(i) = 0; + if (!cpu_online(i)) + continue; + package_index = CPU_TO_PACKAGEINDEX(i); + for (j = 0; j < NR_IRQS; j++) { + unsigned long value_now, delta; + /* Is this an active IRQ? */ + if (!irq_desc[j].action) + continue; + if ( package_index == i ) + IRQ_DELTA(package_index,j) = 0; + /* Determine the total count per processor per IRQ */ + value_now = (unsigned long) kstat_cpu(i).irqs[j]; + + /* Determine the activity per processor per IRQ */ + delta = value_now - LAST_CPU_IRQ(i,j); + + /* Update last_cpu_irq[][] for the next time */ + LAST_CPU_IRQ(i,j) = value_now; + + /* Ignore IRQs whose rate is less than the clock */ + if (delta < useful_load_threshold) + continue; + /* update the load for the processor or package total */ + IRQ_DELTA(package_index,j) += delta; + + /* Keep track of the higher numbered sibling as well */ + if (i != package_index) + CPU_IRQ(i) += delta; + /* + * We have sibling A and sibling B in the package + * + * cpu_irq[A] = load for cpu A + load for cpu B + * cpu_irq[B] = load for cpu B + */ + CPU_IRQ(package_index) += delta; + } + } + /* Find the least loaded processor package */ + for (i = 0; i < NR_CPUS; i++) { + if (!cpu_online(i)) + continue; + if (i != CPU_TO_PACKAGEINDEX(i)) + continue; + if (min_cpu_irq > CPU_IRQ(i)) { + min_cpu_irq = CPU_IRQ(i); + min_loaded = i; + } + } + max_cpu_irq = ULONG_MAX; + +tryanothercpu: + /* Look for heaviest loaded processor. + * We may come back to get the next heaviest loaded processor. + * Skip processors with trivial loads. + */ + tmp_cpu_irq = 0; + tmp_loaded = -1; + for (i = 0; i < NR_CPUS; i++) { + if (!cpu_online(i)) + continue; + if (i != CPU_TO_PACKAGEINDEX(i)) + continue; + if (max_cpu_irq <= CPU_IRQ(i)) + continue; + if (tmp_cpu_irq < CPU_IRQ(i)) { + tmp_cpu_irq = CPU_IRQ(i); + tmp_loaded = i; + } + } + + if (tmp_loaded == -1) { + /* In the case of small number of heavy interrupt sources, + * loading some of the cpus too much. We use Ingo's original + * approach to rotate them around. + */ + if (!first_attempt && imbalance >= useful_load_threshold) { + rotate_irqs_among_cpus(useful_load_threshold); + return; + } + goto not_worth_the_effort; + } + + first_attempt = 0; /* heaviest search */ + max_cpu_irq = tmp_cpu_irq; /* load */ + max_loaded = tmp_loaded; /* processor */ + imbalance = (max_cpu_irq - min_cpu_irq) / 2; + + Dprintk("max_loaded cpu = %d\n", max_loaded); + Dprintk("min_loaded cpu = %d\n", min_loaded); + Dprintk("max_cpu_irq load = %ld\n", max_cpu_irq); + Dprintk("min_cpu_irq load = %ld\n", min_cpu_irq); + Dprintk("load imbalance = %lu\n", imbalance); + + /* if imbalance is less than approx 10% of max load, then + * observe diminishing returns action. - quit + */ + if (imbalance < (max_cpu_irq >> 3)) { + Dprintk("Imbalance too trivial\n"); + goto not_worth_the_effort; + } + +tryanotherirq: + /* if we select an IRQ to move that can't go where we want, then + * see if there is another one to try. + */ + move_this_load = 0; + selected_irq = -1; + for (j = 0; j < NR_IRQS; j++) { + /* Is this an active IRQ? */ + if (!irq_desc[j].action) + continue; + if (imbalance <= IRQ_DELTA(max_loaded,j)) + continue; + /* Try to find the IRQ that is closest to the imbalance + * without going over. + */ + if (move_this_load < IRQ_DELTA(max_loaded,j)) { + move_this_load = IRQ_DELTA(max_loaded,j); + selected_irq = j; + } + } + if (selected_irq == -1) { + goto tryanothercpu; + } + + imbalance = move_this_load; + + /* For physical_balance case, we accumlated both load + * values in the one of the siblings cpu_irq[], + * to use the same code for physical and logical processors + * as much as possible. + * + * NOTE: the cpu_irq[] array holds the sum of the load for + * sibling A and sibling B in the slot for the lowest numbered + * sibling (A), _AND_ the load for sibling B in the slot for + * the higher numbered sibling. + * + * We seek the least loaded sibling by making the comparison + * (A+B)/2 vs B + */ + load = CPU_IRQ(min_loaded) >> 1; + for_each_cpu_mask(j, cpu_sibling_map[min_loaded]) { + if (load > CPU_IRQ(j)) { + /* This won't change cpu_sibling_map[min_loaded] */ + load = CPU_IRQ(j); + min_loaded = j; + } + } + + cpus_and(allowed_mask, cpu_online_map, irq_affinity[selected_irq]); + target_cpu_mask = cpumask_of_cpu(min_loaded); + cpus_and(tmp, target_cpu_mask, allowed_mask); + + if (!cpus_empty(tmp)) { + irq_desc_t *desc = irq_desc + selected_irq; + unsigned long flags; + + Dprintk("irq = %d moved to cpu = %d\n", + selected_irq, min_loaded); + /* mark for change destination */ + spin_lock_irqsave(&desc->lock, flags); + pending_irq_balance_cpumask[selected_irq] = + cpumask_of_cpu(min_loaded); + spin_unlock_irqrestore(&desc->lock, flags); + /* Since we made a change, come back sooner to + * check for more variation. + */ + balanced_irq_interval = max((long)MIN_BALANCED_IRQ_INTERVAL, + balanced_irq_interval - BALANCED_IRQ_LESS_DELTA); + return; + } + goto tryanotherirq; + +not_worth_the_effort: + /* + * if we did not find an IRQ to move, then adjust the time interval + * upward + */ + balanced_irq_interval = min((long)MAX_BALANCED_IRQ_INTERVAL, + balanced_irq_interval + BALANCED_IRQ_MORE_DELTA); + Dprintk("IRQ worth rotating not found\n"); + return; +} + +static int balanced_irq(void *unused) +{ + int i; + unsigned long prev_balance_time = jiffies; + long time_remaining = balanced_irq_interval; + + daemonize("kirqd"); + + /* push everything to CPU 0 to give us a starting point. */ + for (i = 0 ; i < NR_IRQS ; i++) { + pending_irq_balance_cpumask[i] = cpumask_of_cpu(0); + } + + for ( ; ; ) { + set_current_state(TASK_INTERRUPTIBLE); + time_remaining = schedule_timeout(time_remaining); + try_to_freeze(PF_FREEZE); + if (time_after(jiffies, + prev_balance_time+balanced_irq_interval)) { + do_irq_balance(); + prev_balance_time = jiffies; + time_remaining = balanced_irq_interval; + } + } + return 0; +} + +static int __init balanced_irq_init(void) +{ + int i; + struct cpuinfo_x86 *c; + cpumask_t tmp; + + cpus_shift_right(tmp, cpu_online_map, 2); + c = &boot_cpu_data; + /* When not overwritten by the command line ask subarchitecture. */ + if (irqbalance_disabled == IRQBALANCE_CHECK_ARCH) + irqbalance_disabled = NO_BALANCE_IRQ; + if (irqbalance_disabled) + return 0; + + /* disable irqbalance completely if there is only one processor online */ + if (num_online_cpus() < 2) { + irqbalance_disabled = 1; + return 0; + } + /* + * Enable physical balance only if more than 1 physical processor + * is present + */ + if (smp_num_siblings > 1 && !cpus_empty(tmp)) + physical_balance = 1; + + for (i = 0; i < NR_CPUS; i++) { + if (!cpu_online(i)) + continue; + irq_cpu_data[i].irq_delta = kmalloc(sizeof(unsigned long) * NR_IRQS, GFP_KERNEL); + irq_cpu_data[i].last_irq = kmalloc(sizeof(unsigned long) * NR_IRQS, GFP_KERNEL); + if (irq_cpu_data[i].irq_delta == NULL || irq_cpu_data[i].last_irq == NULL) { + printk(KERN_ERR "balanced_irq_init: out of memory"); + goto failed; + } + memset(irq_cpu_data[i].irq_delta,0,sizeof(unsigned long) * NR_IRQS); + memset(irq_cpu_data[i].last_irq,0,sizeof(unsigned long) * NR_IRQS); + } + + printk(KERN_INFO "Starting balanced_irq\n"); + if (kernel_thread(balanced_irq, NULL, CLONE_KERNEL) >= 0) + return 0; + else + printk(KERN_ERR "balanced_irq_init: failed to spawn balanced_irq"); +failed: + for (i = 0; i < NR_CPUS; i++) { + if(irq_cpu_data[i].irq_delta) + kfree(irq_cpu_data[i].irq_delta); + if(irq_cpu_data[i].last_irq) + kfree(irq_cpu_data[i].last_irq); + } + return 0; +} + +int __init irqbalance_disable(char *str) +{ + irqbalance_disabled = 1; + return 0; +} + +__setup("noirqbalance", irqbalance_disable); + +static inline void move_irq(int irq) +{ + /* note - we hold the desc->lock */ + if (unlikely(!cpus_empty(pending_irq_balance_cpumask[irq]))) { + set_ioapic_affinity_irq(irq, pending_irq_balance_cpumask[irq]); + cpus_clear(pending_irq_balance_cpumask[irq]); + } +} + +late_initcall(balanced_irq_init); + +#else /* !CONFIG_IRQBALANCE */ +static inline void move_irq(int irq) { } +#endif /* CONFIG_IRQBALANCE */ + +#ifndef CONFIG_SMP +void fastcall send_IPI_self(int vector) +{ + unsigned int cfg; + + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + cfg = APIC_DM_FIXED | APIC_DEST_SELF | vector | APIC_DEST_LOGICAL; + /* + * Send the IPI. The write to APIC_ICR fires this off. + */ + apic_write_around(APIC_ICR, cfg); +} +#endif /* !CONFIG_SMP */ + + +/* + * support for broken MP BIOSs, enables hand-redirection of PIRQ0-7 to + * specific CPU-side IRQs. + */ + +#define MAX_PIRQS 8 +static int pirq_entries [MAX_PIRQS]; +static int pirqs_enabled; +int skip_ioapic_setup; + +static int __init ioapic_setup(char *str) +{ + skip_ioapic_setup = 1; + return 1; +} + +__setup("noapic", ioapic_setup); + +static int __init ioapic_pirq_setup(char *str) +{ + int i, max; + int ints[MAX_PIRQS+1]; + + get_options(str, ARRAY_SIZE(ints), ints); + + for (i = 0; i < MAX_PIRQS; i++) + pirq_entries[i] = -1; + + pirqs_enabled = 1; + apic_printk(APIC_VERBOSE, KERN_INFO + "PIRQ redirection, working around broken MP-BIOS.\n"); + max = MAX_PIRQS; + if (ints[0] < MAX_PIRQS) + max = ints[0]; + + for (i = 0; i < max; i++) { + apic_printk(APIC_VERBOSE, KERN_DEBUG + "... PIRQ%d -> IRQ %d\n", i, ints[i+1]); + /* + * PIRQs are mapped upside down, usually. + */ + pirq_entries[MAX_PIRQS-i-1] = ints[i+1]; + } + return 1; +} + +__setup("pirq=", ioapic_pirq_setup); + +/* + * Find the IRQ entry number of a certain pin. + */ +static int find_irq_entry(int apic, int pin, int type) +{ + int i; + + for (i = 0; i < mp_irq_entries; i++) + if (mp_irqs[i].mpc_irqtype == type && + (mp_irqs[i].mpc_dstapic == mp_ioapics[apic].mpc_apicid || + mp_irqs[i].mpc_dstapic == MP_APIC_ALL) && + mp_irqs[i].mpc_dstirq == pin) + return i; + + return -1; +} + +/* + * Find the pin to which IRQ[irq] (ISA) is connected + */ +static int find_isa_irq_pin(int irq, int type) +{ + int i; + + for (i = 0; i < mp_irq_entries; i++) { + int lbus = mp_irqs[i].mpc_srcbus; + + if ((mp_bus_id_to_type[lbus] == MP_BUS_ISA || + mp_bus_id_to_type[lbus] == MP_BUS_EISA || + mp_bus_id_to_type[lbus] == MP_BUS_MCA || + mp_bus_id_to_type[lbus] == MP_BUS_NEC98 + ) && + (mp_irqs[i].mpc_irqtype == type) && + (mp_irqs[i].mpc_srcbusirq == irq)) + + return mp_irqs[i].mpc_dstirq; + } + return -1; +} + +/* + * Find a specific PCI IRQ entry. + * Not an __init, possibly needed by modules + */ +static int pin_2_irq(int idx, int apic, int pin); + +int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin) +{ + int apic, i, best_guess = -1; + + apic_printk(APIC_DEBUG, "querying PCI -> IRQ mapping bus:%d, " + "slot:%d, pin:%d.\n", bus, slot, pin); + if (mp_bus_id_to_pci_bus[bus] == -1) { + printk(KERN_WARNING "PCI BIOS passed nonexistent PCI bus %d!\n", bus); + return -1; + } + for (i = 0; i < mp_irq_entries; i++) { + int lbus = mp_irqs[i].mpc_srcbus; + + for (apic = 0; apic < nr_ioapics; apic++) + if (mp_ioapics[apic].mpc_apicid == mp_irqs[i].mpc_dstapic || + mp_irqs[i].mpc_dstapic == MP_APIC_ALL) + break; + + if ((mp_bus_id_to_type[lbus] == MP_BUS_PCI) && + !mp_irqs[i].mpc_irqtype && + (bus == lbus) && + (slot == ((mp_irqs[i].mpc_srcbusirq >> 2) & 0x1f))) { + int irq = pin_2_irq(i,apic,mp_irqs[i].mpc_dstirq); + + if (!(apic || IO_APIC_IRQ(irq))) + continue; + + if (pin == (mp_irqs[i].mpc_srcbusirq & 3)) + return irq; + /* + * Use the first all-but-pin matching entry as a + * best-guess fuzzy result for broken mptables. + */ + if (best_guess < 0) + best_guess = irq; + } + } + return best_guess; +} + +/* + * This function currently is only a helper for the i386 smp boot process where + * we need to reprogram the ioredtbls to cater for the cpus which have come online + * so mask in all cases should simply be TARGET_CPUS + */ +void __init setup_ioapic_dest(void) +{ + int pin, ioapic, irq, irq_entry; + + if (skip_ioapic_setup == 1) + return; + + for (ioapic = 0; ioapic < nr_ioapics; ioapic++) { + for (pin = 0; pin < nr_ioapic_registers[ioapic]; pin++) { + irq_entry = find_irq_entry(ioapic, pin, mp_INT); + if (irq_entry == -1) + continue; + irq = pin_2_irq(irq_entry, ioapic, pin); + set_ioapic_affinity_irq(irq, TARGET_CPUS); + } + + } +} + +/* + * EISA Edge/Level control register, ELCR + */ +static int EISA_ELCR(unsigned int irq) +{ + if (irq < 16) { + unsigned int port = 0x4d0 + (irq >> 3); + return (inb(port) >> (irq & 7)) & 1; + } + apic_printk(APIC_VERBOSE, KERN_INFO + "Broken MPtable reports ISA irq %d\n", irq); + return 0; +} + +/* EISA interrupts are always polarity zero and can be edge or level + * trigger depending on the ELCR value. If an interrupt is listed as + * EISA conforming in the MP table, that means its trigger type must + * be read in from the ELCR */ + +#define default_EISA_trigger(idx) (EISA_ELCR(mp_irqs[idx].mpc_srcbusirq)) +#define default_EISA_polarity(idx) (0) + +/* ISA interrupts are always polarity zero edge triggered, + * when listed as conforming in the MP table. */ + +#define default_ISA_trigger(idx) (0) +#define default_ISA_polarity(idx) (0) + +/* PCI interrupts are always polarity one level triggered, + * when listed as conforming in the MP table. */ + +#define default_PCI_trigger(idx) (1) +#define default_PCI_polarity(idx) (1) + +/* MCA interrupts are always polarity zero level triggered, + * when listed as conforming in the MP table. */ + +#define default_MCA_trigger(idx) (1) +#define default_MCA_polarity(idx) (0) + +/* NEC98 interrupts are always polarity zero edge triggered, + * when listed as conforming in the MP table. */ + +#define default_NEC98_trigger(idx) (0) +#define default_NEC98_polarity(idx) (0) + +static int __init MPBIOS_polarity(int idx) +{ + int bus = mp_irqs[idx].mpc_srcbus; + int polarity; + + /* + * Determine IRQ line polarity (high active or low active): + */ + switch (mp_irqs[idx].mpc_irqflag & 3) + { + case 0: /* conforms, ie. bus-type dependent polarity */ + { + switch (mp_bus_id_to_type[bus]) + { + case MP_BUS_ISA: /* ISA pin */ + { + polarity = default_ISA_polarity(idx); + break; + } + case MP_BUS_EISA: /* EISA pin */ + { + polarity = default_EISA_polarity(idx); + break; + } + case MP_BUS_PCI: /* PCI pin */ + { + polarity = default_PCI_polarity(idx); + break; + } + case MP_BUS_MCA: /* MCA pin */ + { + polarity = default_MCA_polarity(idx); + break; + } + case MP_BUS_NEC98: /* NEC 98 pin */ + { + polarity = default_NEC98_polarity(idx); + break; + } + default: + { + printk(KERN_WARNING "broken BIOS!!\n"); + polarity = 1; + break; + } + } + break; + } + case 1: /* high active */ + { + polarity = 0; + break; + } + case 2: /* reserved */ + { + printk(KERN_WARNING "broken BIOS!!\n"); + polarity = 1; + break; + } + case 3: /* low active */ + { + polarity = 1; + break; + } + default: /* invalid */ + { + printk(KERN_WARNING "broken BIOS!!\n"); + polarity = 1; + break; + } + } + return polarity; +} + +static int MPBIOS_trigger(int idx) +{ + int bus = mp_irqs[idx].mpc_srcbus; + int trigger; + + /* + * Determine IRQ trigger mode (edge or level sensitive): + */ + switch ((mp_irqs[idx].mpc_irqflag>>2) & 3) + { + case 0: /* conforms, ie. bus-type dependent */ + { + switch (mp_bus_id_to_type[bus]) + { + case MP_BUS_ISA: /* ISA pin */ + { + trigger = default_ISA_trigger(idx); + break; + } + case MP_BUS_EISA: /* EISA pin */ + { + trigger = default_EISA_trigger(idx); + break; + } + case MP_BUS_PCI: /* PCI pin */ + { + trigger = default_PCI_trigger(idx); + break; + } + case MP_BUS_MCA: /* MCA pin */ + { + trigger = default_MCA_trigger(idx); + break; + } + case MP_BUS_NEC98: /* NEC 98 pin */ + { + trigger = default_NEC98_trigger(idx); + break; + } + default: + { + printk(KERN_WARNING "broken BIOS!!\n"); + trigger = 1; + break; + } + } + break; + } + case 1: /* edge */ + { + trigger = 0; + break; + } + case 2: /* reserved */ + { + printk(KERN_WARNING "broken BIOS!!\n"); + trigger = 1; + break; + } + case 3: /* level */ + { + trigger = 1; + break; + } + default: /* invalid */ + { + printk(KERN_WARNING "broken BIOS!!\n"); + trigger = 0; + break; + } + } + return trigger; +} + +static inline int irq_polarity(int idx) +{ + return MPBIOS_polarity(idx); +} + +static inline int irq_trigger(int idx) +{ + return MPBIOS_trigger(idx); +} + +static int pin_2_irq(int idx, int apic, int pin) +{ + int irq, i; + int bus = mp_irqs[idx].mpc_srcbus; + + /* + * Debugging check, we are in big trouble if this message pops up! + */ + if (mp_irqs[idx].mpc_dstirq != pin) + printk(KERN_ERR "broken BIOS or MPTABLE parser, ayiee!!\n"); + + switch (mp_bus_id_to_type[bus]) + { + case MP_BUS_ISA: /* ISA pin */ + case MP_BUS_EISA: + case MP_BUS_MCA: + case MP_BUS_NEC98: + { + irq = mp_irqs[idx].mpc_srcbusirq; + break; + } + case MP_BUS_PCI: /* PCI pin */ + { + /* + * PCI IRQs are mapped in order + */ + i = irq = 0; + while (i < apic) + irq += nr_ioapic_registers[i++]; + irq += pin; + + /* + * For MPS mode, so far only needed by ES7000 platform + */ + if (ioapic_renumber_irq) + irq = ioapic_renumber_irq(apic, irq); + + break; + } + default: + { + printk(KERN_ERR "unknown bus type %d.\n",bus); + irq = 0; + break; + } + } + + /* + * PCI IRQ command line redirection. Yes, limits are hardcoded. + */ + if ((pin >= 16) && (pin <= 23)) { + if (pirq_entries[pin-16] != -1) { + if (!pirq_entries[pin-16]) { + apic_printk(APIC_VERBOSE, KERN_DEBUG + "disabling PIRQ%d\n", pin-16); + } else { + irq = pirq_entries[pin-16]; + apic_printk(APIC_VERBOSE, KERN_DEBUG + "using PIRQ%d -> IRQ %d\n", + pin-16, irq); + } + } + } + return irq; +} + +static inline int IO_APIC_irq_trigger(int irq) +{ + int apic, idx, pin; + + for (apic = 0; apic < nr_ioapics; apic++) { + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) { + idx = find_irq_entry(apic,pin,mp_INT); + if ((idx != -1) && (irq == pin_2_irq(idx,apic,pin))) + return irq_trigger(idx); + } + } + /* + * nonexistent IRQs are edge default + */ + return 0; +} + +/* irq_vectors is indexed by the sum of all RTEs in all I/O APICs. */ +u8 irq_vector[NR_IRQ_VECTORS] = { FIRST_DEVICE_VECTOR , 0 }; + +int assign_irq_vector(int irq) +{ + static int current_vector = FIRST_DEVICE_VECTOR, offset = 0; + + BUG_ON(irq >= NR_IRQ_VECTORS); + if (irq != AUTO_ASSIGN && IO_APIC_VECTOR(irq) > 0) + return IO_APIC_VECTOR(irq); +next: + current_vector += 8; + if (current_vector == SYSCALL_VECTOR) + goto next; + + if (current_vector >= FIRST_SYSTEM_VECTOR) { + offset++; + if (!(offset%8)) + return -ENOSPC; + current_vector = FIRST_DEVICE_VECTOR + offset; + } + + vector_irq[current_vector] = irq; + if (irq != AUTO_ASSIGN) + IO_APIC_VECTOR(irq) = current_vector; + + return current_vector; +} + +static struct hw_interrupt_type ioapic_level_type; +static struct hw_interrupt_type ioapic_edge_type; + +#define IOAPIC_AUTO -1 +#define IOAPIC_EDGE 0 +#define IOAPIC_LEVEL 1 + +static inline void ioapic_register_intr(int irq, int vector, unsigned long trigger) +{ + if (use_pci_vector() && !platform_legacy_irq(irq)) { + if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || + trigger == IOAPIC_LEVEL) + irq_desc[vector].handler = &ioapic_level_type; + else + irq_desc[vector].handler = &ioapic_edge_type; + set_intr_gate(vector, interrupt[vector]); + } else { + if ((trigger == IOAPIC_AUTO && IO_APIC_irq_trigger(irq)) || + trigger == IOAPIC_LEVEL) + irq_desc[irq].handler = &ioapic_level_type; + else + irq_desc[irq].handler = &ioapic_edge_type; + set_intr_gate(vector, interrupt[irq]); + } +} + +static void __init setup_IO_APIC_irqs(void) +{ + struct IO_APIC_route_entry entry; + int apic, pin, idx, irq, first_notcon = 1, vector; + unsigned long flags; + + apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); + + for (apic = 0; apic < nr_ioapics; apic++) { + for (pin = 0; pin < nr_ioapic_registers[apic]; pin++) { + + /* + * add it to the IO-APIC irq-routing table: + */ + memset(&entry,0,sizeof(entry)); + + entry.delivery_mode = INT_DELIVERY_MODE; + entry.dest_mode = INT_DEST_MODE; + entry.mask = 0; /* enable IRQ */ + entry.dest.logical.logical_dest = + cpu_mask_to_apicid(TARGET_CPUS); + + idx = find_irq_entry(apic,pin,mp_INT); + if (idx == -1) { + if (first_notcon) { + apic_printk(APIC_VERBOSE, KERN_DEBUG + " IO-APIC (apicid-pin) %d-%d", + mp_ioapics[apic].mpc_apicid, + pin); + first_notcon = 0; + } else + apic_printk(APIC_VERBOSE, ", %d-%d", + mp_ioapics[apic].mpc_apicid, pin); + continue; + } + + entry.trigger = irq_trigger(idx); + entry.polarity = irq_polarity(idx); + + if (irq_trigger(idx)) { + entry.trigger = 1; + entry.mask = 1; + } + + irq = pin_2_irq(idx, apic, pin); + /* + * skip adding the timer int on secondary nodes, which causes + * a small but painful rift in the time-space continuum + */ + if (multi_timer_check(apic, irq)) + continue; + else + add_pin_to_irq(irq, apic, pin); + + if (!apic && !IO_APIC_IRQ(irq)) + continue; + + if (IO_APIC_IRQ(irq)) { + vector = assign_irq_vector(irq); + entry.vector = vector; + ioapic_register_intr(irq, vector, IOAPIC_AUTO); + + if (!apic && (irq < 16)) + disable_8259A_irq(irq); + } + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(apic, 0x11+2*pin, *(((int *)&entry)+1)); + io_apic_write(apic, 0x10+2*pin, *(((int *)&entry)+0)); + spin_unlock_irqrestore(&ioapic_lock, flags); + } + } + + if (!first_notcon) + apic_printk(APIC_VERBOSE, " not connected.\n"); +} + +/* + * Set up the 8259A-master output pin: + */ +static void __init setup_ExtINT_IRQ0_pin(unsigned int pin, int vector) +{ + struct IO_APIC_route_entry entry; + unsigned long flags; + + memset(&entry,0,sizeof(entry)); + + disable_8259A_irq(0); + + /* mask LVT0 */ + apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); + + /* + * We use logical delivery to get the timer IRQ + * to the first CPU. + */ + entry.dest_mode = INT_DEST_MODE; + entry.mask = 0; /* unmask IRQ now */ + entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS); + entry.delivery_mode = INT_DELIVERY_MODE; + entry.polarity = 0; + entry.trigger = 0; + entry.vector = vector; + + /* + * The timer IRQ doesn't have to know that behind the + * scene we have a 8259A-master in AEOI mode ... + */ + irq_desc[0].handler = &ioapic_edge_type; + + /* + * Add it to the IO-APIC irq-routing table: + */ + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(0, 0x11+2*pin, *(((int *)&entry)+1)); + io_apic_write(0, 0x10+2*pin, *(((int *)&entry)+0)); + spin_unlock_irqrestore(&ioapic_lock, flags); + + enable_8259A_irq(0); +} + +static inline void UNEXPECTED_IO_APIC(void) +{ +} + +void __init print_IO_APIC(void) +{ + int apic, i; + union IO_APIC_reg_00 reg_00; + union IO_APIC_reg_01 reg_01; + union IO_APIC_reg_02 reg_02; + union IO_APIC_reg_03 reg_03; + unsigned long flags; + + if (apic_verbosity == APIC_QUIET) + return; + + printk(KERN_DEBUG "number of MP IRQ sources: %d.\n", mp_irq_entries); + for (i = 0; i < nr_ioapics; i++) + printk(KERN_DEBUG "number of IO-APIC #%d registers: %d.\n", + mp_ioapics[i].mpc_apicid, nr_ioapic_registers[i]); + + /* + * We are a bit conservative about what we expect. We have to + * know about every hardware change ASAP. + */ + printk(KERN_INFO "testing the IO APIC.......................\n"); + + for (apic = 0; apic < nr_ioapics; apic++) { + + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(apic, 0); + reg_01.raw = io_apic_read(apic, 1); + if (reg_01.bits.version >= 0x10) + reg_02.raw = io_apic_read(apic, 2); + if (reg_01.bits.version >= 0x20) + reg_03.raw = io_apic_read(apic, 3); + spin_unlock_irqrestore(&ioapic_lock, flags); + + printk(KERN_DEBUG "IO APIC #%d......\n", mp_ioapics[apic].mpc_apicid); + printk(KERN_DEBUG ".... register #00: %08X\n", reg_00.raw); + printk(KERN_DEBUG "....... : physical APIC id: %02X\n", reg_00.bits.ID); + printk(KERN_DEBUG "....... : Delivery Type: %X\n", reg_00.bits.delivery_type); + printk(KERN_DEBUG "....... : LTS : %X\n", reg_00.bits.LTS); + if (reg_00.bits.ID >= get_physical_broadcast()) + UNEXPECTED_IO_APIC(); + if (reg_00.bits.__reserved_1 || reg_00.bits.__reserved_2) + UNEXPECTED_IO_APIC(); + + printk(KERN_DEBUG ".... register #01: %08X\n", reg_01.raw); + printk(KERN_DEBUG "....... : max redirection entries: %04X\n", reg_01.bits.entries); + if ( (reg_01.bits.entries != 0x0f) && /* older (Neptune) boards */ + (reg_01.bits.entries != 0x17) && /* typical ISA+PCI boards */ + (reg_01.bits.entries != 0x1b) && /* Compaq Proliant boards */ + (reg_01.bits.entries != 0x1f) && /* dual Xeon boards */ + (reg_01.bits.entries != 0x22) && /* bigger Xeon boards */ + (reg_01.bits.entries != 0x2E) && + (reg_01.bits.entries != 0x3F) + ) + UNEXPECTED_IO_APIC(); + + printk(KERN_DEBUG "....... : PRQ implemented: %X\n", reg_01.bits.PRQ); + printk(KERN_DEBUG "....... : IO APIC version: %04X\n", reg_01.bits.version); + if ( (reg_01.bits.version != 0x01) && /* 82489DX IO-APICs */ + (reg_01.bits.version != 0x10) && /* oldest IO-APICs */ + (reg_01.bits.version != 0x11) && /* Pentium/Pro IO-APICs */ + (reg_01.bits.version != 0x13) && /* Xeon IO-APICs */ + (reg_01.bits.version != 0x20) /* Intel P64H (82806 AA) */ + ) + UNEXPECTED_IO_APIC(); + if (reg_01.bits.__reserved_1 || reg_01.bits.__reserved_2) + UNEXPECTED_IO_APIC(); + + /* + * Some Intel chipsets with IO APIC VERSION of 0x1? don't have reg_02, + * but the value of reg_02 is read as the previous read register + * value, so ignore it if reg_02 == reg_01. + */ + if (reg_01.bits.version >= 0x10 && reg_02.raw != reg_01.raw) { + printk(KERN_DEBUG ".... register #02: %08X\n", reg_02.raw); + printk(KERN_DEBUG "....... : arbitration: %02X\n", reg_02.bits.arbitration); + if (reg_02.bits.__reserved_1 || reg_02.bits.__reserved_2) + UNEXPECTED_IO_APIC(); + } + + /* + * Some Intel chipsets with IO APIC VERSION of 0x2? don't have reg_02 + * or reg_03, but the value of reg_0[23] is read as the previous read + * register value, so ignore it if reg_03 == reg_0[12]. + */ + if (reg_01.bits.version >= 0x20 && reg_03.raw != reg_02.raw && + reg_03.raw != reg_01.raw) { + printk(KERN_DEBUG ".... register #03: %08X\n", reg_03.raw); + printk(KERN_DEBUG "....... : Boot DT : %X\n", reg_03.bits.boot_DT); + if (reg_03.bits.__reserved_1) + UNEXPECTED_IO_APIC(); + } + + printk(KERN_DEBUG ".... IRQ redirection table:\n"); + + printk(KERN_DEBUG " NR Log Phy Mask Trig IRR Pol" + " Stat Dest Deli Vect: \n"); + + for (i = 0; i <= reg_01.bits.entries; i++) { + struct IO_APIC_route_entry entry; + + spin_lock_irqsave(&ioapic_lock, flags); + *(((int *)&entry)+0) = io_apic_read(apic, 0x10+i*2); + *(((int *)&entry)+1) = io_apic_read(apic, 0x11+i*2); + spin_unlock_irqrestore(&ioapic_lock, flags); + + printk(KERN_DEBUG " %02x %03X %02X ", + i, + entry.dest.logical.logical_dest, + entry.dest.physical.physical_dest + ); + + printk("%1d %1d %1d %1d %1d %1d %1d %02X\n", + entry.mask, + entry.trigger, + entry.irr, + entry.polarity, + entry.delivery_status, + entry.dest_mode, + entry.delivery_mode, + entry.vector + ); + } + } + if (use_pci_vector()) + printk(KERN_INFO "Using vector-based indexing\n"); + printk(KERN_DEBUG "IRQ to pin mappings:\n"); + for (i = 0; i < NR_IRQS; i++) { + struct irq_pin_list *entry = irq_2_pin + i; + if (entry->pin < 0) + continue; + if (use_pci_vector() && !platform_legacy_irq(i)) + printk(KERN_DEBUG "IRQ%d ", IO_APIC_VECTOR(i)); + else + printk(KERN_DEBUG "IRQ%d ", i); + for (;;) { + printk("-> %d:%d", entry->apic, entry->pin); + if (!entry->next) + break; + entry = irq_2_pin + entry->next; + } + printk("\n"); + } + + printk(KERN_INFO ".................................... done.\n"); + + return; +} + +#if 0 + +static void print_APIC_bitfield (int base) +{ + unsigned int v; + int i, j; + + if (apic_verbosity == APIC_QUIET) + return; + + printk(KERN_DEBUG "0123456789abcdef0123456789abcdef\n" KERN_DEBUG); + for (i = 0; i < 8; i++) { + v = apic_read(base + i*0x10); + for (j = 0; j < 32; j++) { + if (v & (1<<j)) + printk("1"); + else + printk("0"); + } + printk("\n"); + } +} + +void /*__init*/ print_local_APIC(void * dummy) +{ + unsigned int v, ver, maxlvt; + + if (apic_verbosity == APIC_QUIET) + return; + + printk("\n" KERN_DEBUG "printing local APIC contents on CPU#%d/%d:\n", + smp_processor_id(), hard_smp_processor_id()); + v = apic_read(APIC_ID); + printk(KERN_INFO "... APIC ID: %08x (%01x)\n", v, GET_APIC_ID(v)); + v = apic_read(APIC_LVR); + printk(KERN_INFO "... APIC VERSION: %08x\n", v); + ver = GET_APIC_VERSION(v); + maxlvt = get_maxlvt(); + + v = apic_read(APIC_TASKPRI); + printk(KERN_DEBUG "... APIC TASKPRI: %08x (%02x)\n", v, v & APIC_TPRI_MASK); + + if (APIC_INTEGRATED(ver)) { /* !82489DX */ + v = apic_read(APIC_ARBPRI); + printk(KERN_DEBUG "... APIC ARBPRI: %08x (%02x)\n", v, + v & APIC_ARBPRI_MASK); + v = apic_read(APIC_PROCPRI); + printk(KERN_DEBUG "... APIC PROCPRI: %08x\n", v); + } + + v = apic_read(APIC_EOI); + printk(KERN_DEBUG "... APIC EOI: %08x\n", v); + v = apic_read(APIC_RRR); + printk(KERN_DEBUG "... APIC RRR: %08x\n", v); + v = apic_read(APIC_LDR); + printk(KERN_DEBUG "... APIC LDR: %08x\n", v); + v = apic_read(APIC_DFR); + printk(KERN_DEBUG "... APIC DFR: %08x\n", v); + v = apic_read(APIC_SPIV); + printk(KERN_DEBUG "... APIC SPIV: %08x\n", v); + + printk(KERN_DEBUG "... APIC ISR field:\n"); + print_APIC_bitfield(APIC_ISR); + printk(KERN_DEBUG "... APIC TMR field:\n"); + print_APIC_bitfield(APIC_TMR); + printk(KERN_DEBUG "... APIC IRR field:\n"); + print_APIC_bitfield(APIC_IRR); + + if (APIC_INTEGRATED(ver)) { /* !82489DX */ + if (maxlvt > 3) /* Due to the Pentium erratum 3AP. */ + apic_write(APIC_ESR, 0); + v = apic_read(APIC_ESR); + printk(KERN_DEBUG "... APIC ESR: %08x\n", v); + } + + v = apic_read(APIC_ICR); + printk(KERN_DEBUG "... APIC ICR: %08x\n", v); + v = apic_read(APIC_ICR2); + printk(KERN_DEBUG "... APIC ICR2: %08x\n", v); + + v = apic_read(APIC_LVTT); + printk(KERN_DEBUG "... APIC LVTT: %08x\n", v); + + if (maxlvt > 3) { /* PC is LVT#4. */ + v = apic_read(APIC_LVTPC); + printk(KERN_DEBUG "... APIC LVTPC: %08x\n", v); + } + v = apic_read(APIC_LVT0); + printk(KERN_DEBUG "... APIC LVT0: %08x\n", v); + v = apic_read(APIC_LVT1); + printk(KERN_DEBUG "... APIC LVT1: %08x\n", v); + + if (maxlvt > 2) { /* ERR is LVT#3. */ + v = apic_read(APIC_LVTERR); + printk(KERN_DEBUG "... APIC LVTERR: %08x\n", v); + } + + v = apic_read(APIC_TMICT); + printk(KERN_DEBUG "... APIC TMICT: %08x\n", v); + v = apic_read(APIC_TMCCT); + printk(KERN_DEBUG "... APIC TMCCT: %08x\n", v); + v = apic_read(APIC_TDCR); + printk(KERN_DEBUG "... APIC TDCR: %08x\n", v); + printk("\n"); +} + +void print_all_local_APICs (void) +{ + on_each_cpu(print_local_APIC, NULL, 1, 1); +} + +void /*__init*/ print_PIC(void) +{ + extern spinlock_t i8259A_lock; + unsigned int v; + unsigned long flags; + + if (apic_verbosity == APIC_QUIET) + return; + + printk(KERN_DEBUG "\nprinting PIC contents\n"); + + spin_lock_irqsave(&i8259A_lock, flags); + + v = inb(0xa1) << 8 | inb(0x21); + printk(KERN_DEBUG "... PIC IMR: %04x\n", v); + + v = inb(0xa0) << 8 | inb(0x20); + printk(KERN_DEBUG "... PIC IRR: %04x\n", v); + + outb(0x0b,0xa0); + outb(0x0b,0x20); + v = inb(0xa0) << 8 | inb(0x20); + outb(0x0a,0xa0); + outb(0x0a,0x20); + + spin_unlock_irqrestore(&i8259A_lock, flags); + + printk(KERN_DEBUG "... PIC ISR: %04x\n", v); + + v = inb(0x4d1) << 8 | inb(0x4d0); + printk(KERN_DEBUG "... PIC ELCR: %04x\n", v); +} + +#endif /* 0 */ + +static void __init enable_IO_APIC(void) +{ + union IO_APIC_reg_01 reg_01; + int i; + unsigned long flags; + + for (i = 0; i < PIN_MAP_SIZE; i++) { + irq_2_pin[i].pin = -1; + irq_2_pin[i].next = 0; + } + if (!pirqs_enabled) + for (i = 0; i < MAX_PIRQS; i++) + pirq_entries[i] = -1; + + /* + * The number of IO-APIC IRQ registers (== #pins): + */ + for (i = 0; i < nr_ioapics; i++) { + spin_lock_irqsave(&ioapic_lock, flags); + reg_01.raw = io_apic_read(i, 1); + spin_unlock_irqrestore(&ioapic_lock, flags); + nr_ioapic_registers[i] = reg_01.bits.entries+1; + } + + /* + * Do not trust the IO-APIC being empty at bootup + */ + clear_IO_APIC(); +} + +/* + * Not an __init, needed by the reboot code + */ +void disable_IO_APIC(void) +{ + /* + * Clear the IO-APIC before rebooting: + */ + clear_IO_APIC(); + + disconnect_bsp_APIC(); +} + +/* + * function to set the IO-APIC physical IDs based on the + * values stored in the MPC table. + * + * by Matt Domsch <Matt_Domsch@dell.com> Tue Dec 21 12:25:05 CST 1999 + */ + +#ifndef CONFIG_X86_NUMAQ +static void __init setup_ioapic_ids_from_mpc(void) +{ + union IO_APIC_reg_00 reg_00; + physid_mask_t phys_id_present_map; + int apic; + int i; + unsigned char old_id; + unsigned long flags; + + /* + * This is broken; anything with a real cpu count has to + * circumvent this idiocy regardless. + */ + phys_id_present_map = ioapic_phys_id_map(phys_cpu_present_map); + + /* + * Set the IOAPIC ID to the value stored in the MPC table. + */ + for (apic = 0; apic < nr_ioapics; apic++) { + + /* Read the register 0 value */ + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(apic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + old_id = mp_ioapics[apic].mpc_apicid; + + if (mp_ioapics[apic].mpc_apicid >= get_physical_broadcast()) { + printk(KERN_ERR "BIOS bug, IO-APIC#%d ID is %d in the MPC table!...\n", + apic, mp_ioapics[apic].mpc_apicid); + printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n", + reg_00.bits.ID); + mp_ioapics[apic].mpc_apicid = reg_00.bits.ID; + } + + /* Don't check I/O APIC IDs for some xAPIC systems. They have + * no meaning without the serial APIC bus. */ + if (NO_IOAPIC_CHECK) + continue; + /* + * Sanity check, is the ID really free? Every APIC in a + * system must have a unique ID or we get lots of nice + * 'stuck on smp_invalidate_needed IPI wait' messages. + */ + if (check_apicid_used(phys_id_present_map, + mp_ioapics[apic].mpc_apicid)) { + printk(KERN_ERR "BIOS bug, IO-APIC#%d ID %d is already used!...\n", + apic, mp_ioapics[apic].mpc_apicid); + for (i = 0; i < get_physical_broadcast(); i++) + if (!physid_isset(i, phys_id_present_map)) + break; + if (i >= get_physical_broadcast()) + panic("Max APIC ID exceeded!\n"); + printk(KERN_ERR "... fixing up to %d. (tell your hw vendor)\n", + i); + physid_set(i, phys_id_present_map); + mp_ioapics[apic].mpc_apicid = i; + } else { + physid_mask_t tmp; + tmp = apicid_to_cpu_present(mp_ioapics[apic].mpc_apicid); + apic_printk(APIC_VERBOSE, "Setting %d in the " + "phys_id_present_map\n", + mp_ioapics[apic].mpc_apicid); + physids_or(phys_id_present_map, phys_id_present_map, tmp); + } + + + /* + * We need to adjust the IRQ routing table + * if the ID changed. + */ + if (old_id != mp_ioapics[apic].mpc_apicid) + for (i = 0; i < mp_irq_entries; i++) + if (mp_irqs[i].mpc_dstapic == old_id) + mp_irqs[i].mpc_dstapic + = mp_ioapics[apic].mpc_apicid; + + /* + * Read the right value from the MPC table and + * write it into the ID register. + */ + apic_printk(APIC_VERBOSE, KERN_INFO + "...changing IO-APIC physical APIC ID to %d ...", + mp_ioapics[apic].mpc_apicid); + + reg_00.bits.ID = mp_ioapics[apic].mpc_apicid; + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(apic, 0, reg_00.raw); + spin_unlock_irqrestore(&ioapic_lock, flags); + + /* + * Sanity check + */ + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(apic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + if (reg_00.bits.ID != mp_ioapics[apic].mpc_apicid) + printk("could not set ID!\n"); + else + apic_printk(APIC_VERBOSE, " ok.\n"); + } +} +#else +static void __init setup_ioapic_ids_from_mpc(void) { } +#endif + +/* + * There is a nasty bug in some older SMP boards, their mptable lies + * about the timer IRQ. We do the following to work around the situation: + * + * - timer IRQ defaults to IO-APIC IRQ + * - if this function detects that timer IRQs are defunct, then we fall + * back to ISA timer IRQs + */ +static int __init timer_irq_works(void) +{ + unsigned long t1 = jiffies; + + local_irq_enable(); + /* Let ten ticks pass... */ + mdelay((10 * 1000) / HZ); + + /* + * Expect a few ticks at least, to be sure some possible + * glue logic does not lock up after one or two first + * ticks in a non-ExtINT mode. Also the local APIC + * might have cached one ExtINT interrupt. Finally, at + * least one tick may be lost due to delays. + */ + if (jiffies - t1 > 4) + return 1; + + return 0; +} + +/* + * In the SMP+IOAPIC case it might happen that there are an unspecified + * number of pending IRQ events unhandled. These cases are very rare, + * so we 'resend' these IRQs via IPIs, to the same CPU. It's much + * better to do it this way as thus we do not have to be aware of + * 'pending' interrupts in the IRQ path, except at this point. + */ +/* + * Edge triggered needs to resend any interrupt + * that was delayed but this is now handled in the device + * independent code. + */ + +/* + * Starting up a edge-triggered IO-APIC interrupt is + * nasty - we need to make sure that we get the edge. + * If it is already asserted for some reason, we need + * return 1 to indicate that is was pending. + * + * This is not complete - we should be able to fake + * an edge even if it isn't on the 8259A... + */ +static unsigned int startup_edge_ioapic_irq(unsigned int irq) +{ + int was_pending = 0; + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + if (irq < 16) { + disable_8259A_irq(irq); + if (i8259A_irq_pending(irq)) + was_pending = 1; + } + __unmask_IO_APIC_irq(irq); + spin_unlock_irqrestore(&ioapic_lock, flags); + + return was_pending; +} + +/* + * Once we have recorded IRQ_PENDING already, we can mask the + * interrupt for real. This prevents IRQ storms from unhandled + * devices. + */ +static void ack_edge_ioapic_irq(unsigned int irq) +{ + move_irq(irq); + if ((irq_desc[irq].status & (IRQ_PENDING | IRQ_DISABLED)) + == (IRQ_PENDING | IRQ_DISABLED)) + mask_IO_APIC_irq(irq); + ack_APIC_irq(); +} + +/* + * Level triggered interrupts can just be masked, + * and shutting down and starting up the interrupt + * is the same as enabling and disabling them -- except + * with a startup need to return a "was pending" value. + * + * Level triggered interrupts are special because we + * do not touch any IO-APIC register while handling + * them. We ack the APIC in the end-IRQ handler, not + * in the start-IRQ-handler. Protection against reentrance + * from the same interrupt is still provided, both by the + * generic IRQ layer and by the fact that an unacked local + * APIC does not accept IRQs. + */ +static unsigned int startup_level_ioapic_irq (unsigned int irq) +{ + unmask_IO_APIC_irq(irq); + + return 0; /* don't check for pending */ +} + +static void end_level_ioapic_irq (unsigned int irq) +{ + unsigned long v; + int i; + + move_irq(irq); +/* + * It appears there is an erratum which affects at least version 0x11 + * of I/O APIC (that's the 82093AA and cores integrated into various + * chipsets). Under certain conditions a level-triggered interrupt is + * erroneously delivered as edge-triggered one but the respective IRR + * bit gets set nevertheless. As a result the I/O unit expects an EOI + * message but it will never arrive and further interrupts are blocked + * from the source. The exact reason is so far unknown, but the + * phenomenon was observed when two consecutive interrupt requests + * from a given source get delivered to the same CPU and the source is + * temporarily disabled in between. + * + * A workaround is to simulate an EOI message manually. We achieve it + * by setting the trigger mode to edge and then to level when the edge + * trigger mode gets detected in the TMR of a local APIC for a + * level-triggered interrupt. We mask the source for the time of the + * operation to prevent an edge-triggered interrupt escaping meanwhile. + * The idea is from Manfred Spraul. --macro + */ + i = IO_APIC_VECTOR(irq); + + v = apic_read(APIC_TMR + ((i & ~0x1f) >> 1)); + + ack_APIC_irq(); + + if (!(v & (1 << (i & 0x1f)))) { + atomic_inc(&irq_mis_count); + spin_lock(&ioapic_lock); + __mask_and_edge_IO_APIC_irq(irq); + __unmask_and_level_IO_APIC_irq(irq); + spin_unlock(&ioapic_lock); + } +} + +#ifdef CONFIG_PCI_MSI +static unsigned int startup_edge_ioapic_vector(unsigned int vector) +{ + int irq = vector_to_irq(vector); + + return startup_edge_ioapic_irq(irq); +} + +static void ack_edge_ioapic_vector(unsigned int vector) +{ + int irq = vector_to_irq(vector); + + ack_edge_ioapic_irq(irq); +} + +static unsigned int startup_level_ioapic_vector (unsigned int vector) +{ + int irq = vector_to_irq(vector); + + return startup_level_ioapic_irq (irq); +} + +static void end_level_ioapic_vector (unsigned int vector) +{ + int irq = vector_to_irq(vector); + + end_level_ioapic_irq(irq); +} + +static void mask_IO_APIC_vector (unsigned int vector) +{ + int irq = vector_to_irq(vector); + + mask_IO_APIC_irq(irq); +} + +static void unmask_IO_APIC_vector (unsigned int vector) +{ + int irq = vector_to_irq(vector); + + unmask_IO_APIC_irq(irq); +} + +static void set_ioapic_affinity_vector (unsigned int vector, + cpumask_t cpu_mask) +{ + int irq = vector_to_irq(vector); + + set_ioapic_affinity_irq(irq, cpu_mask); +} +#endif + +/* + * Level and edge triggered IO-APIC interrupts need different handling, + * so we use two separate IRQ descriptors. Edge triggered IRQs can be + * handled with the level-triggered descriptor, but that one has slightly + * more overhead. Level-triggered interrupts cannot be handled with the + * edge-triggered handler, without risking IRQ storms and other ugly + * races. + */ +static struct hw_interrupt_type ioapic_edge_type = { + .typename = "IO-APIC-edge", + .startup = startup_edge_ioapic, + .shutdown = shutdown_edge_ioapic, + .enable = enable_edge_ioapic, + .disable = disable_edge_ioapic, + .ack = ack_edge_ioapic, + .end = end_edge_ioapic, + .set_affinity = set_ioapic_affinity, +}; + +static struct hw_interrupt_type ioapic_level_type = { + .typename = "IO-APIC-level", + .startup = startup_level_ioapic, + .shutdown = shutdown_level_ioapic, + .enable = enable_level_ioapic, + .disable = disable_level_ioapic, + .ack = mask_and_ack_level_ioapic, + .end = end_level_ioapic, + .set_affinity = set_ioapic_affinity, +}; + +static inline void init_IO_APIC_traps(void) +{ + int irq; + + /* + * NOTE! The local APIC isn't very good at handling + * multiple interrupts at the same interrupt level. + * As the interrupt level is determined by taking the + * vector number and shifting that right by 4, we + * want to spread these out a bit so that they don't + * all fall in the same interrupt level. + * + * Also, we've got to be careful not to trash gate + * 0x80, because int 0x80 is hm, kind of importantish. ;) + */ + for (irq = 0; irq < NR_IRQS ; irq++) { + int tmp = irq; + if (use_pci_vector()) { + if (!platform_legacy_irq(tmp)) + if ((tmp = vector_to_irq(tmp)) == -1) + continue; + } + if (IO_APIC_IRQ(tmp) && !IO_APIC_VECTOR(tmp)) { + /* + * Hmm.. We don't have an entry for this, + * so default to an old-fashioned 8259 + * interrupt if we can.. + */ + if (irq < 16) + make_8259A_irq(irq); + else + /* Strange. Oh, well.. */ + irq_desc[irq].handler = &no_irq_type; + } + } +} + +static void enable_lapic_irq (unsigned int irq) +{ + unsigned long v; + + v = apic_read(APIC_LVT0); + apic_write_around(APIC_LVT0, v & ~APIC_LVT_MASKED); +} + +static void disable_lapic_irq (unsigned int irq) +{ + unsigned long v; + + v = apic_read(APIC_LVT0); + apic_write_around(APIC_LVT0, v | APIC_LVT_MASKED); +} + +static void ack_lapic_irq (unsigned int irq) +{ + ack_APIC_irq(); +} + +static void end_lapic_irq (unsigned int i) { /* nothing */ } + +static struct hw_interrupt_type lapic_irq_type = { + .typename = "local-APIC-edge", + .startup = NULL, /* startup_irq() not used for IRQ0 */ + .shutdown = NULL, /* shutdown_irq() not used for IRQ0 */ + .enable = enable_lapic_irq, + .disable = disable_lapic_irq, + .ack = ack_lapic_irq, + .end = end_lapic_irq +}; + +static void setup_nmi (void) +{ + /* + * Dirty trick to enable the NMI watchdog ... + * We put the 8259A master into AEOI mode and + * unmask on all local APICs LVT0 as NMI. + * + * The idea to use the 8259A in AEOI mode ('8259A Virtual Wire') + * is from Maciej W. Rozycki - so we do not have to EOI from + * the NMI handler or the timer interrupt. + */ + apic_printk(APIC_VERBOSE, KERN_INFO "activating NMI Watchdog ..."); + + on_each_cpu(enable_NMI_through_LVT0, NULL, 1, 1); + + apic_printk(APIC_VERBOSE, " done.\n"); +} + +/* + * This looks a bit hackish but it's about the only one way of sending + * a few INTA cycles to 8259As and any associated glue logic. ICR does + * not support the ExtINT mode, unfortunately. We need to send these + * cycles as some i82489DX-based boards have glue logic that keeps the + * 8259A interrupt line asserted until INTA. --macro + */ +static inline void unlock_ExtINT_logic(void) +{ + int pin, i; + struct IO_APIC_route_entry entry0, entry1; + unsigned char save_control, save_freq_select; + unsigned long flags; + + pin = find_isa_irq_pin(8, mp_INT); + if (pin == -1) + return; + + spin_lock_irqsave(&ioapic_lock, flags); + *(((int *)&entry0) + 1) = io_apic_read(0, 0x11 + 2 * pin); + *(((int *)&entry0) + 0) = io_apic_read(0, 0x10 + 2 * pin); + spin_unlock_irqrestore(&ioapic_lock, flags); + clear_IO_APIC_pin(0, pin); + + memset(&entry1, 0, sizeof(entry1)); + + entry1.dest_mode = 0; /* physical delivery */ + entry1.mask = 0; /* unmask IRQ now */ + entry1.dest.physical.physical_dest = hard_smp_processor_id(); + entry1.delivery_mode = dest_ExtINT; + entry1.polarity = entry0.polarity; + entry1.trigger = 0; + entry1.vector = 0; + + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(0, 0x11 + 2 * pin, *(((int *)&entry1) + 1)); + io_apic_write(0, 0x10 + 2 * pin, *(((int *)&entry1) + 0)); + spin_unlock_irqrestore(&ioapic_lock, flags); + + save_control = CMOS_READ(RTC_CONTROL); + save_freq_select = CMOS_READ(RTC_FREQ_SELECT); + CMOS_WRITE((save_freq_select & ~RTC_RATE_SELECT) | 0x6, + RTC_FREQ_SELECT); + CMOS_WRITE(save_control | RTC_PIE, RTC_CONTROL); + + i = 100; + while (i-- > 0) { + mdelay(10); + if ((CMOS_READ(RTC_INTR_FLAGS) & RTC_PF) == RTC_PF) + i -= 10; + } + + CMOS_WRITE(save_control, RTC_CONTROL); + CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT); + clear_IO_APIC_pin(0, pin); + + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(0, 0x11 + 2 * pin, *(((int *)&entry0) + 1)); + io_apic_write(0, 0x10 + 2 * pin, *(((int *)&entry0) + 0)); + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +/* + * This code may look a bit paranoid, but it's supposed to cooperate with + * a wide range of boards and BIOS bugs. Fortunately only the timer IRQ + * is so screwy. Thanks to Brian Perkins for testing/hacking this beast + * fanatically on his truly buggy board. + */ +static inline void check_timer(void) +{ + int pin1, pin2; + int vector; + + /* + * get/set the timer IRQ vector: + */ + disable_8259A_irq(0); + vector = assign_irq_vector(0); + set_intr_gate(vector, interrupt[0]); + + /* + * Subtle, code in do_timer_interrupt() expects an AEOI + * mode for the 8259A whenever interrupts are routed + * through I/O APICs. Also IRQ0 has to be enabled in + * the 8259A which implies the virtual wire has to be + * disabled in the local APIC. + */ + apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_EXTINT); + init_8259A(1); + timer_ack = 1; + enable_8259A_irq(0); + + pin1 = find_isa_irq_pin(0, mp_INT); + pin2 = find_isa_irq_pin(0, mp_ExtINT); + + printk(KERN_INFO "..TIMER: vector=0x%02X pin1=%d pin2=%d\n", vector, pin1, pin2); + + if (pin1 != -1) { + /* + * Ok, does IRQ0 through the IOAPIC work? + */ + unmask_IO_APIC_irq(0); + if (timer_irq_works()) { + if (nmi_watchdog == NMI_IO_APIC) { + disable_8259A_irq(0); + setup_nmi(); + enable_8259A_irq(0); + check_nmi_watchdog(); + } + return; + } + clear_IO_APIC_pin(0, pin1); + printk(KERN_ERR "..MP-BIOS bug: 8254 timer not connected to IO-APIC\n"); + } + + printk(KERN_INFO "...trying to set up timer (IRQ0) through the 8259A ... "); + if (pin2 != -1) { + printk("\n..... (found pin %d) ...", pin2); + /* + * legacy devices should be connected to IO APIC #0 + */ + setup_ExtINT_IRQ0_pin(pin2, vector); + if (timer_irq_works()) { + printk("works.\n"); + if (pin1 != -1) + replace_pin_at_irq(0, 0, pin1, 0, pin2); + else + add_pin_to_irq(0, 0, pin2); + if (nmi_watchdog == NMI_IO_APIC) { + setup_nmi(); + check_nmi_watchdog(); + } + return; + } + /* + * Cleanup, just in case ... + */ + clear_IO_APIC_pin(0, pin2); + } + printk(" failed.\n"); + + if (nmi_watchdog == NMI_IO_APIC) { + printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n"); + nmi_watchdog = 0; + } + + printk(KERN_INFO "...trying to set up timer as Virtual Wire IRQ..."); + + disable_8259A_irq(0); + irq_desc[0].handler = &lapic_irq_type; + apic_write_around(APIC_LVT0, APIC_DM_FIXED | vector); /* Fixed mode */ + enable_8259A_irq(0); + + if (timer_irq_works()) { + printk(" works.\n"); + return; + } + apic_write_around(APIC_LVT0, APIC_LVT_MASKED | APIC_DM_FIXED | vector); + printk(" failed.\n"); + + printk(KERN_INFO "...trying to set up timer as ExtINT IRQ..."); + + timer_ack = 0; + init_8259A(0); + make_8259A_irq(0); + apic_write_around(APIC_LVT0, APIC_DM_EXTINT); + + unlock_ExtINT_logic(); + + if (timer_irq_works()) { + printk(" works.\n"); + return; + } + printk(" failed :(.\n"); + panic("IO-APIC + timer doesn't work! Boot with apic=debug and send a " + "report. Then try booting with the 'noapic' option"); +} + +/* + * + * IRQ's that are handled by the PIC in the MPS IOAPIC case. + * - IRQ2 is the cascade IRQ, and cannot be a io-apic IRQ. + * Linux doesn't really care, as it's not actually used + * for any interrupt handling anyway. + */ +#define PIC_IRQS (1 << PIC_CASCADE_IR) + +void __init setup_IO_APIC(void) +{ + enable_IO_APIC(); + + if (acpi_ioapic) + io_apic_irqs = ~0; /* all IRQs go through IOAPIC */ + else + io_apic_irqs = ~PIC_IRQS; + + printk("ENABLING IO-APIC IRQs\n"); + + /* + * Set up IO-APIC IRQ routing. + */ + if (!acpi_ioapic) + setup_ioapic_ids_from_mpc(); + sync_Arb_IDs(); + setup_IO_APIC_irqs(); + init_IO_APIC_traps(); + check_timer(); + if (!acpi_ioapic) + print_IO_APIC(); +} + +/* + * Called after all the initialization is done. If we didnt find any + * APIC bugs then we can allow the modify fast path + */ + +static int __init io_apic_bug_finalize(void) +{ + if(sis_apic_bug == -1) + sis_apic_bug = 0; + return 0; +} + +late_initcall(io_apic_bug_finalize); + +struct sysfs_ioapic_data { + struct sys_device dev; + struct IO_APIC_route_entry entry[0]; +}; +static struct sysfs_ioapic_data * mp_ioapic_data[MAX_IO_APICS]; + +static int ioapic_suspend(struct sys_device *dev, u32 state) +{ + struct IO_APIC_route_entry *entry; + struct sysfs_ioapic_data *data; + unsigned long flags; + int i; + + data = container_of(dev, struct sysfs_ioapic_data, dev); + entry = data->entry; + spin_lock_irqsave(&ioapic_lock, flags); + for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) { + *(((int *)entry) + 1) = io_apic_read(dev->id, 0x11 + 2 * i); + *(((int *)entry) + 0) = io_apic_read(dev->id, 0x10 + 2 * i); + } + spin_unlock_irqrestore(&ioapic_lock, flags); + + return 0; +} + +static int ioapic_resume(struct sys_device *dev) +{ + struct IO_APIC_route_entry *entry; + struct sysfs_ioapic_data *data; + unsigned long flags; + union IO_APIC_reg_00 reg_00; + int i; + + data = container_of(dev, struct sysfs_ioapic_data, dev); + entry = data->entry; + + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(dev->id, 0); + if (reg_00.bits.ID != mp_ioapics[dev->id].mpc_apicid) { + reg_00.bits.ID = mp_ioapics[dev->id].mpc_apicid; + io_apic_write(dev->id, 0, reg_00.raw); + } + for (i = 0; i < nr_ioapic_registers[dev->id]; i ++, entry ++ ) { + io_apic_write(dev->id, 0x11+2*i, *(((int *)entry)+1)); + io_apic_write(dev->id, 0x10+2*i, *(((int *)entry)+0)); + } + spin_unlock_irqrestore(&ioapic_lock, flags); + + return 0; +} + +static struct sysdev_class ioapic_sysdev_class = { + set_kset_name("ioapic"), + .suspend = ioapic_suspend, + .resume = ioapic_resume, +}; + +static int __init ioapic_init_sysfs(void) +{ + struct sys_device * dev; + int i, size, error = 0; + + error = sysdev_class_register(&ioapic_sysdev_class); + if (error) + return error; + + for (i = 0; i < nr_ioapics; i++ ) { + size = sizeof(struct sys_device) + nr_ioapic_registers[i] + * sizeof(struct IO_APIC_route_entry); + mp_ioapic_data[i] = kmalloc(size, GFP_KERNEL); + if (!mp_ioapic_data[i]) { + printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i); + continue; + } + memset(mp_ioapic_data[i], 0, size); + dev = &mp_ioapic_data[i]->dev; + dev->id = i; + dev->cls = &ioapic_sysdev_class; + error = sysdev_register(dev); + if (error) { + kfree(mp_ioapic_data[i]); + mp_ioapic_data[i] = NULL; + printk(KERN_ERR "Can't suspend/resume IOAPIC %d\n", i); + continue; + } + } + + return 0; +} + +device_initcall(ioapic_init_sysfs); + +/* -------------------------------------------------------------------------- + ACPI-based IOAPIC Configuration + -------------------------------------------------------------------------- */ + +#ifdef CONFIG_ACPI_BOOT + +int __init io_apic_get_unique_id (int ioapic, int apic_id) +{ + union IO_APIC_reg_00 reg_00; + static physid_mask_t apic_id_map = PHYSID_MASK_NONE; + physid_mask_t tmp; + unsigned long flags; + int i = 0; + + /* + * The P4 platform supports up to 256 APIC IDs on two separate APIC + * buses (one for LAPICs, one for IOAPICs), where predecessors only + * supports up to 16 on one shared APIC bus. + * + * TBD: Expand LAPIC/IOAPIC support on P4-class systems to take full + * advantage of new APIC bus architecture. + */ + + if (physids_empty(apic_id_map)) + apic_id_map = ioapic_phys_id_map(phys_cpu_present_map); + + spin_lock_irqsave(&ioapic_lock, flags); + reg_00.raw = io_apic_read(ioapic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + if (apic_id >= get_physical_broadcast()) { + printk(KERN_WARNING "IOAPIC[%d]: Invalid apic_id %d, trying " + "%d\n", ioapic, apic_id, reg_00.bits.ID); + apic_id = reg_00.bits.ID; + } + + /* + * Every APIC in a system must have a unique ID or we get lots of nice + * 'stuck on smp_invalidate_needed IPI wait' messages. + */ + if (check_apicid_used(apic_id_map, apic_id)) { + + for (i = 0; i < get_physical_broadcast(); i++) { + if (!check_apicid_used(apic_id_map, i)) + break; + } + + if (i == get_physical_broadcast()) + panic("Max apic_id exceeded!\n"); + + printk(KERN_WARNING "IOAPIC[%d]: apic_id %d already used, " + "trying %d\n", ioapic, apic_id, i); + + apic_id = i; + } + + tmp = apicid_to_cpu_present(apic_id); + physids_or(apic_id_map, apic_id_map, tmp); + + if (reg_00.bits.ID != apic_id) { + reg_00.bits.ID = apic_id; + + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(ioapic, 0, reg_00.raw); + reg_00.raw = io_apic_read(ioapic, 0); + spin_unlock_irqrestore(&ioapic_lock, flags); + + /* Sanity check */ + if (reg_00.bits.ID != apic_id) + panic("IOAPIC[%d]: Unable change apic_id!\n", ioapic); + } + + apic_printk(APIC_VERBOSE, KERN_INFO + "IOAPIC[%d]: Assigned apic_id %d\n", ioapic, apic_id); + + return apic_id; +} + + +int __init io_apic_get_version (int ioapic) +{ + union IO_APIC_reg_01 reg_01; + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + reg_01.raw = io_apic_read(ioapic, 1); + spin_unlock_irqrestore(&ioapic_lock, flags); + + return reg_01.bits.version; +} + + +int __init io_apic_get_redir_entries (int ioapic) +{ + union IO_APIC_reg_01 reg_01; + unsigned long flags; + + spin_lock_irqsave(&ioapic_lock, flags); + reg_01.raw = io_apic_read(ioapic, 1); + spin_unlock_irqrestore(&ioapic_lock, flags); + + return reg_01.bits.entries; +} + + +int io_apic_set_pci_routing (int ioapic, int pin, int irq, int edge_level, int active_high_low) +{ + struct IO_APIC_route_entry entry; + unsigned long flags; + + if (!IO_APIC_IRQ(irq)) { + printk(KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", + ioapic); + return -EINVAL; + } + + /* + * Generate a PCI IRQ routing entry and program the IOAPIC accordingly. + * Note that we mask (disable) IRQs now -- these get enabled when the + * corresponding device driver registers for this IRQ. + */ + + memset(&entry,0,sizeof(entry)); + + entry.delivery_mode = INT_DELIVERY_MODE; + entry.dest_mode = INT_DEST_MODE; + entry.dest.logical.logical_dest = cpu_mask_to_apicid(TARGET_CPUS); + entry.trigger = edge_level; + entry.polarity = active_high_low; + entry.mask = 1; + + /* + * IRQs < 16 are already in the irq_2_pin[] map + */ + if (irq >= 16) + add_pin_to_irq(irq, ioapic, pin); + + entry.vector = assign_irq_vector(irq); + + apic_printk(APIC_DEBUG, KERN_DEBUG "IOAPIC[%d]: Set PCI routing entry " + "(%d-%d -> 0x%x -> IRQ %d Mode:%i Active:%i)\n", ioapic, + mp_ioapics[ioapic].mpc_apicid, pin, entry.vector, irq, + edge_level, active_high_low); + + ioapic_register_intr(irq, entry.vector, edge_level); + + if (!ioapic && (irq < 16)) + disable_8259A_irq(irq); + + spin_lock_irqsave(&ioapic_lock, flags); + io_apic_write(ioapic, 0x11+2*pin, *(((int *)&entry)+1)); + io_apic_write(ioapic, 0x10+2*pin, *(((int *)&entry)+0)); + spin_unlock_irqrestore(&ioapic_lock, flags); + + return 0; +} + +#endif /*CONFIG_ACPI_BOOT*/ diff --git a/arch/i386/kernel/ioport.c b/arch/i386/kernel/ioport.c new file mode 100644 index 00000000000..8b25160393c --- /dev/null +++ b/arch/i386/kernel/ioport.c @@ -0,0 +1,147 @@ +/* + * linux/arch/i386/kernel/ioport.c + * + * This contains the io-permission bitmap code - written by obz, with changes + * by Linus. + */ + +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/errno.h> +#include <linux/types.h> +#include <linux/ioport.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/stddef.h> +#include <linux/slab.h> +#include <linux/thread_info.h> + +/* Set EXTENT bits starting at BASE in BITMAP to value TURN_ON. */ +static void set_bitmap(unsigned long *bitmap, unsigned int base, unsigned int extent, int new_value) +{ + unsigned long mask; + unsigned long *bitmap_base = bitmap + (base / BITS_PER_LONG); + unsigned int low_index = base & (BITS_PER_LONG-1); + int length = low_index + extent; + + if (low_index != 0) { + mask = (~0UL << low_index); + if (length < BITS_PER_LONG) + mask &= ~(~0UL << length); + if (new_value) + *bitmap_base++ |= mask; + else + *bitmap_base++ &= ~mask; + length -= BITS_PER_LONG; + } + + mask = (new_value ? ~0UL : 0UL); + while (length >= BITS_PER_LONG) { + *bitmap_base++ = mask; + length -= BITS_PER_LONG; + } + + if (length > 0) { + mask = ~(~0UL << length); + if (new_value) + *bitmap_base++ |= mask; + else + *bitmap_base++ &= ~mask; + } +} + + +/* + * this changes the io permissions bitmap in the current task. + */ +asmlinkage long sys_ioperm(unsigned long from, unsigned long num, int turn_on) +{ + unsigned long i, max_long, bytes, bytes_updated; + struct thread_struct * t = ¤t->thread; + struct tss_struct * tss; + unsigned long *bitmap; + + if ((from + num <= from) || (from + num > IO_BITMAP_BITS)) + return -EINVAL; + if (turn_on && !capable(CAP_SYS_RAWIO)) + return -EPERM; + + /* + * If it's the first ioperm() call in this thread's lifetime, set the + * IO bitmap up. ioperm() is much less timing critical than clone(), + * this is why we delay this operation until now: + */ + if (!t->io_bitmap_ptr) { + bitmap = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL); + if (!bitmap) + return -ENOMEM; + + memset(bitmap, 0xff, IO_BITMAP_BYTES); + t->io_bitmap_ptr = bitmap; + } + + /* + * do it in the per-thread copy and in the TSS ... + * + * Disable preemption via get_cpu() - we must not switch away + * because the ->io_bitmap_max value must match the bitmap + * contents: + */ + tss = &per_cpu(init_tss, get_cpu()); + + set_bitmap(t->io_bitmap_ptr, from, num, !turn_on); + + /* + * Search for a (possibly new) maximum. This is simple and stupid, + * to keep it obviously correct: + */ + max_long = 0; + for (i = 0; i < IO_BITMAP_LONGS; i++) + if (t->io_bitmap_ptr[i] != ~0UL) + max_long = i; + + bytes = (max_long + 1) * sizeof(long); + bytes_updated = max(bytes, t->io_bitmap_max); + + t->io_bitmap_max = bytes; + + /* + * Sets the lazy trigger so that the next I/O operation will + * reload the correct bitmap. + */ + tss->io_bitmap_base = INVALID_IO_BITMAP_OFFSET_LAZY; + + put_cpu(); + + return 0; +} + +/* + * sys_iopl has to be used when you want to access the IO ports + * beyond the 0x3ff range: to get the full 65536 ports bitmapped + * you'd need 8kB of bitmaps/process, which is a bit excessive. + * + * Here we just change the eflags value on the stack: we allow + * only the super-user to do it. This depends on the stack-layout + * on system-call entry - see also fork() and the signal handling + * code. + */ + +asmlinkage long sys_iopl(unsigned long unused) +{ + volatile struct pt_regs * regs = (struct pt_regs *) &unused; + unsigned int level = regs->ebx; + unsigned int old = (regs->eflags >> 12) & 3; + + if (level > 3) + return -EINVAL; + /* Trying to gain more privileges? */ + if (level > old) { + if (!capable(CAP_SYS_RAWIO)) + return -EPERM; + } + regs->eflags = (regs->eflags &~ 0x3000UL) | (level << 12); + /* Make sure we return the long way (not sysenter) */ + set_thread_flag(TIF_IRET); + return 0; +} diff --git a/arch/i386/kernel/irq.c b/arch/i386/kernel/irq.c new file mode 100644 index 00000000000..73945a3c53c --- /dev/null +++ b/arch/i386/kernel/irq.c @@ -0,0 +1,261 @@ +/* + * linux/arch/i386/kernel/irq.c + * + * Copyright (C) 1992, 1998 Linus Torvalds, Ingo Molnar + * + * This file contains the lowest level x86-specific interrupt + * entry, irq-stacks and irq statistics code. All the remaining + * irq logic is done by the generic kernel/irq/ code and + * by the x86-specific irq controller code. (e.g. i8259.c and + * io_apic.c.) + */ + +#include <asm/uaccess.h> +#include <linux/module.h> +#include <linux/seq_file.h> +#include <linux/interrupt.h> +#include <linux/kernel_stat.h> + +DEFINE_PER_CPU(irq_cpustat_t, irq_stat) ____cacheline_maxaligned_in_smp; +EXPORT_PER_CPU_SYMBOL(irq_stat); + +#ifndef CONFIG_X86_LOCAL_APIC +/* + * 'what should we do if we get a hw irq event on an illegal vector'. + * each architecture has to answer this themselves. + */ +void ack_bad_irq(unsigned int irq) +{ + printk("unexpected IRQ trap at vector %02x\n", irq); +} +#endif + +#ifdef CONFIG_4KSTACKS +/* + * per-CPU IRQ handling contexts (thread information and stack) + */ +union irq_ctx { + struct thread_info tinfo; + u32 stack[THREAD_SIZE/sizeof(u32)]; +}; + +static union irq_ctx *hardirq_ctx[NR_CPUS]; +static union irq_ctx *softirq_ctx[NR_CPUS]; +#endif + +/* + * do_IRQ handles all normal device IRQ's (the special + * SMP cross-CPU interrupts have their own specific + * handlers). + */ +fastcall unsigned int do_IRQ(struct pt_regs *regs) +{ + /* high bits used in ret_from_ code */ + int irq = regs->orig_eax & 0xff; +#ifdef CONFIG_4KSTACKS + union irq_ctx *curctx, *irqctx; + u32 *isp; +#endif + + irq_enter(); +#ifdef CONFIG_DEBUG_STACKOVERFLOW + /* Debugging check for stack overflow: is there less than 1KB free? */ + { + long esp; + + __asm__ __volatile__("andl %%esp,%0" : + "=r" (esp) : "0" (THREAD_SIZE - 1)); + if (unlikely(esp < (sizeof(struct thread_info) + STACK_WARN))) { + printk("do_IRQ: stack overflow: %ld\n", + esp - sizeof(struct thread_info)); + dump_stack(); + } + } +#endif + +#ifdef CONFIG_4KSTACKS + + curctx = (union irq_ctx *) current_thread_info(); + irqctx = hardirq_ctx[smp_processor_id()]; + + /* + * this is where we switch to the IRQ stack. However, if we are + * already using the IRQ stack (because we interrupted a hardirq + * handler) we can't do that and just have to keep using the + * current stack (which is the irq stack already after all) + */ + if (curctx != irqctx) { + int arg1, arg2, ebx; + + /* build the stack frame on the IRQ stack */ + isp = (u32*) ((char*)irqctx + sizeof(*irqctx)); + irqctx->tinfo.task = curctx->tinfo.task; + irqctx->tinfo.previous_esp = current_stack_pointer; + + asm volatile( + " xchgl %%ebx,%%esp \n" + " call __do_IRQ \n" + " movl %%ebx,%%esp \n" + : "=a" (arg1), "=d" (arg2), "=b" (ebx) + : "0" (irq), "1" (regs), "2" (isp) + : "memory", "cc", "ecx" + ); + } else +#endif + __do_IRQ(irq, regs); + + irq_exit(); + + return 1; +} + +#ifdef CONFIG_4KSTACKS + +/* + * These should really be __section__(".bss.page_aligned") as well, but + * gcc's 3.0 and earlier don't handle that correctly. + */ +static char softirq_stack[NR_CPUS * THREAD_SIZE] + __attribute__((__aligned__(THREAD_SIZE))); + +static char hardirq_stack[NR_CPUS * THREAD_SIZE] + __attribute__((__aligned__(THREAD_SIZE))); + +/* + * allocate per-cpu stacks for hardirq and for softirq processing + */ +void irq_ctx_init(int cpu) +{ + union irq_ctx *irqctx; + + if (hardirq_ctx[cpu]) + return; + + irqctx = (union irq_ctx*) &hardirq_stack[cpu*THREAD_SIZE]; + irqctx->tinfo.task = NULL; + irqctx->tinfo.exec_domain = NULL; + irqctx->tinfo.cpu = cpu; + irqctx->tinfo.preempt_count = HARDIRQ_OFFSET; + irqctx->tinfo.addr_limit = MAKE_MM_SEG(0); + + hardirq_ctx[cpu] = irqctx; + + irqctx = (union irq_ctx*) &softirq_stack[cpu*THREAD_SIZE]; + irqctx->tinfo.task = NULL; + irqctx->tinfo.exec_domain = NULL; + irqctx->tinfo.cpu = cpu; + irqctx->tinfo.preempt_count = SOFTIRQ_OFFSET; + irqctx->tinfo.addr_limit = MAKE_MM_SEG(0); + + softirq_ctx[cpu] = irqctx; + + printk("CPU %u irqstacks, hard=%p soft=%p\n", + cpu,hardirq_ctx[cpu],softirq_ctx[cpu]); +} + +extern asmlinkage void __do_softirq(void); + +asmlinkage void do_softirq(void) +{ + unsigned long flags; + struct thread_info *curctx; + union irq_ctx *irqctx; + u32 *isp; + + if (in_interrupt()) + return; + + local_irq_save(flags); + + if (local_softirq_pending()) { + curctx = current_thread_info(); + irqctx = softirq_ctx[smp_processor_id()]; + irqctx->tinfo.task = curctx->task; + irqctx->tinfo.previous_esp = current_stack_pointer; + + /* build the stack frame on the softirq stack */ + isp = (u32*) ((char*)irqctx + sizeof(*irqctx)); + + asm volatile( + " xchgl %%ebx,%%esp \n" + " call __do_softirq \n" + " movl %%ebx,%%esp \n" + : "=b"(isp) + : "0"(isp) + : "memory", "cc", "edx", "ecx", "eax" + ); + } + + local_irq_restore(flags); +} + +EXPORT_SYMBOL(do_softirq); +#endif + +/* + * Interrupt statistics: + */ + +atomic_t irq_err_count; + +/* + * /proc/interrupts printing: + */ + +int show_interrupts(struct seq_file *p, void *v) +{ + int i = *(loff_t *) v, j; + struct irqaction * action; + unsigned long flags; + + if (i == 0) { + seq_printf(p, " "); + for (j=0; j<NR_CPUS; j++) + if (cpu_online(j)) + seq_printf(p, "CPU%d ",j); + seq_putc(p, '\n'); + } + + if (i < NR_IRQS) { + spin_lock_irqsave(&irq_desc[i].lock, flags); + action = irq_desc[i].action; + if (!action) + goto skip; + seq_printf(p, "%3d: ",i); +#ifndef CONFIG_SMP + seq_printf(p, "%10u ", kstat_irqs(i)); +#else + for (j = 0; j < NR_CPUS; j++) + if (cpu_online(j)) + seq_printf(p, "%10u ", kstat_cpu(j).irqs[i]); +#endif + seq_printf(p, " %14s", irq_desc[i].handler->typename); + seq_printf(p, " %s", action->name); + + for (action=action->next; action; action = action->next) + seq_printf(p, ", %s", action->name); + + seq_putc(p, '\n'); +skip: + spin_unlock_irqrestore(&irq_desc[i].lock, flags); + } else if (i == NR_IRQS) { + seq_printf(p, "NMI: "); + for (j = 0; j < NR_CPUS; j++) + if (cpu_online(j)) + seq_printf(p, "%10u ", nmi_count(j)); + seq_putc(p, '\n'); +#ifdef CONFIG_X86_LOCAL_APIC + seq_printf(p, "LOC: "); + for (j = 0; j < NR_CPUS; j++) + if (cpu_online(j)) + seq_printf(p, "%10u ", + per_cpu(irq_stat,j).apic_timer_irqs); + seq_putc(p, '\n'); +#endif + seq_printf(p, "ERR: %10u\n", atomic_read(&irq_err_count)); +#if defined(CONFIG_X86_IO_APIC) + seq_printf(p, "MIS: %10u\n", atomic_read(&irq_mis_count)); +#endif + } + return 0; +} diff --git a/arch/i386/kernel/kprobes.c b/arch/i386/kernel/kprobes.c new file mode 100644 index 00000000000..67168165924 --- /dev/null +++ b/arch/i386/kernel/kprobes.c @@ -0,0 +1,385 @@ +/* + * Kernel Probes (KProbes) + * arch/i386/kernel/kprobes.c + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. + * + * Copyright (C) IBM Corporation, 2002, 2004 + * + * 2002-Oct Created by Vamsi Krishna S <vamsi_krishna@in.ibm.com> Kernel + * Probes initial implementation ( includes contributions from + * Rusty Russell). + * 2004-July Suparna Bhattacharya <suparna@in.ibm.com> added jumper probes + * interface to access function arguments. + */ + +#include <linux/config.h> +#include <linux/kprobes.h> +#include <linux/ptrace.h> +#include <linux/spinlock.h> +#include <linux/preempt.h> +#include <asm/kdebug.h> +#include <asm/desc.h> + +/* kprobe_status settings */ +#define KPROBE_HIT_ACTIVE 0x00000001 +#define KPROBE_HIT_SS 0x00000002 + +static struct kprobe *current_kprobe; +static unsigned long kprobe_status, kprobe_old_eflags, kprobe_saved_eflags; +static struct pt_regs jprobe_saved_regs; +static long *jprobe_saved_esp; +/* copy of the kernel stack at the probe fire time */ +static kprobe_opcode_t jprobes_stack[MAX_STACK_SIZE]; +void jprobe_return_end(void); + +/* + * returns non-zero if opcode modifies the interrupt flag. + */ +static inline int is_IF_modifier(kprobe_opcode_t opcode) +{ + switch (opcode) { + case 0xfa: /* cli */ + case 0xfb: /* sti */ + case 0xcf: /* iret/iretd */ + case 0x9d: /* popf/popfd */ + return 1; + } + return 0; +} + +int arch_prepare_kprobe(struct kprobe *p) +{ + return 0; +} + +void arch_copy_kprobe(struct kprobe *p) +{ + memcpy(p->ainsn.insn, p->addr, MAX_INSN_SIZE * sizeof(kprobe_opcode_t)); +} + +void arch_remove_kprobe(struct kprobe *p) +{ +} + +static inline void disarm_kprobe(struct kprobe *p, struct pt_regs *regs) +{ + *p->addr = p->opcode; + regs->eip = (unsigned long)p->addr; +} + +static inline void prepare_singlestep(struct kprobe *p, struct pt_regs *regs) +{ + regs->eflags |= TF_MASK; + regs->eflags &= ~IF_MASK; + /*single step inline if the instruction is an int3*/ + if (p->opcode == BREAKPOINT_INSTRUCTION) + regs->eip = (unsigned long)p->addr; + else + regs->eip = (unsigned long)&p->ainsn.insn; +} + +/* + * Interrupts are disabled on entry as trap3 is an interrupt gate and they + * remain disabled thorough out this function. + */ +static int kprobe_handler(struct pt_regs *regs) +{ + struct kprobe *p; + int ret = 0; + kprobe_opcode_t *addr = NULL; + unsigned long *lp; + + /* We're in an interrupt, but this is clear and BUG()-safe. */ + preempt_disable(); + /* Check if the application is using LDT entry for its code segment and + * calculate the address by reading the base address from the LDT entry. + */ + if ((regs->xcs & 4) && (current->mm)) { + lp = (unsigned long *) ((unsigned long)((regs->xcs >> 3) * 8) + + (char *) current->mm->context.ldt); + addr = (kprobe_opcode_t *) (get_desc_base(lp) + regs->eip - + sizeof(kprobe_opcode_t)); + } else { + addr = (kprobe_opcode_t *)(regs->eip - sizeof(kprobe_opcode_t)); + } + /* Check we're not actually recursing */ + if (kprobe_running()) { + /* We *are* holding lock here, so this is safe. + Disarm the probe we just hit, and ignore it. */ + p = get_kprobe(addr); + if (p) { + if (kprobe_status == KPROBE_HIT_SS) { + regs->eflags &= ~TF_MASK; + regs->eflags |= kprobe_saved_eflags; + unlock_kprobes(); + goto no_kprobe; + } + disarm_kprobe(p, regs); + ret = 1; + } else { + p = current_kprobe; + if (p->break_handler && p->break_handler(p, regs)) { + goto ss_probe; + } + } + /* If it's not ours, can't be delete race, (we hold lock). */ + goto no_kprobe; + } + + lock_kprobes(); + p = get_kprobe(addr); + if (!p) { + unlock_kprobes(); + if (regs->eflags & VM_MASK) { + /* We are in virtual-8086 mode. Return 0 */ + goto no_kprobe; + } + + if (*addr != BREAKPOINT_INSTRUCTION) { + /* + * The breakpoint instruction was removed right + * after we hit it. Another cpu has removed + * either a probepoint or a debugger breakpoint + * at this address. In either case, no further + * handling of this interrupt is appropriate. + */ + ret = 1; + } + /* Not one of ours: let kernel handle it */ + goto no_kprobe; + } + + kprobe_status = KPROBE_HIT_ACTIVE; + current_kprobe = p; + kprobe_saved_eflags = kprobe_old_eflags + = (regs->eflags & (TF_MASK | IF_MASK)); + if (is_IF_modifier(p->opcode)) + kprobe_saved_eflags &= ~IF_MASK; + + if (p->pre_handler && p->pre_handler(p, regs)) + /* handler has already set things up, so skip ss setup */ + return 1; + +ss_probe: + prepare_singlestep(p, regs); + kprobe_status = KPROBE_HIT_SS; + return 1; + +no_kprobe: + preempt_enable_no_resched(); + return ret; +} + +/* + * Called after single-stepping. p->addr is the address of the + * instruction whose first byte has been replaced by the "int 3" + * instruction. To avoid the SMP problems that can occur when we + * temporarily put back the original opcode to single-step, we + * single-stepped a copy of the instruction. The address of this + * copy is p->ainsn.insn. + * + * This function prepares to return from the post-single-step + * interrupt. We have to fix up the stack as follows: + * + * 0) Except in the case of absolute or indirect jump or call instructions, + * the new eip is relative to the copied instruction. We need to make + * it relative to the original instruction. + * + * 1) If the single-stepped instruction was pushfl, then the TF and IF + * flags are set in the just-pushed eflags, and may need to be cleared. + * + * 2) If the single-stepped instruction was a call, the return address + * that is atop the stack is the address following the copied instruction. + * We need to make it the address following the original instruction. + */ +static void resume_execution(struct kprobe *p, struct pt_regs *regs) +{ + unsigned long *tos = (unsigned long *)®s->esp; + unsigned long next_eip = 0; + unsigned long copy_eip = (unsigned long)&p->ainsn.insn; + unsigned long orig_eip = (unsigned long)p->addr; + + switch (p->ainsn.insn[0]) { + case 0x9c: /* pushfl */ + *tos &= ~(TF_MASK | IF_MASK); + *tos |= kprobe_old_eflags; + break; + case 0xe8: /* call relative - Fix return addr */ + *tos = orig_eip + (*tos - copy_eip); + break; + case 0xff: + if ((p->ainsn.insn[1] & 0x30) == 0x10) { + /* call absolute, indirect */ + /* Fix return addr; eip is correct. */ + next_eip = regs->eip; + *tos = orig_eip + (*tos - copy_eip); + } else if (((p->ainsn.insn[1] & 0x31) == 0x20) || /* jmp near, absolute indirect */ + ((p->ainsn.insn[1] & 0x31) == 0x21)) { /* jmp far, absolute indirect */ + /* eip is correct. */ + next_eip = regs->eip; + } + break; + case 0xea: /* jmp absolute -- eip is correct */ + next_eip = regs->eip; + break; + default: + break; + } + + regs->eflags &= ~TF_MASK; + if (next_eip) { + regs->eip = next_eip; + } else { + regs->eip = orig_eip + (regs->eip - copy_eip); + } +} + +/* + * Interrupts are disabled on entry as trap1 is an interrupt gate and they + * remain disabled thoroughout this function. And we hold kprobe lock. + */ +static inline int post_kprobe_handler(struct pt_regs *regs) +{ + if (!kprobe_running()) + return 0; + + if (current_kprobe->post_handler) + current_kprobe->post_handler(current_kprobe, regs, 0); + + resume_execution(current_kprobe, regs); + regs->eflags |= kprobe_saved_eflags; + + unlock_kprobes(); + preempt_enable_no_resched(); + + /* + * if somebody else is singlestepping across a probe point, eflags + * will have TF set, in which case, continue the remaining processing + * of do_debug, as if this is not a probe hit. + */ + if (regs->eflags & TF_MASK) + return 0; + + return 1; +} + +/* Interrupts disabled, kprobe_lock held. */ +static inline int kprobe_fault_handler(struct pt_regs *regs, int trapnr) +{ + if (current_kprobe->fault_handler + && current_kprobe->fault_handler(current_kprobe, regs, trapnr)) + return 1; + + if (kprobe_status & KPROBE_HIT_SS) { + resume_execution(current_kprobe, regs); + regs->eflags |= kprobe_old_eflags; + + unlock_kprobes(); + preempt_enable_no_resched(); + } + return 0; +} + +/* + * Wrapper routine to for handling exceptions. + */ +int kprobe_exceptions_notify(struct notifier_block *self, unsigned long val, + void *data) +{ + struct die_args *args = (struct die_args *)data; + switch (val) { + case DIE_INT3: + if (kprobe_handler(args->regs)) + return NOTIFY_STOP; + break; + case DIE_DEBUG: + if (post_kprobe_handler(args->regs)) + return NOTIFY_STOP; + break; + case DIE_GPF: + if (kprobe_running() && + kprobe_fault_handler(args->regs, args->trapnr)) + return NOTIFY_STOP; + break; + case DIE_PAGE_FAULT: + if (kprobe_running() && + kprobe_fault_handler(args->regs, args->trapnr)) + return NOTIFY_STOP; + break; + default: + break; + } + return NOTIFY_DONE; +} + +int setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs) +{ + struct jprobe *jp = container_of(p, struct jprobe, kp); + unsigned long addr; + + jprobe_saved_regs = *regs; + jprobe_saved_esp = ®s->esp; + addr = (unsigned long)jprobe_saved_esp; + + /* + * TBD: As Linus pointed out, gcc assumes that the callee + * owns the argument space and could overwrite it, e.g. + * tailcall optimization. So, to be absolutely safe + * we also save and restore enough stack bytes to cover + * the argument area. + */ + memcpy(jprobes_stack, (kprobe_opcode_t *) addr, MIN_STACK_SIZE(addr)); + regs->eflags &= ~IF_MASK; + regs->eip = (unsigned long)(jp->entry); + return 1; +} + +void jprobe_return(void) +{ + preempt_enable_no_resched(); + asm volatile (" xchgl %%ebx,%%esp \n" + " int3 \n" + " .globl jprobe_return_end \n" + " jprobe_return_end: \n" + " nop \n"::"b" + (jprobe_saved_esp):"memory"); +} + +int longjmp_break_handler(struct kprobe *p, struct pt_regs *regs) +{ + u8 *addr = (u8 *) (regs->eip - 1); + unsigned long stack_addr = (unsigned long)jprobe_saved_esp; + struct jprobe *jp = container_of(p, struct jprobe, kp); + + if ((addr > (u8 *) jprobe_return) && (addr < (u8 *) jprobe_return_end)) { + if (®s->esp != jprobe_saved_esp) { + struct pt_regs *saved_regs = + container_of(jprobe_saved_esp, struct pt_regs, esp); + printk("current esp %p does not match saved esp %p\n", + ®s->esp, jprobe_saved_esp); + printk("Saved registers for jprobe %p\n", jp); + show_registers(saved_regs); + printk("Current registers\n"); + show_registers(regs); + BUG(); + } + *regs = jprobe_saved_regs; + memcpy((kprobe_opcode_t *) stack_addr, jprobes_stack, + MIN_STACK_SIZE(stack_addr)); + return 1; + } + return 0; +} diff --git a/arch/i386/kernel/ldt.c b/arch/i386/kernel/ldt.c new file mode 100644 index 00000000000..bb50afbee92 --- /dev/null +++ b/arch/i386/kernel/ldt.c @@ -0,0 +1,255 @@ +/* + * linux/kernel/ldt.c + * + * Copyright (C) 1992 Krishna Balasubramanian and Linus Torvalds + * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com> + */ + +#include <linux/errno.h> +#include <linux/sched.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/vmalloc.h> +#include <linux/slab.h> + +#include <asm/uaccess.h> +#include <asm/system.h> +#include <asm/ldt.h> +#include <asm/desc.h> + +#ifdef CONFIG_SMP /* avoids "defined but not used" warnig */ +static void flush_ldt(void *null) +{ + if (current->active_mm) + load_LDT(¤t->active_mm->context); +} +#endif + +static int alloc_ldt(mm_context_t *pc, int mincount, int reload) +{ + void *oldldt; + void *newldt; + int oldsize; + + if (mincount <= pc->size) + return 0; + oldsize = pc->size; + mincount = (mincount+511)&(~511); + if (mincount*LDT_ENTRY_SIZE > PAGE_SIZE) + newldt = vmalloc(mincount*LDT_ENTRY_SIZE); + else + newldt = kmalloc(mincount*LDT_ENTRY_SIZE, GFP_KERNEL); + + if (!newldt) + return -ENOMEM; + + if (oldsize) + memcpy(newldt, pc->ldt, oldsize*LDT_ENTRY_SIZE); + oldldt = pc->ldt; + memset(newldt+oldsize*LDT_ENTRY_SIZE, 0, (mincount-oldsize)*LDT_ENTRY_SIZE); + pc->ldt = newldt; + wmb(); + pc->size = mincount; + wmb(); + + if (reload) { +#ifdef CONFIG_SMP + cpumask_t mask; + preempt_disable(); + load_LDT(pc); + mask = cpumask_of_cpu(smp_processor_id()); + if (!cpus_equal(current->mm->cpu_vm_mask, mask)) + smp_call_function(flush_ldt, NULL, 1, 1); + preempt_enable(); +#else + load_LDT(pc); +#endif + } + if (oldsize) { + if (oldsize*LDT_ENTRY_SIZE > PAGE_SIZE) + vfree(oldldt); + else + kfree(oldldt); + } + return 0; +} + +static inline int copy_ldt(mm_context_t *new, mm_context_t *old) +{ + int err = alloc_ldt(new, old->size, 0); + if (err < 0) + return err; + memcpy(new->ldt, old->ldt, old->size*LDT_ENTRY_SIZE); + return 0; +} + +/* + * we do not have to muck with descriptors here, that is + * done in switch_mm() as needed. + */ +int init_new_context(struct task_struct *tsk, struct mm_struct *mm) +{ + struct mm_struct * old_mm; + int retval = 0; + + init_MUTEX(&mm->context.sem); + mm->context.size = 0; + old_mm = current->mm; + if (old_mm && old_mm->context.size > 0) { + down(&old_mm->context.sem); + retval = copy_ldt(&mm->context, &old_mm->context); + up(&old_mm->context.sem); + } + return retval; +} + +/* + * No need to lock the MM as we are the last user + */ +void destroy_context(struct mm_struct *mm) +{ + if (mm->context.size) { + if (mm == current->active_mm) + clear_LDT(); + if (mm->context.size*LDT_ENTRY_SIZE > PAGE_SIZE) + vfree(mm->context.ldt); + else + kfree(mm->context.ldt); + mm->context.size = 0; + } +} + +static int read_ldt(void __user * ptr, unsigned long bytecount) +{ + int err; + unsigned long size; + struct mm_struct * mm = current->mm; + + if (!mm->context.size) + return 0; + if (bytecount > LDT_ENTRY_SIZE*LDT_ENTRIES) + bytecount = LDT_ENTRY_SIZE*LDT_ENTRIES; + + down(&mm->context.sem); + size = mm->context.size*LDT_ENTRY_SIZE; + if (size > bytecount) + size = bytecount; + + err = 0; + if (copy_to_user(ptr, mm->context.ldt, size)) + err = -EFAULT; + up(&mm->context.sem); + if (err < 0) + goto error_return; + if (size != bytecount) { + /* zero-fill the rest */ + if (clear_user(ptr+size, bytecount-size) != 0) { + err = -EFAULT; + goto error_return; + } + } + return bytecount; +error_return: + return err; +} + +static int read_default_ldt(void __user * ptr, unsigned long bytecount) +{ + int err; + unsigned long size; + void *address; + + err = 0; + address = &default_ldt[0]; + size = 5*sizeof(struct desc_struct); + if (size > bytecount) + size = bytecount; + + err = size; + if (copy_to_user(ptr, address, size)) + err = -EFAULT; + + return err; +} + +static int write_ldt(void __user * ptr, unsigned long bytecount, int oldmode) +{ + struct mm_struct * mm = current->mm; + __u32 entry_1, entry_2, *lp; + int error; + struct user_desc ldt_info; + + error = -EINVAL; + if (bytecount != sizeof(ldt_info)) + goto out; + error = -EFAULT; + if (copy_from_user(&ldt_info, ptr, sizeof(ldt_info))) + goto out; + + error = -EINVAL; + if (ldt_info.entry_number >= LDT_ENTRIES) + goto out; + if (ldt_info.contents == 3) { + if (oldmode) + goto out; + if (ldt_info.seg_not_present == 0) + goto out; + } + + down(&mm->context.sem); + if (ldt_info.entry_number >= mm->context.size) { + error = alloc_ldt(¤t->mm->context, ldt_info.entry_number+1, 1); + if (error < 0) + goto out_unlock; + } + + lp = (__u32 *) ((ldt_info.entry_number << 3) + (char *) mm->context.ldt); + + /* Allow LDTs to be cleared by the user. */ + if (ldt_info.base_addr == 0 && ldt_info.limit == 0) { + if (oldmode || LDT_empty(&ldt_info)) { + entry_1 = 0; + entry_2 = 0; + goto install; + } + } + + entry_1 = LDT_entry_a(&ldt_info); + entry_2 = LDT_entry_b(&ldt_info); + if (oldmode) + entry_2 &= ~(1 << 20); + + /* Install the new entry ... */ +install: + *lp = entry_1; + *(lp+1) = entry_2; + error = 0; + +out_unlock: + up(&mm->context.sem); +out: + return error; +} + +asmlinkage int sys_modify_ldt(int func, void __user *ptr, unsigned long bytecount) +{ + int ret = -ENOSYS; + + switch (func) { + case 0: + ret = read_ldt(ptr, bytecount); + break; + case 1: + ret = write_ldt(ptr, bytecount, 1); + break; + case 2: + ret = read_default_ldt(ptr, bytecount); + break; + case 0x11: + ret = write_ldt(ptr, bytecount, 0); + break; + } + return ret; +} diff --git a/arch/i386/kernel/mca.c b/arch/i386/kernel/mca.c new file mode 100644 index 00000000000..8600faeea29 --- /dev/null +++ b/arch/i386/kernel/mca.c @@ -0,0 +1,474 @@ +/* + * linux/arch/i386/kernel/mca.c + * Written by Martin Kolinek, February 1996 + * + * Changes: + * + * Chris Beauregard July 28th, 1996 + * - Fixed up integrated SCSI detection + * + * Chris Beauregard August 3rd, 1996 + * - Made mca_info local + * - Made integrated registers accessible through standard function calls + * - Added name field + * - More sanity checking + * + * Chris Beauregard August 9th, 1996 + * - Rewrote /proc/mca + * + * Chris Beauregard January 7th, 1997 + * - Added basic NMI-processing + * - Added more information to mca_info structure + * + * David Weinehall October 12th, 1998 + * - Made a lot of cleaning up in the source + * - Added use of save_flags / restore_flags + * - Added the 'driver_loaded' flag in MCA_adapter + * - Added an alternative implemention of ZP Gu's mca_find_unused_adapter + * + * David Weinehall March 24th, 1999 + * - Fixed the output of 'Driver Installed' in /proc/mca/pos + * - Made the Integrated Video & SCSI show up even if they have id 0000 + * + * Alexander Viro November 9th, 1999 + * - Switched to regular procfs methods + * + * Alfred Arnold & David Weinehall August 23rd, 2000 + * - Added support for Planar POS-registers + */ + +#include <linux/module.h> +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/mca.h> +#include <asm/system.h> +#include <asm/io.h> +#include <linux/proc_fs.h> +#include <linux/mman.h> +#include <linux/config.h> +#include <linux/mm.h> +#include <linux/pagemap.h> +#include <linux/ioport.h> +#include <asm/uaccess.h> +#include <linux/init.h> +#include <asm/arch_hooks.h> + +static unsigned char which_scsi = 0; + +int MCA_bus = 0; +EXPORT_SYMBOL(MCA_bus); + +/* + * Motherboard register spinlock. Untested on SMP at the moment, but + * are there any MCA SMP boxes? + * + * Yes - Alan + */ +static DEFINE_SPINLOCK(mca_lock); + +/* Build the status info for the adapter */ + +static void mca_configure_adapter_status(struct mca_device *mca_dev) { + mca_dev->status = MCA_ADAPTER_NONE; + + mca_dev->pos_id = mca_dev->pos[0] + + (mca_dev->pos[1] << 8); + + if(!mca_dev->pos_id && mca_dev->slot < MCA_MAX_SLOT_NR) { + + /* id = 0x0000 usually indicates hardware failure, + * however, ZP Gu (zpg@castle.net> reports that his 9556 + * has 0x0000 as id and everything still works. There + * also seem to be an adapter with id = 0x0000; the + * NCR Parallel Bus Memory Card. Until this is confirmed, + * however, this code will stay. + */ + + mca_dev->status = MCA_ADAPTER_ERROR; + + return; + } else if(mca_dev->pos_id != 0xffff) { + + /* 0xffff usually indicates that there's no adapter, + * however, some integrated adapters may have 0xffff as + * their id and still be valid. Examples are on-board + * VGA of the 55sx, the integrated SCSI of the 56 & 57, + * and possibly also the 95 ULTIMEDIA. + */ + + mca_dev->status = MCA_ADAPTER_NORMAL; + } + + if((mca_dev->pos_id == 0xffff || + mca_dev->pos_id == 0x0000) && mca_dev->slot >= MCA_MAX_SLOT_NR) { + int j; + + for(j = 2; j < 8; j++) { + if(mca_dev->pos[j] != 0xff) { + mca_dev->status = MCA_ADAPTER_NORMAL; + break; + } + } + } + + if(!(mca_dev->pos[2] & MCA_ENABLED)) { + + /* enabled bit is in POS 2 */ + + mca_dev->status = MCA_ADAPTER_DISABLED; + } +} /* mca_configure_adapter_status */ + +/*--------------------------------------------------------------------*/ + +static struct resource mca_standard_resources[] = { + { .start = 0x60, .end = 0x60, .name = "system control port B (MCA)" }, + { .start = 0x90, .end = 0x90, .name = "arbitration (MCA)" }, + { .start = 0x91, .end = 0x91, .name = "card Select Feedback (MCA)" }, + { .start = 0x92, .end = 0x92, .name = "system Control port A (MCA)" }, + { .start = 0x94, .end = 0x94, .name = "system board setup (MCA)" }, + { .start = 0x96, .end = 0x97, .name = "POS (MCA)" }, + { .start = 0x100, .end = 0x107, .name = "POS (MCA)" } +}; + +#define MCA_STANDARD_RESOURCES (sizeof(mca_standard_resources)/sizeof(struct resource)) + +/** + * mca_read_and_store_pos - read the POS registers into a memory buffer + * @pos: a char pointer to 8 bytes, contains the POS register value on + * successful return + * + * Returns 1 if a card actually exists (i.e. the pos isn't + * all 0xff) or 0 otherwise + */ +static int mca_read_and_store_pos(unsigned char *pos) { + int j; + int found = 0; + + for(j=0; j<8; j++) { + if((pos[j] = inb_p(MCA_POS_REG(j))) != 0xff) { + /* 0xff all across means no device. 0x00 means + * something's broken, but a device is + * probably there. However, if you get 0x00 + * from a motherboard register it won't matter + * what we find. For the record, on the + * 57SLC, the integrated SCSI adapter has + * 0xffff for the adapter ID, but nonzero for + * other registers. */ + + found = 1; + } + } + return found; +} + +static unsigned char mca_pc_read_pos(struct mca_device *mca_dev, int reg) +{ + unsigned char byte; + unsigned long flags; + + if(reg < 0 || reg >= 8) + return 0; + + spin_lock_irqsave(&mca_lock, flags); + if(mca_dev->pos_register) { + /* Disable adapter setup, enable motherboard setup */ + + outb_p(0, MCA_ADAPTER_SETUP_REG); + outb_p(mca_dev->pos_register, MCA_MOTHERBOARD_SETUP_REG); + + byte = inb_p(MCA_POS_REG(reg)); + outb_p(0xff, MCA_MOTHERBOARD_SETUP_REG); + } else { + + /* Make sure motherboard setup is off */ + + outb_p(0xff, MCA_MOTHERBOARD_SETUP_REG); + + /* Read the appropriate register */ + + outb_p(0x8|(mca_dev->slot & 0xf), MCA_ADAPTER_SETUP_REG); + byte = inb_p(MCA_POS_REG(reg)); + outb_p(0, MCA_ADAPTER_SETUP_REG); + } + spin_unlock_irqrestore(&mca_lock, flags); + + mca_dev->pos[reg] = byte; + + return byte; +} + +static void mca_pc_write_pos(struct mca_device *mca_dev, int reg, + unsigned char byte) +{ + unsigned long flags; + + if(reg < 0 || reg >= 8) + return; + + spin_lock_irqsave(&mca_lock, flags); + + /* Make sure motherboard setup is off */ + + outb_p(0xff, MCA_MOTHERBOARD_SETUP_REG); + + /* Read in the appropriate register */ + + outb_p(0x8|(mca_dev->slot&0xf), MCA_ADAPTER_SETUP_REG); + outb_p(byte, MCA_POS_REG(reg)); + outb_p(0, MCA_ADAPTER_SETUP_REG); + + spin_unlock_irqrestore(&mca_lock, flags); + + /* Update the global register list, while we have the byte */ + + mca_dev->pos[reg] = byte; + +} + +/* for the primary MCA bus, we have identity transforms */ +static int mca_dummy_transform_irq(struct mca_device * mca_dev, int irq) +{ + return irq; +} + +static int mca_dummy_transform_ioport(struct mca_device * mca_dev, int port) +{ + return port; +} + +static void *mca_dummy_transform_memory(struct mca_device * mca_dev, void *mem) +{ + return mem; +} + + +static int __init mca_init(void) +{ + unsigned int i, j; + struct mca_device *mca_dev; + unsigned char pos[8]; + short mca_builtin_scsi_ports[] = {0xf7, 0xfd, 0x00}; + struct mca_bus *bus; + + /* WARNING: Be careful when making changes here. Putting an adapter + * and the motherboard simultaneously into setup mode may result in + * damage to chips (according to The Indispensible PC Hardware Book + * by Hans-Peter Messmer). Also, we disable system interrupts (so + * that we are not disturbed in the middle of this). + */ + + /* Make sure the MCA bus is present */ + + if (mca_system_init()) { + printk(KERN_ERR "MCA bus system initialisation failed\n"); + return -ENODEV; + } + + if (!MCA_bus) + return -ENODEV; + + printk(KERN_INFO "Micro Channel bus detected.\n"); + + /* All MCA systems have at least a primary bus */ + bus = mca_attach_bus(MCA_PRIMARY_BUS); + if (!bus) + goto out_nomem; + bus->default_dma_mask = 0xffffffffLL; + bus->f.mca_write_pos = mca_pc_write_pos; + bus->f.mca_read_pos = mca_pc_read_pos; + bus->f.mca_transform_irq = mca_dummy_transform_irq; + bus->f.mca_transform_ioport = mca_dummy_transform_ioport; + bus->f.mca_transform_memory = mca_dummy_transform_memory; + + /* get the motherboard device */ + mca_dev = kmalloc(sizeof(struct mca_device), GFP_KERNEL); + if(unlikely(!mca_dev)) + goto out_nomem; + memset(mca_dev, 0, sizeof(struct mca_device)); + + /* + * We do not expect many MCA interrupts during initialization, + * but let us be safe: + */ + spin_lock_irq(&mca_lock); + + /* Make sure adapter setup is off */ + + outb_p(0, MCA_ADAPTER_SETUP_REG); + + /* Read motherboard POS registers */ + + mca_dev->pos_register = 0x7f; + outb_p(mca_dev->pos_register, MCA_MOTHERBOARD_SETUP_REG); + mca_dev->name[0] = 0; + mca_read_and_store_pos(mca_dev->pos); + mca_configure_adapter_status(mca_dev); + /* fake POS and slot for a motherboard */ + mca_dev->pos_id = MCA_MOTHERBOARD_POS; + mca_dev->slot = MCA_MOTHERBOARD; + mca_register_device(MCA_PRIMARY_BUS, mca_dev); + + mca_dev = kmalloc(sizeof(struct mca_device), GFP_ATOMIC); + if(unlikely(!mca_dev)) + goto out_unlock_nomem; + memset(mca_dev, 0, sizeof(struct mca_device)); + + + /* Put motherboard into video setup mode, read integrated video + * POS registers, and turn motherboard setup off. + */ + + mca_dev->pos_register = 0xdf; + outb_p(mca_dev->pos_register, MCA_MOTHERBOARD_SETUP_REG); + mca_dev->name[0] = 0; + mca_read_and_store_pos(mca_dev->pos); + mca_configure_adapter_status(mca_dev); + /* fake POS and slot for the integrated video */ + mca_dev->pos_id = MCA_INTEGVIDEO_POS; + mca_dev->slot = MCA_INTEGVIDEO; + mca_register_device(MCA_PRIMARY_BUS, mca_dev); + + /* Put motherboard into scsi setup mode, read integrated scsi + * POS registers, and turn motherboard setup off. + * + * It seems there are two possible SCSI registers. Martin says that + * for the 56,57, 0xf7 is the one, but fails on the 76. + * Alfredo (apena@vnet.ibm.com) says + * 0xfd works on his machine. We'll try both of them. I figure it's + * a good bet that only one could be valid at a time. This could + * screw up though if one is used for something else on the other + * machine. + */ + + for(i = 0; (which_scsi = mca_builtin_scsi_ports[i]) != 0; i++) { + outb_p(which_scsi, MCA_MOTHERBOARD_SETUP_REG); + if(mca_read_and_store_pos(pos)) + break; + } + if(which_scsi) { + /* found a scsi card */ + mca_dev = kmalloc(sizeof(struct mca_device), GFP_ATOMIC); + if(unlikely(!mca_dev)) + goto out_unlock_nomem; + memset(mca_dev, 0, sizeof(struct mca_device)); + + for(j = 0; j < 8; j++) + mca_dev->pos[j] = pos[j]; + + mca_configure_adapter_status(mca_dev); + /* fake POS and slot for integrated SCSI controller */ + mca_dev->pos_id = MCA_INTEGSCSI_POS; + mca_dev->slot = MCA_INTEGSCSI; + mca_dev->pos_register = which_scsi; + mca_register_device(MCA_PRIMARY_BUS, mca_dev); + } + + /* Turn off motherboard setup */ + + outb_p(0xff, MCA_MOTHERBOARD_SETUP_REG); + + /* Now loop over MCA slots: put each adapter into setup mode, and + * read its POS registers. Then put adapter setup off. + */ + + for(i=0; i<MCA_MAX_SLOT_NR; i++) { + outb_p(0x8|(i&0xf), MCA_ADAPTER_SETUP_REG); + if(!mca_read_and_store_pos(pos)) + continue; + + mca_dev = kmalloc(sizeof(struct mca_device), GFP_ATOMIC); + if(unlikely(!mca_dev)) + goto out_unlock_nomem; + memset(mca_dev, 0, sizeof(struct mca_device)); + + for(j=0; j<8; j++) + mca_dev->pos[j]=pos[j]; + + mca_dev->driver_loaded = 0; + mca_dev->slot = i; + mca_dev->pos_register = 0; + mca_configure_adapter_status(mca_dev); + mca_register_device(MCA_PRIMARY_BUS, mca_dev); + } + outb_p(0, MCA_ADAPTER_SETUP_REG); + + /* Enable interrupts and return memory start */ + spin_unlock_irq(&mca_lock); + + for (i = 0; i < MCA_STANDARD_RESOURCES; i++) + request_resource(&ioport_resource, mca_standard_resources + i); + + mca_do_proc_init(); + + return 0; + + out_unlock_nomem: + spin_unlock_irq(&mca_lock); + out_nomem: + printk(KERN_EMERG "Failed memory allocation in MCA setup!\n"); + return -ENOMEM; +} + +subsys_initcall(mca_init); + +/*--------------------------------------------------------------------*/ + +static void mca_handle_nmi_device(struct mca_device *mca_dev, int check_flag) +{ + int slot = mca_dev->slot; + + if(slot == MCA_INTEGSCSI) { + printk(KERN_CRIT "NMI: caused by MCA integrated SCSI adapter (%s)\n", + mca_dev->name); + } else if(slot == MCA_INTEGVIDEO) { + printk(KERN_CRIT "NMI: caused by MCA integrated video adapter (%s)\n", + mca_dev->name); + } else if(slot == MCA_MOTHERBOARD) { + printk(KERN_CRIT "NMI: caused by motherboard (%s)\n", + mca_dev->name); + } + + /* More info available in POS 6 and 7? */ + + if(check_flag) { + unsigned char pos6, pos7; + + pos6 = mca_device_read_pos(mca_dev, 6); + pos7 = mca_device_read_pos(mca_dev, 7); + + printk(KERN_CRIT "NMI: POS 6 = 0x%x, POS 7 = 0x%x\n", pos6, pos7); + } + +} /* mca_handle_nmi_slot */ + +/*--------------------------------------------------------------------*/ + +static int mca_handle_nmi_callback(struct device *dev, void *data) +{ + struct mca_device *mca_dev = to_mca_device(dev); + unsigned char pos5; + + pos5 = mca_device_read_pos(mca_dev, 5); + + if(!(pos5 & 0x80)) { + /* Bit 7 of POS 5 is reset when this adapter has a hardware + * error. Bit 7 it reset if there's error information + * available in POS 6 and 7. + */ + mca_handle_nmi_device(mca_dev, !(pos5 & 0x40)); + return 1; + } + return 0; +} + +void mca_handle_nmi(void) +{ + /* First try - scan the various adapters and see if a specific + * adapter was responsible for the error. + */ + bus_for_each_dev(&mca_bus_type, NULL, NULL, mca_handle_nmi_callback); + + mca_nmi_hook(); +} /* mca_handle_nmi */ diff --git a/arch/i386/kernel/microcode.c b/arch/i386/kernel/microcode.c new file mode 100644 index 00000000000..a77c612aad0 --- /dev/null +++ b/arch/i386/kernel/microcode.c @@ -0,0 +1,512 @@ +/* + * Intel CPU Microcode Update Driver for Linux + * + * Copyright (C) 2000-2004 Tigran Aivazian + * + * This driver allows to upgrade microcode on Intel processors + * belonging to IA-32 family - PentiumPro, Pentium II, + * Pentium III, Xeon, Pentium 4, etc. + * + * Reference: Section 8.10 of Volume III, Intel Pentium 4 Manual, + * Order Number 245472 or free download from: + * + * http://developer.intel.com/design/pentium4/manuals/245472.htm + * + * For more information, go to http://www.urbanmyth.org/microcode + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * 1.0 16 Feb 2000, Tigran Aivazian <tigran@sco.com> + * Initial release. + * 1.01 18 Feb 2000, Tigran Aivazian <tigran@sco.com> + * Added read() support + cleanups. + * 1.02 21 Feb 2000, Tigran Aivazian <tigran@sco.com> + * Added 'device trimming' support. open(O_WRONLY) zeroes + * and frees the saved copy of applied microcode. + * 1.03 29 Feb 2000, Tigran Aivazian <tigran@sco.com> + * Made to use devfs (/dev/cpu/microcode) + cleanups. + * 1.04 06 Jun 2000, Simon Trimmer <simon@veritas.com> + * Added misc device support (now uses both devfs and misc). + * Added MICROCODE_IOCFREE ioctl to clear memory. + * 1.05 09 Jun 2000, Simon Trimmer <simon@veritas.com> + * Messages for error cases (non Intel & no suitable microcode). + * 1.06 03 Aug 2000, Tigran Aivazian <tigran@veritas.com> + * Removed ->release(). Removed exclusive open and status bitmap. + * Added microcode_rwsem to serialize read()/write()/ioctl(). + * Removed global kernel lock usage. + * 1.07 07 Sep 2000, Tigran Aivazian <tigran@veritas.com> + * Write 0 to 0x8B msr and then cpuid before reading revision, + * so that it works even if there were no update done by the + * BIOS. Otherwise, reading from 0x8B gives junk (which happened + * to be 0 on my machine which is why it worked even when I + * disabled update by the BIOS) + * Thanks to Eric W. Biederman <ebiederman@lnxi.com> for the fix. + * 1.08 11 Dec 2000, Richard Schaal <richard.schaal@intel.com> and + * Tigran Aivazian <tigran@veritas.com> + * Intel Pentium 4 processor support and bugfixes. + * 1.09 30 Oct 2001, Tigran Aivazian <tigran@veritas.com> + * Bugfix for HT (Hyper-Threading) enabled processors + * whereby processor resources are shared by all logical processors + * in a single CPU package. + * 1.10 28 Feb 2002 Asit K Mallick <asit.k.mallick@intel.com> and + * Tigran Aivazian <tigran@veritas.com>, + * Serialize updates as required on HT processors due to speculative + * nature of implementation. + * 1.11 22 Mar 2002 Tigran Aivazian <tigran@veritas.com> + * Fix the panic when writing zero-length microcode chunk. + * 1.12 29 Sep 2003 Nitin Kamble <nitin.a.kamble@intel.com>, + * Jun Nakajima <jun.nakajima@intel.com> + * Support for the microcode updates in the new format. + * 1.13 10 Oct 2003 Tigran Aivazian <tigran@veritas.com> + * Removed ->read() method and obsoleted MICROCODE_IOCFREE ioctl + * because we no longer hold a copy of applied microcode + * in kernel memory. + * 1.14 25 Jun 2004 Tigran Aivazian <tigran@veritas.com> + * Fix sigmatch() macro to handle old CPUs with pf == 0. + * Thanks to Stuart Swales for pointing out this bug. + */ + +//#define DEBUG /* pr_debug */ +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/sched.h> +#include <linux/module.h> +#include <linux/slab.h> +#include <linux/vmalloc.h> +#include <linux/miscdevice.h> +#include <linux/spinlock.h> +#include <linux/mm.h> + +#include <asm/msr.h> +#include <asm/uaccess.h> +#include <asm/processor.h> + +MODULE_DESCRIPTION("Intel CPU (IA-32) Microcode Update Driver"); +MODULE_AUTHOR("Tigran Aivazian <tigran@veritas.com>"); +MODULE_LICENSE("GPL"); + +#define MICROCODE_VERSION "1.14" + +#define DEFAULT_UCODE_DATASIZE (2000) /* 2000 bytes */ +#define MC_HEADER_SIZE (sizeof (microcode_header_t)) /* 48 bytes */ +#define DEFAULT_UCODE_TOTALSIZE (DEFAULT_UCODE_DATASIZE + MC_HEADER_SIZE) /* 2048 bytes */ +#define EXT_HEADER_SIZE (sizeof (struct extended_sigtable)) /* 20 bytes */ +#define EXT_SIGNATURE_SIZE (sizeof (struct extended_signature)) /* 12 bytes */ +#define DWSIZE (sizeof (u32)) +#define get_totalsize(mc) \ + (((microcode_t *)mc)->hdr.totalsize ? \ + ((microcode_t *)mc)->hdr.totalsize : DEFAULT_UCODE_TOTALSIZE) +#define get_datasize(mc) \ + (((microcode_t *)mc)->hdr.datasize ? \ + ((microcode_t *)mc)->hdr.datasize : DEFAULT_UCODE_DATASIZE) + +#define sigmatch(s1, s2, p1, p2) \ + (((s1) == (s2)) && (((p1) & (p2)) || (((p1) == 0) && ((p2) == 0)))) + +#define exttable_size(et) ((et)->count * EXT_SIGNATURE_SIZE + EXT_HEADER_SIZE) + +/* serialize access to the physical write to MSR 0x79 */ +static DEFINE_SPINLOCK(microcode_update_lock); + +/* no concurrent ->write()s are allowed on /dev/cpu/microcode */ +static DECLARE_MUTEX(microcode_sem); + +static void __user *user_buffer; /* user area microcode data buffer */ +static unsigned int user_buffer_size; /* it's size */ + +typedef enum mc_error_code { + MC_SUCCESS = 0, + MC_NOTFOUND = 1, + MC_MARKED = 2, + MC_ALLOCATED = 3, +} mc_error_code_t; + +static struct ucode_cpu_info { + unsigned int sig; + unsigned int pf; + unsigned int rev; + unsigned int cksum; + mc_error_code_t err; + microcode_t *mc; +} ucode_cpu_info[NR_CPUS]; + +static int microcode_open (struct inode *unused1, struct file *unused2) +{ + return capable(CAP_SYS_RAWIO) ? 0 : -EPERM; +} + +static void collect_cpu_info (void *unused) +{ + int cpu_num = smp_processor_id(); + struct cpuinfo_x86 *c = cpu_data + cpu_num; + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + unsigned int val[2]; + + uci->sig = uci->pf = uci->rev = uci->cksum = 0; + uci->err = MC_NOTFOUND; + uci->mc = NULL; + + if (c->x86_vendor != X86_VENDOR_INTEL || c->x86 < 6 || + cpu_has(c, X86_FEATURE_IA64)) { + printk(KERN_ERR "microcode: CPU%d not a capable Intel processor\n", cpu_num); + return; + } else { + uci->sig = cpuid_eax(0x00000001); + + if ((c->x86_model >= 5) || (c->x86 > 6)) { + /* get processor flags from MSR 0x17 */ + rdmsr(MSR_IA32_PLATFORM_ID, val[0], val[1]); + uci->pf = 1 << ((val[1] >> 18) & 7); + } + } + + wrmsr(MSR_IA32_UCODE_REV, 0, 0); + __asm__ __volatile__ ("cpuid" : : : "ax", "bx", "cx", "dx"); + /* get the current revision from MSR 0x8B */ + rdmsr(MSR_IA32_UCODE_REV, val[0], uci->rev); + pr_debug("microcode: collect_cpu_info : sig=0x%x, pf=0x%x, rev=0x%x\n", + uci->sig, uci->pf, uci->rev); +} + +static inline void mark_microcode_update (int cpu_num, microcode_header_t *mc_header, int sig, int pf, int cksum) +{ + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + + pr_debug("Microcode Found.\n"); + pr_debug(" Header Revision 0x%x\n", mc_header->hdrver); + pr_debug(" Loader Revision 0x%x\n", mc_header->ldrver); + pr_debug(" Revision 0x%x \n", mc_header->rev); + pr_debug(" Date %x/%x/%x\n", + ((mc_header->date >> 24 ) & 0xff), + ((mc_header->date >> 16 ) & 0xff), + (mc_header->date & 0xFFFF)); + pr_debug(" Signature 0x%x\n", sig); + pr_debug(" Type 0x%x Family 0x%x Model 0x%x Stepping 0x%x\n", + ((sig >> 12) & 0x3), + ((sig >> 8) & 0xf), + ((sig >> 4) & 0xf), + ((sig & 0xf))); + pr_debug(" Processor Flags 0x%x\n", pf); + pr_debug(" Checksum 0x%x\n", cksum); + + if (mc_header->rev < uci->rev) { + printk(KERN_ERR "microcode: CPU%d not 'upgrading' to earlier revision" + " 0x%x (current=0x%x)\n", cpu_num, mc_header->rev, uci->rev); + goto out; + } else if (mc_header->rev == uci->rev) { + /* notify the caller of success on this cpu */ + uci->err = MC_SUCCESS; + printk(KERN_ERR "microcode: CPU%d already at revision" + " 0x%x (current=0x%x)\n", cpu_num, mc_header->rev, uci->rev); + goto out; + } + + pr_debug("microcode: CPU%d found a matching microcode update with " + " revision 0x%x (current=0x%x)\n", cpu_num, mc_header->rev, uci->rev); + uci->cksum = cksum; + uci->pf = pf; /* keep the original mc pf for cksum calculation */ + uci->err = MC_MARKED; /* found the match */ +out: + return; +} + +static int find_matching_ucodes (void) +{ + int cursor = 0; + int error = 0; + + while (cursor + MC_HEADER_SIZE < user_buffer_size) { + microcode_header_t mc_header; + void *newmc = NULL; + int i, sum, cpu_num, allocated_flag, total_size, data_size, ext_table_size; + + if (copy_from_user(&mc_header, user_buffer + cursor, MC_HEADER_SIZE)) { + printk(KERN_ERR "microcode: error! Can not read user data\n"); + error = -EFAULT; + goto out; + } + + total_size = get_totalsize(&mc_header); + if ((cursor + total_size > user_buffer_size) || (total_size < DEFAULT_UCODE_TOTALSIZE)) { + printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + error = -EINVAL; + goto out; + } + + data_size = get_datasize(&mc_header); + if ((data_size + MC_HEADER_SIZE > total_size) || (data_size < DEFAULT_UCODE_DATASIZE)) { + printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + error = -EINVAL; + goto out; + } + + if (mc_header.ldrver != 1 || mc_header.hdrver != 1) { + printk(KERN_ERR "microcode: error! Unknown microcode update format\n"); + error = -EINVAL; + goto out; + } + + for (cpu_num = 0; cpu_num < num_online_cpus(); cpu_num++) { + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + if (uci->err != MC_NOTFOUND) /* already found a match or not an online cpu*/ + continue; + + if (sigmatch(mc_header.sig, uci->sig, mc_header.pf, uci->pf)) + mark_microcode_update(cpu_num, &mc_header, mc_header.sig, mc_header.pf, mc_header.cksum); + } + + ext_table_size = total_size - (MC_HEADER_SIZE + data_size); + if (ext_table_size) { + struct extended_sigtable ext_header; + struct extended_signature ext_sig; + int ext_sigcount; + + if ((ext_table_size < EXT_HEADER_SIZE) + || ((ext_table_size - EXT_HEADER_SIZE) % EXT_SIGNATURE_SIZE)) { + printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + error = -EINVAL; + goto out; + } + if (copy_from_user(&ext_header, user_buffer + cursor + + MC_HEADER_SIZE + data_size, EXT_HEADER_SIZE)) { + printk(KERN_ERR "microcode: error! Can not read user data\n"); + error = -EFAULT; + goto out; + } + if (ext_table_size != exttable_size(&ext_header)) { + printk(KERN_ERR "microcode: error! Bad data in microcode data file\n"); + error = -EFAULT; + goto out; + } + + ext_sigcount = ext_header.count; + + for (i = 0; i < ext_sigcount; i++) { + if (copy_from_user(&ext_sig, user_buffer + cursor + MC_HEADER_SIZE + data_size + EXT_HEADER_SIZE + + EXT_SIGNATURE_SIZE * i, EXT_SIGNATURE_SIZE)) { + printk(KERN_ERR "microcode: error! Can not read user data\n"); + error = -EFAULT; + goto out; + } + for (cpu_num = 0; cpu_num < num_online_cpus(); cpu_num++) { + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + if (uci->err != MC_NOTFOUND) /* already found a match or not an online cpu*/ + continue; + if (sigmatch(ext_sig.sig, uci->sig, ext_sig.pf, uci->pf)) { + mark_microcode_update(cpu_num, &mc_header, ext_sig.sig, ext_sig.pf, ext_sig.cksum); + } + } + } + } + /* now check if any cpu has matched */ + for (cpu_num = 0, allocated_flag = 0, sum = 0; cpu_num < num_online_cpus(); cpu_num++) { + if (ucode_cpu_info[cpu_num].err == MC_MARKED) { + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + if (!allocated_flag) { + allocated_flag = 1; + newmc = vmalloc(total_size); + if (!newmc) { + printk(KERN_ERR "microcode: error! Can not allocate memory\n"); + error = -ENOMEM; + goto out; + } + if (copy_from_user(newmc + MC_HEADER_SIZE, + user_buffer + cursor + MC_HEADER_SIZE, + total_size - MC_HEADER_SIZE)) { + printk(KERN_ERR "microcode: error! Can not read user data\n"); + vfree(newmc); + error = -EFAULT; + goto out; + } + memcpy(newmc, &mc_header, MC_HEADER_SIZE); + /* check extended table checksum */ + if (ext_table_size) { + int ext_table_sum = 0; + int * ext_tablep = (((void *) newmc) + MC_HEADER_SIZE + data_size); + i = ext_table_size / DWSIZE; + while (i--) ext_table_sum += ext_tablep[i]; + if (ext_table_sum) { + printk(KERN_WARNING "microcode: aborting, bad extended signature table checksum\n"); + vfree(newmc); + error = -EINVAL; + goto out; + } + } + + /* calculate the checksum */ + i = (MC_HEADER_SIZE + data_size) / DWSIZE; + while (i--) sum += ((int *)newmc)[i]; + sum -= (mc_header.sig + mc_header.pf + mc_header.cksum); + } + ucode_cpu_info[cpu_num].mc = newmc; + ucode_cpu_info[cpu_num].err = MC_ALLOCATED; /* mc updated */ + if (sum + uci->sig + uci->pf + uci->cksum != 0) { + printk(KERN_ERR "microcode: CPU%d aborting, bad checksum\n", cpu_num); + error = -EINVAL; + goto out; + } + } + } + cursor += total_size; /* goto the next update patch */ + } /* end of while */ +out: + return error; +} + +static void do_update_one (void * unused) +{ + unsigned long flags; + unsigned int val[2]; + int cpu_num = smp_processor_id(); + struct ucode_cpu_info *uci = ucode_cpu_info + cpu_num; + + if (uci->mc == NULL) { + printk(KERN_INFO "microcode: No new microcode data for CPU%d\n", cpu_num); + return; + } + + /* serialize access to the physical write to MSR 0x79 */ + spin_lock_irqsave(µcode_update_lock, flags); + + /* write microcode via MSR 0x79 */ + wrmsr(MSR_IA32_UCODE_WRITE, + (unsigned long) uci->mc->bits, + (unsigned long) uci->mc->bits >> 16 >> 16); + wrmsr(MSR_IA32_UCODE_REV, 0, 0); + + __asm__ __volatile__ ("cpuid" : : : "ax", "bx", "cx", "dx"); + /* get the current revision from MSR 0x8B */ + rdmsr(MSR_IA32_UCODE_REV, val[0], val[1]); + + /* notify the caller of success on this cpu */ + uci->err = MC_SUCCESS; + spin_unlock_irqrestore(µcode_update_lock, flags); + printk(KERN_INFO "microcode: CPU%d updated from revision " + "0x%x to 0x%x, date = %08x \n", + cpu_num, uci->rev, val[1], uci->mc->hdr.date); + return; +} + +static int do_microcode_update (void) +{ + int i, error; + + if (on_each_cpu(collect_cpu_info, NULL, 1, 1) != 0) { + printk(KERN_ERR "microcode: Error! Could not run on all processors\n"); + error = -EIO; + goto out; + } + + if ((error = find_matching_ucodes())) { + printk(KERN_ERR "microcode: Error in the microcode data\n"); + goto out_free; + } + + if (on_each_cpu(do_update_one, NULL, 1, 1) != 0) { + printk(KERN_ERR "microcode: Error! Could not run on all processors\n"); + error = -EIO; + } + +out_free: + for (i = 0; i < num_online_cpus(); i++) { + if (ucode_cpu_info[i].mc) { + int j; + void *tmp = ucode_cpu_info[i].mc; + vfree(tmp); + for (j = i; j < num_online_cpus(); j++) { + if (ucode_cpu_info[j].mc == tmp) + ucode_cpu_info[j].mc = NULL; + } + } + } +out: + return error; +} + +static ssize_t microcode_write (struct file *file, const char __user *buf, size_t len, loff_t *ppos) +{ + ssize_t ret; + + if (len < DEFAULT_UCODE_TOTALSIZE) { + printk(KERN_ERR "microcode: not enough data\n"); + return -EINVAL; + } + + if ((len >> PAGE_SHIFT) > num_physpages) { + printk(KERN_ERR "microcode: too much data (max %ld pages)\n", num_physpages); + return -EINVAL; + } + + down(µcode_sem); + + user_buffer = (void __user *) buf; + user_buffer_size = (int) len; + + ret = do_microcode_update(); + if (!ret) + ret = (ssize_t)len; + + up(µcode_sem); + + return ret; +} + +static int microcode_ioctl (struct inode *inode, struct file *file, + unsigned int cmd, unsigned long arg) +{ + switch (cmd) { + /* + * XXX: will be removed after microcode_ctl + * is updated to ignore failure of this ioctl() + */ + case MICROCODE_IOCFREE: + return 0; + default: + return -EINVAL; + } + return -EINVAL; +} + +static struct file_operations microcode_fops = { + .owner = THIS_MODULE, + .write = microcode_write, + .ioctl = microcode_ioctl, + .open = microcode_open, +}; + +static struct miscdevice microcode_dev = { + .minor = MICROCODE_MINOR, + .name = "microcode", + .devfs_name = "cpu/microcode", + .fops = µcode_fops, +}; + +static int __init microcode_init (void) +{ + int error; + + error = misc_register(µcode_dev); + if (error) { + printk(KERN_ERR + "microcode: can't misc_register on minor=%d\n", + MICROCODE_MINOR); + return error; + } + + printk(KERN_INFO + "IA-32 Microcode Update Driver: v" MICROCODE_VERSION " <tigran@veritas.com>\n"); + return 0; +} + +static void __exit microcode_exit (void) +{ + misc_deregister(µcode_dev); + printk(KERN_INFO "IA-32 Microcode Update Driver v" MICROCODE_VERSION " unregistered\n"); +} + +module_init(microcode_init) +module_exit(microcode_exit) +MODULE_ALIAS_MISCDEV(MICROCODE_MINOR); diff --git a/arch/i386/kernel/module.c b/arch/i386/kernel/module.c new file mode 100644 index 00000000000..5149c8a621f --- /dev/null +++ b/arch/i386/kernel/module.c @@ -0,0 +1,129 @@ +/* Kernel module help for i386. + Copyright (C) 2001 Rusty Russell. + + This program is free software; you can redistribute it and/or modify + it under the terms of the GNU General Public License as published by + the Free Software Foundation; either version 2 of the License, or + (at your option) any later version. + + This program is distributed in the hope that it will be useful, + but WITHOUT ANY WARRANTY; without even the implied warranty of + MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + GNU General Public License for more details. + + You should have received a copy of the GNU General Public License + along with this program; if not, write to the Free Software + Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA +*/ +#include <linux/moduleloader.h> +#include <linux/elf.h> +#include <linux/vmalloc.h> +#include <linux/fs.h> +#include <linux/string.h> +#include <linux/kernel.h> + +#if 0 +#define DEBUGP printk +#else +#define DEBUGP(fmt...) +#endif + +void *module_alloc(unsigned long size) +{ + if (size == 0) + return NULL; + return vmalloc_exec(size); +} + + +/* Free memory returned from module_alloc */ +void module_free(struct module *mod, void *module_region) +{ + vfree(module_region); + /* FIXME: If module_region == mod->init_region, trim exception + table entries. */ +} + +/* We don't need anything special. */ +int module_frob_arch_sections(Elf_Ehdr *hdr, + Elf_Shdr *sechdrs, + char *secstrings, + struct module *mod) +{ + return 0; +} + +int apply_relocate(Elf32_Shdr *sechdrs, + const char *strtab, + unsigned int symindex, + unsigned int relsec, + struct module *me) +{ + unsigned int i; + Elf32_Rel *rel = (void *)sechdrs[relsec].sh_addr; + Elf32_Sym *sym; + uint32_t *location; + + DEBUGP("Applying relocate section %u to %u\n", relsec, + sechdrs[relsec].sh_info); + for (i = 0; i < sechdrs[relsec].sh_size / sizeof(*rel); i++) { + /* This is where to make the change */ + location = (void *)sechdrs[sechdrs[relsec].sh_info].sh_addr + + rel[i].r_offset; + /* This is the symbol it is referring to. Note that all + undefined symbols have been resolved. */ + sym = (Elf32_Sym *)sechdrs[symindex].sh_addr + + ELF32_R_SYM(rel[i].r_info); + + switch (ELF32_R_TYPE(rel[i].r_info)) { + case R_386_32: + /* We add the value into the location given */ + *location += sym->st_value; + break; + case R_386_PC32: + /* Add the value, subtract its postition */ + *location += sym->st_value - (uint32_t)location; + break; + default: + printk(KERN_ERR "module %s: Unknown relocation: %u\n", + me->name, ELF32_R_TYPE(rel[i].r_info)); + return -ENOEXEC; + } + } + return 0; +} + +int apply_relocate_add(Elf32_Shdr *sechdrs, + const char *strtab, + unsigned int symindex, + unsigned int relsec, + struct module *me) +{ + printk(KERN_ERR "module %s: ADD RELOCATION unsupported\n", + me->name); + return -ENOEXEC; +} + +extern void apply_alternatives(void *start, void *end); + +int module_finalize(const Elf_Ehdr *hdr, + const Elf_Shdr *sechdrs, + struct module *me) +{ + const Elf_Shdr *s; + char *secstrings = (void *)hdr + sechdrs[hdr->e_shstrndx].sh_offset; + + /* look for .altinstructions to patch */ + for (s = sechdrs; s < sechdrs + hdr->e_shnum; s++) { + void *seg; + if (strcmp(".altinstructions", secstrings + s->sh_name)) + continue; + seg = (void *)s->sh_addr; + apply_alternatives(seg, seg + s->sh_size); + } + return 0; +} + +void module_arch_cleanup(struct module *mod) +{ +} diff --git a/arch/i386/kernel/mpparse.c b/arch/i386/kernel/mpparse.c new file mode 100644 index 00000000000..1347ab4939e --- /dev/null +++ b/arch/i386/kernel/mpparse.c @@ -0,0 +1,1109 @@ +/* + * Intel Multiprocessor Specification 1.1 and 1.4 + * compliant MP-table parsing routines. + * + * (c) 1995 Alan Cox, Building #3 <alan@redhat.com> + * (c) 1998, 1999, 2000 Ingo Molnar <mingo@redhat.com> + * + * Fixes + * Erich Boleyn : MP v1.4 and additional changes. + * Alan Cox : Added EBDA scanning + * Ingo Molnar : various cleanups and rewrites + * Maciej W. Rozycki: Bits for default MP configurations + * Paul Diefenbaugh: Added full ACPI support + */ + +#include <linux/mm.h> +#include <linux/irq.h> +#include <linux/init.h> +#include <linux/acpi.h> +#include <linux/delay.h> +#include <linux/config.h> +#include <linux/bootmem.h> +#include <linux/smp_lock.h> +#include <linux/kernel_stat.h> +#include <linux/mc146818rtc.h> +#include <linux/bitops.h> + +#include <asm/smp.h> +#include <asm/acpi.h> +#include <asm/mtrr.h> +#include <asm/mpspec.h> +#include <asm/io_apic.h> + +#include <mach_apic.h> +#include <mach_mpparse.h> +#include <bios_ebda.h> + +/* Have we found an MP table */ +int smp_found_config; +unsigned int __initdata maxcpus = NR_CPUS; + +/* + * Various Linux-internal data structures created from the + * MP-table. + */ +int apic_version [MAX_APICS]; +int mp_bus_id_to_type [MAX_MP_BUSSES]; +int mp_bus_id_to_node [MAX_MP_BUSSES]; +int mp_bus_id_to_local [MAX_MP_BUSSES]; +int quad_local_to_mp_bus_id [NR_CPUS/4][4]; +int mp_bus_id_to_pci_bus [MAX_MP_BUSSES] = { [0 ... MAX_MP_BUSSES-1] = -1 }; +static int mp_current_pci_id; + +/* I/O APIC entries */ +struct mpc_config_ioapic mp_ioapics[MAX_IO_APICS]; + +/* # of MP IRQ source entries */ +struct mpc_config_intsrc mp_irqs[MAX_IRQ_SOURCES]; + +/* MP IRQ source entries */ +int mp_irq_entries; + +int nr_ioapics; + +int pic_mode; +unsigned long mp_lapic_addr; + +/* Processor that is doing the boot up */ +unsigned int boot_cpu_physical_apicid = -1U; +unsigned int boot_cpu_logical_apicid = -1U; +/* Internal processor count */ +static unsigned int __initdata num_processors; + +/* Bitmask of physically existing CPUs */ +physid_mask_t phys_cpu_present_map; + +u8 bios_cpu_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID }; + +/* + * Intel MP BIOS table parsing routines: + */ + + +/* + * Checksum an MP configuration block. + */ + +static int __init mpf_checksum(unsigned char *mp, int len) +{ + int sum = 0; + + while (len--) + sum += *mp++; + + return sum & 0xFF; +} + +/* + * Have to match translation table entries to main table entries by counter + * hence the mpc_record variable .... can't see a less disgusting way of + * doing this .... + */ + +static int mpc_record; +static struct mpc_config_translation *translation_table[MAX_MPC_ENTRY] __initdata; + +#ifdef CONFIG_X86_NUMAQ +static int MP_valid_apicid(int apicid, int version) +{ + return hweight_long(apicid & 0xf) == 1 && (apicid >> 4) != 0xf; +} +#else +static int MP_valid_apicid(int apicid, int version) +{ + if (version >= 0x14) + return apicid < 0xff; + else + return apicid < 0xf; +} +#endif + +static void __init MP_processor_info (struct mpc_config_processor *m) +{ + int ver, apicid; + physid_mask_t tmp; + + if (!(m->mpc_cpuflag & CPU_ENABLED)) + return; + + apicid = mpc_apic_id(m, translation_table[mpc_record]); + + if (m->mpc_featureflag&(1<<0)) + Dprintk(" Floating point unit present.\n"); + if (m->mpc_featureflag&(1<<7)) + Dprintk(" Machine Exception supported.\n"); + if (m->mpc_featureflag&(1<<8)) + Dprintk(" 64 bit compare & exchange supported.\n"); + if (m->mpc_featureflag&(1<<9)) + Dprintk(" Internal APIC present.\n"); + if (m->mpc_featureflag&(1<<11)) + Dprintk(" SEP present.\n"); + if (m->mpc_featureflag&(1<<12)) + Dprintk(" MTRR present.\n"); + if (m->mpc_featureflag&(1<<13)) + Dprintk(" PGE present.\n"); + if (m->mpc_featureflag&(1<<14)) + Dprintk(" MCA present.\n"); + if (m->mpc_featureflag&(1<<15)) + Dprintk(" CMOV present.\n"); + if (m->mpc_featureflag&(1<<16)) + Dprintk(" PAT present.\n"); + if (m->mpc_featureflag&(1<<17)) + Dprintk(" PSE present.\n"); + if (m->mpc_featureflag&(1<<18)) + Dprintk(" PSN present.\n"); + if (m->mpc_featureflag&(1<<19)) + Dprintk(" Cache Line Flush Instruction present.\n"); + /* 20 Reserved */ + if (m->mpc_featureflag&(1<<21)) + Dprintk(" Debug Trace and EMON Store present.\n"); + if (m->mpc_featureflag&(1<<22)) + Dprintk(" ACPI Thermal Throttle Registers present.\n"); + if (m->mpc_featureflag&(1<<23)) + Dprintk(" MMX present.\n"); + if (m->mpc_featureflag&(1<<24)) + Dprintk(" FXSR present.\n"); + if (m->mpc_featureflag&(1<<25)) + Dprintk(" XMM present.\n"); + if (m->mpc_featureflag&(1<<26)) + Dprintk(" Willamette New Instructions present.\n"); + if (m->mpc_featureflag&(1<<27)) + Dprintk(" Self Snoop present.\n"); + if (m->mpc_featureflag&(1<<28)) + Dprintk(" HT present.\n"); + if (m->mpc_featureflag&(1<<29)) + Dprintk(" Thermal Monitor present.\n"); + /* 30, 31 Reserved */ + + + if (m->mpc_cpuflag & CPU_BOOTPROCESSOR) { + Dprintk(" Bootup CPU\n"); + boot_cpu_physical_apicid = m->mpc_apicid; + boot_cpu_logical_apicid = apicid; + } + + if (num_processors >= NR_CPUS) { + printk(KERN_WARNING "WARNING: NR_CPUS limit of %i reached." + " Processor ignored.\n", NR_CPUS); + return; + } + + if (num_processors >= maxcpus) { + printk(KERN_WARNING "WARNING: maxcpus limit of %i reached." + " Processor ignored.\n", maxcpus); + return; + } + num_processors++; + ver = m->mpc_apicver; + + if (!MP_valid_apicid(apicid, ver)) { + printk(KERN_WARNING "Processor #%d INVALID. (Max ID: %d).\n", + m->mpc_apicid, MAX_APICS); + --num_processors; + return; + } + + tmp = apicid_to_cpu_present(apicid); + physids_or(phys_cpu_present_map, phys_cpu_present_map, tmp); + + /* + * Validate version + */ + if (ver == 0x0) { + printk(KERN_WARNING "BIOS bug, APIC version is 0 for CPU#%d! fixing up to 0x10. (tell your hw vendor)\n", m->mpc_apicid); + ver = 0x10; + } + apic_version[m->mpc_apicid] = ver; + bios_cpu_apicid[num_processors - 1] = m->mpc_apicid; +} + +static void __init MP_bus_info (struct mpc_config_bus *m) +{ + char str[7]; + + memcpy(str, m->mpc_bustype, 6); + str[6] = 0; + + mpc_oem_bus_info(m, str, translation_table[mpc_record]); + + if (strncmp(str, BUSTYPE_ISA, sizeof(BUSTYPE_ISA)-1) == 0) { + mp_bus_id_to_type[m->mpc_busid] = MP_BUS_ISA; + } else if (strncmp(str, BUSTYPE_EISA, sizeof(BUSTYPE_EISA)-1) == 0) { + mp_bus_id_to_type[m->mpc_busid] = MP_BUS_EISA; + } else if (strncmp(str, BUSTYPE_PCI, sizeof(BUSTYPE_PCI)-1) == 0) { + mpc_oem_pci_bus(m, translation_table[mpc_record]); + mp_bus_id_to_type[m->mpc_busid] = MP_BUS_PCI; + mp_bus_id_to_pci_bus[m->mpc_busid] = mp_current_pci_id; + mp_current_pci_id++; + } else if (strncmp(str, BUSTYPE_MCA, sizeof(BUSTYPE_MCA)-1) == 0) { + mp_bus_id_to_type[m->mpc_busid] = MP_BUS_MCA; + } else if (strncmp(str, BUSTYPE_NEC98, sizeof(BUSTYPE_NEC98)-1) == 0) { + mp_bus_id_to_type[m->mpc_busid] = MP_BUS_NEC98; + } else { + printk(KERN_WARNING "Unknown bustype %s - ignoring\n", str); + } +} + +static void __init MP_ioapic_info (struct mpc_config_ioapic *m) +{ + if (!(m->mpc_flags & MPC_APIC_USABLE)) + return; + + printk(KERN_INFO "I/O APIC #%d Version %d at 0x%lX.\n", + m->mpc_apicid, m->mpc_apicver, m->mpc_apicaddr); + if (nr_ioapics >= MAX_IO_APICS) { + printk(KERN_CRIT "Max # of I/O APICs (%d) exceeded (found %d).\n", + MAX_IO_APICS, nr_ioapics); + panic("Recompile kernel with bigger MAX_IO_APICS!.\n"); + } + if (!m->mpc_apicaddr) { + printk(KERN_ERR "WARNING: bogus zero I/O APIC address" + " found in MP table, skipping!\n"); + return; + } + mp_ioapics[nr_ioapics] = *m; + nr_ioapics++; +} + +static void __init MP_intsrc_info (struct mpc_config_intsrc *m) +{ + mp_irqs [mp_irq_entries] = *m; + Dprintk("Int: type %d, pol %d, trig %d, bus %d," + " IRQ %02x, APIC ID %x, APIC INT %02x\n", + m->mpc_irqtype, m->mpc_irqflag & 3, + (m->mpc_irqflag >> 2) & 3, m->mpc_srcbus, + m->mpc_srcbusirq, m->mpc_dstapic, m->mpc_dstirq); + if (++mp_irq_entries == MAX_IRQ_SOURCES) + panic("Max # of irq sources exceeded!!\n"); +} + +static void __init MP_lintsrc_info (struct mpc_config_lintsrc *m) +{ + Dprintk("Lint: type %d, pol %d, trig %d, bus %d," + " IRQ %02x, APIC ID %x, APIC LINT %02x\n", + m->mpc_irqtype, m->mpc_irqflag & 3, + (m->mpc_irqflag >> 2) &3, m->mpc_srcbusid, + m->mpc_srcbusirq, m->mpc_destapic, m->mpc_destapiclint); + /* + * Well it seems all SMP boards in existence + * use ExtINT/LVT1 == LINT0 and + * NMI/LVT2 == LINT1 - the following check + * will show us if this assumptions is false. + * Until then we do not have to add baggage. + */ + if ((m->mpc_irqtype == mp_ExtINT) && + (m->mpc_destapiclint != 0)) + BUG(); + if ((m->mpc_irqtype == mp_NMI) && + (m->mpc_destapiclint != 1)) + BUG(); +} + +#ifdef CONFIG_X86_NUMAQ +static void __init MP_translation_info (struct mpc_config_translation *m) +{ + printk(KERN_INFO "Translation: record %d, type %d, quad %d, global %d, local %d\n", mpc_record, m->trans_type, m->trans_quad, m->trans_global, m->trans_local); + + if (mpc_record >= MAX_MPC_ENTRY) + printk(KERN_ERR "MAX_MPC_ENTRY exceeded!\n"); + else + translation_table[mpc_record] = m; /* stash this for later */ + if (m->trans_quad < MAX_NUMNODES && !node_online(m->trans_quad)) + node_set_online(m->trans_quad); +} + +/* + * Read/parse the MPC oem tables + */ + +static void __init smp_read_mpc_oem(struct mp_config_oemtable *oemtable, \ + unsigned short oemsize) +{ + int count = sizeof (*oemtable); /* the header size */ + unsigned char *oemptr = ((unsigned char *)oemtable)+count; + + mpc_record = 0; + printk(KERN_INFO "Found an OEM MPC table at %8p - parsing it ... \n", oemtable); + if (memcmp(oemtable->oem_signature,MPC_OEM_SIGNATURE,4)) + { + printk(KERN_WARNING "SMP mpc oemtable: bad signature [%c%c%c%c]!\n", + oemtable->oem_signature[0], + oemtable->oem_signature[1], + oemtable->oem_signature[2], + oemtable->oem_signature[3]); + return; + } + if (mpf_checksum((unsigned char *)oemtable,oemtable->oem_length)) + { + printk(KERN_WARNING "SMP oem mptable: checksum error!\n"); + return; + } + while (count < oemtable->oem_length) { + switch (*oemptr) { + case MP_TRANSLATION: + { + struct mpc_config_translation *m= + (struct mpc_config_translation *)oemptr; + MP_translation_info(m); + oemptr += sizeof(*m); + count += sizeof(*m); + ++mpc_record; + break; + } + default: + { + printk(KERN_WARNING "Unrecognised OEM table entry type! - %d\n", (int) *oemptr); + return; + } + } + } +} + +static inline void mps_oem_check(struct mp_config_table *mpc, char *oem, + char *productid) +{ + if (strncmp(oem, "IBM NUMA", 8)) + printk("Warning! May not be a NUMA-Q system!\n"); + if (mpc->mpc_oemptr) + smp_read_mpc_oem((struct mp_config_oemtable *) mpc->mpc_oemptr, + mpc->mpc_oemsize); +} +#endif /* CONFIG_X86_NUMAQ */ + +/* + * Read/parse the MPC + */ + +static int __init smp_read_mpc(struct mp_config_table *mpc) +{ + char str[16]; + char oem[10]; + int count=sizeof(*mpc); + unsigned char *mpt=((unsigned char *)mpc)+count; + + if (memcmp(mpc->mpc_signature,MPC_SIGNATURE,4)) { + printk(KERN_ERR "SMP mptable: bad signature [0x%x]!\n", + *(u32 *)mpc->mpc_signature); + return 0; + } + if (mpf_checksum((unsigned char *)mpc,mpc->mpc_length)) { + printk(KERN_ERR "SMP mptable: checksum error!\n"); + return 0; + } + if (mpc->mpc_spec!=0x01 && mpc->mpc_spec!=0x04) { + printk(KERN_ERR "SMP mptable: bad table version (%d)!!\n", + mpc->mpc_spec); + return 0; + } + if (!mpc->mpc_lapic) { + printk(KERN_ERR "SMP mptable: null local APIC address!\n"); + return 0; + } + memcpy(oem,mpc->mpc_oem,8); + oem[8]=0; + printk(KERN_INFO "OEM ID: %s ",oem); + + memcpy(str,mpc->mpc_productid,12); + str[12]=0; + printk("Product ID: %s ",str); + + mps_oem_check(mpc, oem, str); + + printk("APIC at: 0x%lX\n",mpc->mpc_lapic); + + /* + * Save the local APIC address (it might be non-default) -- but only + * if we're not using ACPI. + */ + if (!acpi_lapic) + mp_lapic_addr = mpc->mpc_lapic; + + /* + * Now process the configuration blocks. + */ + mpc_record = 0; + while (count < mpc->mpc_length) { + switch(*mpt) { + case MP_PROCESSOR: + { + struct mpc_config_processor *m= + (struct mpc_config_processor *)mpt; + /* ACPI may have already provided this data */ + if (!acpi_lapic) + MP_processor_info(m); + mpt += sizeof(*m); + count += sizeof(*m); + break; + } + case MP_BUS: + { + struct mpc_config_bus *m= + (struct mpc_config_bus *)mpt; + MP_bus_info(m); + mpt += sizeof(*m); + count += sizeof(*m); + break; + } + case MP_IOAPIC: + { + struct mpc_config_ioapic *m= + (struct mpc_config_ioapic *)mpt; + MP_ioapic_info(m); + mpt+=sizeof(*m); + count+=sizeof(*m); + break; + } + case MP_INTSRC: + { + struct mpc_config_intsrc *m= + (struct mpc_config_intsrc *)mpt; + + MP_intsrc_info(m); + mpt+=sizeof(*m); + count+=sizeof(*m); + break; + } + case MP_LINTSRC: + { + struct mpc_config_lintsrc *m= + (struct mpc_config_lintsrc *)mpt; + MP_lintsrc_info(m); + mpt+=sizeof(*m); + count+=sizeof(*m); + break; + } + default: + { + count = mpc->mpc_length; + break; + } + } + ++mpc_record; + } + clustered_apic_check(); + if (!num_processors) + printk(KERN_ERR "SMP mptable: no processors registered!\n"); + return num_processors; +} + +static int __init ELCR_trigger(unsigned int irq) +{ + unsigned int port; + + port = 0x4d0 + (irq >> 3); + return (inb(port) >> (irq & 7)) & 1; +} + +static void __init construct_default_ioirq_mptable(int mpc_default_type) +{ + struct mpc_config_intsrc intsrc; + int i; + int ELCR_fallback = 0; + + intsrc.mpc_type = MP_INTSRC; + intsrc.mpc_irqflag = 0; /* conforming */ + intsrc.mpc_srcbus = 0; + intsrc.mpc_dstapic = mp_ioapics[0].mpc_apicid; + + intsrc.mpc_irqtype = mp_INT; + + /* + * If true, we have an ISA/PCI system with no IRQ entries + * in the MP table. To prevent the PCI interrupts from being set up + * incorrectly, we try to use the ELCR. The sanity check to see if + * there is good ELCR data is very simple - IRQ0, 1, 2 and 13 can + * never be level sensitive, so we simply see if the ELCR agrees. + * If it does, we assume it's valid. + */ + if (mpc_default_type == 5) { + printk(KERN_INFO "ISA/PCI bus type with no IRQ information... falling back to ELCR\n"); + + if (ELCR_trigger(0) || ELCR_trigger(1) || ELCR_trigger(2) || ELCR_trigger(13)) + printk(KERN_WARNING "ELCR contains invalid data... not using ELCR\n"); + else { + printk(KERN_INFO "Using ELCR to identify PCI interrupts\n"); + ELCR_fallback = 1; + } + } + + for (i = 0; i < 16; i++) { + switch (mpc_default_type) { + case 2: + if (i == 0 || i == 13) + continue; /* IRQ0 & IRQ13 not connected */ + /* fall through */ + default: + if (i == 2) + continue; /* IRQ2 is never connected */ + } + + if (ELCR_fallback) { + /* + * If the ELCR indicates a level-sensitive interrupt, we + * copy that information over to the MP table in the + * irqflag field (level sensitive, active high polarity). + */ + if (ELCR_trigger(i)) + intsrc.mpc_irqflag = 13; + else + intsrc.mpc_irqflag = 0; + } + + intsrc.mpc_srcbusirq = i; + intsrc.mpc_dstirq = i ? i : 2; /* IRQ0 to INTIN2 */ + MP_intsrc_info(&intsrc); + } + + intsrc.mpc_irqtype = mp_ExtINT; + intsrc.mpc_srcbusirq = 0; + intsrc.mpc_dstirq = 0; /* 8259A to INTIN0 */ + MP_intsrc_info(&intsrc); +} + +static inline void __init construct_default_ISA_mptable(int mpc_default_type) +{ + struct mpc_config_processor processor; + struct mpc_config_bus bus; + struct mpc_config_ioapic ioapic; + struct mpc_config_lintsrc lintsrc; + int linttypes[2] = { mp_ExtINT, mp_NMI }; + int i; + + /* + * local APIC has default address + */ + mp_lapic_addr = APIC_DEFAULT_PHYS_BASE; + + /* + * 2 CPUs, numbered 0 & 1. + */ + processor.mpc_type = MP_PROCESSOR; + /* Either an integrated APIC or a discrete 82489DX. */ + processor.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01; + processor.mpc_cpuflag = CPU_ENABLED; + processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) | + (boot_cpu_data.x86_model << 4) | + boot_cpu_data.x86_mask; + processor.mpc_featureflag = boot_cpu_data.x86_capability[0]; + processor.mpc_reserved[0] = 0; + processor.mpc_reserved[1] = 0; + for (i = 0; i < 2; i++) { + processor.mpc_apicid = i; + MP_processor_info(&processor); + } + + bus.mpc_type = MP_BUS; + bus.mpc_busid = 0; + switch (mpc_default_type) { + default: + printk("???\n"); + printk(KERN_ERR "Unknown standard configuration %d\n", + mpc_default_type); + /* fall through */ + case 1: + case 5: + memcpy(bus.mpc_bustype, "ISA ", 6); + break; + case 2: + case 6: + case 3: + memcpy(bus.mpc_bustype, "EISA ", 6); + break; + case 4: + case 7: + memcpy(bus.mpc_bustype, "MCA ", 6); + } + MP_bus_info(&bus); + if (mpc_default_type > 4) { + bus.mpc_busid = 1; + memcpy(bus.mpc_bustype, "PCI ", 6); + MP_bus_info(&bus); + } + + ioapic.mpc_type = MP_IOAPIC; + ioapic.mpc_apicid = 2; + ioapic.mpc_apicver = mpc_default_type > 4 ? 0x10 : 0x01; + ioapic.mpc_flags = MPC_APIC_USABLE; + ioapic.mpc_apicaddr = 0xFEC00000; + MP_ioapic_info(&ioapic); + + /* + * We set up most of the low 16 IO-APIC pins according to MPS rules. + */ + construct_default_ioirq_mptable(mpc_default_type); + + lintsrc.mpc_type = MP_LINTSRC; + lintsrc.mpc_irqflag = 0; /* conforming */ + lintsrc.mpc_srcbusid = 0; + lintsrc.mpc_srcbusirq = 0; + lintsrc.mpc_destapic = MP_APIC_ALL; + for (i = 0; i < 2; i++) { + lintsrc.mpc_irqtype = linttypes[i]; + lintsrc.mpc_destapiclint = i; + MP_lintsrc_info(&lintsrc); + } +} + +static struct intel_mp_floating *mpf_found; + +/* + * Scan the memory blocks for an SMP configuration block. + */ +void __init get_smp_config (void) +{ + struct intel_mp_floating *mpf = mpf_found; + + /* + * ACPI may be used to obtain the entire SMP configuration or just to + * enumerate/configure processors (CONFIG_ACPI_BOOT). Note that + * ACPI supports both logical (e.g. Hyper-Threading) and physical + * processors, where MPS only supports physical. + */ + if (acpi_lapic && acpi_ioapic) { + printk(KERN_INFO "Using ACPI (MADT) for SMP configuration information\n"); + return; + } + else if (acpi_lapic) + printk(KERN_INFO "Using ACPI for processor (LAPIC) configuration information\n"); + + printk(KERN_INFO "Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification); + if (mpf->mpf_feature2 & (1<<7)) { + printk(KERN_INFO " IMCR and PIC compatibility mode.\n"); + pic_mode = 1; + } else { + printk(KERN_INFO " Virtual Wire compatibility mode.\n"); + pic_mode = 0; + } + + /* + * Now see if we need to read further. + */ + if (mpf->mpf_feature1 != 0) { + + printk(KERN_INFO "Default MP configuration #%d\n", mpf->mpf_feature1); + construct_default_ISA_mptable(mpf->mpf_feature1); + + } else if (mpf->mpf_physptr) { + + /* + * Read the physical hardware table. Anything here will + * override the defaults. + */ + if (!smp_read_mpc((void *)mpf->mpf_physptr)) { + smp_found_config = 0; + printk(KERN_ERR "BIOS bug, MP table errors detected!...\n"); + printk(KERN_ERR "... disabling SMP support. (tell your hw vendor)\n"); + return; + } + /* + * If there are no explicit MP IRQ entries, then we are + * broken. We set up most of the low 16 IO-APIC pins to + * ISA defaults and hope it will work. + */ + if (!mp_irq_entries) { + struct mpc_config_bus bus; + + printk(KERN_ERR "BIOS bug, no explicit IRQ entries, using default mptable. (tell your hw vendor)\n"); + + bus.mpc_type = MP_BUS; + bus.mpc_busid = 0; + memcpy(bus.mpc_bustype, "ISA ", 6); + MP_bus_info(&bus); + + construct_default_ioirq_mptable(0); + } + + } else + BUG(); + + printk(KERN_INFO "Processors: %d\n", num_processors); + /* + * Only use the first configuration found. + */ +} + +static int __init smp_scan_config (unsigned long base, unsigned long length) +{ + unsigned long *bp = phys_to_virt(base); + struct intel_mp_floating *mpf; + + Dprintk("Scan SMP from %p for %ld bytes.\n", bp,length); + if (sizeof(*mpf) != 16) + printk("Error: MPF size\n"); + + while (length > 0) { + mpf = (struct intel_mp_floating *)bp; + if ((*bp == SMP_MAGIC_IDENT) && + (mpf->mpf_length == 1) && + !mpf_checksum((unsigned char *)bp, 16) && + ((mpf->mpf_specification == 1) + || (mpf->mpf_specification == 4)) ) { + + smp_found_config = 1; + printk(KERN_INFO "found SMP MP-table at %08lx\n", + virt_to_phys(mpf)); + reserve_bootmem(virt_to_phys(mpf), PAGE_SIZE); + if (mpf->mpf_physptr) { + /* + * We cannot access to MPC table to compute + * table size yet, as only few megabytes from + * the bottom is mapped now. + * PC-9800's MPC table places on the very last + * of physical memory; so that simply reserving + * PAGE_SIZE from mpg->mpf_physptr yields BUG() + * in reserve_bootmem. + */ + unsigned long size = PAGE_SIZE; + unsigned long end = max_low_pfn * PAGE_SIZE; + if (mpf->mpf_physptr + size > end) + size = end - mpf->mpf_physptr; + reserve_bootmem(mpf->mpf_physptr, size); + } + + mpf_found = mpf; + return 1; + } + bp += 4; + length -= 16; + } + return 0; +} + +void __init find_smp_config (void) +{ + unsigned int address; + + /* + * FIXME: Linux assumes you have 640K of base ram.. + * this continues the error... + * + * 1) Scan the bottom 1K for a signature + * 2) Scan the top 1K of base RAM + * 3) Scan the 64K of bios + */ + if (smp_scan_config(0x0,0x400) || + smp_scan_config(639*0x400,0x400) || + smp_scan_config(0xF0000,0x10000)) + return; + /* + * If it is an SMP machine we should know now, unless the + * configuration is in an EISA/MCA bus machine with an + * extended bios data area. + * + * there is a real-mode segmented pointer pointing to the + * 4K EBDA area at 0x40E, calculate and scan it here. + * + * NOTE! There are Linux loaders that will corrupt the EBDA + * area, and as such this kind of SMP config may be less + * trustworthy, simply because the SMP table may have been + * stomped on during early boot. These loaders are buggy and + * should be fixed. + * + * MP1.4 SPEC states to only scan first 1K of 4K EBDA. + */ + + address = get_bios_ebda(); + if (address) + smp_scan_config(address, 0x400); +} + +/* -------------------------------------------------------------------------- + ACPI-based MP Configuration + -------------------------------------------------------------------------- */ + +#ifdef CONFIG_ACPI_BOOT + +void __init mp_register_lapic_address ( + u64 address) +{ + mp_lapic_addr = (unsigned long) address; + + set_fixmap_nocache(FIX_APIC_BASE, mp_lapic_addr); + + if (boot_cpu_physical_apicid == -1U) + boot_cpu_physical_apicid = GET_APIC_ID(apic_read(APIC_ID)); + + Dprintk("Boot CPU = %d\n", boot_cpu_physical_apicid); +} + + +void __init mp_register_lapic ( + u8 id, + u8 enabled) +{ + struct mpc_config_processor processor; + int boot_cpu = 0; + + if (MAX_APICS - id <= 0) { + printk(KERN_WARNING "Processor #%d invalid (max %d)\n", + id, MAX_APICS); + return; + } + + if (id == boot_cpu_physical_apicid) + boot_cpu = 1; + + processor.mpc_type = MP_PROCESSOR; + processor.mpc_apicid = id; + processor.mpc_apicver = GET_APIC_VERSION(apic_read(APIC_LVR)); + processor.mpc_cpuflag = (enabled ? CPU_ENABLED : 0); + processor.mpc_cpuflag |= (boot_cpu ? CPU_BOOTPROCESSOR : 0); + processor.mpc_cpufeature = (boot_cpu_data.x86 << 8) | + (boot_cpu_data.x86_model << 4) | boot_cpu_data.x86_mask; + processor.mpc_featureflag = boot_cpu_data.x86_capability[0]; + processor.mpc_reserved[0] = 0; + processor.mpc_reserved[1] = 0; + + MP_processor_info(&processor); +} + +#if defined(CONFIG_X86_IO_APIC) && (defined(CONFIG_ACPI_INTERPRETER) || defined(CONFIG_ACPI_BOOT)) + +#define MP_ISA_BUS 0 +#define MP_MAX_IOAPIC_PIN 127 + +static struct mp_ioapic_routing { + int apic_id; + int gsi_base; + int gsi_end; + u32 pin_programmed[4]; +} mp_ioapic_routing[MAX_IO_APICS]; + + +static int mp_find_ioapic ( + int gsi) +{ + int i = 0; + + /* Find the IOAPIC that manages this GSI. */ + for (i = 0; i < nr_ioapics; i++) { + if ((gsi >= mp_ioapic_routing[i].gsi_base) + && (gsi <= mp_ioapic_routing[i].gsi_end)) + return i; + } + + printk(KERN_ERR "ERROR: Unable to locate IOAPIC for GSI %d\n", gsi); + + return -1; +} + + +void __init mp_register_ioapic ( + u8 id, + u32 address, + u32 gsi_base) +{ + int idx = 0; + + if (nr_ioapics >= MAX_IO_APICS) { + printk(KERN_ERR "ERROR: Max # of I/O APICs (%d) exceeded " + "(found %d)\n", MAX_IO_APICS, nr_ioapics); + panic("Recompile kernel with bigger MAX_IO_APICS!\n"); + } + if (!address) { + printk(KERN_ERR "WARNING: Bogus (zero) I/O APIC address" + " found in MADT table, skipping!\n"); + return; + } + + idx = nr_ioapics++; + + mp_ioapics[idx].mpc_type = MP_IOAPIC; + mp_ioapics[idx].mpc_flags = MPC_APIC_USABLE; + mp_ioapics[idx].mpc_apicaddr = address; + + set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address); + mp_ioapics[idx].mpc_apicid = io_apic_get_unique_id(idx, id); + mp_ioapics[idx].mpc_apicver = io_apic_get_version(idx); + + /* + * Build basic GSI lookup table to facilitate gsi->io_apic lookups + * and to prevent reprogramming of IOAPIC pins (PCI GSIs). + */ + mp_ioapic_routing[idx].apic_id = mp_ioapics[idx].mpc_apicid; + mp_ioapic_routing[idx].gsi_base = gsi_base; + mp_ioapic_routing[idx].gsi_end = gsi_base + + io_apic_get_redir_entries(idx); + + printk("IOAPIC[%d]: apic_id %d, version %d, address 0x%lx, " + "GSI %d-%d\n", idx, mp_ioapics[idx].mpc_apicid, + mp_ioapics[idx].mpc_apicver, mp_ioapics[idx].mpc_apicaddr, + mp_ioapic_routing[idx].gsi_base, + mp_ioapic_routing[idx].gsi_end); + + return; +} + + +void __init mp_override_legacy_irq ( + u8 bus_irq, + u8 polarity, + u8 trigger, + u32 gsi) +{ + struct mpc_config_intsrc intsrc; + int ioapic = -1; + int pin = -1; + + /* + * Convert 'gsi' to 'ioapic.pin'. + */ + ioapic = mp_find_ioapic(gsi); + if (ioapic < 0) + return; + pin = gsi - mp_ioapic_routing[ioapic].gsi_base; + + /* + * TBD: This check is for faulty timer entries, where the override + * erroneously sets the trigger to level, resulting in a HUGE + * increase of timer interrupts! + */ + if ((bus_irq == 0) && (trigger == 3)) + trigger = 1; + + intsrc.mpc_type = MP_INTSRC; + intsrc.mpc_irqtype = mp_INT; + intsrc.mpc_irqflag = (trigger << 2) | polarity; + intsrc.mpc_srcbus = MP_ISA_BUS; + intsrc.mpc_srcbusirq = bus_irq; /* IRQ */ + intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid; /* APIC ID */ + intsrc.mpc_dstirq = pin; /* INTIN# */ + + Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, %d-%d\n", + intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, + (intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, + intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, intsrc.mpc_dstirq); + + mp_irqs[mp_irq_entries] = intsrc; + if (++mp_irq_entries == MAX_IRQ_SOURCES) + panic("Max # of irq sources exceeded!\n"); + + return; +} + +int es7000_plat; + +void __init mp_config_acpi_legacy_irqs (void) +{ + struct mpc_config_intsrc intsrc; + int i = 0; + int ioapic = -1; + + /* + * Fabricate the legacy ISA bus (bus #31). + */ + mp_bus_id_to_type[MP_ISA_BUS] = MP_BUS_ISA; + Dprintk("Bus #%d is ISA\n", MP_ISA_BUS); + + /* + * Older generations of ES7000 have no legacy identity mappings + */ + if (es7000_plat == 1) + return; + + /* + * Locate the IOAPIC that manages the ISA IRQs (0-15). + */ + ioapic = mp_find_ioapic(0); + if (ioapic < 0) + return; + + intsrc.mpc_type = MP_INTSRC; + intsrc.mpc_irqflag = 0; /* Conforming */ + intsrc.mpc_srcbus = MP_ISA_BUS; + intsrc.mpc_dstapic = mp_ioapics[ioapic].mpc_apicid; + + /* + * Use the default configuration for the IRQs 0-15. Unless + * overriden by (MADT) interrupt source override entries. + */ + for (i = 0; i < 16; i++) { + int idx; + + for (idx = 0; idx < mp_irq_entries; idx++) { + struct mpc_config_intsrc *irq = mp_irqs + idx; + + /* Do we already have a mapping for this ISA IRQ? */ + if (irq->mpc_srcbus == MP_ISA_BUS && irq->mpc_srcbusirq == i) + break; + + /* Do we already have a mapping for this IOAPIC pin */ + if ((irq->mpc_dstapic == intsrc.mpc_dstapic) && + (irq->mpc_dstirq == i)) + break; + } + + if (idx != mp_irq_entries) { + printk(KERN_DEBUG "ACPI: IRQ%d used by override.\n", i); + continue; /* IRQ already used */ + } + + intsrc.mpc_irqtype = mp_INT; + intsrc.mpc_srcbusirq = i; /* Identity mapped */ + intsrc.mpc_dstirq = i; + + Dprintk("Int: type %d, pol %d, trig %d, bus %d, irq %d, " + "%d-%d\n", intsrc.mpc_irqtype, intsrc.mpc_irqflag & 3, + (intsrc.mpc_irqflag >> 2) & 3, intsrc.mpc_srcbus, + intsrc.mpc_srcbusirq, intsrc.mpc_dstapic, + intsrc.mpc_dstirq); + + mp_irqs[mp_irq_entries] = intsrc; + if (++mp_irq_entries == MAX_IRQ_SOURCES) + panic("Max # of irq sources exceeded!\n"); + } +} + +int mp_register_gsi (u32 gsi, int edge_level, int active_high_low) +{ + int ioapic = -1; + int ioapic_pin = 0; + int idx, bit = 0; + +#ifdef CONFIG_ACPI_BUS + /* Don't set up the ACPI SCI because it's already set up */ + if (acpi_fadt.sci_int == gsi) + return gsi; +#endif + + ioapic = mp_find_ioapic(gsi); + if (ioapic < 0) { + printk(KERN_WARNING "No IOAPIC for GSI %u\n", gsi); + return gsi; + } + + ioapic_pin = gsi - mp_ioapic_routing[ioapic].gsi_base; + + if (ioapic_renumber_irq) + gsi = ioapic_renumber_irq(ioapic, gsi); + + /* + * Avoid pin reprogramming. PRTs typically include entries + * with redundant pin->gsi mappings (but unique PCI devices); + * we only program the IOAPIC on the first. + */ + bit = ioapic_pin % 32; + idx = (ioapic_pin < 32) ? 0 : (ioapic_pin / 32); + if (idx > 3) { + printk(KERN_ERR "Invalid reference to IOAPIC pin " + "%d-%d\n", mp_ioapic_routing[ioapic].apic_id, + ioapic_pin); + return gsi; + } + if ((1<<bit) & mp_ioapic_routing[ioapic].pin_programmed[idx]) { + Dprintk(KERN_DEBUG "Pin %d-%d already programmed\n", + mp_ioapic_routing[ioapic].apic_id, ioapic_pin); + return gsi; + } + + mp_ioapic_routing[ioapic].pin_programmed[idx] |= (1<<bit); + + io_apic_set_pci_routing(ioapic, ioapic_pin, gsi, + edge_level == ACPI_EDGE_SENSITIVE ? 0 : 1, + active_high_low == ACPI_ACTIVE_HIGH ? 0 : 1); + return gsi; +} + +#endif /*CONFIG_X86_IO_APIC && (CONFIG_ACPI_INTERPRETER || CONFIG_ACPI_BOOT)*/ +#endif /*CONFIG_ACPI_BOOT*/ diff --git a/arch/i386/kernel/msr.c b/arch/i386/kernel/msr.c new file mode 100644 index 00000000000..05d9f8f363a --- /dev/null +++ b/arch/i386/kernel/msr.c @@ -0,0 +1,346 @@ +/* ----------------------------------------------------------------------- * + * + * Copyright 2000 H. Peter Anvin - All Rights Reserved + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation, Inc., 675 Mass Ave, Cambridge MA 02139, + * USA; either version 2 of the License, or (at your option) any later + * version; incorporated herein by reference. + * + * ----------------------------------------------------------------------- */ + +/* + * msr.c + * + * x86 MSR access device + * + * This device is accessed by lseek() to the appropriate register number + * and then read/write in chunks of 8 bytes. A larger size means multiple + * reads or writes of the same register. + * + * This driver uses /dev/cpu/%d/msr where %d is the minor number, and on + * an SMP box will direct the access to CPU %d. + */ + +#include <linux/module.h> +#include <linux/config.h> + +#include <linux/types.h> +#include <linux/errno.h> +#include <linux/fcntl.h> +#include <linux/init.h> +#include <linux/poll.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/major.h> +#include <linux/fs.h> +#include <linux/device.h> +#include <linux/cpu.h> +#include <linux/notifier.h> + +#include <asm/processor.h> +#include <asm/msr.h> +#include <asm/uaccess.h> +#include <asm/system.h> + +static struct class_simple *msr_class; + +/* Note: "err" is handled in a funny way below. Otherwise one version + of gcc or another breaks. */ + +static inline int wrmsr_eio(u32 reg, u32 eax, u32 edx) +{ + int err; + + asm volatile ("1: wrmsr\n" + "2:\n" + ".section .fixup,\"ax\"\n" + "3: movl %4,%0\n" + " jmp 2b\n" + ".previous\n" + ".section __ex_table,\"a\"\n" + " .align 4\n" " .long 1b,3b\n" ".previous":"=&bDS" (err) + :"a"(eax), "d"(edx), "c"(reg), "i"(-EIO), "0"(0)); + + return err; +} + +static inline int rdmsr_eio(u32 reg, u32 *eax, u32 *edx) +{ + int err; + + asm volatile ("1: rdmsr\n" + "2:\n" + ".section .fixup,\"ax\"\n" + "3: movl %4,%0\n" + " jmp 2b\n" + ".previous\n" + ".section __ex_table,\"a\"\n" + " .align 4\n" + " .long 1b,3b\n" + ".previous":"=&bDS" (err), "=a"(*eax), "=d"(*edx) + :"c"(reg), "i"(-EIO), "0"(0)); + + return err; +} + +#ifdef CONFIG_SMP + +struct msr_command { + int cpu; + int err; + u32 reg; + u32 data[2]; +}; + +static void msr_smp_wrmsr(void *cmd_block) +{ + struct msr_command *cmd = (struct msr_command *)cmd_block; + + if (cmd->cpu == smp_processor_id()) + cmd->err = wrmsr_eio(cmd->reg, cmd->data[0], cmd->data[1]); +} + +static void msr_smp_rdmsr(void *cmd_block) +{ + struct msr_command *cmd = (struct msr_command *)cmd_block; + + if (cmd->cpu == smp_processor_id()) + cmd->err = rdmsr_eio(cmd->reg, &cmd->data[0], &cmd->data[1]); +} + +static inline int do_wrmsr(int cpu, u32 reg, u32 eax, u32 edx) +{ + struct msr_command cmd; + int ret; + + preempt_disable(); + if (cpu == smp_processor_id()) { + ret = wrmsr_eio(reg, eax, edx); + } else { + cmd.cpu = cpu; + cmd.reg = reg; + cmd.data[0] = eax; + cmd.data[1] = edx; + + smp_call_function(msr_smp_wrmsr, &cmd, 1, 1); + ret = cmd.err; + } + preempt_enable(); + return ret; +} + +static inline int do_rdmsr(int cpu, u32 reg, u32 * eax, u32 * edx) +{ + struct msr_command cmd; + int ret; + + preempt_disable(); + if (cpu == smp_processor_id()) { + ret = rdmsr_eio(reg, eax, edx); + } else { + cmd.cpu = cpu; + cmd.reg = reg; + + smp_call_function(msr_smp_rdmsr, &cmd, 1, 1); + + *eax = cmd.data[0]; + *edx = cmd.data[1]; + + ret = cmd.err; + } + preempt_enable(); + return ret; +} + +#else /* ! CONFIG_SMP */ + +static inline int do_wrmsr(int cpu, u32 reg, u32 eax, u32 edx) +{ + return wrmsr_eio(reg, eax, edx); +} + +static inline int do_rdmsr(int cpu, u32 reg, u32 *eax, u32 *edx) +{ + return rdmsr_eio(reg, eax, edx); +} + +#endif /* ! CONFIG_SMP */ + +static loff_t msr_seek(struct file *file, loff_t offset, int orig) +{ + loff_t ret = -EINVAL; + + lock_kernel(); + switch (orig) { + case 0: + file->f_pos = offset; + ret = file->f_pos; + break; + case 1: + file->f_pos += offset; + ret = file->f_pos; + } + unlock_kernel(); + return ret; +} + +static ssize_t msr_read(struct file *file, char __user * buf, + size_t count, loff_t * ppos) +{ + u32 __user *tmp = (u32 __user *) buf; + u32 data[2]; + size_t rv; + u32 reg = *ppos; + int cpu = iminor(file->f_dentry->d_inode); + int err; + + if (count % 8) + return -EINVAL; /* Invalid chunk size */ + + for (rv = 0; count; count -= 8) { + err = do_rdmsr(cpu, reg, &data[0], &data[1]); + if (err) + return err; + if (copy_to_user(tmp, &data, 8)) + return -EFAULT; + tmp += 2; + } + + return ((char __user *)tmp) - buf; +} + +static ssize_t msr_write(struct file *file, const char __user *buf, + size_t count, loff_t *ppos) +{ + const u32 __user *tmp = (const u32 __user *)buf; + u32 data[2]; + size_t rv; + u32 reg = *ppos; + int cpu = iminor(file->f_dentry->d_inode); + int err; + + if (count % 8) + return -EINVAL; /* Invalid chunk size */ + + for (rv = 0; count; count -= 8) { + if (copy_from_user(&data, tmp, 8)) + return -EFAULT; + err = do_wrmsr(cpu, reg, data[0], data[1]); + if (err) + return err; + tmp += 2; + } + + return ((char __user *)tmp) - buf; +} + +static int msr_open(struct inode *inode, struct file *file) +{ + unsigned int cpu = iminor(file->f_dentry->d_inode); + struct cpuinfo_x86 *c = &(cpu_data)[cpu]; + + if (cpu >= NR_CPUS || !cpu_online(cpu)) + return -ENXIO; /* No such CPU */ + if (!cpu_has(c, X86_FEATURE_MSR)) + return -EIO; /* MSR not supported */ + + return 0; +} + +/* + * File operations we support + */ +static struct file_operations msr_fops = { + .owner = THIS_MODULE, + .llseek = msr_seek, + .read = msr_read, + .write = msr_write, + .open = msr_open, +}; + +static int msr_class_simple_device_add(int i) +{ + int err = 0; + struct class_device *class_err; + + class_err = class_simple_device_add(msr_class, MKDEV(MSR_MAJOR, i), NULL, "msr%d",i); + if (IS_ERR(class_err)) + err = PTR_ERR(class_err); + return err; +} + +static int __devinit msr_class_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) +{ + unsigned int cpu = (unsigned long)hcpu; + + switch (action) { + case CPU_ONLINE: + msr_class_simple_device_add(cpu); + break; + case CPU_DEAD: + class_simple_device_remove(MKDEV(MSR_MAJOR, cpu)); + break; + } + return NOTIFY_OK; +} + +static struct notifier_block msr_class_cpu_notifier = +{ + .notifier_call = msr_class_cpu_callback, +}; + +static int __init msr_init(void) +{ + int i, err = 0; + i = 0; + + if (register_chrdev(MSR_MAJOR, "cpu/msr", &msr_fops)) { + printk(KERN_ERR "msr: unable to get major %d for msr\n", + MSR_MAJOR); + err = -EBUSY; + goto out; + } + msr_class = class_simple_create(THIS_MODULE, "msr"); + if (IS_ERR(msr_class)) { + err = PTR_ERR(msr_class); + goto out_chrdev; + } + for_each_online_cpu(i) { + err = msr_class_simple_device_add(i); + if (err != 0) + goto out_class; + } + register_cpu_notifier(&msr_class_cpu_notifier); + + err = 0; + goto out; + +out_class: + i = 0; + for_each_online_cpu(i) + class_simple_device_remove(MKDEV(MSR_MAJOR, i)); + class_simple_destroy(msr_class); +out_chrdev: + unregister_chrdev(MSR_MAJOR, "cpu/msr"); +out: + return err; +} + +static void __exit msr_exit(void) +{ + int cpu = 0; + for_each_online_cpu(cpu) + class_simple_device_remove(MKDEV(MSR_MAJOR, cpu)); + class_simple_destroy(msr_class); + unregister_chrdev(MSR_MAJOR, "cpu/msr"); + unregister_cpu_notifier(&msr_class_cpu_notifier); +} + +module_init(msr_init); +module_exit(msr_exit) + +MODULE_AUTHOR("H. Peter Anvin <hpa@zytor.com>"); +MODULE_DESCRIPTION("x86 generic MSR driver"); +MODULE_LICENSE("GPL"); diff --git a/arch/i386/kernel/nmi.c b/arch/i386/kernel/nmi.c new file mode 100644 index 00000000000..f5b0c5081bd --- /dev/null +++ b/arch/i386/kernel/nmi.c @@ -0,0 +1,570 @@ +/* + * linux/arch/i386/nmi.c + * + * NMI watchdog support on APIC systems + * + * Started by Ingo Molnar <mingo@redhat.com> + * + * Fixes: + * Mikael Pettersson : AMD K7 support for local APIC NMI watchdog. + * Mikael Pettersson : Power Management for local APIC NMI watchdog. + * Mikael Pettersson : Pentium 4 support for local APIC NMI watchdog. + * Pavel Machek and + * Mikael Pettersson : PM converted to driver model. Disable/enable API. + */ + +#include <linux/config.h> +#include <linux/mm.h> +#include <linux/irq.h> +#include <linux/delay.h> +#include <linux/bootmem.h> +#include <linux/smp_lock.h> +#include <linux/interrupt.h> +#include <linux/mc146818rtc.h> +#include <linux/kernel_stat.h> +#include <linux/module.h> +#include <linux/nmi.h> +#include <linux/sysdev.h> +#include <linux/sysctl.h> + +#include <asm/smp.h> +#include <asm/mtrr.h> +#include <asm/mpspec.h> +#include <asm/nmi.h> + +#include "mach_traps.h" + +unsigned int nmi_watchdog = NMI_NONE; +extern int unknown_nmi_panic; +static unsigned int nmi_hz = HZ; +static unsigned int nmi_perfctr_msr; /* the MSR to reset in NMI handler */ +static unsigned int nmi_p4_cccr_val; +extern void show_registers(struct pt_regs *regs); + +/* + * lapic_nmi_owner tracks the ownership of the lapic NMI hardware: + * - it may be reserved by some other driver, or not + * - when not reserved by some other driver, it may be used for + * the NMI watchdog, or not + * + * This is maintained separately from nmi_active because the NMI + * watchdog may also be driven from the I/O APIC timer. + */ +static DEFINE_SPINLOCK(lapic_nmi_owner_lock); +static unsigned int lapic_nmi_owner; +#define LAPIC_NMI_WATCHDOG (1<<0) +#define LAPIC_NMI_RESERVED (1<<1) + +/* nmi_active: + * +1: the lapic NMI watchdog is active, but can be disabled + * 0: the lapic NMI watchdog has not been set up, and cannot + * be enabled + * -1: the lapic NMI watchdog is disabled, but can be enabled + */ +int nmi_active; + +#define K7_EVNTSEL_ENABLE (1 << 22) +#define K7_EVNTSEL_INT (1 << 20) +#define K7_EVNTSEL_OS (1 << 17) +#define K7_EVNTSEL_USR (1 << 16) +#define K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING 0x76 +#define K7_NMI_EVENT K7_EVENT_CYCLES_PROCESSOR_IS_RUNNING + +#define P6_EVNTSEL0_ENABLE (1 << 22) +#define P6_EVNTSEL_INT (1 << 20) +#define P6_EVNTSEL_OS (1 << 17) +#define P6_EVNTSEL_USR (1 << 16) +#define P6_EVENT_CPU_CLOCKS_NOT_HALTED 0x79 +#define P6_NMI_EVENT P6_EVENT_CPU_CLOCKS_NOT_HALTED + +#define MSR_P4_MISC_ENABLE 0x1A0 +#define MSR_P4_MISC_ENABLE_PERF_AVAIL (1<<7) +#define MSR_P4_MISC_ENABLE_PEBS_UNAVAIL (1<<12) +#define MSR_P4_PERFCTR0 0x300 +#define MSR_P4_CCCR0 0x360 +#define P4_ESCR_EVENT_SELECT(N) ((N)<<25) +#define P4_ESCR_OS (1<<3) +#define P4_ESCR_USR (1<<2) +#define P4_CCCR_OVF_PMI0 (1<<26) +#define P4_CCCR_OVF_PMI1 (1<<27) +#define P4_CCCR_THRESHOLD(N) ((N)<<20) +#define P4_CCCR_COMPLEMENT (1<<19) +#define P4_CCCR_COMPARE (1<<18) +#define P4_CCCR_REQUIRED (3<<16) +#define P4_CCCR_ESCR_SELECT(N) ((N)<<13) +#define P4_CCCR_ENABLE (1<<12) +/* Set up IQ_COUNTER0 to behave like a clock, by having IQ_CCCR0 filter + CRU_ESCR0 (with any non-null event selector) through a complemented + max threshold. [IA32-Vol3, Section 14.9.9] */ +#define MSR_P4_IQ_COUNTER0 0x30C +#define P4_NMI_CRU_ESCR0 (P4_ESCR_EVENT_SELECT(0x3F)|P4_ESCR_OS|P4_ESCR_USR) +#define P4_NMI_IQ_CCCR0 \ + (P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT| \ + P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE) + +int __init check_nmi_watchdog (void) +{ + unsigned int prev_nmi_count[NR_CPUS]; + int cpu; + + printk(KERN_INFO "testing NMI watchdog ... "); + + for (cpu = 0; cpu < NR_CPUS; cpu++) + prev_nmi_count[cpu] = per_cpu(irq_stat, cpu).__nmi_count; + local_irq_enable(); + mdelay((10*1000)/nmi_hz); // wait 10 ticks + + /* FIXME: Only boot CPU is online at this stage. Check CPUs + as they come up. */ + for (cpu = 0; cpu < NR_CPUS; cpu++) { +#ifdef CONFIG_SMP + /* Check cpu_callin_map here because that is set + after the timer is started. */ + if (!cpu_isset(cpu, cpu_callin_map)) + continue; +#endif + if (nmi_count(cpu) - prev_nmi_count[cpu] <= 5) { + printk("CPU#%d: NMI appears to be stuck!\n", cpu); + nmi_active = 0; + lapic_nmi_owner &= ~LAPIC_NMI_WATCHDOG; + return -1; + } + } + printk("OK.\n"); + + /* now that we know it works we can reduce NMI frequency to + something more reasonable; makes a difference in some configs */ + if (nmi_watchdog == NMI_LOCAL_APIC) + nmi_hz = 1; + + return 0; +} + +static int __init setup_nmi_watchdog(char *str) +{ + int nmi; + + get_option(&str, &nmi); + + if (nmi >= NMI_INVALID) + return 0; + if (nmi == NMI_NONE) + nmi_watchdog = nmi; + /* + * If any other x86 CPU has a local APIC, then + * please test the NMI stuff there and send me the + * missing bits. Right now Intel P6/P4 and AMD K7 only. + */ + if ((nmi == NMI_LOCAL_APIC) && + (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL) && + (boot_cpu_data.x86 == 6 || boot_cpu_data.x86 == 15)) + nmi_watchdog = nmi; + if ((nmi == NMI_LOCAL_APIC) && + (boot_cpu_data.x86_vendor == X86_VENDOR_AMD) && + (boot_cpu_data.x86 == 6 || boot_cpu_data.x86 == 15)) + nmi_watchdog = nmi; + /* + * We can enable the IO-APIC watchdog + * unconditionally. + */ + if (nmi == NMI_IO_APIC) { + nmi_active = 1; + nmi_watchdog = nmi; + } + return 1; +} + +__setup("nmi_watchdog=", setup_nmi_watchdog); + +static void disable_lapic_nmi_watchdog(void) +{ + if (nmi_active <= 0) + return; + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + wrmsr(MSR_K7_EVNTSEL0, 0, 0); + break; + case X86_VENDOR_INTEL: + switch (boot_cpu_data.x86) { + case 6: + if (boot_cpu_data.x86_model > 0xd) + break; + + wrmsr(MSR_P6_EVNTSEL0, 0, 0); + break; + case 15: + if (boot_cpu_data.x86_model > 0x3) + break; + + wrmsr(MSR_P4_IQ_CCCR0, 0, 0); + wrmsr(MSR_P4_CRU_ESCR0, 0, 0); + break; + } + break; + } + nmi_active = -1; + /* tell do_nmi() and others that we're not active any more */ + nmi_watchdog = 0; +} + +static void enable_lapic_nmi_watchdog(void) +{ + if (nmi_active < 0) { + nmi_watchdog = NMI_LOCAL_APIC; + setup_apic_nmi_watchdog(); + } +} + +int reserve_lapic_nmi(void) +{ + unsigned int old_owner; + + spin_lock(&lapic_nmi_owner_lock); + old_owner = lapic_nmi_owner; + lapic_nmi_owner |= LAPIC_NMI_RESERVED; + spin_unlock(&lapic_nmi_owner_lock); + if (old_owner & LAPIC_NMI_RESERVED) + return -EBUSY; + if (old_owner & LAPIC_NMI_WATCHDOG) + disable_lapic_nmi_watchdog(); + return 0; +} + +void release_lapic_nmi(void) +{ + unsigned int new_owner; + + spin_lock(&lapic_nmi_owner_lock); + new_owner = lapic_nmi_owner & ~LAPIC_NMI_RESERVED; + lapic_nmi_owner = new_owner; + spin_unlock(&lapic_nmi_owner_lock); + if (new_owner & LAPIC_NMI_WATCHDOG) + enable_lapic_nmi_watchdog(); +} + +void disable_timer_nmi_watchdog(void) +{ + if ((nmi_watchdog != NMI_IO_APIC) || (nmi_active <= 0)) + return; + + unset_nmi_callback(); + nmi_active = -1; + nmi_watchdog = NMI_NONE; +} + +void enable_timer_nmi_watchdog(void) +{ + if (nmi_active < 0) { + nmi_watchdog = NMI_IO_APIC; + touch_nmi_watchdog(); + nmi_active = 1; + } +} + +#ifdef CONFIG_PM + +static int nmi_pm_active; /* nmi_active before suspend */ + +static int lapic_nmi_suspend(struct sys_device *dev, u32 state) +{ + nmi_pm_active = nmi_active; + disable_lapic_nmi_watchdog(); + return 0; +} + +static int lapic_nmi_resume(struct sys_device *dev) +{ + if (nmi_pm_active > 0) + enable_lapic_nmi_watchdog(); + return 0; +} + + +static struct sysdev_class nmi_sysclass = { + set_kset_name("lapic_nmi"), + .resume = lapic_nmi_resume, + .suspend = lapic_nmi_suspend, +}; + +static struct sys_device device_lapic_nmi = { + .id = 0, + .cls = &nmi_sysclass, +}; + +static int __init init_lapic_nmi_sysfs(void) +{ + int error; + + if (nmi_active == 0 || nmi_watchdog != NMI_LOCAL_APIC) + return 0; + + error = sysdev_class_register(&nmi_sysclass); + if (!error) + error = sysdev_register(&device_lapic_nmi); + return error; +} +/* must come after the local APIC's device_initcall() */ +late_initcall(init_lapic_nmi_sysfs); + +#endif /* CONFIG_PM */ + +/* + * Activate the NMI watchdog via the local APIC. + * Original code written by Keith Owens. + */ + +static void clear_msr_range(unsigned int base, unsigned int n) +{ + unsigned int i; + + for(i = 0; i < n; ++i) + wrmsr(base+i, 0, 0); +} + +static void setup_k7_watchdog(void) +{ + unsigned int evntsel; + + nmi_perfctr_msr = MSR_K7_PERFCTR0; + + clear_msr_range(MSR_K7_EVNTSEL0, 4); + clear_msr_range(MSR_K7_PERFCTR0, 4); + + evntsel = K7_EVNTSEL_INT + | K7_EVNTSEL_OS + | K7_EVNTSEL_USR + | K7_NMI_EVENT; + + wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); + Dprintk("setting K7_PERFCTR0 to %08lx\n", -(cpu_khz/nmi_hz*1000)); + wrmsr(MSR_K7_PERFCTR0, -(cpu_khz/nmi_hz*1000), -1); + apic_write(APIC_LVTPC, APIC_DM_NMI); + evntsel |= K7_EVNTSEL_ENABLE; + wrmsr(MSR_K7_EVNTSEL0, evntsel, 0); +} + +static void setup_p6_watchdog(void) +{ + unsigned int evntsel; + + nmi_perfctr_msr = MSR_P6_PERFCTR0; + + clear_msr_range(MSR_P6_EVNTSEL0, 2); + clear_msr_range(MSR_P6_PERFCTR0, 2); + + evntsel = P6_EVNTSEL_INT + | P6_EVNTSEL_OS + | P6_EVNTSEL_USR + | P6_NMI_EVENT; + + wrmsr(MSR_P6_EVNTSEL0, evntsel, 0); + Dprintk("setting P6_PERFCTR0 to %08lx\n", -(cpu_khz/nmi_hz*1000)); + wrmsr(MSR_P6_PERFCTR0, -(cpu_khz/nmi_hz*1000), 0); + apic_write(APIC_LVTPC, APIC_DM_NMI); + evntsel |= P6_EVNTSEL0_ENABLE; + wrmsr(MSR_P6_EVNTSEL0, evntsel, 0); +} + +static int setup_p4_watchdog(void) +{ + unsigned int misc_enable, dummy; + + rdmsr(MSR_P4_MISC_ENABLE, misc_enable, dummy); + if (!(misc_enable & MSR_P4_MISC_ENABLE_PERF_AVAIL)) + return 0; + + nmi_perfctr_msr = MSR_P4_IQ_COUNTER0; + nmi_p4_cccr_val = P4_NMI_IQ_CCCR0; +#ifdef CONFIG_SMP + if (smp_num_siblings == 2) + nmi_p4_cccr_val |= P4_CCCR_OVF_PMI1; +#endif + + if (!(misc_enable & MSR_P4_MISC_ENABLE_PEBS_UNAVAIL)) + clear_msr_range(0x3F1, 2); + /* MSR 0x3F0 seems to have a default value of 0xFC00, but current + docs doesn't fully define it, so leave it alone for now. */ + if (boot_cpu_data.x86_model >= 0x3) { + /* MSR_P4_IQ_ESCR0/1 (0x3ba/0x3bb) removed */ + clear_msr_range(0x3A0, 26); + clear_msr_range(0x3BC, 3); + } else { + clear_msr_range(0x3A0, 31); + } + clear_msr_range(0x3C0, 6); + clear_msr_range(0x3C8, 6); + clear_msr_range(0x3E0, 2); + clear_msr_range(MSR_P4_CCCR0, 18); + clear_msr_range(MSR_P4_PERFCTR0, 18); + + wrmsr(MSR_P4_CRU_ESCR0, P4_NMI_CRU_ESCR0, 0); + wrmsr(MSR_P4_IQ_CCCR0, P4_NMI_IQ_CCCR0 & ~P4_CCCR_ENABLE, 0); + Dprintk("setting P4_IQ_COUNTER0 to 0x%08lx\n", -(cpu_khz/nmi_hz*1000)); + wrmsr(MSR_P4_IQ_COUNTER0, -(cpu_khz/nmi_hz*1000), -1); + apic_write(APIC_LVTPC, APIC_DM_NMI); + wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); + return 1; +} + +void setup_apic_nmi_watchdog (void) +{ + switch (boot_cpu_data.x86_vendor) { + case X86_VENDOR_AMD: + if (boot_cpu_data.x86 != 6 && boot_cpu_data.x86 != 15) + return; + setup_k7_watchdog(); + break; + case X86_VENDOR_INTEL: + switch (boot_cpu_data.x86) { + case 6: + if (boot_cpu_data.x86_model > 0xd) + return; + + setup_p6_watchdog(); + break; + case 15: + if (boot_cpu_data.x86_model > 0x3) + return; + + if (!setup_p4_watchdog()) + return; + break; + default: + return; + } + break; + default: + return; + } + lapic_nmi_owner = LAPIC_NMI_WATCHDOG; + nmi_active = 1; +} + +/* + * the best way to detect whether a CPU has a 'hard lockup' problem + * is to check it's local APIC timer IRQ counts. If they are not + * changing then that CPU has some problem. + * + * as these watchdog NMI IRQs are generated on every CPU, we only + * have to check the current processor. + * + * since NMIs don't listen to _any_ locks, we have to be extremely + * careful not to rely on unsafe variables. The printk might lock + * up though, so we have to break up any console locks first ... + * [when there will be more tty-related locks, break them up + * here too!] + */ + +static unsigned int + last_irq_sums [NR_CPUS], + alert_counter [NR_CPUS]; + +void touch_nmi_watchdog (void) +{ + int i; + + /* + * Just reset the alert counters, (other CPUs might be + * spinning on locks we hold): + */ + for (i = 0; i < NR_CPUS; i++) + alert_counter[i] = 0; +} + +extern void die_nmi(struct pt_regs *, const char *msg); + +void nmi_watchdog_tick (struct pt_regs * regs) +{ + + /* + * Since current_thread_info()-> is always on the stack, and we + * always switch the stack NMI-atomically, it's safe to use + * smp_processor_id(). + */ + int sum, cpu = smp_processor_id(); + + sum = per_cpu(irq_stat, cpu).apic_timer_irqs; + + if (last_irq_sums[cpu] == sum) { + /* + * Ayiee, looks like this CPU is stuck ... + * wait a few IRQs (5 seconds) before doing the oops ... + */ + alert_counter[cpu]++; + if (alert_counter[cpu] == 5*nmi_hz) + die_nmi(regs, "NMI Watchdog detected LOCKUP"); + } else { + last_irq_sums[cpu] = sum; + alert_counter[cpu] = 0; + } + if (nmi_perfctr_msr) { + if (nmi_perfctr_msr == MSR_P4_IQ_COUNTER0) { + /* + * P4 quirks: + * - An overflown perfctr will assert its interrupt + * until the OVF flag in its CCCR is cleared. + * - LVTPC is masked on interrupt and must be + * unmasked by the LVTPC handler. + */ + wrmsr(MSR_P4_IQ_CCCR0, nmi_p4_cccr_val, 0); + apic_write(APIC_LVTPC, APIC_DM_NMI); + } + else if (nmi_perfctr_msr == MSR_P6_PERFCTR0) { + /* Only P6 based Pentium M need to re-unmask + * the apic vector but it doesn't hurt + * other P6 variant */ + apic_write(APIC_LVTPC, APIC_DM_NMI); + } + wrmsr(nmi_perfctr_msr, -(cpu_khz/nmi_hz*1000), -1); + } +} + +#ifdef CONFIG_SYSCTL + +static int unknown_nmi_panic_callback(struct pt_regs *regs, int cpu) +{ + unsigned char reason = get_nmi_reason(); + char buf[64]; + + if (!(reason & 0xc0)) { + sprintf(buf, "NMI received for unknown reason %02x\n", reason); + die_nmi(regs, buf); + } + return 0; +} + +/* + * proc handler for /proc/sys/kernel/unknown_nmi_panic + */ +int proc_unknown_nmi_panic(ctl_table *table, int write, struct file *file, + void __user *buffer, size_t *length, loff_t *ppos) +{ + int old_state; + + old_state = unknown_nmi_panic; + proc_dointvec(table, write, file, buffer, length, ppos); + if (!!old_state == !!unknown_nmi_panic) + return 0; + + if (unknown_nmi_panic) { + if (reserve_lapic_nmi() < 0) { + unknown_nmi_panic = 0; + return -EBUSY; + } else { + set_nmi_callback(unknown_nmi_panic_callback); + } + } else { + release_lapic_nmi(); + unset_nmi_callback(); + } + return 0; +} + +#endif + +EXPORT_SYMBOL(nmi_active); +EXPORT_SYMBOL(nmi_watchdog); +EXPORT_SYMBOL(reserve_lapic_nmi); +EXPORT_SYMBOL(release_lapic_nmi); +EXPORT_SYMBOL(disable_timer_nmi_watchdog); +EXPORT_SYMBOL(enable_timer_nmi_watchdog); diff --git a/arch/i386/kernel/numaq.c b/arch/i386/kernel/numaq.c new file mode 100644 index 00000000000..e51edf0a656 --- /dev/null +++ b/arch/i386/kernel/numaq.c @@ -0,0 +1,79 @@ +/* + * Written by: Patricia Gaughen, IBM Corporation + * + * Copyright (C) 2002, IBM Corp. + * + * All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * + * Send feedback to <gone@us.ibm.com> + */ + +#include <linux/config.h> +#include <linux/mm.h> +#include <linux/bootmem.h> +#include <linux/mmzone.h> +#include <linux/module.h> +#include <linux/nodemask.h> +#include <asm/numaq.h> +#include <asm/topology.h> + +#define MB_TO_PAGES(addr) ((addr) << (20 - PAGE_SHIFT)) + +/* + * Function: smp_dump_qct() + * + * Description: gets memory layout from the quad config table. This + * function also updates node_online_map with the nodes (quads) present. + */ +static void __init smp_dump_qct(void) +{ + int node; + struct eachquadmem *eq; + struct sys_cfg_data *scd = + (struct sys_cfg_data *)__va(SYS_CFG_DATA_PRIV_ADDR); + + nodes_clear(node_online_map); + for_each_node(node) { + if (scd->quads_present31_0 & (1 << node)) { + node_set_online(node); + eq = &scd->eq[node]; + /* Convert to pages */ + node_start_pfn[node] = MB_TO_PAGES( + eq->hi_shrd_mem_start - eq->priv_mem_size); + node_end_pfn[node] = MB_TO_PAGES( + eq->hi_shrd_mem_start + eq->hi_shrd_mem_size); + + memory_present(node, + node_start_pfn[node], node_end_pfn[node]); + node_remap_size[node] = node_memmap_size_bytes(node, + node_start_pfn[node], + node_end_pfn[node]); + } + } +} + +/* + * Unlike Summit, we don't really care to let the NUMA-Q + * fall back to flat mode. Don't compile for NUMA-Q + * unless you really need it! + */ +int __init get_memcfg_numaq(void) +{ + smp_dump_qct(); + return 1; +} diff --git a/arch/i386/kernel/pci-dma.c b/arch/i386/kernel/pci-dma.c new file mode 100644 index 00000000000..4de2e03c7b4 --- /dev/null +++ b/arch/i386/kernel/pci-dma.c @@ -0,0 +1,147 @@ +/* + * Dynamic DMA mapping support. + * + * On i386 there is no hardware dynamic DMA address translation, + * so consistent alloc/free are merely page allocation/freeing. + * The rest of the dynamic DMA mapping interface is implemented + * in asm/pci.h. + */ + +#include <linux/types.h> +#include <linux/mm.h> +#include <linux/string.h> +#include <linux/pci.h> +#include <asm/io.h> + +struct dma_coherent_mem { + void *virt_base; + u32 device_base; + int size; + int flags; + unsigned long *bitmap; +}; + +void *dma_alloc_coherent(struct device *dev, size_t size, + dma_addr_t *dma_handle, unsigned int __nocast gfp) +{ + void *ret; + struct dma_coherent_mem *mem = dev ? dev->dma_mem : NULL; + int order = get_order(size); + /* ignore region specifiers */ + gfp &= ~(__GFP_DMA | __GFP_HIGHMEM); + + if (mem) { + int page = bitmap_find_free_region(mem->bitmap, mem->size, + order); + if (page >= 0) { + *dma_handle = mem->device_base + (page << PAGE_SHIFT); + ret = mem->virt_base + (page << PAGE_SHIFT); + memset(ret, 0, size); + return ret; + } + if (mem->flags & DMA_MEMORY_EXCLUSIVE) + return NULL; + } + + if (dev == NULL || (dev->coherent_dma_mask < 0xffffffff)) + gfp |= GFP_DMA; + + ret = (void *)__get_free_pages(gfp, order); + + if (ret != NULL) { + memset(ret, 0, size); + *dma_handle = virt_to_phys(ret); + } + return ret; +} + +void dma_free_coherent(struct device *dev, size_t size, + void *vaddr, dma_addr_t dma_handle) +{ + struct dma_coherent_mem *mem = dev ? dev->dma_mem : NULL; + int order = get_order(size); + + if (mem && vaddr >= mem->virt_base && vaddr < (mem->virt_base + (mem->size << PAGE_SHIFT))) { + int page = (vaddr - mem->virt_base) >> PAGE_SHIFT; + + bitmap_release_region(mem->bitmap, page, order); + } else + free_pages((unsigned long)vaddr, order); +} + +int dma_declare_coherent_memory(struct device *dev, dma_addr_t bus_addr, + dma_addr_t device_addr, size_t size, int flags) +{ + void __iomem *mem_base; + int pages = size >> PAGE_SHIFT; + int bitmap_size = (pages + 31)/32; + + if ((flags & (DMA_MEMORY_MAP | DMA_MEMORY_IO)) == 0) + goto out; + if (!size) + goto out; + if (dev->dma_mem) + goto out; + + /* FIXME: this routine just ignores DMA_MEMORY_INCLUDES_CHILDREN */ + + mem_base = ioremap(bus_addr, size); + if (!mem_base) + goto out; + + dev->dma_mem = kmalloc(sizeof(struct dma_coherent_mem), GFP_KERNEL); + if (!dev->dma_mem) + goto out; + memset(dev->dma_mem, 0, sizeof(struct dma_coherent_mem)); + dev->dma_mem->bitmap = kmalloc(bitmap_size, GFP_KERNEL); + if (!dev->dma_mem->bitmap) + goto free1_out; + memset(dev->dma_mem->bitmap, 0, bitmap_size); + + dev->dma_mem->virt_base = mem_base; + dev->dma_mem->device_base = device_addr; + dev->dma_mem->size = pages; + dev->dma_mem->flags = flags; + + if (flags & DMA_MEMORY_MAP) + return DMA_MEMORY_MAP; + + return DMA_MEMORY_IO; + + free1_out: + kfree(dev->dma_mem->bitmap); + out: + return 0; +} +EXPORT_SYMBOL(dma_declare_coherent_memory); + +void dma_release_declared_memory(struct device *dev) +{ + struct dma_coherent_mem *mem = dev->dma_mem; + + if(!mem) + return; + dev->dma_mem = NULL; + iounmap(mem->virt_base); + kfree(mem->bitmap); + kfree(mem); +} +EXPORT_SYMBOL(dma_release_declared_memory); + +void *dma_mark_declared_memory_occupied(struct device *dev, + dma_addr_t device_addr, size_t size) +{ + struct dma_coherent_mem *mem = dev->dma_mem; + int pages = (size + (device_addr & ~PAGE_MASK) + PAGE_SIZE - 1) >> PAGE_SHIFT; + int pos, err; + + if (!mem) + return ERR_PTR(-EINVAL); + + pos = (device_addr - mem->device_base) >> PAGE_SHIFT; + err = bitmap_allocate_region(mem->bitmap, pos, get_order(pages)); + if (err != 0) + return ERR_PTR(err); + return mem->virt_base + (pos << PAGE_SHIFT); +} +EXPORT_SYMBOL(dma_mark_declared_memory_occupied); diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c new file mode 100644 index 00000000000..c36fedf40e9 --- /dev/null +++ b/arch/i386/kernel/process.c @@ -0,0 +1,848 @@ +/* + * linux/arch/i386/kernel/process.c + * + * Copyright (C) 1995 Linus Torvalds + * + * Pentium III FXSR, SSE support + * Gareth Hughes <gareth@valinux.com>, May 2000 + */ + +/* + * This file handles the architecture-dependent parts of process handling.. + */ + +#include <stdarg.h> + +#include <linux/errno.h> +#include <linux/sched.h> +#include <linux/fs.h> +#include <linux/kernel.h> +#include <linux/mm.h> +#include <linux/elfcore.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/stddef.h> +#include <linux/slab.h> +#include <linux/vmalloc.h> +#include <linux/user.h> +#include <linux/a.out.h> +#include <linux/interrupt.h> +#include <linux/config.h> +#include <linux/utsname.h> +#include <linux/delay.h> +#include <linux/reboot.h> +#include <linux/init.h> +#include <linux/mc146818rtc.h> +#include <linux/module.h> +#include <linux/kallsyms.h> +#include <linux/ptrace.h> +#include <linux/random.h> + +#include <asm/uaccess.h> +#include <asm/pgtable.h> +#include <asm/system.h> +#include <asm/io.h> +#include <asm/ldt.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/irq.h> +#include <asm/desc.h> +#ifdef CONFIG_MATH_EMULATION +#include <asm/math_emu.h> +#endif + +#include <linux/irq.h> +#include <linux/err.h> + +asmlinkage void ret_from_fork(void) __asm__("ret_from_fork"); + +static int hlt_counter; + +unsigned long boot_option_idle_override = 0; +EXPORT_SYMBOL(boot_option_idle_override); + +/* + * Return saved PC of a blocked thread. + */ +unsigned long thread_saved_pc(struct task_struct *tsk) +{ + return ((unsigned long *)tsk->thread.esp)[3]; +} + +/* + * Powermanagement idle function, if any.. + */ +void (*pm_idle)(void); +static DEFINE_PER_CPU(unsigned int, cpu_idle_state); + +void disable_hlt(void) +{ + hlt_counter++; +} + +EXPORT_SYMBOL(disable_hlt); + +void enable_hlt(void) +{ + hlt_counter--; +} + +EXPORT_SYMBOL(enable_hlt); + +/* + * We use this if we don't have any better + * idle routine.. + */ +void default_idle(void) +{ + if (!hlt_counter && boot_cpu_data.hlt_works_ok) { + local_irq_disable(); + if (!need_resched()) + safe_halt(); + else + local_irq_enable(); + } else { + cpu_relax(); + } +} + +/* + * On SMP it's slightly faster (but much more power-consuming!) + * to poll the ->work.need_resched flag instead of waiting for the + * cross-CPU IPI to arrive. Use this option with caution. + */ +static void poll_idle (void) +{ + int oldval; + + local_irq_enable(); + + /* + * Deal with another CPU just having chosen a thread to + * run here: + */ + oldval = test_and_clear_thread_flag(TIF_NEED_RESCHED); + + if (!oldval) { + set_thread_flag(TIF_POLLING_NRFLAG); + asm volatile( + "2:" + "testl %0, %1;" + "rep; nop;" + "je 2b;" + : : "i"(_TIF_NEED_RESCHED), "m" (current_thread_info()->flags)); + + clear_thread_flag(TIF_POLLING_NRFLAG); + } else { + set_need_resched(); + } +} + +/* + * The idle thread. There's no useful work to be + * done, so just try to conserve power and have a + * low exit latency (ie sit in a loop waiting for + * somebody to say that they'd like to reschedule) + */ +void cpu_idle (void) +{ + /* endless idle loop with no priority at all */ + while (1) { + while (!need_resched()) { + void (*idle)(void); + + if (__get_cpu_var(cpu_idle_state)) + __get_cpu_var(cpu_idle_state) = 0; + + rmb(); + idle = pm_idle; + + if (!idle) + idle = default_idle; + + __get_cpu_var(irq_stat).idle_timestamp = jiffies; + idle(); + } + schedule(); + } +} + +void cpu_idle_wait(void) +{ + unsigned int cpu, this_cpu = get_cpu(); + cpumask_t map; + + set_cpus_allowed(current, cpumask_of_cpu(this_cpu)); + put_cpu(); + + cpus_clear(map); + for_each_online_cpu(cpu) { + per_cpu(cpu_idle_state, cpu) = 1; + cpu_set(cpu, map); + } + + __get_cpu_var(cpu_idle_state) = 0; + + wmb(); + do { + ssleep(1); + for_each_online_cpu(cpu) { + if (cpu_isset(cpu, map) && !per_cpu(cpu_idle_state, cpu)) + cpu_clear(cpu, map); + } + cpus_and(map, map, cpu_online_map); + } while (!cpus_empty(map)); +} +EXPORT_SYMBOL_GPL(cpu_idle_wait); + +/* + * This uses new MONITOR/MWAIT instructions on P4 processors with PNI, + * which can obviate IPI to trigger checking of need_resched. + * We execute MONITOR against need_resched and enter optimized wait state + * through MWAIT. Whenever someone changes need_resched, we would be woken + * up from MWAIT (without an IPI). + */ +static void mwait_idle(void) +{ + local_irq_enable(); + + if (!need_resched()) { + set_thread_flag(TIF_POLLING_NRFLAG); + do { + __monitor((void *)¤t_thread_info()->flags, 0, 0); + if (need_resched()) + break; + __mwait(0, 0); + } while (!need_resched()); + clear_thread_flag(TIF_POLLING_NRFLAG); + } +} + +void __init select_idle_routine(const struct cpuinfo_x86 *c) +{ + if (cpu_has(c, X86_FEATURE_MWAIT)) { + printk("monitor/mwait feature present.\n"); + /* + * Skip, if setup has overridden idle. + * One CPU supports mwait => All CPUs supports mwait + */ + if (!pm_idle) { + printk("using mwait in idle threads.\n"); + pm_idle = mwait_idle; + } + } +} + +static int __init idle_setup (char *str) +{ + if (!strncmp(str, "poll", 4)) { + printk("using polling idle threads.\n"); + pm_idle = poll_idle; +#ifdef CONFIG_X86_SMP + if (smp_num_siblings > 1) + printk("WARNING: polling idle and HT enabled, performance may degrade.\n"); +#endif + } else if (!strncmp(str, "halt", 4)) { + printk("using halt in idle threads.\n"); + pm_idle = default_idle; + } + + boot_option_idle_override = 1; + return 1; +} + +__setup("idle=", idle_setup); + +void show_regs(struct pt_regs * regs) +{ + unsigned long cr0 = 0L, cr2 = 0L, cr3 = 0L, cr4 = 0L; + + printk("\n"); + printk("Pid: %d, comm: %20s\n", current->pid, current->comm); + printk("EIP: %04x:[<%08lx>] CPU: %d\n",0xffff & regs->xcs,regs->eip, smp_processor_id()); + print_symbol("EIP is at %s\n", regs->eip); + + if (regs->xcs & 3) + printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp); + printk(" EFLAGS: %08lx %s (%s)\n", + regs->eflags, print_tainted(), system_utsname.release); + printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n", + regs->eax,regs->ebx,regs->ecx,regs->edx); + printk("ESI: %08lx EDI: %08lx EBP: %08lx", + regs->esi, regs->edi, regs->ebp); + printk(" DS: %04x ES: %04x\n", + 0xffff & regs->xds,0xffff & regs->xes); + + __asm__("movl %%cr0, %0": "=r" (cr0)); + __asm__("movl %%cr2, %0": "=r" (cr2)); + __asm__("movl %%cr3, %0": "=r" (cr3)); + /* This could fault if %cr4 does not exist */ + __asm__("1: movl %%cr4, %0 \n" + "2: \n" + ".section __ex_table,\"a\" \n" + ".long 1b,2b \n" + ".previous \n" + : "=r" (cr4): "0" (0)); + printk("CR0: %08lx CR2: %08lx CR3: %08lx CR4: %08lx\n", cr0, cr2, cr3, cr4); + show_trace(NULL, ®s->esp); +} + +/* + * This gets run with %ebx containing the + * function to call, and %edx containing + * the "args". + */ +extern void kernel_thread_helper(void); +__asm__(".section .text\n" + ".align 4\n" + "kernel_thread_helper:\n\t" + "movl %edx,%eax\n\t" + "pushl %edx\n\t" + "call *%ebx\n\t" + "pushl %eax\n\t" + "call do_exit\n" + ".previous"); + +/* + * Create a kernel thread + */ +int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags) +{ + struct pt_regs regs; + + memset(®s, 0, sizeof(regs)); + + regs.ebx = (unsigned long) fn; + regs.edx = (unsigned long) arg; + + regs.xds = __USER_DS; + regs.xes = __USER_DS; + regs.orig_eax = -1; + regs.eip = (unsigned long) kernel_thread_helper; + regs.xcs = __KERNEL_CS; + regs.eflags = X86_EFLAGS_IF | X86_EFLAGS_SF | X86_EFLAGS_PF | 0x2; + + /* Ok, create the new process.. */ + return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, ®s, 0, NULL, NULL); +} + +/* + * Free current thread data structures etc.. + */ +void exit_thread(void) +{ + struct task_struct *tsk = current; + struct thread_struct *t = &tsk->thread; + + /* The process may have allocated an io port bitmap... nuke it. */ + if (unlikely(NULL != t->io_bitmap_ptr)) { + int cpu = get_cpu(); + struct tss_struct *tss = &per_cpu(init_tss, cpu); + + kfree(t->io_bitmap_ptr); + t->io_bitmap_ptr = NULL; + /* + * Careful, clear this in the TSS too: + */ + memset(tss->io_bitmap, 0xff, tss->io_bitmap_max); + t->io_bitmap_max = 0; + tss->io_bitmap_owner = NULL; + tss->io_bitmap_max = 0; + tss->io_bitmap_base = INVALID_IO_BITMAP_OFFSET; + put_cpu(); + } +} + +void flush_thread(void) +{ + struct task_struct *tsk = current; + + memset(tsk->thread.debugreg, 0, sizeof(unsigned long)*8); + memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array)); + /* + * Forget coprocessor state.. + */ + clear_fpu(tsk); + clear_used_math(); +} + +void release_thread(struct task_struct *dead_task) +{ + if (dead_task->mm) { + // temporary debugging check + if (dead_task->mm->context.size) { + printk("WARNING: dead process %8s still has LDT? <%p/%d>\n", + dead_task->comm, + dead_task->mm->context.ldt, + dead_task->mm->context.size); + BUG(); + } + } + + release_vm86_irqs(dead_task); +} + +/* + * This gets called before we allocate a new thread and copy + * the current task into it. + */ +void prepare_to_copy(struct task_struct *tsk) +{ + unlazy_fpu(tsk); +} + +int copy_thread(int nr, unsigned long clone_flags, unsigned long esp, + unsigned long unused, + struct task_struct * p, struct pt_regs * regs) +{ + struct pt_regs * childregs; + struct task_struct *tsk; + int err; + + childregs = ((struct pt_regs *) (THREAD_SIZE + (unsigned long) p->thread_info)) - 1; + *childregs = *regs; + childregs->eax = 0; + childregs->esp = esp; + + p->thread.esp = (unsigned long) childregs; + p->thread.esp0 = (unsigned long) (childregs+1); + + p->thread.eip = (unsigned long) ret_from_fork; + + savesegment(fs,p->thread.fs); + savesegment(gs,p->thread.gs); + + tsk = current; + if (unlikely(NULL != tsk->thread.io_bitmap_ptr)) { + p->thread.io_bitmap_ptr = kmalloc(IO_BITMAP_BYTES, GFP_KERNEL); + if (!p->thread.io_bitmap_ptr) { + p->thread.io_bitmap_max = 0; + return -ENOMEM; + } + memcpy(p->thread.io_bitmap_ptr, tsk->thread.io_bitmap_ptr, + IO_BITMAP_BYTES); + } + + /* + * Set a new TLS for the child thread? + */ + if (clone_flags & CLONE_SETTLS) { + struct desc_struct *desc; + struct user_desc info; + int idx; + + err = -EFAULT; + if (copy_from_user(&info, (void __user *)childregs->esi, sizeof(info))) + goto out; + err = -EINVAL; + if (LDT_empty(&info)) + goto out; + + idx = info.entry_number; + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) + goto out; + + desc = p->thread.tls_array + idx - GDT_ENTRY_TLS_MIN; + desc->a = LDT_entry_a(&info); + desc->b = LDT_entry_b(&info); + } + + err = 0; + out: + if (err && p->thread.io_bitmap_ptr) { + kfree(p->thread.io_bitmap_ptr); + p->thread.io_bitmap_max = 0; + } + return err; +} + +/* + * fill in the user structure for a core dump.. + */ +void dump_thread(struct pt_regs * regs, struct user * dump) +{ + int i; + +/* changed the size calculations - should hopefully work better. lbt */ + dump->magic = CMAGIC; + dump->start_code = 0; + dump->start_stack = regs->esp & ~(PAGE_SIZE - 1); + dump->u_tsize = ((unsigned long) current->mm->end_code) >> PAGE_SHIFT; + dump->u_dsize = ((unsigned long) (current->mm->brk + (PAGE_SIZE-1))) >> PAGE_SHIFT; + dump->u_dsize -= dump->u_tsize; + dump->u_ssize = 0; + for (i = 0; i < 8; i++) + dump->u_debugreg[i] = current->thread.debugreg[i]; + + if (dump->start_stack < TASK_SIZE) + dump->u_ssize = ((unsigned long) (TASK_SIZE - dump->start_stack)) >> PAGE_SHIFT; + + dump->regs.ebx = regs->ebx; + dump->regs.ecx = regs->ecx; + dump->regs.edx = regs->edx; + dump->regs.esi = regs->esi; + dump->regs.edi = regs->edi; + dump->regs.ebp = regs->ebp; + dump->regs.eax = regs->eax; + dump->regs.ds = regs->xds; + dump->regs.es = regs->xes; + savesegment(fs,dump->regs.fs); + savesegment(gs,dump->regs.gs); + dump->regs.orig_eax = regs->orig_eax; + dump->regs.eip = regs->eip; + dump->regs.cs = regs->xcs; + dump->regs.eflags = regs->eflags; + dump->regs.esp = regs->esp; + dump->regs.ss = regs->xss; + + dump->u_fpvalid = dump_fpu (regs, &dump->i387); +} + +/* + * Capture the user space registers if the task is not running (in user space) + */ +int dump_task_regs(struct task_struct *tsk, elf_gregset_t *regs) +{ + struct pt_regs ptregs; + + ptregs = *(struct pt_regs *) + ((unsigned long)tsk->thread_info+THREAD_SIZE - sizeof(ptregs)); + ptregs.xcs &= 0xffff; + ptregs.xds &= 0xffff; + ptregs.xes &= 0xffff; + ptregs.xss &= 0xffff; + + elf_core_copy_regs(regs, &ptregs); + + return 1; +} + +static inline void +handle_io_bitmap(struct thread_struct *next, struct tss_struct *tss) +{ + if (!next->io_bitmap_ptr) { + /* + * Disable the bitmap via an invalid offset. We still cache + * the previous bitmap owner and the IO bitmap contents: + */ + tss->io_bitmap_base = INVALID_IO_BITMAP_OFFSET; + return; + } + if (likely(next == tss->io_bitmap_owner)) { + /* + * Previous owner of the bitmap (hence the bitmap content) + * matches the next task, we dont have to do anything but + * to set a valid offset in the TSS: + */ + tss->io_bitmap_base = IO_BITMAP_OFFSET; + return; + } + /* + * Lazy TSS's I/O bitmap copy. We set an invalid offset here + * and we let the task to get a GPF in case an I/O instruction + * is performed. The handler of the GPF will verify that the + * faulting task has a valid I/O bitmap and, it true, does the + * real copy and restart the instruction. This will save us + * redundant copies when the currently switched task does not + * perform any I/O during its timeslice. + */ + tss->io_bitmap_base = INVALID_IO_BITMAP_OFFSET_LAZY; +} +/* + * This special macro can be used to load a debugging register + */ +#define loaddebug(thread,register) \ + __asm__("movl %0,%%db" #register \ + : /* no output */ \ + :"r" (thread->debugreg[register])) + +/* + * switch_to(x,yn) should switch tasks from x to y. + * + * We fsave/fwait so that an exception goes off at the right time + * (as a call from the fsave or fwait in effect) rather than to + * the wrong process. Lazy FP saving no longer makes any sense + * with modern CPU's, and this simplifies a lot of things (SMP + * and UP become the same). + * + * NOTE! We used to use the x86 hardware context switching. The + * reason for not using it any more becomes apparent when you + * try to recover gracefully from saved state that is no longer + * valid (stale segment register values in particular). With the + * hardware task-switch, there is no way to fix up bad state in + * a reasonable manner. + * + * The fact that Intel documents the hardware task-switching to + * be slow is a fairly red herring - this code is not noticeably + * faster. However, there _is_ some room for improvement here, + * so the performance issues may eventually be a valid point. + * More important, however, is the fact that this allows us much + * more flexibility. + * + * The return value (in %eax) will be the "prev" task after + * the task-switch, and shows up in ret_from_fork in entry.S, + * for example. + */ +struct task_struct fastcall * __switch_to(struct task_struct *prev_p, struct task_struct *next_p) +{ + struct thread_struct *prev = &prev_p->thread, + *next = &next_p->thread; + int cpu = smp_processor_id(); + struct tss_struct *tss = &per_cpu(init_tss, cpu); + + /* never put a printk in __switch_to... printk() calls wake_up*() indirectly */ + + __unlazy_fpu(prev_p); + + /* + * Reload esp0, LDT and the page table pointer: + */ + load_esp0(tss, next); + + /* + * Load the per-thread Thread-Local Storage descriptor. + */ + load_TLS(next, cpu); + + /* + * Save away %fs and %gs. No need to save %es and %ds, as + * those are always kernel segments while inside the kernel. + */ + asm volatile("movl %%fs,%0":"=m" (*(int *)&prev->fs)); + asm volatile("movl %%gs,%0":"=m" (*(int *)&prev->gs)); + + /* + * Restore %fs and %gs if needed. + */ + if (unlikely(prev->fs | prev->gs | next->fs | next->gs)) { + loadsegment(fs, next->fs); + loadsegment(gs, next->gs); + } + + /* + * Now maybe reload the debug registers + */ + if (unlikely(next->debugreg[7])) { + loaddebug(next, 0); + loaddebug(next, 1); + loaddebug(next, 2); + loaddebug(next, 3); + /* no 4 and 5 */ + loaddebug(next, 6); + loaddebug(next, 7); + } + + if (unlikely(prev->io_bitmap_ptr || next->io_bitmap_ptr)) + handle_io_bitmap(next, tss); + + return prev_p; +} + +asmlinkage int sys_fork(struct pt_regs regs) +{ + return do_fork(SIGCHLD, regs.esp, ®s, 0, NULL, NULL); +} + +asmlinkage int sys_clone(struct pt_regs regs) +{ + unsigned long clone_flags; + unsigned long newsp; + int __user *parent_tidptr, *child_tidptr; + + clone_flags = regs.ebx; + newsp = regs.ecx; + parent_tidptr = (int __user *)regs.edx; + child_tidptr = (int __user *)regs.edi; + if (!newsp) + newsp = regs.esp; + return do_fork(clone_flags, newsp, ®s, 0, parent_tidptr, child_tidptr); +} + +/* + * This is trivial, and on the face of it looks like it + * could equally well be done in user mode. + * + * Not so, for quite unobvious reasons - register pressure. + * In user mode vfork() cannot have a stack frame, and if + * done by calling the "clone()" system call directly, you + * do not have enough call-clobbered registers to hold all + * the information you need. + */ +asmlinkage int sys_vfork(struct pt_regs regs) +{ + return do_fork(CLONE_VFORK | CLONE_VM | SIGCHLD, regs.esp, ®s, 0, NULL, NULL); +} + +/* + * sys_execve() executes a new program. + */ +asmlinkage int sys_execve(struct pt_regs regs) +{ + int error; + char * filename; + + filename = getname((char __user *) regs.ebx); + error = PTR_ERR(filename); + if (IS_ERR(filename)) + goto out; + error = do_execve(filename, + (char __user * __user *) regs.ecx, + (char __user * __user *) regs.edx, + ®s); + if (error == 0) { + task_lock(current); + current->ptrace &= ~PT_DTRACE; + task_unlock(current); + /* Make sure we don't return using sysenter.. */ + set_thread_flag(TIF_IRET); + } + putname(filename); +out: + return error; +} + +#define top_esp (THREAD_SIZE - sizeof(unsigned long)) +#define top_ebp (THREAD_SIZE - 2*sizeof(unsigned long)) + +unsigned long get_wchan(struct task_struct *p) +{ + unsigned long ebp, esp, eip; + unsigned long stack_page; + int count = 0; + if (!p || p == current || p->state == TASK_RUNNING) + return 0; + stack_page = (unsigned long)p->thread_info; + esp = p->thread.esp; + if (!stack_page || esp < stack_page || esp > top_esp+stack_page) + return 0; + /* include/asm-i386/system.h:switch_to() pushes ebp last. */ + ebp = *(unsigned long *) esp; + do { + if (ebp < stack_page || ebp > top_ebp+stack_page) + return 0; + eip = *(unsigned long *) (ebp+4); + if (!in_sched_functions(eip)) + return eip; + ebp = *(unsigned long *) ebp; + } while (count++ < 16); + return 0; +} + +/* + * sys_alloc_thread_area: get a yet unused TLS descriptor index. + */ +static int get_free_idx(void) +{ + struct thread_struct *t = ¤t->thread; + int idx; + + for (idx = 0; idx < GDT_ENTRY_TLS_ENTRIES; idx++) + if (desc_empty(t->tls_array + idx)) + return idx + GDT_ENTRY_TLS_MIN; + return -ESRCH; +} + +/* + * Set a given TLS descriptor: + */ +asmlinkage int sys_set_thread_area(struct user_desc __user *u_info) +{ + struct thread_struct *t = ¤t->thread; + struct user_desc info; + struct desc_struct *desc; + int cpu, idx; + + if (copy_from_user(&info, u_info, sizeof(info))) + return -EFAULT; + idx = info.entry_number; + + /* + * index -1 means the kernel should try to find and + * allocate an empty descriptor: + */ + if (idx == -1) { + idx = get_free_idx(); + if (idx < 0) + return idx; + if (put_user(idx, &u_info->entry_number)) + return -EFAULT; + } + + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) + return -EINVAL; + + desc = t->tls_array + idx - GDT_ENTRY_TLS_MIN; + + /* + * We must not get preempted while modifying the TLS. + */ + cpu = get_cpu(); + + if (LDT_empty(&info)) { + desc->a = 0; + desc->b = 0; + } else { + desc->a = LDT_entry_a(&info); + desc->b = LDT_entry_b(&info); + } + load_TLS(t, cpu); + + put_cpu(); + + return 0; +} + +/* + * Get the current Thread-Local Storage area: + */ + +#define GET_BASE(desc) ( \ + (((desc)->a >> 16) & 0x0000ffff) | \ + (((desc)->b << 16) & 0x00ff0000) | \ + ( (desc)->b & 0xff000000) ) + +#define GET_LIMIT(desc) ( \ + ((desc)->a & 0x0ffff) | \ + ((desc)->b & 0xf0000) ) + +#define GET_32BIT(desc) (((desc)->b >> 22) & 1) +#define GET_CONTENTS(desc) (((desc)->b >> 10) & 3) +#define GET_WRITABLE(desc) (((desc)->b >> 9) & 1) +#define GET_LIMIT_PAGES(desc) (((desc)->b >> 23) & 1) +#define GET_PRESENT(desc) (((desc)->b >> 15) & 1) +#define GET_USEABLE(desc) (((desc)->b >> 20) & 1) + +asmlinkage int sys_get_thread_area(struct user_desc __user *u_info) +{ + struct user_desc info; + struct desc_struct *desc; + int idx; + + if (get_user(idx, &u_info->entry_number)) + return -EFAULT; + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) + return -EINVAL; + + desc = current->thread.tls_array + idx - GDT_ENTRY_TLS_MIN; + + info.entry_number = idx; + info.base_addr = GET_BASE(desc); + info.limit = GET_LIMIT(desc); + info.seg_32bit = GET_32BIT(desc); + info.contents = GET_CONTENTS(desc); + info.read_exec_only = !GET_WRITABLE(desc); + info.limit_in_pages = GET_LIMIT_PAGES(desc); + info.seg_not_present = !GET_PRESENT(desc); + info.useable = GET_USEABLE(desc); + + if (copy_to_user(u_info, &info, sizeof(info))) + return -EFAULT; + return 0; +} + +unsigned long arch_align_stack(unsigned long sp) +{ + if (randomize_va_space) + sp -= get_random_int() % 8192; + return sp & ~0xf; +} diff --git a/arch/i386/kernel/ptrace.c b/arch/i386/kernel/ptrace.c new file mode 100644 index 00000000000..b2f17640cef --- /dev/null +++ b/arch/i386/kernel/ptrace.c @@ -0,0 +1,717 @@ +/* ptrace.c */ +/* By Ross Biro 1/23/92 */ +/* + * Pentium III FXSR, SSE support + * Gareth Hughes <gareth@valinux.com>, May 2000 + */ + +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/errno.h> +#include <linux/ptrace.h> +#include <linux/user.h> +#include <linux/security.h> +#include <linux/audit.h> +#include <linux/seccomp.h> + +#include <asm/uaccess.h> +#include <asm/pgtable.h> +#include <asm/system.h> +#include <asm/processor.h> +#include <asm/i387.h> +#include <asm/debugreg.h> +#include <asm/ldt.h> +#include <asm/desc.h> + +/* + * does not yet catch signals sent when the child dies. + * in exit.c or in signal.c. + */ + +/* determines which flags the user has access to. */ +/* 1 = access 0 = no access */ +#define FLAG_MASK 0x00044dd5 + +/* set's the trap flag. */ +#define TRAP_FLAG 0x100 + +/* + * Offset of eflags on child stack.. + */ +#define EFL_OFFSET ((EFL-2)*4-sizeof(struct pt_regs)) + +static inline struct pt_regs *get_child_regs(struct task_struct *task) +{ + void *stack_top = (void *)task->thread.esp0; + return stack_top - sizeof(struct pt_regs); +} + +/* + * this routine will get a word off of the processes privileged stack. + * the offset is how far from the base addr as stored in the TSS. + * this routine assumes that all the privileged stacks are in our + * data space. + */ +static inline int get_stack_long(struct task_struct *task, int offset) +{ + unsigned char *stack; + + stack = (unsigned char *)task->thread.esp0; + stack += offset; + return (*((int *)stack)); +} + +/* + * this routine will put a word on the processes privileged stack. + * the offset is how far from the base addr as stored in the TSS. + * this routine assumes that all the privileged stacks are in our + * data space. + */ +static inline int put_stack_long(struct task_struct *task, int offset, + unsigned long data) +{ + unsigned char * stack; + + stack = (unsigned char *) task->thread.esp0; + stack += offset; + *(unsigned long *) stack = data; + return 0; +} + +static int putreg(struct task_struct *child, + unsigned long regno, unsigned long value) +{ + switch (regno >> 2) { + case FS: + if (value && (value & 3) != 3) + return -EIO; + child->thread.fs = value; + return 0; + case GS: + if (value && (value & 3) != 3) + return -EIO; + child->thread.gs = value; + return 0; + case DS: + case ES: + if (value && (value & 3) != 3) + return -EIO; + value &= 0xffff; + break; + case SS: + case CS: + if ((value & 3) != 3) + return -EIO; + value &= 0xffff; + break; + case EFL: + value &= FLAG_MASK; + value |= get_stack_long(child, EFL_OFFSET) & ~FLAG_MASK; + break; + } + if (regno > GS*4) + regno -= 2*4; + put_stack_long(child, regno - sizeof(struct pt_regs), value); + return 0; +} + +static unsigned long getreg(struct task_struct *child, + unsigned long regno) +{ + unsigned long retval = ~0UL; + + switch (regno >> 2) { + case FS: + retval = child->thread.fs; + break; + case GS: + retval = child->thread.gs; + break; + case DS: + case ES: + case SS: + case CS: + retval = 0xffff; + /* fall through */ + default: + if (regno > GS*4) + regno -= 2*4; + regno = regno - sizeof(struct pt_regs); + retval &= get_stack_long(child, regno); + } + return retval; +} + +#define LDT_SEGMENT 4 + +static unsigned long convert_eip_to_linear(struct task_struct *child, struct pt_regs *regs) +{ + unsigned long addr, seg; + + addr = regs->eip; + seg = regs->xcs & 0xffff; + if (regs->eflags & VM_MASK) { + addr = (addr & 0xffff) + (seg << 4); + return addr; + } + + /* + * We'll assume that the code segments in the GDT + * are all zero-based. That is largely true: the + * TLS segments are used for data, and the PNPBIOS + * and APM bios ones we just ignore here. + */ + if (seg & LDT_SEGMENT) { + u32 *desc; + unsigned long base; + + down(&child->mm->context.sem); + desc = child->mm->context.ldt + (seg & ~7); + base = (desc[0] >> 16) | ((desc[1] & 0xff) << 16) | (desc[1] & 0xff000000); + + /* 16-bit code segment? */ + if (!((desc[1] >> 22) & 1)) + addr &= 0xffff; + addr += base; + up(&child->mm->context.sem); + } + return addr; +} + +static inline int is_at_popf(struct task_struct *child, struct pt_regs *regs) +{ + int i, copied; + unsigned char opcode[16]; + unsigned long addr = convert_eip_to_linear(child, regs); + + copied = access_process_vm(child, addr, opcode, sizeof(opcode), 0); + for (i = 0; i < copied; i++) { + switch (opcode[i]) { + /* popf */ + case 0x9d: + return 1; + /* opcode and address size prefixes */ + case 0x66: case 0x67: + continue; + /* irrelevant prefixes (segment overrides and repeats) */ + case 0x26: case 0x2e: + case 0x36: case 0x3e: + case 0x64: case 0x65: + case 0xf0: case 0xf2: case 0xf3: + continue; + + /* + * pushf: NOTE! We should probably not let + * the user see the TF bit being set. But + * it's more pain than it's worth to avoid + * it, and a debugger could emulate this + * all in user space if it _really_ cares. + */ + case 0x9c: + default: + return 0; + } + } + return 0; +} + +static void set_singlestep(struct task_struct *child) +{ + struct pt_regs *regs = get_child_regs(child); + + /* + * Always set TIF_SINGLESTEP - this guarantees that + * we single-step system calls etc.. This will also + * cause us to set TF when returning to user mode. + */ + set_tsk_thread_flag(child, TIF_SINGLESTEP); + + /* + * If TF was already set, don't do anything else + */ + if (regs->eflags & TRAP_FLAG) + return; + + /* Set TF on the kernel stack.. */ + regs->eflags |= TRAP_FLAG; + + /* + * ..but if TF is changed by the instruction we will trace, + * don't mark it as being "us" that set it, so that we + * won't clear it by hand later. + */ + if (is_at_popf(child, regs)) + return; + + child->ptrace |= PT_DTRACE; +} + +static void clear_singlestep(struct task_struct *child) +{ + /* Always clear TIF_SINGLESTEP... */ + clear_tsk_thread_flag(child, TIF_SINGLESTEP); + + /* But touch TF only if it was set by us.. */ + if (child->ptrace & PT_DTRACE) { + struct pt_regs *regs = get_child_regs(child); + regs->eflags &= ~TRAP_FLAG; + child->ptrace &= ~PT_DTRACE; + } +} + +/* + * Called by kernel/ptrace.c when detaching.. + * + * Make sure the single step bit is not set. + */ +void ptrace_disable(struct task_struct *child) +{ + clear_singlestep(child); +} + +/* + * Perform get_thread_area on behalf of the traced child. + */ +static int +ptrace_get_thread_area(struct task_struct *child, + int idx, struct user_desc __user *user_desc) +{ + struct user_desc info; + struct desc_struct *desc; + +/* + * Get the current Thread-Local Storage area: + */ + +#define GET_BASE(desc) ( \ + (((desc)->a >> 16) & 0x0000ffff) | \ + (((desc)->b << 16) & 0x00ff0000) | \ + ( (desc)->b & 0xff000000) ) + +#define GET_LIMIT(desc) ( \ + ((desc)->a & 0x0ffff) | \ + ((desc)->b & 0xf0000) ) + +#define GET_32BIT(desc) (((desc)->b >> 22) & 1) +#define GET_CONTENTS(desc) (((desc)->b >> 10) & 3) +#define GET_WRITABLE(desc) (((desc)->b >> 9) & 1) +#define GET_LIMIT_PAGES(desc) (((desc)->b >> 23) & 1) +#define GET_PRESENT(desc) (((desc)->b >> 15) & 1) +#define GET_USEABLE(desc) (((desc)->b >> 20) & 1) + + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) + return -EINVAL; + + desc = child->thread.tls_array + idx - GDT_ENTRY_TLS_MIN; + + info.entry_number = idx; + info.base_addr = GET_BASE(desc); + info.limit = GET_LIMIT(desc); + info.seg_32bit = GET_32BIT(desc); + info.contents = GET_CONTENTS(desc); + info.read_exec_only = !GET_WRITABLE(desc); + info.limit_in_pages = GET_LIMIT_PAGES(desc); + info.seg_not_present = !GET_PRESENT(desc); + info.useable = GET_USEABLE(desc); + + if (copy_to_user(user_desc, &info, sizeof(info))) + return -EFAULT; + + return 0; +} + +/* + * Perform set_thread_area on behalf of the traced child. + */ +static int +ptrace_set_thread_area(struct task_struct *child, + int idx, struct user_desc __user *user_desc) +{ + struct user_desc info; + struct desc_struct *desc; + + if (copy_from_user(&info, user_desc, sizeof(info))) + return -EFAULT; + + if (idx < GDT_ENTRY_TLS_MIN || idx > GDT_ENTRY_TLS_MAX) + return -EINVAL; + + desc = child->thread.tls_array + idx - GDT_ENTRY_TLS_MIN; + if (LDT_empty(&info)) { + desc->a = 0; + desc->b = 0; + } else { + desc->a = LDT_entry_a(&info); + desc->b = LDT_entry_b(&info); + } + + return 0; +} + +asmlinkage int sys_ptrace(long request, long pid, long addr, long data) +{ + struct task_struct *child; + struct user * dummy = NULL; + int i, ret; + unsigned long __user *datap = (unsigned long __user *)data; + + lock_kernel(); + ret = -EPERM; + if (request == PTRACE_TRACEME) { + /* are we already being traced? */ + if (current->ptrace & PT_PTRACED) + goto out; + ret = security_ptrace(current->parent, current); + if (ret) + goto out; + /* set the ptrace bit in the process flags. */ + current->ptrace |= PT_PTRACED; + ret = 0; + goto out; + } + ret = -ESRCH; + read_lock(&tasklist_lock); + child = find_task_by_pid(pid); + if (child) + get_task_struct(child); + read_unlock(&tasklist_lock); + if (!child) + goto out; + + ret = -EPERM; + if (pid == 1) /* you may not mess with init */ + goto out_tsk; + + if (request == PTRACE_ATTACH) { + ret = ptrace_attach(child); + goto out_tsk; + } + + ret = ptrace_check_attach(child, request == PTRACE_KILL); + if (ret < 0) + goto out_tsk; + + switch (request) { + /* when I and D space are separate, these will need to be fixed. */ + case PTRACE_PEEKTEXT: /* read word at location addr. */ + case PTRACE_PEEKDATA: { + unsigned long tmp; + int copied; + + copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0); + ret = -EIO; + if (copied != sizeof(tmp)) + break; + ret = put_user(tmp, datap); + break; + } + + /* read the word at location addr in the USER area. */ + case PTRACE_PEEKUSR: { + unsigned long tmp; + + ret = -EIO; + if ((addr & 3) || addr < 0 || + addr > sizeof(struct user) - 3) + break; + + tmp = 0; /* Default return condition */ + if(addr < FRAME_SIZE*sizeof(long)) + tmp = getreg(child, addr); + if(addr >= (long) &dummy->u_debugreg[0] && + addr <= (long) &dummy->u_debugreg[7]){ + addr -= (long) &dummy->u_debugreg[0]; + addr = addr >> 2; + tmp = child->thread.debugreg[addr]; + } + ret = put_user(tmp, datap); + break; + } + + /* when I and D space are separate, this will have to be fixed. */ + case PTRACE_POKETEXT: /* write the word at location addr. */ + case PTRACE_POKEDATA: + ret = 0; + if (access_process_vm(child, addr, &data, sizeof(data), 1) == sizeof(data)) + break; + ret = -EIO; + break; + + case PTRACE_POKEUSR: /* write the word at location addr in the USER area */ + ret = -EIO; + if ((addr & 3) || addr < 0 || + addr > sizeof(struct user) - 3) + break; + + if (addr < FRAME_SIZE*sizeof(long)) { + ret = putreg(child, addr, data); + break; + } + /* We need to be very careful here. We implicitly + want to modify a portion of the task_struct, and we + have to be selective about what portions we allow someone + to modify. */ + + ret = -EIO; + if(addr >= (long) &dummy->u_debugreg[0] && + addr <= (long) &dummy->u_debugreg[7]){ + + if(addr == (long) &dummy->u_debugreg[4]) break; + if(addr == (long) &dummy->u_debugreg[5]) break; + if(addr < (long) &dummy->u_debugreg[4] && + ((unsigned long) data) >= TASK_SIZE-3) break; + + /* Sanity-check data. Take one half-byte at once with + * check = (val >> (16 + 4*i)) & 0xf. It contains the + * R/Wi and LENi bits; bits 0 and 1 are R/Wi, and bits + * 2 and 3 are LENi. Given a list of invalid values, + * we do mask |= 1 << invalid_value, so that + * (mask >> check) & 1 is a correct test for invalid + * values. + * + * R/Wi contains the type of the breakpoint / + * watchpoint, LENi contains the length of the watched + * data in the watchpoint case. + * + * The invalid values are: + * - LENi == 0x10 (undefined), so mask |= 0x0f00. + * - R/Wi == 0x10 (break on I/O reads or writes), so + * mask |= 0x4444. + * - R/Wi == 0x00 && LENi != 0x00, so we have mask |= + * 0x1110. + * + * Finally, mask = 0x0f00 | 0x4444 | 0x1110 == 0x5f54. + * + * See the Intel Manual "System Programming Guide", + * 15.2.4 + * + * Note that LENi == 0x10 is defined on x86_64 in long + * mode (i.e. even for 32-bit userspace software, but + * 64-bit kernel), so the x86_64 mask value is 0x5454. + * See the AMD manual no. 24593 (AMD64 System + * Programming)*/ + + if(addr == (long) &dummy->u_debugreg[7]) { + data &= ~DR_CONTROL_RESERVED; + for(i=0; i<4; i++) + if ((0x5f54 >> ((data >> (16 + 4*i)) & 0xf)) & 1) + goto out_tsk; + } + + addr -= (long) &dummy->u_debugreg; + addr = addr >> 2; + child->thread.debugreg[addr] = data; + ret = 0; + } + break; + + case PTRACE_SYSCALL: /* continue and stop at next (return from) syscall */ + case PTRACE_CONT: /* restart after signal. */ + ret = -EIO; + if ((unsigned long) data > _NSIG) + break; + if (request == PTRACE_SYSCALL) { + set_tsk_thread_flag(child, TIF_SYSCALL_TRACE); + } + else { + clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); + } + child->exit_code = data; + /* make sure the single step bit is not set. */ + clear_singlestep(child); + wake_up_process(child); + ret = 0; + break; + +/* + * make the child exit. Best I can do is send it a sigkill. + * perhaps it should be put in the status that it wants to + * exit. + */ + case PTRACE_KILL: + ret = 0; + if (child->exit_state == EXIT_ZOMBIE) /* already dead */ + break; + child->exit_code = SIGKILL; + /* make sure the single step bit is not set. */ + clear_singlestep(child); + wake_up_process(child); + break; + + case PTRACE_SINGLESTEP: /* set the trap flag. */ + ret = -EIO; + if ((unsigned long) data > _NSIG) + break; + clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE); + set_singlestep(child); + child->exit_code = data; + /* give it a chance to run. */ + wake_up_process(child); + ret = 0; + break; + + case PTRACE_DETACH: + /* detach a process that was attached. */ + ret = ptrace_detach(child, data); + break; + + case PTRACE_GETREGS: { /* Get all gp regs from the child. */ + if (!access_ok(VERIFY_WRITE, datap, FRAME_SIZE*sizeof(long))) { + ret = -EIO; + break; + } + for ( i = 0; i < FRAME_SIZE*sizeof(long); i += sizeof(long) ) { + __put_user(getreg(child, i), datap); + datap++; + } + ret = 0; + break; + } + + case PTRACE_SETREGS: { /* Set all gp regs in the child. */ + unsigned long tmp; + if (!access_ok(VERIFY_READ, datap, FRAME_SIZE*sizeof(long))) { + ret = -EIO; + break; + } + for ( i = 0; i < FRAME_SIZE*sizeof(long); i += sizeof(long) ) { + __get_user(tmp, datap); + putreg(child, i, tmp); + datap++; + } + ret = 0; + break; + } + + case PTRACE_GETFPREGS: { /* Get the child FPU state. */ + if (!access_ok(VERIFY_WRITE, datap, + sizeof(struct user_i387_struct))) { + ret = -EIO; + break; + } + ret = 0; + if (!tsk_used_math(child)) + init_fpu(child); + get_fpregs((struct user_i387_struct __user *)data, child); + break; + } + + case PTRACE_SETFPREGS: { /* Set the child FPU state. */ + if (!access_ok(VERIFY_READ, datap, + sizeof(struct user_i387_struct))) { + ret = -EIO; + break; + } + set_stopped_child_used_math(child); + set_fpregs(child, (struct user_i387_struct __user *)data); + ret = 0; + break; + } + + case PTRACE_GETFPXREGS: { /* Get the child extended FPU state. */ + if (!access_ok(VERIFY_WRITE, datap, + sizeof(struct user_fxsr_struct))) { + ret = -EIO; + break; + } + if (!tsk_used_math(child)) + init_fpu(child); + ret = get_fpxregs((struct user_fxsr_struct __user *)data, child); + break; + } + + case PTRACE_SETFPXREGS: { /* Set the child extended FPU state. */ + if (!access_ok(VERIFY_READ, datap, + sizeof(struct user_fxsr_struct))) { + ret = -EIO; + break; + } + set_stopped_child_used_math(child); + ret = set_fpxregs(child, (struct user_fxsr_struct __user *)data); + break; + } + + case PTRACE_GET_THREAD_AREA: + ret = ptrace_get_thread_area(child, addr, + (struct user_desc __user *) data); + break; + + case PTRACE_SET_THREAD_AREA: + ret = ptrace_set_thread_area(child, addr, + (struct user_desc __user *) data); + break; + + default: + ret = ptrace_request(child, request, addr, data); + break; + } +out_tsk: + put_task_struct(child); +out: + unlock_kernel(); + return ret; +} + +void send_sigtrap(struct task_struct *tsk, struct pt_regs *regs, int error_code) +{ + struct siginfo info; + + tsk->thread.trap_no = 1; + tsk->thread.error_code = error_code; + + memset(&info, 0, sizeof(info)); + info.si_signo = SIGTRAP; + info.si_code = TRAP_BRKPT; + + /* User-mode eip? */ + info.si_addr = user_mode(regs) ? (void __user *) regs->eip : NULL; + + /* Send us the fakey SIGTRAP */ + force_sig_info(SIGTRAP, &info, tsk); +} + +/* notification of system call entry/exit + * - triggered by current->work.syscall_trace + */ +__attribute__((regparm(3))) +void do_syscall_trace(struct pt_regs *regs, int entryexit) +{ + /* do the secure computing check first */ + secure_computing(regs->orig_eax); + + if (unlikely(current->audit_context)) { + if (!entryexit) + audit_syscall_entry(current, regs->orig_eax, + regs->ebx, regs->ecx, + regs->edx, regs->esi); + else + audit_syscall_exit(current, regs->eax); + } + + if (!(current->ptrace & PT_PTRACED)) + return; + + /* Fake a debug trap */ + if (test_thread_flag(TIF_SINGLESTEP)) + send_sigtrap(current, regs, 0); + + if (!test_thread_flag(TIF_SYSCALL_TRACE)) + return; + + /* the 0x80 provides a way for the tracing parent to distinguish + between a syscall stop and SIGTRAP delivery */ + ptrace_notify(SIGTRAP | ((current->ptrace & PT_TRACESYSGOOD) ? 0x80 : 0)); + + /* + * this isn't the same as continuing with a signal, but it will do + * for normal use. strace only continues with a signal if the + * stopping signal is not SIGTRAP. -brl + */ + if (current->exit_code) { + send_sig(current->exit_code, current, 1); + current->exit_code = 0; + } +} diff --git a/arch/i386/kernel/quirks.c b/arch/i386/kernel/quirks.c new file mode 100644 index 00000000000..aaf89cb2bc5 --- /dev/null +++ b/arch/i386/kernel/quirks.c @@ -0,0 +1,52 @@ +/* + * This file contains work-arounds for x86 and x86_64 platform bugs. + */ +#include <linux/config.h> +#include <linux/pci.h> +#include <linux/irq.h> + +#if defined(CONFIG_X86_IO_APIC) && defined(CONFIG_SMP) && defined(CONFIG_PCI) + +static void __devinit quirk_intel_irqbalance(struct pci_dev *dev) +{ + u8 config, rev; + u32 word; + + /* BIOS may enable hardware IRQ balancing for + * E7520/E7320/E7525(revision ID 0x9 and below) + * based platforms. + * Disable SW irqbalance/affinity on those platforms. + */ + pci_read_config_byte(dev, PCI_CLASS_REVISION, &rev); + if (rev > 0x9) + return; + + printk(KERN_INFO "Intel E7520/7320/7525 detected."); + + /* enable access to config space*/ + pci_read_config_byte(dev, 0xf4, &config); + config |= 0x2; + pci_write_config_byte(dev, 0xf4, config); + + /* read xTPR register */ + raw_pci_ops->read(0, 0, 0x40, 0x4c, 2, &word); + + if (!(word & (1 << 13))) { + printk(KERN_INFO "Disabling irq balancing and affinity\n"); +#ifdef CONFIG_IRQBALANCE + irqbalance_disable(""); +#endif + noirqdebug_setup(""); +#ifdef CONFIG_PROC_FS + no_irq_affinity = 1; +#endif + } + + config &= ~0x2; + /* disable access to config space*/ + pci_write_config_byte(dev, 0xf4, config); +} +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_E7320_MCH, quirk_intel_irqbalance); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_E7525_MCH, quirk_intel_irqbalance); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_E7520_MCH, quirk_intel_irqbalance); +#endif diff --git a/arch/i386/kernel/reboot.c b/arch/i386/kernel/reboot.c new file mode 100644 index 00000000000..3d7e994563d --- /dev/null +++ b/arch/i386/kernel/reboot.c @@ -0,0 +1,382 @@ +/* + * linux/arch/i386/kernel/reboot.c + */ + +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/mc146818rtc.h> +#include <linux/efi.h> +#include <linux/dmi.h> +#include <asm/uaccess.h> +#include <asm/apic.h> +#include "mach_reboot.h" + +/* + * Power off function, if any + */ +void (*pm_power_off)(void); + +static int reboot_mode; +static int reboot_thru_bios; + +#ifdef CONFIG_SMP +int reboot_smp = 0; +static int reboot_cpu = -1; +/* shamelessly grabbed from lib/vsprintf.c for readability */ +#define is_digit(c) ((c) >= '0' && (c) <= '9') +#endif +static int __init reboot_setup(char *str) +{ + while(1) { + switch (*str) { + case 'w': /* "warm" reboot (no memory testing etc) */ + reboot_mode = 0x1234; + break; + case 'c': /* "cold" reboot (with memory testing etc) */ + reboot_mode = 0x0; + break; + case 'b': /* "bios" reboot by jumping through the BIOS */ + reboot_thru_bios = 1; + break; + case 'h': /* "hard" reboot by toggling RESET and/or crashing the CPU */ + reboot_thru_bios = 0; + break; +#ifdef CONFIG_SMP + case 's': /* "smp" reboot by executing reset on BSP or other CPU*/ + reboot_smp = 1; + if (is_digit(*(str+1))) { + reboot_cpu = (int) (*(str+1) - '0'); + if (is_digit(*(str+2))) + reboot_cpu = reboot_cpu*10 + (int)(*(str+2) - '0'); + } + /* we will leave sorting out the final value + when we are ready to reboot, since we might not + have set up boot_cpu_id or smp_num_cpu */ + break; +#endif + } + if((str = strchr(str,',')) != NULL) + str++; + else + break; + } + return 1; +} + +__setup("reboot=", reboot_setup); + +/* + * Reboot options and system auto-detection code provided by + * Dell Inc. so their systems "just work". :-) + */ + +/* + * Some machines require the "reboot=b" commandline option, this quirk makes that automatic. + */ +static int __init set_bios_reboot(struct dmi_system_id *d) +{ + if (!reboot_thru_bios) { + reboot_thru_bios = 1; + printk(KERN_INFO "%s series board detected. Selecting BIOS-method for reboots.\n", d->ident); + } + return 0; +} + +/* + * Some machines require the "reboot=s" commandline option, this quirk makes that automatic. + */ +static int __init set_smp_reboot(struct dmi_system_id *d) +{ +#ifdef CONFIG_SMP + if (!reboot_smp) { + reboot_smp = 1; + printk(KERN_INFO "%s series board detected. Selecting SMP-method for reboots.\n", d->ident); + } +#endif + return 0; +} + +/* + * Some machines require the "reboot=b,s" commandline option, this quirk makes that automatic. + */ +static int __init set_smp_bios_reboot(struct dmi_system_id *d) +{ + set_smp_reboot(d); + set_bios_reboot(d); + return 0; +} + +static struct dmi_system_id __initdata reboot_dmi_table[] = { + { /* Handle problems with rebooting on Dell 1300's */ + .callback = set_smp_bios_reboot, + .ident = "Dell PowerEdge 1300", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 1300/"), + }, + }, + { /* Handle problems with rebooting on Dell 300's */ + .callback = set_bios_reboot, + .ident = "Dell PowerEdge 300", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 300/"), + }, + }, + { /* Handle problems with rebooting on Dell 2400's */ + .callback = set_bios_reboot, + .ident = "Dell PowerEdge 2400", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Computer Corporation"), + DMI_MATCH(DMI_PRODUCT_NAME, "PowerEdge 2400"), + }, + }, + { } +}; + +static int __init reboot_init(void) +{ + dmi_check_system(reboot_dmi_table); + return 0; +} + +core_initcall(reboot_init); + +/* The following code and data reboots the machine by switching to real + mode and jumping to the BIOS reset entry point, as if the CPU has + really been reset. The previous version asked the keyboard + controller to pulse the CPU reset line, which is more thorough, but + doesn't work with at least one type of 486 motherboard. It is easy + to stop this code working; hence the copious comments. */ + +static unsigned long long +real_mode_gdt_entries [3] = +{ + 0x0000000000000000ULL, /* Null descriptor */ + 0x00009a000000ffffULL, /* 16-bit real-mode 64k code at 0x00000000 */ + 0x000092000100ffffULL /* 16-bit real-mode 64k data at 0x00000100 */ +}; + +static struct +{ + unsigned short size __attribute__ ((packed)); + unsigned long long * base __attribute__ ((packed)); +} +real_mode_gdt = { sizeof (real_mode_gdt_entries) - 1, real_mode_gdt_entries }, +real_mode_idt = { 0x3ff, NULL }, +no_idt = { 0, NULL }; + + +/* This is 16-bit protected mode code to disable paging and the cache, + switch to real mode and jump to the BIOS reset code. + + The instruction that switches to real mode by writing to CR0 must be + followed immediately by a far jump instruction, which set CS to a + valid value for real mode, and flushes the prefetch queue to avoid + running instructions that have already been decoded in protected + mode. + + Clears all the flags except ET, especially PG (paging), PE + (protected-mode enable) and TS (task switch for coprocessor state + save). Flushes the TLB after paging has been disabled. Sets CD and + NW, to disable the cache on a 486, and invalidates the cache. This + is more like the state of a 486 after reset. I don't know if + something else should be done for other chips. + + More could be done here to set up the registers as if a CPU reset had + occurred; hopefully real BIOSs don't assume much. */ + +static unsigned char real_mode_switch [] = +{ + 0x66, 0x0f, 0x20, 0xc0, /* movl %cr0,%eax */ + 0x66, 0x83, 0xe0, 0x11, /* andl $0x00000011,%eax */ + 0x66, 0x0d, 0x00, 0x00, 0x00, 0x60, /* orl $0x60000000,%eax */ + 0x66, 0x0f, 0x22, 0xc0, /* movl %eax,%cr0 */ + 0x66, 0x0f, 0x22, 0xd8, /* movl %eax,%cr3 */ + 0x66, 0x0f, 0x20, 0xc3, /* movl %cr0,%ebx */ + 0x66, 0x81, 0xe3, 0x00, 0x00, 0x00, 0x60, /* andl $0x60000000,%ebx */ + 0x74, 0x02, /* jz f */ + 0x0f, 0x09, /* wbinvd */ + 0x24, 0x10, /* f: andb $0x10,al */ + 0x66, 0x0f, 0x22, 0xc0 /* movl %eax,%cr0 */ +}; +static unsigned char jump_to_bios [] = +{ + 0xea, 0x00, 0x00, 0xff, 0xff /* ljmp $0xffff,$0x0000 */ +}; + +/* + * Switch to real mode and then execute the code + * specified by the code and length parameters. + * We assume that length will aways be less that 100! + */ +void machine_real_restart(unsigned char *code, int length) +{ + unsigned long flags; + + local_irq_disable(); + + /* Write zero to CMOS register number 0x0f, which the BIOS POST + routine will recognize as telling it to do a proper reboot. (Well + that's what this book in front of me says -- it may only apply to + the Phoenix BIOS though, it's not clear). At the same time, + disable NMIs by setting the top bit in the CMOS address register, + as we're about to do peculiar things to the CPU. I'm not sure if + `outb_p' is needed instead of just `outb'. Use it to be on the + safe side. (Yes, CMOS_WRITE does outb_p's. - Paul G.) + */ + + spin_lock_irqsave(&rtc_lock, flags); + CMOS_WRITE(0x00, 0x8f); + spin_unlock_irqrestore(&rtc_lock, flags); + + /* Remap the kernel at virtual address zero, as well as offset zero + from the kernel segment. This assumes the kernel segment starts at + virtual address PAGE_OFFSET. */ + + memcpy (swapper_pg_dir, swapper_pg_dir + USER_PGD_PTRS, + sizeof (swapper_pg_dir [0]) * KERNEL_PGD_PTRS); + + /* + * Use `swapper_pg_dir' as our page directory. + */ + load_cr3(swapper_pg_dir); + + /* Write 0x1234 to absolute memory location 0x472. The BIOS reads + this on booting to tell it to "Bypass memory test (also warm + boot)". This seems like a fairly standard thing that gets set by + REBOOT.COM programs, and the previous reset routine did this + too. */ + + *((unsigned short *)0x472) = reboot_mode; + + /* For the switch to real mode, copy some code to low memory. It has + to be in the first 64k because it is running in 16-bit mode, and it + has to have the same physical and virtual address, because it turns + off paging. Copy it near the end of the first page, out of the way + of BIOS variables. */ + + memcpy ((void *) (0x1000 - sizeof (real_mode_switch) - 100), + real_mode_switch, sizeof (real_mode_switch)); + memcpy ((void *) (0x1000 - 100), code, length); + + /* Set up the IDT for real mode. */ + + __asm__ __volatile__ ("lidt %0" : : "m" (real_mode_idt)); + + /* Set up a GDT from which we can load segment descriptors for real + mode. The GDT is not used in real mode; it is just needed here to + prepare the descriptors. */ + + __asm__ __volatile__ ("lgdt %0" : : "m" (real_mode_gdt)); + + /* Load the data segment registers, and thus the descriptors ready for + real mode. The base address of each segment is 0x100, 16 times the + selector value being loaded here. This is so that the segment + registers don't have to be reloaded after switching to real mode: + the values are consistent for real mode operation already. */ + + __asm__ __volatile__ ("movl $0x0010,%%eax\n" + "\tmovl %%eax,%%ds\n" + "\tmovl %%eax,%%es\n" + "\tmovl %%eax,%%fs\n" + "\tmovl %%eax,%%gs\n" + "\tmovl %%eax,%%ss" : : : "eax"); + + /* Jump to the 16-bit code that we copied earlier. It disables paging + and the cache, switches to real mode, and jumps to the BIOS reset + entry point. */ + + __asm__ __volatile__ ("ljmp $0x0008,%0" + : + : "i" ((void *) (0x1000 - sizeof (real_mode_switch) - 100))); +} + +void machine_restart(char * __unused) +{ +#ifdef CONFIG_SMP + int cpuid; + + cpuid = GET_APIC_ID(apic_read(APIC_ID)); + + if (reboot_smp) { + + /* check to see if reboot_cpu is valid + if its not, default to the BSP */ + if ((reboot_cpu == -1) || + (reboot_cpu > (NR_CPUS -1)) || + !physid_isset(cpuid, phys_cpu_present_map)) + reboot_cpu = boot_cpu_physical_apicid; + + reboot_smp = 0; /* use this as a flag to only go through this once*/ + /* re-run this function on the other CPUs + it will fall though this section since we have + cleared reboot_smp, and do the reboot if it is the + correct CPU, otherwise it halts. */ + if (reboot_cpu != cpuid) + smp_call_function((void *)machine_restart , NULL, 1, 0); + } + + /* if reboot_cpu is still -1, then we want a tradional reboot, + and if we are not running on the reboot_cpu,, halt */ + if ((reboot_cpu != -1) && (cpuid != reboot_cpu)) { + for (;;) + __asm__ __volatile__ ("hlt"); + } + /* + * Stop all CPUs and turn off local APICs and the IO-APIC, so + * other OSs see a clean IRQ state. + */ + smp_send_stop(); +#endif /* CONFIG_SMP */ + + lapic_shutdown(); + +#ifdef CONFIG_X86_IO_APIC + disable_IO_APIC(); +#endif + + if (!reboot_thru_bios) { + if (efi_enabled) { + efi.reset_system(EFI_RESET_COLD, EFI_SUCCESS, 0, NULL); + __asm__ __volatile__("lidt %0": :"m" (no_idt)); + __asm__ __volatile__("int3"); + } + /* rebooting needs to touch the page at absolute addr 0 */ + *((unsigned short *)__va(0x472)) = reboot_mode; + for (;;) { + mach_reboot(); + /* That didn't work - force a triple fault.. */ + __asm__ __volatile__("lidt %0": :"m" (no_idt)); + __asm__ __volatile__("int3"); + } + } + if (efi_enabled) + efi.reset_system(EFI_RESET_WARM, EFI_SUCCESS, 0, NULL); + + machine_real_restart(jump_to_bios, sizeof(jump_to_bios)); +} + +EXPORT_SYMBOL(machine_restart); + +void machine_halt(void) +{ +} + +EXPORT_SYMBOL(machine_halt); + +void machine_power_off(void) +{ + lapic_shutdown(); + + if (efi_enabled) + efi.reset_system(EFI_RESET_SHUTDOWN, EFI_SUCCESS, 0, NULL); + if (pm_power_off) + pm_power_off(); +} + +EXPORT_SYMBOL(machine_power_off); + diff --git a/arch/i386/kernel/scx200.c b/arch/i386/kernel/scx200.c new file mode 100644 index 00000000000..69e203a0d33 --- /dev/null +++ b/arch/i386/kernel/scx200.c @@ -0,0 +1,167 @@ +/* linux/arch/i386/kernel/scx200.c + + Copyright (c) 2001,2002 Christer Weinigel <wingel@nano-system.com> + + National Semiconductor SCx200 support. */ + +#include <linux/config.h> +#include <linux/module.h> +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/pci.h> + +#include <linux/scx200.h> + +/* Verify that the configuration block really is there */ +#define scx200_cb_probe(base) (inw((base) + SCx200_CBA) == (base)) + +#define NAME "scx200" + +MODULE_AUTHOR("Christer Weinigel <wingel@nano-system.com>"); +MODULE_DESCRIPTION("NatSemi SCx200 Driver"); +MODULE_LICENSE("GPL"); + +unsigned scx200_gpio_base = 0; +long scx200_gpio_shadow[2]; + +unsigned scx200_cb_base = 0; + +static struct pci_device_id scx200_tbl[] = { + { PCI_DEVICE(PCI_VENDOR_ID_NS, PCI_DEVICE_ID_NS_SCx200_BRIDGE) }, + { PCI_DEVICE(PCI_VENDOR_ID_NS, PCI_DEVICE_ID_NS_SC1100_BRIDGE) }, + { PCI_DEVICE(PCI_VENDOR_ID_NS, PCI_DEVICE_ID_NS_SCx200_XBUS) }, + { PCI_DEVICE(PCI_VENDOR_ID_NS, PCI_DEVICE_ID_NS_SC1100_XBUS) }, + { }, +}; +MODULE_DEVICE_TABLE(pci,scx200_tbl); + +static int __devinit scx200_probe(struct pci_dev *, const struct pci_device_id *); + +static struct pci_driver scx200_pci_driver = { + .name = "scx200", + .id_table = scx200_tbl, + .probe = scx200_probe, +}; + +static DEFINE_SPINLOCK(scx200_gpio_config_lock); + +static int __devinit scx200_probe(struct pci_dev *pdev, const struct pci_device_id *ent) +{ + int bank; + unsigned base; + + if (pdev->device == PCI_DEVICE_ID_NS_SCx200_BRIDGE || + pdev->device == PCI_DEVICE_ID_NS_SC1100_BRIDGE) { + base = pci_resource_start(pdev, 0); + printk(KERN_INFO NAME ": GPIO base 0x%x\n", base); + + if (request_region(base, SCx200_GPIO_SIZE, "NatSemi SCx200 GPIO") == 0) { + printk(KERN_ERR NAME ": can't allocate I/O for GPIOs\n"); + return -EBUSY; + } + + scx200_gpio_base = base; + + /* read the current values driven on the GPIO signals */ + for (bank = 0; bank < 2; ++bank) + scx200_gpio_shadow[bank] = inl(scx200_gpio_base + 0x10 * bank); + + } else { + /* find the base of the Configuration Block */ + if (scx200_cb_probe(SCx200_CB_BASE_FIXED)) { + scx200_cb_base = SCx200_CB_BASE_FIXED; + } else { + pci_read_config_dword(pdev, SCx200_CBA_SCRATCH, &base); + if (scx200_cb_probe(base)) { + scx200_cb_base = base; + } else { + printk(KERN_WARNING NAME ": Configuration Block not found\n"); + return -ENODEV; + } + } + printk(KERN_INFO NAME ": Configuration Block base 0x%x\n", scx200_cb_base); + } + + return 0; +} + +u32 scx200_gpio_configure(int index, u32 mask, u32 bits) +{ + u32 config, new_config; + unsigned long flags; + + spin_lock_irqsave(&scx200_gpio_config_lock, flags); + + outl(index, scx200_gpio_base + 0x20); + config = inl(scx200_gpio_base + 0x24); + + new_config = (config & mask) | bits; + outl(new_config, scx200_gpio_base + 0x24); + + spin_unlock_irqrestore(&scx200_gpio_config_lock, flags); + + return config; +} + +#if 0 +void scx200_gpio_dump(unsigned index) +{ + u32 config = scx200_gpio_configure(index, ~0, 0); + printk(KERN_DEBUG "GPIO%02u: 0x%08lx", index, (unsigned long)config); + + if (config & 1) + printk(" OE"); /* output enabled */ + else + printk(" TS"); /* tristate */ + if (config & 2) + printk(" PP"); /* push pull */ + else + printk(" OD"); /* open drain */ + if (config & 4) + printk(" PUE"); /* pull up enabled */ + else + printk(" PUD"); /* pull up disabled */ + if (config & 8) + printk(" LOCKED"); /* locked */ + if (config & 16) + printk(" LEVEL"); /* level input */ + else + printk(" EDGE"); /* edge input */ + if (config & 32) + printk(" HI"); /* trigger on rising edge */ + else + printk(" LO"); /* trigger on falling edge */ + if (config & 64) + printk(" DEBOUNCE"); /* debounce */ + printk("\n"); +} +#endif /* 0 */ + +static int __init scx200_init(void) +{ + printk(KERN_INFO NAME ": NatSemi SCx200 Driver\n"); + + return pci_module_init(&scx200_pci_driver); +} + +static void __exit scx200_cleanup(void) +{ + pci_unregister_driver(&scx200_pci_driver); + release_region(scx200_gpio_base, SCx200_GPIO_SIZE); +} + +module_init(scx200_init); +module_exit(scx200_cleanup); + +EXPORT_SYMBOL(scx200_gpio_base); +EXPORT_SYMBOL(scx200_gpio_shadow); +EXPORT_SYMBOL(scx200_gpio_configure); +EXPORT_SYMBOL(scx200_cb_base); + +/* + Local variables: + compile-command: "make -k -C ../../.. SUBDIRS=arch/i386/kernel modules" + c-basic-offset: 8 + End: +*/ diff --git a/arch/i386/kernel/semaphore.c b/arch/i386/kernel/semaphore.c new file mode 100644 index 00000000000..469f496e55c --- /dev/null +++ b/arch/i386/kernel/semaphore.c @@ -0,0 +1,297 @@ +/* + * i386 semaphore implementation. + * + * (C) Copyright 1999 Linus Torvalds + * + * Portions Copyright 1999 Red Hat, Inc. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation; either version + * 2 of the License, or (at your option) any later version. + * + * rw semaphores implemented November 1999 by Benjamin LaHaise <bcrl@kvack.org> + */ +#include <linux/config.h> +#include <linux/sched.h> +#include <linux/err.h> +#include <linux/init.h> +#include <asm/semaphore.h> + +/* + * Semaphores are implemented using a two-way counter: + * The "count" variable is decremented for each process + * that tries to acquire the semaphore, while the "sleeping" + * variable is a count of such acquires. + * + * Notably, the inline "up()" and "down()" functions can + * efficiently test if they need to do any extra work (up + * needs to do something only if count was negative before + * the increment operation. + * + * "sleeping" and the contention routine ordering is protected + * by the spinlock in the semaphore's waitqueue head. + * + * Note that these functions are only called when there is + * contention on the lock, and as such all this is the + * "non-critical" part of the whole semaphore business. The + * critical part is the inline stuff in <asm/semaphore.h> + * where we want to avoid any extra jumps and calls. + */ + +/* + * Logic: + * - only on a boundary condition do we need to care. When we go + * from a negative count to a non-negative, we wake people up. + * - when we go from a non-negative count to a negative do we + * (a) synchronize with the "sleeper" count and (b) make sure + * that we're on the wakeup list before we synchronize so that + * we cannot lose wakeup events. + */ + +static fastcall void __attribute_used__ __up(struct semaphore *sem) +{ + wake_up(&sem->wait); +} + +static fastcall void __attribute_used__ __sched __down(struct semaphore * sem) +{ + struct task_struct *tsk = current; + DECLARE_WAITQUEUE(wait, tsk); + unsigned long flags; + + tsk->state = TASK_UNINTERRUPTIBLE; + spin_lock_irqsave(&sem->wait.lock, flags); + add_wait_queue_exclusive_locked(&sem->wait, &wait); + + sem->sleepers++; + for (;;) { + int sleepers = sem->sleepers; + + /* + * Add "everybody else" into it. They aren't + * playing, because we own the spinlock in + * the wait_queue_head. + */ + if (!atomic_add_negative(sleepers - 1, &sem->count)) { + sem->sleepers = 0; + break; + } + sem->sleepers = 1; /* us - see -1 above */ + spin_unlock_irqrestore(&sem->wait.lock, flags); + + schedule(); + + spin_lock_irqsave(&sem->wait.lock, flags); + tsk->state = TASK_UNINTERRUPTIBLE; + } + remove_wait_queue_locked(&sem->wait, &wait); + wake_up_locked(&sem->wait); + spin_unlock_irqrestore(&sem->wait.lock, flags); + tsk->state = TASK_RUNNING; +} + +static fastcall int __attribute_used__ __sched __down_interruptible(struct semaphore * sem) +{ + int retval = 0; + struct task_struct *tsk = current; + DECLARE_WAITQUEUE(wait, tsk); + unsigned long flags; + + tsk->state = TASK_INTERRUPTIBLE; + spin_lock_irqsave(&sem->wait.lock, flags); + add_wait_queue_exclusive_locked(&sem->wait, &wait); + + sem->sleepers++; + for (;;) { + int sleepers = sem->sleepers; + + /* + * With signals pending, this turns into + * the trylock failure case - we won't be + * sleeping, and we* can't get the lock as + * it has contention. Just correct the count + * and exit. + */ + if (signal_pending(current)) { + retval = -EINTR; + sem->sleepers = 0; + atomic_add(sleepers, &sem->count); + break; + } + + /* + * Add "everybody else" into it. They aren't + * playing, because we own the spinlock in + * wait_queue_head. The "-1" is because we're + * still hoping to get the semaphore. + */ + if (!atomic_add_negative(sleepers - 1, &sem->count)) { + sem->sleepers = 0; + break; + } + sem->sleepers = 1; /* us - see -1 above */ + spin_unlock_irqrestore(&sem->wait.lock, flags); + + schedule(); + + spin_lock_irqsave(&sem->wait.lock, flags); + tsk->state = TASK_INTERRUPTIBLE; + } + remove_wait_queue_locked(&sem->wait, &wait); + wake_up_locked(&sem->wait); + spin_unlock_irqrestore(&sem->wait.lock, flags); + + tsk->state = TASK_RUNNING; + return retval; +} + +/* + * Trylock failed - make sure we correct for + * having decremented the count. + * + * We could have done the trylock with a + * single "cmpxchg" without failure cases, + * but then it wouldn't work on a 386. + */ +static fastcall int __attribute_used__ __down_trylock(struct semaphore * sem) +{ + int sleepers; + unsigned long flags; + + spin_lock_irqsave(&sem->wait.lock, flags); + sleepers = sem->sleepers + 1; + sem->sleepers = 0; + + /* + * Add "everybody else" and us into it. They aren't + * playing, because we own the spinlock in the + * wait_queue_head. + */ + if (!atomic_add_negative(sleepers, &sem->count)) { + wake_up_locked(&sem->wait); + } + + spin_unlock_irqrestore(&sem->wait.lock, flags); + return 1; +} + + +/* + * The semaphore operations have a special calling sequence that + * allow us to do a simpler in-line version of them. These routines + * need to convert that sequence back into the C sequence when + * there is contention on the semaphore. + * + * %eax contains the semaphore pointer on entry. Save the C-clobbered + * registers (%eax, %edx and %ecx) except %eax whish is either a return + * value or just clobbered.. + */ +asm( +".section .sched.text\n" +".align 4\n" +".globl __down_failed\n" +"__down_failed:\n\t" +#if defined(CONFIG_FRAME_POINTER) + "pushl %ebp\n\t" + "movl %esp,%ebp\n\t" +#endif + "pushl %edx\n\t" + "pushl %ecx\n\t" + "call __down\n\t" + "popl %ecx\n\t" + "popl %edx\n\t" +#if defined(CONFIG_FRAME_POINTER) + "movl %ebp,%esp\n\t" + "popl %ebp\n\t" +#endif + "ret" +); + +asm( +".section .sched.text\n" +".align 4\n" +".globl __down_failed_interruptible\n" +"__down_failed_interruptible:\n\t" +#if defined(CONFIG_FRAME_POINTER) + "pushl %ebp\n\t" + "movl %esp,%ebp\n\t" +#endif + "pushl %edx\n\t" + "pushl %ecx\n\t" + "call __down_interruptible\n\t" + "popl %ecx\n\t" + "popl %edx\n\t" +#if defined(CONFIG_FRAME_POINTER) + "movl %ebp,%esp\n\t" + "popl %ebp\n\t" +#endif + "ret" +); + +asm( +".section .sched.text\n" +".align 4\n" +".globl __down_failed_trylock\n" +"__down_failed_trylock:\n\t" +#if defined(CONFIG_FRAME_POINTER) + "pushl %ebp\n\t" + "movl %esp,%ebp\n\t" +#endif + "pushl %edx\n\t" + "pushl %ecx\n\t" + "call __down_trylock\n\t" + "popl %ecx\n\t" + "popl %edx\n\t" +#if defined(CONFIG_FRAME_POINTER) + "movl %ebp,%esp\n\t" + "popl %ebp\n\t" +#endif + "ret" +); + +asm( +".section .sched.text\n" +".align 4\n" +".globl __up_wakeup\n" +"__up_wakeup:\n\t" + "pushl %edx\n\t" + "pushl %ecx\n\t" + "call __up\n\t" + "popl %ecx\n\t" + "popl %edx\n\t" + "ret" +); + +/* + * rw spinlock fallbacks + */ +#if defined(CONFIG_SMP) +asm( +".section .sched.text\n" +".align 4\n" +".globl __write_lock_failed\n" +"__write_lock_failed:\n\t" + LOCK "addl $" RW_LOCK_BIAS_STR ",(%eax)\n" +"1: rep; nop\n\t" + "cmpl $" RW_LOCK_BIAS_STR ",(%eax)\n\t" + "jne 1b\n\t" + LOCK "subl $" RW_LOCK_BIAS_STR ",(%eax)\n\t" + "jnz __write_lock_failed\n\t" + "ret" +); + +asm( +".section .sched.text\n" +".align 4\n" +".globl __read_lock_failed\n" +"__read_lock_failed:\n\t" + LOCK "incl (%eax)\n" +"1: rep; nop\n\t" + "cmpl $1,(%eax)\n\t" + "js 1b\n\t" + LOCK "decl (%eax)\n\t" + "js __read_lock_failed\n\t" + "ret" +); +#endif diff --git a/arch/i386/kernel/setup.c b/arch/i386/kernel/setup.c new file mode 100644 index 00000000000..945ec73163c --- /dev/null +++ b/arch/i386/kernel/setup.c @@ -0,0 +1,1535 @@ +/* + * linux/arch/i386/kernel/setup.c + * + * Copyright (C) 1995 Linus Torvalds + * + * Support of BIGMEM added by Gerhard Wichert, Siemens AG, July 1999 + * + * Memory region support + * David Parsons <orc@pell.chi.il.us>, July-August 1999 + * + * Added E820 sanitization routine (removes overlapping memory regions); + * Brian Moyle <bmoyle@mvista.com>, February 2001 + * + * Moved CPU detection code to cpu/${cpu}.c + * Patrick Mochel <mochel@osdl.org>, March 2002 + * + * Provisions for empty E820 memory regions (reported by certain BIOSes). + * Alex Achenbach <xela@slit.de>, December 2002. + * + */ + +/* + * This file handles the architecture-dependent parts of initialization + */ + +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/tty.h> +#include <linux/ioport.h> +#include <linux/acpi.h> +#include <linux/apm_bios.h> +#include <linux/initrd.h> +#include <linux/bootmem.h> +#include <linux/seq_file.h> +#include <linux/console.h> +#include <linux/mca.h> +#include <linux/root_dev.h> +#include <linux/highmem.h> +#include <linux/module.h> +#include <linux/efi.h> +#include <linux/init.h> +#include <linux/edd.h> +#include <linux/nodemask.h> +#include <video/edid.h> +#include <asm/e820.h> +#include <asm/mpspec.h> +#include <asm/setup.h> +#include <asm/arch_hooks.h> +#include <asm/sections.h> +#include <asm/io_apic.h> +#include <asm/ist.h> +#include <asm/io.h> +#include "setup_arch_pre.h" +#include <bios_ebda.h> + +/* This value is set up by the early boot code to point to the value + immediately after the boot time page tables. It contains a *physical* + address, and must not be in the .bss segment! */ +unsigned long init_pg_tables_end __initdata = ~0UL; + +int disable_pse __initdata = 0; + +/* + * Machine setup.. + */ + +#ifdef CONFIG_EFI +int efi_enabled = 0; +EXPORT_SYMBOL(efi_enabled); +#endif + +/* cpu data as detected by the assembly code in head.S */ +struct cpuinfo_x86 new_cpu_data __initdata = { 0, 0, 0, 0, -1, 1, 0, 0, -1 }; +/* common cpu data for all cpus */ +struct cpuinfo_x86 boot_cpu_data = { 0, 0, 0, 0, -1, 1, 0, 0, -1 }; + +unsigned long mmu_cr4_features; + +#ifdef CONFIG_ACPI_INTERPRETER + int acpi_disabled = 0; +#else + int acpi_disabled = 1; +#endif +EXPORT_SYMBOL(acpi_disabled); + +#ifdef CONFIG_ACPI_BOOT +int __initdata acpi_force = 0; +extern acpi_interrupt_flags acpi_sci_flags; +#endif + +/* for MCA, but anyone else can use it if they want */ +unsigned int machine_id; +unsigned int machine_submodel_id; +unsigned int BIOS_revision; +unsigned int mca_pentium_flag; + +/* For PCI or other memory-mapped resources */ +unsigned long pci_mem_start = 0x10000000; + +/* Boot loader ID as an integer, for the benefit of proc_dointvec */ +int bootloader_type; + +/* user-defined highmem size */ +static unsigned int highmem_pages = -1; + +/* + * Setup options + */ +struct drive_info_struct { char dummy[32]; } drive_info; +struct screen_info screen_info; +struct apm_info apm_info; +struct sys_desc_table_struct { + unsigned short length; + unsigned char table[0]; +}; +struct edid_info edid_info; +struct ist_info ist_info; +struct e820map e820; + +extern void early_cpu_init(void); +extern void dmi_scan_machine(void); +extern void generic_apic_probe(char *); +extern int root_mountflags; + +unsigned long saved_videomode; + +#define RAMDISK_IMAGE_START_MASK 0x07FF +#define RAMDISK_PROMPT_FLAG 0x8000 +#define RAMDISK_LOAD_FLAG 0x4000 + +static char command_line[COMMAND_LINE_SIZE]; + +unsigned char __initdata boot_params[PARAM_SIZE]; + +static struct resource data_resource = { + .name = "Kernel data", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_MEM +}; + +static struct resource code_resource = { + .name = "Kernel code", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_MEM +}; + +static struct resource system_rom_resource = { + .name = "System ROM", + .start = 0xf0000, + .end = 0xfffff, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}; + +static struct resource extension_rom_resource = { + .name = "Extension ROM", + .start = 0xe0000, + .end = 0xeffff, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}; + +static struct resource adapter_rom_resources[] = { { + .name = "Adapter ROM", + .start = 0xc8000, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}, { + .name = "Adapter ROM", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}, { + .name = "Adapter ROM", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}, { + .name = "Adapter ROM", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}, { + .name = "Adapter ROM", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}, { + .name = "Adapter ROM", + .start = 0, + .end = 0, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +} }; + +#define ADAPTER_ROM_RESOURCES \ + (sizeof adapter_rom_resources / sizeof adapter_rom_resources[0]) + +static struct resource video_rom_resource = { + .name = "Video ROM", + .start = 0xc0000, + .end = 0xc7fff, + .flags = IORESOURCE_BUSY | IORESOURCE_READONLY | IORESOURCE_MEM +}; + +static struct resource video_ram_resource = { + .name = "Video RAM area", + .start = 0xa0000, + .end = 0xbffff, + .flags = IORESOURCE_BUSY | IORESOURCE_MEM +}; + +static struct resource standard_io_resources[] = { { + .name = "dma1", + .start = 0x0000, + .end = 0x001f, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "pic1", + .start = 0x0020, + .end = 0x0021, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "timer0", + .start = 0x0040, + .end = 0x0043, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "timer1", + .start = 0x0050, + .end = 0x0053, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "keyboard", + .start = 0x0060, + .end = 0x006f, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "dma page reg", + .start = 0x0080, + .end = 0x008f, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "pic2", + .start = 0x00a0, + .end = 0x00a1, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "dma2", + .start = 0x00c0, + .end = 0x00df, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +}, { + .name = "fpu", + .start = 0x00f0, + .end = 0x00ff, + .flags = IORESOURCE_BUSY | IORESOURCE_IO +} }; + +#define STANDARD_IO_RESOURCES \ + (sizeof standard_io_resources / sizeof standard_io_resources[0]) + +#define romsignature(x) (*(unsigned short *)(x) == 0xaa55) + +static int __init romchecksum(unsigned char *rom, unsigned long length) +{ + unsigned char *p, sum = 0; + + for (p = rom; p < rom + length; p++) + sum += *p; + return sum == 0; +} + +static void __init probe_roms(void) +{ + unsigned long start, length, upper; + unsigned char *rom; + int i; + + /* video rom */ + upper = adapter_rom_resources[0].start; + for (start = video_rom_resource.start; start < upper; start += 2048) { + rom = isa_bus_to_virt(start); + if (!romsignature(rom)) + continue; + + video_rom_resource.start = start; + + /* 0 < length <= 0x7f * 512, historically */ + length = rom[2] * 512; + + /* if checksum okay, trust length byte */ + if (length && romchecksum(rom, length)) + video_rom_resource.end = start + length - 1; + + request_resource(&iomem_resource, &video_rom_resource); + break; + } + + start = (video_rom_resource.end + 1 + 2047) & ~2047UL; + if (start < upper) + start = upper; + + /* system rom */ + request_resource(&iomem_resource, &system_rom_resource); + upper = system_rom_resource.start; + + /* check for extension rom (ignore length byte!) */ + rom = isa_bus_to_virt(extension_rom_resource.start); + if (romsignature(rom)) { + length = extension_rom_resource.end - extension_rom_resource.start + 1; + if (romchecksum(rom, length)) { + request_resource(&iomem_resource, &extension_rom_resource); + upper = extension_rom_resource.start; + } + } + + /* check for adapter roms on 2k boundaries */ + for (i = 0; i < ADAPTER_ROM_RESOURCES && start < upper; start += 2048) { + rom = isa_bus_to_virt(start); + if (!romsignature(rom)) + continue; + + /* 0 < length <= 0x7f * 512, historically */ + length = rom[2] * 512; + + /* but accept any length that fits if checksum okay */ + if (!length || start + length > upper || !romchecksum(rom, length)) + continue; + + adapter_rom_resources[i].start = start; + adapter_rom_resources[i].end = start + length - 1; + request_resource(&iomem_resource, &adapter_rom_resources[i]); + + start = adapter_rom_resources[i++].end & ~2047UL; + } +} + +static void __init limit_regions(unsigned long long size) +{ + unsigned long long current_addr = 0; + int i; + + if (efi_enabled) { + for (i = 0; i < memmap.nr_map; i++) { + current_addr = memmap.map[i].phys_addr + + (memmap.map[i].num_pages << 12); + if (memmap.map[i].type == EFI_CONVENTIONAL_MEMORY) { + if (current_addr >= size) { + memmap.map[i].num_pages -= + (((current_addr-size) + PAGE_SIZE-1) >> PAGE_SHIFT); + memmap.nr_map = i + 1; + return; + } + } + } + } + for (i = 0; i < e820.nr_map; i++) { + if (e820.map[i].type == E820_RAM) { + current_addr = e820.map[i].addr + e820.map[i].size; + if (current_addr >= size) { + e820.map[i].size -= current_addr-size; + e820.nr_map = i + 1; + return; + } + } + } +} + +static void __init add_memory_region(unsigned long long start, + unsigned long long size, int type) +{ + int x; + + if (!efi_enabled) { + x = e820.nr_map; + + if (x == E820MAX) { + printk(KERN_ERR "Ooops! Too many entries in the memory map!\n"); + return; + } + + e820.map[x].addr = start; + e820.map[x].size = size; + e820.map[x].type = type; + e820.nr_map++; + } +} /* add_memory_region */ + +#define E820_DEBUG 1 + +static void __init print_memory_map(char *who) +{ + int i; + + for (i = 0; i < e820.nr_map; i++) { + printk(" %s: %016Lx - %016Lx ", who, + e820.map[i].addr, + e820.map[i].addr + e820.map[i].size); + switch (e820.map[i].type) { + case E820_RAM: printk("(usable)\n"); + break; + case E820_RESERVED: + printk("(reserved)\n"); + break; + case E820_ACPI: + printk("(ACPI data)\n"); + break; + case E820_NVS: + printk("(ACPI NVS)\n"); + break; + default: printk("type %lu\n", e820.map[i].type); + break; + } + } +} + +/* + * Sanitize the BIOS e820 map. + * + * Some e820 responses include overlapping entries. The following + * replaces the original e820 map with a new one, removing overlaps. + * + */ +struct change_member { + struct e820entry *pbios; /* pointer to original bios entry */ + unsigned long long addr; /* address for this change point */ +}; +static struct change_member change_point_list[2*E820MAX] __initdata; +static struct change_member *change_point[2*E820MAX] __initdata; +static struct e820entry *overlap_list[E820MAX] __initdata; +static struct e820entry new_bios[E820MAX] __initdata; + +static int __init sanitize_e820_map(struct e820entry * biosmap, char * pnr_map) +{ + struct change_member *change_tmp; + unsigned long current_type, last_type; + unsigned long long last_addr; + int chgidx, still_changing; + int overlap_entries; + int new_bios_entry; + int old_nr, new_nr, chg_nr; + int i; + + /* + Visually we're performing the following (1,2,3,4 = memory types)... + + Sample memory map (w/overlaps): + ____22__________________ + ______________________4_ + ____1111________________ + _44_____________________ + 11111111________________ + ____________________33__ + ___________44___________ + __________33333_________ + ______________22________ + ___________________2222_ + _________111111111______ + _____________________11_ + _________________4______ + + Sanitized equivalent (no overlap): + 1_______________________ + _44_____________________ + ___1____________________ + ____22__________________ + ______11________________ + _________1______________ + __________3_____________ + ___________44___________ + _____________33_________ + _______________2________ + ________________1_______ + _________________4______ + ___________________2____ + ____________________33__ + ______________________4_ + */ + + /* if there's only one memory region, don't bother */ + if (*pnr_map < 2) + return -1; + + old_nr = *pnr_map; + + /* bail out if we find any unreasonable addresses in bios map */ + for (i=0; i<old_nr; i++) + if (biosmap[i].addr + biosmap[i].size < biosmap[i].addr) + return -1; + + /* create pointers for initial change-point information (for sorting) */ + for (i=0; i < 2*old_nr; i++) + change_point[i] = &change_point_list[i]; + + /* record all known change-points (starting and ending addresses), + omitting those that are for empty memory regions */ + chgidx = 0; + for (i=0; i < old_nr; i++) { + if (biosmap[i].size != 0) { + change_point[chgidx]->addr = biosmap[i].addr; + change_point[chgidx++]->pbios = &biosmap[i]; + change_point[chgidx]->addr = biosmap[i].addr + biosmap[i].size; + change_point[chgidx++]->pbios = &biosmap[i]; + } + } + chg_nr = chgidx; /* true number of change-points */ + + /* sort change-point list by memory addresses (low -> high) */ + still_changing = 1; + while (still_changing) { + still_changing = 0; + for (i=1; i < chg_nr; i++) { + /* if <current_addr> > <last_addr>, swap */ + /* or, if current=<start_addr> & last=<end_addr>, swap */ + if ((change_point[i]->addr < change_point[i-1]->addr) || + ((change_point[i]->addr == change_point[i-1]->addr) && + (change_point[i]->addr == change_point[i]->pbios->addr) && + (change_point[i-1]->addr != change_point[i-1]->pbios->addr)) + ) + { + change_tmp = change_point[i]; + change_point[i] = change_point[i-1]; + change_point[i-1] = change_tmp; + still_changing=1; + } + } + } + + /* create a new bios memory map, removing overlaps */ + overlap_entries=0; /* number of entries in the overlap table */ + new_bios_entry=0; /* index for creating new bios map entries */ + last_type = 0; /* start with undefined memory type */ + last_addr = 0; /* start with 0 as last starting address */ + /* loop through change-points, determining affect on the new bios map */ + for (chgidx=0; chgidx < chg_nr; chgidx++) + { + /* keep track of all overlapping bios entries */ + if (change_point[chgidx]->addr == change_point[chgidx]->pbios->addr) + { + /* add map entry to overlap list (> 1 entry implies an overlap) */ + overlap_list[overlap_entries++]=change_point[chgidx]->pbios; + } + else + { + /* remove entry from list (order independent, so swap with last) */ + for (i=0; i<overlap_entries; i++) + { + if (overlap_list[i] == change_point[chgidx]->pbios) + overlap_list[i] = overlap_list[overlap_entries-1]; + } + overlap_entries--; + } + /* if there are overlapping entries, decide which "type" to use */ + /* (larger value takes precedence -- 1=usable, 2,3,4,4+=unusable) */ + current_type = 0; + for (i=0; i<overlap_entries; i++) + if (overlap_list[i]->type > current_type) + current_type = overlap_list[i]->type; + /* continue building up new bios map based on this information */ + if (current_type != last_type) { + if (last_type != 0) { + new_bios[new_bios_entry].size = + change_point[chgidx]->addr - last_addr; + /* move forward only if the new size was non-zero */ + if (new_bios[new_bios_entry].size != 0) + if (++new_bios_entry >= E820MAX) + break; /* no more space left for new bios entries */ + } + if (current_type != 0) { + new_bios[new_bios_entry].addr = change_point[chgidx]->addr; + new_bios[new_bios_entry].type = current_type; + last_addr=change_point[chgidx]->addr; + } + last_type = current_type; + } + } + new_nr = new_bios_entry; /* retain count for new bios entries */ + + /* copy new bios mapping into original location */ + memcpy(biosmap, new_bios, new_nr*sizeof(struct e820entry)); + *pnr_map = new_nr; + + return 0; +} + +/* + * Copy the BIOS e820 map into a safe place. + * + * Sanity-check it while we're at it.. + * + * If we're lucky and live on a modern system, the setup code + * will have given us a memory map that we can use to properly + * set up memory. If we aren't, we'll fake a memory map. + * + * We check to see that the memory map contains at least 2 elements + * before we'll use it, because the detection code in setup.S may + * not be perfect and most every PC known to man has two memory + * regions: one from 0 to 640k, and one from 1mb up. (The IBM + * thinkpad 560x, for example, does not cooperate with the memory + * detection code.) + */ +static int __init copy_e820_map(struct e820entry * biosmap, int nr_map) +{ + /* Only one memory region (or negative)? Ignore it */ + if (nr_map < 2) + return -1; + + do { + unsigned long long start = biosmap->addr; + unsigned long long size = biosmap->size; + unsigned long long end = start + size; + unsigned long type = biosmap->type; + + /* Overflow in 64 bits? Ignore the memory map. */ + if (start > end) + return -1; + + /* + * Some BIOSes claim RAM in the 640k - 1M region. + * Not right. Fix it up. + */ + if (type == E820_RAM) { + if (start < 0x100000ULL && end > 0xA0000ULL) { + if (start < 0xA0000ULL) + add_memory_region(start, 0xA0000ULL-start, type); + if (end <= 0x100000ULL) + continue; + start = 0x100000ULL; + size = end - start; + } + } + add_memory_region(start, size, type); + } while (biosmap++,--nr_map); + return 0; +} + +#if defined(CONFIG_EDD) || defined(CONFIG_EDD_MODULE) +struct edd edd; +#ifdef CONFIG_EDD_MODULE +EXPORT_SYMBOL(edd); +#endif +/** + * copy_edd() - Copy the BIOS EDD information + * from boot_params into a safe place. + * + */ +static inline void copy_edd(void) +{ + memcpy(edd.mbr_signature, EDD_MBR_SIGNATURE, sizeof(edd.mbr_signature)); + memcpy(edd.edd_info, EDD_BUF, sizeof(edd.edd_info)); + edd.mbr_signature_nr = EDD_MBR_SIG_NR; + edd.edd_info_nr = EDD_NR; +} +#else +static inline void copy_edd(void) +{ +} +#endif + +/* + * Do NOT EVER look at the BIOS memory size location. + * It does not work on many machines. + */ +#define LOWMEMSIZE() (0x9f000) + +static void __init parse_cmdline_early (char ** cmdline_p) +{ + char c = ' ', *to = command_line, *from = saved_command_line; + int len = 0; + int userdef = 0; + + /* Save unparsed command line copy for /proc/cmdline */ + saved_command_line[COMMAND_LINE_SIZE-1] = '\0'; + + for (;;) { + if (c != ' ') + goto next_char; + /* + * "mem=nopentium" disables the 4MB page tables. + * "mem=XXX[kKmM]" defines a memory region from HIGH_MEM + * to <mem>, overriding the bios size. + * "memmap=XXX[KkmM]@XXX[KkmM]" defines a memory region from + * <start> to <start>+<mem>, overriding the bios size. + * + * HPA tells me bootloaders need to parse mem=, so no new + * option should be mem= [also see Documentation/i386/boot.txt] + */ + if (!memcmp(from, "mem=", 4)) { + if (to != command_line) + to--; + if (!memcmp(from+4, "nopentium", 9)) { + from += 9+4; + clear_bit(X86_FEATURE_PSE, boot_cpu_data.x86_capability); + disable_pse = 1; + } else { + /* If the user specifies memory size, we + * limit the BIOS-provided memory map to + * that size. exactmap can be used to specify + * the exact map. mem=number can be used to + * trim the existing memory map. + */ + unsigned long long mem_size; + + mem_size = memparse(from+4, &from); + limit_regions(mem_size); + userdef=1; + } + } + + else if (!memcmp(from, "memmap=", 7)) { + if (to != command_line) + to--; + if (!memcmp(from+7, "exactmap", 8)) { + from += 8+7; + e820.nr_map = 0; + userdef = 1; + } else { + /* If the user specifies memory size, we + * limit the BIOS-provided memory map to + * that size. exactmap can be used to specify + * the exact map. mem=number can be used to + * trim the existing memory map. + */ + unsigned long long start_at, mem_size; + + mem_size = memparse(from+7, &from); + if (*from == '@') { + start_at = memparse(from+1, &from); + add_memory_region(start_at, mem_size, E820_RAM); + } else if (*from == '#') { + start_at = memparse(from+1, &from); + add_memory_region(start_at, mem_size, E820_ACPI); + } else if (*from == '$') { + start_at = memparse(from+1, &from); + add_memory_region(start_at, mem_size, E820_RESERVED); + } else { + limit_regions(mem_size); + userdef=1; + } + } + } + + else if (!memcmp(from, "noexec=", 7)) + noexec_setup(from + 7); + + +#ifdef CONFIG_X86_SMP + /* + * If the BIOS enumerates physical processors before logical, + * maxcpus=N at enumeration-time can be used to disable HT. + */ + else if (!memcmp(from, "maxcpus=", 8)) { + extern unsigned int maxcpus; + + maxcpus = simple_strtoul(from + 8, NULL, 0); + } +#endif + +#ifdef CONFIG_ACPI_BOOT + /* "acpi=off" disables both ACPI table parsing and interpreter */ + else if (!memcmp(from, "acpi=off", 8)) { + disable_acpi(); + } + + /* acpi=force to over-ride black-list */ + else if (!memcmp(from, "acpi=force", 10)) { + acpi_force = 1; + acpi_ht = 1; + acpi_disabled = 0; + } + + /* acpi=strict disables out-of-spec workarounds */ + else if (!memcmp(from, "acpi=strict", 11)) { + acpi_strict = 1; + } + + /* Limit ACPI just to boot-time to enable HT */ + else if (!memcmp(from, "acpi=ht", 7)) { + if (!acpi_force) + disable_acpi(); + acpi_ht = 1; + } + + /* "pci=noacpi" disable ACPI IRQ routing and PCI scan */ + else if (!memcmp(from, "pci=noacpi", 10)) { + acpi_disable_pci(); + } + /* "acpi=noirq" disables ACPI interrupt routing */ + else if (!memcmp(from, "acpi=noirq", 10)) { + acpi_noirq_set(); + } + + else if (!memcmp(from, "acpi_sci=edge", 13)) + acpi_sci_flags.trigger = 1; + + else if (!memcmp(from, "acpi_sci=level", 14)) + acpi_sci_flags.trigger = 3; + + else if (!memcmp(from, "acpi_sci=high", 13)) + acpi_sci_flags.polarity = 1; + + else if (!memcmp(from, "acpi_sci=low", 12)) + acpi_sci_flags.polarity = 3; + +#ifdef CONFIG_X86_IO_APIC + else if (!memcmp(from, "acpi_skip_timer_override", 24)) + acpi_skip_timer_override = 1; +#endif + +#ifdef CONFIG_X86_LOCAL_APIC + /* disable IO-APIC */ + else if (!memcmp(from, "noapic", 6)) + disable_ioapic_setup(); +#endif /* CONFIG_X86_LOCAL_APIC */ +#endif /* CONFIG_ACPI_BOOT */ + + /* + * highmem=size forces highmem to be exactly 'size' bytes. + * This works even on boxes that have no highmem otherwise. + * This also works to reduce highmem size on bigger boxes. + */ + else if (!memcmp(from, "highmem=", 8)) + highmem_pages = memparse(from+8, &from) >> PAGE_SHIFT; + + /* + * vmalloc=size forces the vmalloc area to be exactly 'size' + * bytes. This can be used to increase (or decrease) the + * vmalloc area - the default is 128m. + */ + else if (!memcmp(from, "vmalloc=", 8)) + __VMALLOC_RESERVE = memparse(from+8, &from); + + next_char: + c = *(from++); + if (!c) + break; + if (COMMAND_LINE_SIZE <= ++len) + break; + *(to++) = c; + } + *to = '\0'; + *cmdline_p = command_line; + if (userdef) { + printk(KERN_INFO "user-defined physical RAM map:\n"); + print_memory_map("user"); + } +} + +/* + * Callback for efi_memory_walk. + */ +static int __init +efi_find_max_pfn(unsigned long start, unsigned long end, void *arg) +{ + unsigned long *max_pfn = arg, pfn; + + if (start < end) { + pfn = PFN_UP(end -1); + if (pfn > *max_pfn) + *max_pfn = pfn; + } + return 0; +} + + +/* + * Find the highest page frame number we have available + */ +void __init find_max_pfn(void) +{ + int i; + + max_pfn = 0; + if (efi_enabled) { + efi_memmap_walk(efi_find_max_pfn, &max_pfn); + return; + } + + for (i = 0; i < e820.nr_map; i++) { + unsigned long start, end; + /* RAM? */ + if (e820.map[i].type != E820_RAM) + continue; + start = PFN_UP(e820.map[i].addr); + end = PFN_DOWN(e820.map[i].addr + e820.map[i].size); + if (start >= end) + continue; + if (end > max_pfn) + max_pfn = end; + } +} + +/* + * Determine low and high memory ranges: + */ +unsigned long __init find_max_low_pfn(void) +{ + unsigned long max_low_pfn; + + max_low_pfn = max_pfn; + if (max_low_pfn > MAXMEM_PFN) { + if (highmem_pages == -1) + highmem_pages = max_pfn - MAXMEM_PFN; + if (highmem_pages + MAXMEM_PFN < max_pfn) + max_pfn = MAXMEM_PFN + highmem_pages; + if (highmem_pages + MAXMEM_PFN > max_pfn) { + printk("only %luMB highmem pages available, ignoring highmem size of %uMB.\n", pages_to_mb(max_pfn - MAXMEM_PFN), pages_to_mb(highmem_pages)); + highmem_pages = 0; + } + max_low_pfn = MAXMEM_PFN; +#ifndef CONFIG_HIGHMEM + /* Maximum memory usable is what is directly addressable */ + printk(KERN_WARNING "Warning only %ldMB will be used.\n", + MAXMEM>>20); + if (max_pfn > MAX_NONPAE_PFN) + printk(KERN_WARNING "Use a PAE enabled kernel.\n"); + else + printk(KERN_WARNING "Use a HIGHMEM enabled kernel.\n"); + max_pfn = MAXMEM_PFN; +#else /* !CONFIG_HIGHMEM */ +#ifndef CONFIG_X86_PAE + if (max_pfn > MAX_NONPAE_PFN) { + max_pfn = MAX_NONPAE_PFN; + printk(KERN_WARNING "Warning only 4GB will be used.\n"); + printk(KERN_WARNING "Use a PAE enabled kernel.\n"); + } +#endif /* !CONFIG_X86_PAE */ +#endif /* !CONFIG_HIGHMEM */ + } else { + if (highmem_pages == -1) + highmem_pages = 0; +#ifdef CONFIG_HIGHMEM + if (highmem_pages >= max_pfn) { + printk(KERN_ERR "highmem size specified (%uMB) is bigger than pages available (%luMB)!.\n", pages_to_mb(highmem_pages), pages_to_mb(max_pfn)); + highmem_pages = 0; + } + if (highmem_pages) { + if (max_low_pfn-highmem_pages < 64*1024*1024/PAGE_SIZE){ + printk(KERN_ERR "highmem size %uMB results in smaller than 64MB lowmem, ignoring it.\n", pages_to_mb(highmem_pages)); + highmem_pages = 0; + } + max_low_pfn -= highmem_pages; + } +#else + if (highmem_pages) + printk(KERN_ERR "ignoring highmem size on non-highmem kernel!\n"); +#endif + } + return max_low_pfn; +} + +/* + * Free all available memory for boot time allocation. Used + * as a callback function by efi_memory_walk() + */ + +static int __init +free_available_memory(unsigned long start, unsigned long end, void *arg) +{ + /* check max_low_pfn */ + if (start >= ((max_low_pfn + 1) << PAGE_SHIFT)) + return 0; + if (end >= ((max_low_pfn + 1) << PAGE_SHIFT)) + end = (max_low_pfn + 1) << PAGE_SHIFT; + if (start < end) + free_bootmem(start, end - start); + + return 0; +} +/* + * Register fully available low RAM pages with the bootmem allocator. + */ +static void __init register_bootmem_low_pages(unsigned long max_low_pfn) +{ + int i; + + if (efi_enabled) { + efi_memmap_walk(free_available_memory, NULL); + return; + } + for (i = 0; i < e820.nr_map; i++) { + unsigned long curr_pfn, last_pfn, size; + /* + * Reserve usable low memory + */ + if (e820.map[i].type != E820_RAM) + continue; + /* + * We are rounding up the start address of usable memory: + */ + curr_pfn = PFN_UP(e820.map[i].addr); + if (curr_pfn >= max_low_pfn) + continue; + /* + * ... and at the end of the usable range downwards: + */ + last_pfn = PFN_DOWN(e820.map[i].addr + e820.map[i].size); + + if (last_pfn > max_low_pfn) + last_pfn = max_low_pfn; + + /* + * .. finally, did all the rounding and playing + * around just make the area go away? + */ + if (last_pfn <= curr_pfn) + continue; + + size = last_pfn - curr_pfn; + free_bootmem(PFN_PHYS(curr_pfn), PFN_PHYS(size)); + } +} + +/* + * workaround for Dell systems that neglect to reserve EBDA + */ +static void __init reserve_ebda_region(void) +{ + unsigned int addr; + addr = get_bios_ebda(); + if (addr) + reserve_bootmem(addr, PAGE_SIZE); +} + +#ifndef CONFIG_DISCONTIGMEM +void __init setup_bootmem_allocator(void); +static unsigned long __init setup_memory(void) +{ + /* + * partially used pages are not usable - thus + * we are rounding upwards: + */ + min_low_pfn = PFN_UP(init_pg_tables_end); + + find_max_pfn(); + + max_low_pfn = find_max_low_pfn(); + +#ifdef CONFIG_HIGHMEM + highstart_pfn = highend_pfn = max_pfn; + if (max_pfn > max_low_pfn) { + highstart_pfn = max_low_pfn; + } + printk(KERN_NOTICE "%ldMB HIGHMEM available.\n", + pages_to_mb(highend_pfn - highstart_pfn)); +#endif + printk(KERN_NOTICE "%ldMB LOWMEM available.\n", + pages_to_mb(max_low_pfn)); + + setup_bootmem_allocator(); + + return max_low_pfn; +} + +void __init zone_sizes_init(void) +{ + unsigned long zones_size[MAX_NR_ZONES] = {0, 0, 0}; + unsigned int max_dma, low; + + max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; + low = max_low_pfn; + + if (low < max_dma) + zones_size[ZONE_DMA] = low; + else { + zones_size[ZONE_DMA] = max_dma; + zones_size[ZONE_NORMAL] = low - max_dma; +#ifdef CONFIG_HIGHMEM + zones_size[ZONE_HIGHMEM] = highend_pfn - low; +#endif + } + free_area_init(zones_size); +} +#else +extern unsigned long setup_memory(void); +extern void zone_sizes_init(void); +#endif /* !CONFIG_DISCONTIGMEM */ + +void __init setup_bootmem_allocator(void) +{ + unsigned long bootmap_size; + /* + * Initialize the boot-time allocator (with low memory only): + */ + bootmap_size = init_bootmem(min_low_pfn, max_low_pfn); + + register_bootmem_low_pages(max_low_pfn); + + /* + * Reserve the bootmem bitmap itself as well. We do this in two + * steps (first step was init_bootmem()) because this catches + * the (very unlikely) case of us accidentally initializing the + * bootmem allocator with an invalid RAM area. + */ + reserve_bootmem(HIGH_MEMORY, (PFN_PHYS(min_low_pfn) + + bootmap_size + PAGE_SIZE-1) - (HIGH_MEMORY)); + + /* + * reserve physical page 0 - it's a special BIOS page on many boxes, + * enabling clean reboots, SMP operation, laptop functions. + */ + reserve_bootmem(0, PAGE_SIZE); + + /* reserve EBDA region, it's a 4K region */ + reserve_ebda_region(); + + /* could be an AMD 768MPX chipset. Reserve a page before VGA to prevent + PCI prefetch into it (errata #56). Usually the page is reserved anyways, + unless you have no PS/2 mouse plugged in. */ + if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD && + boot_cpu_data.x86 == 6) + reserve_bootmem(0xa0000 - 4096, 4096); + +#ifdef CONFIG_SMP + /* + * But first pinch a few for the stack/trampoline stuff + * FIXME: Don't need the extra page at 4K, but need to fix + * trampoline before removing it. (see the GDT stuff) + */ + reserve_bootmem(PAGE_SIZE, PAGE_SIZE); +#endif +#ifdef CONFIG_ACPI_SLEEP + /* + * Reserve low memory region for sleep support. + */ + acpi_reserve_bootmem(); +#endif +#ifdef CONFIG_X86_FIND_SMP_CONFIG + /* + * Find and reserve possible boot-time SMP configuration: + */ + find_smp_config(); +#endif + +#ifdef CONFIG_BLK_DEV_INITRD + if (LOADER_TYPE && INITRD_START) { + if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) { + reserve_bootmem(INITRD_START, INITRD_SIZE); + initrd_start = + INITRD_START ? INITRD_START + PAGE_OFFSET : 0; + initrd_end = initrd_start+INITRD_SIZE; + } + else { + printk(KERN_ERR "initrd extends beyond end of memory " + "(0x%08lx > 0x%08lx)\ndisabling initrd\n", + INITRD_START + INITRD_SIZE, + max_low_pfn << PAGE_SHIFT); + initrd_start = 0; + } + } +#endif +} + +/* + * The node 0 pgdat is initialized before all of these because + * it's needed for bootmem. node>0 pgdats have their virtual + * space allocated before the pagetables are in place to access + * them, so they can't be cleared then. + * + * This should all compile down to nothing when NUMA is off. + */ +void __init remapped_pgdat_init(void) +{ + int nid; + + for_each_online_node(nid) { + if (nid != 0) + memset(NODE_DATA(nid), 0, sizeof(struct pglist_data)); + } +} + +/* + * Request address space for all standard RAM and ROM resources + * and also for regions reported as reserved by the e820. + */ +static void __init +legacy_init_iomem_resources(struct resource *code_resource, struct resource *data_resource) +{ + int i; + + probe_roms(); + for (i = 0; i < e820.nr_map; i++) { + struct resource *res; + if (e820.map[i].addr + e820.map[i].size > 0x100000000ULL) + continue; + res = alloc_bootmem_low(sizeof(struct resource)); + switch (e820.map[i].type) { + case E820_RAM: res->name = "System RAM"; break; + case E820_ACPI: res->name = "ACPI Tables"; break; + case E820_NVS: res->name = "ACPI Non-volatile Storage"; break; + default: res->name = "reserved"; + } + res->start = e820.map[i].addr; + res->end = res->start + e820.map[i].size - 1; + res->flags = IORESOURCE_MEM | IORESOURCE_BUSY; + request_resource(&iomem_resource, res); + if (e820.map[i].type == E820_RAM) { + /* + * We don't know which RAM region contains kernel data, + * so we try it repeatedly and let the resource manager + * test it. + */ + request_resource(res, code_resource); + request_resource(res, data_resource); + } + } +} + +/* + * Request address space for all standard resources + */ +static void __init register_memory(void) +{ + unsigned long gapstart, gapsize; + unsigned long long last; + int i; + + if (efi_enabled) + efi_initialize_iomem_resources(&code_resource, &data_resource); + else + legacy_init_iomem_resources(&code_resource, &data_resource); + + /* EFI systems may still have VGA */ + request_resource(&iomem_resource, &video_ram_resource); + + /* request I/O space for devices used on all i[345]86 PCs */ + for (i = 0; i < STANDARD_IO_RESOURCES; i++) + request_resource(&ioport_resource, &standard_io_resources[i]); + + /* + * Search for the bigest gap in the low 32 bits of the e820 + * memory space. + */ + last = 0x100000000ull; + gapstart = 0x10000000; + gapsize = 0x400000; + i = e820.nr_map; + while (--i >= 0) { + unsigned long long start = e820.map[i].addr; + unsigned long long end = start + e820.map[i].size; + + /* + * Since "last" is at most 4GB, we know we'll + * fit in 32 bits if this condition is true + */ + if (last > end) { + unsigned long gap = last - end; + + if (gap > gapsize) { + gapsize = gap; + gapstart = end; + } + } + if (start < last) + last = start; + } + + /* + * Start allocating dynamic PCI memory a bit into the gap, + * aligned up to the nearest megabyte. + * + * Question: should we try to pad it up a bit (do something + * like " + (gapsize >> 3)" in there too?). We now have the + * technology. + */ + pci_mem_start = (gapstart + 0xfffff) & ~0xfffff; + + printk("Allocating PCI resources starting at %08lx (gap: %08lx:%08lx)\n", + pci_mem_start, gapstart, gapsize); +} + +/* Use inline assembly to define this because the nops are defined + as inline assembly strings in the include files and we cannot + get them easily into strings. */ +asm("\t.data\nintelnops: " + GENERIC_NOP1 GENERIC_NOP2 GENERIC_NOP3 GENERIC_NOP4 GENERIC_NOP5 GENERIC_NOP6 + GENERIC_NOP7 GENERIC_NOP8); +asm("\t.data\nk8nops: " + K8_NOP1 K8_NOP2 K8_NOP3 K8_NOP4 K8_NOP5 K8_NOP6 + K8_NOP7 K8_NOP8); +asm("\t.data\nk7nops: " + K7_NOP1 K7_NOP2 K7_NOP3 K7_NOP4 K7_NOP5 K7_NOP6 + K7_NOP7 K7_NOP8); + +extern unsigned char intelnops[], k8nops[], k7nops[]; +static unsigned char *intel_nops[ASM_NOP_MAX+1] = { + NULL, + intelnops, + intelnops + 1, + intelnops + 1 + 2, + intelnops + 1 + 2 + 3, + intelnops + 1 + 2 + 3 + 4, + intelnops + 1 + 2 + 3 + 4 + 5, + intelnops + 1 + 2 + 3 + 4 + 5 + 6, + intelnops + 1 + 2 + 3 + 4 + 5 + 6 + 7, +}; +static unsigned char *k8_nops[ASM_NOP_MAX+1] = { + NULL, + k8nops, + k8nops + 1, + k8nops + 1 + 2, + k8nops + 1 + 2 + 3, + k8nops + 1 + 2 + 3 + 4, + k8nops + 1 + 2 + 3 + 4 + 5, + k8nops + 1 + 2 + 3 + 4 + 5 + 6, + k8nops + 1 + 2 + 3 + 4 + 5 + 6 + 7, +}; +static unsigned char *k7_nops[ASM_NOP_MAX+1] = { + NULL, + k7nops, + k7nops + 1, + k7nops + 1 + 2, + k7nops + 1 + 2 + 3, + k7nops + 1 + 2 + 3 + 4, + k7nops + 1 + 2 + 3 + 4 + 5, + k7nops + 1 + 2 + 3 + 4 + 5 + 6, + k7nops + 1 + 2 + 3 + 4 + 5 + 6 + 7, +}; +static struct nop { + int cpuid; + unsigned char **noptable; +} noptypes[] = { + { X86_FEATURE_K8, k8_nops }, + { X86_FEATURE_K7, k7_nops }, + { -1, NULL } +}; + +/* Replace instructions with better alternatives for this CPU type. + + This runs before SMP is initialized to avoid SMP problems with + self modifying code. This implies that assymetric systems where + APs have less capabilities than the boot processor are not handled. + In this case boot with "noreplacement". */ +void apply_alternatives(void *start, void *end) +{ + struct alt_instr *a; + int diff, i, k; + unsigned char **noptable = intel_nops; + for (i = 0; noptypes[i].cpuid >= 0; i++) { + if (boot_cpu_has(noptypes[i].cpuid)) { + noptable = noptypes[i].noptable; + break; + } + } + for (a = start; (void *)a < end; a++) { + if (!boot_cpu_has(a->cpuid)) + continue; + BUG_ON(a->replacementlen > a->instrlen); + memcpy(a->instr, a->replacement, a->replacementlen); + diff = a->instrlen - a->replacementlen; + /* Pad the rest with nops */ + for (i = a->replacementlen; diff > 0; diff -= k, i += k) { + k = diff; + if (k > ASM_NOP_MAX) + k = ASM_NOP_MAX; + memcpy(a->instr + i, noptable[k], k); + } + } +} + +static int no_replacement __initdata = 0; + +void __init alternative_instructions(void) +{ + extern struct alt_instr __alt_instructions[], __alt_instructions_end[]; + if (no_replacement) + return; + apply_alternatives(__alt_instructions, __alt_instructions_end); +} + +static int __init noreplacement_setup(char *s) +{ + no_replacement = 1; + return 0; +} + +__setup("noreplacement", noreplacement_setup); + +static char * __init machine_specific_memory_setup(void); + +#ifdef CONFIG_MCA +static void set_mca_bus(int x) +{ + MCA_bus = x; +} +#else +static void set_mca_bus(int x) { } +#endif + +/* + * Determine if we were loaded by an EFI loader. If so, then we have also been + * passed the efi memmap, systab, etc., so we should use these data structures + * for initialization. Note, the efi init code path is determined by the + * global efi_enabled. This allows the same kernel image to be used on existing + * systems (with a traditional BIOS) as well as on EFI systems. + */ +void __init setup_arch(char **cmdline_p) +{ + unsigned long max_low_pfn; + + memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data)); + pre_setup_arch_hook(); + early_cpu_init(); + + /* + * FIXME: This isn't an official loader_type right + * now but does currently work with elilo. + * If we were configured as an EFI kernel, check to make + * sure that we were loaded correctly from elilo and that + * the system table is valid. If not, then initialize normally. + */ +#ifdef CONFIG_EFI + if ((LOADER_TYPE == 0x50) && EFI_SYSTAB) + efi_enabled = 1; +#endif + + ROOT_DEV = old_decode_dev(ORIG_ROOT_DEV); + drive_info = DRIVE_INFO; + screen_info = SCREEN_INFO; + edid_info = EDID_INFO; + apm_info.bios = APM_BIOS_INFO; + ist_info = IST_INFO; + saved_videomode = VIDEO_MODE; + if( SYS_DESC_TABLE.length != 0 ) { + set_mca_bus(SYS_DESC_TABLE.table[3] & 0x2); + machine_id = SYS_DESC_TABLE.table[0]; + machine_submodel_id = SYS_DESC_TABLE.table[1]; + BIOS_revision = SYS_DESC_TABLE.table[2]; + } + bootloader_type = LOADER_TYPE; + +#ifdef CONFIG_BLK_DEV_RAM + rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK; + rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0); + rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0); +#endif + ARCH_SETUP + if (efi_enabled) + efi_init(); + else { + printk(KERN_INFO "BIOS-provided physical RAM map:\n"); + print_memory_map(machine_specific_memory_setup()); + } + + copy_edd(); + + if (!MOUNT_ROOT_RDONLY) + root_mountflags &= ~MS_RDONLY; + init_mm.start_code = (unsigned long) _text; + init_mm.end_code = (unsigned long) _etext; + init_mm.end_data = (unsigned long) _edata; + init_mm.brk = init_pg_tables_end + PAGE_OFFSET; + + code_resource.start = virt_to_phys(_text); + code_resource.end = virt_to_phys(_etext)-1; + data_resource.start = virt_to_phys(_etext); + data_resource.end = virt_to_phys(_edata)-1; + + parse_cmdline_early(cmdline_p); + + max_low_pfn = setup_memory(); + + /* + * NOTE: before this point _nobody_ is allowed to allocate + * any memory using the bootmem allocator. Although the + * alloctor is now initialised only the first 8Mb of the kernel + * virtual address space has been mapped. All allocations before + * paging_init() has completed must use the alloc_bootmem_low_pages() + * variant (which allocates DMA'able memory) and care must be taken + * not to exceed the 8Mb limit. + */ + +#ifdef CONFIG_SMP + smp_alloc_memory(); /* AP processor realmode stacks in low memory*/ +#endif + paging_init(); + remapped_pgdat_init(); + zone_sizes_init(); + + /* + * NOTE: at this point the bootmem allocator is fully available. + */ + +#ifdef CONFIG_EARLY_PRINTK + { + char *s = strstr(*cmdline_p, "earlyprintk="); + if (s) { + extern void setup_early_printk(char *); + + setup_early_printk(s); + printk("early console enabled\n"); + } + } +#endif + + + dmi_scan_machine(); + +#ifdef CONFIG_X86_GENERICARCH + generic_apic_probe(*cmdline_p); +#endif + if (efi_enabled) + efi_map_memmap(); + + /* + * Parse the ACPI tables for possible boot-time SMP configuration. + */ + acpi_boot_table_init(); + acpi_boot_init(); + +#ifdef CONFIG_X86_LOCAL_APIC + if (smp_found_config) + get_smp_config(); +#endif + + register_memory(); + +#ifdef CONFIG_VT +#if defined(CONFIG_VGA_CONSOLE) + if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY)) + conswitchp = &vga_con; +#elif defined(CONFIG_DUMMY_CONSOLE) + conswitchp = &dummy_con; +#endif +#endif +} + +#include "setup_arch_post.h" +/* + * Local Variables: + * mode:c + * c-file-style:"k&r" + * c-basic-offset:8 + * End: + */ diff --git a/arch/i386/kernel/sigframe.h b/arch/i386/kernel/sigframe.h new file mode 100644 index 00000000000..d21b14f5c25 --- /dev/null +++ b/arch/i386/kernel/sigframe.h @@ -0,0 +1,21 @@ +struct sigframe +{ + char *pretcode; + int sig; + struct sigcontext sc; + struct _fpstate fpstate; + unsigned long extramask[_NSIG_WORDS-1]; + char retcode[8]; +}; + +struct rt_sigframe +{ + char *pretcode; + int sig; + struct siginfo *pinfo; + void *puc; + struct siginfo info; + struct ucontext uc; + struct _fpstate fpstate; + char retcode[8]; +}; diff --git a/arch/i386/kernel/signal.c b/arch/i386/kernel/signal.c new file mode 100644 index 00000000000..ef3602e1c05 --- /dev/null +++ b/arch/i386/kernel/signal.c @@ -0,0 +1,665 @@ +/* + * linux/arch/i386/kernel/signal.c + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * 1997-11-28 Modified for POSIX.1b signals by Richard Henderson + * 2000-06-20 Pentium III FXSR, SSE support by Gareth Hughes + */ + +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/kernel.h> +#include <linux/signal.h> +#include <linux/errno.h> +#include <linux/wait.h> +#include <linux/unistd.h> +#include <linux/stddef.h> +#include <linux/personality.h> +#include <linux/suspend.h> +#include <linux/ptrace.h> +#include <linux/elf.h> +#include <asm/processor.h> +#include <asm/ucontext.h> +#include <asm/uaccess.h> +#include <asm/i387.h> +#include "sigframe.h" + +#define DEBUG_SIG 0 + +#define _BLOCKABLE (~(sigmask(SIGKILL) | sigmask(SIGSTOP))) + +/* + * Atomically swap in the new signal mask, and wait for a signal. + */ +asmlinkage int +sys_sigsuspend(int history0, int history1, old_sigset_t mask) +{ + struct pt_regs * regs = (struct pt_regs *) &history0; + sigset_t saveset; + + mask &= _BLOCKABLE; + spin_lock_irq(¤t->sighand->siglock); + saveset = current->blocked; + siginitset(¤t->blocked, mask); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + regs->eax = -EINTR; + while (1) { + current->state = TASK_INTERRUPTIBLE; + schedule(); + if (do_signal(regs, &saveset)) + return -EINTR; + } +} + +asmlinkage int +sys_rt_sigsuspend(struct pt_regs regs) +{ + sigset_t saveset, newset; + + /* XXX: Don't preclude handling different sized sigset_t's. */ + if (regs.ecx != sizeof(sigset_t)) + return -EINVAL; + + if (copy_from_user(&newset, (sigset_t __user *)regs.ebx, sizeof(newset))) + return -EFAULT; + sigdelsetmask(&newset, ~_BLOCKABLE); + + spin_lock_irq(¤t->sighand->siglock); + saveset = current->blocked; + current->blocked = newset; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + regs.eax = -EINTR; + while (1) { + current->state = TASK_INTERRUPTIBLE; + schedule(); + if (do_signal(®s, &saveset)) + return -EINTR; + } +} + +asmlinkage int +sys_sigaction(int sig, const struct old_sigaction __user *act, + struct old_sigaction __user *oact) +{ + struct k_sigaction new_ka, old_ka; + int ret; + + if (act) { + old_sigset_t mask; + if (!access_ok(VERIFY_READ, act, sizeof(*act)) || + __get_user(new_ka.sa.sa_handler, &act->sa_handler) || + __get_user(new_ka.sa.sa_restorer, &act->sa_restorer)) + return -EFAULT; + __get_user(new_ka.sa.sa_flags, &act->sa_flags); + __get_user(mask, &act->sa_mask); + siginitset(&new_ka.sa.sa_mask, mask); + } + + ret = do_sigaction(sig, act ? &new_ka : NULL, oact ? &old_ka : NULL); + + if (!ret && oact) { + if (!access_ok(VERIFY_WRITE, oact, sizeof(*oact)) || + __put_user(old_ka.sa.sa_handler, &oact->sa_handler) || + __put_user(old_ka.sa.sa_restorer, &oact->sa_restorer)) + return -EFAULT; + __put_user(old_ka.sa.sa_flags, &oact->sa_flags); + __put_user(old_ka.sa.sa_mask.sig[0], &oact->sa_mask); + } + + return ret; +} + +asmlinkage int +sys_sigaltstack(unsigned long ebx) +{ + /* This is needed to make gcc realize it doesn't own the "struct pt_regs" */ + struct pt_regs *regs = (struct pt_regs *)&ebx; + const stack_t __user *uss = (const stack_t __user *)ebx; + stack_t __user *uoss = (stack_t __user *)regs->ecx; + + return do_sigaltstack(uss, uoss, regs->esp); +} + + +/* + * Do a signal return; undo the signal stack. + */ + +static int +restore_sigcontext(struct pt_regs *regs, struct sigcontext __user *sc, int *peax) +{ + unsigned int err = 0; + + /* Always make any pending restarted system calls return -EINTR */ + current_thread_info()->restart_block.fn = do_no_restart_syscall; + +#define COPY(x) err |= __get_user(regs->x, &sc->x) + +#define COPY_SEG(seg) \ + { unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->x##seg = tmp; } + +#define COPY_SEG_STRICT(seg) \ + { unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + regs->x##seg = tmp|3; } + +#define GET_SEG(seg) \ + { unsigned short tmp; \ + err |= __get_user(tmp, &sc->seg); \ + loadsegment(seg,tmp); } + +#define FIX_EFLAGS (X86_EFLAGS_AC | X86_EFLAGS_OF | X86_EFLAGS_DF | \ + X86_EFLAGS_TF | X86_EFLAGS_SF | X86_EFLAGS_ZF | \ + X86_EFLAGS_AF | X86_EFLAGS_PF | X86_EFLAGS_CF) + + GET_SEG(gs); + GET_SEG(fs); + COPY_SEG(es); + COPY_SEG(ds); + COPY(edi); + COPY(esi); + COPY(ebp); + COPY(esp); + COPY(ebx); + COPY(edx); + COPY(ecx); + COPY(eip); + COPY_SEG_STRICT(cs); + COPY_SEG_STRICT(ss); + + { + unsigned int tmpflags; + err |= __get_user(tmpflags, &sc->eflags); + regs->eflags = (regs->eflags & ~FIX_EFLAGS) | (tmpflags & FIX_EFLAGS); + regs->orig_eax = -1; /* disable syscall checks */ + } + + { + struct _fpstate __user * buf; + err |= __get_user(buf, &sc->fpstate); + if (buf) { + if (!access_ok(VERIFY_READ, buf, sizeof(*buf))) + goto badframe; + err |= restore_i387(buf); + } else { + struct task_struct *me = current; + if (used_math()) { + clear_fpu(me); + clear_used_math(); + } + } + } + + err |= __get_user(*peax, &sc->eax); + return err; + +badframe: + return 1; +} + +asmlinkage int sys_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *) &__unused; + struct sigframe __user *frame = (struct sigframe __user *)(regs->esp - 8); + sigset_t set; + int eax; + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__get_user(set.sig[0], &frame->sc.oldmask) + || (_NSIG_WORDS > 1 + && __copy_from_user(&set.sig[1], &frame->extramask, + sizeof(frame->extramask)))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->sc, &eax)) + goto badframe; + return eax; + +badframe: + force_sig(SIGSEGV, current); + return 0; +} + +asmlinkage int sys_rt_sigreturn(unsigned long __unused) +{ + struct pt_regs *regs = (struct pt_regs *) &__unused; + struct rt_sigframe __user *frame = (struct rt_sigframe __user *)(regs->esp - 4); + sigset_t set; + int eax; + + if (!access_ok(VERIFY_READ, frame, sizeof(*frame))) + goto badframe; + if (__copy_from_user(&set, &frame->uc.uc_sigmask, sizeof(set))) + goto badframe; + + sigdelsetmask(&set, ~_BLOCKABLE); + spin_lock_irq(¤t->sighand->siglock); + current->blocked = set; + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + + if (restore_sigcontext(regs, &frame->uc.uc_mcontext, &eax)) + goto badframe; + + if (do_sigaltstack(&frame->uc.uc_stack, NULL, regs->esp) == -EFAULT) + goto badframe; + + return eax; + +badframe: + force_sig(SIGSEGV, current); + return 0; +} + +/* + * Set up a signal frame. + */ + +static int +setup_sigcontext(struct sigcontext __user *sc, struct _fpstate __user *fpstate, + struct pt_regs *regs, unsigned long mask) +{ + int tmp, err = 0; + + tmp = 0; + __asm__("movl %%gs,%0" : "=r"(tmp): "0"(tmp)); + err |= __put_user(tmp, (unsigned int __user *)&sc->gs); + __asm__("movl %%fs,%0" : "=r"(tmp): "0"(tmp)); + err |= __put_user(tmp, (unsigned int __user *)&sc->fs); + + err |= __put_user(regs->xes, (unsigned int __user *)&sc->es); + err |= __put_user(regs->xds, (unsigned int __user *)&sc->ds); + err |= __put_user(regs->edi, &sc->edi); + err |= __put_user(regs->esi, &sc->esi); + err |= __put_user(regs->ebp, &sc->ebp); + err |= __put_user(regs->esp, &sc->esp); + err |= __put_user(regs->ebx, &sc->ebx); + err |= __put_user(regs->edx, &sc->edx); + err |= __put_user(regs->ecx, &sc->ecx); + err |= __put_user(regs->eax, &sc->eax); + err |= __put_user(current->thread.trap_no, &sc->trapno); + err |= __put_user(current->thread.error_code, &sc->err); + err |= __put_user(regs->eip, &sc->eip); + err |= __put_user(regs->xcs, (unsigned int __user *)&sc->cs); + err |= __put_user(regs->eflags, &sc->eflags); + err |= __put_user(regs->esp, &sc->esp_at_signal); + err |= __put_user(regs->xss, (unsigned int __user *)&sc->ss); + + tmp = save_i387(fpstate); + if (tmp < 0) + err = 1; + else + err |= __put_user(tmp ? fpstate : NULL, &sc->fpstate); + + /* non-iBCS2 extensions.. */ + err |= __put_user(mask, &sc->oldmask); + err |= __put_user(current->thread.cr2, &sc->cr2); + + return err; +} + +/* + * Determine which stack to use.. + */ +static inline void __user * +get_sigframe(struct k_sigaction *ka, struct pt_regs * regs, size_t frame_size) +{ + unsigned long esp; + + /* Default to using normal stack */ + esp = regs->esp; + + /* This is the X/Open sanctioned signal stack switching. */ + if (ka->sa.sa_flags & SA_ONSTACK) { + if (sas_ss_flags(esp) == 0) + esp = current->sas_ss_sp + current->sas_ss_size; + } + + /* This is the legacy signal stack switching. */ + else if ((regs->xss & 0xffff) != __USER_DS && + !(ka->sa.sa_flags & SA_RESTORER) && + ka->sa.sa_restorer) { + esp = (unsigned long) ka->sa.sa_restorer; + } + + return (void __user *)((esp - frame_size) & -8ul); +} + +/* These symbols are defined with the addresses in the vsyscall page. + See vsyscall-sigreturn.S. */ +extern void __user __kernel_sigreturn; +extern void __user __kernel_rt_sigreturn; + +static void setup_frame(int sig, struct k_sigaction *ka, + sigset_t *set, struct pt_regs * regs) +{ + void __user *restorer; + struct sigframe __user *frame; + int err = 0; + int usig; + + frame = get_sigframe(ka, regs, sizeof(*frame)); + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + goto give_sigsegv; + + usig = current_thread_info()->exec_domain + && current_thread_info()->exec_domain->signal_invmap + && sig < 32 + ? current_thread_info()->exec_domain->signal_invmap[sig] + : sig; + + err = __put_user(usig, &frame->sig); + if (err) + goto give_sigsegv; + + err = setup_sigcontext(&frame->sc, &frame->fpstate, regs, set->sig[0]); + if (err) + goto give_sigsegv; + + if (_NSIG_WORDS > 1) { + err = __copy_to_user(&frame->extramask, &set->sig[1], + sizeof(frame->extramask)); + if (err) + goto give_sigsegv; + } + + restorer = &__kernel_sigreturn; + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + + /* Set up to return from userspace. */ + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is popl %eax ; movl $,%eax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(0xb858, (short __user *)(frame->retcode+0)); + err |= __put_user(__NR_sigreturn, (int __user *)(frame->retcode+2)); + err |= __put_user(0x80cd, (short __user *)(frame->retcode+6)); + + if (err) + goto give_sigsegv; + + /* Set up registers for signal handler */ + regs->esp = (unsigned long) frame; + regs->eip = (unsigned long) ka->sa.sa_handler; + regs->eax = (unsigned long) sig; + regs->edx = (unsigned long) 0; + regs->ecx = (unsigned long) 0; + + set_fs(USER_DS); + regs->xds = __USER_DS; + regs->xes = __USER_DS; + regs->xss = __USER_DS; + regs->xcs = __USER_CS; + + /* + * Clear TF when entering the signal handler, but + * notify any tracer that was single-stepping it. + * The tracer may want to single-step inside the + * handler too. + */ + regs->eflags &= ~TF_MASK; + if (test_thread_flag(TIF_SINGLESTEP)) + ptrace_notify(SIGTRAP); + +#if DEBUG_SIG + printk("SIG deliver (%s:%d): sp=%p pc=%p ra=%p\n", + current->comm, current->pid, frame, regs->eip, frame->pretcode); +#endif + + return; + +give_sigsegv: + force_sigsegv(sig, current); +} + +static void setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info, + sigset_t *set, struct pt_regs * regs) +{ + void __user *restorer; + struct rt_sigframe __user *frame; + int err = 0; + int usig; + + frame = get_sigframe(ka, regs, sizeof(*frame)); + + if (!access_ok(VERIFY_WRITE, frame, sizeof(*frame))) + goto give_sigsegv; + + usig = current_thread_info()->exec_domain + && current_thread_info()->exec_domain->signal_invmap + && sig < 32 + ? current_thread_info()->exec_domain->signal_invmap[sig] + : sig; + + err |= __put_user(usig, &frame->sig); + err |= __put_user(&frame->info, &frame->pinfo); + err |= __put_user(&frame->uc, &frame->puc); + err |= copy_siginfo_to_user(&frame->info, info); + if (err) + goto give_sigsegv; + + /* Create the ucontext. */ + err |= __put_user(0, &frame->uc.uc_flags); + err |= __put_user(0, &frame->uc.uc_link); + err |= __put_user(current->sas_ss_sp, &frame->uc.uc_stack.ss_sp); + err |= __put_user(sas_ss_flags(regs->esp), + &frame->uc.uc_stack.ss_flags); + err |= __put_user(current->sas_ss_size, &frame->uc.uc_stack.ss_size); + err |= setup_sigcontext(&frame->uc.uc_mcontext, &frame->fpstate, + regs, set->sig[0]); + err |= __copy_to_user(&frame->uc.uc_sigmask, set, sizeof(*set)); + if (err) + goto give_sigsegv; + + /* Set up to return from userspace. */ + restorer = &__kernel_rt_sigreturn; + if (ka->sa.sa_flags & SA_RESTORER) + restorer = ka->sa.sa_restorer; + err |= __put_user(restorer, &frame->pretcode); + + /* + * This is movl $,%eax ; int $0x80 + * + * WE DO NOT USE IT ANY MORE! It's only left here for historical + * reasons and because gdb uses it as a signature to notice + * signal handler stack frames. + */ + err |= __put_user(0xb8, (char __user *)(frame->retcode+0)); + err |= __put_user(__NR_rt_sigreturn, (int __user *)(frame->retcode+1)); + err |= __put_user(0x80cd, (short __user *)(frame->retcode+5)); + + if (err) + goto give_sigsegv; + + /* Set up registers for signal handler */ + regs->esp = (unsigned long) frame; + regs->eip = (unsigned long) ka->sa.sa_handler; + regs->eax = (unsigned long) usig; + regs->edx = (unsigned long) &frame->info; + regs->ecx = (unsigned long) &frame->uc; + + set_fs(USER_DS); + regs->xds = __USER_DS; + regs->xes = __USER_DS; + regs->xss = __USER_DS; + regs->xcs = __USER_CS; + + /* + * Clear TF when entering the signal handler, but + * notify any tracer that was single-stepping it. + * The tracer may want to single-step inside the + * handler too. + */ + regs->eflags &= ~TF_MASK; + if (test_thread_flag(TIF_SINGLESTEP)) + ptrace_notify(SIGTRAP); + +#if DEBUG_SIG + printk("SIG deliver (%s:%d): sp=%p pc=%p ra=%p\n", + current->comm, current->pid, frame, regs->eip, frame->pretcode); +#endif + + return; + +give_sigsegv: + force_sigsegv(sig, current); +} + +/* + * OK, we're invoking a handler + */ + +static void +handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka, + sigset_t *oldset, struct pt_regs * regs) +{ + /* Are we from a system call? */ + if (regs->orig_eax >= 0) { + /* If so, check system call restarting.. */ + switch (regs->eax) { + case -ERESTART_RESTARTBLOCK: + case -ERESTARTNOHAND: + regs->eax = -EINTR; + break; + + case -ERESTARTSYS: + if (!(ka->sa.sa_flags & SA_RESTART)) { + regs->eax = -EINTR; + break; + } + /* fallthrough */ + case -ERESTARTNOINTR: + regs->eax = regs->orig_eax; + regs->eip -= 2; + } + } + + /* + * If TF is set due to a debugger (PT_DTRACE), clear the TF flag so + * that register information in the sigcontext is correct. + */ + if (unlikely(regs->eflags & TF_MASK) + && likely(current->ptrace & PT_DTRACE)) { + current->ptrace &= ~PT_DTRACE; + regs->eflags &= ~TF_MASK; + } + + /* Set up the stack frame */ + if (ka->sa.sa_flags & SA_SIGINFO) + setup_rt_frame(sig, ka, info, oldset, regs); + else + setup_frame(sig, ka, oldset, regs); + + if (!(ka->sa.sa_flags & SA_NODEFER)) { + spin_lock_irq(¤t->sighand->siglock); + sigorsets(¤t->blocked,¤t->blocked,&ka->sa.sa_mask); + sigaddset(¤t->blocked,sig); + recalc_sigpending(); + spin_unlock_irq(¤t->sighand->siglock); + } +} + +/* + * Note that 'init' is a special process: it doesn't get signals it doesn't + * want to handle. Thus you cannot kill init even with a SIGKILL even by + * mistake. + */ +int fastcall do_signal(struct pt_regs *regs, sigset_t *oldset) +{ + siginfo_t info; + int signr; + struct k_sigaction ka; + + /* + * We want the common case to go fast, which + * is why we may in certain cases get here from + * kernel mode. Just return without doing anything + * if so. + */ + if ((regs->xcs & 3) != 3) + return 1; + + if (current->flags & PF_FREEZE) { + refrigerator(0); + goto no_signal; + } + + if (!oldset) + oldset = ¤t->blocked; + + signr = get_signal_to_deliver(&info, &ka, regs, NULL); + if (signr > 0) { + /* Reenable any watchpoints before delivering the + * signal to user space. The processor register will + * have been cleared if the watchpoint triggered + * inside the kernel. + */ + if (unlikely(current->thread.debugreg[7])) { + __asm__("movl %0,%%db7" : : "r" (current->thread.debugreg[7])); + } + + /* Whee! Actually deliver the signal. */ + handle_signal(signr, &info, &ka, oldset, regs); + return 1; + } + + no_signal: + /* Did we come from a system call? */ + if (regs->orig_eax >= 0) { + /* Restart the system call - no handlers present */ + if (regs->eax == -ERESTARTNOHAND || + regs->eax == -ERESTARTSYS || + regs->eax == -ERESTARTNOINTR) { + regs->eax = regs->orig_eax; + regs->eip -= 2; + } + if (regs->eax == -ERESTART_RESTARTBLOCK){ + regs->eax = __NR_restart_syscall; + regs->eip -= 2; + } + } + return 0; +} + +/* + * notification of userspace execution resumption + * - triggered by current->work.notify_resume + */ +__attribute__((regparm(3))) +void do_notify_resume(struct pt_regs *regs, sigset_t *oldset, + __u32 thread_info_flags) +{ + /* Pending single-step? */ + if (thread_info_flags & _TIF_SINGLESTEP) { + regs->eflags |= TF_MASK; + clear_thread_flag(TIF_SINGLESTEP); + } + /* deal with pending signal delivery */ + if (thread_info_flags & _TIF_SIGPENDING) + do_signal(regs,oldset); + + clear_thread_flag(TIF_IRET); +} diff --git a/arch/i386/kernel/smp.c b/arch/i386/kernel/smp.c new file mode 100644 index 00000000000..6223c33ac91 --- /dev/null +++ b/arch/i386/kernel/smp.c @@ -0,0 +1,612 @@ +/* + * Intel SMP support routines. + * + * (c) 1995 Alan Cox, Building #3 <alan@redhat.com> + * (c) 1998-99, 2000 Ingo Molnar <mingo@redhat.com> + * + * This code is released under the GNU General Public License version 2 or + * later. + */ + +#include <linux/init.h> + +#include <linux/mm.h> +#include <linux/irq.h> +#include <linux/delay.h> +#include <linux/spinlock.h> +#include <linux/smp_lock.h> +#include <linux/kernel_stat.h> +#include <linux/mc146818rtc.h> +#include <linux/cache.h> +#include <linux/interrupt.h> + +#include <asm/mtrr.h> +#include <asm/tlbflush.h> +#include <mach_apic.h> + +/* + * Some notes on x86 processor bugs affecting SMP operation: + * + * Pentium, Pentium Pro, II, III (and all CPUs) have bugs. + * The Linux implications for SMP are handled as follows: + * + * Pentium III / [Xeon] + * None of the E1AP-E3AP errata are visible to the user. + * + * E1AP. see PII A1AP + * E2AP. see PII A2AP + * E3AP. see PII A3AP + * + * Pentium II / [Xeon] + * None of the A1AP-A3AP errata are visible to the user. + * + * A1AP. see PPro 1AP + * A2AP. see PPro 2AP + * A3AP. see PPro 7AP + * + * Pentium Pro + * None of 1AP-9AP errata are visible to the normal user, + * except occasional delivery of 'spurious interrupt' as trap #15. + * This is very rare and a non-problem. + * + * 1AP. Linux maps APIC as non-cacheable + * 2AP. worked around in hardware + * 3AP. fixed in C0 and above steppings microcode update. + * Linux does not use excessive STARTUP_IPIs. + * 4AP. worked around in hardware + * 5AP. symmetric IO mode (normal Linux operation) not affected. + * 'noapic' mode has vector 0xf filled out properly. + * 6AP. 'noapic' mode might be affected - fixed in later steppings + * 7AP. We do not assume writes to the LVT deassering IRQs + * 8AP. We do not enable low power mode (deep sleep) during MP bootup + * 9AP. We do not use mixed mode + * + * Pentium + * There is a marginal case where REP MOVS on 100MHz SMP + * machines with B stepping processors can fail. XXX should provide + * an L1cache=Writethrough or L1cache=off option. + * + * B stepping CPUs may hang. There are hardware work arounds + * for this. We warn about it in case your board doesn't have the work + * arounds. Basically thats so I can tell anyone with a B stepping + * CPU and SMP problems "tough". + * + * Specific items [From Pentium Processor Specification Update] + * + * 1AP. Linux doesn't use remote read + * 2AP. Linux doesn't trust APIC errors + * 3AP. We work around this + * 4AP. Linux never generated 3 interrupts of the same priority + * to cause a lost local interrupt. + * 5AP. Remote read is never used + * 6AP. not affected - worked around in hardware + * 7AP. not affected - worked around in hardware + * 8AP. worked around in hardware - we get explicit CS errors if not + * 9AP. only 'noapic' mode affected. Might generate spurious + * interrupts, we log only the first one and count the + * rest silently. + * 10AP. not affected - worked around in hardware + * 11AP. Linux reads the APIC between writes to avoid this, as per + * the documentation. Make sure you preserve this as it affects + * the C stepping chips too. + * 12AP. not affected - worked around in hardware + * 13AP. not affected - worked around in hardware + * 14AP. we always deassert INIT during bootup + * 15AP. not affected - worked around in hardware + * 16AP. not affected - worked around in hardware + * 17AP. not affected - worked around in hardware + * 18AP. not affected - worked around in hardware + * 19AP. not affected - worked around in BIOS + * + * If this sounds worrying believe me these bugs are either ___RARE___, + * or are signal timing bugs worked around in hardware and there's + * about nothing of note with C stepping upwards. + */ + +DEFINE_PER_CPU(struct tlb_state, cpu_tlbstate) ____cacheline_aligned = { &init_mm, 0, }; + +/* + * the following functions deal with sending IPIs between CPUs. + * + * We use 'broadcast', CPU->CPU IPIs and self-IPIs too. + */ + +static inline int __prepare_ICR (unsigned int shortcut, int vector) +{ + return APIC_DM_FIXED | shortcut | vector | APIC_DEST_LOGICAL; +} + +static inline int __prepare_ICR2 (unsigned int mask) +{ + return SET_APIC_DEST_FIELD(mask); +} + +void __send_IPI_shortcut(unsigned int shortcut, int vector) +{ + /* + * Subtle. In the case of the 'never do double writes' workaround + * we have to lock out interrupts to be safe. As we don't care + * of the value read we use an atomic rmw access to avoid costly + * cli/sti. Otherwise we use an even cheaper single atomic write + * to the APIC. + */ + unsigned int cfg; + + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + + /* + * No need to touch the target chip field + */ + cfg = __prepare_ICR(shortcut, vector); + + /* + * Send the IPI. The write to APIC_ICR fires this off. + */ + apic_write_around(APIC_ICR, cfg); +} + +void fastcall send_IPI_self(int vector) +{ + __send_IPI_shortcut(APIC_DEST_SELF, vector); +} + +/* + * This is only used on smaller machines. + */ +void send_IPI_mask_bitmask(cpumask_t cpumask, int vector) +{ + unsigned long mask = cpus_addr(cpumask)[0]; + unsigned long cfg; + unsigned long flags; + + local_irq_save(flags); + + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + + /* + * prepare target chip field + */ + cfg = __prepare_ICR2(mask); + apic_write_around(APIC_ICR2, cfg); + + /* + * program the ICR + */ + cfg = __prepare_ICR(0, vector); + + /* + * Send the IPI. The write to APIC_ICR fires this off. + */ + apic_write_around(APIC_ICR, cfg); + + local_irq_restore(flags); +} + +void send_IPI_mask_sequence(cpumask_t mask, int vector) +{ + unsigned long cfg, flags; + unsigned int query_cpu; + + /* + * Hack. The clustered APIC addressing mode doesn't allow us to send + * to an arbitrary mask, so I do a unicasts to each CPU instead. This + * should be modified to do 1 message per cluster ID - mbligh + */ + + local_irq_save(flags); + + for (query_cpu = 0; query_cpu < NR_CPUS; ++query_cpu) { + if (cpu_isset(query_cpu, mask)) { + + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + + /* + * prepare target chip field + */ + cfg = __prepare_ICR2(cpu_to_logical_apicid(query_cpu)); + apic_write_around(APIC_ICR2, cfg); + + /* + * program the ICR + */ + cfg = __prepare_ICR(0, vector); + + /* + * Send the IPI. The write to APIC_ICR fires this off. + */ + apic_write_around(APIC_ICR, cfg); + } + } + local_irq_restore(flags); +} + +#include <mach_ipi.h> /* must come after the send_IPI functions above for inlining */ + +/* + * Smarter SMP flushing macros. + * c/o Linus Torvalds. + * + * These mean you can really definitely utterly forget about + * writing to user space from interrupts. (Its not allowed anyway). + * + * Optimizations Manfred Spraul <manfred@colorfullife.com> + */ + +static cpumask_t flush_cpumask; +static struct mm_struct * flush_mm; +static unsigned long flush_va; +static DEFINE_SPINLOCK(tlbstate_lock); +#define FLUSH_ALL 0xffffffff + +/* + * We cannot call mmdrop() because we are in interrupt context, + * instead update mm->cpu_vm_mask. + * + * We need to reload %cr3 since the page tables may be going + * away from under us.. + */ +static inline void leave_mm (unsigned long cpu) +{ + if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) + BUG(); + cpu_clear(cpu, per_cpu(cpu_tlbstate, cpu).active_mm->cpu_vm_mask); + load_cr3(swapper_pg_dir); +} + +/* + * + * The flush IPI assumes that a thread switch happens in this order: + * [cpu0: the cpu that switches] + * 1) switch_mm() either 1a) or 1b) + * 1a) thread switch to a different mm + * 1a1) cpu_clear(cpu, old_mm->cpu_vm_mask); + * Stop ipi delivery for the old mm. This is not synchronized with + * the other cpus, but smp_invalidate_interrupt ignore flush ipis + * for the wrong mm, and in the worst case we perform a superflous + * tlb flush. + * 1a2) set cpu_tlbstate to TLBSTATE_OK + * Now the smp_invalidate_interrupt won't call leave_mm if cpu0 + * was in lazy tlb mode. + * 1a3) update cpu_tlbstate[].active_mm + * Now cpu0 accepts tlb flushes for the new mm. + * 1a4) cpu_set(cpu, new_mm->cpu_vm_mask); + * Now the other cpus will send tlb flush ipis. + * 1a4) change cr3. + * 1b) thread switch without mm change + * cpu_tlbstate[].active_mm is correct, cpu0 already handles + * flush ipis. + * 1b1) set cpu_tlbstate to TLBSTATE_OK + * 1b2) test_and_set the cpu bit in cpu_vm_mask. + * Atomically set the bit [other cpus will start sending flush ipis], + * and test the bit. + * 1b3) if the bit was 0: leave_mm was called, flush the tlb. + * 2) switch %%esp, ie current + * + * The interrupt must handle 2 special cases: + * - cr3 is changed before %%esp, ie. it cannot use current->{active_,}mm. + * - the cpu performs speculative tlb reads, i.e. even if the cpu only + * runs in kernel space, the cpu could load tlb entries for user space + * pages. + * + * The good news is that cpu_tlbstate is local to each cpu, no + * write/read ordering problems. + */ + +/* + * TLB flush IPI: + * + * 1) Flush the tlb entries if the cpu uses the mm that's being flushed. + * 2) Leave the mm if we are in the lazy tlb mode. + */ + +fastcall void smp_invalidate_interrupt(struct pt_regs *regs) +{ + unsigned long cpu; + + cpu = get_cpu(); + + if (!cpu_isset(cpu, flush_cpumask)) + goto out; + /* + * This was a BUG() but until someone can quote me the + * line from the intel manual that guarantees an IPI to + * multiple CPUs is retried _only_ on the erroring CPUs + * its staying as a return + * + * BUG(); + */ + + if (flush_mm == per_cpu(cpu_tlbstate, cpu).active_mm) { + if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_OK) { + if (flush_va == FLUSH_ALL) + local_flush_tlb(); + else + __flush_tlb_one(flush_va); + } else + leave_mm(cpu); + } + ack_APIC_irq(); + smp_mb__before_clear_bit(); + cpu_clear(cpu, flush_cpumask); + smp_mb__after_clear_bit(); +out: + put_cpu_no_resched(); +} + +static void flush_tlb_others(cpumask_t cpumask, struct mm_struct *mm, + unsigned long va) +{ + cpumask_t tmp; + /* + * A couple of (to be removed) sanity checks: + * + * - we do not send IPIs to not-yet booted CPUs. + * - current CPU must not be in mask + * - mask must exist :) + */ + BUG_ON(cpus_empty(cpumask)); + + cpus_and(tmp, cpumask, cpu_online_map); + BUG_ON(!cpus_equal(cpumask, tmp)); + BUG_ON(cpu_isset(smp_processor_id(), cpumask)); + BUG_ON(!mm); + + /* + * i'm not happy about this global shared spinlock in the + * MM hot path, but we'll see how contended it is. + * Temporarily this turns IRQs off, so that lockups are + * detected by the NMI watchdog. + */ + spin_lock(&tlbstate_lock); + + flush_mm = mm; + flush_va = va; +#if NR_CPUS <= BITS_PER_LONG + atomic_set_mask(cpumask, &flush_cpumask); +#else + { + int k; + unsigned long *flush_mask = (unsigned long *)&flush_cpumask; + unsigned long *cpu_mask = (unsigned long *)&cpumask; + for (k = 0; k < BITS_TO_LONGS(NR_CPUS); ++k) + atomic_set_mask(cpu_mask[k], &flush_mask[k]); + } +#endif + /* + * We have to send the IPI only to + * CPUs affected. + */ + send_IPI_mask(cpumask, INVALIDATE_TLB_VECTOR); + + while (!cpus_empty(flush_cpumask)) + /* nothing. lockup detection does not belong here */ + mb(); + + flush_mm = NULL; + flush_va = 0; + spin_unlock(&tlbstate_lock); +} + +void flush_tlb_current_task(void) +{ + struct mm_struct *mm = current->mm; + cpumask_t cpu_mask; + + preempt_disable(); + cpu_mask = mm->cpu_vm_mask; + cpu_clear(smp_processor_id(), cpu_mask); + + local_flush_tlb(); + if (!cpus_empty(cpu_mask)) + flush_tlb_others(cpu_mask, mm, FLUSH_ALL); + preempt_enable(); +} + +void flush_tlb_mm (struct mm_struct * mm) +{ + cpumask_t cpu_mask; + + preempt_disable(); + cpu_mask = mm->cpu_vm_mask; + cpu_clear(smp_processor_id(), cpu_mask); + + if (current->active_mm == mm) { + if (current->mm) + local_flush_tlb(); + else + leave_mm(smp_processor_id()); + } + if (!cpus_empty(cpu_mask)) + flush_tlb_others(cpu_mask, mm, FLUSH_ALL); + + preempt_enable(); +} + +void flush_tlb_page(struct vm_area_struct * vma, unsigned long va) +{ + struct mm_struct *mm = vma->vm_mm; + cpumask_t cpu_mask; + + preempt_disable(); + cpu_mask = mm->cpu_vm_mask; + cpu_clear(smp_processor_id(), cpu_mask); + + if (current->active_mm == mm) { + if(current->mm) + __flush_tlb_one(va); + else + leave_mm(smp_processor_id()); + } + + if (!cpus_empty(cpu_mask)) + flush_tlb_others(cpu_mask, mm, va); + + preempt_enable(); +} + +static void do_flush_tlb_all(void* info) +{ + unsigned long cpu = smp_processor_id(); + + __flush_tlb_all(); + if (per_cpu(cpu_tlbstate, cpu).state == TLBSTATE_LAZY) + leave_mm(cpu); +} + +void flush_tlb_all(void) +{ + on_each_cpu(do_flush_tlb_all, NULL, 1, 1); +} + +/* + * this function sends a 'reschedule' IPI to another CPU. + * it goes straight through and wastes no time serializing + * anything. Worst case is that we lose a reschedule ... + */ +void smp_send_reschedule(int cpu) +{ + send_IPI_mask(cpumask_of_cpu(cpu), RESCHEDULE_VECTOR); +} + +/* + * Structure and data for smp_call_function(). This is designed to minimise + * static memory requirements. It also looks cleaner. + */ +static DEFINE_SPINLOCK(call_lock); + +struct call_data_struct { + void (*func) (void *info); + void *info; + atomic_t started; + atomic_t finished; + int wait; +}; + +static struct call_data_struct * call_data; + +/* + * this function sends a 'generic call function' IPI to all other CPUs + * in the system. + */ + +int smp_call_function (void (*func) (void *info), void *info, int nonatomic, + int wait) +/* + * [SUMMARY] Run a function on all other CPUs. + * <func> The function to run. This must be fast and non-blocking. + * <info> An arbitrary pointer to pass to the function. + * <nonatomic> currently unused. + * <wait> If true, wait (atomically) until function has completed on other CPUs. + * [RETURNS] 0 on success, else a negative status code. Does not return until + * remote CPUs are nearly ready to execute <<func>> or are or have executed. + * + * You must not call this function with disabled interrupts or from a + * hardware interrupt handler or from a bottom half handler. + */ +{ + struct call_data_struct data; + int cpus = num_online_cpus()-1; + + if (!cpus) + return 0; + + /* Can deadlock when called with interrupts disabled */ + WARN_ON(irqs_disabled()); + + data.func = func; + data.info = info; + atomic_set(&data.started, 0); + data.wait = wait; + if (wait) + atomic_set(&data.finished, 0); + + spin_lock(&call_lock); + call_data = &data; + mb(); + + /* Send a message to all other CPUs and wait for them to respond */ + send_IPI_allbutself(CALL_FUNCTION_VECTOR); + + /* Wait for response */ + while (atomic_read(&data.started) != cpus) + cpu_relax(); + + if (wait) + while (atomic_read(&data.finished) != cpus) + cpu_relax(); + spin_unlock(&call_lock); + + return 0; +} + +static void stop_this_cpu (void * dummy) +{ + /* + * Remove this CPU: + */ + cpu_clear(smp_processor_id(), cpu_online_map); + local_irq_disable(); + disable_local_APIC(); + if (cpu_data[smp_processor_id()].hlt_works_ok) + for(;;) __asm__("hlt"); + for (;;); +} + +/* + * this function calls the 'stop' function on all other CPUs in the system. + */ + +void smp_send_stop(void) +{ + smp_call_function(stop_this_cpu, NULL, 1, 0); + + local_irq_disable(); + disable_local_APIC(); + local_irq_enable(); +} + +/* + * Reschedule call back. Nothing to do, + * all the work is done automatically when + * we return from the interrupt. + */ +fastcall void smp_reschedule_interrupt(struct pt_regs *regs) +{ + ack_APIC_irq(); +} + +fastcall void smp_call_function_interrupt(struct pt_regs *regs) +{ + void (*func) (void *info) = call_data->func; + void *info = call_data->info; + int wait = call_data->wait; + + ack_APIC_irq(); + /* + * Notify initiating CPU that I've grabbed the data and am + * about to execute the function + */ + mb(); + atomic_inc(&call_data->started); + /* + * At this point the info structure may be out of scope unless wait==1 + */ + irq_enter(); + (*func)(info); + irq_exit(); + + if (wait) { + mb(); + atomic_inc(&call_data->finished); + } +} + diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c new file mode 100644 index 00000000000..332ee7a1d1a --- /dev/null +++ b/arch/i386/kernel/smpboot.c @@ -0,0 +1,1145 @@ +/* + * x86 SMP booting functions + * + * (c) 1995 Alan Cox, Building #3 <alan@redhat.com> + * (c) 1998, 1999, 2000 Ingo Molnar <mingo@redhat.com> + * + * Much of the core SMP work is based on previous work by Thomas Radke, to + * whom a great many thanks are extended. + * + * Thanks to Intel for making available several different Pentium, + * Pentium Pro and Pentium-II/Xeon MP machines. + * Original development of Linux SMP code supported by Caldera. + * + * This code is released under the GNU General Public License version 2 or + * later. + * + * Fixes + * Felix Koop : NR_CPUS used properly + * Jose Renau : Handle single CPU case. + * Alan Cox : By repeated request 8) - Total BogoMIPS report. + * Greg Wright : Fix for kernel stacks panic. + * Erich Boleyn : MP v1.4 and additional changes. + * Matthias Sattler : Changes for 2.1 kernel map. + * Michel Lespinasse : Changes for 2.1 kernel map. + * Michael Chastain : Change trampoline.S to gnu as. + * Alan Cox : Dumb bug: 'B' step PPro's are fine + * Ingo Molnar : Added APIC timers, based on code + * from Jose Renau + * Ingo Molnar : various cleanups and rewrites + * Tigran Aivazian : fixed "0.00 in /proc/uptime on SMP" bug. + * Maciej W. Rozycki : Bits for genuine 82489DX APICs + * Martin J. Bligh : Added support for multi-quad systems + * Dave Jones : Report invalid combinations of Athlon CPUs. +* Rusty Russell : Hacked into shape for new "hotplug" boot process. */ + +#include <linux/module.h> +#include <linux/config.h> +#include <linux/init.h> +#include <linux/kernel.h> + +#include <linux/mm.h> +#include <linux/sched.h> +#include <linux/kernel_stat.h> +#include <linux/smp_lock.h> +#include <linux/irq.h> +#include <linux/bootmem.h> + +#include <linux/delay.h> +#include <linux/mc146818rtc.h> +#include <asm/tlbflush.h> +#include <asm/desc.h> +#include <asm/arch_hooks.h> + +#include <mach_apic.h> +#include <mach_wakecpu.h> +#include <smpboot_hooks.h> + +/* Set if we find a B stepping CPU */ +static int __initdata smp_b_stepping; + +/* Number of siblings per CPU package */ +int smp_num_siblings = 1; +int phys_proc_id[NR_CPUS]; /* Package ID of each logical CPU */ +EXPORT_SYMBOL(phys_proc_id); + +/* bitmap of online cpus */ +cpumask_t cpu_online_map; + +cpumask_t cpu_callin_map; +cpumask_t cpu_callout_map; +static cpumask_t smp_commenced_mask; + +/* Per CPU bogomips and other parameters */ +struct cpuinfo_x86 cpu_data[NR_CPUS] __cacheline_aligned; + +u8 x86_cpu_to_apicid[NR_CPUS] = + { [0 ... NR_CPUS-1] = 0xff }; +EXPORT_SYMBOL(x86_cpu_to_apicid); + +/* + * Trampoline 80x86 program as an array. + */ + +extern unsigned char trampoline_data []; +extern unsigned char trampoline_end []; +static unsigned char *trampoline_base; +static int trampoline_exec; + +static void map_cpu_to_logical_apicid(void); + +/* + * Currently trivial. Write the real->protected mode + * bootstrap into the page concerned. The caller + * has made sure it's suitably aligned. + */ + +static unsigned long __init setup_trampoline(void) +{ + memcpy(trampoline_base, trampoline_data, trampoline_end - trampoline_data); + return virt_to_phys(trampoline_base); +} + +/* + * We are called very early to get the low memory for the + * SMP bootup trampoline page. + */ +void __init smp_alloc_memory(void) +{ + trampoline_base = (void *) alloc_bootmem_low_pages(PAGE_SIZE); + /* + * Has to be in very low memory so we can execute + * real-mode AP code. + */ + if (__pa(trampoline_base) >= 0x9F000) + BUG(); + /* + * Make the SMP trampoline executable: + */ + trampoline_exec = set_kernel_exec((unsigned long)trampoline_base, 1); +} + +/* + * The bootstrap kernel entry code has set these up. Save them for + * a given CPU + */ + +static void __init smp_store_cpu_info(int id) +{ + struct cpuinfo_x86 *c = cpu_data + id; + + *c = boot_cpu_data; + if (id!=0) + identify_cpu(c); + /* + * Mask B, Pentium, but not Pentium MMX + */ + if (c->x86_vendor == X86_VENDOR_INTEL && + c->x86 == 5 && + c->x86_mask >= 1 && c->x86_mask <= 4 && + c->x86_model <= 3) + /* + * Remember we have B step Pentia with bugs + */ + smp_b_stepping = 1; + + /* + * Certain Athlons might work (for various values of 'work') in SMP + * but they are not certified as MP capable. + */ + if ((c->x86_vendor == X86_VENDOR_AMD) && (c->x86 == 6)) { + + /* Athlon 660/661 is valid. */ + if ((c->x86_model==6) && ((c->x86_mask==0) || (c->x86_mask==1))) + goto valid_k7; + + /* Duron 670 is valid */ + if ((c->x86_model==7) && (c->x86_mask==0)) + goto valid_k7; + + /* + * Athlon 662, Duron 671, and Athlon >model 7 have capability bit. + * It's worth noting that the A5 stepping (662) of some Athlon XP's + * have the MP bit set. + * See http://www.heise.de/newsticker/data/jow-18.10.01-000 for more. + */ + if (((c->x86_model==6) && (c->x86_mask>=2)) || + ((c->x86_model==7) && (c->x86_mask>=1)) || + (c->x86_model> 7)) + if (cpu_has_mp) + goto valid_k7; + + /* If we get here, it's not a certified SMP capable AMD system. */ + tainted |= TAINT_UNSAFE_SMP; + } + +valid_k7: + ; +} + +/* + * TSC synchronization. + * + * We first check whether all CPUs have their TSC's synchronized, + * then we print a warning if not, and always resync. + */ + +static atomic_t tsc_start_flag = ATOMIC_INIT(0); +static atomic_t tsc_count_start = ATOMIC_INIT(0); +static atomic_t tsc_count_stop = ATOMIC_INIT(0); +static unsigned long long tsc_values[NR_CPUS]; + +#define NR_LOOPS 5 + +static void __init synchronize_tsc_bp (void) +{ + int i; + unsigned long long t0; + unsigned long long sum, avg; + long long delta; + unsigned long one_usec; + int buggy = 0; + + printk(KERN_INFO "checking TSC synchronization across %u CPUs: ", num_booting_cpus()); + + /* convert from kcyc/sec to cyc/usec */ + one_usec = cpu_khz / 1000; + + atomic_set(&tsc_start_flag, 1); + wmb(); + + /* + * We loop a few times to get a primed instruction cache, + * then the last pass is more or less synchronized and + * the BP and APs set their cycle counters to zero all at + * once. This reduces the chance of having random offsets + * between the processors, and guarantees that the maximum + * delay between the cycle counters is never bigger than + * the latency of information-passing (cachelines) between + * two CPUs. + */ + for (i = 0; i < NR_LOOPS; i++) { + /* + * all APs synchronize but they loop on '== num_cpus' + */ + while (atomic_read(&tsc_count_start) != num_booting_cpus()-1) + mb(); + atomic_set(&tsc_count_stop, 0); + wmb(); + /* + * this lets the APs save their current TSC: + */ + atomic_inc(&tsc_count_start); + + rdtscll(tsc_values[smp_processor_id()]); + /* + * We clear the TSC in the last loop: + */ + if (i == NR_LOOPS-1) + write_tsc(0, 0); + + /* + * Wait for all APs to leave the synchronization point: + */ + while (atomic_read(&tsc_count_stop) != num_booting_cpus()-1) + mb(); + atomic_set(&tsc_count_start, 0); + wmb(); + atomic_inc(&tsc_count_stop); + } + + sum = 0; + for (i = 0; i < NR_CPUS; i++) { + if (cpu_isset(i, cpu_callout_map)) { + t0 = tsc_values[i]; + sum += t0; + } + } + avg = sum; + do_div(avg, num_booting_cpus()); + + sum = 0; + for (i = 0; i < NR_CPUS; i++) { + if (!cpu_isset(i, cpu_callout_map)) + continue; + delta = tsc_values[i] - avg; + if (delta < 0) + delta = -delta; + /* + * We report bigger than 2 microseconds clock differences. + */ + if (delta > 2*one_usec) { + long realdelta; + if (!buggy) { + buggy = 1; + printk("\n"); + } + realdelta = delta; + do_div(realdelta, one_usec); + if (tsc_values[i] < avg) + realdelta = -realdelta; + + printk(KERN_INFO "CPU#%d had %ld usecs TSC skew, fixed it up.\n", i, realdelta); + } + + sum += delta; + } + if (!buggy) + printk("passed.\n"); +} + +static void __init synchronize_tsc_ap (void) +{ + int i; + + /* + * Not every cpu is online at the time + * this gets called, so we first wait for the BP to + * finish SMP initialization: + */ + while (!atomic_read(&tsc_start_flag)) mb(); + + for (i = 0; i < NR_LOOPS; i++) { + atomic_inc(&tsc_count_start); + while (atomic_read(&tsc_count_start) != num_booting_cpus()) + mb(); + + rdtscll(tsc_values[smp_processor_id()]); + if (i == NR_LOOPS-1) + write_tsc(0, 0); + + atomic_inc(&tsc_count_stop); + while (atomic_read(&tsc_count_stop) != num_booting_cpus()) mb(); + } +} +#undef NR_LOOPS + +extern void calibrate_delay(void); + +static atomic_t init_deasserted; + +static void __init smp_callin(void) +{ + int cpuid, phys_id; + unsigned long timeout; + + /* + * If waken up by an INIT in an 82489DX configuration + * we may get here before an INIT-deassert IPI reaches + * our local APIC. We have to wait for the IPI or we'll + * lock up on an APIC access. + */ + wait_for_init_deassert(&init_deasserted); + + /* + * (This works even if the APIC is not enabled.) + */ + phys_id = GET_APIC_ID(apic_read(APIC_ID)); + cpuid = smp_processor_id(); + if (cpu_isset(cpuid, cpu_callin_map)) { + printk("huh, phys CPU#%d, CPU#%d already present??\n", + phys_id, cpuid); + BUG(); + } + Dprintk("CPU#%d (phys ID: %d) waiting for CALLOUT\n", cpuid, phys_id); + + /* + * STARTUP IPIs are fragile beasts as they might sometimes + * trigger some glue motherboard logic. Complete APIC bus + * silence for 1 second, this overestimates the time the + * boot CPU is spending to send the up to 2 STARTUP IPIs + * by a factor of two. This should be enough. + */ + + /* + * Waiting 2s total for startup (udelay is not yet working) + */ + timeout = jiffies + 2*HZ; + while (time_before(jiffies, timeout)) { + /* + * Has the boot CPU finished it's STARTUP sequence? + */ + if (cpu_isset(cpuid, cpu_callout_map)) + break; + rep_nop(); + } + + if (!time_before(jiffies, timeout)) { + printk("BUG: CPU%d started up but did not get a callout!\n", + cpuid); + BUG(); + } + + /* + * the boot CPU has finished the init stage and is spinning + * on callin_map until we finish. We are free to set up this + * CPU, first the APIC. (this is probably redundant on most + * boards) + */ + + Dprintk("CALLIN, before setup_local_APIC().\n"); + smp_callin_clear_local_apic(); + setup_local_APIC(); + map_cpu_to_logical_apicid(); + + /* + * Get our bogomips. + */ + calibrate_delay(); + Dprintk("Stack at about %p\n",&cpuid); + + /* + * Save our processor parameters + */ + smp_store_cpu_info(cpuid); + + disable_APIC_timer(); + + /* + * Allow the master to continue. + */ + cpu_set(cpuid, cpu_callin_map); + + /* + * Synchronize the TSC with the BP + */ + if (cpu_has_tsc && cpu_khz) + synchronize_tsc_ap(); +} + +static int cpucount; + +/* + * Activate a secondary processor. + */ +static void __init start_secondary(void *unused) +{ + /* + * Dont put anything before smp_callin(), SMP + * booting is too fragile that we want to limit the + * things done here to the most necessary things. + */ + cpu_init(); + smp_callin(); + while (!cpu_isset(smp_processor_id(), smp_commenced_mask)) + rep_nop(); + setup_secondary_APIC_clock(); + if (nmi_watchdog == NMI_IO_APIC) { + disable_8259A_irq(0); + enable_NMI_through_LVT0(NULL); + enable_8259A_irq(0); + } + enable_APIC_timer(); + /* + * low-memory mappings have been cleared, flush them from + * the local TLBs too. + */ + local_flush_tlb(); + cpu_set(smp_processor_id(), cpu_online_map); + + /* We can take interrupts now: we're officially "up". */ + local_irq_enable(); + + wmb(); + cpu_idle(); +} + +/* + * Everything has been set up for the secondary + * CPUs - they just need to reload everything + * from the task structure + * This function must not return. + */ +void __init initialize_secondary(void) +{ + /* + * We don't actually need to load the full TSS, + * basically just the stack pointer and the eip. + */ + + asm volatile( + "movl %0,%%esp\n\t" + "jmp *%1" + : + :"r" (current->thread.esp),"r" (current->thread.eip)); +} + +extern struct { + void * esp; + unsigned short ss; +} stack_start; + +#ifdef CONFIG_NUMA + +/* which logical CPUs are on which nodes */ +cpumask_t node_2_cpu_mask[MAX_NUMNODES] = + { [0 ... MAX_NUMNODES-1] = CPU_MASK_NONE }; +/* which node each logical CPU is on */ +int cpu_2_node[NR_CPUS] = { [0 ... NR_CPUS-1] = 0 }; +EXPORT_SYMBOL(cpu_2_node); + +/* set up a mapping between cpu and node. */ +static inline void map_cpu_to_node(int cpu, int node) +{ + printk("Mapping cpu %d to node %d\n", cpu, node); + cpu_set(cpu, node_2_cpu_mask[node]); + cpu_2_node[cpu] = node; +} + +/* undo a mapping between cpu and node. */ +static inline void unmap_cpu_to_node(int cpu) +{ + int node; + + printk("Unmapping cpu %d from all nodes\n", cpu); + for (node = 0; node < MAX_NUMNODES; node ++) + cpu_clear(cpu, node_2_cpu_mask[node]); + cpu_2_node[cpu] = 0; +} +#else /* !CONFIG_NUMA */ + +#define map_cpu_to_node(cpu, node) ({}) +#define unmap_cpu_to_node(cpu) ({}) + +#endif /* CONFIG_NUMA */ + +u8 cpu_2_logical_apicid[NR_CPUS] = { [0 ... NR_CPUS-1] = BAD_APICID }; + +static void map_cpu_to_logical_apicid(void) +{ + int cpu = smp_processor_id(); + int apicid = logical_smp_processor_id(); + + cpu_2_logical_apicid[cpu] = apicid; + map_cpu_to_node(cpu, apicid_to_node(apicid)); +} + +static void unmap_cpu_to_logical_apicid(int cpu) +{ + cpu_2_logical_apicid[cpu] = BAD_APICID; + unmap_cpu_to_node(cpu); +} + +#if APIC_DEBUG +static inline void __inquire_remote_apic(int apicid) +{ + int i, regs[] = { APIC_ID >> 4, APIC_LVR >> 4, APIC_SPIV >> 4 }; + char *names[] = { "ID", "VERSION", "SPIV" }; + int timeout, status; + + printk("Inquiring remote APIC #%d...\n", apicid); + + for (i = 0; i < sizeof(regs) / sizeof(*regs); i++) { + printk("... APIC #%d %s: ", apicid, names[i]); + + /* + * Wait for idle. + */ + apic_wait_icr_idle(); + + apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(apicid)); + apic_write_around(APIC_ICR, APIC_DM_REMRD | regs[i]); + + timeout = 0; + do { + udelay(100); + status = apic_read(APIC_ICR) & APIC_ICR_RR_MASK; + } while (status == APIC_ICR_RR_INPROG && timeout++ < 1000); + + switch (status) { + case APIC_ICR_RR_VALID: + status = apic_read(APIC_RRR); + printk("%08x\n", status); + break; + default: + printk("failed\n"); + } + } +} +#endif + +#ifdef WAKE_SECONDARY_VIA_NMI +/* + * Poke the other CPU in the eye via NMI to wake it up. Remember that the normal + * INIT, INIT, STARTUP sequence will reset the chip hard for us, and this + * won't ... remember to clear down the APIC, etc later. + */ +static int __init +wakeup_secondary_cpu(int logical_apicid, unsigned long start_eip) +{ + unsigned long send_status = 0, accept_status = 0; + int timeout, maxlvt; + + /* Target chip */ + apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(logical_apicid)); + + /* Boot on the stack */ + /* Kick the second */ + apic_write_around(APIC_ICR, APIC_DM_NMI | APIC_DEST_LOGICAL); + + Dprintk("Waiting for send to finish...\n"); + timeout = 0; + do { + Dprintk("+"); + udelay(100); + send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY; + } while (send_status && (timeout++ < 1000)); + + /* + * Give the other CPU some time to accept the IPI. + */ + udelay(200); + /* + * Due to the Pentium erratum 3AP. + */ + maxlvt = get_maxlvt(); + if (maxlvt > 3) { + apic_read_around(APIC_SPIV); + apic_write(APIC_ESR, 0); + } + accept_status = (apic_read(APIC_ESR) & 0xEF); + Dprintk("NMI sent.\n"); + + if (send_status) + printk("APIC never delivered???\n"); + if (accept_status) + printk("APIC delivery error (%lx).\n", accept_status); + + return (send_status | accept_status); +} +#endif /* WAKE_SECONDARY_VIA_NMI */ + +#ifdef WAKE_SECONDARY_VIA_INIT +static int __init +wakeup_secondary_cpu(int phys_apicid, unsigned long start_eip) +{ + unsigned long send_status = 0, accept_status = 0; + int maxlvt, timeout, num_starts, j; + + /* + * Be paranoid about clearing APIC errors. + */ + if (APIC_INTEGRATED(apic_version[phys_apicid])) { + apic_read_around(APIC_SPIV); + apic_write(APIC_ESR, 0); + apic_read(APIC_ESR); + } + + Dprintk("Asserting INIT.\n"); + + /* + * Turn INIT on target chip + */ + apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid)); + + /* + * Send IPI + */ + apic_write_around(APIC_ICR, APIC_INT_LEVELTRIG | APIC_INT_ASSERT + | APIC_DM_INIT); + + Dprintk("Waiting for send to finish...\n"); + timeout = 0; + do { + Dprintk("+"); + udelay(100); + send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY; + } while (send_status && (timeout++ < 1000)); + + mdelay(10); + + Dprintk("Deasserting INIT.\n"); + + /* Target chip */ + apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid)); + + /* Send IPI */ + apic_write_around(APIC_ICR, APIC_INT_LEVELTRIG | APIC_DM_INIT); + + Dprintk("Waiting for send to finish...\n"); + timeout = 0; + do { + Dprintk("+"); + udelay(100); + send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY; + } while (send_status && (timeout++ < 1000)); + + atomic_set(&init_deasserted, 1); + + /* + * Should we send STARTUP IPIs ? + * + * Determine this based on the APIC version. + * If we don't have an integrated APIC, don't send the STARTUP IPIs. + */ + if (APIC_INTEGRATED(apic_version[phys_apicid])) + num_starts = 2; + else + num_starts = 0; + + /* + * Run STARTUP IPI loop. + */ + Dprintk("#startup loops: %d.\n", num_starts); + + maxlvt = get_maxlvt(); + + for (j = 1; j <= num_starts; j++) { + Dprintk("Sending STARTUP #%d.\n",j); + apic_read_around(APIC_SPIV); + apic_write(APIC_ESR, 0); + apic_read(APIC_ESR); + Dprintk("After apic_write.\n"); + + /* + * STARTUP IPI + */ + + /* Target chip */ + apic_write_around(APIC_ICR2, SET_APIC_DEST_FIELD(phys_apicid)); + + /* Boot on the stack */ + /* Kick the second */ + apic_write_around(APIC_ICR, APIC_DM_STARTUP + | (start_eip >> 12)); + + /* + * Give the other CPU some time to accept the IPI. + */ + udelay(300); + + Dprintk("Startup point 1.\n"); + + Dprintk("Waiting for send to finish...\n"); + timeout = 0; + do { + Dprintk("+"); + udelay(100); + send_status = apic_read(APIC_ICR) & APIC_ICR_BUSY; + } while (send_status && (timeout++ < 1000)); + + /* + * Give the other CPU some time to accept the IPI. + */ + udelay(200); + /* + * Due to the Pentium erratum 3AP. + */ + if (maxlvt > 3) { + apic_read_around(APIC_SPIV); + apic_write(APIC_ESR, 0); + } + accept_status = (apic_read(APIC_ESR) & 0xEF); + if (send_status || accept_status) + break; + } + Dprintk("After Startup.\n"); + + if (send_status) + printk("APIC never delivered???\n"); + if (accept_status) + printk("APIC delivery error (%lx).\n", accept_status); + + return (send_status | accept_status); +} +#endif /* WAKE_SECONDARY_VIA_INIT */ + +extern cpumask_t cpu_initialized; + +static int __init do_boot_cpu(int apicid) +/* + * NOTE - on most systems this is a PHYSICAL apic ID, but on multiquad + * (ie clustered apic addressing mode), this is a LOGICAL apic ID. + * Returns zero if CPU booted OK, else error code from wakeup_secondary_cpu. + */ +{ + struct task_struct *idle; + unsigned long boot_error; + int timeout, cpu; + unsigned long start_eip; + unsigned short nmi_high = 0, nmi_low = 0; + + cpu = ++cpucount; + /* + * We can't use kernel_thread since we must avoid to + * reschedule the child. + */ + idle = fork_idle(cpu); + if (IS_ERR(idle)) + panic("failed fork for CPU %d", cpu); + idle->thread.eip = (unsigned long) start_secondary; + /* start_eip had better be page-aligned! */ + start_eip = setup_trampoline(); + + /* So we see what's up */ + printk("Booting processor %d/%d eip %lx\n", cpu, apicid, start_eip); + /* Stack for startup_32 can be just as for start_secondary onwards */ + stack_start.esp = (void *) idle->thread.esp; + + irq_ctx_init(cpu); + + /* + * This grunge runs the startup process for + * the targeted processor. + */ + + atomic_set(&init_deasserted, 0); + + Dprintk("Setting warm reset code and vector.\n"); + + store_NMI_vector(&nmi_high, &nmi_low); + + smpboot_setup_warm_reset_vector(start_eip); + + /* + * Starting actual IPI sequence... + */ + boot_error = wakeup_secondary_cpu(apicid, start_eip); + + if (!boot_error) { + /* + * allow APs to start initializing. + */ + Dprintk("Before Callout %d.\n", cpu); + cpu_set(cpu, cpu_callout_map); + Dprintk("After Callout %d.\n", cpu); + + /* + * Wait 5s total for a response + */ + for (timeout = 0; timeout < 50000; timeout++) { + if (cpu_isset(cpu, cpu_callin_map)) + break; /* It has booted */ + udelay(100); + } + + if (cpu_isset(cpu, cpu_callin_map)) { + /* number CPUs logically, starting from 1 (BSP is 0) */ + Dprintk("OK.\n"); + printk("CPU%d: ", cpu); + print_cpu_info(&cpu_data[cpu]); + Dprintk("CPU has booted.\n"); + } else { + boot_error= 1; + if (*((volatile unsigned char *)trampoline_base) + == 0xA5) + /* trampoline started but...? */ + printk("Stuck ??\n"); + else + /* trampoline code not run */ + printk("Not responding.\n"); + inquire_remote_apic(apicid); + } + } + x86_cpu_to_apicid[cpu] = apicid; + if (boot_error) { + /* Try to put things back the way they were before ... */ + unmap_cpu_to_logical_apicid(cpu); + cpu_clear(cpu, cpu_callout_map); /* was set here (do_boot_cpu()) */ + cpu_clear(cpu, cpu_initialized); /* was set by cpu_init() */ + cpucount--; + } + + /* mark "stuck" area as not stuck */ + *((volatile unsigned long *)trampoline_base) = 0; + + return boot_error; +} + +static void smp_tune_scheduling (void) +{ + unsigned long cachesize; /* kB */ + unsigned long bandwidth = 350; /* MB/s */ + /* + * Rough estimation for SMP scheduling, this is the number of + * cycles it takes for a fully memory-limited process to flush + * the SMP-local cache. + * + * (For a P5 this pretty much means we will choose another idle + * CPU almost always at wakeup time (this is due to the small + * L1 cache), on PIIs it's around 50-100 usecs, depending on + * the cache size) + */ + + if (!cpu_khz) { + /* + * this basically disables processor-affinity + * scheduling on SMP without a TSC. + */ + return; + } else { + cachesize = boot_cpu_data.x86_cache_size; + if (cachesize == -1) { + cachesize = 16; /* Pentiums, 2x8kB cache */ + bandwidth = 100; + } + } +} + +/* + * Cycle through the processors sending APIC IPIs to boot each. + */ + +static int boot_cpu_logical_apicid; +/* Where the IO area was mapped on multiquad, always 0 otherwise */ +void *xquad_portio; + +cpumask_t cpu_sibling_map[NR_CPUS] __cacheline_aligned; + +static void __init smp_boot_cpus(unsigned int max_cpus) +{ + int apicid, cpu, bit, kicked; + unsigned long bogosum = 0; + + /* + * Setup boot CPU information + */ + smp_store_cpu_info(0); /* Final full version of the data */ + printk("CPU%d: ", 0); + print_cpu_info(&cpu_data[0]); + + boot_cpu_physical_apicid = GET_APIC_ID(apic_read(APIC_ID)); + boot_cpu_logical_apicid = logical_smp_processor_id(); + x86_cpu_to_apicid[0] = boot_cpu_physical_apicid; + + current_thread_info()->cpu = 0; + smp_tune_scheduling(); + cpus_clear(cpu_sibling_map[0]); + cpu_set(0, cpu_sibling_map[0]); + + /* + * If we couldn't find an SMP configuration at boot time, + * get out of here now! + */ + if (!smp_found_config && !acpi_lapic) { + printk(KERN_NOTICE "SMP motherboard not detected.\n"); + smpboot_clear_io_apic_irqs(); + phys_cpu_present_map = physid_mask_of_physid(0); + if (APIC_init_uniprocessor()) + printk(KERN_NOTICE "Local APIC not detected." + " Using dummy APIC emulation.\n"); + map_cpu_to_logical_apicid(); + return; + } + + /* + * Should not be necessary because the MP table should list the boot + * CPU too, but we do it for the sake of robustness anyway. + * Makes no sense to do this check in clustered apic mode, so skip it + */ + if (!check_phys_apicid_present(boot_cpu_physical_apicid)) { + printk("weird, boot CPU (#%d) not listed by the BIOS.\n", + boot_cpu_physical_apicid); + physid_set(hard_smp_processor_id(), phys_cpu_present_map); + } + + /* + * If we couldn't find a local APIC, then get out of here now! + */ + if (APIC_INTEGRATED(apic_version[boot_cpu_physical_apicid]) && !cpu_has_apic) { + printk(KERN_ERR "BIOS bug, local APIC #%d not detected!...\n", + boot_cpu_physical_apicid); + printk(KERN_ERR "... forcing use of dummy APIC emulation. (tell your hw vendor)\n"); + smpboot_clear_io_apic_irqs(); + phys_cpu_present_map = physid_mask_of_physid(0); + return; + } + + verify_local_APIC(); + + /* + * If SMP should be disabled, then really disable it! + */ + if (!max_cpus) { + smp_found_config = 0; + printk(KERN_INFO "SMP mode deactivated, forcing use of dummy APIC emulation.\n"); + smpboot_clear_io_apic_irqs(); + phys_cpu_present_map = physid_mask_of_physid(0); + return; + } + + connect_bsp_APIC(); + setup_local_APIC(); + map_cpu_to_logical_apicid(); + + + setup_portio_remap(); + + /* + * Scan the CPU present map and fire up the other CPUs via do_boot_cpu + * + * In clustered apic mode, phys_cpu_present_map is a constructed thus: + * bits 0-3 are quad0, 4-7 are quad1, etc. A perverse twist on the + * clustered apic ID. + */ + Dprintk("CPU present map: %lx\n", physids_coerce(phys_cpu_present_map)); + + kicked = 1; + for (bit = 0; kicked < NR_CPUS && bit < MAX_APICS; bit++) { + apicid = cpu_present_to_apicid(bit); + /* + * Don't even attempt to start the boot CPU! + */ + if ((apicid == boot_cpu_apicid) || (apicid == BAD_APICID)) + continue; + + if (!check_apicid_present(bit)) + continue; + if (max_cpus <= cpucount+1) + continue; + + if (do_boot_cpu(apicid)) + printk("CPU #%d not responding - cannot use it.\n", + apicid); + else + ++kicked; + } + + /* + * Cleanup possible dangling ends... + */ + smpboot_restore_warm_reset_vector(); + + /* + * Allow the user to impress friends. + */ + Dprintk("Before bogomips.\n"); + for (cpu = 0; cpu < NR_CPUS; cpu++) + if (cpu_isset(cpu, cpu_callout_map)) + bogosum += cpu_data[cpu].loops_per_jiffy; + printk(KERN_INFO + "Total of %d processors activated (%lu.%02lu BogoMIPS).\n", + cpucount+1, + bogosum/(500000/HZ), + (bogosum/(5000/HZ))%100); + + Dprintk("Before bogocount - setting activated=1.\n"); + + if (smp_b_stepping) + printk(KERN_WARNING "WARNING: SMP operation may be unreliable with B stepping processors.\n"); + + /* + * Don't taint if we are running SMP kernel on a single non-MP + * approved Athlon + */ + if (tainted & TAINT_UNSAFE_SMP) { + if (cpucount) + printk (KERN_INFO "WARNING: This combination of AMD processors is not suitable for SMP.\n"); + else + tainted &= ~TAINT_UNSAFE_SMP; + } + + Dprintk("Boot done.\n"); + + /* + * construct cpu_sibling_map[], so that we can tell sibling CPUs + * efficiently. + */ + for (cpu = 0; cpu < NR_CPUS; cpu++) + cpus_clear(cpu_sibling_map[cpu]); + + for (cpu = 0; cpu < NR_CPUS; cpu++) { + int siblings = 0; + int i; + if (!cpu_isset(cpu, cpu_callout_map)) + continue; + + if (smp_num_siblings > 1) { + for (i = 0; i < NR_CPUS; i++) { + if (!cpu_isset(i, cpu_callout_map)) + continue; + if (phys_proc_id[cpu] == phys_proc_id[i]) { + siblings++; + cpu_set(i, cpu_sibling_map[cpu]); + } + } + } else { + siblings++; + cpu_set(cpu, cpu_sibling_map[cpu]); + } + + if (siblings != smp_num_siblings) + printk(KERN_WARNING "WARNING: %d siblings found for CPU%d, should be %d\n", siblings, cpu, smp_num_siblings); + } + + if (nmi_watchdog == NMI_LOCAL_APIC) + check_nmi_watchdog(); + + smpboot_setup_io_apic(); + + setup_boot_APIC_clock(); + + /* + * Synchronize the TSC with the AP + */ + if (cpu_has_tsc && cpucount && cpu_khz) + synchronize_tsc_bp(); +} + +/* These are wrappers to interface to the new boot process. Someone + who understands all this stuff should rewrite it properly. --RR 15/Jul/02 */ +void __init smp_prepare_cpus(unsigned int max_cpus) +{ + smp_boot_cpus(max_cpus); +} + +void __devinit smp_prepare_boot_cpu(void) +{ + cpu_set(smp_processor_id(), cpu_online_map); + cpu_set(smp_processor_id(), cpu_callout_map); +} + +int __devinit __cpu_up(unsigned int cpu) +{ + /* This only works at boot for x86. See "rewrite" above. */ + if (cpu_isset(cpu, smp_commenced_mask)) { + local_irq_enable(); + return -ENOSYS; + } + + /* In case one didn't come up */ + if (!cpu_isset(cpu, cpu_callin_map)) { + local_irq_enable(); + return -EIO; + } + + local_irq_enable(); + /* Unleash the CPU! */ + cpu_set(cpu, smp_commenced_mask); + while (!cpu_isset(cpu, cpu_online_map)) + mb(); + return 0; +} + +void __init smp_cpus_done(unsigned int max_cpus) +{ +#ifdef CONFIG_X86_IO_APIC + setup_ioapic_dest(); +#endif + zap_low_mappings(); + /* + * Disable executability of the SMP trampoline: + */ + set_kernel_exec((unsigned long)trampoline_base, trampoline_exec); +} + +void __init smp_intr_init(void) +{ + /* + * IRQ0 must be given a fixed assignment and initialized, + * because it's used before the IO-APIC is set up. + */ + set_intr_gate(FIRST_DEVICE_VECTOR, interrupt[0]); + + /* + * The reschedule interrupt is a CPU-to-CPU reschedule-helper + * IPI, driven by wakeup. + */ + set_intr_gate(RESCHEDULE_VECTOR, reschedule_interrupt); + + /* IPI for invalidation */ + set_intr_gate(INVALIDATE_TLB_VECTOR, invalidate_interrupt); + + /* IPI for generic function call */ + set_intr_gate(CALL_FUNCTION_VECTOR, call_function_interrupt); +} diff --git a/arch/i386/kernel/srat.c b/arch/i386/kernel/srat.c new file mode 100644 index 00000000000..7b3b27d6440 --- /dev/null +++ b/arch/i386/kernel/srat.c @@ -0,0 +1,456 @@ +/* + * Some of the code in this file has been gleaned from the 64 bit + * discontigmem support code base. + * + * Copyright (C) 2002, IBM Corp. + * + * All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * + * Send feedback to Pat Gaughen <gone@us.ibm.com> + */ +#include <linux/config.h> +#include <linux/mm.h> +#include <linux/bootmem.h> +#include <linux/mmzone.h> +#include <linux/acpi.h> +#include <linux/nodemask.h> +#include <asm/srat.h> +#include <asm/topology.h> + +/* + * proximity macros and definitions + */ +#define NODE_ARRAY_INDEX(x) ((x) / 8) /* 8 bits/char */ +#define NODE_ARRAY_OFFSET(x) ((x) % 8) /* 8 bits/char */ +#define BMAP_SET(bmap, bit) ((bmap)[NODE_ARRAY_INDEX(bit)] |= 1 << NODE_ARRAY_OFFSET(bit)) +#define BMAP_TEST(bmap, bit) ((bmap)[NODE_ARRAY_INDEX(bit)] & (1 << NODE_ARRAY_OFFSET(bit))) +#define MAX_PXM_DOMAINS 256 /* 1 byte and no promises about values */ +/* bitmap length; _PXM is at most 255 */ +#define PXM_BITMAP_LEN (MAX_PXM_DOMAINS / 8) +static u8 pxm_bitmap[PXM_BITMAP_LEN]; /* bitmap of proximity domains */ + +#define MAX_CHUNKS_PER_NODE 4 +#define MAXCHUNKS (MAX_CHUNKS_PER_NODE * MAX_NUMNODES) +struct node_memory_chunk_s { + unsigned long start_pfn; + unsigned long end_pfn; + u8 pxm; // proximity domain of node + u8 nid; // which cnode contains this chunk? + u8 bank; // which mem bank on this node +}; +static struct node_memory_chunk_s node_memory_chunk[MAXCHUNKS]; + +static int num_memory_chunks; /* total number of memory chunks */ +static int zholes_size_init; +static unsigned long zholes_size[MAX_NUMNODES * MAX_NR_ZONES]; + +extern void * boot_ioremap(unsigned long, unsigned long); + +/* Identify CPU proximity domains */ +static void __init parse_cpu_affinity_structure(char *p) +{ + struct acpi_table_processor_affinity *cpu_affinity = + (struct acpi_table_processor_affinity *) p; + + if (!cpu_affinity->flags.enabled) + return; /* empty entry */ + + /* mark this node as "seen" in node bitmap */ + BMAP_SET(pxm_bitmap, cpu_affinity->proximity_domain); + + printk("CPU 0x%02X in proximity domain 0x%02X\n", + cpu_affinity->apic_id, cpu_affinity->proximity_domain); +} + +/* + * Identify memory proximity domains and hot-remove capabilities. + * Fill node memory chunk list structure. + */ +static void __init parse_memory_affinity_structure (char *sratp) +{ + unsigned long long paddr, size; + unsigned long start_pfn, end_pfn; + u8 pxm; + struct node_memory_chunk_s *p, *q, *pend; + struct acpi_table_memory_affinity *memory_affinity = + (struct acpi_table_memory_affinity *) sratp; + + if (!memory_affinity->flags.enabled) + return; /* empty entry */ + + /* mark this node as "seen" in node bitmap */ + BMAP_SET(pxm_bitmap, memory_affinity->proximity_domain); + + /* calculate info for memory chunk structure */ + paddr = memory_affinity->base_addr_hi; + paddr = (paddr << 32) | memory_affinity->base_addr_lo; + size = memory_affinity->length_hi; + size = (size << 32) | memory_affinity->length_lo; + + start_pfn = paddr >> PAGE_SHIFT; + end_pfn = (paddr + size) >> PAGE_SHIFT; + + pxm = memory_affinity->proximity_domain; + + if (num_memory_chunks >= MAXCHUNKS) { + printk("Too many mem chunks in SRAT. Ignoring %lld MBytes at %llx\n", + size/(1024*1024), paddr); + return; + } + + /* Insertion sort based on base address */ + pend = &node_memory_chunk[num_memory_chunks]; + for (p = &node_memory_chunk[0]; p < pend; p++) { + if (start_pfn < p->start_pfn) + break; + } + if (p < pend) { + for (q = pend; q >= p; q--) + *(q + 1) = *q; + } + p->start_pfn = start_pfn; + p->end_pfn = end_pfn; + p->pxm = pxm; + + num_memory_chunks++; + + printk("Memory range 0x%lX to 0x%lX (type 0x%X) in proximity domain 0x%02X %s\n", + start_pfn, end_pfn, + memory_affinity->memory_type, + memory_affinity->proximity_domain, + (memory_affinity->flags.hot_pluggable ? + "enabled and removable" : "enabled" ) ); +} + +#if MAX_NR_ZONES != 3 +#error "MAX_NR_ZONES != 3, chunk_to_zone requires review" +#endif +/* Take a chunk of pages from page frame cstart to cend and count the number + * of pages in each zone, returned via zones[]. + */ +static __init void chunk_to_zones(unsigned long cstart, unsigned long cend, + unsigned long *zones) +{ + unsigned long max_dma; + extern unsigned long max_low_pfn; + + int z; + unsigned long rend; + + /* FIXME: MAX_DMA_ADDRESS and max_low_pfn are trying to provide + * similarly scoped information and should be handled in a consistant + * manner. + */ + max_dma = virt_to_phys((char *)MAX_DMA_ADDRESS) >> PAGE_SHIFT; + + /* Split the hole into the zones in which it falls. Repeatedly + * take the segment in which the remaining hole starts, round it + * to the end of that zone. + */ + memset(zones, 0, MAX_NR_ZONES * sizeof(long)); + while (cstart < cend) { + if (cstart < max_dma) { + z = ZONE_DMA; + rend = (cend < max_dma)? cend : max_dma; + + } else if (cstart < max_low_pfn) { + z = ZONE_NORMAL; + rend = (cend < max_low_pfn)? cend : max_low_pfn; + + } else { + z = ZONE_HIGHMEM; + rend = cend; + } + zones[z] += rend - cstart; + cstart = rend; + } +} + +/* + * The SRAT table always lists ascending addresses, so can always + * assume that the first "start" address that you see is the real + * start of the node, and that the current "end" address is after + * the previous one. + */ +static __init void node_read_chunk(int nid, struct node_memory_chunk_s *memory_chunk) +{ + /* + * Only add present memory as told by the e820. + * There is no guarantee from the SRAT that the memory it + * enumerates is present at boot time because it represents + * *possible* memory hotplug areas the same as normal RAM. + */ + if (memory_chunk->start_pfn >= max_pfn) { + printk (KERN_INFO "Ignoring SRAT pfns: 0x%08lx -> %08lx\n", + memory_chunk->start_pfn, memory_chunk->end_pfn); + return; + } + if (memory_chunk->nid != nid) + return; + + if (!node_has_online_mem(nid)) + node_start_pfn[nid] = memory_chunk->start_pfn; + + if (node_start_pfn[nid] > memory_chunk->start_pfn) + node_start_pfn[nid] = memory_chunk->start_pfn; + + if (node_end_pfn[nid] < memory_chunk->end_pfn) + node_end_pfn[nid] = memory_chunk->end_pfn; +} + +/* Parse the ACPI Static Resource Affinity Table */ +static int __init acpi20_parse_srat(struct acpi_table_srat *sratp) +{ + u8 *start, *end, *p; + int i, j, nid; + u8 pxm_to_nid_map[MAX_PXM_DOMAINS];/* _PXM to logical node ID map */ + u8 nid_to_pxm_map[MAX_NUMNODES];/* logical node ID to _PXM map */ + + start = (u8 *)(&(sratp->reserved) + 1); /* skip header */ + p = start; + end = (u8 *)sratp + sratp->header.length; + + memset(pxm_bitmap, 0, sizeof(pxm_bitmap)); /* init proximity domain bitmap */ + memset(node_memory_chunk, 0, sizeof(node_memory_chunk)); + memset(zholes_size, 0, sizeof(zholes_size)); + + /* -1 in these maps means not available */ + memset(pxm_to_nid_map, -1, sizeof(pxm_to_nid_map)); + memset(nid_to_pxm_map, -1, sizeof(nid_to_pxm_map)); + + num_memory_chunks = 0; + while (p < end) { + switch (*p) { + case ACPI_SRAT_PROCESSOR_AFFINITY: + parse_cpu_affinity_structure(p); + break; + case ACPI_SRAT_MEMORY_AFFINITY: + parse_memory_affinity_structure(p); + break; + default: + printk("ACPI 2.0 SRAT: unknown entry skipped: type=0x%02X, len=%d\n", p[0], p[1]); + break; + } + p += p[1]; + if (p[1] == 0) { + printk("acpi20_parse_srat: Entry length value is zero;" + " can't parse any further!\n"); + break; + } + } + + if (num_memory_chunks == 0) { + printk("could not finy any ACPI SRAT memory areas.\n"); + goto out_fail; + } + + /* Calculate total number of nodes in system from PXM bitmap and create + * a set of sequential node IDs starting at zero. (ACPI doesn't seem + * to specify the range of _PXM values.) + */ + /* + * MCD - we no longer HAVE to number nodes sequentially. PXM domain + * numbers could go as high as 256, and MAX_NUMNODES for i386 is typically + * 32, so we will continue numbering them in this manner until MAX_NUMNODES + * approaches MAX_PXM_DOMAINS for i386. + */ + nodes_clear(node_online_map); + for (i = 0; i < MAX_PXM_DOMAINS; i++) { + if (BMAP_TEST(pxm_bitmap, i)) { + nid = num_online_nodes(); + pxm_to_nid_map[i] = nid; + nid_to_pxm_map[nid] = i; + node_set_online(nid); + } + } + BUG_ON(num_online_nodes() == 0); + + /* set cnode id in memory chunk structure */ + for (i = 0; i < num_memory_chunks; i++) + node_memory_chunk[i].nid = pxm_to_nid_map[node_memory_chunk[i].pxm]; + + printk("pxm bitmap: "); + for (i = 0; i < sizeof(pxm_bitmap); i++) { + printk("%02X ", pxm_bitmap[i]); + } + printk("\n"); + printk("Number of logical nodes in system = %d\n", num_online_nodes()); + printk("Number of memory chunks in system = %d\n", num_memory_chunks); + + for (j = 0; j < num_memory_chunks; j++){ + struct node_memory_chunk_s * chunk = &node_memory_chunk[j]; + printk("chunk %d nid %d start_pfn %08lx end_pfn %08lx\n", + j, chunk->nid, chunk->start_pfn, chunk->end_pfn); + node_read_chunk(chunk->nid, chunk); + } + + for_each_online_node(nid) { + unsigned long start = node_start_pfn[nid]; + unsigned long end = node_end_pfn[nid]; + + memory_present(nid, start, end); + node_remap_size[nid] = node_memmap_size_bytes(nid, start, end); + } + return 1; +out_fail: + return 0; +} + +int __init get_memcfg_from_srat(void) +{ + struct acpi_table_header *header = NULL; + struct acpi_table_rsdp *rsdp = NULL; + struct acpi_table_rsdt *rsdt = NULL; + struct acpi_pointer *rsdp_address = NULL; + struct acpi_table_rsdt saved_rsdt; + int tables = 0; + int i = 0; + + acpi_find_root_pointer(ACPI_PHYSICAL_ADDRESSING, rsdp_address); + + if (rsdp_address->pointer_type == ACPI_PHYSICAL_POINTER) { + printk("%s: assigning address to rsdp\n", __FUNCTION__); + rsdp = (struct acpi_table_rsdp *) + (u32)rsdp_address->pointer.physical; + } else { + printk("%s: rsdp_address is not a physical pointer\n", __FUNCTION__); + goto out_err; + } + if (!rsdp) { + printk("%s: Didn't find ACPI root!\n", __FUNCTION__); + goto out_err; + } + + printk(KERN_INFO "%.8s v%d [%.6s]\n", rsdp->signature, rsdp->revision, + rsdp->oem_id); + + if (strncmp(rsdp->signature, RSDP_SIG,strlen(RSDP_SIG))) { + printk(KERN_WARNING "%s: RSDP table signature incorrect\n", __FUNCTION__); + goto out_err; + } + + rsdt = (struct acpi_table_rsdt *) + boot_ioremap(rsdp->rsdt_address, sizeof(struct acpi_table_rsdt)); + + if (!rsdt) { + printk(KERN_WARNING + "%s: ACPI: Invalid root system description tables (RSDT)\n", + __FUNCTION__); + goto out_err; + } + + header = & rsdt->header; + + if (strncmp(header->signature, RSDT_SIG, strlen(RSDT_SIG))) { + printk(KERN_WARNING "ACPI: RSDT signature incorrect\n"); + goto out_err; + } + + /* + * The number of tables is computed by taking the + * size of all entries (header size minus total + * size of RSDT) divided by the size of each entry + * (4-byte table pointers). + */ + tables = (header->length - sizeof(struct acpi_table_header)) / 4; + + if (!tables) + goto out_err; + + memcpy(&saved_rsdt, rsdt, sizeof(saved_rsdt)); + + if (saved_rsdt.header.length > sizeof(saved_rsdt)) { + printk(KERN_WARNING "ACPI: Too big length in RSDT: %d\n", + saved_rsdt.header.length); + goto out_err; + } + + printk("Begin SRAT table scan....\n"); + + for (i = 0; i < tables; i++) { + /* Map in header, then map in full table length. */ + header = (struct acpi_table_header *) + boot_ioremap(saved_rsdt.entry[i], sizeof(struct acpi_table_header)); + if (!header) + break; + header = (struct acpi_table_header *) + boot_ioremap(saved_rsdt.entry[i], header->length); + if (!header) + break; + + if (strncmp((char *) &header->signature, "SRAT", 4)) + continue; + + /* we've found the srat table. don't need to look at any more tables */ + return acpi20_parse_srat((struct acpi_table_srat *)header); + } +out_err: + printk("failed to get NUMA memory information from SRAT table\n"); + return 0; +} + +/* For each node run the memory list to determine whether there are + * any memory holes. For each hole determine which ZONE they fall + * into. + * + * NOTE#1: this requires knowledge of the zone boundries and so + * _cannot_ be performed before those are calculated in setup_memory. + * + * NOTE#2: we rely on the fact that the memory chunks are ordered by + * start pfn number during setup. + */ +static void __init get_zholes_init(void) +{ + int nid; + int c; + int first; + unsigned long end = 0; + + for_each_online_node(nid) { + first = 1; + for (c = 0; c < num_memory_chunks; c++){ + if (node_memory_chunk[c].nid == nid) { + if (first) { + end = node_memory_chunk[c].end_pfn; + first = 0; + + } else { + /* Record any gap between this chunk + * and the previous chunk on this node + * against the zones it spans. + */ + chunk_to_zones(end, + node_memory_chunk[c].start_pfn, + &zholes_size[nid * MAX_NR_ZONES]); + } + } + } + } +} + +unsigned long * __init get_zholes_size(int nid) +{ + if (!zholes_size_init) { + zholes_size_init++; + get_zholes_init(); + } + if (nid >= MAX_NUMNODES || !node_online(nid)) + printk("%s: nid = %d is invalid/offline. num_online_nodes = %d", + __FUNCTION__, nid, num_online_nodes()); + return &zholes_size[nid * MAX_NR_ZONES]; +} diff --git a/arch/i386/kernel/summit.c b/arch/i386/kernel/summit.c new file mode 100644 index 00000000000..d0e01a3acf3 --- /dev/null +++ b/arch/i386/kernel/summit.c @@ -0,0 +1,180 @@ +/* + * arch/i386/kernel/summit.c - IBM Summit-Specific Code + * + * Written By: Matthew Dobson, IBM Corporation + * + * Copyright (c) 2003 IBM Corp. + * + * All rights reserved. + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or (at + * your option) any later version. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for more + * details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * + * Send feedback to <colpatch@us.ibm.com> + * + */ + +#include <linux/mm.h> +#include <linux/init.h> +#include <asm/io.h> +#include <asm/mach-summit/mach_mpparse.h> + +static struct rio_table_hdr *rio_table_hdr __initdata; +static struct scal_detail *scal_devs[MAX_NUMNODES] __initdata; +static struct rio_detail *rio_devs[MAX_NUMNODES*4] __initdata; + +static int __init setup_pci_node_map_for_wpeg(int wpeg_num, int last_bus) +{ + int twister = 0, node = 0; + int i, bus, num_buses; + + for(i = 0; i < rio_table_hdr->num_rio_dev; i++){ + if (rio_devs[i]->node_id == rio_devs[wpeg_num]->owner_id){ + twister = rio_devs[i]->owner_id; + break; + } + } + if (i == rio_table_hdr->num_rio_dev){ + printk(KERN_ERR "%s: Couldn't find owner Cyclone for Winnipeg!\n", __FUNCTION__); + return last_bus; + } + + for(i = 0; i < rio_table_hdr->num_scal_dev; i++){ + if (scal_devs[i]->node_id == twister){ + node = scal_devs[i]->node_id; + break; + } + } + if (i == rio_table_hdr->num_scal_dev){ + printk(KERN_ERR "%s: Couldn't find owner Twister for Cyclone!\n", __FUNCTION__); + return last_bus; + } + + switch (rio_devs[wpeg_num]->type){ + case CompatWPEG: + /* The Compatability Winnipeg controls the 2 legacy buses, + * the 66MHz PCI bus [2 slots] and the 2 "extra" buses in case + * a PCI-PCI bridge card is used in either slot: total 5 buses. + */ + num_buses = 5; + break; + case AltWPEG: + /* The Alternate Winnipeg controls the 2 133MHz buses [1 slot + * each], their 2 "extra" buses, the 100MHz bus [2 slots] and + * the "extra" buses for each of those slots: total 7 buses. + */ + num_buses = 7; + break; + case LookOutAWPEG: + case LookOutBWPEG: + /* A Lookout Winnipeg controls 3 100MHz buses [2 slots each] + * & the "extra" buses for each of those slots: total 9 buses. + */ + num_buses = 9; + break; + default: + printk(KERN_INFO "%s: Unsupported Winnipeg type!\n", __FUNCTION__); + return last_bus; + } + + for(bus = last_bus; bus < last_bus + num_buses; bus++) + mp_bus_id_to_node[bus] = node; + return bus; +} + +static int __init build_detail_arrays(void) +{ + unsigned long ptr; + int i, scal_detail_size, rio_detail_size; + + if (rio_table_hdr->num_scal_dev > MAX_NUMNODES){ + printk(KERN_WARNING "%s: MAX_NUMNODES too low! Defined as %d, but system has %d nodes.\n", __FUNCTION__, MAX_NUMNODES, rio_table_hdr->num_scal_dev); + return 0; + } + + switch (rio_table_hdr->version){ + default: + printk(KERN_WARNING "%s: Invalid Rio Grande Table Version: %d\n", __FUNCTION__, rio_table_hdr->version); + return 0; + case 2: + scal_detail_size = 11; + rio_detail_size = 13; + break; + case 3: + scal_detail_size = 12; + rio_detail_size = 15; + break; + } + + ptr = (unsigned long)rio_table_hdr + 3; + for(i = 0; i < rio_table_hdr->num_scal_dev; i++, ptr += scal_detail_size) + scal_devs[i] = (struct scal_detail *)ptr; + + for(i = 0; i < rio_table_hdr->num_rio_dev; i++, ptr += rio_detail_size) + rio_devs[i] = (struct rio_detail *)ptr; + + return 1; +} + +void __init setup_summit(void) +{ + unsigned long ptr; + unsigned short offset; + int i, next_wpeg, next_bus = 0; + + /* The pointer to the EBDA is stored in the word @ phys 0x40E(40:0E) */ + ptr = *(unsigned short *)phys_to_virt(0x40Eul); + ptr = (unsigned long)phys_to_virt(ptr << 4); + + rio_table_hdr = NULL; + offset = 0x180; + while (offset){ + /* The block id is stored in the 2nd word */ + if (*((unsigned short *)(ptr + offset + 2)) == 0x4752){ + /* set the pointer past the offset & block id */ + rio_table_hdr = (struct rio_table_hdr *)(ptr + offset + 4); + break; + } + /* The next offset is stored in the 1st word. 0 means no more */ + offset = *((unsigned short *)(ptr + offset)); + } + if (!rio_table_hdr){ + printk(KERN_ERR "%s: Unable to locate Rio Grande Table in EBDA - bailing!\n", __FUNCTION__); + return; + } + + if (!build_detail_arrays()) + return; + + /* The first Winnipeg we're looking for has an index of 0 */ + next_wpeg = 0; + do { + for(i = 0; i < rio_table_hdr->num_rio_dev; i++){ + if (is_WPEG(rio_devs[i]) && rio_devs[i]->WP_index == next_wpeg){ + /* It's the Winnipeg we're looking for! */ + next_bus = setup_pci_node_map_for_wpeg(i, next_bus); + next_wpeg++; + break; + } + } + /* + * If we go through all Rio devices and don't find one with + * the next index, it means we've found all the Winnipegs, + * and thus all the PCI buses. + */ + if (i == rio_table_hdr->num_rio_dev) + next_wpeg = 0; + } while (next_wpeg != 0); +} diff --git a/arch/i386/kernel/sys_i386.c b/arch/i386/kernel/sys_i386.c new file mode 100644 index 00000000000..a4a61976ecb --- /dev/null +++ b/arch/i386/kernel/sys_i386.c @@ -0,0 +1,252 @@ +/* + * linux/arch/i386/kernel/sys_i386.c + * + * This file contains various random system calls that + * have a non-standard calling sequence on the Linux/i386 + * platform. + */ + +#include <linux/errno.h> +#include <linux/sched.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/sem.h> +#include <linux/msg.h> +#include <linux/shm.h> +#include <linux/stat.h> +#include <linux/syscalls.h> +#include <linux/mman.h> +#include <linux/file.h> +#include <linux/utsname.h> + +#include <asm/uaccess.h> +#include <asm/ipc.h> + +/* + * sys_pipe() is the normal C calling standard for creating + * a pipe. It's not the way Unix traditionally does this, though. + */ +asmlinkage int sys_pipe(unsigned long __user * fildes) +{ + int fd[2]; + int error; + + error = do_pipe(fd); + if (!error) { + if (copy_to_user(fildes, fd, 2*sizeof(int))) + error = -EFAULT; + } + return error; +} + +/* common code for old and new mmaps */ +static inline long do_mmap2( + unsigned long addr, unsigned long len, + unsigned long prot, unsigned long flags, + unsigned long fd, unsigned long pgoff) +{ + int error = -EBADF; + struct file * file = NULL; + + flags &= ~(MAP_EXECUTABLE | MAP_DENYWRITE); + if (!(flags & MAP_ANONYMOUS)) { + file = fget(fd); + if (!file) + goto out; + } + + down_write(¤t->mm->mmap_sem); + error = do_mmap_pgoff(file, addr, len, prot, flags, pgoff); + up_write(¤t->mm->mmap_sem); + + if (file) + fput(file); +out: + return error; +} + +asmlinkage long sys_mmap2(unsigned long addr, unsigned long len, + unsigned long prot, unsigned long flags, + unsigned long fd, unsigned long pgoff) +{ + return do_mmap2(addr, len, prot, flags, fd, pgoff); +} + +/* + * Perform the select(nd, in, out, ex, tv) and mmap() system + * calls. Linux/i386 didn't use to be able to handle more than + * 4 system call parameters, so these system calls used a memory + * block for parameter passing.. + */ + +struct mmap_arg_struct { + unsigned long addr; + unsigned long len; + unsigned long prot; + unsigned long flags; + unsigned long fd; + unsigned long offset; +}; + +asmlinkage int old_mmap(struct mmap_arg_struct __user *arg) +{ + struct mmap_arg_struct a; + int err = -EFAULT; + + if (copy_from_user(&a, arg, sizeof(a))) + goto out; + + err = -EINVAL; + if (a.offset & ~PAGE_MASK) + goto out; + + err = do_mmap2(a.addr, a.len, a.prot, a.flags, a.fd, a.offset >> PAGE_SHIFT); +out: + return err; +} + + +struct sel_arg_struct { + unsigned long n; + fd_set __user *inp, *outp, *exp; + struct timeval __user *tvp; +}; + +asmlinkage int old_select(struct sel_arg_struct __user *arg) +{ + struct sel_arg_struct a; + + if (copy_from_user(&a, arg, sizeof(a))) + return -EFAULT; + /* sys_select() does the appropriate kernel locking */ + return sys_select(a.n, a.inp, a.outp, a.exp, a.tvp); +} + +/* + * sys_ipc() is the de-multiplexer for the SysV IPC calls.. + * + * This is really horribly ugly. + */ +asmlinkage int sys_ipc (uint call, int first, int second, + int third, void __user *ptr, long fifth) +{ + int version, ret; + + version = call >> 16; /* hack for backward compatibility */ + call &= 0xffff; + + switch (call) { + case SEMOP: + return sys_semtimedop (first, (struct sembuf __user *)ptr, second, NULL); + case SEMTIMEDOP: + return sys_semtimedop(first, (struct sembuf __user *)ptr, second, + (const struct timespec __user *)fifth); + + case SEMGET: + return sys_semget (first, second, third); + case SEMCTL: { + union semun fourth; + if (!ptr) + return -EINVAL; + if (get_user(fourth.__pad, (void __user * __user *) ptr)) + return -EFAULT; + return sys_semctl (first, second, third, fourth); + } + + case MSGSND: + return sys_msgsnd (first, (struct msgbuf __user *) ptr, + second, third); + case MSGRCV: + switch (version) { + case 0: { + struct ipc_kludge tmp; + if (!ptr) + return -EINVAL; + + if (copy_from_user(&tmp, + (struct ipc_kludge __user *) ptr, + sizeof (tmp))) + return -EFAULT; + return sys_msgrcv (first, tmp.msgp, second, + tmp.msgtyp, third); + } + default: + return sys_msgrcv (first, + (struct msgbuf __user *) ptr, + second, fifth, third); + } + case MSGGET: + return sys_msgget ((key_t) first, second); + case MSGCTL: + return sys_msgctl (first, second, (struct msqid_ds __user *) ptr); + + case SHMAT: + switch (version) { + default: { + ulong raddr; + ret = do_shmat (first, (char __user *) ptr, second, &raddr); + if (ret) + return ret; + return put_user (raddr, (ulong __user *) third); + } + case 1: /* iBCS2 emulator entry point */ + if (!segment_eq(get_fs(), get_ds())) + return -EINVAL; + /* The "(ulong *) third" is valid _only_ because of the kernel segment thing */ + return do_shmat (first, (char __user *) ptr, second, (ulong *) third); + } + case SHMDT: + return sys_shmdt ((char __user *)ptr); + case SHMGET: + return sys_shmget (first, second, third); + case SHMCTL: + return sys_shmctl (first, second, + (struct shmid_ds __user *) ptr); + default: + return -ENOSYS; + } +} + +/* + * Old cruft + */ +asmlinkage int sys_uname(struct old_utsname __user * name) +{ + int err; + if (!name) + return -EFAULT; + down_read(&uts_sem); + err=copy_to_user(name, &system_utsname, sizeof (*name)); + up_read(&uts_sem); + return err?-EFAULT:0; +} + +asmlinkage int sys_olduname(struct oldold_utsname __user * name) +{ + int error; + + if (!name) + return -EFAULT; + if (!access_ok(VERIFY_WRITE,name,sizeof(struct oldold_utsname))) + return -EFAULT; + + down_read(&uts_sem); + + error = __copy_to_user(&name->sysname,&system_utsname.sysname,__OLD_UTS_LEN); + error |= __put_user(0,name->sysname+__OLD_UTS_LEN); + error |= __copy_to_user(&name->nodename,&system_utsname.nodename,__OLD_UTS_LEN); + error |= __put_user(0,name->nodename+__OLD_UTS_LEN); + error |= __copy_to_user(&name->release,&system_utsname.release,__OLD_UTS_LEN); + error |= __put_user(0,name->release+__OLD_UTS_LEN); + error |= __copy_to_user(&name->version,&system_utsname.version,__OLD_UTS_LEN); + error |= __put_user(0,name->version+__OLD_UTS_LEN); + error |= __copy_to_user(&name->machine,&system_utsname.machine,__OLD_UTS_LEN); + error |= __put_user(0,name->machine+__OLD_UTS_LEN); + + up_read(&uts_sem); + + error = error ? -EFAULT : 0; + + return error; +} diff --git a/arch/i386/kernel/sysenter.c b/arch/i386/kernel/sysenter.c new file mode 100644 index 00000000000..960d8bd137d --- /dev/null +++ b/arch/i386/kernel/sysenter.c @@ -0,0 +1,65 @@ +/* + * linux/arch/i386/kernel/sysenter.c + * + * (C) Copyright 2002 Linus Torvalds + * + * This file contains the needed initializations to support sysenter. + */ + +#include <linux/init.h> +#include <linux/smp.h> +#include <linux/thread_info.h> +#include <linux/sched.h> +#include <linux/gfp.h> +#include <linux/string.h> +#include <linux/elf.h> + +#include <asm/cpufeature.h> +#include <asm/msr.h> +#include <asm/pgtable.h> +#include <asm/unistd.h> + +extern asmlinkage void sysenter_entry(void); + +void enable_sep_cpu(void *info) +{ + int cpu = get_cpu(); + struct tss_struct *tss = &per_cpu(init_tss, cpu); + + tss->ss1 = __KERNEL_CS; + tss->esp1 = sizeof(struct tss_struct) + (unsigned long) tss; + wrmsr(MSR_IA32_SYSENTER_CS, __KERNEL_CS, 0); + wrmsr(MSR_IA32_SYSENTER_ESP, tss->esp1, 0); + wrmsr(MSR_IA32_SYSENTER_EIP, (unsigned long) sysenter_entry, 0); + put_cpu(); +} + +/* + * These symbols are defined by vsyscall.o to mark the bounds + * of the ELF DSO images included therein. + */ +extern const char vsyscall_int80_start, vsyscall_int80_end; +extern const char vsyscall_sysenter_start, vsyscall_sysenter_end; + +static int __init sysenter_setup(void) +{ + void *page = (void *)get_zeroed_page(GFP_ATOMIC); + + __set_fixmap(FIX_VSYSCALL, __pa(page), PAGE_READONLY_EXEC); + + if (!boot_cpu_has(X86_FEATURE_SEP)) { + memcpy(page, + &vsyscall_int80_start, + &vsyscall_int80_end - &vsyscall_int80_start); + return 0; + } + + memcpy(page, + &vsyscall_sysenter_start, + &vsyscall_sysenter_end - &vsyscall_sysenter_start); + + on_each_cpu(enable_sep_cpu, NULL, 1, 1); + return 0; +} + +__initcall(sysenter_setup); diff --git a/arch/i386/kernel/time.c b/arch/i386/kernel/time.c new file mode 100644 index 00000000000..9b55e30e449 --- /dev/null +++ b/arch/i386/kernel/time.c @@ -0,0 +1,476 @@ +/* + * linux/arch/i386/kernel/time.c + * + * Copyright (C) 1991, 1992, 1995 Linus Torvalds + * + * This file contains the PC-specific time handling details: + * reading the RTC at bootup, etc.. + * 1994-07-02 Alan Modra + * fixed set_rtc_mmss, fixed time.year for >= 2000, new mktime + * 1995-03-26 Markus Kuhn + * fixed 500 ms bug at call to set_rtc_mmss, fixed DS12887 + * precision CMOS clock update + * 1996-05-03 Ingo Molnar + * fixed time warps in do_[slow|fast]_gettimeoffset() + * 1997-09-10 Updated NTP code according to technical memorandum Jan '96 + * "A Kernel Model for Precision Timekeeping" by Dave Mills + * 1998-09-05 (Various) + * More robust do_fast_gettimeoffset() algorithm implemented + * (works with APM, Cyrix 6x86MX and Centaur C6), + * monotonic gettimeofday() with fast_get_timeoffset(), + * drift-proof precision TSC calibration on boot + * (C. Scott Ananian <cananian@alumni.princeton.edu>, Andrew D. + * Balsa <andrebalsa@altern.org>, Philip Gladstone <philip@raptor.com>; + * ported from 2.0.35 Jumbo-9 by Michael Krause <m.krause@tu-harburg.de>). + * 1998-12-16 Andrea Arcangeli + * Fixed Jumbo-9 code in 2.1.131: do_gettimeofday was missing 1 jiffy + * because was not accounting lost_ticks. + * 1998-12-24 Copyright (C) 1998 Andrea Arcangeli + * Fixed a xtime SMP race (we need the xtime_lock rw spinlock to + * serialize accesses to xtime/lost_ticks). + */ + +#include <linux/errno.h> +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/param.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/interrupt.h> +#include <linux/time.h> +#include <linux/delay.h> +#include <linux/init.h> +#include <linux/smp.h> +#include <linux/module.h> +#include <linux/sysdev.h> +#include <linux/bcd.h> +#include <linux/efi.h> +#include <linux/mca.h> + +#include <asm/io.h> +#include <asm/smp.h> +#include <asm/irq.h> +#include <asm/msr.h> +#include <asm/delay.h> +#include <asm/mpspec.h> +#include <asm/uaccess.h> +#include <asm/processor.h> +#include <asm/timer.h> + +#include "mach_time.h" + +#include <linux/timex.h> +#include <linux/config.h> + +#include <asm/hpet.h> + +#include <asm/arch_hooks.h> + +#include "io_ports.h" + +extern spinlock_t i8259A_lock; +int pit_latch_buggy; /* extern */ + +#include "do_timer.h" + +u64 jiffies_64 = INITIAL_JIFFIES; + +EXPORT_SYMBOL(jiffies_64); + +unsigned long cpu_khz; /* Detected as we calibrate the TSC */ + +extern unsigned long wall_jiffies; + +DEFINE_SPINLOCK(rtc_lock); + +DEFINE_SPINLOCK(i8253_lock); +EXPORT_SYMBOL(i8253_lock); + +struct timer_opts *cur_timer = &timer_none; + +/* + * This is a special lock that is owned by the CPU and holds the index + * register we are working with. It is required for NMI access to the + * CMOS/RTC registers. See include/asm-i386/mc146818rtc.h for details. + */ +volatile unsigned long cmos_lock = 0; +EXPORT_SYMBOL(cmos_lock); + +/* Routines for accessing the CMOS RAM/RTC. */ +unsigned char rtc_cmos_read(unsigned char addr) +{ + unsigned char val; + lock_cmos_prefix(addr); + outb_p(addr, RTC_PORT(0)); + val = inb_p(RTC_PORT(1)); + lock_cmos_suffix(addr); + return val; +} +EXPORT_SYMBOL(rtc_cmos_read); + +void rtc_cmos_write(unsigned char val, unsigned char addr) +{ + lock_cmos_prefix(addr); + outb_p(addr, RTC_PORT(0)); + outb_p(val, RTC_PORT(1)); + lock_cmos_suffix(addr); +} +EXPORT_SYMBOL(rtc_cmos_write); + +/* + * This version of gettimeofday has microsecond resolution + * and better than microsecond precision on fast x86 machines with TSC. + */ +void do_gettimeofday(struct timeval *tv) +{ + unsigned long seq; + unsigned long usec, sec; + unsigned long max_ntp_tick; + + do { + unsigned long lost; + + seq = read_seqbegin(&xtime_lock); + + usec = cur_timer->get_offset(); + lost = jiffies - wall_jiffies; + + /* + * If time_adjust is negative then NTP is slowing the clock + * so make sure not to go into next possible interval. + * Better to lose some accuracy than have time go backwards.. + */ + if (unlikely(time_adjust < 0)) { + max_ntp_tick = (USEC_PER_SEC / HZ) - tickadj; + usec = min(usec, max_ntp_tick); + + if (lost) + usec += lost * max_ntp_tick; + } + else if (unlikely(lost)) + usec += lost * (USEC_PER_SEC / HZ); + + sec = xtime.tv_sec; + usec += (xtime.tv_nsec / 1000); + } while (read_seqretry(&xtime_lock, seq)); + + while (usec >= 1000000) { + usec -= 1000000; + sec++; + } + + tv->tv_sec = sec; + tv->tv_usec = usec; +} + +EXPORT_SYMBOL(do_gettimeofday); + +int do_settimeofday(struct timespec *tv) +{ + time_t wtm_sec, sec = tv->tv_sec; + long wtm_nsec, nsec = tv->tv_nsec; + + if ((unsigned long)tv->tv_nsec >= NSEC_PER_SEC) + return -EINVAL; + + write_seqlock_irq(&xtime_lock); + /* + * This is revolting. We need to set "xtime" correctly. However, the + * value in this location is the value at the most recent update of + * wall time. Discover what correction gettimeofday() would have + * made, and then undo it! + */ + nsec -= cur_timer->get_offset() * NSEC_PER_USEC; + nsec -= (jiffies - wall_jiffies) * TICK_NSEC; + + wtm_sec = wall_to_monotonic.tv_sec + (xtime.tv_sec - sec); + wtm_nsec = wall_to_monotonic.tv_nsec + (xtime.tv_nsec - nsec); + + set_normalized_timespec(&xtime, sec, nsec); + set_normalized_timespec(&wall_to_monotonic, wtm_sec, wtm_nsec); + + time_adjust = 0; /* stop active adjtime() */ + time_status |= STA_UNSYNC; + time_maxerror = NTP_PHASE_LIMIT; + time_esterror = NTP_PHASE_LIMIT; + write_sequnlock_irq(&xtime_lock); + clock_was_set(); + return 0; +} + +EXPORT_SYMBOL(do_settimeofday); + +static int set_rtc_mmss(unsigned long nowtime) +{ + int retval; + + WARN_ON(irqs_disabled()); + + /* gets recalled with irq locally disabled */ + spin_lock_irq(&rtc_lock); + if (efi_enabled) + retval = efi_set_rtc_mmss(nowtime); + else + retval = mach_set_rtc_mmss(nowtime); + spin_unlock_irq(&rtc_lock); + + return retval; +} + + +int timer_ack; + +/* monotonic_clock(): returns # of nanoseconds passed since time_init() + * Note: This function is required to return accurate + * time even in the absence of multiple timer ticks. + */ +unsigned long long monotonic_clock(void) +{ + return cur_timer->monotonic_clock(); +} +EXPORT_SYMBOL(monotonic_clock); + +#if defined(CONFIG_SMP) && defined(CONFIG_FRAME_POINTER) +unsigned long profile_pc(struct pt_regs *regs) +{ + unsigned long pc = instruction_pointer(regs); + + if (in_lock_functions(pc)) + return *(unsigned long *)(regs->ebp + 4); + + return pc; +} +EXPORT_SYMBOL(profile_pc); +#endif + +/* + * timer_interrupt() needs to keep up the real-time clock, + * as well as call the "do_timer()" routine every clocktick + */ +static inline void do_timer_interrupt(int irq, void *dev_id, + struct pt_regs *regs) +{ +#ifdef CONFIG_X86_IO_APIC + if (timer_ack) { + /* + * Subtle, when I/O APICs are used we have to ack timer IRQ + * manually to reset the IRR bit for do_slow_gettimeoffset(). + * This will also deassert NMI lines for the watchdog if run + * on an 82489DX-based system. + */ + spin_lock(&i8259A_lock); + outb(0x0c, PIC_MASTER_OCW3); + /* Ack the IRQ; AEOI will end it automatically. */ + inb(PIC_MASTER_POLL); + spin_unlock(&i8259A_lock); + } +#endif + + do_timer_interrupt_hook(regs); + + + if (MCA_bus) { + /* The PS/2 uses level-triggered interrupts. You can't + turn them off, nor would you want to (any attempt to + enable edge-triggered interrupts usually gets intercepted by a + special hardware circuit). Hence we have to acknowledge + the timer interrupt. Through some incredibly stupid + design idea, the reset for IRQ 0 is done by setting the + high bit of the PPI port B (0x61). Note that some PS/2s, + notably the 55SX, work fine if this is removed. */ + + irq = inb_p( 0x61 ); /* read the current state */ + outb_p( irq|0x80, 0x61 ); /* reset the IRQ */ + } +} + +/* + * This is the same as the above, except we _also_ save the current + * Time Stamp Counter value at the time of the timer interrupt, so that + * we later on can estimate the time of day more exactly. + */ +irqreturn_t timer_interrupt(int irq, void *dev_id, struct pt_regs *regs) +{ + /* + * Here we are in the timer irq handler. We just have irqs locally + * disabled but we don't know if the timer_bh is running on the other + * CPU. We need to avoid to SMP race with it. NOTE: we don' t need + * the irq version of write_lock because as just said we have irq + * locally disabled. -arca + */ + write_seqlock(&xtime_lock); + + cur_timer->mark_offset(); + + do_timer_interrupt(irq, NULL, regs); + + write_sequnlock(&xtime_lock); + return IRQ_HANDLED; +} + +/* not static: needed by APM */ +unsigned long get_cmos_time(void) +{ + unsigned long retval; + + spin_lock(&rtc_lock); + + if (efi_enabled) + retval = efi_get_time(); + else + retval = mach_get_cmos_time(); + + spin_unlock(&rtc_lock); + + return retval; +} +static void sync_cmos_clock(unsigned long dummy); + +static struct timer_list sync_cmos_timer = + TIMER_INITIALIZER(sync_cmos_clock, 0, 0); + +static void sync_cmos_clock(unsigned long dummy) +{ + struct timeval now, next; + int fail = 1; + + /* + * If we have an externally synchronized Linux clock, then update + * CMOS clock accordingly every ~11 minutes. Set_rtc_mmss() has to be + * called as close as possible to 500 ms before the new second starts. + * This code is run on a timer. If the clock is set, that timer + * may not expire at the correct time. Thus, we adjust... + */ + if ((time_status & STA_UNSYNC) != 0) + /* + * Not synced, exit, do not restart a timer (if one is + * running, let it run out). + */ + return; + + do_gettimeofday(&now); + if (now.tv_usec >= USEC_AFTER - ((unsigned) TICK_SIZE) / 2 && + now.tv_usec <= USEC_BEFORE + ((unsigned) TICK_SIZE) / 2) + fail = set_rtc_mmss(now.tv_sec); + + next.tv_usec = USEC_AFTER - now.tv_usec; + if (next.tv_usec <= 0) + next.tv_usec += USEC_PER_SEC; + + if (!fail) + next.tv_sec = 659; + else + next.tv_sec = 0; + + if (next.tv_usec >= USEC_PER_SEC) { + next.tv_sec++; + next.tv_usec -= USEC_PER_SEC; + } + mod_timer(&sync_cmos_timer, jiffies + timeval_to_jiffies(&next)); +} + +void notify_arch_cmos_timer(void) +{ + mod_timer(&sync_cmos_timer, jiffies + 1); +} + +static long clock_cmos_diff, sleep_start; + +static int timer_suspend(struct sys_device *dev, u32 state) +{ + /* + * Estimate time zone so that set_time can update the clock + */ + clock_cmos_diff = -get_cmos_time(); + clock_cmos_diff += get_seconds(); + sleep_start = get_cmos_time(); + return 0; +} + +static int timer_resume(struct sys_device *dev) +{ + unsigned long flags; + unsigned long sec; + unsigned long sleep_length; + +#ifdef CONFIG_HPET_TIMER + if (is_hpet_enabled()) + hpet_reenable(); +#endif + sec = get_cmos_time() + clock_cmos_diff; + sleep_length = (get_cmos_time() - sleep_start) * HZ; + write_seqlock_irqsave(&xtime_lock, flags); + xtime.tv_sec = sec; + xtime.tv_nsec = 0; + write_sequnlock_irqrestore(&xtime_lock, flags); + jiffies += sleep_length; + wall_jiffies += sleep_length; + return 0; +} + +static struct sysdev_class timer_sysclass = { + .resume = timer_resume, + .suspend = timer_suspend, + set_kset_name("timer"), +}; + + +/* XXX this driverfs stuff should probably go elsewhere later -john */ +static struct sys_device device_timer = { + .id = 0, + .cls = &timer_sysclass, +}; + +static int time_init_device(void) +{ + int error = sysdev_class_register(&timer_sysclass); + if (!error) + error = sysdev_register(&device_timer); + return error; +} + +device_initcall(time_init_device); + +#ifdef CONFIG_HPET_TIMER +extern void (*late_time_init)(void); +/* Duplicate of time_init() below, with hpet_enable part added */ +static void __init hpet_time_init(void) +{ + xtime.tv_sec = get_cmos_time(); + xtime.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ); + set_normalized_timespec(&wall_to_monotonic, + -xtime.tv_sec, -xtime.tv_nsec); + + if (hpet_enable() >= 0) { + printk("Using HPET for base-timer\n"); + } + + cur_timer = select_timer(); + printk(KERN_INFO "Using %s for high-res timesource\n",cur_timer->name); + + time_init_hook(); +} +#endif + +void __init time_init(void) +{ +#ifdef CONFIG_HPET_TIMER + if (is_hpet_capable()) { + /* + * HPET initialization needs to do memory-mapped io. So, let + * us do a late initialization after mem_init(). + */ + late_time_init = hpet_time_init; + return; + } +#endif + xtime.tv_sec = get_cmos_time(); + xtime.tv_nsec = (INITIAL_JIFFIES % HZ) * (NSEC_PER_SEC / HZ); + set_normalized_timespec(&wall_to_monotonic, + -xtime.tv_sec, -xtime.tv_nsec); + + cur_timer = select_timer(); + printk(KERN_INFO "Using %s for high-res timesource\n",cur_timer->name); + + time_init_hook(); +} diff --git a/arch/i386/kernel/time_hpet.c b/arch/i386/kernel/time_hpet.c new file mode 100644 index 00000000000..244a31b04be --- /dev/null +++ b/arch/i386/kernel/time_hpet.c @@ -0,0 +1,458 @@ +/* + * linux/arch/i386/kernel/time_hpet.c + * This code largely copied from arch/x86_64/kernel/time.c + * See that file for credits. + * + * 2003-06-30 Venkatesh Pallipadi - Additional changes for HPET support + */ + +#include <linux/errno.h> +#include <linux/kernel.h> +#include <linux/param.h> +#include <linux/string.h> +#include <linux/init.h> +#include <linux/smp.h> + +#include <asm/timer.h> +#include <asm/fixmap.h> +#include <asm/apic.h> + +#include <linux/timex.h> +#include <linux/config.h> + +#include <asm/hpet.h> +#include <linux/hpet.h> + +static unsigned long hpet_period; /* fsecs / HPET clock */ +unsigned long hpet_tick; /* hpet clks count per tick */ +unsigned long hpet_address; /* hpet memory map physical address */ + +static int use_hpet; /* can be used for runtime check of hpet */ +static int boot_hpet_disable; /* boottime override for HPET timer */ +static void __iomem * hpet_virt_address; /* hpet kernel virtual address */ + +#define FSEC_TO_USEC (1000000000UL) + +int hpet_readl(unsigned long a) +{ + return readl(hpet_virt_address + a); +} + +static void hpet_writel(unsigned long d, unsigned long a) +{ + writel(d, hpet_virt_address + a); +} + +#ifdef CONFIG_X86_LOCAL_APIC +/* + * HPET counters dont wrap around on every tick. They just change the + * comparator value and continue. Next tick can be caught by checking + * for a change in the comparator value. Used in apic.c. + */ +static void __init wait_hpet_tick(void) +{ + unsigned int start_cmp_val, end_cmp_val; + + start_cmp_val = hpet_readl(HPET_T0_CMP); + do { + end_cmp_val = hpet_readl(HPET_T0_CMP); + } while (start_cmp_val == end_cmp_val); +} +#endif + +static int hpet_timer_stop_set_go(unsigned long tick) +{ + unsigned int cfg; + + /* + * Stop the timers and reset the main counter. + */ + cfg = hpet_readl(HPET_CFG); + cfg &= ~HPET_CFG_ENABLE; + hpet_writel(cfg, HPET_CFG); + hpet_writel(0, HPET_COUNTER); + hpet_writel(0, HPET_COUNTER + 4); + + /* + * Set up timer 0, as periodic with first interrupt to happen at + * hpet_tick, and period also hpet_tick. + */ + cfg = hpet_readl(HPET_T0_CFG); + cfg |= HPET_TN_ENABLE | HPET_TN_PERIODIC | + HPET_TN_SETVAL | HPET_TN_32BIT; + hpet_writel(cfg, HPET_T0_CFG); + + /* + * The first write after writing TN_SETVAL to the config register sets + * the counter value, the second write sets the threshold. + */ + hpet_writel(tick, HPET_T0_CMP); + hpet_writel(tick, HPET_T0_CMP); + + /* + * Go! + */ + cfg = hpet_readl(HPET_CFG); + cfg |= HPET_CFG_ENABLE | HPET_CFG_LEGACY; + hpet_writel(cfg, HPET_CFG); + + return 0; +} + +/* + * Check whether HPET was found by ACPI boot parse. If yes setup HPET + * counter 0 for kernel base timer. + */ +int __init hpet_enable(void) +{ + unsigned int id; + unsigned long tick_fsec_low, tick_fsec_high; /* tick in femto sec */ + unsigned long hpet_tick_rem; + + if (boot_hpet_disable) + return -1; + + if (!hpet_address) { + return -1; + } + hpet_virt_address = ioremap_nocache(hpet_address, HPET_MMAP_SIZE); + /* + * Read the period, compute tick and quotient. + */ + id = hpet_readl(HPET_ID); + + /* + * We are checking for value '1' or more in number field if + * CONFIG_HPET_EMULATE_RTC is set because we will need an + * additional timer for RTC emulation. + * However, we can do with one timer otherwise using the + * the single HPET timer for system time. + */ + if ( +#ifdef CONFIG_HPET_EMULATE_RTC + !(id & HPET_ID_NUMBER) || +#endif + !(id & HPET_ID_LEGSUP)) + return -1; + + hpet_period = hpet_readl(HPET_PERIOD); + if ((hpet_period < HPET_MIN_PERIOD) || (hpet_period > HPET_MAX_PERIOD)) + return -1; + + /* + * 64 bit math + * First changing tick into fsec + * Then 64 bit div to find number of hpet clk per tick + */ + ASM_MUL64_REG(tick_fsec_low, tick_fsec_high, + KERNEL_TICK_USEC, FSEC_TO_USEC); + ASM_DIV64_REG(hpet_tick, hpet_tick_rem, + hpet_period, tick_fsec_low, tick_fsec_high); + + if (hpet_tick_rem > (hpet_period >> 1)) + hpet_tick++; /* rounding the result */ + + if (hpet_timer_stop_set_go(hpet_tick)) + return -1; + + use_hpet = 1; + +#ifdef CONFIG_HPET + { + struct hpet_data hd; + unsigned int ntimer; + + memset(&hd, 0, sizeof (hd)); + + ntimer = hpet_readl(HPET_ID); + ntimer = (ntimer & HPET_ID_NUMBER) >> HPET_ID_NUMBER_SHIFT; + ntimer++; + + /* + * Register with driver. + * Timer0 and Timer1 is used by platform. + */ + hd.hd_phys_address = hpet_address; + hd.hd_address = hpet_virt_address; + hd.hd_nirqs = ntimer; + hd.hd_flags = HPET_DATA_PLATFORM; + hpet_reserve_timer(&hd, 0); +#ifdef CONFIG_HPET_EMULATE_RTC + hpet_reserve_timer(&hd, 1); +#endif + hd.hd_irq[0] = HPET_LEGACY_8254; + hd.hd_irq[1] = HPET_LEGACY_RTC; + if (ntimer > 2) { + struct hpet __iomem *hpet; + struct hpet_timer __iomem *timer; + int i; + + hpet = hpet_virt_address; + + for (i = 2, timer = &hpet->hpet_timers[2]; i < ntimer; + timer++, i++) + hd.hd_irq[i] = (timer->hpet_config & + Tn_INT_ROUTE_CNF_MASK) >> + Tn_INT_ROUTE_CNF_SHIFT; + + } + + hpet_alloc(&hd); + } +#endif + +#ifdef CONFIG_X86_LOCAL_APIC + wait_timer_tick = wait_hpet_tick; +#endif + return 0; +} + +int hpet_reenable(void) +{ + return hpet_timer_stop_set_go(hpet_tick); +} + +int is_hpet_enabled(void) +{ + return use_hpet; +} + +int is_hpet_capable(void) +{ + if (!boot_hpet_disable && hpet_address) + return 1; + return 0; +} + +static int __init hpet_setup(char* str) +{ + if (str) { + if (!strncmp("disable", str, 7)) + boot_hpet_disable = 1; + } + return 1; +} + +__setup("hpet=", hpet_setup); + +#ifdef CONFIG_HPET_EMULATE_RTC +/* HPET in LegacyReplacement Mode eats up RTC interrupt line. When, HPET + * is enabled, we support RTC interrupt functionality in software. + * RTC has 3 kinds of interrupts: + * 1) Update Interrupt - generate an interrupt, every sec, when RTC clock + * is updated + * 2) Alarm Interrupt - generate an interrupt at a specific time of day + * 3) Periodic Interrupt - generate periodic interrupt, with frequencies + * 2Hz-8192Hz (2Hz-64Hz for non-root user) (all freqs in powers of 2) + * (1) and (2) above are implemented using polling at a frequency of + * 64 Hz. The exact frequency is a tradeoff between accuracy and interrupt + * overhead. (DEFAULT_RTC_INT_FREQ) + * For (3), we use interrupts at 64Hz or user specified periodic + * frequency, whichever is higher. + */ +#include <linux/mc146818rtc.h> +#include <linux/rtc.h> + +extern irqreturn_t rtc_interrupt(int irq, void *dev_id, struct pt_regs *regs); + +#define DEFAULT_RTC_INT_FREQ 64 +#define RTC_NUM_INTS 1 + +static unsigned long UIE_on; +static unsigned long prev_update_sec; + +static unsigned long AIE_on; +static struct rtc_time alarm_time; + +static unsigned long PIE_on; +static unsigned long PIE_freq = DEFAULT_RTC_INT_FREQ; +static unsigned long PIE_count; + +static unsigned long hpet_rtc_int_freq; /* RTC interrupt frequency */ + +/* + * Timer 1 for RTC, we do not use periodic interrupt feature, + * even if HPET supports periodic interrupts on Timer 1. + * The reason being, to set up a periodic interrupt in HPET, we need to + * stop the main counter. And if we do that everytime someone diables/enables + * RTC, we will have adverse effect on main kernel timer running on Timer 0. + * So, for the time being, simulate the periodic interrupt in software. + * + * hpet_rtc_timer_init() is called for the first time and during subsequent + * interuppts reinit happens through hpet_rtc_timer_reinit(). + */ +int hpet_rtc_timer_init(void) +{ + unsigned int cfg, cnt; + unsigned long flags; + + if (!is_hpet_enabled()) + return 0; + /* + * Set the counter 1 and enable the interrupts. + */ + if (PIE_on && (PIE_freq > DEFAULT_RTC_INT_FREQ)) + hpet_rtc_int_freq = PIE_freq; + else + hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ; + + local_irq_save(flags); + cnt = hpet_readl(HPET_COUNTER); + cnt += ((hpet_tick*HZ)/hpet_rtc_int_freq); + hpet_writel(cnt, HPET_T1_CMP); + local_irq_restore(flags); + + cfg = hpet_readl(HPET_T1_CFG); + cfg |= HPET_TN_ENABLE | HPET_TN_SETVAL | HPET_TN_32BIT; + hpet_writel(cfg, HPET_T1_CFG); + + return 1; +} + +static void hpet_rtc_timer_reinit(void) +{ + unsigned int cfg, cnt; + + if (!(PIE_on | AIE_on | UIE_on)) + return; + + if (PIE_on && (PIE_freq > DEFAULT_RTC_INT_FREQ)) + hpet_rtc_int_freq = PIE_freq; + else + hpet_rtc_int_freq = DEFAULT_RTC_INT_FREQ; + + /* It is more accurate to use the comparator value than current count.*/ + cnt = hpet_readl(HPET_T1_CMP); + cnt += hpet_tick*HZ/hpet_rtc_int_freq; + hpet_writel(cnt, HPET_T1_CMP); + + cfg = hpet_readl(HPET_T1_CFG); + cfg |= HPET_TN_ENABLE | HPET_TN_SETVAL | HPET_TN_32BIT; + hpet_writel(cfg, HPET_T1_CFG); + + return; +} + +/* + * The functions below are called from rtc driver. + * Return 0 if HPET is not being used. + * Otherwise do the necessary changes and return 1. + */ +int hpet_mask_rtc_irq_bit(unsigned long bit_mask) +{ + if (!is_hpet_enabled()) + return 0; + + if (bit_mask & RTC_UIE) + UIE_on = 0; + if (bit_mask & RTC_PIE) + PIE_on = 0; + if (bit_mask & RTC_AIE) + AIE_on = 0; + + return 1; +} + +int hpet_set_rtc_irq_bit(unsigned long bit_mask) +{ + int timer_init_reqd = 0; + + if (!is_hpet_enabled()) + return 0; + + if (!(PIE_on | AIE_on | UIE_on)) + timer_init_reqd = 1; + + if (bit_mask & RTC_UIE) { + UIE_on = 1; + } + if (bit_mask & RTC_PIE) { + PIE_on = 1; + PIE_count = 0; + } + if (bit_mask & RTC_AIE) { + AIE_on = 1; + } + + if (timer_init_reqd) + hpet_rtc_timer_init(); + + return 1; +} + +int hpet_set_alarm_time(unsigned char hrs, unsigned char min, unsigned char sec) +{ + if (!is_hpet_enabled()) + return 0; + + alarm_time.tm_hour = hrs; + alarm_time.tm_min = min; + alarm_time.tm_sec = sec; + + return 1; +} + +int hpet_set_periodic_freq(unsigned long freq) +{ + if (!is_hpet_enabled()) + return 0; + + PIE_freq = freq; + PIE_count = 0; + + return 1; +} + +int hpet_rtc_dropped_irq(void) +{ + if (!is_hpet_enabled()) + return 0; + + return 1; +} + +irqreturn_t hpet_rtc_interrupt(int irq, void *dev_id, struct pt_regs *regs) +{ + struct rtc_time curr_time; + unsigned long rtc_int_flag = 0; + int call_rtc_interrupt = 0; + + hpet_rtc_timer_reinit(); + + if (UIE_on | AIE_on) { + rtc_get_rtc_time(&curr_time); + } + if (UIE_on) { + if (curr_time.tm_sec != prev_update_sec) { + /* Set update int info, call real rtc int routine */ + call_rtc_interrupt = 1; + rtc_int_flag = RTC_UF; + prev_update_sec = curr_time.tm_sec; + } + } + if (PIE_on) { + PIE_count++; + if (PIE_count >= hpet_rtc_int_freq/PIE_freq) { + /* Set periodic int info, call real rtc int routine */ + call_rtc_interrupt = 1; + rtc_int_flag |= RTC_PF; + PIE_count = 0; + } + } + if (AIE_on) { + if ((curr_time.tm_sec == alarm_time.tm_sec) && + (curr_time.tm_min == alarm_time.tm_min) && + (curr_time.tm_hour == alarm_time.tm_hour)) { + /* Set alarm int info, call real rtc int routine */ + call_rtc_interrupt = 1; + rtc_int_flag |= RTC_AF; + } + } + if (call_rtc_interrupt) { + rtc_int_flag |= (RTC_IRQF | (RTC_NUM_INTS << 8)); + rtc_interrupt(rtc_int_flag, dev_id, regs); + } + return IRQ_HANDLED; +} +#endif + diff --git a/arch/i386/kernel/timers/Makefile b/arch/i386/kernel/timers/Makefile new file mode 100644 index 00000000000..8fa12be658d --- /dev/null +++ b/arch/i386/kernel/timers/Makefile @@ -0,0 +1,9 @@ +# +# Makefile for x86 timers +# + +obj-y := timer.o timer_none.o timer_tsc.o timer_pit.o common.o + +obj-$(CONFIG_X86_CYCLONE_TIMER) += timer_cyclone.o +obj-$(CONFIG_HPET_TIMER) += timer_hpet.o +obj-$(CONFIG_X86_PM_TIMER) += timer_pm.o diff --git a/arch/i386/kernel/timers/common.c b/arch/i386/kernel/timers/common.c new file mode 100644 index 00000000000..f7f90005e22 --- /dev/null +++ b/arch/i386/kernel/timers/common.c @@ -0,0 +1,160 @@ +/* + * Common functions used across the timers go here + */ + +#include <linux/init.h> +#include <linux/timex.h> +#include <linux/errno.h> +#include <linux/jiffies.h> + +#include <asm/io.h> +#include <asm/timer.h> +#include <asm/hpet.h> + +#include "mach_timer.h" + +/* ------ Calibrate the TSC ------- + * Return 2^32 * (1 / (TSC clocks per usec)) for do_fast_gettimeoffset(). + * Too much 64-bit arithmetic here to do this cleanly in C, and for + * accuracy's sake we want to keep the overhead on the CTC speaker (channel 2) + * output busy loop as low as possible. We avoid reading the CTC registers + * directly because of the awkward 8-bit access mechanism of the 82C54 + * device. + */ + +#define CALIBRATE_TIME (5 * 1000020/HZ) + +unsigned long __init calibrate_tsc(void) +{ + mach_prepare_counter(); + + { + unsigned long startlow, starthigh; + unsigned long endlow, endhigh; + unsigned long count; + + rdtsc(startlow,starthigh); + mach_countup(&count); + rdtsc(endlow,endhigh); + + + /* Error: ECTCNEVERSET */ + if (count <= 1) + goto bad_ctc; + + /* 64-bit subtract - gcc just messes up with long longs */ + __asm__("subl %2,%0\n\t" + "sbbl %3,%1" + :"=a" (endlow), "=d" (endhigh) + :"g" (startlow), "g" (starthigh), + "0" (endlow), "1" (endhigh)); + + /* Error: ECPUTOOFAST */ + if (endhigh) + goto bad_ctc; + + /* Error: ECPUTOOSLOW */ + if (endlow <= CALIBRATE_TIME) + goto bad_ctc; + + __asm__("divl %2" + :"=a" (endlow), "=d" (endhigh) + :"r" (endlow), "0" (0), "1" (CALIBRATE_TIME)); + + return endlow; + } + + /* + * The CTC wasn't reliable: we got a hit on the very first read, + * or the CPU was so fast/slow that the quotient wouldn't fit in + * 32 bits.. + */ +bad_ctc: + return 0; +} + +#ifdef CONFIG_HPET_TIMER +/* ------ Calibrate the TSC using HPET ------- + * Return 2^32 * (1 / (TSC clocks per usec)) for getting the CPU freq. + * Second output is parameter 1 (when non NULL) + * Set 2^32 * (1 / (tsc per HPET clk)) for delay_hpet(). + * calibrate_tsc() calibrates the processor TSC by comparing + * it to the HPET timer of known frequency. + * Too much 64-bit arithmetic here to do this cleanly in C + */ +#define CALIBRATE_CNT_HPET (5 * hpet_tick) +#define CALIBRATE_TIME_HPET (5 * KERNEL_TICK_USEC) + +unsigned long __init calibrate_tsc_hpet(unsigned long *tsc_hpet_quotient_ptr) +{ + unsigned long tsc_startlow, tsc_starthigh; + unsigned long tsc_endlow, tsc_endhigh; + unsigned long hpet_start, hpet_end; + unsigned long result, remain; + + hpet_start = hpet_readl(HPET_COUNTER); + rdtsc(tsc_startlow, tsc_starthigh); + do { + hpet_end = hpet_readl(HPET_COUNTER); + } while ((hpet_end - hpet_start) < CALIBRATE_CNT_HPET); + rdtsc(tsc_endlow, tsc_endhigh); + + /* 64-bit subtract - gcc just messes up with long longs */ + __asm__("subl %2,%0\n\t" + "sbbl %3,%1" + :"=a" (tsc_endlow), "=d" (tsc_endhigh) + :"g" (tsc_startlow), "g" (tsc_starthigh), + "0" (tsc_endlow), "1" (tsc_endhigh)); + + /* Error: ECPUTOOFAST */ + if (tsc_endhigh) + goto bad_calibration; + + /* Error: ECPUTOOSLOW */ + if (tsc_endlow <= CALIBRATE_TIME_HPET) + goto bad_calibration; + + ASM_DIV64_REG(result, remain, tsc_endlow, 0, CALIBRATE_TIME_HPET); + if (remain > (tsc_endlow >> 1)) + result++; /* rounding the result */ + + if (tsc_hpet_quotient_ptr) { + unsigned long tsc_hpet_quotient; + + ASM_DIV64_REG(tsc_hpet_quotient, remain, tsc_endlow, 0, + CALIBRATE_CNT_HPET); + if (remain > (tsc_endlow >> 1)) + tsc_hpet_quotient++; /* rounding the result */ + *tsc_hpet_quotient_ptr = tsc_hpet_quotient; + } + + return result; +bad_calibration: + /* + * the CPU was so fast/slow that the quotient wouldn't fit in + * 32 bits.. + */ + return 0; +} +#endif + +/* calculate cpu_khz */ +void __init init_cpu_khz(void) +{ + if (cpu_has_tsc) { + unsigned long tsc_quotient = calibrate_tsc(); + if (tsc_quotient) { + /* report CPU clock rate in Hz. + * The formula is (10^6 * 2^32) / (2^32 * 1 / (clocks/us)) = + * clock/second. Our precision is about 100 ppm. + */ + { unsigned long eax=0, edx=1000; + __asm__("divl %2" + :"=a" (cpu_khz), "=d" (edx) + :"r" (tsc_quotient), + "0" (eax), "1" (edx)); + printk("Detected %lu.%03lu MHz processor.\n", cpu_khz / 1000, cpu_khz % 1000); + } + } + } +} diff --git a/arch/i386/kernel/timers/timer.c b/arch/i386/kernel/timers/timer.c new file mode 100644 index 00000000000..a3d6a288088 --- /dev/null +++ b/arch/i386/kernel/timers/timer.c @@ -0,0 +1,66 @@ +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <asm/timer.h> + +#ifdef CONFIG_HPET_TIMER +/* + * HPET memory read is slower than tsc reads, but is more dependable as it + * always runs at constant frequency and reduces complexity due to + * cpufreq. So, we prefer HPET timer to tsc based one. Also, we cannot use + * timer_pit when HPET is active. So, we default to timer_tsc. + */ +#endif +/* list of timers, ordered by preference, NULL terminated */ +static struct init_timer_opts* __initdata timers[] = { +#ifdef CONFIG_X86_CYCLONE_TIMER + &timer_cyclone_init, +#endif +#ifdef CONFIG_HPET_TIMER + &timer_hpet_init, +#endif +#ifdef CONFIG_X86_PM_TIMER + &timer_pmtmr_init, +#endif + &timer_tsc_init, + &timer_pit_init, + NULL, +}; + +static char clock_override[10] __initdata; + +static int __init clock_setup(char* str) +{ + if (str) + strlcpy(clock_override, str, sizeof(clock_override)); + return 1; +} +__setup("clock=", clock_setup); + + +/* The chosen timesource has been found to be bad. + * Fall back to a known good timesource (the PIT) + */ +void clock_fallback(void) +{ + cur_timer = &timer_pit; +} + +/* iterates through the list of timers, returning the first + * one that initializes successfully. + */ +struct timer_opts* __init select_timer(void) +{ + int i = 0; + + /* find most preferred working timer */ + while (timers[i]) { + if (timers[i]->init) + if (timers[i]->init(clock_override) == 0) + return timers[i]->opts; + ++i; + } + + panic("select_timer: Cannot find a suitable timer\n"); + return NULL; +} diff --git a/arch/i386/kernel/timers/timer_cyclone.c b/arch/i386/kernel/timers/timer_cyclone.c new file mode 100644 index 00000000000..f6f1206a11b --- /dev/null +++ b/arch/i386/kernel/timers/timer_cyclone.c @@ -0,0 +1,259 @@ +/* Cyclone-timer: + * This code implements timer_ops for the cyclone counter found + * on IBM x440, x360, and other Summit based systems. + * + * Copyright (C) 2002 IBM, John Stultz (johnstul@us.ibm.com) + */ + + +#include <linux/spinlock.h> +#include <linux/init.h> +#include <linux/timex.h> +#include <linux/errno.h> +#include <linux/string.h> +#include <linux/jiffies.h> + +#include <asm/timer.h> +#include <asm/io.h> +#include <asm/pgtable.h> +#include <asm/fixmap.h> +#include "io_ports.h" + +extern spinlock_t i8253_lock; + +/* Number of usecs that the last interrupt was delayed */ +static int delay_at_last_interrupt; + +#define CYCLONE_CBAR_ADDR 0xFEB00CD0 +#define CYCLONE_PMCC_OFFSET 0x51A0 +#define CYCLONE_MPMC_OFFSET 0x51D0 +#define CYCLONE_MPCS_OFFSET 0x51A8 +#define CYCLONE_TIMER_FREQ 100000000 +#define CYCLONE_TIMER_MASK (((u64)1<<40)-1) /* 40 bit mask */ +int use_cyclone = 0; + +static u32* volatile cyclone_timer; /* Cyclone MPMC0 register */ +static u32 last_cyclone_low; +static u32 last_cyclone_high; +static unsigned long long monotonic_base; +static seqlock_t monotonic_lock = SEQLOCK_UNLOCKED; + +/* helper macro to atomically read both cyclone counter registers */ +#define read_cyclone_counter(low,high) \ + do{ \ + high = cyclone_timer[1]; low = cyclone_timer[0]; \ + } while (high != cyclone_timer[1]); + + +static void mark_offset_cyclone(void) +{ + unsigned long lost, delay; + unsigned long delta = last_cyclone_low; + int count; + unsigned long long this_offset, last_offset; + + write_seqlock(&monotonic_lock); + last_offset = ((unsigned long long)last_cyclone_high<<32)|last_cyclone_low; + + spin_lock(&i8253_lock); + read_cyclone_counter(last_cyclone_low,last_cyclone_high); + + /* read values for delay_at_last_interrupt */ + outb_p(0x00, 0x43); /* latch the count ASAP */ + + count = inb_p(0x40); /* read the latched count */ + count |= inb(0x40) << 8; + + /* + * VIA686a test code... reset the latch if count > max + 1 + * from timer_pit.c - cjb + */ + if (count > LATCH) { + outb_p(0x34, PIT_MODE); + outb_p(LATCH & 0xff, PIT_CH0); + outb(LATCH >> 8, PIT_CH0); + count = LATCH - 1; + } + spin_unlock(&i8253_lock); + + /* lost tick compensation */ + delta = last_cyclone_low - delta; + delta /= (CYCLONE_TIMER_FREQ/1000000); + delta += delay_at_last_interrupt; + lost = delta/(1000000/HZ); + delay = delta%(1000000/HZ); + if (lost >= 2) + jiffies_64 += lost-1; + + /* update the monotonic base value */ + this_offset = ((unsigned long long)last_cyclone_high<<32)|last_cyclone_low; + monotonic_base += (this_offset - last_offset) & CYCLONE_TIMER_MASK; + write_sequnlock(&monotonic_lock); + + /* calculate delay_at_last_interrupt */ + count = ((LATCH-1) - count) * TICK_SIZE; + delay_at_last_interrupt = (count + LATCH/2) / LATCH; + + + /* catch corner case where tick rollover occured + * between cyclone and pit reads (as noted when + * usec delta is > 90% # of usecs/tick) + */ + if (lost && abs(delay - delay_at_last_interrupt) > (900000/HZ)) + jiffies_64++; +} + +static unsigned long get_offset_cyclone(void) +{ + u32 offset; + + if(!cyclone_timer) + return delay_at_last_interrupt; + + /* Read the cyclone timer */ + offset = cyclone_timer[0]; + + /* .. relative to previous jiffy */ + offset = offset - last_cyclone_low; + + /* convert cyclone ticks to microseconds */ + /* XXX slow, can we speed this up? */ + offset = offset/(CYCLONE_TIMER_FREQ/1000000); + + /* our adjusted time offset in microseconds */ + return delay_at_last_interrupt + offset; +} + +static unsigned long long monotonic_clock_cyclone(void) +{ + u32 now_low, now_high; + unsigned long long last_offset, this_offset, base; + unsigned long long ret; + unsigned seq; + + /* atomically read monotonic base & last_offset */ + do { + seq = read_seqbegin(&monotonic_lock); + last_offset = ((unsigned long long)last_cyclone_high<<32)|last_cyclone_low; + base = monotonic_base; + } while (read_seqretry(&monotonic_lock, seq)); + + + /* Read the cyclone counter */ + read_cyclone_counter(now_low,now_high); + this_offset = ((unsigned long long)now_high<<32)|now_low; + + /* convert to nanoseconds */ + ret = base + ((this_offset - last_offset)&CYCLONE_TIMER_MASK); + return ret * (1000000000 / CYCLONE_TIMER_FREQ); +} + +static int __init init_cyclone(char* override) +{ + u32* reg; + u32 base; /* saved cyclone base address */ + u32 pageaddr; /* page that contains cyclone_timer register */ + u32 offset; /* offset from pageaddr to cyclone_timer register */ + int i; + + /* check clock override */ + if (override[0] && strncmp(override,"cyclone",7)) + return -ENODEV; + + /*make sure we're on a summit box*/ + if(!use_cyclone) return -ENODEV; + + printk(KERN_INFO "Summit chipset: Starting Cyclone Counter.\n"); + + /* find base address */ + pageaddr = (CYCLONE_CBAR_ADDR)&PAGE_MASK; + offset = (CYCLONE_CBAR_ADDR)&(~PAGE_MASK); + set_fixmap_nocache(FIX_CYCLONE_TIMER, pageaddr); + reg = (u32*)(fix_to_virt(FIX_CYCLONE_TIMER) + offset); + if(!reg){ + printk(KERN_ERR "Summit chipset: Could not find valid CBAR register.\n"); + return -ENODEV; + } + base = *reg; + if(!base){ + printk(KERN_ERR "Summit chipset: Could not find valid CBAR value.\n"); + return -ENODEV; + } + + /* setup PMCC */ + pageaddr = (base + CYCLONE_PMCC_OFFSET)&PAGE_MASK; + offset = (base + CYCLONE_PMCC_OFFSET)&(~PAGE_MASK); + set_fixmap_nocache(FIX_CYCLONE_TIMER, pageaddr); + reg = (u32*)(fix_to_virt(FIX_CYCLONE_TIMER) + offset); + if(!reg){ + printk(KERN_ERR "Summit chipset: Could not find valid PMCC register.\n"); + return -ENODEV; + } + reg[0] = 0x00000001; + + /* setup MPCS */ + pageaddr = (base + CYCLONE_MPCS_OFFSET)&PAGE_MASK; + offset = (base + CYCLONE_MPCS_OFFSET)&(~PAGE_MASK); + set_fixmap_nocache(FIX_CYCLONE_TIMER, pageaddr); + reg = (u32*)(fix_to_virt(FIX_CYCLONE_TIMER) + offset); + if(!reg){ + printk(KERN_ERR "Summit chipset: Could not find valid MPCS register.\n"); + return -ENODEV; + } + reg[0] = 0x00000001; + + /* map in cyclone_timer */ + pageaddr = (base + CYCLONE_MPMC_OFFSET)&PAGE_MASK; + offset = (base + CYCLONE_MPMC_OFFSET)&(~PAGE_MASK); + set_fixmap_nocache(FIX_CYCLONE_TIMER, pageaddr); + cyclone_timer = (u32*)(fix_to_virt(FIX_CYCLONE_TIMER) + offset); + if(!cyclone_timer){ + printk(KERN_ERR "Summit chipset: Could not find valid MPMC register.\n"); + return -ENODEV; + } + + /*quick test to make sure its ticking*/ + for(i=0; i<3; i++){ + u32 old = cyclone_timer[0]; + int stall = 100; + while(stall--) barrier(); + if(cyclone_timer[0] == old){ + printk(KERN_ERR "Summit chipset: Counter not counting! DISABLED\n"); + cyclone_timer = 0; + return -ENODEV; + } + } + + init_cpu_khz(); + + /* Everything looks good! */ + return 0; +} + + +static void delay_cyclone(unsigned long loops) +{ + unsigned long bclock, now; + if(!cyclone_timer) + return; + bclock = cyclone_timer[0]; + do { + rep_nop(); + now = cyclone_timer[0]; + } while ((now-bclock) < loops); +} +/************************************************************/ + +/* cyclone timer_opts struct */ +static struct timer_opts timer_cyclone = { + .name = "cyclone", + .mark_offset = mark_offset_cyclone, + .get_offset = get_offset_cyclone, + .monotonic_clock = monotonic_clock_cyclone, + .delay = delay_cyclone, +}; + +struct init_timer_opts __initdata timer_cyclone_init = { + .init = init_cyclone, + .opts = &timer_cyclone, +}; diff --git a/arch/i386/kernel/timers/timer_hpet.c b/arch/i386/kernel/timers/timer_hpet.c new file mode 100644 index 00000000000..713134e7184 --- /dev/null +++ b/arch/i386/kernel/timers/timer_hpet.c @@ -0,0 +1,191 @@ +/* + * This code largely moved from arch/i386/kernel/time.c. + * See comments there for proper credits. + */ + +#include <linux/spinlock.h> +#include <linux/init.h> +#include <linux/timex.h> +#include <linux/errno.h> +#include <linux/string.h> +#include <linux/jiffies.h> + +#include <asm/timer.h> +#include <asm/io.h> +#include <asm/processor.h> + +#include "io_ports.h" +#include "mach_timer.h" +#include <asm/hpet.h> + +static unsigned long hpet_usec_quotient; /* convert hpet clks to usec */ +static unsigned long tsc_hpet_quotient; /* convert tsc to hpet clks */ +static unsigned long hpet_last; /* hpet counter value at last tick*/ +static unsigned long last_tsc_low; /* lsb 32 bits of Time Stamp Counter */ +static unsigned long last_tsc_high; /* msb 32 bits of Time Stamp Counter */ +static unsigned long long monotonic_base; +static seqlock_t monotonic_lock = SEQLOCK_UNLOCKED; + +/* convert from cycles(64bits) => nanoseconds (64bits) + * basic equation: + * ns = cycles / (freq / ns_per_sec) + * ns = cycles * (ns_per_sec / freq) + * ns = cycles * (10^9 / (cpu_mhz * 10^6)) + * ns = cycles * (10^3 / cpu_mhz) + * + * Then we use scaling math (suggested by george@mvista.com) to get: + * ns = cycles * (10^3 * SC / cpu_mhz) / SC + * ns = cycles * cyc2ns_scale / SC + * + * And since SC is a constant power of two, we can convert the div + * into a shift. + * -johnstul@us.ibm.com "math is hard, lets go shopping!" + */ +static unsigned long cyc2ns_scale; +#define CYC2NS_SCALE_FACTOR 10 /* 2^10, carefully chosen */ + +static inline void set_cyc2ns_scale(unsigned long cpu_mhz) +{ + cyc2ns_scale = (1000 << CYC2NS_SCALE_FACTOR)/cpu_mhz; +} + +static inline unsigned long long cycles_2_ns(unsigned long long cyc) +{ + return (cyc * cyc2ns_scale) >> CYC2NS_SCALE_FACTOR; +} + +static unsigned long long monotonic_clock_hpet(void) +{ + unsigned long long last_offset, this_offset, base; + unsigned seq; + + /* atomically read monotonic base & last_offset */ + do { + seq = read_seqbegin(&monotonic_lock); + last_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + base = monotonic_base; + } while (read_seqretry(&monotonic_lock, seq)); + + /* Read the Time Stamp Counter */ + rdtscll(this_offset); + + /* return the value in ns */ + return base + cycles_2_ns(this_offset - last_offset); +} + +static unsigned long get_offset_hpet(void) +{ + register unsigned long eax, edx; + + eax = hpet_readl(HPET_COUNTER); + eax -= hpet_last; /* hpet delta */ + + /* + * Time offset = (hpet delta) * ( usecs per HPET clock ) + * = (hpet delta) * ( usecs per tick / HPET clocks per tick) + * = (hpet delta) * ( hpet_usec_quotient ) / (2^32) + * + * Where, + * hpet_usec_quotient = (2^32 * usecs per tick)/HPET clocks per tick + * + * Using a mull instead of a divl saves some cycles in critical path. + */ + ASM_MUL64_REG(eax, edx, hpet_usec_quotient, eax); + + /* our adjusted time offset in microseconds */ + return edx; +} + +static void mark_offset_hpet(void) +{ + unsigned long long this_offset, last_offset; + unsigned long offset; + + write_seqlock(&monotonic_lock); + last_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + rdtsc(last_tsc_low, last_tsc_high); + + offset = hpet_readl(HPET_T0_CMP) - hpet_tick; + if (unlikely(((offset - hpet_last) > hpet_tick) && (hpet_last != 0))) { + int lost_ticks = (offset - hpet_last) / hpet_tick; + jiffies_64 += lost_ticks; + } + hpet_last = offset; + + /* update the monotonic base value */ + this_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + monotonic_base += cycles_2_ns(this_offset - last_offset); + write_sequnlock(&monotonic_lock); +} + +static void delay_hpet(unsigned long loops) +{ + unsigned long hpet_start, hpet_end; + unsigned long eax; + + /* loops is the number of cpu cycles. Convert it to hpet clocks */ + ASM_MUL64_REG(eax, loops, tsc_hpet_quotient, loops); + + hpet_start = hpet_readl(HPET_COUNTER); + do { + rep_nop(); + hpet_end = hpet_readl(HPET_COUNTER); + } while ((hpet_end - hpet_start) < (loops)); +} + +static int __init init_hpet(char* override) +{ + unsigned long result, remain; + + /* check clock override */ + if (override[0] && strncmp(override,"hpet",4)) + return -ENODEV; + + if (!is_hpet_enabled()) + return -ENODEV; + + printk("Using HPET for gettimeofday\n"); + if (cpu_has_tsc) { + unsigned long tsc_quotient = calibrate_tsc_hpet(&tsc_hpet_quotient); + if (tsc_quotient) { + /* report CPU clock rate in Hz. + * The formula is (10^6 * 2^32) / (2^32 * 1 / (clocks/us)) = + * clock/second. Our precision is about 100 ppm. + */ + { unsigned long eax=0, edx=1000; + ASM_DIV64_REG(cpu_khz, edx, tsc_quotient, + eax, edx); + printk("Detected %lu.%03lu MHz processor.\n", + cpu_khz / 1000, cpu_khz % 1000); + } + set_cyc2ns_scale(cpu_khz/1000); + } + } + + /* + * Math to calculate hpet to usec multiplier + * Look for the comments at get_offset_hpet() + */ + ASM_DIV64_REG(result, remain, hpet_tick, 0, KERNEL_TICK_USEC); + if (remain > (hpet_tick >> 1)) + result++; /* rounding the result */ + hpet_usec_quotient = result; + + return 0; +} + +/************************************************************/ + +/* tsc timer_opts struct */ +static struct timer_opts timer_hpet = { + .name = "hpet", + .mark_offset = mark_offset_hpet, + .get_offset = get_offset_hpet, + .monotonic_clock = monotonic_clock_hpet, + .delay = delay_hpet, +}; + +struct init_timer_opts __initdata timer_hpet_init = { + .init = init_hpet, + .opts = &timer_hpet, +}; diff --git a/arch/i386/kernel/timers/timer_none.c b/arch/i386/kernel/timers/timer_none.c new file mode 100644 index 00000000000..4ea2f414dbb --- /dev/null +++ b/arch/i386/kernel/timers/timer_none.c @@ -0,0 +1,39 @@ +#include <linux/init.h> +#include <asm/timer.h> + +static void mark_offset_none(void) +{ + /* nothing needed */ +} + +static unsigned long get_offset_none(void) +{ + return 0; +} + +static unsigned long long monotonic_clock_none(void) +{ + return 0; +} + +static void delay_none(unsigned long loops) +{ + int d0; + __asm__ __volatile__( + "\tjmp 1f\n" + ".align 16\n" + "1:\tjmp 2f\n" + ".align 16\n" + "2:\tdecl %0\n\tjns 2b" + :"=&a" (d0) + :"0" (loops)); +} + +/* none timer_opts struct */ +struct timer_opts timer_none = { + .name = "none", + .mark_offset = mark_offset_none, + .get_offset = get_offset_none, + .monotonic_clock = monotonic_clock_none, + .delay = delay_none, +}; diff --git a/arch/i386/kernel/timers/timer_pit.c b/arch/i386/kernel/timers/timer_pit.c new file mode 100644 index 00000000000..967d5453cd0 --- /dev/null +++ b/arch/i386/kernel/timers/timer_pit.c @@ -0,0 +1,206 @@ +/* + * This code largely moved from arch/i386/kernel/time.c. + * See comments there for proper credits. + */ + +#include <linux/spinlock.h> +#include <linux/module.h> +#include <linux/device.h> +#include <linux/irq.h> +#include <linux/sysdev.h> +#include <linux/timex.h> +#include <asm/delay.h> +#include <asm/mpspec.h> +#include <asm/timer.h> +#include <asm/smp.h> +#include <asm/io.h> +#include <asm/arch_hooks.h> + +extern spinlock_t i8259A_lock; +extern spinlock_t i8253_lock; +#include "do_timer.h" +#include "io_ports.h" + +static int count_p; /* counter in get_offset_pit() */ + +static int __init init_pit(char* override) +{ + /* check clock override */ + if (override[0] && strncmp(override,"pit",3)) + printk(KERN_ERR "Warning: clock= override failed. Defaulting to PIT\n"); + + count_p = LATCH; + return 0; +} + +static void mark_offset_pit(void) +{ + /* nothing needed */ +} + +static unsigned long long monotonic_clock_pit(void) +{ + return 0; +} + +static void delay_pit(unsigned long loops) +{ + int d0; + __asm__ __volatile__( + "\tjmp 1f\n" + ".align 16\n" + "1:\tjmp 2f\n" + ".align 16\n" + "2:\tdecl %0\n\tjns 2b" + :"=&a" (d0) + :"0" (loops)); +} + + +/* This function must be called with xtime_lock held. + * It was inspired by Steve McCanne's microtime-i386 for BSD. -- jrs + * + * However, the pc-audio speaker driver changes the divisor so that + * it gets interrupted rather more often - it loads 64 into the + * counter rather than 11932! This has an adverse impact on + * do_gettimeoffset() -- it stops working! What is also not + * good is that the interval that our timer function gets called + * is no longer 10.0002 ms, but 9.9767 ms. To get around this + * would require using a different timing source. Maybe someone + * could use the RTC - I know that this can interrupt at frequencies + * ranging from 8192Hz to 2Hz. If I had the energy, I'd somehow fix + * it so that at startup, the timer code in sched.c would select + * using either the RTC or the 8253 timer. The decision would be + * based on whether there was any other device around that needed + * to trample on the 8253. I'd set up the RTC to interrupt at 1024 Hz, + * and then do some jiggery to have a version of do_timer that + * advanced the clock by 1/1024 s. Every time that reached over 1/100 + * of a second, then do all the old code. If the time was kept correct + * then do_gettimeoffset could just return 0 - there is no low order + * divider that can be accessed. + * + * Ideally, you would be able to use the RTC for the speaker driver, + * but it appears that the speaker driver really needs interrupt more + * often than every 120 us or so. + * + * Anyway, this needs more thought.... pjsg (1993-08-28) + * + * If you are really that interested, you should be reading + * comp.protocols.time.ntp! + */ + +static unsigned long get_offset_pit(void) +{ + int count; + unsigned long flags; + static unsigned long jiffies_p = 0; + + /* + * cache volatile jiffies temporarily; we have xtime_lock. + */ + unsigned long jiffies_t; + + spin_lock_irqsave(&i8253_lock, flags); + /* timer count may underflow right here */ + outb_p(0x00, PIT_MODE); /* latch the count ASAP */ + + count = inb_p(PIT_CH0); /* read the latched count */ + + /* + * We do this guaranteed double memory access instead of a _p + * postfix in the previous port access. Wheee, hackady hack + */ + jiffies_t = jiffies; + + count |= inb_p(PIT_CH0) << 8; + + /* VIA686a test code... reset the latch if count > max + 1 */ + if (count > LATCH) { + outb_p(0x34, PIT_MODE); + outb_p(LATCH & 0xff, PIT_CH0); + outb(LATCH >> 8, PIT_CH0); + count = LATCH - 1; + } + + /* + * avoiding timer inconsistencies (they are rare, but they happen)... + * there are two kinds of problems that must be avoided here: + * 1. the timer counter underflows + * 2. hardware problem with the timer, not giving us continuous time, + * the counter does small "jumps" upwards on some Pentium systems, + * (see c't 95/10 page 335 for Neptun bug.) + */ + + if( jiffies_t == jiffies_p ) { + if( count > count_p ) { + /* the nutcase */ + count = do_timer_overflow(count); + } + } else + jiffies_p = jiffies_t; + + count_p = count; + + spin_unlock_irqrestore(&i8253_lock, flags); + + count = ((LATCH-1) - count) * TICK_SIZE; + count = (count + LATCH/2) / LATCH; + + return count; +} + + +/* tsc timer_opts struct */ +struct timer_opts timer_pit = { + .name = "pit", + .mark_offset = mark_offset_pit, + .get_offset = get_offset_pit, + .monotonic_clock = monotonic_clock_pit, + .delay = delay_pit, +}; + +struct init_timer_opts __initdata timer_pit_init = { + .init = init_pit, + .opts = &timer_pit, +}; + +void setup_pit_timer(void) +{ + extern spinlock_t i8253_lock; + unsigned long flags; + + spin_lock_irqsave(&i8253_lock, flags); + outb_p(0x34,PIT_MODE); /* binary, mode 2, LSB/MSB, ch 0 */ + udelay(10); + outb_p(LATCH & 0xff , PIT_CH0); /* LSB */ + udelay(10); + outb(LATCH >> 8 , PIT_CH0); /* MSB */ + spin_unlock_irqrestore(&i8253_lock, flags); +} + +static int timer_resume(struct sys_device *dev) +{ + setup_pit_timer(); + return 0; +} + +static struct sysdev_class timer_sysclass = { + set_kset_name("timer_pit"), + .resume = timer_resume, +}; + +static struct sys_device device_timer = { + .id = 0, + .cls = &timer_sysclass, +}; + +static int __init init_timer_sysfs(void) +{ + int error = sysdev_class_register(&timer_sysclass); + if (!error) + error = sysdev_register(&device_timer); + return error; +} + +device_initcall(init_timer_sysfs); + diff --git a/arch/i386/kernel/timers/timer_pm.c b/arch/i386/kernel/timers/timer_pm.c new file mode 100644 index 00000000000..d77f22030fe --- /dev/null +++ b/arch/i386/kernel/timers/timer_pm.c @@ -0,0 +1,258 @@ +/* + * (C) Dominik Brodowski <linux@brodo.de> 2003 + * + * Driver to use the Power Management Timer (PMTMR) available in some + * southbridges as primary timing source for the Linux kernel. + * + * Based on parts of linux/drivers/acpi/hardware/hwtimer.c, timer_pit.c, + * timer_hpet.c, and on Arjan van de Ven's implementation for 2.4. + * + * This file is licensed under the GPL v2. + */ + + +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/device.h> +#include <linux/init.h> +#include <asm/types.h> +#include <asm/timer.h> +#include <asm/smp.h> +#include <asm/io.h> +#include <asm/arch_hooks.h> + +#include <linux/timex.h> +#include "mach_timer.h" + +/* Number of PMTMR ticks expected during calibration run */ +#define PMTMR_TICKS_PER_SEC 3579545 +#define PMTMR_EXPECTED_RATE \ + ((CALIBRATE_LATCH * (PMTMR_TICKS_PER_SEC >> 10)) / (CLOCK_TICK_RATE>>10)) + + +/* The I/O port the PMTMR resides at. + * The location is detected during setup_arch(), + * in arch/i386/acpi/boot.c */ +u32 pmtmr_ioport = 0; + + +/* value of the Power timer at last timer interrupt */ +static u32 offset_tick; +static u32 offset_delay; + +static unsigned long long monotonic_base; +static seqlock_t monotonic_lock = SEQLOCK_UNLOCKED; + +#define ACPI_PM_MASK 0xFFFFFF /* limit it to 24 bits */ + +/*helper function to safely read acpi pm timesource*/ +static inline u32 read_pmtmr(void) +{ + u32 v1=0,v2=0,v3=0; + /* It has been reported that because of various broken + * chipsets (ICH4, PIIX4 and PIIX4E) where the ACPI PM time + * source is not latched, so you must read it multiple + * times to insure a safe value is read. + */ + do { + v1 = inl(pmtmr_ioport); + v2 = inl(pmtmr_ioport); + v3 = inl(pmtmr_ioport); + } while ((v1 > v2 && v1 < v3) || (v2 > v3 && v2 < v1) + || (v3 > v1 && v3 < v2)); + + /* mask the output to 24 bits */ + return v2 & ACPI_PM_MASK; +} + + +/* + * Some boards have the PMTMR running way too fast. We check + * the PMTMR rate against PIT channel 2 to catch these cases. + */ +static int verify_pmtmr_rate(void) +{ + u32 value1, value2; + unsigned long count, delta; + + mach_prepare_counter(); + value1 = read_pmtmr(); + mach_countup(&count); + value2 = read_pmtmr(); + delta = (value2 - value1) & ACPI_PM_MASK; + + /* Check that the PMTMR delta is within 5% of what we expect */ + if (delta < (PMTMR_EXPECTED_RATE * 19) / 20 || + delta > (PMTMR_EXPECTED_RATE * 21) / 20) { + printk(KERN_INFO "PM-Timer running at invalid rate: %lu%% of normal - aborting.\n", 100UL * delta / PMTMR_EXPECTED_RATE); + return -1; + } + + return 0; +} + + +static int init_pmtmr(char* override) +{ + u32 value1, value2; + unsigned int i; + + if (override[0] && strncmp(override,"pmtmr",5)) + return -ENODEV; + + if (!pmtmr_ioport) + return -ENODEV; + + /* we use the TSC for delay_pmtmr, so make sure it exists */ + if (!cpu_has_tsc) + return -ENODEV; + + /* "verify" this timing source */ + value1 = read_pmtmr(); + for (i = 0; i < 10000; i++) { + value2 = read_pmtmr(); + if (value2 == value1) + continue; + if (value2 > value1) + goto pm_good; + if ((value2 < value1) && ((value2) < 0xFFF)) + goto pm_good; + printk(KERN_INFO "PM-Timer had inconsistent results: 0x%#x, 0x%#x - aborting.\n", value1, value2); + return -EINVAL; + } + printk(KERN_INFO "PM-Timer had no reasonable result: 0x%#x - aborting.\n", value1); + return -ENODEV; + +pm_good: + if (verify_pmtmr_rate() != 0) + return -ENODEV; + + init_cpu_khz(); + return 0; +} + +static inline u32 cyc2us(u32 cycles) +{ + /* The Power Management Timer ticks at 3.579545 ticks per microsecond. + * 1 / PM_TIMER_FREQUENCY == 0.27936511 =~ 286/1024 [error: 0.024%] + * + * Even with HZ = 100, delta is at maximum 35796 ticks, so it can + * easily be multiplied with 286 (=0x11E) without having to fear + * u32 overflows. + */ + cycles *= 286; + return (cycles >> 10); +} + +/* + * this gets called during each timer interrupt + * - Called while holding the writer xtime_lock + */ +static void mark_offset_pmtmr(void) +{ + u32 lost, delta, last_offset; + static int first_run = 1; + last_offset = offset_tick; + + write_seqlock(&monotonic_lock); + + offset_tick = read_pmtmr(); + + /* calculate tick interval */ + delta = (offset_tick - last_offset) & ACPI_PM_MASK; + + /* convert to usecs */ + delta = cyc2us(delta); + + /* update the monotonic base value */ + monotonic_base += delta * NSEC_PER_USEC; + write_sequnlock(&monotonic_lock); + + /* convert to ticks */ + delta += offset_delay; + lost = delta / (USEC_PER_SEC / HZ); + offset_delay = delta % (USEC_PER_SEC / HZ); + + + /* compensate for lost ticks */ + if (lost >= 2) + jiffies_64 += lost - 1; + + /* don't calculate delay for first run, + or if we've got less then a tick */ + if (first_run || (lost < 1)) { + first_run = 0; + offset_delay = 0; + } +} + + +static unsigned long long monotonic_clock_pmtmr(void) +{ + u32 last_offset, this_offset; + unsigned long long base, ret; + unsigned seq; + + + /* atomically read monotonic base & last_offset */ + do { + seq = read_seqbegin(&monotonic_lock); + last_offset = offset_tick; + base = monotonic_base; + } while (read_seqretry(&monotonic_lock, seq)); + + /* Read the pmtmr */ + this_offset = read_pmtmr(); + + /* convert to nanoseconds */ + ret = (this_offset - last_offset) & ACPI_PM_MASK; + ret = base + (cyc2us(ret) * NSEC_PER_USEC); + return ret; +} + +static void delay_pmtmr(unsigned long loops) +{ + unsigned long bclock, now; + + rdtscl(bclock); + do + { + rep_nop(); + rdtscl(now); + } while ((now-bclock) < loops); +} + + +/* + * get the offset (in microseconds) from the last call to mark_offset() + * - Called holding a reader xtime_lock + */ +static unsigned long get_offset_pmtmr(void) +{ + u32 now, offset, delta = 0; + + offset = offset_tick; + now = read_pmtmr(); + delta = (now - offset)&ACPI_PM_MASK; + + return (unsigned long) offset_delay + cyc2us(delta); +} + + +/* acpi timer_opts struct */ +static struct timer_opts timer_pmtmr = { + .name = "pmtmr", + .mark_offset = mark_offset_pmtmr, + .get_offset = get_offset_pmtmr, + .monotonic_clock = monotonic_clock_pmtmr, + .delay = delay_pmtmr, +}; + +struct init_timer_opts __initdata timer_pmtmr_init = { + .init = init_pmtmr, + .opts = &timer_pmtmr, +}; + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Dominik Brodowski <linux@brodo.de>"); +MODULE_DESCRIPTION("Power Management Timer (PMTMR) as primary timing source for x86"); diff --git a/arch/i386/kernel/timers/timer_tsc.c b/arch/i386/kernel/timers/timer_tsc.c new file mode 100644 index 00000000000..a685994e5c8 --- /dev/null +++ b/arch/i386/kernel/timers/timer_tsc.c @@ -0,0 +1,560 @@ +/* + * This code largely moved from arch/i386/kernel/time.c. + * See comments there for proper credits. + * + * 2004-06-25 Jesper Juhl + * moved mark_offset_tsc below cpufreq_delayed_get to avoid gcc 3.4 + * failing to inline. + */ + +#include <linux/spinlock.h> +#include <linux/init.h> +#include <linux/timex.h> +#include <linux/errno.h> +#include <linux/cpufreq.h> +#include <linux/string.h> +#include <linux/jiffies.h> + +#include <asm/timer.h> +#include <asm/io.h> +/* processor.h for distable_tsc flag */ +#include <asm/processor.h> + +#include "io_ports.h" +#include "mach_timer.h" + +#include <asm/hpet.h> + +#ifdef CONFIG_HPET_TIMER +static unsigned long hpet_usec_quotient; +static unsigned long hpet_last; +static struct timer_opts timer_tsc; +#endif + +static inline void cpufreq_delayed_get(void); + +int tsc_disable __initdata = 0; + +extern spinlock_t i8253_lock; + +static int use_tsc; +/* Number of usecs that the last interrupt was delayed */ +static int delay_at_last_interrupt; + +static unsigned long last_tsc_low; /* lsb 32 bits of Time Stamp Counter */ +static unsigned long last_tsc_high; /* msb 32 bits of Time Stamp Counter */ +static unsigned long long monotonic_base; +static seqlock_t monotonic_lock = SEQLOCK_UNLOCKED; + +/* convert from cycles(64bits) => nanoseconds (64bits) + * basic equation: + * ns = cycles / (freq / ns_per_sec) + * ns = cycles * (ns_per_sec / freq) + * ns = cycles * (10^9 / (cpu_mhz * 10^6)) + * ns = cycles * (10^3 / cpu_mhz) + * + * Then we use scaling math (suggested by george@mvista.com) to get: + * ns = cycles * (10^3 * SC / cpu_mhz) / SC + * ns = cycles * cyc2ns_scale / SC + * + * And since SC is a constant power of two, we can convert the div + * into a shift. + * -johnstul@us.ibm.com "math is hard, lets go shopping!" + */ +static unsigned long cyc2ns_scale; +#define CYC2NS_SCALE_FACTOR 10 /* 2^10, carefully chosen */ + +static inline void set_cyc2ns_scale(unsigned long cpu_mhz) +{ + cyc2ns_scale = (1000 << CYC2NS_SCALE_FACTOR)/cpu_mhz; +} + +static inline unsigned long long cycles_2_ns(unsigned long long cyc) +{ + return (cyc * cyc2ns_scale) >> CYC2NS_SCALE_FACTOR; +} + +static int count2; /* counter for mark_offset_tsc() */ + +/* Cached *multiplier* to convert TSC counts to microseconds. + * (see the equation below). + * Equal to 2^32 * (1 / (clocks per usec) ). + * Initialized in time_init. + */ +static unsigned long fast_gettimeoffset_quotient; + +static unsigned long get_offset_tsc(void) +{ + register unsigned long eax, edx; + + /* Read the Time Stamp Counter */ + + rdtsc(eax,edx); + + /* .. relative to previous jiffy (32 bits is enough) */ + eax -= last_tsc_low; /* tsc_low delta */ + + /* + * Time offset = (tsc_low delta) * fast_gettimeoffset_quotient + * = (tsc_low delta) * (usecs_per_clock) + * = (tsc_low delta) * (usecs_per_jiffy / clocks_per_jiffy) + * + * Using a mull instead of a divl saves up to 31 clock cycles + * in the critical path. + */ + + __asm__("mull %2" + :"=a" (eax), "=d" (edx) + :"rm" (fast_gettimeoffset_quotient), + "0" (eax)); + + /* our adjusted time offset in microseconds */ + return delay_at_last_interrupt + edx; +} + +static unsigned long long monotonic_clock_tsc(void) +{ + unsigned long long last_offset, this_offset, base; + unsigned seq; + + /* atomically read monotonic base & last_offset */ + do { + seq = read_seqbegin(&monotonic_lock); + last_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + base = monotonic_base; + } while (read_seqretry(&monotonic_lock, seq)); + + /* Read the Time Stamp Counter */ + rdtscll(this_offset); + + /* return the value in ns */ + return base + cycles_2_ns(this_offset - last_offset); +} + +/* + * Scheduler clock - returns current time in nanosec units. + */ +unsigned long long sched_clock(void) +{ + unsigned long long this_offset; + + /* + * In the NUMA case we dont use the TSC as they are not + * synchronized across all CPUs. + */ +#ifndef CONFIG_NUMA + if (!use_tsc) +#endif + /* no locking but a rare wrong value is not a big deal */ + return jiffies_64 * (1000000000 / HZ); + + /* Read the Time Stamp Counter */ + rdtscll(this_offset); + + /* return the value in ns */ + return cycles_2_ns(this_offset); +} + +static void delay_tsc(unsigned long loops) +{ + unsigned long bclock, now; + + rdtscl(bclock); + do + { + rep_nop(); + rdtscl(now); + } while ((now-bclock) < loops); +} + +#ifdef CONFIG_HPET_TIMER +static void mark_offset_tsc_hpet(void) +{ + unsigned long long this_offset, last_offset; + unsigned long offset, temp, hpet_current; + + write_seqlock(&monotonic_lock); + last_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + /* + * It is important that these two operations happen almost at + * the same time. We do the RDTSC stuff first, since it's + * faster. To avoid any inconsistencies, we need interrupts + * disabled locally. + */ + /* + * Interrupts are just disabled locally since the timer irq + * has the SA_INTERRUPT flag set. -arca + */ + /* read Pentium cycle counter */ + + hpet_current = hpet_readl(HPET_COUNTER); + rdtsc(last_tsc_low, last_tsc_high); + + /* lost tick compensation */ + offset = hpet_readl(HPET_T0_CMP) - hpet_tick; + if (unlikely(((offset - hpet_last) > hpet_tick) && (hpet_last != 0))) { + int lost_ticks = (offset - hpet_last) / hpet_tick; + jiffies_64 += lost_ticks; + } + hpet_last = hpet_current; + + /* update the monotonic base value */ + this_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + monotonic_base += cycles_2_ns(this_offset - last_offset); + write_sequnlock(&monotonic_lock); + + /* calculate delay_at_last_interrupt */ + /* + * Time offset = (hpet delta) * ( usecs per HPET clock ) + * = (hpet delta) * ( usecs per tick / HPET clocks per tick) + * = (hpet delta) * ( hpet_usec_quotient ) / (2^32) + * Where, + * hpet_usec_quotient = (2^32 * usecs per tick)/HPET clocks per tick + */ + delay_at_last_interrupt = hpet_current - offset; + ASM_MUL64_REG(temp, delay_at_last_interrupt, + hpet_usec_quotient, delay_at_last_interrupt); +} +#endif + + +#ifdef CONFIG_CPU_FREQ +#include <linux/workqueue.h> + +static unsigned int cpufreq_delayed_issched = 0; +static unsigned int cpufreq_init = 0; +static struct work_struct cpufreq_delayed_get_work; + +static void handle_cpufreq_delayed_get(void *v) +{ + unsigned int cpu; + for_each_online_cpu(cpu) { + cpufreq_get(cpu); + } + cpufreq_delayed_issched = 0; +} + +/* if we notice lost ticks, schedule a call to cpufreq_get() as it tries + * to verify the CPU frequency the timing core thinks the CPU is running + * at is still correct. + */ +static inline void cpufreq_delayed_get(void) +{ + if (cpufreq_init && !cpufreq_delayed_issched) { + cpufreq_delayed_issched = 1; + printk(KERN_DEBUG "Losing some ticks... checking if CPU frequency changed.\n"); + schedule_work(&cpufreq_delayed_get_work); + } +} + +/* If the CPU frequency is scaled, TSC-based delays will need a different + * loops_per_jiffy value to function properly. + */ + +static unsigned int ref_freq = 0; +static unsigned long loops_per_jiffy_ref = 0; + +#ifndef CONFIG_SMP +static unsigned long fast_gettimeoffset_ref = 0; +static unsigned long cpu_khz_ref = 0; +#endif + +static int +time_cpufreq_notifier(struct notifier_block *nb, unsigned long val, + void *data) +{ + struct cpufreq_freqs *freq = data; + + if (val != CPUFREQ_RESUMECHANGE) + write_seqlock_irq(&xtime_lock); + if (!ref_freq) { + ref_freq = freq->old; + loops_per_jiffy_ref = cpu_data[freq->cpu].loops_per_jiffy; +#ifndef CONFIG_SMP + fast_gettimeoffset_ref = fast_gettimeoffset_quotient; + cpu_khz_ref = cpu_khz; +#endif + } + + if ((val == CPUFREQ_PRECHANGE && freq->old < freq->new) || + (val == CPUFREQ_POSTCHANGE && freq->old > freq->new) || + (val == CPUFREQ_RESUMECHANGE)) { + if (!(freq->flags & CPUFREQ_CONST_LOOPS)) + cpu_data[freq->cpu].loops_per_jiffy = cpufreq_scale(loops_per_jiffy_ref, ref_freq, freq->new); +#ifndef CONFIG_SMP + if (cpu_khz) + cpu_khz = cpufreq_scale(cpu_khz_ref, ref_freq, freq->new); + if (use_tsc) { + if (!(freq->flags & CPUFREQ_CONST_LOOPS)) { + fast_gettimeoffset_quotient = cpufreq_scale(fast_gettimeoffset_ref, freq->new, ref_freq); + set_cyc2ns_scale(cpu_khz/1000); + } + } +#endif + } + + if (val != CPUFREQ_RESUMECHANGE) + write_sequnlock_irq(&xtime_lock); + + return 0; +} + +static struct notifier_block time_cpufreq_notifier_block = { + .notifier_call = time_cpufreq_notifier +}; + + +static int __init cpufreq_tsc(void) +{ + int ret; + INIT_WORK(&cpufreq_delayed_get_work, handle_cpufreq_delayed_get, NULL); + ret = cpufreq_register_notifier(&time_cpufreq_notifier_block, + CPUFREQ_TRANSITION_NOTIFIER); + if (!ret) + cpufreq_init = 1; + return ret; +} +core_initcall(cpufreq_tsc); + +#else /* CONFIG_CPU_FREQ */ +static inline void cpufreq_delayed_get(void) { return; } +#endif + +static void mark_offset_tsc(void) +{ + unsigned long lost,delay; + unsigned long delta = last_tsc_low; + int count; + int countmp; + static int count1 = 0; + unsigned long long this_offset, last_offset; + static int lost_count = 0; + + write_seqlock(&monotonic_lock); + last_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + /* + * It is important that these two operations happen almost at + * the same time. We do the RDTSC stuff first, since it's + * faster. To avoid any inconsistencies, we need interrupts + * disabled locally. + */ + + /* + * Interrupts are just disabled locally since the timer irq + * has the SA_INTERRUPT flag set. -arca + */ + + /* read Pentium cycle counter */ + + rdtsc(last_tsc_low, last_tsc_high); + + spin_lock(&i8253_lock); + outb_p(0x00, PIT_MODE); /* latch the count ASAP */ + + count = inb_p(PIT_CH0); /* read the latched count */ + count |= inb(PIT_CH0) << 8; + + /* + * VIA686a test code... reset the latch if count > max + 1 + * from timer_pit.c - cjb + */ + if (count > LATCH) { + outb_p(0x34, PIT_MODE); + outb_p(LATCH & 0xff, PIT_CH0); + outb(LATCH >> 8, PIT_CH0); + count = LATCH - 1; + } + + spin_unlock(&i8253_lock); + + if (pit_latch_buggy) { + /* get center value of last 3 time lutch */ + if ((count2 >= count && count >= count1) + || (count1 >= count && count >= count2)) { + count2 = count1; count1 = count; + } else if ((count1 >= count2 && count2 >= count) + || (count >= count2 && count2 >= count1)) { + countmp = count;count = count2; + count2 = count1;count1 = countmp; + } else { + count2 = count1; count1 = count; count = count1; + } + } + + /* lost tick compensation */ + delta = last_tsc_low - delta; + { + register unsigned long eax, edx; + eax = delta; + __asm__("mull %2" + :"=a" (eax), "=d" (edx) + :"rm" (fast_gettimeoffset_quotient), + "0" (eax)); + delta = edx; + } + delta += delay_at_last_interrupt; + lost = delta/(1000000/HZ); + delay = delta%(1000000/HZ); + if (lost >= 2) { + jiffies_64 += lost-1; + + /* sanity check to ensure we're not always losing ticks */ + if (lost_count++ > 100) { + printk(KERN_WARNING "Losing too many ticks!\n"); + printk(KERN_WARNING "TSC cannot be used as a timesource. \n"); + printk(KERN_WARNING "Possible reasons for this are:\n"); + printk(KERN_WARNING " You're running with Speedstep,\n"); + printk(KERN_WARNING " You don't have DMA enabled for your hard disk (see hdparm),\n"); + printk(KERN_WARNING " Incorrect TSC synchronization on an SMP system (see dmesg).\n"); + printk(KERN_WARNING "Falling back to a sane timesource now.\n"); + + clock_fallback(); + } + /* ... but give the TSC a fair chance */ + if (lost_count > 25) + cpufreq_delayed_get(); + } else + lost_count = 0; + /* update the monotonic base value */ + this_offset = ((unsigned long long)last_tsc_high<<32)|last_tsc_low; + monotonic_base += cycles_2_ns(this_offset - last_offset); + write_sequnlock(&monotonic_lock); + + /* calculate delay_at_last_interrupt */ + count = ((LATCH-1) - count) * TICK_SIZE; + delay_at_last_interrupt = (count + LATCH/2) / LATCH; + + /* catch corner case where tick rollover occured + * between tsc and pit reads (as noted when + * usec delta is > 90% # of usecs/tick) + */ + if (lost && abs(delay - delay_at_last_interrupt) > (900000/HZ)) + jiffies_64++; +} + +static int __init init_tsc(char* override) +{ + + /* check clock override */ + if (override[0] && strncmp(override,"tsc",3)) { +#ifdef CONFIG_HPET_TIMER + if (is_hpet_enabled()) { + printk(KERN_ERR "Warning: clock= override failed. Defaulting to tsc\n"); + } else +#endif + { + return -ENODEV; + } + } + + /* + * If we have APM enabled or the CPU clock speed is variable + * (CPU stops clock on HLT or slows clock to save power) + * then the TSC timestamps may diverge by up to 1 jiffy from + * 'real time' but nothing will break. + * The most frequent case is that the CPU is "woken" from a halt + * state by the timer interrupt itself, so we get 0 error. In the + * rare cases where a driver would "wake" the CPU and request a + * timestamp, the maximum error is < 1 jiffy. But timestamps are + * still perfectly ordered. + * Note that the TSC counter will be reset if APM suspends + * to disk; this won't break the kernel, though, 'cuz we're + * smart. See arch/i386/kernel/apm.c. + */ + /* + * Firstly we have to do a CPU check for chips with + * a potentially buggy TSC. At this point we haven't run + * the ident/bugs checks so we must run this hook as it + * may turn off the TSC flag. + * + * NOTE: this doesn't yet handle SMP 486 machines where only + * some CPU's have a TSC. Thats never worked and nobody has + * moaned if you have the only one in the world - you fix it! + */ + + count2 = LATCH; /* initialize counter for mark_offset_tsc() */ + + if (cpu_has_tsc) { + unsigned long tsc_quotient; +#ifdef CONFIG_HPET_TIMER + if (is_hpet_enabled()){ + unsigned long result, remain; + printk("Using TSC for gettimeofday\n"); + tsc_quotient = calibrate_tsc_hpet(NULL); + timer_tsc.mark_offset = &mark_offset_tsc_hpet; + /* + * Math to calculate hpet to usec multiplier + * Look for the comments at get_offset_tsc_hpet() + */ + ASM_DIV64_REG(result, remain, hpet_tick, + 0, KERNEL_TICK_USEC); + if (remain > (hpet_tick >> 1)) + result++; /* rounding the result */ + + hpet_usec_quotient = result; + } else +#endif + { + tsc_quotient = calibrate_tsc(); + } + + if (tsc_quotient) { + fast_gettimeoffset_quotient = tsc_quotient; + use_tsc = 1; + /* + * We could be more selective here I suspect + * and just enable this for the next intel chips ? + */ + /* report CPU clock rate in Hz. + * The formula is (10^6 * 2^32) / (2^32 * 1 / (clocks/us)) = + * clock/second. Our precision is about 100 ppm. + */ + { unsigned long eax=0, edx=1000; + __asm__("divl %2" + :"=a" (cpu_khz), "=d" (edx) + :"r" (tsc_quotient), + "0" (eax), "1" (edx)); + printk("Detected %lu.%03lu MHz processor.\n", cpu_khz / 1000, cpu_khz % 1000); + } + set_cyc2ns_scale(cpu_khz/1000); + return 0; + } + } + return -ENODEV; +} + +#ifndef CONFIG_X86_TSC +/* disable flag for tsc. Takes effect by clearing the TSC cpu flag + * in cpu/common.c */ +static int __init tsc_setup(char *str) +{ + tsc_disable = 1; + return 1; +} +#else +static int __init tsc_setup(char *str) +{ + printk(KERN_WARNING "notsc: Kernel compiled with CONFIG_X86_TSC, " + "cannot disable TSC.\n"); + return 1; +} +#endif +__setup("notsc", tsc_setup); + + + +/************************************************************/ + +/* tsc timer_opts struct */ +static struct timer_opts timer_tsc = { + .name = "tsc", + .mark_offset = mark_offset_tsc, + .get_offset = get_offset_tsc, + .monotonic_clock = monotonic_clock_tsc, + .delay = delay_tsc, +}; + +struct init_timer_opts __initdata timer_tsc_init = { + .init = init_tsc, + .opts = &timer_tsc, +}; diff --git a/arch/i386/kernel/trampoline.S b/arch/i386/kernel/trampoline.S new file mode 100644 index 00000000000..fcce0e61b0e --- /dev/null +++ b/arch/i386/kernel/trampoline.S @@ -0,0 +1,80 @@ +/* + * + * Trampoline.S Derived from Setup.S by Linus Torvalds + * + * 4 Jan 1997 Michael Chastain: changed to gnu as. + * + * This is only used for booting secondary CPUs in SMP machine + * + * Entry: CS:IP point to the start of our code, we are + * in real mode with no stack, but the rest of the + * trampoline page to make our stack and everything else + * is a mystery. + * + * In fact we don't actually need a stack so we don't + * set one up. + * + * We jump into the boot/compressed/head.S code. So you'd + * better be running a compressed kernel image or you + * won't get very far. + * + * On entry to trampoline_data, the processor is in real mode + * with 16-bit addressing and 16-bit data. CS has some value + * and IP is zero. Thus, data addresses need to be absolute + * (no relocation) and are taken with regard to r_base. + * + * If you work on this file, check the object module with + * objdump --reloc to make sure there are no relocation + * entries except for: + * + * TYPE VALUE + * R_386_32 startup_32_smp + * R_386_32 boot_gdt_table + */ + +#include <linux/linkage.h> +#include <asm/segment.h> +#include <asm/page.h> + +.data + +.code16 + +ENTRY(trampoline_data) +r_base = . + wbinvd # Needed for NUMA-Q should be harmless for others + mov %cs, %ax # Code and data in the same place + mov %ax, %ds + + cli # We should be safe anyway + + movl $0xA5A5A5A5, trampoline_data - r_base + # write marker for master knows we're running + + /* GDT tables in non default location kernel can be beyond 16MB and + * lgdt will not be able to load the address as in real mode default + * operand size is 16bit. Use lgdtl instead to force operand size + * to 32 bit. + */ + + lidtl boot_idt - r_base # load idt with 0, 0 + lgdtl boot_gdt - r_base # load gdt with whatever is appropriate + + xor %ax, %ax + inc %ax # protected mode (PE) bit + lmsw %ax # into protected mode + # flush prefetch and jump to startup_32_smp in arch/i386/kernel/head.S + ljmpl $__BOOT_CS, $(startup_32_smp-__PAGE_OFFSET) + + # These need to be in the same 64K segment as the above; + # hence we don't use the boot_gdt_descr defined in head.S +boot_gdt: + .word __BOOT_DS + 7 # gdt limit + .long boot_gdt_table-__PAGE_OFFSET # gdt base + +boot_idt: + .word 0 # idt limit = 0 + .long 0 # idt base = 0L + +.globl trampoline_end +trampoline_end: diff --git a/arch/i386/kernel/traps.c b/arch/i386/kernel/traps.c new file mode 100644 index 00000000000..6c0e383915b --- /dev/null +++ b/arch/i386/kernel/traps.c @@ -0,0 +1,1084 @@ +/* + * linux/arch/i386/traps.c + * + * Copyright (C) 1991, 1992 Linus Torvalds + * + * Pentium III FXSR, SSE support + * Gareth Hughes <gareth@valinux.com>, May 2000 + */ + +/* + * 'Traps.c' handles hardware traps and faults after we have saved some + * state in 'asm.s'. + */ +#include <linux/config.h> +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/string.h> +#include <linux/errno.h> +#include <linux/timer.h> +#include <linux/mm.h> +#include <linux/init.h> +#include <linux/delay.h> +#include <linux/spinlock.h> +#include <linux/interrupt.h> +#include <linux/highmem.h> +#include <linux/kallsyms.h> +#include <linux/ptrace.h> +#include <linux/utsname.h> +#include <linux/kprobes.h> + +#ifdef CONFIG_EISA +#include <linux/ioport.h> +#include <linux/eisa.h> +#endif + +#ifdef CONFIG_MCA +#include <linux/mca.h> +#endif + +#include <asm/processor.h> +#include <asm/system.h> +#include <asm/uaccess.h> +#include <asm/io.h> +#include <asm/atomic.h> +#include <asm/debugreg.h> +#include <asm/desc.h> +#include <asm/i387.h> +#include <asm/nmi.h> + +#include <asm/smp.h> +#include <asm/arch_hooks.h> +#include <asm/kdebug.h> + +#include <linux/irq.h> +#include <linux/module.h> + +#include "mach_traps.h" + +asmlinkage int system_call(void); + +struct desc_struct default_ldt[] = { { 0, 0 }, { 0, 0 }, { 0, 0 }, + { 0, 0 }, { 0, 0 } }; + +/* Do we ignore FPU interrupts ? */ +char ignore_fpu_irq = 0; + +/* + * The IDT has to be page-aligned to simplify the Pentium + * F0 0F bug workaround.. We have a special link segment + * for this. + */ +struct desc_struct idt_table[256] __attribute__((__section__(".data.idt"))) = { {0, 0}, }; + +asmlinkage void divide_error(void); +asmlinkage void debug(void); +asmlinkage void nmi(void); +asmlinkage void int3(void); +asmlinkage void overflow(void); +asmlinkage void bounds(void); +asmlinkage void invalid_op(void); +asmlinkage void device_not_available(void); +asmlinkage void coprocessor_segment_overrun(void); +asmlinkage void invalid_TSS(void); +asmlinkage void segment_not_present(void); +asmlinkage void stack_segment(void); +asmlinkage void general_protection(void); +asmlinkage void page_fault(void); +asmlinkage void coprocessor_error(void); +asmlinkage void simd_coprocessor_error(void); +asmlinkage void alignment_check(void); +asmlinkage void spurious_interrupt_bug(void); +asmlinkage void machine_check(void); + +static int kstack_depth_to_print = 24; +struct notifier_block *i386die_chain; +static DEFINE_SPINLOCK(die_notifier_lock); + +int register_die_notifier(struct notifier_block *nb) +{ + int err = 0; + unsigned long flags; + spin_lock_irqsave(&die_notifier_lock, flags); + err = notifier_chain_register(&i386die_chain, nb); + spin_unlock_irqrestore(&die_notifier_lock, flags); + return err; +} + +static inline int valid_stack_ptr(struct thread_info *tinfo, void *p) +{ + return p > (void *)tinfo && + p < (void *)tinfo + THREAD_SIZE - 3; +} + +static inline unsigned long print_context_stack(struct thread_info *tinfo, + unsigned long *stack, unsigned long ebp) +{ + unsigned long addr; + +#ifdef CONFIG_FRAME_POINTER + while (valid_stack_ptr(tinfo, (void *)ebp)) { + addr = *(unsigned long *)(ebp + 4); + printk(" [<%08lx>] ", addr); + print_symbol("%s", addr); + printk("\n"); + ebp = *(unsigned long *)ebp; + } +#else + while (valid_stack_ptr(tinfo, stack)) { + addr = *stack++; + if (__kernel_text_address(addr)) { + printk(" [<%08lx>]", addr); + print_symbol(" %s", addr); + printk("\n"); + } + } +#endif + return ebp; +} + +void show_trace(struct task_struct *task, unsigned long * stack) +{ + unsigned long ebp; + + if (!task) + task = current; + + if (task == current) { + /* Grab ebp right from our regs */ + asm ("movl %%ebp, %0" : "=r" (ebp) : ); + } else { + /* ebp is the last reg pushed by switch_to */ + ebp = *(unsigned long *) task->thread.esp; + } + + while (1) { + struct thread_info *context; + context = (struct thread_info *) + ((unsigned long)stack & (~(THREAD_SIZE - 1))); + ebp = print_context_stack(context, stack, ebp); + stack = (unsigned long*)context->previous_esp; + if (!stack) + break; + printk(" =======================\n"); + } +} + +void show_stack(struct task_struct *task, unsigned long *esp) +{ + unsigned long *stack; + int i; + + if (esp == NULL) { + if (task) + esp = (unsigned long*)task->thread.esp; + else + esp = (unsigned long *)&esp; + } + + stack = esp; + for(i = 0; i < kstack_depth_to_print; i++) { + if (kstack_end(stack)) + break; + if (i && ((i % 8) == 0)) + printk("\n "); + printk("%08lx ", *stack++); + } + printk("\nCall Trace:\n"); + show_trace(task, esp); +} + +/* + * The architecture-independent dump_stack generator + */ +void dump_stack(void) +{ + unsigned long stack; + + show_trace(current, &stack); +} + +EXPORT_SYMBOL(dump_stack); + +void show_registers(struct pt_regs *regs) +{ + int i; + int in_kernel = 1; + unsigned long esp; + unsigned short ss; + + esp = (unsigned long) (®s->esp); + ss = __KERNEL_DS; + if (regs->xcs & 3) { + in_kernel = 0; + esp = regs->esp; + ss = regs->xss & 0xffff; + } + print_modules(); + printk("CPU: %d\nEIP: %04x:[<%08lx>] %s VLI\nEFLAGS: %08lx" + " (%s) \n", + smp_processor_id(), 0xffff & regs->xcs, regs->eip, + print_tainted(), regs->eflags, system_utsname.release); + print_symbol("EIP is at %s\n", regs->eip); + printk("eax: %08lx ebx: %08lx ecx: %08lx edx: %08lx\n", + regs->eax, regs->ebx, regs->ecx, regs->edx); + printk("esi: %08lx edi: %08lx ebp: %08lx esp: %08lx\n", + regs->esi, regs->edi, regs->ebp, esp); + printk("ds: %04x es: %04x ss: %04x\n", + regs->xds & 0xffff, regs->xes & 0xffff, ss); + printk("Process %s (pid: %d, threadinfo=%p task=%p)", + current->comm, current->pid, current_thread_info(), current); + /* + * When in-kernel, we also print out the stack and code at the + * time of the fault.. + */ + if (in_kernel) { + u8 *eip; + + printk("\nStack: "); + show_stack(NULL, (unsigned long*)esp); + + printk("Code: "); + + eip = (u8 *)regs->eip - 43; + for (i = 0; i < 64; i++, eip++) { + unsigned char c; + + if (eip < (u8 *)PAGE_OFFSET || __get_user(c, eip)) { + printk(" Bad EIP value."); + break; + } + if (eip == (u8 *)regs->eip) + printk("<%02x> ", c); + else + printk("%02x ", c); + } + } + printk("\n"); +} + +static void handle_BUG(struct pt_regs *regs) +{ + unsigned short ud2; + unsigned short line; + char *file; + char c; + unsigned long eip; + + if (regs->xcs & 3) + goto no_bug; /* Not in kernel */ + + eip = regs->eip; + + if (eip < PAGE_OFFSET) + goto no_bug; + if (__get_user(ud2, (unsigned short *)eip)) + goto no_bug; + if (ud2 != 0x0b0f) + goto no_bug; + if (__get_user(line, (unsigned short *)(eip + 2))) + goto bug; + if (__get_user(file, (char **)(eip + 4)) || + (unsigned long)file < PAGE_OFFSET || __get_user(c, file)) + file = "<bad filename>"; + + printk("------------[ cut here ]------------\n"); + printk(KERN_ALERT "kernel BUG at %s:%d!\n", file, line); + +no_bug: + return; + + /* Here we know it was a BUG but file-n-line is unavailable */ +bug: + printk("Kernel BUG\n"); +} + +void die(const char * str, struct pt_regs * regs, long err) +{ + static struct { + spinlock_t lock; + u32 lock_owner; + int lock_owner_depth; + } die = { + .lock = SPIN_LOCK_UNLOCKED, + .lock_owner = -1, + .lock_owner_depth = 0 + }; + static int die_counter; + + if (die.lock_owner != _smp_processor_id()) { + console_verbose(); + spin_lock_irq(&die.lock); + die.lock_owner = smp_processor_id(); + die.lock_owner_depth = 0; + bust_spinlocks(1); + } + + if (++die.lock_owner_depth < 3) { + int nl = 0; + handle_BUG(regs); + printk(KERN_ALERT "%s: %04lx [#%d]\n", str, err & 0xffff, ++die_counter); +#ifdef CONFIG_PREEMPT + printk("PREEMPT "); + nl = 1; +#endif +#ifdef CONFIG_SMP + printk("SMP "); + nl = 1; +#endif +#ifdef CONFIG_DEBUG_PAGEALLOC + printk("DEBUG_PAGEALLOC"); + nl = 1; +#endif + if (nl) + printk("\n"); + notify_die(DIE_OOPS, (char *)str, regs, err, 255, SIGSEGV); + show_registers(regs); + } else + printk(KERN_ERR "Recursive die() failure, output suppressed\n"); + + bust_spinlocks(0); + die.lock_owner = -1; + spin_unlock_irq(&die.lock); + if (in_interrupt()) + panic("Fatal exception in interrupt"); + + if (panic_on_oops) { + printk(KERN_EMERG "Fatal exception: panic in 5 seconds\n"); + ssleep(5); + panic("Fatal exception"); + } + do_exit(SIGSEGV); +} + +static inline void die_if_kernel(const char * str, struct pt_regs * regs, long err) +{ + if (!(regs->eflags & VM_MASK) && !(3 & regs->xcs)) + die(str, regs, err); +} + +static void do_trap(int trapnr, int signr, char *str, int vm86, + struct pt_regs * regs, long error_code, siginfo_t *info) +{ + if (regs->eflags & VM_MASK) { + if (vm86) + goto vm86_trap; + goto trap_signal; + } + + if (!(regs->xcs & 3)) + goto kernel_trap; + + trap_signal: { + struct task_struct *tsk = current; + tsk->thread.error_code = error_code; + tsk->thread.trap_no = trapnr; + if (info) + force_sig_info(signr, info, tsk); + else + force_sig(signr, tsk); + return; + } + + kernel_trap: { + if (!fixup_exception(regs)) + die(str, regs, error_code); + return; + } + + vm86_trap: { + int ret = handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, trapnr); + if (ret) goto trap_signal; + return; + } +} + +#define DO_ERROR(trapnr, signr, str, name) \ +fastcall void do_##name(struct pt_regs * regs, long error_code) \ +{ \ + if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \ + == NOTIFY_STOP) \ + return; \ + do_trap(trapnr, signr, str, 0, regs, error_code, NULL); \ +} + +#define DO_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \ +fastcall void do_##name(struct pt_regs * regs, long error_code) \ +{ \ + siginfo_t info; \ + info.si_signo = signr; \ + info.si_errno = 0; \ + info.si_code = sicode; \ + info.si_addr = (void __user *)siaddr; \ + if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \ + == NOTIFY_STOP) \ + return; \ + do_trap(trapnr, signr, str, 0, regs, error_code, &info); \ +} + +#define DO_VM86_ERROR(trapnr, signr, str, name) \ +fastcall void do_##name(struct pt_regs * regs, long error_code) \ +{ \ + if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \ + == NOTIFY_STOP) \ + return; \ + do_trap(trapnr, signr, str, 1, regs, error_code, NULL); \ +} + +#define DO_VM86_ERROR_INFO(trapnr, signr, str, name, sicode, siaddr) \ +fastcall void do_##name(struct pt_regs * regs, long error_code) \ +{ \ + siginfo_t info; \ + info.si_signo = signr; \ + info.si_errno = 0; \ + info.si_code = sicode; \ + info.si_addr = (void __user *)siaddr; \ + if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) \ + == NOTIFY_STOP) \ + return; \ + do_trap(trapnr, signr, str, 1, regs, error_code, &info); \ +} + +DO_VM86_ERROR_INFO( 0, SIGFPE, "divide error", divide_error, FPE_INTDIV, regs->eip) +#ifndef CONFIG_KPROBES +DO_VM86_ERROR( 3, SIGTRAP, "int3", int3) +#endif +DO_VM86_ERROR( 4, SIGSEGV, "overflow", overflow) +DO_VM86_ERROR( 5, SIGSEGV, "bounds", bounds) +DO_ERROR_INFO( 6, SIGILL, "invalid operand", invalid_op, ILL_ILLOPN, regs->eip) +DO_ERROR( 9, SIGFPE, "coprocessor segment overrun", coprocessor_segment_overrun) +DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS) +DO_ERROR(11, SIGBUS, "segment not present", segment_not_present) +DO_ERROR(12, SIGBUS, "stack segment", stack_segment) +DO_ERROR_INFO(17, SIGBUS, "alignment check", alignment_check, BUS_ADRALN, 0) + +fastcall void do_general_protection(struct pt_regs * regs, long error_code) +{ + int cpu = get_cpu(); + struct tss_struct *tss = &per_cpu(init_tss, cpu); + struct thread_struct *thread = ¤t->thread; + + /* + * Perform the lazy TSS's I/O bitmap copy. If the TSS has an + * invalid offset set (the LAZY one) and the faulting thread has + * a valid I/O bitmap pointer, we copy the I/O bitmap in the TSS + * and we set the offset field correctly. Then we let the CPU to + * restart the faulting instruction. + */ + if (tss->io_bitmap_base == INVALID_IO_BITMAP_OFFSET_LAZY && + thread->io_bitmap_ptr) { + memcpy(tss->io_bitmap, thread->io_bitmap_ptr, + thread->io_bitmap_max); + /* + * If the previously set map was extending to higher ports + * than the current one, pad extra space with 0xff (no access). + */ + if (thread->io_bitmap_max < tss->io_bitmap_max) + memset((char *) tss->io_bitmap + + thread->io_bitmap_max, 0xff, + tss->io_bitmap_max - thread->io_bitmap_max); + tss->io_bitmap_max = thread->io_bitmap_max; + tss->io_bitmap_base = IO_BITMAP_OFFSET; + put_cpu(); + return; + } + put_cpu(); + + if (regs->eflags & VM_MASK) + goto gp_in_vm86; + + if (!(regs->xcs & 3)) + goto gp_in_kernel; + + current->thread.error_code = error_code; + current->thread.trap_no = 13; + force_sig(SIGSEGV, current); + return; + +gp_in_vm86: + local_irq_enable(); + handle_vm86_fault((struct kernel_vm86_regs *) regs, error_code); + return; + +gp_in_kernel: + if (!fixup_exception(regs)) { + if (notify_die(DIE_GPF, "general protection fault", regs, + error_code, 13, SIGSEGV) == NOTIFY_STOP) + return; + die("general protection fault", regs, error_code); + } +} + +static void mem_parity_error(unsigned char reason, struct pt_regs * regs) +{ + printk("Uhhuh. NMI received. Dazed and confused, but trying to continue\n"); + printk("You probably have a hardware problem with your RAM chips\n"); + + /* Clear and disable the memory parity error line. */ + clear_mem_error(reason); +} + +static void io_check_error(unsigned char reason, struct pt_regs * regs) +{ + unsigned long i; + + printk("NMI: IOCK error (debug interrupt?)\n"); + show_registers(regs); + + /* Re-enable the IOCK line, wait for a few seconds */ + reason = (reason & 0xf) | 8; + outb(reason, 0x61); + i = 2000; + while (--i) udelay(1000); + reason &= ~8; + outb(reason, 0x61); +} + +static void unknown_nmi_error(unsigned char reason, struct pt_regs * regs) +{ +#ifdef CONFIG_MCA + /* Might actually be able to figure out what the guilty party + * is. */ + if( MCA_bus ) { + mca_handle_nmi(); + return; + } +#endif + printk("Uhhuh. NMI received for unknown reason %02x on CPU %d.\n", + reason, smp_processor_id()); + printk("Dazed and confused, but trying to continue\n"); + printk("Do you have a strange power saving mode enabled?\n"); +} + +static DEFINE_SPINLOCK(nmi_print_lock); + +void die_nmi (struct pt_regs *regs, const char *msg) +{ + spin_lock(&nmi_print_lock); + /* + * We are in trouble anyway, lets at least try + * to get a message out. + */ + bust_spinlocks(1); + printk(msg); + printk(" on CPU%d, eip %08lx, registers:\n", + smp_processor_id(), regs->eip); + show_registers(regs); + printk("console shuts up ...\n"); + console_silent(); + spin_unlock(&nmi_print_lock); + bust_spinlocks(0); + do_exit(SIGSEGV); +} + +static void default_do_nmi(struct pt_regs * regs) +{ + unsigned char reason = 0; + + /* Only the BSP gets external NMIs from the system. */ + if (!smp_processor_id()) + reason = get_nmi_reason(); + + if (!(reason & 0xc0)) { + if (notify_die(DIE_NMI_IPI, "nmi_ipi", regs, reason, 0, SIGINT) + == NOTIFY_STOP) + return; +#ifdef CONFIG_X86_LOCAL_APIC + /* + * Ok, so this is none of the documented NMI sources, + * so it must be the NMI watchdog. + */ + if (nmi_watchdog) { + nmi_watchdog_tick(regs); + return; + } +#endif + unknown_nmi_error(reason, regs); + return; + } + if (notify_die(DIE_NMI, "nmi", regs, reason, 0, SIGINT) == NOTIFY_STOP) + return; + if (reason & 0x80) + mem_parity_error(reason, regs); + if (reason & 0x40) + io_check_error(reason, regs); + /* + * Reassert NMI in case it became active meanwhile + * as it's edge-triggered. + */ + reassert_nmi(); +} + +static int dummy_nmi_callback(struct pt_regs * regs, int cpu) +{ + return 0; +} + +static nmi_callback_t nmi_callback = dummy_nmi_callback; + +fastcall void do_nmi(struct pt_regs * regs, long error_code) +{ + int cpu; + + nmi_enter(); + + cpu = smp_processor_id(); + ++nmi_count(cpu); + + if (!nmi_callback(regs, cpu)) + default_do_nmi(regs); + + nmi_exit(); +} + +void set_nmi_callback(nmi_callback_t callback) +{ + nmi_callback = callback; +} + +void unset_nmi_callback(void) +{ + nmi_callback = dummy_nmi_callback; +} + +#ifdef CONFIG_KPROBES +fastcall int do_int3(struct pt_regs *regs, long error_code) +{ + if (notify_die(DIE_INT3, "int3", regs, error_code, 3, SIGTRAP) + == NOTIFY_STOP) + return 1; + /* This is an interrupt gate, because kprobes wants interrupts + disabled. Normal trap handlers don't. */ + restore_interrupts(regs); + do_trap(3, SIGTRAP, "int3", 1, regs, error_code, NULL); + return 0; +} +#endif + +/* + * Our handling of the processor debug registers is non-trivial. + * We do not clear them on entry and exit from the kernel. Therefore + * it is possible to get a watchpoint trap here from inside the kernel. + * However, the code in ./ptrace.c has ensured that the user can + * only set watchpoints on userspace addresses. Therefore the in-kernel + * watchpoint trap can only occur in code which is reading/writing + * from user space. Such code must not hold kernel locks (since it + * can equally take a page fault), therefore it is safe to call + * force_sig_info even though that claims and releases locks. + * + * Code in ./signal.c ensures that the debug control register + * is restored before we deliver any signal, and therefore that + * user code runs with the correct debug control register even though + * we clear it here. + * + * Being careful here means that we don't have to be as careful in a + * lot of more complicated places (task switching can be a bit lazy + * about restoring all the debug state, and ptrace doesn't have to + * find every occurrence of the TF bit that could be saved away even + * by user code) + */ +fastcall void do_debug(struct pt_regs * regs, long error_code) +{ + unsigned int condition; + struct task_struct *tsk = current; + + __asm__ __volatile__("movl %%db6,%0" : "=r" (condition)); + + if (notify_die(DIE_DEBUG, "debug", regs, condition, error_code, + SIGTRAP) == NOTIFY_STOP) + return; + /* It's safe to allow irq's after DR6 has been saved */ + if (regs->eflags & X86_EFLAGS_IF) + local_irq_enable(); + + /* Mask out spurious debug traps due to lazy DR7 setting */ + if (condition & (DR_TRAP0|DR_TRAP1|DR_TRAP2|DR_TRAP3)) { + if (!tsk->thread.debugreg[7]) + goto clear_dr7; + } + + if (regs->eflags & VM_MASK) + goto debug_vm86; + + /* Save debug status register where ptrace can see it */ + tsk->thread.debugreg[6] = condition; + + /* + * Single-stepping through TF: make sure we ignore any events in + * kernel space (but re-enable TF when returning to user mode). + */ + if (condition & DR_STEP) { + /* + * We already checked v86 mode above, so we can + * check for kernel mode by just checking the CPL + * of CS. + */ + if ((regs->xcs & 3) == 0) + goto clear_TF_reenable; + } + + /* Ok, finally something we can handle */ + send_sigtrap(tsk, regs, error_code); + + /* Disable additional traps. They'll be re-enabled when + * the signal is delivered. + */ +clear_dr7: + __asm__("movl %0,%%db7" + : /* no output */ + : "r" (0)); + return; + +debug_vm86: + handle_vm86_trap((struct kernel_vm86_regs *) regs, error_code, 1); + return; + +clear_TF_reenable: + set_tsk_thread_flag(tsk, TIF_SINGLESTEP); + regs->eflags &= ~TF_MASK; + return; +} + +/* + * Note that we play around with the 'TS' bit in an attempt to get + * the correct behaviour even in the presence of the asynchronous + * IRQ13 behaviour + */ +void math_error(void __user *eip) +{ + struct task_struct * task; + siginfo_t info; + unsigned short cwd, swd; + + /* + * Save the info for the exception handler and clear the error. + */ + task = current; + save_init_fpu(task); + task->thread.trap_no = 16; + task->thread.error_code = 0; + info.si_signo = SIGFPE; + info.si_errno = 0; + info.si_code = __SI_FAULT; + info.si_addr = eip; + /* + * (~cwd & swd) will mask out exceptions that are not set to unmasked + * status. 0x3f is the exception bits in these regs, 0x200 is the + * C1 reg you need in case of a stack fault, 0x040 is the stack + * fault bit. We should only be taking one exception at a time, + * so if this combination doesn't produce any single exception, + * then we have a bad program that isn't syncronizing its FPU usage + * and it will suffer the consequences since we won't be able to + * fully reproduce the context of the exception + */ + cwd = get_fpu_cwd(task); + swd = get_fpu_swd(task); + switch (((~cwd) & swd & 0x3f) | (swd & 0x240)) { + case 0x000: + default: + break; + case 0x001: /* Invalid Op */ + case 0x041: /* Stack Fault */ + case 0x241: /* Stack Fault | Direction */ + info.si_code = FPE_FLTINV; + /* Should we clear the SF or let user space do it ???? */ + break; + case 0x002: /* Denormalize */ + case 0x010: /* Underflow */ + info.si_code = FPE_FLTUND; + break; + case 0x004: /* Zero Divide */ + info.si_code = FPE_FLTDIV; + break; + case 0x008: /* Overflow */ + info.si_code = FPE_FLTOVF; + break; + case 0x020: /* Precision */ + info.si_code = FPE_FLTRES; + break; + } + force_sig_info(SIGFPE, &info, task); +} + +fastcall void do_coprocessor_error(struct pt_regs * regs, long error_code) +{ + ignore_fpu_irq = 1; + math_error((void __user *)regs->eip); +} + +static void simd_math_error(void __user *eip) +{ + struct task_struct * task; + siginfo_t info; + unsigned short mxcsr; + + /* + * Save the info for the exception handler and clear the error. + */ + task = current; + save_init_fpu(task); + task->thread.trap_no = 19; + task->thread.error_code = 0; + info.si_signo = SIGFPE; + info.si_errno = 0; + info.si_code = __SI_FAULT; + info.si_addr = eip; + /* + * The SIMD FPU exceptions are handled a little differently, as there + * is only a single status/control register. Thus, to determine which + * unmasked exception was caught we must mask the exception mask bits + * at 0x1f80, and then use these to mask the exception bits at 0x3f. + */ + mxcsr = get_fpu_mxcsr(task); + switch (~((mxcsr & 0x1f80) >> 7) & (mxcsr & 0x3f)) { + case 0x000: + default: + break; + case 0x001: /* Invalid Op */ + info.si_code = FPE_FLTINV; + break; + case 0x002: /* Denormalize */ + case 0x010: /* Underflow */ + info.si_code = FPE_FLTUND; + break; + case 0x004: /* Zero Divide */ + info.si_code = FPE_FLTDIV; + break; + case 0x008: /* Overflow */ + info.si_code = FPE_FLTOVF; + break; + case 0x020: /* Precision */ + info.si_code = FPE_FLTRES; + break; + } + force_sig_info(SIGFPE, &info, task); +} + +fastcall void do_simd_coprocessor_error(struct pt_regs * regs, + long error_code) +{ + if (cpu_has_xmm) { + /* Handle SIMD FPU exceptions on PIII+ processors. */ + ignore_fpu_irq = 1; + simd_math_error((void __user *)regs->eip); + } else { + /* + * Handle strange cache flush from user space exception + * in all other cases. This is undocumented behaviour. + */ + if (regs->eflags & VM_MASK) { + handle_vm86_fault((struct kernel_vm86_regs *)regs, + error_code); + return; + } + die_if_kernel("cache flush denied", regs, error_code); + current->thread.trap_no = 19; + current->thread.error_code = error_code; + force_sig(SIGSEGV, current); + } +} + +fastcall void do_spurious_interrupt_bug(struct pt_regs * regs, + long error_code) +{ +#if 0 + /* No need to warn about this any longer. */ + printk("Ignoring P6 Local APIC Spurious Interrupt Bug...\n"); +#endif +} + +fastcall void setup_x86_bogus_stack(unsigned char * stk) +{ + unsigned long *switch16_ptr, *switch32_ptr; + struct pt_regs *regs; + unsigned long stack_top, stack_bot; + unsigned short iret_frame16_off; + int cpu = smp_processor_id(); + /* reserve the space on 32bit stack for the magic switch16 pointer */ + memmove(stk, stk + 8, sizeof(struct pt_regs)); + switch16_ptr = (unsigned long *)(stk + sizeof(struct pt_regs)); + regs = (struct pt_regs *)stk; + /* now the switch32 on 16bit stack */ + stack_bot = (unsigned long)&per_cpu(cpu_16bit_stack, cpu); + stack_top = stack_bot + CPU_16BIT_STACK_SIZE; + switch32_ptr = (unsigned long *)(stack_top - 8); + iret_frame16_off = CPU_16BIT_STACK_SIZE - 8 - 20; + /* copy iret frame on 16bit stack */ + memcpy((void *)(stack_bot + iret_frame16_off), ®s->eip, 20); + /* fill in the switch pointers */ + switch16_ptr[0] = (regs->esp & 0xffff0000) | iret_frame16_off; + switch16_ptr[1] = __ESPFIX_SS; + switch32_ptr[0] = (unsigned long)stk + sizeof(struct pt_regs) + + 8 - CPU_16BIT_STACK_SIZE; + switch32_ptr[1] = __KERNEL_DS; +} + +fastcall unsigned char * fixup_x86_bogus_stack(unsigned short sp) +{ + unsigned long *switch32_ptr; + unsigned char *stack16, *stack32; + unsigned long stack_top, stack_bot; + int len; + int cpu = smp_processor_id(); + stack_bot = (unsigned long)&per_cpu(cpu_16bit_stack, cpu); + stack_top = stack_bot + CPU_16BIT_STACK_SIZE; + switch32_ptr = (unsigned long *)(stack_top - 8); + /* copy the data from 16bit stack to 32bit stack */ + len = CPU_16BIT_STACK_SIZE - 8 - sp; + stack16 = (unsigned char *)(stack_bot + sp); + stack32 = (unsigned char *) + (switch32_ptr[0] + CPU_16BIT_STACK_SIZE - 8 - len); + memcpy(stack32, stack16, len); + return stack32; +} + +/* + * 'math_state_restore()' saves the current math information in the + * old math state array, and gets the new ones from the current task + * + * Careful.. There are problems with IBM-designed IRQ13 behaviour. + * Don't touch unless you *really* know how it works. + * + * Must be called with kernel preemption disabled (in this case, + * local interrupts are disabled at the call-site in entry.S). + */ +asmlinkage void math_state_restore(struct pt_regs regs) +{ + struct thread_info *thread = current_thread_info(); + struct task_struct *tsk = thread->task; + + clts(); /* Allow maths ops (or we recurse) */ + if (!tsk_used_math(tsk)) + init_fpu(tsk); + restore_fpu(tsk); + thread->status |= TS_USEDFPU; /* So we fnsave on switch_to() */ +} + +#ifndef CONFIG_MATH_EMULATION + +asmlinkage void math_emulate(long arg) +{ + printk("math-emulation not enabled and no coprocessor found.\n"); + printk("killing %s.\n",current->comm); + force_sig(SIGFPE,current); + schedule(); +} + +#endif /* CONFIG_MATH_EMULATION */ + +#ifdef CONFIG_X86_F00F_BUG +void __init trap_init_f00f_bug(void) +{ + __set_fixmap(FIX_F00F_IDT, __pa(&idt_table), PAGE_KERNEL_RO); + + /* + * Update the IDT descriptor and reload the IDT so that + * it uses the read-only mapped virtual address. + */ + idt_descr.address = fix_to_virt(FIX_F00F_IDT); + __asm__ __volatile__("lidt %0" : : "m" (idt_descr)); +} +#endif + +#define _set_gate(gate_addr,type,dpl,addr,seg) \ +do { \ + int __d0, __d1; \ + __asm__ __volatile__ ("movw %%dx,%%ax\n\t" \ + "movw %4,%%dx\n\t" \ + "movl %%eax,%0\n\t" \ + "movl %%edx,%1" \ + :"=m" (*((long *) (gate_addr))), \ + "=m" (*(1+(long *) (gate_addr))), "=&a" (__d0), "=&d" (__d1) \ + :"i" ((short) (0x8000+(dpl<<13)+(type<<8))), \ + "3" ((char *) (addr)),"2" ((seg) << 16)); \ +} while (0) + + +/* + * This needs to use 'idt_table' rather than 'idt', and + * thus use the _nonmapped_ version of the IDT, as the + * Pentium F0 0F bugfix can have resulted in the mapped + * IDT being write-protected. + */ +void set_intr_gate(unsigned int n, void *addr) +{ + _set_gate(idt_table+n,14,0,addr,__KERNEL_CS); +} + +/* + * This routine sets up an interrupt gate at directory privilege level 3. + */ +static inline void set_system_intr_gate(unsigned int n, void *addr) +{ + _set_gate(idt_table+n, 14, 3, addr, __KERNEL_CS); +} + +static void __init set_trap_gate(unsigned int n, void *addr) +{ + _set_gate(idt_table+n,15,0,addr,__KERNEL_CS); +} + +static void __init set_system_gate(unsigned int n, void *addr) +{ + _set_gate(idt_table+n,15,3,addr,__KERNEL_CS); +} + +static void __init set_task_gate(unsigned int n, unsigned int gdt_entry) +{ + _set_gate(idt_table+n,5,0,0,(gdt_entry<<3)); +} + + +void __init trap_init(void) +{ +#ifdef CONFIG_EISA + void __iomem *p = ioremap(0x0FFFD9, 4); + if (readl(p) == 'E'+('I'<<8)+('S'<<16)+('A'<<24)) { + EISA_bus = 1; + } + iounmap(p); +#endif + +#ifdef CONFIG_X86_LOCAL_APIC + init_apic_mappings(); +#endif + + set_trap_gate(0,÷_error); + set_intr_gate(1,&debug); + set_intr_gate(2,&nmi); + set_system_intr_gate(3, &int3); /* int3-5 can be called from all */ + set_system_gate(4,&overflow); + set_system_gate(5,&bounds); + set_trap_gate(6,&invalid_op); + set_trap_gate(7,&device_not_available); + set_task_gate(8,GDT_ENTRY_DOUBLEFAULT_TSS); + set_trap_gate(9,&coprocessor_segment_overrun); + set_trap_gate(10,&invalid_TSS); + set_trap_gate(11,&segment_not_present); + set_trap_gate(12,&stack_segment); + set_trap_gate(13,&general_protection); + set_intr_gate(14,&page_fault); + set_trap_gate(15,&spurious_interrupt_bug); + set_trap_gate(16,&coprocessor_error); + set_trap_gate(17,&alignment_check); +#ifdef CONFIG_X86_MCE + set_trap_gate(18,&machine_check); +#endif + set_trap_gate(19,&simd_coprocessor_error); + + set_system_gate(SYSCALL_VECTOR,&system_call); + + /* + * Should be a barrier for any external CPU state. + */ + cpu_init(); + + trap_init_hook(); +} + +static int __init kstack_setup(char *s) +{ + kstack_depth_to_print = simple_strtoul(s, NULL, 0); + return 0; +} +__setup("kstack=", kstack_setup); diff --git a/arch/i386/kernel/vm86.c b/arch/i386/kernel/vm86.c new file mode 100644 index 00000000000..2f3d52dacff --- /dev/null +++ b/arch/i386/kernel/vm86.c @@ -0,0 +1,804 @@ +/* + * linux/kernel/vm86.c + * + * Copyright (C) 1994 Linus Torvalds + * + * 29 dec 2001 - Fixed oopses caused by unchecked access to the vm86 + * stack - Manfred Spraul <manfreds@colorfullife.com> + * + * 22 mar 2002 - Manfred detected the stackfaults, but didn't handle + * them correctly. Now the emulation will be in a + * consistent state after stackfaults - Kasper Dupont + * <kasperd@daimi.au.dk> + * + * 22 mar 2002 - Added missing clear_IF in set_vflags_* Kasper Dupont + * <kasperd@daimi.au.dk> + * + * ?? ??? 2002 - Fixed premature returns from handle_vm86_fault + * caused by Kasper Dupont's changes - Stas Sergeev + * + * 4 apr 2002 - Fixed CHECK_IF_IN_TRAP broken by Stas' changes. + * Kasper Dupont <kasperd@daimi.au.dk> + * + * 9 apr 2002 - Changed syntax of macros in handle_vm86_fault. + * Kasper Dupont <kasperd@daimi.au.dk> + * + * 9 apr 2002 - Changed stack access macros to jump to a label + * instead of returning to userspace. This simplifies + * do_int, and is needed by handle_vm6_fault. Kasper + * Dupont <kasperd@daimi.au.dk> + * + */ + +#include <linux/config.h> +#include <linux/errno.h> +#include <linux/interrupt.h> +#include <linux/sched.h> +#include <linux/kernel.h> +#include <linux/signal.h> +#include <linux/string.h> +#include <linux/mm.h> +#include <linux/smp.h> +#include <linux/smp_lock.h> +#include <linux/highmem.h> +#include <linux/ptrace.h> + +#include <asm/uaccess.h> +#include <asm/io.h> +#include <asm/tlbflush.h> +#include <asm/irq.h> + +/* + * Known problems: + * + * Interrupt handling is not guaranteed: + * - a real x86 will disable all interrupts for one instruction + * after a "mov ss,xx" to make stack handling atomic even without + * the 'lss' instruction. We can't guarantee this in v86 mode, + * as the next instruction might result in a page fault or similar. + * - a real x86 will have interrupts disabled for one instruction + * past the 'sti' that enables them. We don't bother with all the + * details yet. + * + * Let's hope these problems do not actually matter for anything. + */ + + +#define KVM86 ((struct kernel_vm86_struct *)regs) +#define VMPI KVM86->vm86plus + + +/* + * 8- and 16-bit register defines.. + */ +#define AL(regs) (((unsigned char *)&((regs)->eax))[0]) +#define AH(regs) (((unsigned char *)&((regs)->eax))[1]) +#define IP(regs) (*(unsigned short *)&((regs)->eip)) +#define SP(regs) (*(unsigned short *)&((regs)->esp)) + +/* + * virtual flags (16 and 32-bit versions) + */ +#define VFLAGS (*(unsigned short *)&(current->thread.v86flags)) +#define VEFLAGS (current->thread.v86flags) + +#define set_flags(X,new,mask) \ +((X) = ((X) & ~(mask)) | ((new) & (mask))) + +#define SAFE_MASK (0xDD5) +#define RETURN_MASK (0xDFF) + +#define VM86_REGS_PART2 orig_eax +#define VM86_REGS_SIZE1 \ + ( (unsigned)( & (((struct kernel_vm86_regs *)0)->VM86_REGS_PART2) ) ) +#define VM86_REGS_SIZE2 (sizeof(struct kernel_vm86_regs) - VM86_REGS_SIZE1) + +struct pt_regs * FASTCALL(save_v86_state(struct kernel_vm86_regs * regs)); +struct pt_regs * fastcall save_v86_state(struct kernel_vm86_regs * regs) +{ + struct tss_struct *tss; + struct pt_regs *ret; + unsigned long tmp; + + /* + * This gets called from entry.S with interrupts disabled, but + * from process context. Enable interrupts here, before trying + * to access user space. + */ + local_irq_enable(); + + if (!current->thread.vm86_info) { + printk("no vm86_info: BAD\n"); + do_exit(SIGSEGV); + } + set_flags(regs->eflags, VEFLAGS, VIF_MASK | current->thread.v86mask); + tmp = copy_to_user(¤t->thread.vm86_info->regs,regs, VM86_REGS_SIZE1); + tmp += copy_to_user(¤t->thread.vm86_info->regs.VM86_REGS_PART2, + ®s->VM86_REGS_PART2, VM86_REGS_SIZE2); + tmp += put_user(current->thread.screen_bitmap,¤t->thread.vm86_info->screen_bitmap); + if (tmp) { + printk("vm86: could not access userspace vm86_info\n"); + do_exit(SIGSEGV); + } + + tss = &per_cpu(init_tss, get_cpu()); + current->thread.esp0 = current->thread.saved_esp0; + current->thread.sysenter_cs = __KERNEL_CS; + load_esp0(tss, ¤t->thread); + current->thread.saved_esp0 = 0; + put_cpu(); + + loadsegment(fs, current->thread.saved_fs); + loadsegment(gs, current->thread.saved_gs); + ret = KVM86->regs32; + return ret; +} + +static void mark_screen_rdonly(struct task_struct * tsk) +{ + pgd_t *pgd; + pud_t *pud; + pmd_t *pmd; + pte_t *pte, *mapped; + int i; + + preempt_disable(); + spin_lock(&tsk->mm->page_table_lock); + pgd = pgd_offset(tsk->mm, 0xA0000); + if (pgd_none_or_clear_bad(pgd)) + goto out; + pud = pud_offset(pgd, 0xA0000); + if (pud_none_or_clear_bad(pud)) + goto out; + pmd = pmd_offset(pud, 0xA0000); + if (pmd_none_or_clear_bad(pmd)) + goto out; + pte = mapped = pte_offset_map(pmd, 0xA0000); + for (i = 0; i < 32; i++) { + if (pte_present(*pte)) + set_pte(pte, pte_wrprotect(*pte)); + pte++; + } + pte_unmap(mapped); +out: + spin_unlock(&tsk->mm->page_table_lock); + preempt_enable(); + flush_tlb(); +} + + + +static int do_vm86_irq_handling(int subfunction, int irqnumber); +static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk); + +asmlinkage int sys_vm86old(struct pt_regs regs) +{ + struct vm86_struct __user *v86 = (struct vm86_struct __user *)regs.ebx; + struct kernel_vm86_struct info; /* declare this _on top_, + * this avoids wasting of stack space. + * This remains on the stack until we + * return to 32 bit user space. + */ + struct task_struct *tsk; + int tmp, ret = -EPERM; + + tsk = current; + if (tsk->thread.saved_esp0) + goto out; + tmp = copy_from_user(&info, v86, VM86_REGS_SIZE1); + tmp += copy_from_user(&info.regs.VM86_REGS_PART2, &v86->regs.VM86_REGS_PART2, + (long)&info.vm86plus - (long)&info.regs.VM86_REGS_PART2); + ret = -EFAULT; + if (tmp) + goto out; + memset(&info.vm86plus, 0, (int)&info.regs32 - (int)&info.vm86plus); + info.regs32 = ®s; + tsk->thread.vm86_info = v86; + do_sys_vm86(&info, tsk); + ret = 0; /* we never return here */ +out: + return ret; +} + + +asmlinkage int sys_vm86(struct pt_regs regs) +{ + struct kernel_vm86_struct info; /* declare this _on top_, + * this avoids wasting of stack space. + * This remains on the stack until we + * return to 32 bit user space. + */ + struct task_struct *tsk; + int tmp, ret; + struct vm86plus_struct __user *v86; + + tsk = current; + switch (regs.ebx) { + case VM86_REQUEST_IRQ: + case VM86_FREE_IRQ: + case VM86_GET_IRQ_BITS: + case VM86_GET_AND_RESET_IRQ: + ret = do_vm86_irq_handling(regs.ebx, (int)regs.ecx); + goto out; + case VM86_PLUS_INSTALL_CHECK: + /* NOTE: on old vm86 stuff this will return the error + from verify_area(), because the subfunction is + interpreted as (invalid) address to vm86_struct. + So the installation check works. + */ + ret = 0; + goto out; + } + + /* we come here only for functions VM86_ENTER, VM86_ENTER_NO_BYPASS */ + ret = -EPERM; + if (tsk->thread.saved_esp0) + goto out; + v86 = (struct vm86plus_struct __user *)regs.ecx; + tmp = copy_from_user(&info, v86, VM86_REGS_SIZE1); + tmp += copy_from_user(&info.regs.VM86_REGS_PART2, &v86->regs.VM86_REGS_PART2, + (long)&info.regs32 - (long)&info.regs.VM86_REGS_PART2); + ret = -EFAULT; + if (tmp) + goto out; + info.regs32 = ®s; + info.vm86plus.is_vm86pus = 1; + tsk->thread.vm86_info = (struct vm86_struct __user *)v86; + do_sys_vm86(&info, tsk); + ret = 0; /* we never return here */ +out: + return ret; +} + + +static void do_sys_vm86(struct kernel_vm86_struct *info, struct task_struct *tsk) +{ + struct tss_struct *tss; +/* + * make sure the vm86() system call doesn't try to do anything silly + */ + info->regs.__null_ds = 0; + info->regs.__null_es = 0; + +/* we are clearing fs,gs later just before "jmp resume_userspace", + * because starting with Linux 2.1.x they aren't no longer saved/restored + */ + +/* + * The eflags register is also special: we cannot trust that the user + * has set it up safely, so this makes sure interrupt etc flags are + * inherited from protected mode. + */ + VEFLAGS = info->regs.eflags; + info->regs.eflags &= SAFE_MASK; + info->regs.eflags |= info->regs32->eflags & ~SAFE_MASK; + info->regs.eflags |= VM_MASK; + + switch (info->cpu_type) { + case CPU_286: + tsk->thread.v86mask = 0; + break; + case CPU_386: + tsk->thread.v86mask = NT_MASK | IOPL_MASK; + break; + case CPU_486: + tsk->thread.v86mask = AC_MASK | NT_MASK | IOPL_MASK; + break; + default: + tsk->thread.v86mask = ID_MASK | AC_MASK | NT_MASK | IOPL_MASK; + break; + } + +/* + * Save old state, set default return value (%eax) to 0 + */ + info->regs32->eax = 0; + tsk->thread.saved_esp0 = tsk->thread.esp0; + asm volatile("movl %%fs,%0":"=m" (tsk->thread.saved_fs)); + asm volatile("movl %%gs,%0":"=m" (tsk->thread.saved_gs)); + + tss = &per_cpu(init_tss, get_cpu()); + tsk->thread.esp0 = (unsigned long) &info->VM86_TSS_ESP0; + if (cpu_has_sep) + tsk->thread.sysenter_cs = 0; + load_esp0(tss, &tsk->thread); + put_cpu(); + + tsk->thread.screen_bitmap = info->screen_bitmap; + if (info->flags & VM86_SCREEN_BITMAP) + mark_screen_rdonly(tsk); + __asm__ __volatile__( + "xorl %%eax,%%eax; movl %%eax,%%fs; movl %%eax,%%gs\n\t" + "movl %0,%%esp\n\t" + "movl %1,%%ebp\n\t" + "jmp resume_userspace" + : /* no outputs */ + :"r" (&info->regs), "r" (tsk->thread_info) : "ax"); + /* we never return here */ +} + +static inline void return_to_32bit(struct kernel_vm86_regs * regs16, int retval) +{ + struct pt_regs * regs32; + + regs32 = save_v86_state(regs16); + regs32->eax = retval; + __asm__ __volatile__("movl %0,%%esp\n\t" + "movl %1,%%ebp\n\t" + "jmp resume_userspace" + : : "r" (regs32), "r" (current_thread_info())); +} + +static inline void set_IF(struct kernel_vm86_regs * regs) +{ + VEFLAGS |= VIF_MASK; + if (VEFLAGS & VIP_MASK) + return_to_32bit(regs, VM86_STI); +} + +static inline void clear_IF(struct kernel_vm86_regs * regs) +{ + VEFLAGS &= ~VIF_MASK; +} + +static inline void clear_TF(struct kernel_vm86_regs * regs) +{ + regs->eflags &= ~TF_MASK; +} + +static inline void clear_AC(struct kernel_vm86_regs * regs) +{ + regs->eflags &= ~AC_MASK; +} + +/* It is correct to call set_IF(regs) from the set_vflags_* + * functions. However someone forgot to call clear_IF(regs) + * in the opposite case. + * After the command sequence CLI PUSHF STI POPF you should + * end up with interrups disabled, but you ended up with + * interrupts enabled. + * ( I was testing my own changes, but the only bug I + * could find was in a function I had not changed. ) + * [KD] + */ + +static inline void set_vflags_long(unsigned long eflags, struct kernel_vm86_regs * regs) +{ + set_flags(VEFLAGS, eflags, current->thread.v86mask); + set_flags(regs->eflags, eflags, SAFE_MASK); + if (eflags & IF_MASK) + set_IF(regs); + else + clear_IF(regs); +} + +static inline void set_vflags_short(unsigned short flags, struct kernel_vm86_regs * regs) +{ + set_flags(VFLAGS, flags, current->thread.v86mask); + set_flags(regs->eflags, flags, SAFE_MASK); + if (flags & IF_MASK) + set_IF(regs); + else + clear_IF(regs); +} + +static inline unsigned long get_vflags(struct kernel_vm86_regs * regs) +{ + unsigned long flags = regs->eflags & RETURN_MASK; + + if (VEFLAGS & VIF_MASK) + flags |= IF_MASK; + flags |= IOPL_MASK; + return flags | (VEFLAGS & current->thread.v86mask); +} + +static inline int is_revectored(int nr, struct revectored_struct * bitmap) +{ + __asm__ __volatile__("btl %2,%1\n\tsbbl %0,%0" + :"=r" (nr) + :"m" (*bitmap),"r" (nr)); + return nr; +} + +#define val_byte(val, n) (((__u8 *)&val)[n]) + +#define pushb(base, ptr, val, err_label) \ + do { \ + __u8 __val = val; \ + ptr--; \ + if (put_user(__val, base + ptr) < 0) \ + goto err_label; \ + } while(0) + +#define pushw(base, ptr, val, err_label) \ + do { \ + __u16 __val = val; \ + ptr--; \ + if (put_user(val_byte(__val, 1), base + ptr) < 0) \ + goto err_label; \ + ptr--; \ + if (put_user(val_byte(__val, 0), base + ptr) < 0) \ + goto err_label; \ + } while(0) + +#define pushl(base, ptr, val, err_label) \ + do { \ + __u32 __val = val; \ + ptr--; \ + if (put_user(val_byte(__val, 3), base + ptr) < 0) \ + goto err_label; \ + ptr--; \ + if (put_user(val_byte(__val, 2), base + ptr) < 0) \ + goto err_label; \ + ptr--; \ + if (put_user(val_byte(__val, 1), base + ptr) < 0) \ + goto err_label; \ + ptr--; \ + if (put_user(val_byte(__val, 0), base + ptr) < 0) \ + goto err_label; \ + } while(0) + +#define popb(base, ptr, err_label) \ + ({ \ + __u8 __res; \ + if (get_user(__res, base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + __res; \ + }) + +#define popw(base, ptr, err_label) \ + ({ \ + __u16 __res; \ + if (get_user(val_byte(__res, 0), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + if (get_user(val_byte(__res, 1), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + __res; \ + }) + +#define popl(base, ptr, err_label) \ + ({ \ + __u32 __res; \ + if (get_user(val_byte(__res, 0), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + if (get_user(val_byte(__res, 1), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + if (get_user(val_byte(__res, 2), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + if (get_user(val_byte(__res, 3), base + ptr) < 0) \ + goto err_label; \ + ptr++; \ + __res; \ + }) + +/* There are so many possible reasons for this function to return + * VM86_INTx, so adding another doesn't bother me. We can expect + * userspace programs to be able to handle it. (Getting a problem + * in userspace is always better than an Oops anyway.) [KD] + */ +static void do_int(struct kernel_vm86_regs *regs, int i, + unsigned char __user * ssp, unsigned short sp) +{ + unsigned long __user *intr_ptr; + unsigned long segoffs; + + if (regs->cs == BIOSSEG) + goto cannot_handle; + if (is_revectored(i, &KVM86->int_revectored)) + goto cannot_handle; + if (i==0x21 && is_revectored(AH(regs),&KVM86->int21_revectored)) + goto cannot_handle; + intr_ptr = (unsigned long __user *) (i << 2); + if (get_user(segoffs, intr_ptr)) + goto cannot_handle; + if ((segoffs >> 16) == BIOSSEG) + goto cannot_handle; + pushw(ssp, sp, get_vflags(regs), cannot_handle); + pushw(ssp, sp, regs->cs, cannot_handle); + pushw(ssp, sp, IP(regs), cannot_handle); + regs->cs = segoffs >> 16; + SP(regs) -= 6; + IP(regs) = segoffs & 0xffff; + clear_TF(regs); + clear_IF(regs); + clear_AC(regs); + return; + +cannot_handle: + return_to_32bit(regs, VM86_INTx + (i << 8)); +} + +int handle_vm86_trap(struct kernel_vm86_regs * regs, long error_code, int trapno) +{ + if (VMPI.is_vm86pus) { + if ( (trapno==3) || (trapno==1) ) + return_to_32bit(regs, VM86_TRAP + (trapno << 8)); + do_int(regs, trapno, (unsigned char __user *) (regs->ss << 4), SP(regs)); + return 0; + } + if (trapno !=1) + return 1; /* we let this handle by the calling routine */ + if (current->ptrace & PT_PTRACED) { + unsigned long flags; + spin_lock_irqsave(¤t->sighand->siglock, flags); + sigdelset(¤t->blocked, SIGTRAP); + recalc_sigpending(); + spin_unlock_irqrestore(¤t->sighand->siglock, flags); + } + send_sig(SIGTRAP, current, 1); + current->thread.trap_no = trapno; + current->thread.error_code = error_code; + return 0; +} + +void handle_vm86_fault(struct kernel_vm86_regs * regs, long error_code) +{ + unsigned char opcode; + unsigned char __user *csp; + unsigned char __user *ssp; + unsigned short ip, sp; + int data32, pref_done; + +#define CHECK_IF_IN_TRAP \ + if (VMPI.vm86dbg_active && VMPI.vm86dbg_TFpendig) \ + newflags |= TF_MASK +#define VM86_FAULT_RETURN do { \ + if (VMPI.force_return_for_pic && (VEFLAGS & (IF_MASK | VIF_MASK))) \ + return_to_32bit(regs, VM86_PICRETURN); \ + return; } while (0) + + csp = (unsigned char __user *) (regs->cs << 4); + ssp = (unsigned char __user *) (regs->ss << 4); + sp = SP(regs); + ip = IP(regs); + + data32 = 0; + pref_done = 0; + do { + switch (opcode = popb(csp, ip, simulate_sigsegv)) { + case 0x66: /* 32-bit data */ data32=1; break; + case 0x67: /* 32-bit address */ break; + case 0x2e: /* CS */ break; + case 0x3e: /* DS */ break; + case 0x26: /* ES */ break; + case 0x36: /* SS */ break; + case 0x65: /* GS */ break; + case 0x64: /* FS */ break; + case 0xf2: /* repnz */ break; + case 0xf3: /* rep */ break; + default: pref_done = 1; + } + } while (!pref_done); + + switch (opcode) { + + /* pushf */ + case 0x9c: + if (data32) { + pushl(ssp, sp, get_vflags(regs), simulate_sigsegv); + SP(regs) -= 4; + } else { + pushw(ssp, sp, get_vflags(regs), simulate_sigsegv); + SP(regs) -= 2; + } + IP(regs) = ip; + VM86_FAULT_RETURN; + + /* popf */ + case 0x9d: + { + unsigned long newflags; + if (data32) { + newflags=popl(ssp, sp, simulate_sigsegv); + SP(regs) += 4; + } else { + newflags = popw(ssp, sp, simulate_sigsegv); + SP(regs) += 2; + } + IP(regs) = ip; + CHECK_IF_IN_TRAP; + if (data32) { + set_vflags_long(newflags, regs); + } else { + set_vflags_short(newflags, regs); + } + VM86_FAULT_RETURN; + } + + /* int xx */ + case 0xcd: { + int intno=popb(csp, ip, simulate_sigsegv); + IP(regs) = ip; + if (VMPI.vm86dbg_active) { + if ( (1 << (intno &7)) & VMPI.vm86dbg_intxxtab[intno >> 3] ) + return_to_32bit(regs, VM86_INTx + (intno << 8)); + } + do_int(regs, intno, ssp, sp); + return; + } + + /* iret */ + case 0xcf: + { + unsigned long newip; + unsigned long newcs; + unsigned long newflags; + if (data32) { + newip=popl(ssp, sp, simulate_sigsegv); + newcs=popl(ssp, sp, simulate_sigsegv); + newflags=popl(ssp, sp, simulate_sigsegv); + SP(regs) += 12; + } else { + newip = popw(ssp, sp, simulate_sigsegv); + newcs = popw(ssp, sp, simulate_sigsegv); + newflags = popw(ssp, sp, simulate_sigsegv); + SP(regs) += 6; + } + IP(regs) = newip; + regs->cs = newcs; + CHECK_IF_IN_TRAP; + if (data32) { + set_vflags_long(newflags, regs); + } else { + set_vflags_short(newflags, regs); + } + VM86_FAULT_RETURN; + } + + /* cli */ + case 0xfa: + IP(regs) = ip; + clear_IF(regs); + VM86_FAULT_RETURN; + + /* sti */ + /* + * Damn. This is incorrect: the 'sti' instruction should actually + * enable interrupts after the /next/ instruction. Not good. + * + * Probably needs some horsing around with the TF flag. Aiee.. + */ + case 0xfb: + IP(regs) = ip; + set_IF(regs); + VM86_FAULT_RETURN; + + default: + return_to_32bit(regs, VM86_UNKNOWN); + } + + return; + +simulate_sigsegv: + /* FIXME: After a long discussion with Stas we finally + * agreed, that this is wrong. Here we should + * really send a SIGSEGV to the user program. + * But how do we create the correct context? We + * are inside a general protection fault handler + * and has just returned from a page fault handler. + * The correct context for the signal handler + * should be a mixture of the two, but how do we + * get the information? [KD] + */ + return_to_32bit(regs, VM86_UNKNOWN); +} + +/* ---------------- vm86 special IRQ passing stuff ----------------- */ + +#define VM86_IRQNAME "vm86irq" + +static struct vm86_irqs { + struct task_struct *tsk; + int sig; +} vm86_irqs[16]; + +static DEFINE_SPINLOCK(irqbits_lock); +static int irqbits; + +#define ALLOWED_SIGS ( 1 /* 0 = don't send a signal */ \ + | (1 << SIGUSR1) | (1 << SIGUSR2) | (1 << SIGIO) | (1 << SIGURG) \ + | (1 << SIGUNUSED) ) + +static irqreturn_t irq_handler(int intno, void *dev_id, struct pt_regs * regs) +{ + int irq_bit; + unsigned long flags; + + spin_lock_irqsave(&irqbits_lock, flags); + irq_bit = 1 << intno; + if ((irqbits & irq_bit) || ! vm86_irqs[intno].tsk) + goto out; + irqbits |= irq_bit; + if (vm86_irqs[intno].sig) + send_sig(vm86_irqs[intno].sig, vm86_irqs[intno].tsk, 1); + spin_unlock_irqrestore(&irqbits_lock, flags); + /* + * IRQ will be re-enabled when user asks for the irq (whether + * polling or as a result of the signal) + */ + disable_irq(intno); + return IRQ_HANDLED; + +out: + spin_unlock_irqrestore(&irqbits_lock, flags); + return IRQ_NONE; +} + +static inline void free_vm86_irq(int irqnumber) +{ + unsigned long flags; + + free_irq(irqnumber, NULL); + vm86_irqs[irqnumber].tsk = NULL; + + spin_lock_irqsave(&irqbits_lock, flags); + irqbits &= ~(1 << irqnumber); + spin_unlock_irqrestore(&irqbits_lock, flags); +} + +void release_vm86_irqs(struct task_struct *task) +{ + int i; + for (i = FIRST_VM86_IRQ ; i <= LAST_VM86_IRQ; i++) + if (vm86_irqs[i].tsk == task) + free_vm86_irq(i); +} + +static inline int get_and_reset_irq(int irqnumber) +{ + int bit; + unsigned long flags; + + if (invalid_vm86_irq(irqnumber)) return 0; + if (vm86_irqs[irqnumber].tsk != current) return 0; + spin_lock_irqsave(&irqbits_lock, flags); + bit = irqbits & (1 << irqnumber); + irqbits &= ~bit; + spin_unlock_irqrestore(&irqbits_lock, flags); + if (!bit) + return 0; + enable_irq(irqnumber); + return 1; +} + + +static int do_vm86_irq_handling(int subfunction, int irqnumber) +{ + int ret; + switch (subfunction) { + case VM86_GET_AND_RESET_IRQ: { + return get_and_reset_irq(irqnumber); + } + case VM86_GET_IRQ_BITS: { + return irqbits; + } + case VM86_REQUEST_IRQ: { + int sig = irqnumber >> 8; + int irq = irqnumber & 255; + if (!capable(CAP_SYS_ADMIN)) return -EPERM; + if (!((1 << sig) & ALLOWED_SIGS)) return -EPERM; + if (invalid_vm86_irq(irq)) return -EPERM; + if (vm86_irqs[irq].tsk) return -EPERM; + ret = request_irq(irq, &irq_handler, 0, VM86_IRQNAME, NULL); + if (ret) return ret; + vm86_irqs[irq].sig = sig; + vm86_irqs[irq].tsk = current; + return irq; + } + case VM86_FREE_IRQ: { + if (invalid_vm86_irq(irqnumber)) return -EPERM; + if (!vm86_irqs[irqnumber].tsk) return 0; + if (vm86_irqs[irqnumber].tsk != current) return -EPERM; + free_vm86_irq(irqnumber); + return 0; + } + } + return -EINVAL; +} + diff --git a/arch/i386/kernel/vmlinux.lds.S b/arch/i386/kernel/vmlinux.lds.S new file mode 100644 index 00000000000..e0512cc8bea --- /dev/null +++ b/arch/i386/kernel/vmlinux.lds.S @@ -0,0 +1,134 @@ +/* ld script to make i386 Linux kernel + * Written by Martin Mares <mj@atrey.karlin.mff.cuni.cz>; + */ + +#include <asm-generic/vmlinux.lds.h> +#include <asm/thread_info.h> +#include <asm/page.h> + +OUTPUT_FORMAT("elf32-i386", "elf32-i386", "elf32-i386") +OUTPUT_ARCH(i386) +ENTRY(startup_32) +jiffies = jiffies_64; +SECTIONS +{ + . = __PAGE_OFFSET + 0x100000; + /* read-only */ + _text = .; /* Text and read-only data */ + .text : { + *(.text) + SCHED_TEXT + LOCK_TEXT + *(.fixup) + *(.gnu.warning) + } = 0x9090 + + _etext = .; /* End of text section */ + + . = ALIGN(16); /* Exception table */ + __start___ex_table = .; + __ex_table : { *(__ex_table) } + __stop___ex_table = .; + + RODATA + + /* writeable */ + .data : { /* Data */ + *(.data) + CONSTRUCTORS + } + + . = ALIGN(4096); + __nosave_begin = .; + .data_nosave : { *(.data.nosave) } + . = ALIGN(4096); + __nosave_end = .; + + . = ALIGN(4096); + .data.page_aligned : { *(.data.idt) } + + . = ALIGN(32); + .data.cacheline_aligned : { *(.data.cacheline_aligned) } + + _edata = .; /* End of data section */ + + . = ALIGN(THREAD_SIZE); /* init_task */ + .data.init_task : { *(.data.init_task) } + + /* will be freed after init */ + . = ALIGN(4096); /* Init code and data */ + __init_begin = .; + .init.text : { + _sinittext = .; + *(.init.text) + _einittext = .; + } + .init.data : { *(.init.data) } + . = ALIGN(16); + __setup_start = .; + .init.setup : { *(.init.setup) } + __setup_end = .; + __initcall_start = .; + .initcall.init : { + *(.initcall1.init) + *(.initcall2.init) + *(.initcall3.init) + *(.initcall4.init) + *(.initcall5.init) + *(.initcall6.init) + *(.initcall7.init) + } + __initcall_end = .; + __con_initcall_start = .; + .con_initcall.init : { *(.con_initcall.init) } + __con_initcall_end = .; + SECURITY_INIT + . = ALIGN(4); + __alt_instructions = .; + .altinstructions : { *(.altinstructions) } + __alt_instructions_end = .; + .altinstr_replacement : { *(.altinstr_replacement) } + /* .exit.text is discard at runtime, not link time, to deal with references + from .altinstructions and .eh_frame */ + .exit.text : { *(.exit.text) } + .exit.data : { *(.exit.data) } + . = ALIGN(4096); + __initramfs_start = .; + .init.ramfs : { *(.init.ramfs) } + __initramfs_end = .; + . = ALIGN(32); + __per_cpu_start = .; + .data.percpu : { *(.data.percpu) } + __per_cpu_end = .; + . = ALIGN(4096); + __init_end = .; + /* freed after init ends here */ + + __bss_start = .; /* BSS */ + .bss : { + *(.bss.page_aligned) + *(.bss) + } + . = ALIGN(4); + __bss_stop = .; + + _end = . ; + + /* This is where the kernel creates the early boot page tables */ + . = ALIGN(4096); + pg0 = .; + + /* Sections to be discarded */ + /DISCARD/ : { + *(.exitcall.exit) + } + + /* Stabs debugging sections. */ + .stab 0 : { *(.stab) } + .stabstr 0 : { *(.stabstr) } + .stab.excl 0 : { *(.stab.excl) } + .stab.exclstr 0 : { *(.stab.exclstr) } + .stab.index 0 : { *(.stab.index) } + .stab.indexstr 0 : { *(.stab.indexstr) } + .comment 0 : { *(.comment) } +} diff --git a/arch/i386/kernel/vsyscall-int80.S b/arch/i386/kernel/vsyscall-int80.S new file mode 100644 index 00000000000..530d0525e5e --- /dev/null +++ b/arch/i386/kernel/vsyscall-int80.S @@ -0,0 +1,53 @@ +/* + * Code for the vsyscall page. This version uses the old int $0x80 method. + * + * NOTE: + * 1) __kernel_vsyscall _must_ be first in this page. + * 2) there are alignment constraints on this stub, see vsyscall-sigreturn.S + * for details. + */ + + .text + .globl __kernel_vsyscall + .type __kernel_vsyscall,@function +__kernel_vsyscall: +.LSTART_vsyscall: + int $0x80 + ret +.LEND_vsyscall: + .size __kernel_vsyscall,.-.LSTART_vsyscall + .previous + + .section .eh_frame,"a",@progbits +.LSTARTFRAMEDLSI: + .long .LENDCIEDLSI-.LSTARTCIEDLSI +.LSTARTCIEDLSI: + .long 0 /* CIE ID */ + .byte 1 /* Version number */ + .string "zR" /* NUL-terminated augmentation string */ + .uleb128 1 /* Code alignment factor */ + .sleb128 -4 /* Data alignment factor */ + .byte 8 /* Return address register column */ + .uleb128 1 /* Augmentation value length */ + .byte 0x1b /* DW_EH_PE_pcrel|DW_EH_PE_sdata4. */ + .byte 0x0c /* DW_CFA_def_cfa */ + .uleb128 4 + .uleb128 4 + .byte 0x88 /* DW_CFA_offset, column 0x8 */ + .uleb128 1 + .align 4 +.LENDCIEDLSI: + .long .LENDFDEDLSI-.LSTARTFDEDLSI /* Length FDE */ +.LSTARTFDEDLSI: + .long .LSTARTFDEDLSI-.LSTARTFRAMEDLSI /* CIE pointer */ + .long .LSTART_vsyscall-. /* PC-relative start address */ + .long .LEND_vsyscall-.LSTART_vsyscall + .uleb128 0 + .align 4 +.LENDFDEDLSI: + .previous + +/* + * Get the common code for the sigreturn entry points. + */ +#include "vsyscall-sigreturn.S" diff --git a/arch/i386/kernel/vsyscall-sigreturn.S b/arch/i386/kernel/vsyscall-sigreturn.S new file mode 100644 index 00000000000..c8fcf75b9be --- /dev/null +++ b/arch/i386/kernel/vsyscall-sigreturn.S @@ -0,0 +1,142 @@ +/* + * Common code for the sigreturn entry points on the vsyscall page. + * So far this code is the same for both int80 and sysenter versions. + * This file is #include'd by vsyscall-*.S to define them after the + * vsyscall entry point. The kernel assumes that the addresses of these + * routines are constant for all vsyscall implementations. + */ + +#include <asm/unistd.h> +#include <asm/asm_offsets.h> + + +/* XXX + Should these be named "_sigtramp" or something? +*/ + + .text + .org __kernel_vsyscall+32 + .globl __kernel_sigreturn + .type __kernel_sigreturn,@function +__kernel_sigreturn: +.LSTART_sigreturn: + popl %eax /* XXX does this mean it needs unwind info? */ + movl $__NR_sigreturn, %eax + int $0x80 +.LEND_sigreturn: + .size __kernel_sigreturn,.-.LSTART_sigreturn + + .balign 32 + .globl __kernel_rt_sigreturn + .type __kernel_rt_sigreturn,@function +__kernel_rt_sigreturn: +.LSTART_rt_sigreturn: + movl $__NR_rt_sigreturn, %eax + int $0x80 +.LEND_rt_sigreturn: + .size __kernel_rt_sigreturn,.-.LSTART_rt_sigreturn + .previous + + .section .eh_frame,"a",@progbits +.LSTARTFRAMEDLSI1: + .long .LENDCIEDLSI1-.LSTARTCIEDLSI1 +.LSTARTCIEDLSI1: + .long 0 /* CIE ID */ + .byte 1 /* Version number */ + .string "zR" /* NUL-terminated augmentation string */ + .uleb128 1 /* Code alignment factor */ + .sleb128 -4 /* Data alignment factor */ + .byte 8 /* Return address register column */ + .uleb128 1 /* Augmentation value length */ + .byte 0x1b /* DW_EH_PE_pcrel|DW_EH_PE_sdata4. */ + .byte 0 /* DW_CFA_nop */ + .align 4 +.LENDCIEDLSI1: + .long .LENDFDEDLSI1-.LSTARTFDEDLSI1 /* Length FDE */ +.LSTARTFDEDLSI1: + .long .LSTARTFDEDLSI1-.LSTARTFRAMEDLSI1 /* CIE pointer */ + /* HACK: The dwarf2 unwind routines will subtract 1 from the + return address to get an address in the middle of the + presumed call instruction. Since we didn't get here via + a call, we need to include the nop before the real start + to make up for it. */ + .long .LSTART_sigreturn-1-. /* PC-relative start address */ + .long .LEND_sigreturn-.LSTART_sigreturn+1 + .uleb128 0 /* Augmentation */ + /* What follows are the instructions for the table generation. + We record the locations of each register saved. This is + complicated by the fact that the "CFA" is always assumed to + be the value of the stack pointer in the caller. This means + that we must define the CFA of this body of code to be the + saved value of the stack pointer in the sigcontext. Which + also means that there is no fixed relation to the other + saved registers, which means that we must use DW_CFA_expression + to compute their addresses. It also means that when we + adjust the stack with the popl, we have to do it all over again. */ + +#define do_cfa_expr(offset) \ + .byte 0x0f; /* DW_CFA_def_cfa_expression */ \ + .uleb128 1f-0f; /* length */ \ +0: .byte 0x74; /* DW_OP_breg4 */ \ + .sleb128 offset; /* offset */ \ + .byte 0x06; /* DW_OP_deref */ \ +1: + +#define do_expr(regno, offset) \ + .byte 0x10; /* DW_CFA_expression */ \ + .uleb128 regno; /* regno */ \ + .uleb128 1f-0f; /* length */ \ +0: .byte 0x74; /* DW_OP_breg4 */ \ + .sleb128 offset; /* offset */ \ +1: + + do_cfa_expr(SIGCONTEXT_esp+4) + do_expr(0, SIGCONTEXT_eax+4) + do_expr(1, SIGCONTEXT_ecx+4) + do_expr(2, SIGCONTEXT_edx+4) + do_expr(3, SIGCONTEXT_ebx+4) + do_expr(5, SIGCONTEXT_ebp+4) + do_expr(6, SIGCONTEXT_esi+4) + do_expr(7, SIGCONTEXT_edi+4) + do_expr(8, SIGCONTEXT_eip+4) + + .byte 0x42 /* DW_CFA_advance_loc 2 -- nop; popl eax. */ + + do_cfa_expr(SIGCONTEXT_esp) + do_expr(0, SIGCONTEXT_eax) + do_expr(1, SIGCONTEXT_ecx) + do_expr(2, SIGCONTEXT_edx) + do_expr(3, SIGCONTEXT_ebx) + do_expr(5, SIGCONTEXT_ebp) + do_expr(6, SIGCONTEXT_esi) + do_expr(7, SIGCONTEXT_edi) + do_expr(8, SIGCONTEXT_eip) + + .align 4 +.LENDFDEDLSI1: + + .long .LENDFDEDLSI2-.LSTARTFDEDLSI2 /* Length FDE */ +.LSTARTFDEDLSI2: + .long .LSTARTFDEDLSI2-.LSTARTFRAMEDLSI1 /* CIE pointer */ + /* HACK: See above wrt unwind library assumptions. */ + .long .LSTART_rt_sigreturn-1-. /* PC-relative start address */ + .long .LEND_rt_sigreturn-.LSTART_rt_sigreturn+1 + .uleb128 0 /* Augmentation */ + /* What follows are the instructions for the table generation. + We record the locations of each register saved. This is + slightly less complicated than the above, since we don't + modify the stack pointer in the process. */ + + do_cfa_expr(RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_esp) + do_expr(0, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_eax) + do_expr(1, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_ecx) + do_expr(2, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_edx) + do_expr(3, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_ebx) + do_expr(5, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_ebp) + do_expr(6, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_esi) + do_expr(7, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_edi) + do_expr(8, RT_SIGFRAME_sigcontext-4 + SIGCONTEXT_eip) + + .align 4 +.LENDFDEDLSI2: + .previous diff --git a/arch/i386/kernel/vsyscall-sysenter.S b/arch/i386/kernel/vsyscall-sysenter.S new file mode 100644 index 00000000000..4daefb2ec1b --- /dev/null +++ b/arch/i386/kernel/vsyscall-sysenter.S @@ -0,0 +1,104 @@ +/* + * Code for the vsyscall page. This version uses the sysenter instruction. + * + * NOTE: + * 1) __kernel_vsyscall _must_ be first in this page. + * 2) there are alignment constraints on this stub, see vsyscall-sigreturn.S + * for details. + */ + + .text + .globl __kernel_vsyscall + .type __kernel_vsyscall,@function +__kernel_vsyscall: +.LSTART_vsyscall: + push %ecx +.Lpush_ecx: + push %edx +.Lpush_edx: + push %ebp +.Lenter_kernel: + movl %esp,%ebp + sysenter + + /* 7: align return point with nop's to make disassembly easier */ + .space 7,0x90 + + /* 14: System call restart point is here! (SYSENTER_RETURN - 2) */ + jmp .Lenter_kernel + /* 16: System call normal return point is here! */ + .globl SYSENTER_RETURN /* Symbol used by entry.S. */ +SYSENTER_RETURN: + pop %ebp +.Lpop_ebp: + pop %edx +.Lpop_edx: + pop %ecx +.Lpop_ecx: + ret +.LEND_vsyscall: + .size __kernel_vsyscall,.-.LSTART_vsyscall + .previous + + .section .eh_frame,"a",@progbits +.LSTARTFRAMEDLSI: + .long .LENDCIEDLSI-.LSTARTCIEDLSI +.LSTARTCIEDLSI: + .long 0 /* CIE ID */ + .byte 1 /* Version number */ + .string "zR" /* NUL-terminated augmentation string */ + .uleb128 1 /* Code alignment factor */ + .sleb128 -4 /* Data alignment factor */ + .byte 8 /* Return address register column */ + .uleb128 1 /* Augmentation value length */ + .byte 0x1b /* DW_EH_PE_pcrel|DW_EH_PE_sdata4. */ + .byte 0x0c /* DW_CFA_def_cfa */ + .uleb128 4 + .uleb128 4 + .byte 0x88 /* DW_CFA_offset, column 0x8 */ + .uleb128 1 + .align 4 +.LENDCIEDLSI: + .long .LENDFDEDLSI-.LSTARTFDEDLSI /* Length FDE */ +.LSTARTFDEDLSI: + .long .LSTARTFDEDLSI-.LSTARTFRAMEDLSI /* CIE pointer */ + .long .LSTART_vsyscall-. /* PC-relative start address */ + .long .LEND_vsyscall-.LSTART_vsyscall + .uleb128 0 + /* What follows are the instructions for the table generation. + We have to record all changes of the stack pointer. */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lpush_ecx-.LSTART_vsyscall + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x08 /* RA at offset 8 now */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lpush_edx-.Lpush_ecx + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x0c /* RA at offset 12 now */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lenter_kernel-.Lpush_edx + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x10 /* RA at offset 16 now */ + .byte 0x85, 0x04 /* DW_CFA_offset %ebp -16 */ + /* Finally the epilogue. */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lpop_ebp-.Lenter_kernel + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x0c /* RA at offset 12 now */ + .byte 0xc5 /* DW_CFA_restore %ebp */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lpop_edx-.Lpop_ebp + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x08 /* RA at offset 8 now */ + .byte 0x04 /* DW_CFA_advance_loc4 */ + .long .Lpop_ecx-.Lpop_edx + .byte 0x0e /* DW_CFA_def_cfa_offset */ + .byte 0x04 /* RA at offset 4 now */ + .align 4 +.LENDFDEDLSI: + .previous + +/* + * Get the common code for the sigreturn entry points. + */ +#include "vsyscall-sigreturn.S" diff --git a/arch/i386/kernel/vsyscall.S b/arch/i386/kernel/vsyscall.S new file mode 100644 index 00000000000..b403890fe39 --- /dev/null +++ b/arch/i386/kernel/vsyscall.S @@ -0,0 +1,15 @@ +#include <linux/init.h> + +__INITDATA + + .globl vsyscall_int80_start, vsyscall_int80_end +vsyscall_int80_start: + .incbin "arch/i386/kernel/vsyscall-int80.so" +vsyscall_int80_end: + + .globl vsyscall_sysenter_start, vsyscall_sysenter_end +vsyscall_sysenter_start: + .incbin "arch/i386/kernel/vsyscall-sysenter.so" +vsyscall_sysenter_end: + +__FINIT diff --git a/arch/i386/kernel/vsyscall.lds.S b/arch/i386/kernel/vsyscall.lds.S new file mode 100644 index 00000000000..3a8329d6536 --- /dev/null +++ b/arch/i386/kernel/vsyscall.lds.S @@ -0,0 +1,65 @@ +/* + * Linker script for vsyscall DSO. The vsyscall page is an ELF shared + * object prelinked to its virtual address, and with only one read-only + * segment (that fits in one page). This script controls its layout. + */ +#include <asm/asm_offsets.h> + +SECTIONS +{ + . = VSYSCALL_BASE + SIZEOF_HEADERS; + + .hash : { *(.hash) } :text + .dynsym : { *(.dynsym) } + .dynstr : { *(.dynstr) } + .gnu.version : { *(.gnu.version) } + .gnu.version_d : { *(.gnu.version_d) } + .gnu.version_r : { *(.gnu.version_r) } + + /* This linker script is used both with -r and with -shared. + For the layouts to match, we need to skip more than enough + space for the dynamic symbol table et al. If this amount + is insufficient, ld -shared will barf. Just increase it here. */ + . = VSYSCALL_BASE + 0x400; + + .text : { *(.text) } :text =0x90909090 + + .eh_frame_hdr : { *(.eh_frame_hdr) } :text :eh_frame_hdr + .eh_frame : { KEEP (*(.eh_frame)) } :text + .dynamic : { *(.dynamic) } :text :dynamic + .useless : { + *(.got.plt) *(.got) + *(.data .data.* .gnu.linkonce.d.*) + *(.dynbss) + *(.bss .bss.* .gnu.linkonce.b.*) + } :text +} + +/* + * We must supply the ELF program headers explicitly to get just one + * PT_LOAD segment, and set the flags explicitly to make segments read-only. + */ +PHDRS +{ + text PT_LOAD FILEHDR PHDRS FLAGS(5); /* PF_R|PF_X */ + dynamic PT_DYNAMIC FLAGS(4); /* PF_R */ + eh_frame_hdr 0x6474e550; /* PT_GNU_EH_FRAME, but ld doesn't match the name */ +} + +/* + * This controls what symbols we export from the DSO. + */ +VERSION +{ + LINUX_2.5 { + global: + __kernel_vsyscall; + __kernel_sigreturn; + __kernel_rt_sigreturn; + + local: *; + }; +} + +/* The ELF entry point can be used to set the AT_SYSINFO value. */ +ENTRY(__kernel_vsyscall); |