From e85abf8f432bb2a13733ab7609fbb8e1500af51d Mon Sep 17 00:00:00 2001 From: Gary Hade Date: Wed, 8 Apr 2009 14:07:25 -0700 Subject: x86: consolidate SMP code in io_apic.c Impact: Cleanup Reorganizes the code in arch/x86/kernel/io_apic.c by combining two '#ifdef CONFIG_SMP' regions. In addition to making the code easier to understand the first '#ifdef CONFIG_SMP' region is moved to a location later in the file which will reduce the need for function forward declarations when the code subsequently revised. The only changes other than relocating code to a different position in the file were the removal of the assign_irq_vector() forward declaration which was no longer needed and some line length reduction formatting changes. Signed-off-by: Gary Hade Cc: lcm@us.ibm.com LKML-Reference: <20090408210725.GC11159@us.ibm.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 223 ++++++++++++++++++++--------------------- 1 file changed, 109 insertions(+), 114 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 767fe7e46d6..7c9d045ac83 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -518,120 +518,6 @@ static void ioapic_mask_entry(int apic, int pin) spin_unlock_irqrestore(&ioapic_lock, flags); } -#ifdef CONFIG_SMP -static void send_cleanup_vector(struct irq_cfg *cfg) -{ - cpumask_var_t cleanup_mask; - - if (unlikely(!alloc_cpumask_var(&cleanup_mask, GFP_ATOMIC))) { - unsigned int i; - cfg->move_cleanup_count = 0; - for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) - cfg->move_cleanup_count++; - for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) - apic->send_IPI_mask(cpumask_of(i), IRQ_MOVE_CLEANUP_VECTOR); - } else { - cpumask_and(cleanup_mask, cfg->old_domain, cpu_online_mask); - cfg->move_cleanup_count = cpumask_weight(cleanup_mask); - apic->send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); - free_cpumask_var(cleanup_mask); - } - cfg->move_in_progress = 0; -} - -static void __target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq_cfg *cfg) -{ - int apic, pin; - struct irq_pin_list *entry; - u8 vector = cfg->vector; - - entry = cfg->irq_2_pin; - for (;;) { - unsigned int reg; - - if (!entry) - break; - - apic = entry->apic; - pin = entry->pin; - /* - * With interrupt-remapping, destination information comes - * from interrupt-remapping table entry. - */ - if (!irq_remapped(irq)) - io_apic_write(apic, 0x11 + pin*2, dest); - reg = io_apic_read(apic, 0x10 + pin*2); - reg &= ~IO_APIC_REDIR_VECTOR_MASK; - reg |= vector; - io_apic_modify(apic, 0x10 + pin*2, reg); - if (!entry->next) - break; - entry = entry->next; - } -} - -static int -assign_irq_vector(int irq, struct irq_cfg *cfg, const struct cpumask *mask); - -/* - * Either sets desc->affinity to a valid value, and returns - * ->cpu_mask_to_apicid of that, or returns BAD_APICID and - * leaves desc->affinity untouched. - */ -static unsigned int -set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) -{ - struct irq_cfg *cfg; - unsigned int irq; - - if (!cpumask_intersects(mask, cpu_online_mask)) - return BAD_APICID; - - irq = desc->irq; - cfg = desc->chip_data; - if (assign_irq_vector(irq, cfg, mask)) - return BAD_APICID; - - /* check that before desc->addinity get updated */ - set_extra_move_desc(desc, mask); - - cpumask_copy(desc->affinity, mask); - - return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain); -} - -static void -set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) -{ - struct irq_cfg *cfg; - unsigned long flags; - unsigned int dest; - unsigned int irq; - - irq = desc->irq; - cfg = desc->chip_data; - - spin_lock_irqsave(&ioapic_lock, flags); - dest = set_desc_affinity(desc, mask); - if (dest != BAD_APICID) { - /* Only the high 8 bits are valid. */ - dest = SET_APIC_LOGICAL_ID(dest); - __target_IO_APIC_irq(irq, dest, cfg); - } - spin_unlock_irqrestore(&ioapic_lock, flags); -} - -static void -set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) -{ - struct irq_desc *desc; - - desc = irq_to_desc(irq); - - set_ioapic_affinity_irq_desc(desc, mask); -} -#endif /* CONFIG_SMP */ - /* * 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 @@ -2360,6 +2246,115 @@ static int ioapic_retrigger_irq(unsigned int irq) */ #ifdef CONFIG_SMP +static void send_cleanup_vector(struct irq_cfg *cfg) +{ + cpumask_var_t cleanup_mask; + + if (unlikely(!alloc_cpumask_var(&cleanup_mask, GFP_ATOMIC))) { + unsigned int i; + cfg->move_cleanup_count = 0; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + cfg->move_cleanup_count++; + for_each_cpu_and(i, cfg->old_domain, cpu_online_mask) + apic->send_IPI_mask(cpumask_of(i), IRQ_MOVE_CLEANUP_VECTOR); + } else { + cpumask_and(cleanup_mask, cfg->old_domain, cpu_online_mask); + cfg->move_cleanup_count = cpumask_weight(cleanup_mask); + apic->send_IPI_mask(cleanup_mask, IRQ_MOVE_CLEANUP_VECTOR); + free_cpumask_var(cleanup_mask); + } + cfg->move_in_progress = 0; +} + +static void +__target_IO_APIC_irq(unsigned int irq, unsigned int dest, struct irq_cfg *cfg) +{ + int apic, pin; + struct irq_pin_list *entry; + u8 vector = cfg->vector; + + entry = cfg->irq_2_pin; + for (;;) { + unsigned int reg; + + if (!entry) + break; + + apic = entry->apic; + pin = entry->pin; + /* + * With interrupt-remapping, destination information comes + * from interrupt-remapping table entry. + */ + if (!irq_remapped(irq)) + io_apic_write(apic, 0x11 + pin*2, dest); + reg = io_apic_read(apic, 0x10 + pin*2); + reg &= ~IO_APIC_REDIR_VECTOR_MASK; + reg |= vector; + io_apic_modify(apic, 0x10 + pin*2, reg); + if (!entry->next) + break; + entry = entry->next; + } +} + +/* + * Either sets desc->affinity to a valid value, and returns + * ->cpu_mask_to_apicid of that, or returns BAD_APICID and + * leaves desc->affinity untouched. + */ +static unsigned int +set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) +{ + struct irq_cfg *cfg; + unsigned int irq; + + if (!cpumask_intersects(mask, cpu_online_mask)) + return BAD_APICID; + + irq = desc->irq; + cfg = desc->chip_data; + if (assign_irq_vector(irq, cfg, mask)) + return BAD_APICID; + + /* check that before desc->addinity get updated */ + set_extra_move_desc(desc, mask); + + cpumask_copy(desc->affinity, mask); + + return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain); +} + +static void +set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) +{ + struct irq_cfg *cfg; + unsigned long flags; + unsigned int dest; + unsigned int irq; + + irq = desc->irq; + cfg = desc->chip_data; + + spin_lock_irqsave(&ioapic_lock, flags); + dest = set_desc_affinity(desc, mask); + if (dest != BAD_APICID) { + /* Only the high 8 bits are valid. */ + dest = SET_APIC_LOGICAL_ID(dest); + __target_IO_APIC_irq(irq, dest, cfg); + } + spin_unlock_irqrestore(&ioapic_lock, flags); +} + +static void +set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) +{ + struct irq_desc *desc; + + desc = irq_to_desc(irq); + + set_ioapic_affinity_irq_desc(desc, mask); +} #ifdef CONFIG_INTR_REMAP -- cgit v1.2.3-70-g09d2 From 2de1f33e99cec5fd79542a1d0e26efb9c36a98bb Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Sat, 11 Apr 2009 12:55:26 +0530 Subject: x86: apic/x2apic_cluster.c x86_cpu_to_logical_apicid should be static Impact: reduce kernel size a bit, address sparse warning Addresses the problem pointed out by this sparse warning: arch/x86/kernel/apic/x2apic_cluster.c:13:1: warning: symbol 'per_cpu__x86_cpu_to_logical_apicid' was not declared. Should it be static? Signed-off-by: Jaswinder Singh Rajput Cc: Suresh Siddha LKML-Reference: <1239434726.4418.24.camel@localhost.localdomain> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/x2apic_cluster.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/x2apic_cluster.c b/arch/x86/kernel/apic/x2apic_cluster.c index 4a903e2f0d1..8e4cbb255c3 100644 --- a/arch/x86/kernel/apic/x2apic_cluster.c +++ b/arch/x86/kernel/apic/x2apic_cluster.c @@ -10,7 +10,7 @@ #include #include -DEFINE_PER_CPU(u32, x86_cpu_to_logical_apicid); +static DEFINE_PER_CPU(u32, x86_cpu_to_logical_apicid); static int x2apic_acpi_madt_oem_check(char *oem_id, char *oem_table_id) { -- cgit v1.2.3-70-g09d2 From 2c1b284e4fa260fd922b9a65c99169e2630c6862 Mon Sep 17 00:00:00 2001 From: Jaswinder Singh Rajput Date: Sat, 11 Apr 2009 00:03:10 +0530 Subject: x86: clean up declarations and variables Impact: cleanup, no code changed - syscalls.h update declarations due to unifications - irq.c declare smp_generic_interrupt() before it gets used - process.c declare sys_fork() and sys_vfork() before they get used - tsc.c rename tsc_khz shadowed variable - apic/probe_32.c declare apic_default before it gets used - apic/nmi.c prev_nmi_count should be unsigned - apic/io_apic.c declare smp_irq_move_cleanup_interrupt() before it gets used - mm/init.c declare direct_gbpages and free_initrd_mem before they get used Signed-off-by: Jaswinder Singh Rajput Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 3 +++ arch/x86/include/asm/hw_irq.h | 4 ++++ arch/x86/include/asm/pgtable.h | 2 ++ arch/x86/include/asm/pgtable_64.h | 6 ------ arch/x86/include/asm/syscalls.h | 45 +++++++++++++++++++++------------------ arch/x86/kernel/apic/io_apic.c | 1 + arch/x86/kernel/apic/nmi.c | 2 +- arch/x86/kernel/apic/probe_32.c | 1 - arch/x86/kernel/irq.c | 1 + arch/x86/kernel/process.c | 1 + arch/x86/kernel/tsc.c | 8 +++---- arch/x86/mm/init.c | 1 + 12 files changed, 42 insertions(+), 33 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 42f2f837742..5773660c8cd 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -478,6 +478,9 @@ static inline unsigned int read_apic_id(void) extern void default_setup_apic_routing(void); #ifdef CONFIG_X86_32 + +extern struct apic apic_default; + /* * Set up the logical destination ID. * diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index b762ea49bd7..be9ae4111c9 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -78,7 +78,11 @@ extern void eisa_set_level_irq(unsigned int irq); /* SMP */ extern void smp_apic_timer_interrupt(struct pt_regs *); extern void smp_spurious_interrupt(struct pt_regs *); +extern void smp_generic_interrupt(struct pt_regs *); extern void smp_error_interrupt(struct pt_regs *); +#ifdef CONFIG_X86_IO_APIC +extern asmlinkage void smp_irq_move_cleanup_interrupt(void); +#endif #ifdef CONFIG_SMP extern void smp_reschedule_interrupt(struct pt_regs *); extern void smp_call_function_interrupt(struct pt_regs *); diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 29d96d168bc..3f8d09d94eb 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -503,6 +503,8 @@ static inline int pgd_none(pgd_t pgd) #ifndef __ASSEMBLY__ +extern int direct_gbpages; + /* local pte updates need not use xchg for locking */ static inline pte_t native_local_ptep_get_and_clear(pte_t *ptep) { diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 6b87bc6d501..abde308fdb0 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -25,10 +25,6 @@ extern pgd_t init_level4_pgt[]; extern void paging_init(void); -#endif /* !__ASSEMBLY__ */ - -#ifndef __ASSEMBLY__ - #define pte_ERROR(e) \ printk("%s:%d: bad pte %p(%016lx).\n", \ __FILE__, __LINE__, &(e), pte_val(e)) @@ -135,8 +131,6 @@ static inline int pgd_large(pgd_t pgd) { return 0; } #define update_mmu_cache(vma, address, pte) do { } while (0) -extern int direct_gbpages; - /* Encode and de-code a swap entry */ #if _PAGE_BIT_FILE < _PAGE_BIT_PROTNONE #define SWP_TYPE_BITS (_PAGE_BIT_FILE - _PAGE_BIT_PRESENT - 1) diff --git a/arch/x86/include/asm/syscalls.h b/arch/x86/include/asm/syscalls.h index 7043408f690..372b76edd63 100644 --- a/arch/x86/include/asm/syscalls.h +++ b/arch/x86/include/asm/syscalls.h @@ -1,7 +1,7 @@ /* * syscalls.h - Linux syscall interfaces (arch-specific) * - * Copyright (c) 2008 Jaswinder Singh + * Copyright (c) 2008 Jaswinder Singh Rajput * * This file is released under the GPLv2. * See the file COPYING for more details. @@ -12,50 +12,55 @@ #include #include -#include #include +#include /* Common in X86_32 and X86_64 */ /* kernel/ioport.c */ asmlinkage long sys_ioperm(unsigned long, unsigned long, int); +/* kernel/process.c */ +int sys_fork(struct pt_regs *); +int sys_vfork(struct pt_regs *); + /* kernel/ldt.c */ asmlinkage int sys_modify_ldt(int, void __user *, unsigned long); +/* kernel/signal.c */ +long sys_rt_sigreturn(struct pt_regs *); + /* kernel/tls.c */ asmlinkage int sys_set_thread_area(struct user_desc __user *); asmlinkage int sys_get_thread_area(struct user_desc __user *); /* X86_32 only */ #ifdef CONFIG_X86_32 +/* kernel/ioport.c */ +long sys_iopl(struct pt_regs *); + /* kernel/process_32.c */ -int sys_fork(struct pt_regs *); int sys_clone(struct pt_regs *); -int sys_vfork(struct pt_regs *); int sys_execve(struct pt_regs *); -/* kernel/signal_32.c */ +/* kernel/signal.c */ asmlinkage int sys_sigsuspend(int, int, old_sigset_t); asmlinkage int sys_sigaction(int, const struct old_sigaction __user *, struct old_sigaction __user *); int sys_sigaltstack(struct pt_regs *); unsigned long sys_sigreturn(struct pt_regs *); -long sys_rt_sigreturn(struct pt_regs *); - -/* kernel/ioport.c */ -long sys_iopl(struct pt_regs *); /* kernel/sys_i386_32.c */ +struct mmap_arg_struct; +struct sel_arg_struct; +struct oldold_utsname; +struct old_utsname; + asmlinkage long sys_mmap2(unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); -struct mmap_arg_struct; asmlinkage int old_mmap(struct mmap_arg_struct __user *); -struct sel_arg_struct; asmlinkage int old_select(struct sel_arg_struct __user *); asmlinkage int sys_ipc(uint, int, int, int, void __user *, long); -struct old_utsname; asmlinkage int sys_uname(struct old_utsname __user *); -struct oldold_utsname; asmlinkage int sys_olduname(struct oldold_utsname __user *); /* kernel/vm86_32.c */ @@ -65,29 +70,27 @@ int sys_vm86(struct pt_regs *); #else /* CONFIG_X86_32 */ /* X86_64 only */ +/* kernel/ioport.c */ +asmlinkage long sys_iopl(unsigned int, struct pt_regs *); + /* kernel/process_64.c */ -asmlinkage long sys_fork(struct pt_regs *); asmlinkage long sys_clone(unsigned long, unsigned long, void __user *, void __user *, struct pt_regs *); -asmlinkage long sys_vfork(struct pt_regs *); asmlinkage long sys_execve(char __user *, char __user * __user *, char __user * __user *, struct pt_regs *); long sys_arch_prctl(int, unsigned long); -/* kernel/ioport.c */ -asmlinkage long sys_iopl(unsigned int, struct pt_regs *); - -/* kernel/signal_64.c */ +/* kernel/signal.c */ asmlinkage long sys_sigaltstack(const stack_t __user *, stack_t __user *, struct pt_regs *); -long sys_rt_sigreturn(struct pt_regs *); /* kernel/sys_x86_64.c */ +struct new_utsname; + asmlinkage long sys_mmap(unsigned long, unsigned long, unsigned long, unsigned long, unsigned long, unsigned long); -struct new_utsname; asmlinkage long sys_uname(struct new_utsname __user *); #endif /* CONFIG_X86_32 */ diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 767fe7e46d6..870c92ddaf9 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -59,6 +59,7 @@ #include #include #include +#include #include #include diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c index d6bd6240715..02056310f2f 100644 --- a/arch/x86/kernel/apic/nmi.c +++ b/arch/x86/kernel/apic/nmi.c @@ -104,7 +104,7 @@ static __init void nmi_cpu_busy(void *data) } #endif -static void report_broken_nmi(int cpu, int *prev_nmi_count) +static void report_broken_nmi(int cpu, unsigned int *prev_nmi_count) { printk(KERN_CONT "\n"); diff --git a/arch/x86/kernel/apic/probe_32.c b/arch/x86/kernel/apic/probe_32.c index 01eda2ac65e..440a8bccd91 100644 --- a/arch/x86/kernel/apic/probe_32.c +++ b/arch/x86/kernel/apic/probe_32.c @@ -160,7 +160,6 @@ extern struct apic apic_summit; extern struct apic apic_bigsmp; extern struct apic apic_es7000; extern struct apic apic_es7000_cluster; -extern struct apic apic_default; struct apic *apic = &apic_default; EXPORT_SYMBOL_GPL(apic); diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 3aaf7b9e3a8..2188267f523 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -12,6 +12,7 @@ #include #include #include +#include atomic_t irq_err_count; diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index ca989158e84..3e21e38d7e3 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index 7a567ebe636..a8dc0d00b83 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -384,13 +384,13 @@ unsigned long native_calibrate_tsc(void) { u64 tsc1, tsc2, delta, ref1, ref2; unsigned long tsc_pit_min = ULONG_MAX, tsc_ref_min = ULONG_MAX; - unsigned long flags, latch, ms, fast_calibrate, tsc_khz; + unsigned long flags, latch, ms, fast_calibrate, hv_tsc_khz; int hpet = is_hpet_enabled(), i, loopmin; - tsc_khz = get_hypervisor_tsc_freq(); - if (tsc_khz) { + hv_tsc_khz = get_hypervisor_tsc_freq(); + if (hv_tsc_khz) { printk(KERN_INFO "TSC: Frequency read from the hypervisor\n"); - return tsc_khz; + return hv_tsc_khz; } local_irq_save(flags); diff --git a/arch/x86/mm/init.c b/arch/x86/mm/init.c index fd3da1dda1c..40924e445f5 100644 --- a/arch/x86/mm/init.c +++ b/arch/x86/mm/init.c @@ -1,3 +1,4 @@ +#include #include #include -- cgit v1.2.3-70-g09d2 From c0eaa4536f08b98fbcfa7fce5b7b0de1bebcb0e1 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 12 Apr 2009 20:47:40 +0400 Subject: x86: apic - introduce imcr_ helpers Impact: cleanup Distinguish port writting magic into helpers with comments. Signed-off-by: Cyrill Gorcunov LKML-Reference: <20090412165058.535921550@openvz.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 29 +++++++++++++++++++++++++---- 1 file changed, 25 insertions(+), 4 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 098ec84b8c0..c3be10f5773 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -98,6 +98,29 @@ early_param("lapic", parse_lapic); /* Local APIC was disabled by the BIOS and enabled by the kernel */ static int enabled_via_apicbase; +/* + * Handle interrupt mode configuration register (IMCR). + * This register controls whether the interrupt signals + * that reach the BSP come from the master PIC or from the + * local APIC. Before entering Symmetric I/O Mode, either + * the BIOS or the operating system must switch out of + * PIC Mode by changing the IMCR. + */ +static inline imcr_pic_to_apic(void) +{ + /* select IMCR register */ + outb(0x70, 0x22); + /* NMI and 8259 INTR go through APIC */ + outb(0x01, 0x23); +} + +static inline imcr_apic_to_pic(void) +{ + /* select IMCR register */ + outb(0x70, 0x22); + /* NMI and 8259 INTR go directly to BSP */ + outb(0x00, 0x23); +} #endif #ifdef CONFIG_X86_64 @@ -1727,8 +1750,7 @@ void __init connect_bsp_APIC(void) */ apic_printk(APIC_VERBOSE, "leaving PIC mode, " "enabling APIC mode.\n"); - outb(0x70, 0x22); - outb(0x01, 0x23); + imcr_pic_to_apic(); } #endif if (apic->enable_apic_mode) @@ -1756,8 +1778,7 @@ void disconnect_bsp_APIC(int virt_wire_setup) */ apic_printk(APIC_VERBOSE, "disabling APIC mode, " "entering PIC mode.\n"); - outb(0x70, 0x22); - outb(0x00, 0x23); + imcr_apic_to_pic(); return; } #endif -- cgit v1.2.3-70-g09d2 From 08306ce61d6848e6fbf74fa4cc693c3fb29e943f Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 12 Apr 2009 20:47:41 +0400 Subject: x86: apic - introduce dummy apic operations Impact: refactor, speed up and robustize code In case if apic was disabled by kernel option or by hardware limits we can use dummy operations in apic->write to simplify the ack_APIC_irq() code. At the lame time the patch fixes the missed EOI in do_IRQ function (which has place if kernel is compiled as X86-32 and interrupt without handler happens where apic was not asked to be disabled via kernel option). Note that native_apic_write_dummy() consists of WARN_ON_ONCE to catch any buggy writes on enabled APICs. Could be removed after some time of testing. Signed-off-by: Cyrill Gorcunov LKML-Reference: <20090412165058.724788431@openvz.org> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 3 ++- arch/x86/kernel/apic/apic.c | 24 ++++++++++++++++++++++++ arch/x86/kernel/irq.c | 10 ++-------- 3 files changed, 28 insertions(+), 9 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 42f2f837742..2bd5a463fd1 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -212,6 +212,7 @@ static inline void ack_x2APIC_irq(void) } #endif +extern void apic_disable(void); extern int lapic_get_maxlvt(void); extern void clear_local_APIC(void); extern void connect_bsp_APIC(void); @@ -252,7 +253,7 @@ static inline void lapic_shutdown(void) { } #define local_apic_timer_c2_ok 1 static inline void init_apic_mappings(void) { } static inline void disable_local_APIC(void) { } - +static inline void apic_disable(void) { } #endif /* !CONFIG_X86_LOCAL_APIC */ #ifdef CONFIG_X86_64 diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index c3be10f5773..9b849d4957d 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -232,6 +232,24 @@ static int modern_apic(void) return lapic_get_version() >= 0x14; } +/* + * bare function to substitute write operation + * and it's _that_ fast :) + */ +void native_apic_write_dummy(u32 reg, u32 v) +{ + WARN_ON_ONCE((cpu_has_apic || !disable_apic)); +} + +/* + * right after this call apic->write doesn't do anything + * note that there is no restore operation it works one way + */ +void apic_disable(void) +{ + apic->write = native_apic_write_dummy; +} + void native_apic_wait_icr_idle(void) { while (apic_read(APIC_ICR) & APIC_ICR_BUSY) @@ -1582,6 +1600,12 @@ void __init init_apic_mappings(void) */ if (boot_cpu_physical_apicid == -1U) boot_cpu_physical_apicid = read_apic_id(); + + /* lets check if we may to NOP'ify apic operations */ + if (!cpu_has_apic) { + pr_info("APIC: disable apic facility\n"); + apic_disable(); + } } /* diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index 6603492e8b7..fd57bf35d0f 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -27,7 +27,6 @@ void ack_bad_irq(unsigned int irq) if (printk_ratelimit()) pr_err("unexpected IRQ trap at vector %02x\n", irq); -#ifdef CONFIG_X86_LOCAL_APIC /* * Currently unexpected vectors happen only on SMP and APIC. * We _must_ ack these because every local APIC has only N @@ -37,9 +36,7 @@ void ack_bad_irq(unsigned int irq) * completely. * But only ack when the APIC is enabled -AK */ - if (cpu_has_apic) - ack_APIC_irq(); -#endif + ack_APIC_irq(); } #define irq_stats(x) (&per_cpu(irq_stat, x)) @@ -214,10 +211,7 @@ unsigned int __irq_entry do_IRQ(struct pt_regs *regs) irq = __get_cpu_var(vector_irq)[vector]; if (!handle_irq(irq, regs)) { -#ifdef CONFIG_X86_64 - if (!disable_apic) - ack_APIC_irq(); -#endif + ack_APIC_irq(); if (printk_ratelimit()) pr_emerg("%s: %d.%d No irq handler for vector (irq %d)\n", -- cgit v1.2.3-70-g09d2 From 5cda395f4a262788d8ed79ac8a26a2b821e5f751 Mon Sep 17 00:00:00 2001 From: Alexander van Heukelum Date: Mon, 13 Apr 2009 17:39:24 +0200 Subject: x86: fix function definitions after: x86: apic - introduce imcr_ helpers The patch "introduce imcr_ helpers" introduced good comments, but also a few new compile warnings. This fixes the function definitions to have a 'void' return type. Signed-off-by: Alexander van Heukelum Acked-by: Cyrill Gorcunov LKML-Reference: <20090413153924.GA20287@mailshack.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 9b849d4957d..4b48ff9163c 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -106,7 +106,7 @@ static int enabled_via_apicbase; * the BIOS or the operating system must switch out of * PIC Mode by changing the IMCR. */ -static inline imcr_pic_to_apic(void) +static inline void imcr_pic_to_apic(void) { /* select IMCR register */ outb(0x70, 0x22); @@ -114,7 +114,7 @@ static inline imcr_pic_to_apic(void) outb(0x01, 0x23); } -static inline imcr_apic_to_pic(void) +static inline void imcr_apic_to_pic(void) { /* select IMCR register */ outb(0x70, 0x22); -- cgit v1.2.3-70-g09d2 From 5d0ae2db6deac4f15dac4f42f23bc56448fc8d4d Mon Sep 17 00:00:00 2001 From: Weidong Han Date: Fri, 17 Apr 2009 16:42:13 +0800 Subject: x86, intr-remap: fix ack for interrupt remapping Shouldn't call ack_apic_edge() in ir_ack_apic_edge(), because ack_apic_edge() does more than just ack: it also does irq migration in the non-interrupt-remapping case. But there is no such need for interrupt-remapping case, as irq migration is done in the process context. Similarly, ir_ack_apic_level() shouldn't call ack_apic_level, and instead should do the local cpu's EOI + directed EOI to the io-apic. ack_x2APIC_irq() is not neccessary, because ack_APIC_irq() will use MSR write for x2apic, and uncached write for non-x2apic. [ Impact: simplify/standardize intr-remap IRQ acking, fix on !x2apic ] Signed-off-by: Suresh Siddha Signed-off-by: Weidong Han Acked-by: David Woodhouse Cc: iommu@lists.linux-foundation.org Cc: allen.m.kay@intel.com Cc: fenghua.yu@intel.com LKML-Reference: <1239957736-6161-3-git-send-email-weidong.han@intel.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 8 -------- arch/x86/kernel/apic/io_apic.c | 32 +++++--------------------------- 2 files changed, 5 insertions(+), 35 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 2bd5a463fd1..d4cb7e590c0 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -204,14 +204,6 @@ static inline int x2apic_enabled(void) extern int get_physical_broadcast(void); -#ifdef CONFIG_X86_X2APIC -static inline void ack_x2APIC_irq(void) -{ - /* Docs say use 0 for future compatibility */ - native_apic_msr_write(APIC_EOI, 0); -} -#endif - extern void apic_disable(void); extern int lapic_get_maxlvt(void); extern void clear_local_APIC(void); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 84990002240..ea22a86e3cd 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -2552,20 +2552,6 @@ eoi_ioapic_irq(struct irq_desc *desc) spin_unlock_irqrestore(&ioapic_lock, flags); } -#ifdef CONFIG_X86_X2APIC -static void ack_x2apic_level(unsigned int irq) -{ - struct irq_desc *desc = irq_to_desc(irq); - ack_x2APIC_irq(); - eoi_ioapic_irq(desc); -} - -static void ack_x2apic_edge(unsigned int irq) -{ - ack_x2APIC_irq(); -} -#endif - static void ack_apic_edge(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); @@ -2629,9 +2615,6 @@ static void ack_apic_level(unsigned int irq) */ ack_APIC_irq(); - if (irq_remapped(irq)) - eoi_ioapic_irq(desc); - /* Now we can move and renable the irq */ if (unlikely(do_unmask_irq)) { /* Only migrate the irq if the ack has been received. @@ -2680,20 +2663,15 @@ static void ack_apic_level(unsigned int irq) #ifdef CONFIG_INTR_REMAP static void ir_ack_apic_edge(unsigned int irq) { -#ifdef CONFIG_X86_X2APIC - if (x2apic_enabled()) - return ack_x2apic_edge(irq); -#endif - return ack_apic_edge(irq); + ack_APIC_irq(); } static void ir_ack_apic_level(unsigned int irq) { -#ifdef CONFIG_X86_X2APIC - if (x2apic_enabled()) - return ack_x2apic_level(irq); -#endif - return ack_apic_level(irq); + struct irq_desc *desc = irq_to_desc(irq); + + ack_APIC_irq(); + eoi_ioapic_irq(desc); } #endif /* CONFIG_INTR_REMAP */ -- cgit v1.2.3-70-g09d2 From 937582382c71b75b29fbb92615629494e1a05ac0 Mon Sep 17 00:00:00 2001 From: Weidong Han Date: Fri, 17 Apr 2009 16:42:14 +0800 Subject: x86, intr-remap: enable interrupt remapping early Currently, when x2apic is not enabled, interrupt remapping will be enabled in init_dmars(), where it is too late to remap ioapic interrupts, that is, ioapic interrupts are really in compatibility mode, not remappable mode. This patch always enables interrupt remapping before ioapic setup, it guarantees all interrupts will be remapped when interrupt remapping is enabled. Thus it doesn't need to set the compatibility interrupt bit. [ Impact: refactor intr-remap init sequence, enable fuller remap mode ] Signed-off-by: Suresh Siddha Signed-off-by: Weidong Han Acked-by: David Woodhouse Cc: iommu@lists.linux-foundation.org Cc: allen.m.kay@intel.com Cc: fenghua.yu@intel.com LKML-Reference: <1239957736-6161-4-git-send-email-weidong.han@intel.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 7 ++-- arch/x86/kernel/apic/apic.c | 76 +++++++++++++++++++++----------------------- drivers/pci/intel-iommu.c | 9 ------ drivers/pci/intr_remapping.c | 28 ++++++++-------- include/linux/dmar.h | 1 + 5 files changed, 54 insertions(+), 67 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index d4cb7e590c0..fbdd65446c7 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -169,7 +169,6 @@ static inline u64 native_x2apic_icr_read(void) extern int x2apic, x2apic_phys; extern void check_x2apic(void); extern void enable_x2apic(void); -extern void enable_IR_x2apic(void); extern void x2apic_icr_write(u32 low, u32 id); static inline int x2apic_enabled(void) { @@ -190,18 +189,18 @@ static inline void check_x2apic(void) static inline void enable_x2apic(void) { } -static inline void enable_IR_x2apic(void) -{ -} static inline int x2apic_enabled(void) { return 0; } #define x2apic 0 +#define x2apic_preenabled 0 #endif +extern void enable_IR_x2apic(void); + extern int get_physical_broadcast(void); extern void apic_disable(void); diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 83e47febcc8..0cf1eea750c 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -141,6 +141,8 @@ static int x2apic_preenabled; static int disable_x2apic; static __init int setup_nox2apic(char *str) { + if (x2apic_enabled()) + panic("Bios already enabled x2apic, can't enforce nox2apic"); disable_x2apic = 1; setup_clear_cpu_cap(X86_FEATURE_X2APIC); return 0; @@ -1345,6 +1347,7 @@ void enable_x2apic(void) wrmsr(MSR_IA32_APICBASE, msr | X2APIC_ENABLE, 0); } } +#endif /* CONFIG_X86_X2APIC */ void __init enable_IR_x2apic(void) { @@ -1353,32 +1356,21 @@ void __init enable_IR_x2apic(void) unsigned long flags; struct IO_APIC_route_entry **ioapic_entries = NULL; - if (!cpu_has_x2apic) - return; - - if (!x2apic_preenabled && disable_x2apic) { - pr_info("Skipped enabling x2apic and Interrupt-remapping " - "because of nox2apic\n"); - return; + ret = dmar_table_init(); + if (ret) { + pr_debug("dmar_table_init() failed with %d:\n", ret); + goto ir_failed; } - if (x2apic_preenabled && disable_x2apic) - panic("Bios already enabled x2apic, can't enforce nox2apic"); - - if (!x2apic_preenabled && skip_ioapic_setup) { - pr_info("Skipped enabling x2apic and Interrupt-remapping " - "because of skipping io-apic setup\n"); - return; + if (!intr_remapping_supported()) { + pr_debug("intr-remapping not supported\n"); + goto ir_failed; } - ret = dmar_table_init(); - if (ret) { - pr_info("dmar_table_init() failed with %d:\n", ret); - if (x2apic_preenabled) - panic("x2apic enabled by bios. But IR enabling failed"); - else - pr_info("Not enabling x2apic,Intr-remapping\n"); + if (!x2apic_preenabled && skip_ioapic_setup) { + pr_info("Skipped enabling intr-remap because of skipping " + "io-apic setup\n"); return; } @@ -1398,20 +1390,25 @@ void __init enable_IR_x2apic(void) mask_IO_APIC_setup(ioapic_entries); mask_8259A(); - ret = enable_intr_remapping(EIM_32BIT_APIC_ID); - - if (ret && x2apic_preenabled) { - local_irq_restore(flags); - panic("x2apic enabled by bios. But IR enabling failed"); - } +#ifdef CONFIG_X86_X2APIC + if (cpu_has_x2apic) + ret = enable_intr_remapping(EIM_32BIT_APIC_ID); + else +#endif + ret = enable_intr_remapping(EIM_8BIT_APIC_ID); if (ret) goto end_restore; - if (!x2apic) { + pr_info("Enabled Interrupt-remapping\n"); + +#ifdef CONFIG_X86_X2APIC + if (cpu_has_x2apic && !x2apic) { x2apic = 1; enable_x2apic(); + pr_info("Enabled x2apic\n"); } +#endif end_restore: if (ret) @@ -1426,30 +1423,29 @@ end_restore: local_irq_restore(flags); end: - if (!ret) { - if (!x2apic_preenabled) - pr_info("Enabled x2apic and interrupt-remapping\n"); - else - pr_info("Enabled Interrupt-remapping\n"); - } else - pr_err("Failed to enable Interrupt-remapping and x2apic\n"); if (ioapic_entries) free_ioapic_entries(ioapic_entries); + + if (!ret) + return; + +ir_failed: + if (x2apic_preenabled) + panic("x2apic enabled by bios. But IR enabling failed"); + else if (cpu_has_x2apic) + pr_info("Not enabling x2apic,Intr-remapping\n"); #else if (!cpu_has_x2apic) return; if (x2apic_preenabled) panic("x2apic enabled prior OS handover," - " enable CONFIG_INTR_REMAP"); - - pr_info("Enable CONFIG_INTR_REMAP for enabling intr-remapping " - " and x2apic\n"); + " enable CONFIG_X86_X2APIC, CONFIG_INTR_REMAP"); #endif return; } -#endif /* CONFIG_X86_X2APIC */ + #ifdef CONFIG_X86_64 /* diff --git a/drivers/pci/intel-iommu.c b/drivers/pci/intel-iommu.c index 001b328adf8..9ce8f0764be 100644 --- a/drivers/pci/intel-iommu.c +++ b/drivers/pci/intel-iommu.c @@ -1968,15 +1968,6 @@ static int __init init_dmars(void) } } -#ifdef CONFIG_INTR_REMAP - if (!intr_remapping_enabled) { - ret = enable_intr_remapping(0); - if (ret) - printk(KERN_ERR - "IOMMU: enable interrupt remapping failed\n"); - } -#endif - /* * For each rmrr * for each dev attached to rmrr diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index f5e0ea724a6..5c2142656e9 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -423,20 +423,6 @@ static void iommu_set_intr_remapping(struct intel_iommu *iommu, int mode) readl, (sts & DMA_GSTS_IRTPS), sts); spin_unlock_irqrestore(&iommu->register_lock, flags); - if (mode == 0) { - spin_lock_irqsave(&iommu->register_lock, flags); - - /* enable comaptiblity format interrupt pass through */ - cmd = iommu->gcmd | DMA_GCMD_CFI; - iommu->gcmd |= DMA_GCMD_CFI; - writel(cmd, iommu->reg + DMAR_GCMD_REG); - - IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, - readl, (sts & DMA_GSTS_CFIS), sts); - - spin_unlock_irqrestore(&iommu->register_lock, flags); - } - /* * global invalidation of interrupt entry cache before enabling * interrupt-remapping. @@ -516,6 +502,20 @@ end: spin_unlock_irqrestore(&iommu->register_lock, flags); } +int __init intr_remapping_supported(void) +{ + struct dmar_drhd_unit *drhd; + + for_each_drhd_unit(drhd) { + struct intel_iommu *iommu = drhd->iommu; + + if (!ecap_ir_support(iommu->ecap)) + return 0; + } + + return 1; +} + int __init enable_intr_remapping(int eim) { struct dmar_drhd_unit *drhd; diff --git a/include/linux/dmar.h b/include/linux/dmar.h index e397dc342cd..06f592a7f73 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -108,6 +108,7 @@ struct irte { }; #ifdef CONFIG_INTR_REMAP extern int intr_remapping_enabled; +extern int intr_remapping_supported(void); extern int enable_intr_remapping(int); extern void disable_intr_remapping(void); extern int reenable_intr_remapping(int); -- cgit v1.2.3-70-g09d2 From 9a2755c3569e4db92bd9b1daadeddb4045b0cccd Mon Sep 17 00:00:00 2001 From: Weidong Han Date: Fri, 17 Apr 2009 16:42:16 +0800 Subject: x86, intr-remap: fix x2apic/intr-remap resume Interrupt remapping was decoupled from x2apic. Shouldn't check x2apic before resume interrupt remapping. Otherwise, interrupt remapping won't be resumed when x2apic is not enabled. [ Impact: fix potential intr-remap resume hang on !x2apic ] Signed-off-by: Suresh Siddha Signed-off-by: Weidong Han Acked-by: David Woodhouse Cc: iommu@lists.linux-foundation.org Cc: allen.m.kay@intel.com Cc: fenghua.yu@intel.com LKML-Reference: <1239957736-6161-6-git-send-email-weidong.han@intel.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 0cf1eea750c..7b41a32339e 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -2032,7 +2032,7 @@ static int lapic_resume(struct sys_device *dev) return 0; local_irq_save(flags); - if (x2apic) { + if (intr_remapping_enabled) { ioapic_entries = alloc_ioapic_entries(); if (!ioapic_entries) { WARN(1, "Alloc ioapic_entries in lapic resume failed."); @@ -2048,8 +2048,10 @@ static int lapic_resume(struct sys_device *dev) mask_IO_APIC_setup(ioapic_entries); mask_8259A(); - enable_x2apic(); } + + if (x2apic) + enable_x2apic(); #else if (!apic_pm_state.active) return 0; @@ -2097,10 +2099,12 @@ static int lapic_resume(struct sys_device *dev) apic_read(APIC_ESR); #ifdef CONFIG_INTR_REMAP - if (intr_remapping_enabled) - reenable_intr_remapping(EIM_32BIT_APIC_ID); + if (intr_remapping_enabled) { + if (x2apic) + reenable_intr_remapping(EIM_32BIT_APIC_ID); + else + reenable_intr_remapping(EIM_8BIT_APIC_ID); - if (x2apic) { unmask_8259A(); restore_IO_APIC_setup(ioapic_entries); free_ioapic_entries(ioapic_entries); @@ -2109,7 +2113,6 @@ static int lapic_resume(struct sys_device *dev) local_irq_restore(flags); - return 0; } -- cgit v1.2.3-70-g09d2 From 667c5296cc76fefe0abcb79228952b28d9af45e3 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 19 Apr 2009 11:43:11 +0400 Subject: x86: es7000, uv - use __cpuinit for kicking secondary cpus The caller already has __cpuinit attribute. [ Impact: save memory, address section mismatch warning ] Signed-off-by: Cyrill Gorcunov Cc: Yinghai Lu Cc: Pavel Emelyanov LKML-Reference: <20090419074311.GA8670@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/es7000_32.c | 2 +- arch/x86/kernel/apic/x2apic_uv_x.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/es7000_32.c b/arch/x86/kernel/apic/es7000_32.c index 1c11b819f24..8e07c141866 100644 --- a/arch/x86/kernel/apic/es7000_32.c +++ b/arch/x86/kernel/apic/es7000_32.c @@ -145,7 +145,7 @@ es7000_rename_gsi(int ioapic, int gsi) return gsi; } -static int wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) +static int __cpuinit wakeup_secondary_cpu_via_mip(int cpu, unsigned long eip) { unsigned long vect = 0, psaival = 0; diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index de1a50af807..873bf7121e8 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -91,7 +91,7 @@ static void uv_vector_allocation_domain(int cpu, struct cpumask *retmask) cpumask_set_cpu(cpu, retmask); } -static int uv_wakeup_secondary(int phys_apicid, unsigned long start_rip) +static int __cpuinit uv_wakeup_secondary(int phys_apicid, unsigned long start_rip) { #ifdef CONFIG_SMP unsigned long val; -- cgit v1.2.3-70-g09d2 From fc1edaf9e7cc4d4696f83dee495b8f158d01c4eb Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 20 Apr 2009 13:02:27 -0700 Subject: x86: x2apic, IR: Clean up X86_X2APIC and INTR_REMAP config checks Add x2apic_supported() to clean up CONFIG_X86_X2APIC checks. Fix CONFIG_INTR_REMAP checks. [ Impact: cleanup ] Signed-off-by: Suresh Siddha Cc: dwmw2@infradead.org Cc: Suresh Siddha Cc: Weidong Han LKML-Reference: <20090420200450.128993000@linux-os.sc.intel.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apic.h | 10 ++++---- arch/x86/include/asm/io_apic.h | 2 -- arch/x86/include/asm/irq_remapping.h | 2 +- arch/x86/kernel/apic/apic.c | 49 +++++++++--------------------------- arch/x86/kernel/apic/io_apic.c | 2 -- arch/x86/kernel/apic/probe_64.c | 2 +- include/linux/dmar.h | 2 ++ 7 files changed, 21 insertions(+), 48 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index fbdd65446c7..3738438a91f 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -107,8 +107,7 @@ extern u32 native_safe_apic_wait_icr_idle(void); extern void native_apic_icr_write(u32 low, u32 id); extern u64 native_apic_icr_read(void); -#define EIM_8BIT_APIC_ID 0 -#define EIM_32BIT_APIC_ID 1 +extern int x2apic_mode; #ifdef CONFIG_X86_X2APIC /* @@ -166,7 +165,7 @@ static inline u64 native_x2apic_icr_read(void) return val; } -extern int x2apic, x2apic_phys; +extern int x2apic_phys; extern void check_x2apic(void); extern void enable_x2apic(void); extern void x2apic_icr_write(u32 low, u32 id); @@ -182,6 +181,8 @@ static inline int x2apic_enabled(void) return 1; return 0; } + +#define x2apic_supported() (cpu_has_x2apic) #else static inline void check_x2apic(void) { @@ -194,9 +195,8 @@ static inline int x2apic_enabled(void) return 0; } -#define x2apic 0 #define x2apic_preenabled 0 - +#define x2apic_supported() 0 #endif extern void enable_IR_x2apic(void); diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 9d826e43601..34eaa37f7ad 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -161,7 +161,6 @@ extern int io_apic_set_pci_routing(int ioapic, int pin, int irq, extern int (*ioapic_renumber_irq)(int ioapic, int irq); extern void ioapic_init_mappings(void); -#ifdef CONFIG_X86_64 extern struct IO_APIC_route_entry **alloc_ioapic_entries(void); extern void free_ioapic_entries(struct IO_APIC_route_entry **ioapic_entries); extern int save_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); @@ -169,7 +168,6 @@ extern void mask_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); extern int restore_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); extern void reinit_intr_remapped_IO_APIC(int intr_remapping, struct IO_APIC_route_entry **ioapic_entries); -#endif extern void probe_nr_irqs_gsi(void); diff --git a/arch/x86/include/asm/irq_remapping.h b/arch/x86/include/asm/irq_remapping.h index 0396760fccb..f275e224450 100644 --- a/arch/x86/include/asm/irq_remapping.h +++ b/arch/x86/include/asm/irq_remapping.h @@ -1,6 +1,6 @@ #ifndef _ASM_X86_IRQ_REMAPPING_H #define _ASM_X86_IRQ_REMAPPING_H -#define IRTE_DEST(dest) ((x2apic) ? dest : dest << 8) +#define IRTE_DEST(dest) ((x2apic_mode) ? dest : dest << 8) #endif /* _ASM_X86_IRQ_REMAPPING_H */ diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 7b41a32339e..2b30e520dce 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -134,8 +134,8 @@ static __init int setup_apicpmtimer(char *s) __setup("apicpmtimer", setup_apicpmtimer); #endif +int x2apic_mode; #ifdef CONFIG_X86_X2APIC -int x2apic; /* x2apic enabled before OS handover */ static int x2apic_preenabled; static int disable_x2apic; @@ -858,7 +858,7 @@ void clear_local_APIC(void) u32 v; /* APIC hasn't been mapped yet */ - if (!x2apic && !apic_phys) + if (!x2apic_mode && !apic_phys) return; maxlvt = lapic_get_maxlvt(); @@ -1330,7 +1330,7 @@ void check_x2apic(void) { if (x2apic_enabled()) { pr_info("x2apic enabled by BIOS, switching to x2apic ops\n"); - x2apic_preenabled = x2apic = 1; + x2apic_preenabled = x2apic_mode = 1; } } @@ -1338,7 +1338,7 @@ void enable_x2apic(void) { int msr, msr2; - if (!x2apic) + if (!x2apic_mode) return; rdmsr(MSR_IA32_APICBASE, msr, msr2); @@ -1390,25 +1390,17 @@ void __init enable_IR_x2apic(void) mask_IO_APIC_setup(ioapic_entries); mask_8259A(); -#ifdef CONFIG_X86_X2APIC - if (cpu_has_x2apic) - ret = enable_intr_remapping(EIM_32BIT_APIC_ID); - else -#endif - ret = enable_intr_remapping(EIM_8BIT_APIC_ID); - + ret = enable_intr_remapping(x2apic_supported()); if (ret) goto end_restore; pr_info("Enabled Interrupt-remapping\n"); -#ifdef CONFIG_X86_X2APIC - if (cpu_has_x2apic && !x2apic) { - x2apic = 1; + if (x2apic_supported() && !x2apic_mode) { + x2apic_mode = 1; enable_x2apic(); pr_info("Enabled x2apic\n"); } -#endif end_restore: if (ret) @@ -1576,7 +1568,7 @@ void __init early_init_lapic_mapping(void) */ void __init init_apic_mappings(void) { - if (x2apic) { + if (x2apic_mode) { boot_cpu_physical_apicid = read_apic_id(); return; } @@ -2010,10 +2002,10 @@ static int lapic_suspend(struct sys_device *dev, pm_message_t state) local_irq_save(flags); disable_local_APIC(); -#ifdef CONFIG_INTR_REMAP + if (intr_remapping_enabled) disable_intr_remapping(); -#endif + local_irq_restore(flags); return 0; } @@ -2023,8 +2015,6 @@ static int lapic_resume(struct sys_device *dev) unsigned int l, h; unsigned long flags; int maxlvt; - -#ifdef CONFIG_INTR_REMAP int ret; struct IO_APIC_route_entry **ioapic_entries = NULL; @@ -2050,17 +2040,8 @@ static int lapic_resume(struct sys_device *dev) mask_8259A(); } - if (x2apic) + if (x2apic_mode) enable_x2apic(); -#else - if (!apic_pm_state.active) - return 0; - - local_irq_save(flags); - if (x2apic) - enable_x2apic(); -#endif - else { /* * Make sure the APICBASE points to the right address @@ -2098,18 +2079,12 @@ static int lapic_resume(struct sys_device *dev) apic_write(APIC_ESR, 0); apic_read(APIC_ESR); -#ifdef CONFIG_INTR_REMAP if (intr_remapping_enabled) { - if (x2apic) - reenable_intr_remapping(EIM_32BIT_APIC_ID); - else - reenable_intr_remapping(EIM_8BIT_APIC_ID); - + reenable_intr_remapping(x2apic_mode); unmask_8259A(); restore_IO_APIC_setup(ioapic_entries); free_ioapic_entries(ioapic_entries); } -#endif local_irq_restore(flags); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index ea22a86e3cd..3a45d2ec974 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -736,7 +736,6 @@ static int __init ioapic_pirq_setup(char *str) __setup("pirq=", ioapic_pirq_setup); #endif /* CONFIG_X86_32 */ -#ifdef CONFIG_INTR_REMAP struct IO_APIC_route_entry **alloc_ioapic_entries(void) { int apic; @@ -857,7 +856,6 @@ void free_ioapic_entries(struct IO_APIC_route_entry **ioapic_entries) kfree(ioapic_entries); } -#endif /* * Find the IRQ entry number of a certain pin. diff --git a/arch/x86/kernel/apic/probe_64.c b/arch/x86/kernel/apic/probe_64.c index 1783652bb0e..bc3e880f9b8 100644 --- a/arch/x86/kernel/apic/probe_64.c +++ b/arch/x86/kernel/apic/probe_64.c @@ -50,7 +50,7 @@ static struct apic *apic_probe[] __initdata = { void __init default_setup_apic_routing(void) { #ifdef CONFIG_X86_X2APIC - if (x2apic && (apic != &apic_x2apic_phys && + if (x2apic_mode && (apic != &apic_x2apic_phys && #ifdef CONFIG_X86_UV apic != &apic_x2apic_uv_x && #endif diff --git a/include/linux/dmar.h b/include/linux/dmar.h index 06f592a7f73..10ff5c49882 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -158,6 +158,8 @@ static inline struct intel_iommu *map_ioapic_to_ir(int apic) } #define irq_remapped(irq) (0) #define enable_intr_remapping(mode) (-1) +#define disable_intr_remapping() (0) +#define reenable_intr_remapping(mode) (0) #define intr_remapping_enabled (0) #endif -- cgit v1.2.3-70-g09d2 From 25629d810a52176758401184d9b437fbb7f79195 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 20 Apr 2009 13:02:28 -0700 Subject: x86: x2apic, IR: Move eoi_ioapic_irq() into a CONFIG_INTR_REMAP section Address the following complier warning: arch/x86/kernel/apic/io_apic.c:2543: warning: `eoi_ioapic_irq' defined but not used By moving that function (and eoi_ioapic_irq()) into an existing #ifdef CONFIG_INTR_REMAP section of the code. [ Impact: cleanup ] Signed-off-by: Suresh Siddha Cc: dwmw2@infradead.org Cc: Weidong Han LKML-Reference: <20090420200450.271099000@linux-os.sc.intel.com> Signed-off-by: Ingo Molnar Cc: Weidong Han --- arch/x86/kernel/apic/io_apic.c | 66 +++++++++++++++++++++--------------------- 1 file changed, 33 insertions(+), 33 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 3a45d2ec974..4baa9cbd630 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -2517,39 +2517,6 @@ static void irq_complete_move(struct irq_desc **descp) static inline void irq_complete_move(struct irq_desc **descp) {} #endif -static void __eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) -{ - int apic, pin; - struct irq_pin_list *entry; - - entry = cfg->irq_2_pin; - for (;;) { - - if (!entry) - break; - - apic = entry->apic; - pin = entry->pin; - io_apic_eoi(apic, pin); - entry = entry->next; - } -} - -static void -eoi_ioapic_irq(struct irq_desc *desc) -{ - struct irq_cfg *cfg; - unsigned long flags; - unsigned int irq; - - irq = desc->irq; - cfg = desc->chip_data; - - spin_lock_irqsave(&ioapic_lock, flags); - __eoi_ioapic_irq(irq, cfg); - spin_unlock_irqrestore(&ioapic_lock, flags); -} - static void ack_apic_edge(unsigned int irq) { struct irq_desc *desc = irq_to_desc(irq); @@ -2659,6 +2626,39 @@ static void ack_apic_level(unsigned int irq) } #ifdef CONFIG_INTR_REMAP +static void __eoi_ioapic_irq(unsigned int irq, struct irq_cfg *cfg) +{ + int apic, pin; + struct irq_pin_list *entry; + + entry = cfg->irq_2_pin; + for (;;) { + + if (!entry) + break; + + apic = entry->apic; + pin = entry->pin; + io_apic_eoi(apic, pin); + entry = entry->next; + } +} + +static void +eoi_ioapic_irq(struct irq_desc *desc) +{ + struct irq_cfg *cfg; + unsigned long flags; + unsigned int irq; + + irq = desc->irq; + cfg = desc->chip_data; + + spin_lock_irqsave(&ioapic_lock, flags); + __eoi_ioapic_irq(irq, cfg); + spin_unlock_irqrestore(&ioapic_lock, flags); +} + static void ir_ack_apic_edge(unsigned int irq) { ack_APIC_irq(); -- cgit v1.2.3-70-g09d2 From 39d83a5d684a457046aa2a6dac60f105966e78e9 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 20 Apr 2009 13:02:29 -0700 Subject: x86: x2apic, IR: Clean up panic() with nox2apic boot option Instead of panic() ignore the "nox2apic" boot option when BIOS has already enabled x2apic prior to OS handover. [ Impact: printk warning instead of panic() when BIOS has enabled x2apic already ] Signed-off-by: Suresh Siddha Cc: dwmw2@infradead.org Cc: Weidong Han LKML-Reference: <20090420200450.425091000@linux-os.sc.intel.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 2b30e520dce..d32f5589f1d 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -141,8 +141,12 @@ static int x2apic_preenabled; static int disable_x2apic; static __init int setup_nox2apic(char *str) { - if (x2apic_enabled()) - panic("Bios already enabled x2apic, can't enforce nox2apic"); + if (x2apic_enabled()) { + pr_warning("Bios already enabled x2apic, " + "can't enforce nox2apic"); + return 0; + } + disable_x2apic = 1; setup_clear_cpu_cap(X86_FEATURE_X2APIC); return 0; -- cgit v1.2.3-70-g09d2 From ff166cb57a17124af75714a9c11f448f56f1a4a3 Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Mon, 20 Apr 2009 13:02:30 -0700 Subject: x86: x2apic, IR: remove reinit_intr_remapped_IO_APIC() When interrupt-remapping is enabled, we are relying on setup_IO_APIC_irqs() to configure remapped entries in the IO-APIC, which comes little bit later after enabling interrupt-remapping. Meanwhile, restoration of old io-apic entries after enabling interrupt-remapping will not make the interrupts through io-apic functional anyway. So remove the unnecessary reinit_intr_remapped_IO_APIC() step. The longer story: When interrupt-remapping is enabled, IO-APIC entries need to be setup in the re-mappable format (pointing to interrupt-remapping table entries setup by the OS). This remapping configuration is happening in the same place where we traditionally configure IO-APIC (i.e., in setup_IO_APIC_irqs()). So when we enable interrupt-remapping successfully, there is no need to restore old io-apic RTE entries before we actually do a complete configuration shortly in setup_IO_APIC_irqs(). Old IO-APIC RTE's may be in traditional format (non re-mappable) or in re-mappable format pointing to interrupt-remapping table entries setup by BIOS. Restoring both of these will not make IO-APIC functional. We have to rely on setup_IO_APIC_irqs() for proper configuration by OS. So I am removing this unnecessary and broken step. [ Impact: remove unnecessary/broken IO-APIC setup step ] Signed-off-by: Suresh Siddha Acked-by: Weidong Han Cc: dwmw2@infradead.org LKML-Reference: <20090420200450.552359000@linux-os.sc.intel.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io_apic.h | 2 -- arch/x86/kernel/apic/apic.c | 2 -- arch/x86/kernel/apic/io_apic.c | 14 -------------- 3 files changed, 18 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 34eaa37f7ad..1cf145039ee 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -166,8 +166,6 @@ extern void free_ioapic_entries(struct IO_APIC_route_entry **ioapic_entries); extern int save_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); extern void mask_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); extern int restore_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries); -extern void reinit_intr_remapped_IO_APIC(int intr_remapping, - struct IO_APIC_route_entry **ioapic_entries); extern void probe_nr_irqs_gsi(void); diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index d32f5589f1d..1386dbec552 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1412,8 +1412,6 @@ end_restore: * IR enabling failed */ restore_IO_APIC_setup(ioapic_entries); - else - reinit_intr_remapped_IO_APIC(x2apic_preenabled, ioapic_entries); unmask_8259A(); local_irq_restore(flags); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 4baa9cbd630..8aef5f9d947 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -833,20 +833,6 @@ int restore_IO_APIC_setup(struct IO_APIC_route_entry **ioapic_entries) return 0; } -void reinit_intr_remapped_IO_APIC(int intr_remapping, - struct IO_APIC_route_entry **ioapic_entries) - -{ - /* - * for now plain restore of previous settings. - * TBD: In the case of OS enabling interrupt-remapping, - * IO-APIC RTE's need to be setup to point to interrupt-remapping - * table entries. for now, do a plain restore, and wait for - * the setup_IO_APIC_irqs() to do proper initialization. - */ - restore_IO_APIC_setup(ioapic_entries); -} - void free_ioapic_entries(struct IO_APIC_route_entry **ioapic_entries) { int apic; -- cgit v1.2.3-70-g09d2 From b2ba83ff4f4405cebc10884121ee71338a1a6c94 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 26 Apr 2009 23:38:08 -0700 Subject: x86: apic: Remove duplicated macros XAPIC_DEST_* is dupliicated to the one in apicdef.h [ Impact: cleanup ] Signed-off-by: Yinghai Lu LKML-Reference: <49F552D0.5050505@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/summit_32.c | 7 ------- 1 file changed, 7 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/summit_32.c b/arch/x86/kernel/apic/summit_32.c index 9cfe1f415d8..344eee4ac0a 100644 --- a/arch/x86/kernel/apic/summit_32.c +++ b/arch/x86/kernel/apic/summit_32.c @@ -173,13 +173,6 @@ static inline int is_WPEG(struct rio_detail *rio){ rio->type == LookOutAWPEG || rio->type == LookOutBWPEG); } - -/* In clustered mode, the high nibble of APIC ID is a cluster number. - * The low nibble is a 4-bit bitmap. */ -#define XAPIC_DEST_CPUS_SHIFT 4 -#define XAPIC_DEST_CPUS_MASK ((1u << XAPIC_DEST_CPUS_SHIFT) - 1) -#define XAPIC_DEST_CLUSTER_MASK (XAPIC_DEST_CPUS_MASK << XAPIC_DEST_CPUS_SHIFT) - #define SUMMIT_APIC_DFR_VALUE (APIC_DFR_CLUSTER) static const struct cpumask *summit_target_cpus(void) -- cgit v1.2.3-70-g09d2 From e0e42142bab96404de535cceb85d6533d5ad7942 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 26 Apr 2009 23:39:38 -0700 Subject: x86: Use dmi check in apic_is_clustered() on 64-bit to mark the TSC unstable We will have systems with 2 and more sockets 8cores/2thread, but we treat them as multi chassis - while they could have a stable TSC domain. Use DMI check instead. [ Impact: do not turn possibly stable TSCs off incorrectly ] Signed-off-by: Yinghai Lu Cc: Ravikiran Thirumalai LKML-Reference: <49F5532A.5000802@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 86 +++++++++++++++++++++++++++++++-------------- 1 file changed, 59 insertions(+), 27 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 1386dbec552..28f747d61d7 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -2138,31 +2138,14 @@ static void apic_pm_activate(void) { } #endif /* CONFIG_PM */ #ifdef CONFIG_X86_64 -/* - * apic_is_clustered_box() -- Check if we can expect good TSC - * - * Thus far, the major user of this is IBM's Summit2 series: - * - * Clustered boxes may have unsynced TSC problems if they are - * multi-chassis. Use available data to take a good guess. - * If in doubt, go HPET. - */ -__cpuinit int apic_is_clustered_box(void) + +static int __cpuinit apic_cluster_num(void) { int i, clusters, zeros; unsigned id; u16 *bios_cpu_apicid; DECLARE_BITMAP(clustermap, NUM_APIC_CLUSTERS); - /* - * there is not this kind of box with AMD CPU yet. - * Some AMD box with quadcore cpu and 8 sockets apicid - * will be [4, 0x23] or [8, 0x27] could be thought to - * vsmp box still need checking... - */ - if ((boot_cpu_data.x86_vendor == X86_VENDOR_AMD) && !is_vsmp_box()) - return 0; - bios_cpu_apicid = early_per_cpu_ptr(x86_bios_cpu_apicid); bitmap_zero(clustermap, NUM_APIC_CLUSTERS); @@ -2198,18 +2181,67 @@ __cpuinit int apic_is_clustered_box(void) ++zeros; } - /* ScaleMP vSMPowered boxes have one cluster per board and TSCs are - * not guaranteed to be synced between boards - */ - if (is_vsmp_box() && clusters > 1) + return clusters; +} + +static int __cpuinitdata multi_checked; +static int __cpuinitdata multi; + +static int __cpuinit set_multi(const struct dmi_system_id *d) +{ + if (multi) + return 0; + printk(KERN_INFO "APIC: %s detected, Multi Chassis\n", d->ident); + multi = 1; + return 0; +} + +static const __cpuinitconst struct dmi_system_id multi_dmi_table[] = { + { + .callback = set_multi, + .ident = "IBM System Summit2", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "IBM"), + DMI_MATCH(DMI_PRODUCT_NAME, "Summit2"), + }, + }, + {} +}; + +static void __cpuinit dmi_check_multi(void) +{ + if (multi_checked) + return; + + dmi_check_system(multi_dmi_table); + multi_checked = 1; +} + +/* + * apic_is_clustered_box() -- Check if we can expect good TSC + * + * Thus far, the major user of this is IBM's Summit2 series: + * Clustered boxes may have unsynced TSC problems if they are + * multi-chassis. + * Use DMI to check them + */ +__cpuinit int apic_is_clustered_box(void) +{ + dmi_check_multi(); + if (multi) return 1; + if (!is_vsmp_box()) + return 0; + /* - * If clusters > 2, then should be multi-chassis. - * May have to revisit this when multi-core + hyperthreaded CPUs come - * out, but AFAIK this will work even for them. + * ScaleMP vSMPowered boxes have one cluster per board and TSCs are + * not guaranteed to be synced between boards */ - return (clusters > 2); + if (apic_cluster_num() > 1) + return 1; + + return 0; } #endif -- cgit v1.2.3-70-g09d2 From fcef5911c7ea89b80d5bfc727f402f37c9eefd57 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 17:58:23 -0700 Subject: x86/irq: remove leftover code from NUMA_MIGRATE_IRQ_DESC The original feature of migrating irq_desc dynamic was too fragile and was causing problems: it caused crashes on systems with lots of cards with MSI-X when user-space irq-balancer was enabled. We now have new patches that create irq_desc according to device numa node. This patch removes the leftover bits of the dynamic balancer. [ Impact: remove dead code ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell LKML-Reference: <49F654AF.8000808@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 10 ------- arch/x86/configs/x86_64_defconfig | 1 - arch/x86/kernel/apic/io_apic.c | 56 +++------------------------------------ include/linux/irq.h | 10 ------- kernel/irq/Makefile | 2 +- kernel/irq/chip.c | 12 ++------- kernel/irq/handle.c | 9 ++----- kernel/irq/numa_migrate.c | 2 -- 8 files changed, 9 insertions(+), 93 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index df9e885eee1..e1b2543f8ed 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -274,16 +274,6 @@ config SPARSE_IRQ If you don't know what to do here, say N. -config NUMA_MIGRATE_IRQ_DESC - bool "Move irq desc when changing irq smp_affinity" - depends on SPARSE_IRQ && NUMA - depends on BROKEN - default n - ---help--- - This enables moving irq_desc to cpu/node that irq will use handled. - - If you don't know what to do here, say N. - config X86_MPPARSE bool "Enable MPS table" if ACPI default y diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig index 9fe5d212ab4..27b8ce0f590 100644 --- a/arch/x86/configs/x86_64_defconfig +++ b/arch/x86/configs/x86_64_defconfig @@ -195,7 +195,6 @@ CONFIG_HIGH_RES_TIMERS=y CONFIG_GENERIC_CLOCKEVENTS_BUILD=y CONFIG_SMP=y CONFIG_SPARSE_IRQ=y -# CONFIG_NUMA_MIGRATE_IRQ_DESC is not set CONFIG_X86_FIND_SMP_CONFIG=y CONFIG_X86_MPPARSE=y # CONFIG_X86_ELAN is not set diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 30da617d18e..9fbf0f7ec7e 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -148,9 +148,6 @@ struct irq_cfg { unsigned move_cleanup_count; u8 vector; u8 move_in_progress : 1; -#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC - u8 move_desc_pending : 1; -#endif }; /* irq_cfg is indexed by the sum of all RTEs in all I/O APICs. */ @@ -254,8 +251,7 @@ int arch_init_chip_data(struct irq_desc *desc, int cpu) return 0; } -#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC - +/* for move_irq_desc */ static void init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) { @@ -356,19 +352,7 @@ void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc) old_desc->chip_data = NULL; } } - -static void -set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) -{ - struct irq_cfg *cfg = desc->chip_data; - - if (!cfg->move_in_progress) { - /* it means that domain is not changed */ - if (!cpumask_intersects(desc->affinity, mask)) - cfg->move_desc_pending = 1; - } -} -#endif +/* end for move_irq_desc */ #else static struct irq_cfg *irq_cfg(unsigned int irq) @@ -378,13 +362,6 @@ static struct irq_cfg *irq_cfg(unsigned int irq) #endif -#ifndef CONFIG_NUMA_MIGRATE_IRQ_DESC -static inline void -set_extra_move_desc(struct irq_desc *desc, const struct cpumask *mask) -{ -} -#endif - struct io_apic { unsigned int index; unsigned int unused[3]; @@ -592,9 +569,6 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) if (assign_irq_vector(irq, cfg, mask)) return BAD_APICID; - /* check that before desc->addinity get updated */ - set_extra_move_desc(desc, mask); - cpumask_copy(desc->affinity, mask); return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain); @@ -2393,8 +2367,6 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) if (assign_irq_vector(irq, cfg, mask)) return; - set_extra_move_desc(desc, mask); - dest = apic->cpu_mask_to_apicid_and(cfg->domain, mask); irte.vector = cfg->vector; @@ -2491,34 +2463,14 @@ static void irq_complete_move(struct irq_desc **descp) struct irq_cfg *cfg = desc->chip_data; unsigned vector, me; - if (likely(!cfg->move_in_progress)) { -#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC - if (likely(!cfg->move_desc_pending)) - return; - - /* domain has not changed, but affinity did */ - me = smp_processor_id(); - if (cpumask_test_cpu(me, desc->affinity)) { - *descp = desc = move_irq_desc(desc, me); - /* get the new one */ - cfg = desc->chip_data; - cfg->move_desc_pending = 0; - } -#endif + if (likely(!cfg->move_in_progress)) return; - } vector = ~get_irq_regs()->orig_ax; me = smp_processor_id(); - if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) { -#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC - *descp = desc = move_irq_desc(desc, me); - /* get the new one */ - cfg = desc->chip_data; -#endif + if (vector == cfg->vector && cpumask_test_cpu(me, cfg->domain)) send_cleanup_vector(cfg); - } } #else static inline void irq_complete_move(struct irq_desc **descp) {} diff --git a/include/linux/irq.h b/include/linux/irq.h index c4953cf27e5..2a34cd6281d 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -212,16 +212,6 @@ extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu); extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu); -static inline struct irq_desc * -irq_remap_to_desc(unsigned int irq, struct irq_desc *desc) -{ -#ifdef CONFIG_NUMA_MIGRATE_IRQ_DESC - return irq_to_desc(irq); -#else - return desc; -#endif -} - /* * Migration helpers for obsolete names, they will go away: */ diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile index 3394f8f5296..2f065277f8e 100644 --- a/kernel/irq/Makefile +++ b/kernel/irq/Makefile @@ -3,5 +3,5 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o obj-$(CONFIG_PROC_FS) += proc.o obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o -obj-$(CONFIG_NUMA_MIGRATE_IRQ_DESC) += numa_migrate.o +obj-$(CONFIG_SPARSE_IRQ) += numa_migrate.o obj-$(CONFIG_PM_SLEEP) += pm.o diff --git a/kernel/irq/chip.c b/kernel/irq/chip.c index c687ba4363f..13c68e71b72 100644 --- a/kernel/irq/chip.c +++ b/kernel/irq/chip.c @@ -359,7 +359,6 @@ handle_level_irq(unsigned int irq, struct irq_desc *desc) spin_lock(&desc->lock); mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); if (unlikely(desc->status & IRQ_INPROGRESS)) goto out_unlock; @@ -438,7 +437,6 @@ handle_fasteoi_irq(unsigned int irq, struct irq_desc *desc) desc->status &= ~IRQ_INPROGRESS; out: desc->chip->eoi(irq); - desc = irq_remap_to_desc(irq, desc); spin_unlock(&desc->lock); } @@ -475,7 +473,6 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) !desc->action)) { desc->status |= (IRQ_PENDING | IRQ_MASKED); mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); goto out_unlock; } kstat_incr_irqs_this_cpu(irq, desc); @@ -483,7 +480,6 @@ handle_edge_irq(unsigned int irq, struct irq_desc *desc) /* Start handling the irq */ if (desc->chip->ack) desc->chip->ack(irq); - desc = irq_remap_to_desc(irq, desc); /* Mark the IRQ currently in progress.*/ desc->status |= IRQ_INPROGRESS; @@ -544,10 +540,8 @@ handle_percpu_irq(unsigned int irq, struct irq_desc *desc) if (!noirqdebug) note_interrupt(irq, desc, action_ret); - if (desc->chip->eoi) { + if (desc->chip->eoi) desc->chip->eoi(irq); - desc = irq_remap_to_desc(irq, desc); - } } void @@ -582,10 +576,8 @@ __set_irq_handler(unsigned int irq, irq_flow_handler_t handle, int is_chained, /* Uninstall? */ if (handle == handle_bad_irq) { - if (desc->chip != &no_irq_chip) { + if (desc->chip != &no_irq_chip) mask_ack_irq(desc, irq); - desc = irq_remap_to_desc(irq, desc); - } desc->status |= IRQ_DISABLED; desc->depth = 1; } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 882c7980010..3e0cbc44bd7 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -458,11 +458,8 @@ unsigned int __do_IRQ(unsigned int irq) /* * No locking required for CPU-local interrupts: */ - if (desc->chip->ack) { + if (desc->chip->ack) desc->chip->ack(irq); - /* get new one */ - desc = irq_remap_to_desc(irq, desc); - } if (likely(!(desc->status & IRQ_DISABLED))) { action_ret = handle_IRQ_event(irq, desc->action); if (!noirqdebug) @@ -473,10 +470,8 @@ unsigned int __do_IRQ(unsigned int irq) } spin_lock(&desc->lock); - if (desc->chip->ack) { + if (desc->chip->ack) desc->chip->ack(irq); - desc = irq_remap_to_desc(irq, desc); - } /* * REPLAY is when Linux resends an IRQ that was dropped earlier * WAITING is used by probe to mark irqs that are being tested diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index 5760d725162..ce72bc3f4ce 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -97,9 +97,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, /* free the old one */ free_one_irq_desc(old_desc, desc); - spin_unlock(&old_desc->lock); kfree(old_desc); - spin_lock(&desc->lock); return desc; -- cgit v1.2.3-70-g09d2 From d5dedd4507d307eb3f35f21b6e16f336fdc0d82a Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 17:59:21 -0700 Subject: irq: change ->set_affinity() to return status according to Ingo, change set_affinity() in irq_chip should return int, because that way we can handle failure cases in a much cleaner way, in the genirq layer. v2: fix two typos [ Impact: extend API ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell Cc: linux-arch@vger.kernel.org LKML-Reference: <49F654E9.4070809@kernel.org> Signed-off-by: Ingo Molnar --- arch/alpha/kernel/sys_dp264.c | 8 +++-- arch/alpha/kernel/sys_titan.c | 4 ++- arch/arm/common/gic.c | 4 ++- arch/cris/arch-v32/kernel/irq.c | 4 ++- arch/ia64/hp/sim/hpsim_irq.c | 3 +- arch/ia64/kernel/iosapic.c | 10 +++--- arch/ia64/kernel/msi_ia64.c | 16 +++++---- arch/ia64/sn/kernel/irq.c | 4 ++- arch/ia64/sn/kernel/msi_sn.c | 8 +++-- arch/mips/cavium-octeon/octeon-irq.c | 8 +++-- arch/mips/include/asm/irq.h | 2 +- arch/mips/kernel/irq-gic.c | 5 +-- arch/mips/mti-malta/malta-smtc.c | 4 ++- arch/mips/sibyte/bcm1480/irq.c | 8 +++-- arch/mips/sibyte/sb1250/irq.c | 8 +++-- arch/parisc/kernel/irq.c | 6 ++-- arch/powerpc/platforms/pseries/xics.c | 12 ++++--- arch/powerpc/sysdev/mpic.c | 4 ++- arch/sparc/kernel/irq_64.c | 12 +++++-- arch/x86/kernel/apic/io_apic.c | 64 ++++++++++++++++++++++------------- drivers/parisc/iosapic.c | 6 ++-- drivers/xen/events.c | 12 ++++--- include/linux/irq.h | 2 +- 23 files changed, 140 insertions(+), 74 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/alpha/kernel/sys_dp264.c b/arch/alpha/kernel/sys_dp264.c index 9c9d1fd4155..5bd5259324b 100644 --- a/arch/alpha/kernel/sys_dp264.c +++ b/arch/alpha/kernel/sys_dp264.c @@ -176,22 +176,26 @@ cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity) } } -static void +static int dp264_set_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&dp264_irq_lock); cpu_set_irq_affinity(irq, *affinity); tsunami_update_irq_hw(cached_irq_mask); spin_unlock(&dp264_irq_lock); + + return 0; } -static void +static int clipper_set_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&dp264_irq_lock); cpu_set_irq_affinity(irq - 16, *affinity); tsunami_update_irq_hw(cached_irq_mask); spin_unlock(&dp264_irq_lock); + + return 0; } static struct hw_interrupt_type dp264_irq_type = { diff --git a/arch/alpha/kernel/sys_titan.c b/arch/alpha/kernel/sys_titan.c index 27f840a4ad3..8dd239ebdb9 100644 --- a/arch/alpha/kernel/sys_titan.c +++ b/arch/alpha/kernel/sys_titan.c @@ -157,13 +157,15 @@ titan_cpu_set_irq_affinity(unsigned int irq, cpumask_t affinity) } -static void +static int titan_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) { spin_lock(&titan_irq_lock); titan_cpu_set_irq_affinity(irq - 16, *affinity); titan_update_irq_hw(titan_cached_irq_mask); spin_unlock(&titan_irq_lock); + + return 0; } static void diff --git a/arch/arm/common/gic.c b/arch/arm/common/gic.c index c6884ba1d5e..90f6b7f52d4 100644 --- a/arch/arm/common/gic.c +++ b/arch/arm/common/gic.c @@ -109,7 +109,7 @@ static void gic_unmask_irq(unsigned int irq) } #ifdef CONFIG_SMP -static void gic_set_cpu(unsigned int irq, const struct cpumask *mask_val) +static int gic_set_cpu(unsigned int irq, const struct cpumask *mask_val) { void __iomem *reg = gic_dist_base(irq) + GIC_DIST_TARGET + (gic_irq(irq) & ~3); unsigned int shift = (irq % 4) * 8; @@ -122,6 +122,8 @@ static void gic_set_cpu(unsigned int irq, const struct cpumask *mask_val) val |= 1 << (cpu + shift); writel(val, reg); spin_unlock(&irq_controller_lock); + + return 0; } #endif diff --git a/arch/cris/arch-v32/kernel/irq.c b/arch/cris/arch-v32/kernel/irq.c index df3925cb1c7..d70b445f4a8 100644 --- a/arch/cris/arch-v32/kernel/irq.c +++ b/arch/cris/arch-v32/kernel/irq.c @@ -325,12 +325,14 @@ static void end_crisv32_irq(unsigned int irq) { } -void set_affinity_crisv32_irq(unsigned int irq, const struct cpumask *dest) +int set_affinity_crisv32_irq(unsigned int irq, const struct cpumask *dest) { unsigned long flags; spin_lock_irqsave(&irq_lock, flags); irq_allocations[irq - FIRST_IRQ].mask = *dest; spin_unlock_irqrestore(&irq_lock, flags); + + return 0; } static struct irq_chip crisv32_irq_type = { diff --git a/arch/ia64/hp/sim/hpsim_irq.c b/arch/ia64/hp/sim/hpsim_irq.c index cc0a3182db3..acb5047ab57 100644 --- a/arch/ia64/hp/sim/hpsim_irq.c +++ b/arch/ia64/hp/sim/hpsim_irq.c @@ -21,9 +21,10 @@ hpsim_irq_noop (unsigned int irq) { } -static void +static int hpsim_set_affinity_noop(unsigned int a, const struct cpumask *b) { + return 0; } static struct hw_interrupt_type irq_type_hp_sim = { diff --git a/arch/ia64/kernel/iosapic.c b/arch/ia64/kernel/iosapic.c index 166e0d839fa..f92cef47bf8 100644 --- a/arch/ia64/kernel/iosapic.c +++ b/arch/ia64/kernel/iosapic.c @@ -329,7 +329,7 @@ unmask_irq (unsigned int irq) } -static void +static int iosapic_set_affinity(unsigned int irq, const struct cpumask *mask) { #ifdef CONFIG_SMP @@ -343,15 +343,15 @@ iosapic_set_affinity(unsigned int irq, const struct cpumask *mask) cpu = cpumask_first_and(cpu_online_mask, mask); if (cpu >= nr_cpu_ids) - return; + return -1; if (irq_prepare_move(irq, cpu)) - return; + return -1; dest = cpu_physical_id(cpu); if (!iosapic_intr_info[irq].count) - return; /* not an IOSAPIC interrupt */ + return -1; /* not an IOSAPIC interrupt */ set_irq_affinity_info(irq, dest, redir); @@ -376,7 +376,9 @@ iosapic_set_affinity(unsigned int irq, const struct cpumask *mask) iosapic_write(iosapic, IOSAPIC_RTE_HIGH(rte_index), high32); iosapic_write(iosapic, IOSAPIC_RTE_LOW(rte_index), low32); } + #endif + return 0; } /* diff --git a/arch/ia64/kernel/msi_ia64.c b/arch/ia64/kernel/msi_ia64.c index 2b15e233f7f..0f8ade9331b 100644 --- a/arch/ia64/kernel/msi_ia64.c +++ b/arch/ia64/kernel/msi_ia64.c @@ -12,7 +12,7 @@ static struct irq_chip ia64_msi_chip; #ifdef CONFIG_SMP -static void ia64_set_msi_irq_affinity(unsigned int irq, +static int ia64_set_msi_irq_affinity(unsigned int irq, const cpumask_t *cpu_mask) { struct msi_msg msg; @@ -20,10 +20,10 @@ static void ia64_set_msi_irq_affinity(unsigned int irq, int cpu = first_cpu(*cpu_mask); if (!cpu_online(cpu)) - return; + return -1; if (irq_prepare_move(irq, cpu)) - return; + return -1; read_msi_msg(irq, &msg); @@ -39,6 +39,8 @@ static void ia64_set_msi_irq_affinity(unsigned int irq, write_msi_msg(irq, &msg); cpumask_copy(irq_desc[irq].affinity, cpumask_of(cpu)); + + return 0; } #endif /* CONFIG_SMP */ @@ -130,17 +132,17 @@ void arch_teardown_msi_irq(unsigned int irq) #ifdef CONFIG_DMAR #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_cfg *cfg = irq_cfg + irq; struct msi_msg msg; int cpu = cpumask_first(mask); if (!cpu_online(cpu)) - return; + return -1; if (irq_prepare_move(irq, cpu)) - return; + return -1; dmar_msi_read(irq, &msg); @@ -151,6 +153,8 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) dmar_msi_write(irq, &msg); cpumask_copy(irq_desc[irq].affinity, mask); + + return 0; } #endif /* CONFIG_SMP */ diff --git a/arch/ia64/sn/kernel/irq.c b/arch/ia64/sn/kernel/irq.c index 66fd705e82c..764f26abac0 100644 --- a/arch/ia64/sn/kernel/irq.c +++ b/arch/ia64/sn/kernel/irq.c @@ -227,7 +227,7 @@ finish_up: return new_irq_info; } -static void sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask) +static int sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask) { struct sn_irq_info *sn_irq_info, *sn_irq_info_safe; nasid_t nasid; @@ -239,6 +239,8 @@ static void sn_set_affinity_irq(unsigned int irq, const struct cpumask *mask) list_for_each_entry_safe(sn_irq_info, sn_irq_info_safe, sn_irq_lh[irq], list) (void)sn_retarget_vector(sn_irq_info, nasid, slice); + + return 0; } #ifdef CONFIG_SMP diff --git a/arch/ia64/sn/kernel/msi_sn.c b/arch/ia64/sn/kernel/msi_sn.c index 81e428943d7..fbbfb970120 100644 --- a/arch/ia64/sn/kernel/msi_sn.c +++ b/arch/ia64/sn/kernel/msi_sn.c @@ -151,7 +151,7 @@ int sn_setup_msi_irq(struct pci_dev *pdev, struct msi_desc *entry) } #ifdef CONFIG_SMP -static void sn_set_msi_irq_affinity(unsigned int irq, +static int sn_set_msi_irq_affinity(unsigned int irq, const struct cpumask *cpu_mask) { struct msi_msg msg; @@ -168,7 +168,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, cpu = cpumask_first(cpu_mask); sn_irq_info = sn_msi_info[irq].sn_irq_info; if (sn_irq_info == NULL || sn_irq_info->irq_int_bit >= 0) - return; + return -1; /* * Release XIO resources for the old MSI PCI address @@ -189,7 +189,7 @@ static void sn_set_msi_irq_affinity(unsigned int irq, new_irq_info = sn_retarget_vector(sn_irq_info, nasid, slice); sn_msi_info[irq].sn_irq_info = new_irq_info; if (new_irq_info == NULL) - return; + return -1; /* * Map the xio address into bus space @@ -206,6 +206,8 @@ static void sn_set_msi_irq_affinity(unsigned int irq, write_msi_msg(irq, &msg); cpumask_copy(irq_desc[irq].affinity, cpu_mask); + + return 0; } #endif /* CONFIG_SMP */ diff --git a/arch/mips/cavium-octeon/octeon-irq.c b/arch/mips/cavium-octeon/octeon-irq.c index 1c19af8daa6..d3a0c8154be 100644 --- a/arch/mips/cavium-octeon/octeon-irq.c +++ b/arch/mips/cavium-octeon/octeon-irq.c @@ -177,7 +177,7 @@ static void octeon_irq_ciu0_disable(unsigned int irq) } #ifdef CONFIG_SMP -static void octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask *dest) +static int octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask *dest) { int cpu; int bit = irq - OCTEON_IRQ_WORKQ0; /* Bit 0-63 of EN0 */ @@ -199,6 +199,8 @@ static void octeon_irq_ciu0_set_affinity(unsigned int irq, const struct cpumask */ cvmx_read_csr(CVMX_CIU_INTX_EN0(cvmx_get_core_num() * 2)); write_unlock(&octeon_irq_ciu0_rwlock); + + return 0; } #endif @@ -292,7 +294,7 @@ static void octeon_irq_ciu1_disable(unsigned int irq) } #ifdef CONFIG_SMP -static void octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask *dest) +static int octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask *dest) { int cpu; int bit = irq - OCTEON_IRQ_WDOG0; /* Bit 0-63 of EN1 */ @@ -315,6 +317,8 @@ static void octeon_irq_ciu1_set_affinity(unsigned int irq, const struct cpumask */ cvmx_read_csr(CVMX_CIU_INTX_EN1(cvmx_get_core_num() * 2 + 1)); write_unlock(&octeon_irq_ciu1_rwlock); + + return 0; } #endif diff --git a/arch/mips/include/asm/irq.h b/arch/mips/include/asm/irq.h index 3214ade02d1..4f1eed107b0 100644 --- a/arch/mips/include/asm/irq.h +++ b/arch/mips/include/asm/irq.h @@ -49,7 +49,7 @@ static inline void smtc_im_ack_irq(unsigned int irq) #ifdef CONFIG_MIPS_MT_SMTC_IRQAFF #include -extern void plat_set_irq_affinity(unsigned int irq, +extern int plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity); extern void smtc_forward_irq(unsigned int irq); diff --git a/arch/mips/kernel/irq-gic.c b/arch/mips/kernel/irq-gic.c index 87deb8f6c45..3f43c2e3aa5 100644 --- a/arch/mips/kernel/irq-gic.c +++ b/arch/mips/kernel/irq-gic.c @@ -155,7 +155,7 @@ static void gic_unmask_irq(unsigned int irq) static DEFINE_SPINLOCK(gic_lock); -static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask) +static int gic_set_affinity(unsigned int irq, const struct cpumask *cpumask) { cpumask_t tmp = CPU_MASK_NONE; unsigned long flags; @@ -166,7 +166,7 @@ static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask) cpumask_and(&tmp, cpumask, cpu_online_mask); if (cpus_empty(tmp)) - return; + return -1; /* Assumption : cpumask refers to a single CPU */ spin_lock_irqsave(&gic_lock, flags); @@ -190,6 +190,7 @@ static void gic_set_affinity(unsigned int irq, const struct cpumask *cpumask) cpumask_copy(irq_desc[irq].affinity, cpumask); spin_unlock_irqrestore(&gic_lock, flags); + return 0; } #endif diff --git a/arch/mips/mti-malta/malta-smtc.c b/arch/mips/mti-malta/malta-smtc.c index 5ba31888fef..499ffe5475d 100644 --- a/arch/mips/mti-malta/malta-smtc.c +++ b/arch/mips/mti-malta/malta-smtc.c @@ -114,7 +114,7 @@ struct plat_smp_ops msmtc_smp_ops = { */ -void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) +int plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) { cpumask_t tmask; int cpu = 0; @@ -156,5 +156,7 @@ void plat_set_irq_affinity(unsigned int irq, const struct cpumask *affinity) /* Do any generic SMTC IRQ affinity setup */ smtc_set_irq_affinity(irq, tmask); + + return 0; } #endif /* CONFIG_MIPS_MT_SMTC_IRQAFF */ diff --git a/arch/mips/sibyte/bcm1480/irq.c b/arch/mips/sibyte/bcm1480/irq.c index 352352b3cb2..4f256a131bf 100644 --- a/arch/mips/sibyte/bcm1480/irq.c +++ b/arch/mips/sibyte/bcm1480/irq.c @@ -50,7 +50,7 @@ static void enable_bcm1480_irq(unsigned int irq); static void disable_bcm1480_irq(unsigned int irq); static void ack_bcm1480_irq(unsigned int irq); #ifdef CONFIG_SMP -static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask); +static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask); #endif #ifdef CONFIG_PCI @@ -109,7 +109,7 @@ void bcm1480_unmask_irq(int cpu, int irq) } #ifdef CONFIG_SMP -static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) +static int bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) { int i = 0, old_cpu, cpu, int_on, k; u64 cur_ints; @@ -119,7 +119,7 @@ static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) if (cpumask_weight(mask) != 1) { printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq); - return; + return -1; } i = cpumask_first(mask); @@ -155,6 +155,8 @@ static void bcm1480_set_affinity(unsigned int irq, const struct cpumask *mask) } spin_unlock(&bcm1480_imr_lock); spin_unlock_irqrestore(&desc->lock, flags); + + return 0; } #endif diff --git a/arch/mips/sibyte/sb1250/irq.c b/arch/mips/sibyte/sb1250/irq.c index c08ff582da6..e389507f1f9 100644 --- a/arch/mips/sibyte/sb1250/irq.c +++ b/arch/mips/sibyte/sb1250/irq.c @@ -50,7 +50,7 @@ static void enable_sb1250_irq(unsigned int irq); static void disable_sb1250_irq(unsigned int irq); static void ack_sb1250_irq(unsigned int irq); #ifdef CONFIG_SMP -static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask); +static int sb1250_set_affinity(unsigned int irq, const struct cpumask *mask); #endif #ifdef CONFIG_SIBYTE_HAS_LDT @@ -103,7 +103,7 @@ void sb1250_unmask_irq(int cpu, int irq) } #ifdef CONFIG_SMP -static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask) +static int sb1250_set_affinity(unsigned int irq, const struct cpumask *mask) { int i = 0, old_cpu, cpu, int_on; u64 cur_ints; @@ -114,7 +114,7 @@ static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask) if (cpumask_weight(mask) > 1) { printk("attempted to set irq affinity for irq %d to multiple CPUs\n", irq); - return; + return -1; } /* Convert logical CPU to physical CPU */ @@ -146,6 +146,8 @@ static void sb1250_set_affinity(unsigned int irq, const struct cpumask *mask) } spin_unlock(&sb1250_imr_lock); spin_unlock_irqrestore(&desc->lock, flags); + + return 0; } #endif diff --git a/arch/parisc/kernel/irq.c b/arch/parisc/kernel/irq.c index 4ea4229d765..8007f1e6572 100644 --- a/arch/parisc/kernel/irq.c +++ b/arch/parisc/kernel/irq.c @@ -130,15 +130,17 @@ int cpu_check_affinity(unsigned int irq, const struct cpumask *dest) return cpu_dest; } -static void cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest) +static int cpu_set_affinity_irq(unsigned int irq, const struct cpumask *dest) { int cpu_dest; cpu_dest = cpu_check_affinity(irq, dest); if (cpu_dest < 0) - return; + return -1; cpumask_copy(&irq_desc[irq].affinity, dest); + + return 0; } #endif diff --git a/arch/powerpc/platforms/pseries/xics.c b/arch/powerpc/platforms/pseries/xics.c index 80b513449f4..be3581a8c29 100644 --- a/arch/powerpc/platforms/pseries/xics.c +++ b/arch/powerpc/platforms/pseries/xics.c @@ -333,7 +333,7 @@ static void xics_eoi_lpar(unsigned int virq) lpar_xirr_info_set((0xff << 24) | irq); } -static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) +static int xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) { unsigned int irq; int status; @@ -342,14 +342,14 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) irq = (unsigned int)irq_map[virq].hwirq; if (irq == XICS_IPI || irq == XICS_IRQ_SPURIOUS) - return; + return -1; status = rtas_call(ibm_get_xive, 1, 3, xics_status, irq); if (status) { printk(KERN_ERR "%s: ibm,get-xive irq=%u returns %d\n", __func__, irq, status); - return; + return -1; } /* @@ -363,7 +363,7 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) printk(KERN_WARNING "%s: No online cpus in the mask %s for irq %d\n", __func__, cpulist, virq); - return; + return -1; } status = rtas_call(ibm_set_xive, 3, 1, NULL, @@ -372,8 +372,10 @@ static void xics_set_affinity(unsigned int virq, const struct cpumask *cpumask) if (status) { printk(KERN_ERR "%s: ibm,set-xive irq=%u returns %d\n", __func__, irq, status); - return; + return -1; } + + return 0; } static struct irq_chip xics_pic_direct = { diff --git a/arch/powerpc/sysdev/mpic.c b/arch/powerpc/sysdev/mpic.c index 21b95670159..f4cbd15cf22 100644 --- a/arch/powerpc/sysdev/mpic.c +++ b/arch/powerpc/sysdev/mpic.c @@ -807,7 +807,7 @@ static void mpic_end_ipi(unsigned int irq) #endif /* CONFIG_SMP */ -void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask) +int mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask) { struct mpic *mpic = mpic_from_irq(irq); unsigned int src = mpic_irq_to_hw(irq); @@ -824,6 +824,8 @@ void mpic_set_affinity(unsigned int irq, const struct cpumask *cpumask) mpic_irq_write(src, MPIC_INFO(IRQ_DESTINATION), mpic_physmask(cpus_addr(tmp)[0])); } + + return 0; } static unsigned int mpic_type_to_vecpri(struct mpic *mpic, unsigned int type) diff --git a/arch/sparc/kernel/irq_64.c b/arch/sparc/kernel/irq_64.c index 5deabe921a4..e5e78f9cfc9 100644 --- a/arch/sparc/kernel/irq_64.c +++ b/arch/sparc/kernel/irq_64.c @@ -318,10 +318,12 @@ static void sun4u_irq_enable(unsigned int virt_irq) } } -static void sun4u_set_affinity(unsigned int virt_irq, +static int sun4u_set_affinity(unsigned int virt_irq, const struct cpumask *mask) { sun4u_irq_enable(virt_irq); + + return 0; } /* Don't do anything. The desc->status check for IRQ_DISABLED in @@ -377,7 +379,7 @@ static void sun4v_irq_enable(unsigned int virt_irq) ino, err); } -static void sun4v_set_affinity(unsigned int virt_irq, +static int sun4v_set_affinity(unsigned int virt_irq, const struct cpumask *mask) { unsigned int ino = virt_irq_table[virt_irq].dev_ino; @@ -388,6 +390,8 @@ static void sun4v_set_affinity(unsigned int virt_irq, if (err != HV_EOK) printk(KERN_ERR "sun4v_intr_settarget(%x,%lu): " "err(%d)\n", ino, cpuid, err); + + return 0; } static void sun4v_irq_disable(unsigned int virt_irq) @@ -445,7 +449,7 @@ static void sun4v_virq_enable(unsigned int virt_irq) dev_handle, dev_ino, err); } -static void sun4v_virt_set_affinity(unsigned int virt_irq, +static int sun4v_virt_set_affinity(unsigned int virt_irq, const struct cpumask *mask) { unsigned long cpuid, dev_handle, dev_ino; @@ -461,6 +465,8 @@ static void sun4v_virt_set_affinity(unsigned int virt_irq, printk(KERN_ERR "sun4v_vintr_set_target(%lx,%lx,%lu): " "err(%d)\n", dev_handle, dev_ino, cpuid, err); + + return 0; } static void sun4v_virq_disable(unsigned int virt_irq) diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 9fbf0f7ec7e..5c7630b40a5 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -574,13 +574,14 @@ set_desc_affinity(struct irq_desc *desc, const struct cpumask *mask) return apic->cpu_mask_to_apicid_and(desc->affinity, cfg->domain); } -static void +static int set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; unsigned long flags; unsigned int dest; unsigned int irq; + int ret = -1; irq = desc->irq; cfg = desc->chip_data; @@ -591,18 +592,21 @@ set_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) /* Only the high 8 bits are valid. */ dest = SET_APIC_LOGICAL_ID(dest); __target_IO_APIC_irq(irq, dest, cfg); + ret = 0; } spin_unlock_irqrestore(&ioapic_lock, flags); + + return ret; } -static void +static int set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc; desc = irq_to_desc(irq); - set_ioapic_affinity_irq_desc(desc, mask); + return set_ioapic_affinity_irq_desc(desc, mask); } #endif /* CONFIG_SMP */ @@ -2348,24 +2352,25 @@ static int ioapic_retrigger_irq(unsigned int irq) * Real vector that is used for interrupting cpu will be coming from * the interrupt-remapping table entry. */ -static void +static int migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { struct irq_cfg *cfg; struct irte irte; unsigned int dest; unsigned int irq; + int ret = -1; if (!cpumask_intersects(mask, cpu_online_mask)) - return; + return ret; irq = desc->irq; if (get_irte(irq, &irte)) - return; + return ret; cfg = desc->chip_data; if (assign_irq_vector(irq, cfg, mask)) - return; + return ret; dest = apic->cpu_mask_to_apicid_and(cfg->domain, mask); @@ -2381,27 +2386,30 @@ migrate_ioapic_irq_desc(struct irq_desc *desc, const struct cpumask *mask) send_cleanup_vector(cfg); cpumask_copy(desc->affinity, mask); + + return 0; } /* * Migrates the IRQ destination in the process context. */ -static void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, +static int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { - migrate_ioapic_irq_desc(desc, mask); + return migrate_ioapic_irq_desc(desc, mask); } -static void set_ir_ioapic_affinity_irq(unsigned int irq, +static int set_ir_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); - set_ir_ioapic_affinity_irq_desc(desc, mask); + return set_ir_ioapic_affinity_irq_desc(desc, mask); } #else -static inline void set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, +static inline int set_ir_ioapic_affinity_irq_desc(struct irq_desc *desc, const struct cpumask *mask) { + return 0; } #endif @@ -3318,7 +3326,7 @@ static int msi_compose_msg(struct pci_dev *pdev, unsigned int irq, struct msi_ms } #ifdef CONFIG_SMP -static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) +static int set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3327,7 +3335,7 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) dest = set_desc_affinity(desc, mask); if (dest == BAD_APICID) - return; + return -1; cfg = desc->chip_data; @@ -3339,13 +3347,15 @@ static void set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); write_msi_msg_desc(desc, &msg); + + return 0; } #ifdef CONFIG_INTR_REMAP /* * Migrate the MSI irq to another cpumask. This migration is * done in the process context using interrupt-remapping hardware. */ -static void +static int ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); @@ -3354,11 +3364,11 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) struct irte irte; if (get_irte(irq, &irte)) - return; + return -1; dest = set_desc_affinity(desc, mask); if (dest == BAD_APICID) - return; + return -1; irte.vector = cfg->vector; irte.dest_id = IRTE_DEST(dest); @@ -3375,6 +3385,8 @@ ir_set_msi_irq_affinity(unsigned int irq, const struct cpumask *mask) */ if (cfg->move_in_progress) send_cleanup_vector(cfg); + + return 0; } #endif @@ -3528,7 +3540,7 @@ void arch_teardown_msi_irq(unsigned int irq) #if defined (CONFIG_DMAR) || defined (CONFIG_INTR_REMAP) #ifdef CONFIG_SMP -static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static int dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3537,7 +3549,7 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) dest = set_desc_affinity(desc, mask); if (dest == BAD_APICID) - return; + return -1; cfg = desc->chip_data; @@ -3549,6 +3561,8 @@ static void dmar_msi_set_affinity(unsigned int irq, const struct cpumask *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); dmar_msi_write(irq, &msg); + + return 0; } #endif /* CONFIG_SMP */ @@ -3582,7 +3596,7 @@ int arch_setup_dmar_msi(unsigned int irq) #ifdef CONFIG_HPET_TIMER #ifdef CONFIG_SMP -static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) +static int hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3591,7 +3605,7 @@ static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) dest = set_desc_affinity(desc, mask); if (dest == BAD_APICID) - return; + return -1; cfg = desc->chip_data; @@ -3603,6 +3617,8 @@ static void hpet_msi_set_affinity(unsigned int irq, const struct cpumask *mask) msg.address_lo |= MSI_ADDR_DEST_ID(dest); hpet_msi_write(irq, &msg); + + return 0; } #endif /* CONFIG_SMP */ @@ -3659,7 +3675,7 @@ static void target_ht_irq(unsigned int irq, unsigned int dest, u8 vector) write_ht_irq_msg(irq, &msg); } -static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) +static int set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) { struct irq_desc *desc = irq_to_desc(irq); struct irq_cfg *cfg; @@ -3667,11 +3683,13 @@ static void set_ht_irq_affinity(unsigned int irq, const struct cpumask *mask) dest = set_desc_affinity(desc, mask); if (dest == BAD_APICID) - return; + return -1; cfg = desc->chip_data; target_ht_irq(irq, dest, cfg->vector); + + return 0; } #endif diff --git a/drivers/parisc/iosapic.c b/drivers/parisc/iosapic.c index 73348c4047e..4a9cc92d4d1 100644 --- a/drivers/parisc/iosapic.c +++ b/drivers/parisc/iosapic.c @@ -702,7 +702,7 @@ static unsigned int iosapic_startup_irq(unsigned int irq) } #ifdef CONFIG_SMP -static void iosapic_set_affinity_irq(unsigned int irq, +static int iosapic_set_affinity_irq(unsigned int irq, const struct cpumask *dest) { struct vector_info *vi = iosapic_get_vector(irq); @@ -712,7 +712,7 @@ static void iosapic_set_affinity_irq(unsigned int irq, dest_cpu = cpu_check_affinity(irq, dest); if (dest_cpu < 0) - return; + return -1; cpumask_copy(irq_desc[irq].affinity, cpumask_of(dest_cpu)); vi->txn_addr = txn_affinity_addr(irq, dest_cpu); @@ -724,6 +724,8 @@ static void iosapic_set_affinity_irq(unsigned int irq, iosapic_set_irt_data(vi, &dummy_d0, &d1); iosapic_wr_irt_entry(vi, d0, d1); spin_unlock_irqrestore(&iosapic_lock, flags); + + return 0; } #endif diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 30963af5dba..33389880279 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -688,13 +688,13 @@ void rebind_evtchn_irq(int evtchn, int irq) } /* Rebind an evtchn so that it gets delivered to a specific cpu */ -static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu) +static int rebind_irq_to_cpu(unsigned irq, unsigned tcpu) { struct evtchn_bind_vcpu bind_vcpu; int evtchn = evtchn_from_irq(irq); if (!VALID_EVTCHN(evtchn)) - return; + return -1; /* Send future instances of this interrupt to other vcpu. */ bind_vcpu.port = evtchn; @@ -707,13 +707,15 @@ static void rebind_irq_to_cpu(unsigned irq, unsigned tcpu) */ if (HYPERVISOR_event_channel_op(EVTCHNOP_bind_vcpu, &bind_vcpu) >= 0) bind_evtchn_to_cpu(evtchn, tcpu); -} + return 0; +} -static void set_affinity_irq(unsigned irq, const struct cpumask *dest) +static int set_affinity_irq(unsigned irq, const struct cpumask *dest) { unsigned tcpu = cpumask_first(dest); - rebind_irq_to_cpu(irq, tcpu); + + return rebind_irq_to_cpu(irq, tcpu); } int resend_irq_on_evtchn(unsigned int irq) diff --git a/include/linux/irq.h b/include/linux/irq.h index 2a34cd6281d..8e4c18b2915 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -117,7 +117,7 @@ struct irq_chip { void (*eoi)(unsigned int irq); void (*end)(unsigned int irq); - void (*set_affinity)(unsigned int irq, + int (*set_affinity)(unsigned int irq, const struct cpumask *dest); int (*retrigger)(unsigned int irq); int (*set_type)(unsigned int irq, unsigned int flow_type); -- cgit v1.2.3-70-g09d2 From 85ac16d033370caf6f48d743c8dc8103700f5cc5 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 18:00:38 -0700 Subject: x86/irq: change irq_desc_alloc() to take node instead of cpu This simplifies the node awareness of the code. All our allocators only deal with a NUMA node ID locality not with CPU ids anyway - so there's no need to maintain (and transform) a CPU id all across the IRq layer. v2: keep move_irq_desc related [ Impact: cleanup, prepare IRQ code to be NUMA-aware ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell Cc: Jeremy Fitzhardinge LKML-Reference: <49F65536.2020300@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 58 +++++++++++++++++++----------------------- arch/x86/lguest/boot.c | 2 +- drivers/pci/intr_remapping.c | 15 +++++------ drivers/xen/events.c | 2 +- include/linux/interrupt.h | 2 +- include/linux/irq.h | 16 +++++------- kernel/irq/handle.c | 28 ++++++++------------ kernel/irq/internals.h | 2 +- kernel/irq/numa_migrate.c | 36 +++++++++----------------- kernel/softirq.c | 2 +- 10 files changed, 66 insertions(+), 97 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 5c7630b40a5..560b887ba27 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -129,12 +129,9 @@ struct irq_pin_list { struct irq_pin_list *next; }; -static struct irq_pin_list *get_one_free_irq_2_pin(int cpu) +static struct irq_pin_list *get_one_free_irq_2_pin(int node) { struct irq_pin_list *pin; - int node; - - node = cpu_to_node(cpu); pin = kzalloc_node(sizeof(*pin), GFP_ATOMIC, node); @@ -209,12 +206,9 @@ static struct irq_cfg *irq_cfg(unsigned int irq) return cfg; } -static struct irq_cfg *get_one_free_irq_cfg(int cpu) +static struct irq_cfg *get_one_free_irq_cfg(int node) { struct irq_cfg *cfg; - int node; - - node = cpu_to_node(cpu); cfg = kzalloc_node(sizeof(*cfg), GFP_ATOMIC, node); if (cfg) { @@ -235,13 +229,13 @@ static struct irq_cfg *get_one_free_irq_cfg(int cpu) return cfg; } -int arch_init_chip_data(struct irq_desc *desc, int cpu) +int arch_init_chip_data(struct irq_desc *desc, int node) { struct irq_cfg *cfg; cfg = desc->chip_data; if (!cfg) { - desc->chip_data = get_one_free_irq_cfg(cpu); + desc->chip_data = get_one_free_irq_cfg(node); if (!desc->chip_data) { printk(KERN_ERR "can not alloc irq_cfg\n"); BUG_ON(1); @@ -253,7 +247,7 @@ int arch_init_chip_data(struct irq_desc *desc, int cpu) /* for move_irq_desc */ static void -init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) +init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int node) { struct irq_pin_list *old_entry, *head, *tail, *entry; @@ -262,7 +256,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) if (!old_entry) return; - entry = get_one_free_irq_2_pin(cpu); + entry = get_one_free_irq_2_pin(node); if (!entry) return; @@ -272,7 +266,7 @@ init_copy_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg, int cpu) tail = entry; old_entry = old_entry->next; while (old_entry) { - entry = get_one_free_irq_2_pin(cpu); + entry = get_one_free_irq_2_pin(node); if (!entry) { entry = head; while (entry) { @@ -312,12 +306,12 @@ static void free_irq_2_pin(struct irq_cfg *old_cfg, struct irq_cfg *cfg) } void arch_init_copy_chip_data(struct irq_desc *old_desc, - struct irq_desc *desc, int cpu) + struct irq_desc *desc, int node) { struct irq_cfg *cfg; struct irq_cfg *old_cfg; - cfg = get_one_free_irq_cfg(cpu); + cfg = get_one_free_irq_cfg(node); if (!cfg) return; @@ -328,7 +322,7 @@ void arch_init_copy_chip_data(struct irq_desc *old_desc, memcpy(cfg, old_cfg, sizeof(struct irq_cfg)); - init_copy_irq_2_pin(old_cfg, cfg, cpu); + init_copy_irq_2_pin(old_cfg, cfg, node); } static void free_irq_cfg(struct irq_cfg *old_cfg) @@ -615,13 +609,13 @@ set_ioapic_affinity_irq(unsigned int irq, const struct cpumask *mask) * 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_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin) +static void add_pin_to_irq_node(struct irq_cfg *cfg, int node, int apic, int pin) { struct irq_pin_list *entry; entry = cfg->irq_2_pin; if (!entry) { - entry = get_one_free_irq_2_pin(cpu); + entry = get_one_free_irq_2_pin(node); if (!entry) { printk(KERN_ERR "can not alloc irq_2_pin to add %d - %d\n", apic, pin); @@ -641,7 +635,7 @@ static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin) entry = entry->next; } - entry->next = get_one_free_irq_2_pin(cpu); + entry->next = get_one_free_irq_2_pin(node); entry = entry->next; entry->apic = apic; entry->pin = pin; @@ -650,7 +644,7 @@ static void add_pin_to_irq_cpu(struct irq_cfg *cfg, int cpu, int apic, int pin) /* * Reroute an IRQ to a different pin. */ -static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu, +static void __init replace_pin_at_irq_node(struct irq_cfg *cfg, int node, int oldapic, int oldpin, int newapic, int newpin) { @@ -670,7 +664,7 @@ static void __init replace_pin_at_irq_cpu(struct irq_cfg *cfg, int cpu, /* why? call replace before add? */ if (!replaced) - add_pin_to_irq_cpu(cfg, cpu, newapic, newpin); + add_pin_to_irq_node(cfg, node, newapic, newpin); } static inline void io_apic_modify_irq(struct irq_cfg *cfg, @@ -1612,7 +1606,7 @@ static void __init setup_IO_APIC_irqs(void) int notcon = 0; struct irq_desc *desc; struct irq_cfg *cfg; - int cpu = boot_cpu_id; + int node = cpu_to_node(boot_cpu_id); apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); @@ -1647,13 +1641,13 @@ static void __init setup_IO_APIC_irqs(void) apic->multi_timer_check(apic_id, irq)) continue; - desc = irq_to_desc_alloc_cpu(irq, cpu); + desc = irq_to_desc_alloc_node(irq, node); if (!desc) { printk(KERN_INFO "can not get irq_desc for %d\n", irq); continue; } cfg = desc->chip_data; - add_pin_to_irq_cpu(cfg, cpu, apic_id, pin); + add_pin_to_irq_node(cfg, node, apic_id, pin); setup_IO_APIC_irq(apic_id, pin, irq, desc, irq_trigger(idx), irq_polarity(idx)); @@ -2863,7 +2857,7 @@ static inline void __init check_timer(void) { struct irq_desc *desc = irq_to_desc(0); struct irq_cfg *cfg = desc->chip_data; - int cpu = boot_cpu_id; + int node = cpu_to_node(boot_cpu_id); int apic1, pin1, apic2, pin2; unsigned long flags; int no_pin1 = 0; @@ -2929,7 +2923,7 @@ static inline void __init check_timer(void) * Ok, does IRQ0 through the IOAPIC work? */ if (no_pin1) { - add_pin_to_irq_cpu(cfg, cpu, apic1, pin1); + add_pin_to_irq_node(cfg, node, apic1, pin1); setup_timer_IRQ0_pin(apic1, pin1, cfg->vector); } else { /* for edge trigger, setup_IO_APIC_irq already @@ -2966,7 +2960,7 @@ static inline void __init check_timer(void) /* * legacy devices should be connected to IO APIC #0 */ - replace_pin_at_irq_cpu(cfg, cpu, apic1, pin1, apic2, pin2); + replace_pin_at_irq_node(cfg, node, apic1, pin1, apic2, pin2); setup_timer_IRQ0_pin(apic2, pin2, cfg->vector); enable_8259A_irq(0); if (timer_irq_works()) { @@ -3185,7 +3179,7 @@ unsigned int create_irq_nr(unsigned int irq_want) unsigned int new; unsigned long flags; struct irq_cfg *cfg_new = NULL; - int cpu = boot_cpu_id; + int node = cpu_to_node(boot_cpu_id); struct irq_desc *desc_new = NULL; irq = 0; @@ -3194,7 +3188,7 @@ unsigned int create_irq_nr(unsigned int irq_want) spin_lock_irqsave(&vector_lock, flags); for (new = irq_want; new < nr_irqs; new++) { - desc_new = irq_to_desc_alloc_cpu(new, cpu); + desc_new = irq_to_desc_alloc_node(new, node); if (!desc_new) { printk(KERN_INFO "can not get irq_desc for %d\n", new); continue; @@ -3968,7 +3962,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p { struct irq_desc *desc; struct irq_cfg *cfg; - int cpu = boot_cpu_id; + int node = cpu_to_node(boot_cpu_id); if (!IO_APIC_IRQ(irq)) { apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", @@ -3976,7 +3970,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p return -EINVAL; } - desc = irq_to_desc_alloc_cpu(irq, cpu); + desc = irq_to_desc_alloc_node(irq, node); if (!desc) { printk(KERN_INFO "can not get irq_desc %d\n", irq); return 0; @@ -3987,7 +3981,7 @@ int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int p */ if (irq >= NR_IRQS_LEGACY) { cfg = desc->chip_data; - add_pin_to_irq_cpu(cfg, cpu, ioapic, pin); + add_pin_to_irq_node(cfg, node, ioapic, pin); } setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c index ca7ec44bafc..45acbcf2568 100644 --- a/arch/x86/lguest/boot.c +++ b/arch/x86/lguest/boot.c @@ -636,7 +636,7 @@ static void __init lguest_init_IRQ(void) void lguest_setup_irq(unsigned int irq) { - irq_to_desc_alloc_cpu(irq, 0); + irq_to_desc_alloc_node(irq, 0); set_irq_chip_and_handler_name(irq, &lguest_irq_controller, handle_level_irq, "level"); } diff --git a/drivers/pci/intr_remapping.c b/drivers/pci/intr_remapping.c index f5e0ea724a6..9eff36a293e 100644 --- a/drivers/pci/intr_remapping.c +++ b/drivers/pci/intr_remapping.c @@ -23,15 +23,12 @@ struct irq_2_iommu { }; #ifdef CONFIG_GENERIC_HARDIRQS -static struct irq_2_iommu *get_one_free_irq_2_iommu(int cpu) +static struct irq_2_iommu *get_one_free_irq_2_iommu(int node) { struct irq_2_iommu *iommu; - int node; - - node = cpu_to_node(cpu); iommu = kzalloc_node(sizeof(*iommu), GFP_ATOMIC, node); - printk(KERN_DEBUG "alloc irq_2_iommu on cpu %d node %d\n", cpu, node); + printk(KERN_DEBUG "alloc irq_2_iommu on node %d\n", node); return iommu; } @@ -48,7 +45,7 @@ static struct irq_2_iommu *irq_2_iommu(unsigned int irq) return desc->irq_2_iommu; } -static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu) +static struct irq_2_iommu *irq_2_iommu_alloc_node(unsigned int irq, int node) { struct irq_desc *desc; struct irq_2_iommu *irq_iommu; @@ -56,7 +53,7 @@ static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu) /* * alloc irq desc if not allocated already. */ - desc = irq_to_desc_alloc_cpu(irq, cpu); + desc = irq_to_desc_alloc_node(irq, node); if (!desc) { printk(KERN_INFO "can not get irq_desc for %d\n", irq); return NULL; @@ -65,14 +62,14 @@ static struct irq_2_iommu *irq_2_iommu_alloc_cpu(unsigned int irq, int cpu) irq_iommu = desc->irq_2_iommu; if (!irq_iommu) - desc->irq_2_iommu = get_one_free_irq_2_iommu(cpu); + desc->irq_2_iommu = get_one_free_irq_2_iommu(node); return desc->irq_2_iommu; } static struct irq_2_iommu *irq_2_iommu_alloc(unsigned int irq) { - return irq_2_iommu_alloc_cpu(irq, boot_cpu_id); + return irq_2_iommu_alloc_node(irq, cpu_to_node(boot_cpu_id)); } #else /* !CONFIG_SPARSE_IRQ */ diff --git a/drivers/xen/events.c b/drivers/xen/events.c index 33389880279..be437c2bc94 100644 --- a/drivers/xen/events.c +++ b/drivers/xen/events.c @@ -335,7 +335,7 @@ static int find_unbound_irq(void) if (irq == nr_irqs) panic("No available IRQ to bind to: increase nr_irqs!\n"); - desc = irq_to_desc_alloc_cpu(irq, 0); + desc = irq_to_desc_alloc_node(irq, 0); if (WARN_ON(desc == NULL)) return -1; diff --git a/include/linux/interrupt.h b/include/linux/interrupt.h index 91bb76f44f1..ff374ceface 100644 --- a/include/linux/interrupt.h +++ b/include/linux/interrupt.h @@ -566,6 +566,6 @@ struct irq_desc; extern int early_irq_init(void); extern int arch_probe_nr_irqs(void); extern int arch_early_irq_init(void); -extern int arch_init_chip_data(struct irq_desc *desc, int cpu); +extern int arch_init_chip_data(struct irq_desc *desc, int node); #endif diff --git a/include/linux/irq.h b/include/linux/irq.h index 8e4c18b2915..a09baf8f9d9 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -187,7 +187,7 @@ struct irq_desc { spinlock_t lock; #ifdef CONFIG_SMP cpumask_var_t affinity; - unsigned int cpu; + unsigned int node; #ifdef CONFIG_GENERIC_PENDING_IRQ cpumask_var_t pending_mask; #endif @@ -201,16 +201,16 @@ struct irq_desc { } ____cacheline_internodealigned_in_smp; extern void arch_init_copy_chip_data(struct irq_desc *old_desc, - struct irq_desc *desc, int cpu); + struct irq_desc *desc, int node); extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc); #ifndef CONFIG_SPARSE_IRQ extern struct irq_desc irq_desc[NR_IRQS]; #else /* CONFIG_SPARSE_IRQ */ -extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int cpu); +extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int node); #endif /* CONFIG_SPARSE_IRQ */ -extern struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu); +extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node); /* * Migration helpers for obsolete names, they will go away: @@ -422,12 +422,10 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); * Allocates affinity and pending_mask cpumask if required. * Returns true if successful (or not required). */ -static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu, +static inline bool alloc_desc_masks(struct irq_desc *desc, int node, bool boot) { #ifdef CONFIG_CPUMASK_OFFSTACK - int node; - if (boot) { alloc_bootmem_cpumask_var(&desc->affinity); @@ -437,8 +435,6 @@ static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu, return true; } - node = cpu_to_node(cpu); - if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node)) return false; @@ -494,7 +490,7 @@ static inline void free_desc_masks(struct irq_desc *old_desc, #else /* !CONFIG_SMP */ -static inline bool alloc_desc_masks(struct irq_desc *desc, int cpu, +static inline bool alloc_desc_masks(struct irq_desc *desc, int node, bool boot) { return true; diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index 3e0cbc44bd7..a6368db2618 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -81,12 +81,10 @@ static struct irq_desc irq_desc_init = { .lock = __SPIN_LOCK_UNLOCKED(irq_desc_init.lock), }; -void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) +void init_kstat_irqs(struct irq_desc *desc, int node, int nr) { - int node; void *ptr; - node = cpu_to_node(cpu); ptr = kzalloc_node(nr * sizeof(*desc->kstat_irqs), GFP_ATOMIC, node); /* @@ -94,33 +92,32 @@ void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr) * init_copy_kstat_irqs() could still use old one */ if (ptr) { - printk(KERN_DEBUG " alloc kstat_irqs on cpu %d node %d\n", - cpu, node); + printk(KERN_DEBUG " alloc kstat_irqs on node %d\n", node); desc->kstat_irqs = ptr; } } -static void init_one_irq_desc(int irq, struct irq_desc *desc, int cpu) +static void init_one_irq_desc(int irq, struct irq_desc *desc, int node) { memcpy(desc, &irq_desc_init, sizeof(struct irq_desc)); spin_lock_init(&desc->lock); desc->irq = irq; #ifdef CONFIG_SMP - desc->cpu = cpu; + desc->node = node; #endif lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_kstat_irqs(desc, cpu, nr_cpu_ids); + init_kstat_irqs(desc, node, nr_cpu_ids); if (!desc->kstat_irqs) { printk(KERN_ERR "can not alloc kstat_irqs\n"); BUG_ON(1); } - if (!alloc_desc_masks(desc, cpu, false)) { + if (!alloc_desc_masks(desc, node, false)) { printk(KERN_ERR "can not alloc irq_desc cpumasks\n"); BUG_ON(1); } init_desc_masks(desc); - arch_init_chip_data(desc, cpu); + arch_init_chip_data(desc, node); } /* @@ -189,11 +186,10 @@ struct irq_desc *irq_to_desc(unsigned int irq) return NULL; } -struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node) { struct irq_desc *desc; unsigned long flags; - int node; if (irq >= nr_irqs) { WARN(1, "irq (%d) >= nr_irqs (%d) in irq_to_desc_alloc\n", @@ -212,15 +208,13 @@ struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) if (desc) goto out_unlock; - node = cpu_to_node(cpu); desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); - printk(KERN_DEBUG " alloc irq_desc for %d on cpu %d node %d\n", - irq, cpu, node); + printk(KERN_DEBUG " alloc irq_desc for %d on node %d\n", irq, node); if (!desc) { printk(KERN_ERR "can not alloc irq_desc\n"); BUG_ON(1); } - init_one_irq_desc(irq, desc, cpu); + init_one_irq_desc(irq, desc, node); irq_desc_ptrs[irq] = desc; @@ -270,7 +264,7 @@ struct irq_desc *irq_to_desc(unsigned int irq) return (irq < NR_IRQS) ? irq_desc + irq : NULL; } -struct irq_desc *irq_to_desc_alloc_cpu(unsigned int irq, int cpu) +struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node) { return irq_to_desc(irq); } diff --git a/kernel/irq/internals.h b/kernel/irq/internals.h index de5f412f6a9..73468253143 100644 --- a/kernel/irq/internals.h +++ b/kernel/irq/internals.h @@ -16,7 +16,7 @@ extern void __disable_irq(struct irq_desc *desc, unsigned int irq, bool susp); extern void __enable_irq(struct irq_desc *desc, unsigned int irq, bool resume); extern struct lock_class_key irq_desc_lock_class; -extern void init_kstat_irqs(struct irq_desc *desc, int cpu, int nr); +extern void init_kstat_irqs(struct irq_desc *desc, int node, int nr); extern void clear_kstat_irqs(struct irq_desc *desc); extern spinlock_t sparse_irq_lock; diff --git a/kernel/irq/numa_migrate.c b/kernel/irq/numa_migrate.c index ce72bc3f4ce..2f69bee57bf 100644 --- a/kernel/irq/numa_migrate.c +++ b/kernel/irq/numa_migrate.c @@ -15,9 +15,9 @@ static void init_copy_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc, - int cpu, int nr) + int node, int nr) { - init_kstat_irqs(desc, cpu, nr); + init_kstat_irqs(desc, node, nr); if (desc->kstat_irqs != old_desc->kstat_irqs) memcpy(desc->kstat_irqs, old_desc->kstat_irqs, @@ -34,20 +34,20 @@ static void free_kstat_irqs(struct irq_desc *old_desc, struct irq_desc *desc) } static bool init_copy_one_irq_desc(int irq, struct irq_desc *old_desc, - struct irq_desc *desc, int cpu) + struct irq_desc *desc, int node) { memcpy(desc, old_desc, sizeof(struct irq_desc)); - if (!alloc_desc_masks(desc, cpu, false)) { + if (!alloc_desc_masks(desc, node, false)) { printk(KERN_ERR "irq %d: can not get new irq_desc cpumask " "for migration.\n", irq); return false; } spin_lock_init(&desc->lock); - desc->cpu = cpu; + desc->node = node; lockdep_set_class(&desc->lock, &irq_desc_lock_class); - init_copy_kstat_irqs(old_desc, desc, cpu, nr_cpu_ids); + init_copy_kstat_irqs(old_desc, desc, node, nr_cpu_ids); init_copy_desc_masks(old_desc, desc); - arch_init_copy_chip_data(old_desc, desc, cpu); + arch_init_copy_chip_data(old_desc, desc, node); return true; } @@ -59,12 +59,11 @@ static void free_one_irq_desc(struct irq_desc *old_desc, struct irq_desc *desc) } static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, - int cpu) + int node) { struct irq_desc *desc; unsigned int irq; unsigned long flags; - int node; irq = old_desc->irq; @@ -76,7 +75,6 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, if (desc && old_desc != desc) goto out_unlock; - node = cpu_to_node(cpu); desc = kzalloc_node(sizeof(*desc), GFP_ATOMIC, node); if (!desc) { printk(KERN_ERR "irq %d: can not get new irq_desc " @@ -85,7 +83,7 @@ static struct irq_desc *__real_move_irq_desc(struct irq_desc *old_desc, desc = old_desc; goto out_unlock; } - if (!init_copy_one_irq_desc(irq, old_desc, desc, cpu)) { + if (!init_copy_one_irq_desc(irq, old_desc, desc, node)) { /* still use old one */ kfree(desc); desc = old_desc; @@ -107,24 +105,14 @@ out_unlock: return desc; } -struct irq_desc *move_irq_desc(struct irq_desc *desc, int cpu) +struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) { - int old_cpu; - int node, old_node; - /* those all static, do move them */ if (desc->irq < NR_IRQS_LEGACY) return desc; - old_cpu = desc->cpu; - if (old_cpu != cpu) { - node = cpu_to_node(cpu); - old_node = cpu_to_node(old_cpu); - if (old_node != node) - desc = __real_move_irq_desc(desc, cpu); - else - desc->cpu = cpu; - } + if (desc->node != node) + desc = __real_move_irq_desc(desc, node); return desc; } diff --git a/kernel/softirq.c b/kernel/softirq.c index b525dd34851..f674f332a02 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -828,7 +828,7 @@ int __init __weak arch_early_irq_init(void) return 0; } -int __weak arch_init_chip_data(struct irq_desc *desc, int cpu) +int __weak arch_init_chip_data(struct irq_desc *desc, int node) { return 0; } -- cgit v1.2.3-70-g09d2 From a2f809b08ae4dddc1015c7dcd8659e5729e45b3e Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 18:01:20 -0700 Subject: irq: change ACPI GSI APIs to also take a device argument We want to use dev_to_node() later on, to be aware of the 'home node' of the GSI in question. [ Impact: cleanup, prepare the IRQ code to be more NUMA aware ] Signed-off-by: Yinghai Lu Acked-by: Len Brown Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell Cc: Len Brown Cc: Bjorn Helgaas Cc: Tony Luck Cc: linux-acpi@vger.kernel.org Cc: linux-ia64@vger.kernel.org LKML-Reference: <49F65560.20904@kernel.org> Signed-off-by: Ingo Molnar --- arch/ia64/kernel/acpi.c | 5 +++-- arch/x86/include/asm/io_apic.h | 4 ++-- arch/x86/include/asm/mpspec.h | 4 +++- arch/x86/kernel/acpi/boot.c | 8 ++++---- arch/x86/kernel/apic/io_apic.c | 3 ++- drivers/acpi/pci_irq.c | 5 +++-- drivers/char/hpet.c | 4 ++-- drivers/pnp/pnpacpi/rsparser.c | 2 +- include/linux/acpi.h | 2 +- 9 files changed, 21 insertions(+), 16 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c index 5510317db37..baec6f00f7f 100644 --- a/arch/ia64/kernel/acpi.c +++ b/arch/ia64/kernel/acpi.c @@ -636,7 +636,7 @@ void __init acpi_numa_arch_fixup(void) * success: return IRQ number (>=0) * failure: return < 0 */ -int acpi_register_gsi(u32 gsi, int triggering, int polarity) +int acpi_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) { if (acpi_irq_model == ACPI_IRQ_MODEL_PLATFORM) return gsi; @@ -678,7 +678,8 @@ static int __init acpi_parse_fadt(struct acpi_table_header *table) fadt = (struct acpi_table_fadt *)fadt_header; - acpi_register_gsi(fadt->sci_interrupt, ACPI_LEVEL_SENSITIVE, ACPI_ACTIVE_LOW); + acpi_register_gsi(NULL, fadt->sci_interrupt, ACPI_LEVEL_SENSITIVE, + ACPI_ACTIVE_LOW); return 0; } diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 9d826e43601..07f2913ba5d 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -154,8 +154,8 @@ extern int timer_through_8259; extern int io_apic_get_unique_id(int ioapic, int apic_id); extern int io_apic_get_version(int ioapic); extern int io_apic_get_redir_entries(int ioapic); -extern int io_apic_set_pci_routing(int ioapic, int pin, int irq, - int edge_level, int active_high_low); +extern int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, + int irq, int edge_level, int active_high_low); #endif /* CONFIG_ACPI */ extern int (*ioapic_renumber_irq)(int ioapic, int irq); diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index 642fc7fc8cd..3ea1f531f53 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -72,7 +72,9 @@ extern void mp_register_ioapic(int id, u32 address, u32 gsi_base); extern void mp_override_legacy_irq(u8 bus_irq, u8 polarity, u8 trigger, u32 gsi); extern void mp_config_acpi_legacy_irqs(void); -extern int mp_register_gsi(u32 gsi, int edge_level, int active_high_low); +struct device; +extern int mp_register_gsi(struct device *dev, u32 gsi, int edge_level, + int active_high_low); extern int acpi_probe_gsi(void); #ifdef CONFIG_X86_IO_APIC extern int mp_config_acpi_gsi(unsigned char number, unsigned int devfn, u8 pin, diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 723989d7f80..6ee96b5530f 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -522,7 +522,7 @@ int acpi_gsi_to_irq(u32 gsi, unsigned int *irq) * success: return IRQ number (>=0) * failure: return < 0 */ -int acpi_register_gsi(u32 gsi, int triggering, int polarity) +int acpi_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) { unsigned int irq; unsigned int plat_gsi = gsi; @@ -539,7 +539,7 @@ int acpi_register_gsi(u32 gsi, int triggering, int polarity) #ifdef CONFIG_X86_IO_APIC if (acpi_irq_model == ACPI_IRQ_MODEL_IOAPIC) { - plat_gsi = mp_register_gsi(gsi, triggering, polarity); + plat_gsi = mp_register_gsi(dev, gsi, triggering, polarity); } #endif acpi_gsi_to_irq(plat_gsi, &irq); @@ -1158,7 +1158,7 @@ void __init mp_config_acpi_legacy_irqs(void) } } -int mp_register_gsi(u32 gsi, int triggering, int polarity) +int mp_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) { int ioapic; int ioapic_pin; @@ -1253,7 +1253,7 @@ int mp_register_gsi(u32 gsi, int triggering, int polarity) } } #endif - io_apic_set_pci_routing(ioapic, ioapic_pin, gsi, + io_apic_set_pci_routing(dev, ioapic, ioapic_pin, gsi, triggering == ACPI_EDGE_SENSITIVE ? 0 : 1, polarity == ACPI_ACTIVE_HIGH ? 0 : 1); return gsi; diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 560b887ba27..d9346622601 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3958,7 +3958,8 @@ int __init io_apic_get_version(int ioapic) } #endif -int io_apic_set_pci_routing (int ioapic, int pin, int irq, int triggering, int polarity) +int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, + int triggering, int polarity) { struct irq_desc *desc; struct irq_cfg *cfg; diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c index 51b9f8280f8..2faa9e2ac89 100644 --- a/drivers/acpi/pci_irq.c +++ b/drivers/acpi/pci_irq.c @@ -401,7 +401,8 @@ int acpi_pci_irq_enable(struct pci_dev *dev) /* Interrupt Line values above 0xF are forbidden */ if (dev->irq > 0 && (dev->irq <= 0xF)) { printk(" - using IRQ %d\n", dev->irq); - acpi_register_gsi(dev->irq, ACPI_LEVEL_SENSITIVE, + acpi_register_gsi(&dev->dev, dev->irq, + ACPI_LEVEL_SENSITIVE, ACPI_ACTIVE_LOW); return 0; } else { @@ -410,7 +411,7 @@ int acpi_pci_irq_enable(struct pci_dev *dev) } } - rc = acpi_register_gsi(gsi, triggering, polarity); + rc = acpi_register_gsi(&dev->dev, gsi, triggering, polarity); if (rc < 0) { dev_warn(&dev->dev, "PCI INT %c: failed to register GSI\n", pin_name(pin)); diff --git a/drivers/char/hpet.c b/drivers/char/hpet.c index 340ba4f9dc5..4a9f3492b92 100644 --- a/drivers/char/hpet.c +++ b/drivers/char/hpet.c @@ -224,7 +224,7 @@ static void hpet_timer_set_irq(struct hpet_dev *devp) break; } - gsi = acpi_register_gsi(irq, ACPI_LEVEL_SENSITIVE, + gsi = acpi_register_gsi(NULL, irq, ACPI_LEVEL_SENSITIVE, ACPI_ACTIVE_LOW); if (gsi > 0) break; @@ -939,7 +939,7 @@ static acpi_status hpet_resources(struct acpi_resource *res, void *data) irqp = &res->data.extended_irq; for (i = 0; i < irqp->interrupt_count; i++) { - irq = acpi_register_gsi(irqp->interrupts[i], + irq = acpi_register_gsi(NULL, irqp->interrupts[i], irqp->triggering, irqp->polarity); if (irq < 0) return AE_ERROR; diff --git a/drivers/pnp/pnpacpi/rsparser.c b/drivers/pnp/pnpacpi/rsparser.c index adf17856bac..7f207f335be 100644 --- a/drivers/pnp/pnpacpi/rsparser.c +++ b/drivers/pnp/pnpacpi/rsparser.c @@ -123,7 +123,7 @@ static void pnpacpi_parse_allocated_irqresource(struct pnp_dev *dev, } flags = irq_flags(triggering, polarity, shareable); - irq = acpi_register_gsi(gsi, triggering, polarity); + irq = acpi_register_gsi(&dev->dev, gsi, triggering, polarity); if (irq >= 0) pcibios_penalize_isa_irq(irq, 1); else diff --git a/include/linux/acpi.h b/include/linux/acpi.h index 88be890ee3c..51b4b0a5ce8 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -119,7 +119,7 @@ extern int pci_mmcfg_config_num; extern int sbf_port; extern unsigned long acpi_realmode_flags; -int acpi_register_gsi (u32 gsi, int triggering, int polarity); +int acpi_register_gsi (struct device *dev, u32 gsi, int triggering, int polarity); int acpi_gsi_to_irq (u32 gsi, unsigned int *irq); #ifdef CONFIG_X86_IO_APIC -- cgit v1.2.3-70-g09d2 From 024154cfdd802654cb236a18c78b6e37351e2c49 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 18:01:50 -0700 Subject: irq: change io_apic_set_pci_routing() to use device parameter Make actual use of the device parameter passed down to io_apic_set_pci_routing() - to have the IRQ descriptor on the home node of the device. If no device has been passed down, we assume it's a platform device and use the boot node ID for the IRQ descriptor. [ Impact: optimization, make IO-APIC code more NUMA aware ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell LKML-Reference: <49F6557E.3080101@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index d9346622601..82376e021b5 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3963,7 +3963,7 @@ int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, { struct irq_desc *desc; struct irq_cfg *cfg; - int node = cpu_to_node(boot_cpu_id); + int node; if (!IO_APIC_IRQ(irq)) { apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", @@ -3971,6 +3971,11 @@ int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, return -EINVAL; } + if (dev) + node = dev_to_node(dev); + else + node = cpu_to_node(boot_cpu_id); + desc = irq_to_desc_alloc_node(irq, node); if (!desc) { printk(KERN_INFO "can not get irq_desc %d\n", irq); -- cgit v1.2.3-70-g09d2 From d047f53a2ecce37e3bdf79eac5a326fbaadb3628 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 27 Apr 2009 18:02:23 -0700 Subject: x86/irq: change MSI irq_desc to be more numa aware Try to get irq_desc on the home node in create_irq_nr(). v2: don't check if we can move it when sparse_irq is not used v3: use move_irq_des, if that node is not what we want [ Impact: optimization, make MSI IRQ descriptors more NUMA aware ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell LKML-Reference: <49F6559F.7070005@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 17 +++++++++++++---- include/linux/irq.h | 2 +- 2 files changed, 14 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 82376e021b5..9cd4806cdf5 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3172,14 +3172,13 @@ static int nr_irqs_gsi = NR_IRQS_LEGACY; /* * Dynamic irq allocate and deallocation */ -unsigned int create_irq_nr(unsigned int irq_want) +unsigned int create_irq_nr(unsigned int irq_want, int node) { /* Allocate an unused irq */ unsigned int irq; unsigned int new; unsigned long flags; struct irq_cfg *cfg_new = NULL; - int node = cpu_to_node(boot_cpu_id); struct irq_desc *desc_new = NULL; irq = 0; @@ -3197,6 +3196,13 @@ unsigned int create_irq_nr(unsigned int irq_want) if (cfg_new->vector != 0) continue; + +#ifdef CONFIG_NUMA_IRQ_DESC + /* different node ?*/ + if (desc_new->node != node) + desc = move_irq_desc(desc, node); +#endif + if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0) irq = new; break; @@ -3214,11 +3220,12 @@ unsigned int create_irq_nr(unsigned int irq_want) int create_irq(void) { + int node = cpu_to_node(boot_cpu_id); unsigned int irq_want; int irq; irq_want = nr_irqs_gsi; - irq = create_irq_nr(irq_want); + irq = create_irq_nr(irq_want, node); if (irq == 0) irq = -1; @@ -3476,15 +3483,17 @@ int arch_setup_msi_irqs(struct pci_dev *dev, int nvec, int type) unsigned int irq_want; struct intel_iommu *iommu = NULL; int index = 0; + int node; /* x86 doesn't support multiple MSI yet */ if (type == PCI_CAP_ID_MSI && nvec > 1) return 1; + node = dev_to_node(&dev->dev); irq_want = nr_irqs_gsi; sub_handle = 0; list_for_each_entry(msidesc, &dev->msi_list, list) { - irq = create_irq_nr(irq_want); + irq = create_irq_nr(irq_want, node); if (irq == 0) return -1; irq_want = irq + 1; diff --git a/include/linux/irq.h b/include/linux/irq.h index a09baf8f9d9..4b95ddb5304 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -376,7 +376,7 @@ extern void set_irq_noprobe(unsigned int irq); extern void set_irq_probe(unsigned int irq); /* Handle dynamic irq creation and destruction */ -extern unsigned int create_irq_nr(unsigned int irq_want); +extern unsigned int create_irq_nr(unsigned int irq_want, int node); extern int create_irq(void); extern void destroy_irq(unsigned int irq); -- cgit v1.2.3-70-g09d2 From 15e957d08dd4a841359cfec59ecb74041e0097aa Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 30 Apr 2009 01:17:50 -0700 Subject: x86/irq: use move_irq_desc() in create_irq_nr() move_irq_desc() will try to move irq_desc to the home node if the allocated one is not correct, in create_irq_nr(). ( This can happen on devices that are on different nodes that are using MSI, when drivers are loaded and unloaded randomly. ) v2: fix non-smp build v3: add NUMA_IRQ_DESC to eliminate #ifdefs [ Impact: improve irq descriptor locality on NUMA systems ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Suresh Siddha Cc: "Eric W. Biederman" Cc: Rusty Russell LKML-Reference: <49F95EAE.2050903@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/Kconfig | 4 ++++ arch/x86/kernel/apic/io_apic.c | 6 +----- include/linux/irq.h | 11 +++++++++-- kernel/irq/Makefile | 2 +- 4 files changed, 15 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index e1b2543f8ed..674e21e9f0a 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -274,6 +274,10 @@ config SPARSE_IRQ If you don't know what to do here, say N. +config NUMA_IRQ_DESC + def_bool y + depends on SPARSE_IRQ && NUMA + config X86_MPPARSE bool "Enable MPS table" if ACPI default y diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 9cd4806cdf5..e583291fe6c 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3197,11 +3197,7 @@ unsigned int create_irq_nr(unsigned int irq_want, int node) if (cfg_new->vector != 0) continue; -#ifdef CONFIG_NUMA_IRQ_DESC - /* different node ?*/ - if (desc_new->node != node) - desc = move_irq_desc(desc, node); -#endif + desc_new = move_irq_desc(desc_new, node); if (__assign_irq_vector(new, cfg_new, apic->target_cpus()) == 0) irq = new; diff --git a/include/linux/irq.h b/include/linux/irq.h index 4b95ddb5304..eedbb8e5e0c 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -206,9 +206,16 @@ extern void arch_free_chip_data(struct irq_desc *old_desc, struct irq_desc *desc #ifndef CONFIG_SPARSE_IRQ extern struct irq_desc irq_desc[NR_IRQS]; -#else /* CONFIG_SPARSE_IRQ */ +#endif + +#ifdef CONFIG_NUMA_IRQ_DESC extern struct irq_desc *move_irq_desc(struct irq_desc *old_desc, int node); -#endif /* CONFIG_SPARSE_IRQ */ +#else +static inline struct irq_desc *move_irq_desc(struct irq_desc *desc, int node) +{ + return desc; +} +#endif extern struct irq_desc *irq_to_desc_alloc_node(unsigned int irq, int node); diff --git a/kernel/irq/Makefile b/kernel/irq/Makefile index 2f065277f8e..7d047808419 100644 --- a/kernel/irq/Makefile +++ b/kernel/irq/Makefile @@ -3,5 +3,5 @@ obj-y := handle.o manage.o spurious.o resend.o chip.o devres.o obj-$(CONFIG_GENERIC_IRQ_PROBE) += autoprobe.o obj-$(CONFIG_PROC_FS) += proc.o obj-$(CONFIG_GENERIC_PENDING_IRQ) += migration.o -obj-$(CONFIG_SPARSE_IRQ) += numa_migrate.o +obj-$(CONFIG_NUMA_IRQ_DESC) += numa_migrate.o obj-$(CONFIG_PM_SLEEP) += pm.o -- cgit v1.2.3-70-g09d2 From 6f0aced639d346e5f54eea9fcb2784b633493d09 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Fri, 1 May 2009 23:54:25 +0400 Subject: x86, apic: use pr_ macro Replace recenly appeared printk with pr_ macro (the file already use a lot of them). [ Impact: cleanup ] Signed-off-by: Cyrill Gorcunov LKML-Reference: <20090501195425.GB4633@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 28f747d61d7..e258bedce7c 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -2191,7 +2191,7 @@ static int __cpuinit set_multi(const struct dmi_system_id *d) { if (multi) return 0; - printk(KERN_INFO "APIC: %s detected, Multi Chassis\n", d->ident); + pr_info("APIC: %s detected, Multi Chassis\n", d->ident); multi = 1; return 0; } -- cgit v1.2.3-70-g09d2 From 1cbac972ba28e706fa9ce4d4c81830040bc811ee Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sat, 2 May 2009 13:39:56 +0400 Subject: x86: uv io-apic - use BUILD_BUG_ON instead of BUG_ON The expression is known to be true/false at compilation time so we're allowed to use build-time instead of run-time check. Also align 'entry' items assignment. [ Impact: shrink kernel a bit, cleanup ] Signed-off-by: Cyrill Gorcunov Cc: Jack Steiner LKML-Reference: <20090502093956.GB4791@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 8aef5f9d947..a80335ba12c 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3749,6 +3749,8 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, unsigned long flags; int err; + BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); + cfg = irq_cfg(irq); err = assign_irq_vector(irq, cfg, eligible_cpu); @@ -3762,15 +3764,13 @@ int arch_enable_uv_irq(char *irq_name, unsigned int irq, int cpu, int mmr_blade, mmr_value = 0; entry = (struct uv_IO_APIC_route_entry *)&mmr_value; - BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); - - entry->vector = cfg->vector; - entry->delivery_mode = apic->irq_delivery_mode; - entry->dest_mode = apic->irq_dest_mode; - entry->polarity = 0; - entry->trigger = 0; - entry->mask = 0; - entry->dest = apic->cpu_mask_to_apicid(eligible_cpu); + entry->vector = cfg->vector; + entry->delivery_mode = apic->irq_delivery_mode; + entry->dest_mode = apic->irq_dest_mode; + entry->polarity = 0; + entry->trigger = 0; + entry->mask = 0; + entry->dest = apic->cpu_mask_to_apicid(eligible_cpu); mmr_pnode = uv_blade_to_pnode(mmr_blade); uv_write_global_mmr64(mmr_pnode, mmr_offset, mmr_value); @@ -3788,10 +3788,10 @@ void arch_disable_uv_irq(int mmr_blade, unsigned long mmr_offset) struct uv_IO_APIC_route_entry *entry; int mmr_pnode; + BUILD_BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); + mmr_value = 0; entry = (struct uv_IO_APIC_route_entry *)&mmr_value; - BUG_ON(sizeof(struct uv_IO_APIC_route_entry) != sizeof(unsigned long)); - entry->mask = 1; mmr_pnode = uv_blade_to_pnode(mmr_blade); -- cgit v1.2.3-70-g09d2 From 9a8709d44139748fe2e0ab56d20d8c384c8b65ad Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sat, 2 May 2009 00:25:11 +0400 Subject: x86: uv - prevent NULL dereference in uv_system_init() We may reach NULL dereference oops if kmalloc failed. Prevent it with explicit BUG_ON. [ Impact: more controlled assert in 'impossible' scenario ] Signed-off-by: Cyrill Gorcunov Acked-by: Jack Steiner LKML-Reference: <20090501202511.GE4633@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/x2apic_uv_x.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 873bf7121e8..9d9e2281a82 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -569,15 +569,18 @@ void __init uv_system_init(void) bytes = sizeof(struct uv_blade_info) * uv_num_possible_blades(); uv_blade_info = kmalloc(bytes, GFP_KERNEL); + BUG_ON(!uv_blade_info); get_lowmem_redirect(&lowmem_redir_base, &lowmem_redir_size); bytes = sizeof(uv_node_to_blade[0]) * num_possible_nodes(); uv_node_to_blade = kmalloc(bytes, GFP_KERNEL); + BUG_ON(!uv_node_to_blade); memset(uv_node_to_blade, 255, bytes); bytes = sizeof(uv_cpu_to_blade[0]) * num_possible_cpus(); uv_cpu_to_blade = kmalloc(bytes, GFP_KERNEL); + BUG_ON(!uv_cpu_to_blade); memset(uv_cpu_to_blade, 255, bytes); blade = 0; -- cgit v1.2.3-70-g09d2 From b74d446f1f337e3fe906169a3266cb65ffa4179e Mon Sep 17 00:00:00 2001 From: Sam Ravnborg Date: Sat, 9 May 2009 15:35:10 +0600 Subject: x86: Fix false positive section mismatch warnings in the apic code [ Impact: reduce kernel image size a bit, annotate away warnings ] Signed-off-by: Sam Ravnborg [ modified and tested it ] Signed-off-by: Rakib Mullick Cc: Marcin Slusarz LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/es7000_32.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/es7000_32.c b/arch/x86/kernel/apic/es7000_32.c index 1c11b819f24..30294777557 100644 --- a/arch/x86/kernel/apic/es7000_32.c +++ b/arch/x86/kernel/apic/es7000_32.c @@ -254,7 +254,7 @@ static int parse_unisys_oem(char *oemptr) } #ifdef CONFIG_ACPI -static int find_unisys_acpi_oem_table(unsigned long *oem_addr) +static int __init find_unisys_acpi_oem_table(unsigned long *oem_addr) { struct acpi_table_header *header = NULL; struct es7000_oem_table *table; @@ -285,7 +285,7 @@ static int find_unisys_acpi_oem_table(unsigned long *oem_addr) return 0; } -static void unmap_unisys_acpi_oem_table(unsigned long oem_addr) +static void __init unmap_unisys_acpi_oem_table(unsigned long oem_addr) { if (!oem_addr) return; @@ -306,7 +306,7 @@ static int es7000_check_dsdt(void) static int es7000_acpi_ret; /* Hook from generic ACPI tables.c */ -static int es7000_acpi_madt_oem_check(char *oem_id, char *oem_table_id) +static int __init es7000_acpi_madt_oem_check(char *oem_id, char *oem_table_id) { unsigned long oem_addr = 0; int check_dsdt; @@ -717,7 +717,7 @@ struct apic apic_es7000_cluster = { .safe_wait_icr_idle = native_safe_apic_wait_icr_idle, }; -struct apic apic_es7000 = { +struct apic __refdata apic_es7000 = { .name = "es7000", .probe = probe_es7000, -- cgit v1.2.3-70-g09d2 From bdfe8ac153546537ed24de69610ea781a734f785 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 6 May 2009 10:07:41 -0700 Subject: x86/acpi: move pin_programmed bit map to io_apic.c Prepare to call setup_io_apic_routing() in pcibios_irq_enable() also remove not needed member apic_id. [ Impact: clean up, prepare for future change ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Jesse Barnes Cc: Len Brown LKML-Reference: <4A01C3DD.3050104@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/acpi/boot.c | 18 ++---------------- arch/x86/kernel/apic/io_apic.c | 25 ++++++++++++++++++++++++- 2 files changed, 26 insertions(+), 17 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 8019ecf66e9..dcfbc3ab9e4 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -904,10 +904,8 @@ extern int es7000_plat; #endif static struct { - int apic_id; int gsi_base; int gsi_end; - DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); } mp_ioapic_routing[MAX_IO_APICS]; int mp_find_ioapic(int gsi) @@ -996,7 +994,6 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base) * 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].apicid; mp_ioapic_routing[idx].gsi_base = gsi_base; mp_ioapic_routing[idx].gsi_end = gsi_base + io_apic_get_redir_entries(idx); @@ -1189,7 +1186,7 @@ static int mp_config_acpi_gsi(struct device *dev, u32 gsi, int triggering, mp_irq.srcbus = number; mp_irq.srcbusirq = (((devfn >> 3) & 0x1f) << 2) | ((pin - 1) & 3); ioapic = mp_find_ioapic(gsi); - mp_irq.dstapic = mp_ioapic_routing[ioapic].apic_id; + mp_irq.dstapic = mp_ioapics[ioapic].apicid; mp_irq.dstirq = mp_find_ioapic_pin(ioapic, gsi); save_mp_irq(&mp_irq); @@ -1224,23 +1221,12 @@ int mp_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) if (ioapic_pin > MP_MAX_IOAPIC_PIN) { printk(KERN_ERR "Invalid reference to IOAPIC pin " - "%d-%d\n", mp_ioapic_routing[ioapic].apic_id, + "%d-%d\n", mp_ioapics[ioapic].apicid, ioapic_pin); return gsi; } mp_config_acpi_gsi(dev, gsi, triggering, polarity); - /* - * Avoid pin reprogramming. PRTs typically include entries - * with redundant pin->gsi mappings (but unique PCI devices); - * we only program the IOAPIC on the first. - */ - if (test_bit(ioapic_pin, mp_ioapic_routing[ioapic].pin_programmed)) { - pr_debug("Pin %d-%d already programmed\n", - mp_ioapic_routing[ioapic].apic_id, ioapic_pin); - return gsi; - } - set_bit(ioapic_pin, mp_ioapic_routing[ioapic].pin_programmed); io_apic_set_pci_routing(dev, ioapic, ioapic_pin, gsi, triggering == ACPI_EDGE_SENSITIVE ? 0 : 1, polarity == ACPI_ACTIVE_HIGH ? 0 : 1); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 21c30e1121e..e279ae33928 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3922,7 +3922,7 @@ int __init io_apic_get_version(int ioapic) } #endif -int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, +static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, int triggering, int polarity) { struct irq_desc *desc; @@ -3959,6 +3959,29 @@ int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, return 0; } +static struct { + DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); +} mp_ioapic_routing[MAX_IO_APICS]; + +int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, + int triggering, int polarity) +{ + + /* + * Avoid pin reprogramming. PRTs typically include entries + * with redundant pin->gsi mappings (but unique PCI devices); + * we only program the IOAPIC on the first. + */ + if (test_bit(pin, mp_ioapic_routing[ioapic].pin_programmed)) { + pr_debug("Pin %d-%d already programmed\n", + mp_ioapics[ioapic].apicid, pin); + return 0; + } + set_bit(pin, mp_ioapic_routing[ioapic].pin_programmed); + + return __io_apic_set_pci_routing(dev, ioapic, pin, irq, + triggering, polarity); +} int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) { -- cgit v1.2.3-70-g09d2 From e20c06fd6950265a899edd96a02dc2e6ae2d1ce5 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 6 May 2009 10:08:22 -0700 Subject: x86/pci: add 4 more return parameters to IO_APIC_get_PCI_irq_vector() To prepare those params for pcibios_irq_enable() to call setup_io_apic_routing(). [ Impact: extend function call API to prepare for new functionality ] Signed-off-by: Yinghai Lu Acked-by: Jesse Barnes Cc: Len Brown Cc: Andrew Morton LKML-Reference: <4A01C406.2040303@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/hw_irq.h | 4 +- arch/x86/kernel/apic/io_apic.c | 107 +++++++++++++++++++++----------------- arch/x86/pci/irq.c | 24 +++++++-- drivers/pci/hotplug/ibmphp_core.c | 56 +++++++++++--------- 4 files changed, 112 insertions(+), 79 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index b762ea49bd7..26a40ab7013 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -63,7 +63,9 @@ extern unsigned long io_apic_irqs; extern void init_VISWS_APIC_irqs(void); extern void setup_IO_APIC(void); extern void disable_IO_APIC(void); -extern int IO_APIC_get_PCI_irq_vector(int bus, int slot, int fn); +extern int IO_APIC_get_PCI_irq_vector(int bus, int devfn, int pin, + int *ioapic, int *ioapic_pin, + int *trigger, int *polarity); extern void setup_ioapic_dest(void); extern void enable_IO_APIC(void); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index e279ae33928..caf9dbdde05 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -873,54 +873,6 @@ static int __init find_isa_irq_apic(int irq, int type) 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 (test_bit(bus, mp_bus_not_pci)) { - apic_printk(APIC_VERBOSE, "PCI BIOS passed nonexistent PCI bus %d!\n", bus); - return -1; - } - for (i = 0; i < mp_irq_entries; i++) { - int lbus = mp_irqs[i].srcbus; - - for (apic = 0; apic < nr_ioapics; apic++) - if (mp_ioapics[apic].apicid == mp_irqs[i].dstapic || - mp_irqs[i].dstapic == MP_APIC_ALL) - break; - - if (!test_bit(lbus, mp_bus_not_pci) && - !mp_irqs[i].irqtype && - (bus == lbus) && - (slot == ((mp_irqs[i].srcbusirq >> 2) & 0x1f))) { - int irq = pin_2_irq(i, apic, mp_irqs[i].dstirq); - - if (!(apic || IO_APIC_IRQ(irq))) - continue; - - if (pin == (mp_irqs[i].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; -} - -EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); - #if defined(CONFIG_EISA) || defined(CONFIG_MCA) /* * EISA Edge/Level control register, ELCR @@ -1139,6 +1091,65 @@ static int pin_2_irq(int idx, int apic, int pin) return irq; } +/* + * Find a specific PCI IRQ entry. + * Not an __init, possibly needed by modules + */ +int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin, + int *ioapic, int *ioapic_pin, + int *trigger, int *polarity) +{ + 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 (test_bit(bus, mp_bus_not_pci)) { + apic_printk(APIC_VERBOSE, + "PCI BIOS passed nonexistent PCI bus %d!\n", bus); + return -1; + } + for (i = 0; i < mp_irq_entries; i++) { + int lbus = mp_irqs[i].srcbus; + + for (apic = 0; apic < nr_ioapics; apic++) + if (mp_ioapics[apic].apicid == mp_irqs[i].dstapic || + mp_irqs[i].dstapic == MP_APIC_ALL) + break; + + if (!test_bit(lbus, mp_bus_not_pci) && + !mp_irqs[i].irqtype && + (bus == lbus) && + (slot == ((mp_irqs[i].srcbusirq >> 2) & 0x1f))) { + int irq = pin_2_irq(i, apic, mp_irqs[i].dstirq); + + if (!(apic || IO_APIC_IRQ(irq))) + continue; + + if (pin == (mp_irqs[i].srcbusirq & 3)) { + *ioapic = apic; + *ioapic_pin = mp_irqs[i].dstirq; + *trigger = irq_trigger(i); + *polarity = irq_polarity(i); + return irq; + } + /* + * Use the first all-but-pin matching entry as a + * best-guess fuzzy result for broken mptables. + */ + if (best_guess < 0) { + *ioapic = apic; + *ioapic_pin = mp_irqs[i].dstirq; + *trigger = irq_trigger(i); + *polarity = irq_polarity(i); + best_guess = irq; + } + } + } + return best_guess; +} +EXPORT_SYMBOL(IO_APIC_get_PCI_irq_vector); + void lock_vector_lock(void) { /* Used to the online set of cpus does not change diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index fecbce6e7d7..a2f6bde9c4e 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -1051,12 +1051,16 @@ static void __init pcibios_fixup_irqs(void) */ if (io_apic_assign_pci_irqs) { int irq; + int ioapic = -1, ioapic_pin = -1; + int triggering, polarity; /* * interrupt pins are numbered starting from 1 */ irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, - PCI_SLOT(dev->devfn), pin - 1); + PCI_SLOT(dev->devfn), pin - 1, + &ioapic, &ioapic_pin, + &triggering, &polarity); /* * Busses behind bridges are typically not listed in the * MP-table. In this case we have to look up the IRQ @@ -1072,7 +1076,10 @@ static void __init pcibios_fixup_irqs(void) pin = pci_swizzle_interrupt_pin(dev, pin); bus = bridge->bus->number; irq = IO_APIC_get_PCI_irq_vector(bus, - PCI_SLOT(bridge->devfn), pin - 1); + PCI_SLOT(bridge->devfn), + pin - 1, + &ioapic, &ioapic_pin, + &triggering, &polarity); if (irq >= 0) dev_warn(&dev->dev, "using bridge %s INT %c to " @@ -1221,8 +1228,14 @@ static int pirq_enable_irq(struct pci_dev *dev) if (io_apic_assign_pci_irqs) { int irq; + int ioapic = -1, ioapic_pin = -1; + int triggering, polarity; - irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, PCI_SLOT(dev->devfn), pin - 1); + irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, + PCI_SLOT(dev->devfn), + pin - 1, + &ioapic, &ioapic_pin, + &triggering, &polarity); /* * Busses behind bridges are typically not listed in the MP-table. * In this case we have to look up the IRQ based on the parent bus, @@ -1235,7 +1248,10 @@ static int pirq_enable_irq(struct pci_dev *dev) pin = pci_swizzle_interrupt_pin(dev, pin); irq = IO_APIC_get_PCI_irq_vector(bridge->bus->number, - PCI_SLOT(bridge->devfn), pin - 1); + PCI_SLOT(bridge->devfn), + pin - 1, + &ioapic, &ioapic_pin, + &triggering, &polarity); if (irq >= 0) dev_warn(&dev->dev, "using bridge %s " "INT %c to get IRQ %d\n", diff --git a/drivers/pci/hotplug/ibmphp_core.c b/drivers/pci/hotplug/ibmphp_core.c index dd18f857dfb..ef53b05a411 100644 --- a/drivers/pci/hotplug/ibmphp_core.c +++ b/drivers/pci/hotplug/ibmphp_core.c @@ -153,45 +153,49 @@ int ibmphp_init_devno(struct slot **cur_slot) return -1; } for (loop = 0; loop < len; loop++) { - if ((*cur_slot)->number == rtable->slots[loop].slot) { - if ((*cur_slot)->bus == rtable->slots[loop].bus) { + if ((*cur_slot)->number == rtable->slots[loop].slot && + (*cur_slot)->bus == rtable->slots[loop].bus) { + int ioapic = -1, ioapic_pin = -1; + int triggering, polarity; + (*cur_slot)->device = PCI_SLOT(rtable->slots[loop].devfn); for (i = 0; i < 4; i++) (*cur_slot)->irq[i] = IO_APIC_get_PCI_irq_vector((int) (*cur_slot)->bus, - (int) (*cur_slot)->device, i); - - debug("(*cur_slot)->irq[0] = %x\n", - (*cur_slot)->irq[0]); - debug("(*cur_slot)->irq[1] = %x\n", - (*cur_slot)->irq[1]); - debug("(*cur_slot)->irq[2] = %x\n", - (*cur_slot)->irq[2]); - debug("(*cur_slot)->irq[3] = %x\n", - (*cur_slot)->irq[3]); - - debug("rtable->exlusive_irqs = %x\n", + (int) (*cur_slot)->device, i. + &ioapic, &ioapic_pin, + &triggering, &polarity); + + debug("(*cur_slot)->irq[0] = %x\n", + (*cur_slot)->irq[0]); + debug("(*cur_slot)->irq[1] = %x\n", + (*cur_slot)->irq[1]); + debug("(*cur_slot)->irq[2] = %x\n", + (*cur_slot)->irq[2]); + debug("(*cur_slot)->irq[3] = %x\n", + (*cur_slot)->irq[3]); + + debug("rtable->exlusive_irqs = %x\n", rtable->exclusive_irqs); - debug("rtable->slots[loop].irq[0].bitmap = %x\n", + debug("rtable->slots[loop].irq[0].bitmap = %x\n", rtable->slots[loop].irq[0].bitmap); - debug("rtable->slots[loop].irq[1].bitmap = %x\n", + debug("rtable->slots[loop].irq[1].bitmap = %x\n", rtable->slots[loop].irq[1].bitmap); - debug("rtable->slots[loop].irq[2].bitmap = %x\n", + debug("rtable->slots[loop].irq[2].bitmap = %x\n", rtable->slots[loop].irq[2].bitmap); - debug("rtable->slots[loop].irq[3].bitmap = %x\n", + debug("rtable->slots[loop].irq[3].bitmap = %x\n", rtable->slots[loop].irq[3].bitmap); - debug("rtable->slots[loop].irq[0].link = %x\n", + debug("rtable->slots[loop].irq[0].link = %x\n", rtable->slots[loop].irq[0].link); - debug("rtable->slots[loop].irq[1].link = %x\n", + debug("rtable->slots[loop].irq[1].link = %x\n", rtable->slots[loop].irq[1].link); - debug("rtable->slots[loop].irq[2].link = %x\n", + debug("rtable->slots[loop].irq[2].link = %x\n", rtable->slots[loop].irq[2].link); - debug("rtable->slots[loop].irq[3].link = %x\n", + debug("rtable->slots[loop].irq[3].link = %x\n", rtable->slots[loop].irq[3].link); - debug("end of init_devno\n"); - kfree(rtable); - return 0; - } + debug("end of init_devno\n"); + kfree(rtable); + return 0; } } -- cgit v1.2.3-70-g09d2 From 5ef2183768bb7d64b85eccbfa1537a61cbefa97c Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 6 May 2009 10:08:50 -0700 Subject: x86/acpi: move setup io apic routing out of CONFIG_ACPI scope So we could set io apic routing when ACPI is not enabled. [ Impact: prepare for new functionality ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Jesse Barnes Cc: Len Brown LKML-Reference: <4A01C422.5070400@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/io_apic.h | 4 +- arch/x86/kernel/apic/io_apic.c | 122 ++++++++++++++++++++--------------------- 2 files changed, 63 insertions(+), 63 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 27bd2fdd00a..6fd99f96eb0 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -154,10 +154,10 @@ extern int timer_through_8259; extern int io_apic_get_unique_id(int ioapic, int apic_id); extern int io_apic_get_version(int ioapic); extern int io_apic_get_redir_entries(int ioapic); -extern int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, - int irq, int edge_level, int active_high_low); #endif /* CONFIG_ACPI */ +extern int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, + int irq, int edge_level, int active_high_low); extern int (*ioapic_renumber_irq)(int ioapic, int irq); extern void ioapic_init_mappings(void); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index caf9dbdde05..3a68daee0d9 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -3839,6 +3839,67 @@ int __init arch_probe_nr_irqs(void) } #endif +static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, + int triggering, int polarity) +{ + struct irq_desc *desc; + struct irq_cfg *cfg; + int node; + + if (!IO_APIC_IRQ(irq)) { + apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", + ioapic); + return -EINVAL; + } + + if (dev) + node = dev_to_node(dev); + else + node = cpu_to_node(boot_cpu_id); + + desc = irq_to_desc_alloc_node(irq, node); + if (!desc) { + printk(KERN_INFO "can not get irq_desc %d\n", irq); + return 0; + } + + /* + * IRQs < 16 are already in the irq_2_pin[] map + */ + if (irq >= NR_IRQS_LEGACY) { + cfg = desc->chip_data; + add_pin_to_irq_node(cfg, node, ioapic, pin); + } + + setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); + + return 0; +} + +static struct { + DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); +} mp_ioapic_routing[MAX_IO_APICS]; + +int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, + int triggering, int polarity) +{ + + /* + * Avoid pin reprogramming. PRTs typically include entries + * with redundant pin->gsi mappings (but unique PCI devices); + * we only program the IOAPIC on the first. + */ + if (test_bit(pin, mp_ioapic_routing[ioapic].pin_programmed)) { + pr_debug("Pin %d-%d already programmed\n", + mp_ioapics[ioapic].apicid, pin); + return 0; + } + set_bit(pin, mp_ioapic_routing[ioapic].pin_programmed); + + return __io_apic_set_pci_routing(dev, ioapic, pin, irq, + triggering, polarity); +} + /* -------------------------------------------------------------------------- ACPI-based IOAPIC Configuration -------------------------------------------------------------------------- */ @@ -3933,67 +3994,6 @@ int __init io_apic_get_version(int ioapic) } #endif -static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, - int triggering, int polarity) -{ - struct irq_desc *desc; - struct irq_cfg *cfg; - int node; - - if (!IO_APIC_IRQ(irq)) { - apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", - ioapic); - return -EINVAL; - } - - if (dev) - node = dev_to_node(dev); - else - node = cpu_to_node(boot_cpu_id); - - desc = irq_to_desc_alloc_node(irq, node); - if (!desc) { - printk(KERN_INFO "can not get irq_desc %d\n", irq); - return 0; - } - - /* - * IRQs < 16 are already in the irq_2_pin[] map - */ - if (irq >= NR_IRQS_LEGACY) { - cfg = desc->chip_data; - add_pin_to_irq_node(cfg, node, ioapic, pin); - } - - setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); - - return 0; -} - -static struct { - DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); -} mp_ioapic_routing[MAX_IO_APICS]; - -int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, - int triggering, int polarity) -{ - - /* - * Avoid pin reprogramming. PRTs typically include entries - * with redundant pin->gsi mappings (but unique PCI devices); - * we only program the IOAPIC on the first. - */ - if (test_bit(pin, mp_ioapic_routing[ioapic].pin_programmed)) { - pr_debug("Pin %d-%d already programmed\n", - mp_ioapics[ioapic].apicid, pin); - return 0; - } - set_bit(pin, mp_ioapic_routing[ioapic].pin_programmed); - - return __io_apic_set_pci_routing(dev, ioapic, pin, irq, - triggering, polarity); -} - int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) { int i; -- cgit v1.2.3-70-g09d2 From b9c61b70075c87a8612624736faf4a2de5b1ed30 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Wed, 6 May 2009 10:10:06 -0700 Subject: x86/pci: update pirq_enable_irq() to setup io apic routing So we can set io apic routing only when enabling the device irq. This is advantageous for IRQ descriptor allocation affinity: if we set up the IO-APIC entry later, we have a chance to allocate the IRQ descriptor later and know which device it is on and can set affinity accordingly. [ Impact: standardize/enhance irq-enabling sequence for mptable irqs ] Signed-off-by: Yinghai Lu Acked-by: Jesse Barnes Cc: Len Brown Cc: Andrew Morton LKML-Reference: <4A01C46E.8000501@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 150 ++++++++++++++++++++--------------------- arch/x86/pci/irq.c | 84 +++++++++-------------- 2 files changed, 104 insertions(+), 130 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 3a68daee0d9..5d5f4120c74 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1480,9 +1480,13 @@ static void setup_IO_APIC_irq(int apic_id, int pin, unsigned int irq, struct irq ioapic_write_entry(apic_id, pin, entry); } +static struct { + DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); +} mp_ioapic_routing[MAX_IO_APICS]; + static void __init setup_IO_APIC_irqs(void) { - int apic_id, pin, idx, irq; + int apic_id = 0, pin, idx, irq; int notcon = 0; struct irq_desc *desc; struct irq_cfg *cfg; @@ -1490,48 +1494,53 @@ static void __init setup_IO_APIC_irqs(void) apic_printk(APIC_VERBOSE, KERN_DEBUG "init IO_APIC IRQs\n"); - for (apic_id = 0; apic_id < nr_ioapics; apic_id++) { - for (pin = 0; pin < nr_ioapic_registers[apic_id]; pin++) { - - idx = find_irq_entry(apic_id, pin, mp_INT); - if (idx == -1) { - if (!notcon) { - notcon = 1; - apic_printk(APIC_VERBOSE, - KERN_DEBUG " %d-%d", - mp_ioapics[apic_id].apicid, pin); - } else - apic_printk(APIC_VERBOSE, " %d-%d", - mp_ioapics[apic_id].apicid, pin); - continue; - } - if (notcon) { - apic_printk(APIC_VERBOSE, - " (apicid-pin) not connected\n"); - notcon = 0; - } +#ifdef CONFIG_ACPI + if (!acpi_disabled && acpi_ioapic) { + apic_id = mp_find_ioapic(0); + if (apic_id < 0) + apic_id = 0; + } +#endif - irq = pin_2_irq(idx, apic_id, pin); + for (pin = 0; pin < nr_ioapic_registers[apic_id]; pin++) { + idx = find_irq_entry(apic_id, pin, mp_INT); + if (idx == -1) { + if (!notcon) { + notcon = 1; + apic_printk(APIC_VERBOSE, + KERN_DEBUG " %d-%d", + mp_ioapics[apic_id].apicid, pin); + } else + apic_printk(APIC_VERBOSE, " %d-%d", + mp_ioapics[apic_id].apicid, pin); + continue; + } + if (notcon) { + apic_printk(APIC_VERBOSE, + " (apicid-pin) not connected\n"); + notcon = 0; + } - /* - * Skip the timer IRQ if there's a quirk handler - * installed and if it returns 1: - */ - if (apic->multi_timer_check && - apic->multi_timer_check(apic_id, irq)) - continue; + irq = pin_2_irq(idx, apic_id, pin); - desc = irq_to_desc_alloc_node(irq, node); - if (!desc) { - printk(KERN_INFO "can not get irq_desc for %d\n", irq); - continue; - } - cfg = desc->chip_data; - add_pin_to_irq_node(cfg, node, apic_id, pin); + /* + * Skip the timer IRQ if there's a quirk handler + * installed and if it returns 1: + */ + if (apic->multi_timer_check && + apic->multi_timer_check(apic_id, irq)) + continue; - setup_IO_APIC_irq(apic_id, pin, irq, desc, - irq_trigger(idx), irq_polarity(idx)); + desc = irq_to_desc_alloc_node(irq, node); + if (!desc) { + printk(KERN_INFO "can not get irq_desc for %d\n", irq); + continue; } + cfg = desc->chip_data; + add_pin_to_irq_node(cfg, node, apic_id, pin); + set_bit(pin, mp_ioapic_routing[apic_id].pin_programmed); + setup_IO_APIC_irq(apic_id, pin, irq, desc, + irq_trigger(idx), irq_polarity(idx)); } if (notcon) @@ -3876,10 +3885,6 @@ static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, in return 0; } -static struct { - DECLARE_BITMAP(pin_programmed, MP_MAX_IOAPIC_PIN + 1); -} mp_ioapic_routing[MAX_IO_APICS]; - int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, int triggering, int polarity) { @@ -4023,51 +4028,44 @@ int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) #ifdef CONFIG_SMP void __init setup_ioapic_dest(void) { - int pin, ioapic, irq, irq_entry; + int pin, ioapic = 0, irq, irq_entry; struct irq_desc *desc; - struct irq_cfg *cfg; const struct cpumask *mask; 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); - - /* setup_IO_APIC_irqs could fail to get vector for some device - * when you have too many devices, because at that time only boot - * cpu is online. - */ - desc = irq_to_desc(irq); - cfg = desc->chip_data; - if (!cfg->vector) { - setup_IO_APIC_irq(ioapic, pin, irq, desc, - irq_trigger(irq_entry), - irq_polarity(irq_entry)); - continue; +#ifdef CONFIG_ACPI + if (!acpi_disabled && acpi_ioapic) { + ioapic = mp_find_ioapic(0); + if (ioapic < 0) + ioapic = 0; + } +#endif - } + 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); - /* - * Honour affinities which have been set in early boot - */ - if (desc->status & - (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) - mask = desc->affinity; - else - mask = apic->target_cpus(); + desc = irq_to_desc(irq); - if (intr_remapping_enabled) - set_ir_ioapic_affinity_irq_desc(desc, mask); - else - set_ioapic_affinity_irq_desc(desc, mask); - } + /* + * Honour affinities which have been set in early boot + */ + if (desc->status & + (IRQ_NO_BALANCING | IRQ_AFFINITY_SET)) + mask = desc->affinity; + else + mask = apic->target_cpus(); + if (intr_remapping_enabled) + set_ir_ioapic_affinity_irq_desc(desc, mask); + else + set_ioapic_affinity_irq_desc(desc, mask); } + } #endif diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index a2f6bde9c4e..2f3e192615c 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -889,6 +889,9 @@ static int pcibios_lookup_irq(struct pci_dev *dev, int assign) return 0; } + if (io_apic_assign_pci_irqs) + return 0; + /* Find IRQ routing entry */ if (!pirq_table) @@ -1039,63 +1042,15 @@ static void __init pcibios_fixup_irqs(void) pirq_penalty[dev->irq]++; } + if (io_apic_assign_pci_irqs) + return; + dev = NULL; while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) { pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin); if (!pin) continue; -#ifdef CONFIG_X86_IO_APIC - /* - * Recalculate IRQ numbers if we use the I/O APIC. - */ - if (io_apic_assign_pci_irqs) { - int irq; - int ioapic = -1, ioapic_pin = -1; - int triggering, polarity; - - /* - * interrupt pins are numbered starting from 1 - */ - irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, - PCI_SLOT(dev->devfn), pin - 1, - &ioapic, &ioapic_pin, - &triggering, &polarity); - /* - * Busses behind bridges are typically not listed in the - * MP-table. In this case we have to look up the IRQ - * based on the parent bus, parent slot, and pin number. - * The SMP code detects such bridged busses itself so we - * should get into this branch reliably. - */ - if (irq < 0 && dev->bus->parent) { - /* go back to the bridge */ - struct pci_dev *bridge = dev->bus->self; - int bus; - - pin = pci_swizzle_interrupt_pin(dev, pin); - bus = bridge->bus->number; - irq = IO_APIC_get_PCI_irq_vector(bus, - PCI_SLOT(bridge->devfn), - pin - 1, - &ioapic, &ioapic_pin, - &triggering, &polarity); - if (irq >= 0) - dev_warn(&dev->dev, - "using bridge %s INT %c to " - "get IRQ %d\n", - pci_name(bridge), - 'A' + pin - 1, irq); - } - if (irq >= 0) { - dev_info(&dev->dev, - "PCI->APIC IRQ transform: INT %c " - "-> IRQ %d\n", - 'A' + pin - 1, irq); - dev->irq = irq; - } - } -#endif /* * Still no IRQ? Try to lookup one... */ @@ -1190,6 +1145,19 @@ int __init pcibios_irq_init(void) pcibios_enable_irq = pirq_enable_irq; pcibios_fixup_irqs(); + + if (io_apic_assign_pci_irqs && pci_routeirq) { + struct pci_dev *dev = NULL; + /* + * PCI IRQ routing is set up by pci_enable_device(), but we + * also do it here in case there are still broken drivers that + * don't use pci_enable_device(). + */ + printk(KERN_INFO "PCI: Routing PCI interrupts for all devices because \"pci=routeirq\" specified\n"); + for_each_pci_dev(dev) + pirq_enable_irq(dev); + } + return 0; } @@ -1220,13 +1188,17 @@ void pcibios_penalize_isa_irq(int irq, int active) static int pirq_enable_irq(struct pci_dev *dev) { u8 pin; - struct pci_dev *temp_dev; pci_read_config_byte(dev, PCI_INTERRUPT_PIN, &pin); - if (pin && !pcibios_lookup_irq(dev, 1) && !dev->irq) { + if (pin && !pcibios_lookup_irq(dev, 1)) { char *msg = ""; + if (!io_apic_assign_pci_irqs && dev->irq) + return 0; + if (io_apic_assign_pci_irqs) { +#ifdef CONFIG_X86_IO_APIC + struct pci_dev *temp_dev; int irq; int ioapic = -1, ioapic_pin = -1; int triggering, polarity; @@ -1261,12 +1233,16 @@ static int pirq_enable_irq(struct pci_dev *dev) } dev = temp_dev; if (irq >= 0) { + io_apic_set_pci_routing(&dev->dev, ioapic, + ioapic_pin, irq, + triggering, polarity); + dev->irq = irq; dev_info(&dev->dev, "PCI->APIC IRQ transform: " "INT %c -> IRQ %d\n", 'A' + pin - 1, irq); - dev->irq = irq; return 0; } else msg = "; probably buggy MP table"; +#endif } else if (pci_probe & PCI_BIOS_IRQ_SCAN) msg = ""; else -- cgit v1.2.3-70-g09d2 From 61fe91e1319556f32bebfd7ed2c68ef02e2c17f7 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sat, 9 May 2009 23:47:42 -0700 Subject: x86: apic: Check rev 3 fadt correctly for physical_apic bit Impact: fix fadt version checking FADT2_REVISION_ID has value 3 aka rev 3 FADT. So need to use >= instead of >, as other places in the code do. [ Impact: extend scope of APIC boot quirk ] Signed-off-by: Yinghai Lu LKML-Reference: Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic_flat_64.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic_flat_64.c b/arch/x86/kernel/apic/apic_flat_64.c index 306e5e88fb6..744e6d8af27 100644 --- a/arch/x86/kernel/apic/apic_flat_64.c +++ b/arch/x86/kernel/apic/apic_flat_64.c @@ -235,7 +235,7 @@ static int physflat_acpi_madt_oem_check(char *oem_id, char *oem_table_id) * regardless of how many processors are present (x86_64 ES7000 * is an example). */ - if (acpi_gbl_FADT.header.revision > FADT2_REVISION_ID && + if (acpi_gbl_FADT.header.revision >= FADT2_REVISION_ID && (acpi_gbl_FADT.flags & ACPI_FADT_APIC_PHYSICAL)) { printk(KERN_DEBUG "system APIC only can use physical flat"); return 1; -- cgit v1.2.3-70-g09d2 From 4401da6111ac58f94234417427d06a72c4048c74 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sat, 2 May 2009 10:40:57 -0700 Subject: x86: read apic ID in the !acpi_lapic case Ed found that on 32-bit, boot_cpu_physical_apicid is not read right, when the mptable is broken. Interestingly, actually three paths use/set it: 1. acpi: at that time that is already read from reg 2. mptable: only read from mptable 3. no madt, and no mptable, that use default apic id 0 for 64-bit, -1 for 32-bit so we could read the apic id for the 2/3 path. We trust the hardware register more than we trust a BIOS data structure (the mptable). We can also avoid the double set_fixmap() when acpi_lapic is used, and also need to move cpu_has_apic earlier and call apic_disable(). Also when need to update the apic id, we'd better read and set the apic version as well - so that quirks are applied precisely. v2: make path 3 with 64bit, use -1 as apic id, so could read it later. v3: fix whitespace problem pointed out by Ed Swierk [ Impact: get correct apic id for bsp other than acpi path ] Reported-by: Ed Swierk Signed-off-by: Yinghai Lu Acked-by: Cyrill Gorcunov LKML-Reference: <49FC85A9.2070702@kernel.org> [ v4: sanity-check in the ACPI case too ] Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 31 +++++++++++++++++++++---------- 1 file changed, 21 insertions(+), 10 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index e258bedce7c..1ee966f4ae9 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1456,7 +1456,6 @@ static int __init detect_init_APIC(void) } mp_lapic_addr = APIC_DEFAULT_PHYS_BASE; - boot_cpu_physical_apicid = 0; return 0; } #else @@ -1570,6 +1569,8 @@ void __init early_init_lapic_mapping(void) */ void __init init_apic_mappings(void) { + unsigned int new_apicid; + if (x2apic_mode) { boot_cpu_physical_apicid = read_apic_id(); return; @@ -1586,21 +1587,31 @@ void __init init_apic_mappings(void) } else apic_phys = mp_lapic_addr; - set_fixmap_nocache(FIX_APIC_BASE, apic_phys); + /* lets check if we may NOP'ify apic operations */ + if (!cpu_has_apic) { + pr_info("APIC: disable apic facility\n"); + apic_disable(); + return; + } + + /* + * acpi lapic path already maps that address in + * acpi_register_lapic_address() + */ + if (!acpi_lapic) + set_fixmap_nocache(FIX_APIC_BASE, apic_phys); + apic_printk(APIC_VERBOSE, "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 = read_apic_id(); - - /* lets check if we may to NOP'ify apic operations */ - if (!cpu_has_apic) { - pr_info("APIC: disable apic facility\n"); - apic_disable(); + new_apicid = read_apic_id(); + if (boot_cpu_physical_apicid != new_apicid) { + boot_cpu_physical_apicid = new_apicid; + apic_version[new_apicid] = + GET_APIC_VERSION(apic_read(APIC_LVR)); } } -- cgit v1.2.3-70-g09d2 From 97a52714658cd959a3cfa35c5b6f489859f0204b Mon Sep 17 00:00:00 2001 From: Andreas Herrmann Date: Fri, 8 May 2009 18:23:50 +0200 Subject: x86: display extended apic registers with print_local_APIC and cpu_debug code Both print_local_APIC (used when apic=debug kernel param is set) and cpu_debug code missed support for some extended APIC registers that I'd like to see. This adds support to show: - extended APIC feature register - extended APIC control register - extended LVT registers [ Impact: print more debug info ] Signed-off-by: Andreas Herrmann Cc: Jaswinder Singh Rajput Cc: Cyrill Gorcunov LKML-Reference: <20090508162350.GO29045@alberich.amd.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/apicdef.h | 8 ++++---- arch/x86/kernel/apic/apic.c | 2 +- arch/x86/kernel/apic/io_apic.c | 14 +++++++++++++- arch/x86/kernel/cpu/cpu_debug.c | 14 +++++++++++++- 4 files changed, 31 insertions(+), 7 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/apicdef.h b/arch/x86/include/asm/apicdef.h index bc9514fb3b1..7ddb36ab933 100644 --- a/arch/x86/include/asm/apicdef.h +++ b/arch/x86/include/asm/apicdef.h @@ -22,6 +22,7 @@ # define APIC_INTEGRATED(x) (1) #endif #define APIC_XAPIC(x) ((x) >= 0x14) +#define APIC_EXT_SPACE(x) ((x) & 0x80000000) #define APIC_TASKPRI 0x80 #define APIC_TPRI_MASK 0xFFu #define APIC_ARBPRI 0x90 @@ -116,7 +117,9 @@ #define APIC_TDR_DIV_32 0x8 #define APIC_TDR_DIV_64 0x9 #define APIC_TDR_DIV_128 0xA -#define APIC_EILVT0 0x500 +#define APIC_EFEAT 0x400 +#define APIC_ECTRL 0x410 +#define APIC_EILVTn(n) (0x500 + 0x10 * n) #define APIC_EILVT_NR_AMD_K8 1 /* # of extended interrupts */ #define APIC_EILVT_NR_AMD_10H 4 #define APIC_EILVT_LVTOFF(x) (((x) >> 4) & 0xF) @@ -125,9 +128,6 @@ #define APIC_EILVT_MSG_NMI 0x4 #define APIC_EILVT_MSG_EXT 0x7 #define APIC_EILVT_MASKED (1 << 16) -#define APIC_EILVT1 0x510 -#define APIC_EILVT2 0x520 -#define APIC_EILVT3 0x530 #define APIC_BASE (fix_to_virt(FIX_APIC_BASE)) #define APIC_BASE_MSR 0x800 diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 1ee966f4ae9..0e6543fafb5 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -395,7 +395,7 @@ static void __setup_APIC_LVTT(unsigned int clocks, int oneshot, int irqen) static void setup_APIC_eilvt(u8 lvt_off, u8 vector, u8 msg_type, u8 mask) { - unsigned long reg = (lvt_off << 4) + APIC_EILVT0; + unsigned long reg = (lvt_off << 4) + APIC_EILVTn(0); unsigned int v = (mask << 16) | (msg_type << 8) | vector; apic_write(reg, v); diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 2afe145d277..65b824c9c4f 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1739,7 +1739,7 @@ __apicdebuginit(void) print_APIC_bitfield(int base) __apicdebuginit(void) print_local_APIC(void *dummy) { - unsigned int v, ver, maxlvt; + unsigned int i, v, ver, maxlvt; u64 icr; if (apic_verbosity == APIC_QUIET) @@ -1827,6 +1827,18 @@ __apicdebuginit(void) print_local_APIC(void *dummy) printk(KERN_DEBUG "... APIC TMCCT: %08x\n", v); v = apic_read(APIC_TDCR); printk(KERN_DEBUG "... APIC TDCR: %08x\n", v); + + if (boot_cpu_has(X86_FEATURE_EXTAPIC)) { + v = apic_read(APIC_EFEAT); + maxlvt = (v >> 16) & 0xff; + printk(KERN_DEBUG "... APIC EFEAT: %08x\n", v); + v = apic_read(APIC_ECTRL); + printk(KERN_DEBUG "... APIC ECTRL: %08x\n", v); + for (i = 0; i < maxlvt; i++) { + v = apic_read(APIC_EILVTn(i)); + printk(KERN_DEBUG "... APIC EILVT%d: %08x\n", i, v); + } + } printk("\n"); } diff --git a/arch/x86/kernel/cpu/cpu_debug.c b/arch/x86/kernel/cpu/cpu_debug.c index 46e29ab96c6..2fc4f6bb9ca 100644 --- a/arch/x86/kernel/cpu/cpu_debug.c +++ b/arch/x86/kernel/cpu/cpu_debug.c @@ -588,8 +588,20 @@ static void print_apic(void *arg) seq_printf(seq, " TMICT\t\t: %08x\n", apic_read(APIC_TMICT)); seq_printf(seq, " TMCCT\t\t: %08x\n", apic_read(APIC_TMCCT)); seq_printf(seq, " TDCR\t\t: %08x\n", apic_read(APIC_TDCR)); -#endif /* CONFIG_X86_LOCAL_APIC */ + if (boot_cpu_has(X86_FEATURE_EXTAPIC)) { + unsigned int i, v, maxeilvt; + + v = apic_read(APIC_EFEAT); + maxeilvt = (v >> 16) & 0xff; + seq_printf(seq, " EFEAT\t\t: %08x\n", v); + seq_printf(seq, " ECTRL\t\t: %08x\n", apic_read(APIC_ECTRL)); + for (i = 0; i < maxeilvt; i++) { + v = apic_read(APIC_EILVTn(i)); + seq_printf(seq, " EILVT%d\t\t: %08x\n", i, v); + } + } +#endif /* CONFIG_X86_LOCAL_APIC */ seq_printf(seq, "\n MSR\t:\n"); } -- cgit v1.2.3-70-g09d2 From cec6be6d1069d697beb490bbb40a290d5ff554a2 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Mon, 11 May 2009 17:41:40 +0400 Subject: x86: apic: Fixmap apic address even if apic disabled In case if apic were disabled by boot option we still need read_apic operation. So fixmap a fake apic area if needed. [ Impact: fix boot crash ] Signed-off-by: Cyrill Gorcunov Cc: yinghai@kernel.org Cc: eswierk@aristanetworks.com LKML-Reference: <20090511134140.GH4624@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 0e6543fafb5..07cffc1214c 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1587,13 +1587,6 @@ void __init init_apic_mappings(void) } else apic_phys = mp_lapic_addr; - /* lets check if we may NOP'ify apic operations */ - if (!cpu_has_apic) { - pr_info("APIC: disable apic facility\n"); - apic_disable(); - return; - } - /* * acpi lapic path already maps that address in * acpi_register_lapic_address() @@ -1602,7 +1595,15 @@ void __init init_apic_mappings(void) set_fixmap_nocache(FIX_APIC_BASE, apic_phys); apic_printk(APIC_VERBOSE, "mapped APIC to %08lx (%08lx)\n", - APIC_BASE, apic_phys); + APIC_BASE, apic_phys); + + /* lets check if we may NOP'ify apic operations */ + if (!cpu_has_apic) { + pr_info("APIC: disable apic facility\n"); + apic_disable(); + return; + } + /* * Fetch the APIC ID of the BSP in case we have a * default configuration (or the MP table is broken). -- cgit v1.2.3-70-g09d2 From 4797f6b021a3fa399942245d07a1feb30df81bb8 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sat, 2 May 2009 10:40:57 -0700 Subject: x86: read apic ID in the !acpi_lapic case Ed found that on 32-bit, boot_cpu_physical_apicid is not read right, when the mptable is broken. Interestingly, actually three paths use/set it: 1. acpi: at that time that is already read from reg 2. mptable: only read from mptable 3. no madt, and no mptable, that use default apic id 0 for 64-bit, -1 for 32-bit so we could read the apic id for the 2/3 path. We trust the hardware register more than we trust a BIOS data structure (the mptable). We can also avoid the double set_fixmap() when acpi_lapic is used, and also need to move cpu_has_apic earlier and call apic_disable(). Also when need to update the apic id, we'd better read and set the apic version as well - so that quirks are applied precisely. v2: make path 3 with 64bit, use -1 as apic id, so could read it later. v3: fix whitespace problem pointed out by Ed Swierk v5: fix boot crash [ Impact: get correct apic id for bsp other than acpi path ] Reported-by: Ed Swierk Signed-off-by: Yinghai Lu Acked-by: Cyrill Gorcunov LKML-Reference: <49FC85A9.2070702@kernel.org> [ v4: sanity-check in the ACPI case too ] Signed-off-by: Ingo Molnar --- arch/x86/include/asm/smp.h | 2 +- arch/x86/kernel/apic/apic.c | 46 ++++++++++++++++++++---------------------- arch/x86/kernel/apic/io_apic.c | 5 +++++ 3 files changed, 28 insertions(+), 25 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/smp.h b/arch/x86/include/asm/smp.h index 19e0d88b966..6a84ed166ae 100644 --- a/arch/x86/include/asm/smp.h +++ b/arch/x86/include/asm/smp.h @@ -180,7 +180,7 @@ extern int safe_smp_processor_id(void); static inline int logical_smp_processor_id(void) { /* we don't want to mark this access volatile - bad code generation */ - return GET_APIC_LOGICAL_ID(*(u32 *)(APIC_BASE + APIC_LDR)); + return GET_APIC_LOGICAL_ID(apic_read(APIC_LDR)); } #endif diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 07cffc1214c..b0fd26442c4 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -242,17 +242,24 @@ static int modern_apic(void) * bare function to substitute write operation * and it's _that_ fast :) */ -void native_apic_write_dummy(u32 reg, u32 v) +static void native_apic_write_dummy(u32 reg, u32 v) { WARN_ON_ONCE((cpu_has_apic || !disable_apic)); } +static u32 native_apic_read_dummy(u32 reg) +{ + WARN_ON_ONCE((cpu_has_apic || !disable_apic)); + return 0; +} + /* - * right after this call apic->write doesn't do anything + * right after this call apic->write/read doesn't do anything * note that there is no restore operation it works one way */ void apic_disable(void) { + apic->read = native_apic_read_dummy; apic->write = native_apic_write_dummy; } @@ -1576,32 +1583,23 @@ void __init init_apic_mappings(void) return; } - /* - * 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 no local APIC can be found return early */ if (!smp_found_config && detect_init_APIC()) { - apic_phys = (unsigned long) alloc_bootmem_pages(PAGE_SIZE); - apic_phys = __pa(apic_phys); - } else + /* lets NOP'ify apic operations */ + pr_info("APIC: disable apic facility\n"); + apic_disable(); + } else { apic_phys = mp_lapic_addr; - /* - * acpi lapic path already maps that address in - * acpi_register_lapic_address() - */ - if (!acpi_lapic) - set_fixmap_nocache(FIX_APIC_BASE, apic_phys); - - apic_printk(APIC_VERBOSE, "mapped APIC to %08lx (%08lx)\n", - APIC_BASE, apic_phys); + /* + * acpi lapic path already maps that address in + * acpi_register_lapic_address() + */ + if (!acpi_lapic) + set_fixmap_nocache(FIX_APIC_BASE, apic_phys); - /* lets check if we may NOP'ify apic operations */ - if (!cpu_has_apic) { - pr_info("APIC: disable apic facility\n"); - apic_disable(); - return; + apic_printk(APIC_VERBOSE, "mapped APIC to %08lx (%08lx)\n", + APIC_BASE, apic_phys); } /* diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 1f3d3669dae..74d2b480a20 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1878,6 +1878,11 @@ __apicdebuginit(void) print_PIC(void) __apicdebuginit(int) print_all_ICs(void) { print_PIC(); + + /* don't print out if apic is not there */ + if (!cpu_has_apic || disable_apic) + return 0; + print_all_local_APICs(); print_IO_APIC(); -- cgit v1.2.3-70-g09d2 From e5198075c67a22ec9a09565b1ce88d3d3f5ba855 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 15 May 2009 13:05:16 -0700 Subject: x86, apic: introduce io_apic_irq_attr according to Ingo, io_apic irq-setup related functions have too many parameters with a repetitive signature. So reduce related funcs to get less params by passing a pointer to a newly defined io_apic_irq_attr structure. v2: io_apic_irq ==> irq_attr triggering ==> trigger v3: add set_io_apic_irq_attr [ Impact: cleanup ] Signed-off-by: Yinghai Lu Cc: Andrew Morton Cc: Jesse Barnes Cc: Len Brown LKML-Reference: <4A08ACD3.2070401@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/hw_irq.h | 21 +++++++++++++++++-- arch/x86/include/asm/io_apic.h | 5 +++-- arch/x86/kernel/acpi/boot.c | 22 +++++++++++--------- arch/x86/kernel/apic/io_apic.c | 43 +++++++++++++++++++++++---------------- arch/x86/pci/irq.c | 16 +++++---------- drivers/pci/hotplug/ibmphp_core.c | 6 ++---- 6 files changed, 66 insertions(+), 47 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/hw_irq.h b/arch/x86/include/asm/hw_irq.h index 26a40ab7013..a7d14bbae11 100644 --- a/arch/x86/include/asm/hw_irq.h +++ b/arch/x86/include/asm/hw_irq.h @@ -63,9 +63,26 @@ extern unsigned long io_apic_irqs; extern void init_VISWS_APIC_irqs(void); extern void setup_IO_APIC(void); extern void disable_IO_APIC(void); + +struct io_apic_irq_attr { + int ioapic; + int ioapic_pin; + int trigger; + int polarity; +}; + +static inline void set_io_apic_irq_attr(struct io_apic_irq_attr *irq_attr, + int ioapic, int ioapic_pin, + int trigger, int polarity) +{ + irq_attr->ioapic = ioapic; + irq_attr->ioapic_pin = ioapic_pin; + irq_attr->trigger = trigger; + irq_attr->polarity = polarity; +} + extern int IO_APIC_get_PCI_irq_vector(int bus, int devfn, int pin, - int *ioapic, int *ioapic_pin, - int *trigger, int *polarity); + struct io_apic_irq_attr *irq_attr); extern void setup_ioapic_dest(void); extern void enable_IO_APIC(void); diff --git a/arch/x86/include/asm/io_apic.h b/arch/x86/include/asm/io_apic.h index 6fd99f96eb0..daf866ed061 100644 --- a/arch/x86/include/asm/io_apic.h +++ b/arch/x86/include/asm/io_apic.h @@ -156,8 +156,9 @@ extern int io_apic_get_version(int ioapic); extern int io_apic_get_redir_entries(int ioapic); #endif /* CONFIG_ACPI */ -extern int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, - int irq, int edge_level, int active_high_low); +struct io_apic_irq_attr; +extern int io_apic_set_pci_routing(struct device *dev, int irq, + struct io_apic_irq_attr *irq_attr); extern int (*ioapic_renumber_irq)(int ioapic, int irq); extern void ioapic_init_mappings(void); diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index dcfbc3ab9e4..4af63dfb0f0 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -523,7 +523,7 @@ int acpi_gsi_to_irq(u32 gsi, unsigned int *irq) * success: return IRQ number (>=0) * failure: return < 0 */ -int acpi_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) +int acpi_register_gsi(struct device *dev, u32 gsi, int trigger, int polarity) { unsigned int irq; unsigned int plat_gsi = gsi; @@ -533,14 +533,14 @@ int acpi_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) * Make sure all (legacy) PCI IRQs are set as level-triggered. */ if (acpi_irq_model == ACPI_IRQ_MODEL_PIC) { - if (triggering == ACPI_LEVEL_SENSITIVE) + if (trigger == 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(dev, gsi, triggering, polarity); + plat_gsi = mp_register_gsi(dev, gsi, trigger, polarity); } #endif acpi_gsi_to_irq(plat_gsi, &irq); @@ -1156,7 +1156,7 @@ void __init mp_config_acpi_legacy_irqs(void) } } -static int mp_config_acpi_gsi(struct device *dev, u32 gsi, int triggering, +static int mp_config_acpi_gsi(struct device *dev, u32 gsi, int trigger, int polarity) { #ifdef CONFIG_X86_MPPARSE @@ -1181,7 +1181,7 @@ static int mp_config_acpi_gsi(struct device *dev, u32 gsi, int triggering, /* print the entry should happen on mptable identically */ mp_irq.type = MP_INTSRC; mp_irq.irqtype = mp_INT; - mp_irq.irqflag = (triggering == ACPI_EDGE_SENSITIVE ? 4 : 0x0c) | + mp_irq.irqflag = (trigger == ACPI_EDGE_SENSITIVE ? 4 : 0x0c) | (polarity == ACPI_ACTIVE_HIGH ? 1 : 3); mp_irq.srcbus = number; mp_irq.srcbusirq = (((devfn >> 3) & 0x1f) << 2) | ((pin - 1) & 3); @@ -1194,10 +1194,11 @@ static int mp_config_acpi_gsi(struct device *dev, u32 gsi, int triggering, return 0; } -int mp_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) +int mp_register_gsi(struct device *dev, u32 gsi, int trigger, int polarity) { int ioapic; int ioapic_pin; + struct io_apic_irq_attr irq_attr; if (acpi_irq_model != ACPI_IRQ_MODEL_IOAPIC) return gsi; @@ -1225,11 +1226,12 @@ int mp_register_gsi(struct device *dev, u32 gsi, int triggering, int polarity) ioapic_pin); return gsi; } - mp_config_acpi_gsi(dev, gsi, triggering, polarity); + mp_config_acpi_gsi(dev, gsi, trigger, polarity); - io_apic_set_pci_routing(dev, ioapic, ioapic_pin, gsi, - triggering == ACPI_EDGE_SENSITIVE ? 0 : 1, - polarity == ACPI_ACTIVE_HIGH ? 0 : 1); + set_io_apic_irq_attr(&irq_attr, ioapic, ioapic_pin, + trigger == ACPI_EDGE_SENSITIVE ? 0 : 1, + polarity == ACPI_ACTIVE_HIGH ? 0 : 1); + io_apic_set_pci_routing(dev, gsi, &irq_attr); return gsi; } diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 74d2b480a20..ce1ac74baa7 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1096,8 +1096,7 @@ static int pin_2_irq(int idx, int apic, int pin) * Not an __init, possibly needed by modules */ int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin, - int *ioapic, int *ioapic_pin, - int *trigger, int *polarity) + struct io_apic_irq_attr *irq_attr) { int apic, i, best_guess = -1; @@ -1127,10 +1126,10 @@ int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin, continue; if (pin == (mp_irqs[i].srcbusirq & 3)) { - *ioapic = apic; - *ioapic_pin = mp_irqs[i].dstirq; - *trigger = irq_trigger(i); - *polarity = irq_polarity(i); + set_io_apic_irq_attr(irq_attr, apic, + mp_irqs[i].dstirq, + irq_trigger(i), + irq_polarity(i)); return irq; } /* @@ -1138,10 +1137,10 @@ int IO_APIC_get_PCI_irq_vector(int bus, int slot, int pin, * best-guess fuzzy result for broken mptables. */ if (best_guess < 0) { - *ioapic = apic; - *ioapic_pin = mp_irqs[i].dstirq; - *trigger = irq_trigger(i); - *polarity = irq_polarity(i); + set_io_apic_irq_attr(irq_attr, apic, + mp_irqs[i].dstirq, + irq_trigger(i), + irq_polarity(i)); best_guess = irq; } } @@ -3865,13 +3864,16 @@ int __init arch_probe_nr_irqs(void) } #endif -static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, - int triggering, int polarity) +static int __io_apic_set_pci_routing(struct device *dev, int irq, + struct io_apic_irq_attr *irq_attr) { struct irq_desc *desc; struct irq_cfg *cfg; int node; + int ioapic, pin; + int trigger, polarity; + ioapic = irq_attr->ioapic; if (!IO_APIC_IRQ(irq)) { apic_printk(APIC_QUIET,KERN_ERR "IOAPIC[%d]: Invalid reference to IRQ 0\n", ioapic); @@ -3889,6 +3891,10 @@ static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, in return 0; } + pin = irq_attr->ioapic_pin; + trigger = irq_attr->trigger; + polarity = irq_attr->polarity; + /* * IRQs < 16 are already in the irq_2_pin[] map */ @@ -3897,20 +3903,22 @@ static int __io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, in add_pin_to_irq_node(cfg, node, ioapic, pin); } - setup_IO_APIC_irq(ioapic, pin, irq, desc, triggering, polarity); + setup_IO_APIC_irq(ioapic, pin, irq, desc, trigger, polarity); return 0; } -int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, - int triggering, int polarity) +int io_apic_set_pci_routing(struct device *dev, int irq, + struct io_apic_irq_attr *irq_attr) { - + int ioapic, pin; /* * Avoid pin reprogramming. PRTs typically include entries * with redundant pin->gsi mappings (but unique PCI devices); * we only program the IOAPIC on the first. */ + ioapic = irq_attr->ioapic; + pin = irq_attr->ioapic_pin; if (test_bit(pin, mp_ioapic_routing[ioapic].pin_programmed)) { pr_debug("Pin %d-%d already programmed\n", mp_ioapics[ioapic].apicid, pin); @@ -3918,8 +3926,7 @@ int io_apic_set_pci_routing(struct device *dev, int ioapic, int pin, int irq, } set_bit(pin, mp_ioapic_routing[ioapic].pin_programmed); - return __io_apic_set_pci_routing(dev, ioapic, pin, irq, - triggering, polarity); + return __io_apic_set_pci_routing(dev, irq, irq_attr); } /* -------------------------------------------------------------------------- diff --git a/arch/x86/pci/irq.c b/arch/x86/pci/irq.c index 2f3e192615c..0696d506c4a 100644 --- a/arch/x86/pci/irq.c +++ b/arch/x86/pci/irq.c @@ -1200,14 +1200,11 @@ static int pirq_enable_irq(struct pci_dev *dev) #ifdef CONFIG_X86_IO_APIC struct pci_dev *temp_dev; int irq; - int ioapic = -1, ioapic_pin = -1; - int triggering, polarity; + struct io_apic_irq_attr irq_attr; irq = IO_APIC_get_PCI_irq_vector(dev->bus->number, PCI_SLOT(dev->devfn), - pin - 1, - &ioapic, &ioapic_pin, - &triggering, &polarity); + pin - 1, &irq_attr); /* * Busses behind bridges are typically not listed in the MP-table. * In this case we have to look up the IRQ based on the parent bus, @@ -1221,9 +1218,7 @@ static int pirq_enable_irq(struct pci_dev *dev) pin = pci_swizzle_interrupt_pin(dev, pin); irq = IO_APIC_get_PCI_irq_vector(bridge->bus->number, PCI_SLOT(bridge->devfn), - pin - 1, - &ioapic, &ioapic_pin, - &triggering, &polarity); + pin - 1, &irq_attr); if (irq >= 0) dev_warn(&dev->dev, "using bridge %s " "INT %c to get IRQ %d\n", @@ -1233,9 +1228,8 @@ static int pirq_enable_irq(struct pci_dev *dev) } dev = temp_dev; if (irq >= 0) { - io_apic_set_pci_routing(&dev->dev, ioapic, - ioapic_pin, irq, - triggering, polarity); + io_apic_set_pci_routing(&dev->dev, irq, + &irq_attr); dev->irq = irq; dev_info(&dev->dev, "PCI->APIC IRQ transform: " "INT %c -> IRQ %d\n", 'A' + pin - 1, irq); diff --git a/drivers/pci/hotplug/ibmphp_core.c b/drivers/pci/hotplug/ibmphp_core.c index 79901a0db88..42e4260c3b1 100644 --- a/drivers/pci/hotplug/ibmphp_core.c +++ b/drivers/pci/hotplug/ibmphp_core.c @@ -155,15 +155,13 @@ int ibmphp_init_devno(struct slot **cur_slot) for (loop = 0; loop < len; loop++) { if ((*cur_slot)->number == rtable->slots[loop].slot && (*cur_slot)->bus == rtable->slots[loop].bus) { - int ioapic = -1, ioapic_pin = -1; - int triggering, polarity; + struct io_apic_irq_attr irq_attr; (*cur_slot)->device = PCI_SLOT(rtable->slots[loop].devfn); for (i = 0; i < 4; i++) (*cur_slot)->irq[i] = IO_APIC_get_PCI_irq_vector((int) (*cur_slot)->bus, (int) (*cur_slot)->device, i, - &ioapic, &ioapic_pin, - &triggering, &polarity); + &irq_attr); debug("(*cur_slot)->irq[0] = %x\n", (*cur_slot)->irq[0]); -- cgit v1.2.3-70-g09d2 From 2759c3287de27266e06f1f4e82cbd2d65f6a044c Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Fri, 15 May 2009 13:05:16 -0700 Subject: x86: don't call read_apic_id if !cpu_has_apic should not call that if apic is disabled. [ Impact: fix crash on certain UP configs ] Signed-off-by: Yinghai Lu Cc: Cyrill Gorcunov LKML-Reference: <4A09CCBB.2000306@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic_flat_64.c | 2 +- arch/x86/kernel/cpu/amd.c | 2 +- arch/x86/kernel/cpu/common.c | 6 ++++++ arch/x86/kernel/cpu/intel.c | 6 +++--- 4 files changed, 11 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic_flat_64.c b/arch/x86/kernel/apic/apic_flat_64.c index 744e6d8af27..d0c99abc26c 100644 --- a/arch/x86/kernel/apic/apic_flat_64.c +++ b/arch/x86/kernel/apic/apic_flat_64.c @@ -161,7 +161,7 @@ static int flat_apic_id_registered(void) static int flat_phys_pkg_id(int initial_apic_id, int index_msb) { - return hard_smp_processor_id() >> index_msb; + return initial_apic_id >> index_msb; } struct apic apic_flat = { diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 7e4a459daa6..728b3750a3e 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -272,7 +272,7 @@ static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c) #if defined(CONFIG_NUMA) && defined(CONFIG_X86_64) int cpu = smp_processor_id(); int node; - unsigned apicid = hard_smp_processor_id(); + unsigned apicid = cpu_has_apic ? hard_smp_processor_id() : c->apicid; node = c->phys_proc_id; if (apicid_to_node[apicid] != NUMA_NO_NODE) diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index c1caefc82e6..017c600e05a 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -761,6 +761,12 @@ static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) if (this_cpu->c_identify) this_cpu->c_identify(c); + /* Clear/Set all flags overriden by options, after probe */ + for (i = 0; i < NCAPINTS; i++) { + c->x86_capability[i] &= ~cpu_caps_cleared[i]; + c->x86_capability[i] |= cpu_caps_set[i]; + } + #ifdef CONFIG_X86_64 c->apicid = apic->phys_pkg_id(c->initial_apicid, 0); #endif diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 7437fa133c0..daed39ba261 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -229,12 +229,12 @@ static void __cpuinit intel_workarounds(struct cpuinfo_x86 *c) } #endif -static void __cpuinit srat_detect_node(void) +static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c) { #if defined(CONFIG_NUMA) && defined(CONFIG_X86_64) unsigned node; int cpu = smp_processor_id(); - int apicid = hard_smp_processor_id(); + int apicid = cpu_has_apic ? hard_smp_processor_id() : c->apicid; /* Don't do the funky fallback heuristics the AMD version employs for now. */ @@ -400,7 +400,7 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) } /* Work around errata */ - srat_detect_node(); + srat_detect_node(c); if (cpu_has(c, X86_FEATURE_VMX)) detect_vmx_virtcap(c); -- cgit v1.2.3-70-g09d2 From 4c6f18fc81565967da20f2d4a3922cdba33f8e2b Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 18 May 2009 10:23:28 -0700 Subject: x86, io-apic: Don't mark pin_programmed early Peter bisected that: | commit b9c61b70075c87a8612624736faf4a2de5b1ed30 | Date: Wed May 6 10:10:06 2009 -0700 | | x86/pci: update pirq_enable_irq() to setup io apic routing | | So we can set io apic routing only when enabling the device irq. wrecked his opteron box, ata1 interrupts fail to get through. ata1 is using irq 11: [ 1.451839] sata_svw 0000:01:0e.0: version 2.3 [ 1.456333] sata_svw 0000:01:0e.0: PCI INT A -> GSI 11 (level, low) -> IRQ 11 [ 1.463639] scsi0 : sata_svw [ 1.466949] scsi1 : sata_svw [ 1.470022] scsi2 : sata_svw [ 1.473090] scsi3 : sata_svw [ 1.476112] ata1: SATA max UDMA/133 mmio m8192@0xff3fe000 port 0xff3fe000 irq 11 [ 1.483490] ata2: SATA max UDMA/133 mmio m8192@0xff3fe000 port 0xff3fe100 irq 11 [ 1.490870] ata3: SATA max UDMA/133 mmio m8192@0xff3fe000 port 0xff3fe200 irq 11 [ 1.498247] ata4: SATA max UDMA/133 mmio m8192@0xff3fe000 port 0xff3fe300 irq 11 that pin is overlapped with pin with legacy ones. We should not set bits in pin_programmed here, so that those bit could be set later via io_apic_set_pci_routing(). [ Impact: fix boot hang on certain systems ] Reported-by: Peter Zijlstra Signed-off-by: Yinghai Lu Tested-by: Peter Zijlstra Cc: Jack Steiner LKML-Reference: <4A119990.9020606@kernel.org> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/io_apic.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index ce1ac74baa7..ac7f3b6ad58 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1537,7 +1537,10 @@ static void __init setup_IO_APIC_irqs(void) } cfg = desc->chip_data; add_pin_to_irq_node(cfg, node, apic_id, pin); - set_bit(pin, mp_ioapic_routing[apic_id].pin_programmed); + /* + * don't mark it in pin_programmed, so later acpi could + * set it correctly when irq < 16 + */ setup_IO_APIC_irq(apic_id, pin, irq, desc, irq_trigger(idx), irq_polarity(idx)); } -- cgit v1.2.3-70-g09d2 From a63eaf34ae60bdb067a354cc8def2e8f4a01f5f4 Mon Sep 17 00:00:00 2001 From: Paul Mackerras Date: Fri, 22 May 2009 14:17:31 +1000 Subject: perf_counter: Dynamically allocate tasks' perf_counter_context struct This replaces the struct perf_counter_context in the task_struct with a pointer to a dynamically allocated perf_counter_context struct. The main reason for doing is this is to allow us to transfer a perf_counter_context from one task to another when we do lazy PMU switching in a later patch. This has a few side-benefits: the task_struct becomes a little smaller, we save some memory because only tasks that have perf_counters attached get a perf_counter_context allocated for them, and we can remove the inclusion of in sched.h, meaning that we don't end up recompiling nearly everything whenever perf_counter.h changes. The perf_counter_context structures are reference-counted and freed when the last reference is dropped. A context can have references from its task and the counters on its task. Counters can outlive the task so it is possible that a context will be freed well after its task has exited. Contexts are allocated on fork if the parent had a context, or otherwise the first time that a per-task counter is created on a task. In the latter case, we set the context pointer in the task struct locklessly using an atomic compare-and-exchange operation in case we raced with some other task in creating a context for the subject task. This also removes the task pointer from the perf_counter struct. The task pointer was not used anywhere and would make it harder to move a context from one task to another. Anything that needed to know which task a counter was attached to was already using counter->ctx->task. The __perf_counter_init_context function moves up in perf_counter.c so that it can be called from find_get_context, and now initializes the refcount, but is otherwise unchanged. We were potentially calling list_del_counter twice: once from __perf_counter_exit_task when the task exits and once from __perf_counter_remove_from_context when the counter's fd gets closed. This adds a check in list_del_counter so it doesn't do anything if the counter has already been removed from the lists. Since perf_counter_task_sched_in doesn't do anything if the task doesn't have a context, and leaves cpuctx->task_ctx = NULL, this adds code to __perf_install_in_context to set cpuctx->task_ctx if necessary, i.e. in the case where the current task adds the first counter to itself and thus creates a context for itself. This also adds similar code to __perf_counter_enable to handle a similar situation which can arise when the counters have been disabled using prctl; that also leaves cpuctx->task_ctx = NULL. [ Impact: refactor counter context management to prepare for new feature ] Signed-off-by: Paul Mackerras Acked-by: Peter Zijlstra Cc: Corey Ashford Cc: Marcelo Tosatti Cc: Arnaldo Carvalho de Melo LKML-Reference: <18966.10075.781053.231153@cargo.ozlabs.ibm.com> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 1 + include/linux/init_task.h | 13 --- include/linux/perf_counter.h | 4 +- include/linux/sched.h | 6 +- kernel/exit.c | 3 +- kernel/fork.c | 1 + kernel/perf_counter.c | 218 +++++++++++++++++++++++++++---------------- 7 files changed, 145 insertions(+), 101 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index e9021a90802..b4f64402a82 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -14,6 +14,7 @@ * Mikael Pettersson : PM converted to driver model. */ +#include #include #include #include diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 503afaa0afa..d87247d2641 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -108,18 +108,6 @@ extern struct group_info init_groups; extern struct cred init_cred; -#ifdef CONFIG_PERF_COUNTERS -# define INIT_PERF_COUNTERS(tsk) \ - .perf_counter_ctx.counter_list = \ - LIST_HEAD_INIT(tsk.perf_counter_ctx.counter_list), \ - .perf_counter_ctx.event_list = \ - LIST_HEAD_INIT(tsk.perf_counter_ctx.event_list), \ - .perf_counter_ctx.lock = \ - __SPIN_LOCK_UNLOCKED(tsk.perf_counter_ctx.lock), -#else -# define INIT_PERF_COUNTERS(tsk) -#endif - /* * INIT_TASK is used to set up the first task table, touch at * your own risk!. Base=0, limit=0x1fffff (=2MB) @@ -183,7 +171,6 @@ extern struct cred init_cred; }, \ .dirties = INIT_PROP_LOCAL_SINGLE(dirties), \ INIT_IDS \ - INIT_PERF_COUNTERS(tsk) \ INIT_TRACE_IRQFLAGS \ INIT_LOCKDEP \ INIT_FTRACE_GRAPH \ diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h index f612941ef46..07130900546 100644 --- a/include/linux/perf_counter.h +++ b/include/linux/perf_counter.h @@ -449,7 +449,6 @@ struct perf_counter { struct hw_perf_counter hw; struct perf_counter_context *ctx; - struct task_struct *task; struct file *filp; struct perf_counter *parent; @@ -498,7 +497,6 @@ struct perf_counter { * Used as a container for task counters and CPU counters as well: */ struct perf_counter_context { -#ifdef CONFIG_PERF_COUNTERS /* * Protect the states of the counters in the list, * nr_active, and the list: @@ -516,6 +514,7 @@ struct perf_counter_context { int nr_counters; int nr_active; int is_active; + atomic_t refcount; struct task_struct *task; /* @@ -523,7 +522,6 @@ struct perf_counter_context { */ u64 time; u64 timestamp; -#endif }; /** diff --git a/include/linux/sched.h b/include/linux/sched.h index ff59d123151..9714d450f41 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -71,7 +71,6 @@ struct sched_param { #include #include #include -#include #include #include #include @@ -99,6 +98,7 @@ struct robust_list_head; struct bio; struct bts_tracer; struct fs_struct; +struct perf_counter_context; /* * List of flags we want to share for kernel threads, @@ -1387,7 +1387,9 @@ struct task_struct { struct list_head pi_state_list; struct futex_pi_state *pi_state_cache; #endif - struct perf_counter_context perf_counter_ctx; +#ifdef CONFIG_PERF_COUNTERS + struct perf_counter_context *perf_counter_ctxp; +#endif #ifdef CONFIG_NUMA struct mempolicy *mempolicy; short il_next; diff --git a/kernel/exit.c b/kernel/exit.c index f9dfedd94af..99ad4063ee4 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -48,6 +48,7 @@ #include #include #include +#include #include #include @@ -159,7 +160,7 @@ static void delayed_put_task_struct(struct rcu_head *rhp) struct task_struct *tsk = container_of(rhp, struct task_struct, rcu); #ifdef CONFIG_PERF_COUNTERS - WARN_ON_ONCE(!list_empty(&tsk->perf_counter_ctx.counter_list)); + WARN_ON_ONCE(tsk->perf_counter_ctxp); #endif trace_sched_process_free(tsk); put_task_struct(tsk); diff --git a/kernel/fork.c b/kernel/fork.c index d32fef4d38e..e72a09f5355 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -63,6 +63,7 @@ #include #include #include +#include #include #include diff --git a/kernel/perf_counter.c b/kernel/perf_counter.c index 08584c16049..06ea3eae886 100644 --- a/kernel/perf_counter.c +++ b/kernel/perf_counter.c @@ -97,6 +97,17 @@ void perf_enable(void) hw_perf_enable(); } +static void get_ctx(struct perf_counter_context *ctx) +{ + atomic_inc(&ctx->refcount); +} + +static void put_ctx(struct perf_counter_context *ctx) +{ + if (atomic_dec_and_test(&ctx->refcount)) + kfree(ctx); +} + static void list_add_counter(struct perf_counter *counter, struct perf_counter_context *ctx) { @@ -118,11 +129,17 @@ list_add_counter(struct perf_counter *counter, struct perf_counter_context *ctx) ctx->nr_counters++; } +/* + * Remove a counter from the lists for its context. + * Must be called with counter->mutex and ctx->mutex held. + */ static void list_del_counter(struct perf_counter *counter, struct perf_counter_context *ctx) { struct perf_counter *sibling, *tmp; + if (list_empty(&counter->list_entry)) + return; ctx->nr_counters--; list_del_init(&counter->list_entry); @@ -216,8 +233,6 @@ static void __perf_counter_remove_from_context(void *info) counter_sched_out(counter, cpuctx, ctx); - counter->task = NULL; - list_del_counter(counter, ctx); if (!ctx->task) { @@ -279,7 +294,6 @@ retry: */ if (!list_empty(&counter->list_entry)) { list_del_counter(counter, ctx); - counter->task = NULL; } spin_unlock_irq(&ctx->lock); } @@ -568,11 +582,17 @@ static void __perf_install_in_context(void *info) * If this is a task context, we need to check whether it is * the current task context of this cpu. If not it has been * scheduled out before the smp call arrived. + * Or possibly this is the right context but it isn't + * on this cpu because it had no counters. */ - if (ctx->task && cpuctx->task_ctx != ctx) - return; + if (ctx->task && cpuctx->task_ctx != ctx) { + if (cpuctx->task_ctx || ctx->task != current) + return; + cpuctx->task_ctx = ctx; + } spin_lock_irqsave(&ctx->lock, flags); + ctx->is_active = 1; update_context_time(ctx); /* @@ -653,7 +673,6 @@ perf_install_in_context(struct perf_counter_context *ctx, return; } - counter->task = task; retry: task_oncpu_function_call(task, __perf_install_in_context, counter); @@ -693,10 +712,14 @@ static void __perf_counter_enable(void *info) * If this is a per-task counter, need to check whether this * counter's task is the current task on this cpu. */ - if (ctx->task && cpuctx->task_ctx != ctx) - return; + if (ctx->task && cpuctx->task_ctx != ctx) { + if (cpuctx->task_ctx || ctx->task != current) + return; + cpuctx->task_ctx = ctx; + } spin_lock_irqsave(&ctx->lock, flags); + ctx->is_active = 1; update_context_time(ctx); counter->prev_state = counter->state; @@ -852,10 +875,10 @@ void __perf_counter_sched_out(struct perf_counter_context *ctx, void perf_counter_task_sched_out(struct task_struct *task, int cpu) { struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); - struct perf_counter_context *ctx = &task->perf_counter_ctx; + struct perf_counter_context *ctx = task->perf_counter_ctxp; struct pt_regs *regs; - if (likely(!cpuctx->task_ctx)) + if (likely(!ctx || !cpuctx->task_ctx)) return; update_context_time(ctx); @@ -871,6 +894,8 @@ static void __perf_counter_task_sched_out(struct perf_counter_context *ctx) { struct perf_cpu_context *cpuctx = &__get_cpu_var(perf_cpu_context); + if (!cpuctx->task_ctx) + return; __perf_counter_sched_out(ctx, cpuctx); cpuctx->task_ctx = NULL; } @@ -969,8 +994,10 @@ __perf_counter_sched_in(struct perf_counter_context *ctx, void perf_counter_task_sched_in(struct task_struct *task, int cpu) { struct perf_cpu_context *cpuctx = &per_cpu(perf_cpu_context, cpu); - struct perf_counter_context *ctx = &task->perf_counter_ctx; + struct perf_counter_context *ctx = task->perf_counter_ctxp; + if (likely(!ctx)) + return; __perf_counter_sched_in(ctx, cpuctx, cpu); cpuctx->task_ctx = ctx; } @@ -985,11 +1012,11 @@ static void perf_counter_cpu_sched_in(struct perf_cpu_context *cpuctx, int cpu) int perf_counter_task_disable(void) { struct task_struct *curr = current; - struct perf_counter_context *ctx = &curr->perf_counter_ctx; + struct perf_counter_context *ctx = curr->perf_counter_ctxp; struct perf_counter *counter; unsigned long flags; - if (likely(!ctx->nr_counters)) + if (!ctx || !ctx->nr_counters) return 0; local_irq_save(flags); @@ -1020,12 +1047,12 @@ int perf_counter_task_disable(void) int perf_counter_task_enable(void) { struct task_struct *curr = current; - struct perf_counter_context *ctx = &curr->perf_counter_ctx; + struct perf_counter_context *ctx = curr->perf_counter_ctxp; struct perf_counter *counter; unsigned long flags; int cpu; - if (likely(!ctx->nr_counters)) + if (!ctx || !ctx->nr_counters) return 0; local_irq_save(flags); @@ -1128,19 +1155,23 @@ void perf_counter_task_tick(struct task_struct *curr, int cpu) return; cpuctx = &per_cpu(perf_cpu_context, cpu); - ctx = &curr->perf_counter_ctx; + ctx = curr->perf_counter_ctxp; perf_adjust_freq(&cpuctx->ctx); - perf_adjust_freq(ctx); + if (ctx) + perf_adjust_freq(ctx); perf_counter_cpu_sched_out(cpuctx); - __perf_counter_task_sched_out(ctx); + if (ctx) + __perf_counter_task_sched_out(ctx); rotate_ctx(&cpuctx->ctx); - rotate_ctx(ctx); + if (ctx) + rotate_ctx(ctx); perf_counter_cpu_sched_in(cpuctx, cpu); - perf_counter_task_sched_in(curr, cpu); + if (ctx) + perf_counter_task_sched_in(curr, cpu); } /* @@ -1176,6 +1207,22 @@ static u64 perf_counter_read(struct perf_counter *counter) return atomic64_read(&counter->count); } +/* + * Initialize the perf_counter context in a task_struct: + */ +static void +__perf_counter_init_context(struct perf_counter_context *ctx, + struct task_struct *task) +{ + memset(ctx, 0, sizeof(*ctx)); + spin_lock_init(&ctx->lock); + mutex_init(&ctx->mutex); + INIT_LIST_HEAD(&ctx->counter_list); + INIT_LIST_HEAD(&ctx->event_list); + atomic_set(&ctx->refcount, 1); + ctx->task = task; +} + static void put_context(struct perf_counter_context *ctx) { if (ctx->task) @@ -1186,6 +1233,7 @@ static struct perf_counter_context *find_get_context(pid_t pid, int cpu) { struct perf_cpu_context *cpuctx; struct perf_counter_context *ctx; + struct perf_counter_context *tctx; struct task_struct *task; /* @@ -1225,15 +1273,36 @@ static struct perf_counter_context *find_get_context(pid_t pid, int cpu) if (!task) return ERR_PTR(-ESRCH); - ctx = &task->perf_counter_ctx; - ctx->task = task; - /* Reuse ptrace permission checks for now. */ if (!ptrace_may_access(task, PTRACE_MODE_READ)) { - put_context(ctx); + put_task_struct(task); return ERR_PTR(-EACCES); } + ctx = task->perf_counter_ctxp; + if (!ctx) { + ctx = kmalloc(sizeof(struct perf_counter_context), GFP_KERNEL); + if (!ctx) { + put_task_struct(task); + return ERR_PTR(-ENOMEM); + } + __perf_counter_init_context(ctx, task); + /* + * Make sure other cpus see correct values for *ctx + * once task->perf_counter_ctxp is visible to them. + */ + smp_wmb(); + tctx = cmpxchg(&task->perf_counter_ctxp, NULL, ctx); + if (tctx) { + /* + * We raced with some other task; use + * the context they set. + */ + kfree(ctx); + ctx = tctx; + } + } + return ctx; } @@ -1242,6 +1311,7 @@ static void free_counter_rcu(struct rcu_head *head) struct perf_counter *counter; counter = container_of(head, struct perf_counter, rcu_head); + put_ctx(counter->ctx); kfree(counter); } @@ -2247,7 +2317,7 @@ static void perf_counter_comm_event(struct perf_comm_event *comm_event) perf_counter_comm_ctx(&cpuctx->ctx, comm_event); put_cpu_var(perf_cpu_context); - perf_counter_comm_ctx(¤t->perf_counter_ctx, comm_event); + perf_counter_comm_ctx(current->perf_counter_ctxp, comm_event); } void perf_counter_comm(struct task_struct *task) @@ -2256,7 +2326,9 @@ void perf_counter_comm(struct task_struct *task) if (!atomic_read(&nr_comm_tracking)) return; - + if (!current->perf_counter_ctxp) + return; + comm_event = (struct perf_comm_event){ .task = task, .event = { @@ -2372,7 +2444,7 @@ got_name: perf_counter_mmap_ctx(&cpuctx->ctx, mmap_event); put_cpu_var(perf_cpu_context); - perf_counter_mmap_ctx(¤t->perf_counter_ctx, mmap_event); + perf_counter_mmap_ctx(current->perf_counter_ctxp, mmap_event); kfree(buf); } @@ -2384,6 +2456,8 @@ void perf_counter_mmap(unsigned long addr, unsigned long len, if (!atomic_read(&nr_mmap_tracking)) return; + if (!current->perf_counter_ctxp) + return; mmap_event = (struct perf_mmap_event){ .file = file, @@ -2985,6 +3059,7 @@ perf_counter_alloc(struct perf_counter_hw_event *hw_event, counter->group_leader = group_leader; counter->pmu = NULL; counter->ctx = ctx; + get_ctx(ctx); counter->state = PERF_COUNTER_STATE_INACTIVE; if (hw_event->disabled) @@ -3149,21 +3224,6 @@ err_put_context: goto out_fput; } -/* - * Initialize the perf_counter context in a task_struct: - */ -static void -__perf_counter_init_context(struct perf_counter_context *ctx, - struct task_struct *task) -{ - memset(ctx, 0, sizeof(*ctx)); - spin_lock_init(&ctx->lock); - mutex_init(&ctx->mutex); - INIT_LIST_HEAD(&ctx->counter_list); - INIT_LIST_HEAD(&ctx->event_list); - ctx->task = task; -} - /* * inherit a counter from parent task to child task: */ @@ -3195,7 +3255,6 @@ inherit_counter(struct perf_counter *parent_counter, /* * Link it up in the child's context: */ - child_counter->task = child; add_counter_to_ctx(child_counter, child_ctx); child_counter->parent = parent_counter; @@ -3294,40 +3353,15 @@ __perf_counter_exit_task(struct task_struct *child, struct perf_counter *parent_counter; /* - * If we do not self-reap then we have to wait for the - * child task to unschedule (it will happen for sure), - * so that its counter is at its final count. (This - * condition triggers rarely - child tasks usually get - * off their CPU before the parent has a chance to - * get this far into the reaping action) + * Protect against concurrent operations on child_counter + * due its fd getting closed, etc. */ - if (child != current) { - wait_task_inactive(child, 0); - update_counter_times(child_counter); - list_del_counter(child_counter, child_ctx); - } else { - struct perf_cpu_context *cpuctx; - unsigned long flags; - - /* - * Disable and unlink this counter. - * - * Be careful about zapping the list - IRQ/NMI context - * could still be processing it: - */ - local_irq_save(flags); - perf_disable(); - - cpuctx = &__get_cpu_var(perf_cpu_context); + mutex_lock(&child_counter->mutex); - group_sched_out(child_counter, cpuctx, child_ctx); - update_counter_times(child_counter); + update_counter_times(child_counter); + list_del_counter(child_counter, child_ctx); - list_del_counter(child_counter, child_ctx); - - perf_enable(); - local_irq_restore(flags); - } + mutex_unlock(&child_counter->mutex); parent_counter = child_counter->parent; /* @@ -3346,19 +3380,29 @@ __perf_counter_exit_task(struct task_struct *child, * * Note: we may be running in child context, but the PID is not hashed * anymore so new counters will not be added. + * (XXX not sure that is true when we get called from flush_old_exec. + * -- paulus) */ void perf_counter_exit_task(struct task_struct *child) { struct perf_counter *child_counter, *tmp; struct perf_counter_context *child_ctx; + unsigned long flags; WARN_ON_ONCE(child != current); - child_ctx = &child->perf_counter_ctx; + child_ctx = child->perf_counter_ctxp; - if (likely(!child_ctx->nr_counters)) + if (likely(!child_ctx)) return; + local_irq_save(flags); + __perf_counter_task_sched_out(child_ctx); + child->perf_counter_ctxp = NULL; + local_irq_restore(flags); + + mutex_lock(&child_ctx->mutex); + again: list_for_each_entry_safe(child_counter, tmp, &child_ctx->counter_list, list_entry) @@ -3371,6 +3415,10 @@ again: */ if (!list_empty(&child_ctx->counter_list)) goto again; + + mutex_unlock(&child_ctx->mutex); + + put_ctx(child_ctx); } /* @@ -3382,19 +3430,25 @@ void perf_counter_init_task(struct task_struct *child) struct perf_counter *counter; struct task_struct *parent = current; - child_ctx = &child->perf_counter_ctx; - parent_ctx = &parent->perf_counter_ctx; - - __perf_counter_init_context(child_ctx, child); + child->perf_counter_ctxp = NULL; /* * This is executed from the parent task context, so inherit - * counters that have been marked for cloning: + * counters that have been marked for cloning. + * First allocate and initialize a context for the child. */ - if (likely(!parent_ctx->nr_counters)) + child_ctx = kmalloc(sizeof(struct perf_counter_context), GFP_KERNEL); + if (!child_ctx) + return; + + parent_ctx = parent->perf_counter_ctxp; + if (likely(!parent_ctx || !parent_ctx->nr_counters)) return; + __perf_counter_init_context(child_ctx, child); + child->perf_counter_ctxp = child_ctx; + /* * Lock the parent list. No need to lock the child - not PID * hashed yet and not running, so nobody can access it. -- cgit v1.2.3-70-g09d2 From 48e22d56ecdeddd1ffb42a02fccba5c6ef42b133 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 25 May 2009 17:39:04 +0200 Subject: perf_counter: x86: Remove interrupt throttle remove the x86 specific interrupt throttle Signed-off-by: Peter Zijlstra Cc: Paul Mackerras Cc: Corey Ashford Cc: Arnaldo Carvalho de Melo Cc: John Kacur LKML-Reference: <20090525153931.616671838@chello.nl> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 2 -- arch/x86/kernel/cpu/perf_counter.c | 47 ++++---------------------------------- include/linux/perf_counter.h | 2 -- 3 files changed, 5 insertions(+), 46 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index b4f64402a82..89b63b5fad3 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -763,8 +763,6 @@ static void local_apic_timer_interrupt(void) inc_irq_stat(apic_timer_irqs); evt->event_handler(evt); - - perf_counter_unthrottle(); } /* diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index c14437faf5d..8c8177f859f 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -718,11 +718,6 @@ static void intel_pmu_save_and_restart(struct perf_counter *counter) intel_pmu_enable_counter(hwc, idx); } -/* - * Maximum interrupt frequency of 100KHz per CPU - */ -#define PERFMON_MAX_INTERRUPTS (100000/HZ) - /* * This handler is triggered by the local APIC, so the APIC IRQ handling * rules apply: @@ -775,15 +770,14 @@ again: if (status) goto again; - if (++cpuc->interrupts != PERFMON_MAX_INTERRUPTS) - perf_enable(); + perf_enable(); return 1; } static int amd_pmu_handle_irq(struct pt_regs *regs, int nmi) { - int cpu, idx, throttle = 0, handled = 0; + int cpu, idx, handled = 0; struct cpu_hw_counters *cpuc; struct perf_counter *counter; struct hw_perf_counter *hwc; @@ -792,16 +786,7 @@ static int amd_pmu_handle_irq(struct pt_regs *regs, int nmi) cpu = smp_processor_id(); cpuc = &per_cpu(cpu_hw_counters, cpu); - if (++cpuc->interrupts == PERFMON_MAX_INTERRUPTS) { - throttle = 1; - __perf_disable(); - cpuc->enabled = 0; - barrier(); - } - for (idx = 0; idx < x86_pmu.num_counters; idx++) { - int disable = 0; - if (!test_bit(idx, cpuc->active_mask)) continue; @@ -809,45 +794,23 @@ static int amd_pmu_handle_irq(struct pt_regs *regs, int nmi) hwc = &counter->hw; if (counter->hw_event.nmi != nmi) - goto next; + continue; val = x86_perf_counter_update(counter, hwc, idx); if (val & (1ULL << (x86_pmu.counter_bits - 1))) - goto next; + continue; /* counter overflow */ x86_perf_counter_set_period(counter, hwc, idx); handled = 1; inc_irq_stat(apic_perf_irqs); - disable = perf_counter_overflow(counter, nmi, regs, 0); - -next: - if (disable || throttle) + if (perf_counter_overflow(counter, nmi, regs, 0)) amd_pmu_disable_counter(hwc, idx); } return handled; } -void perf_counter_unthrottle(void) -{ - struct cpu_hw_counters *cpuc; - - if (!x86_pmu_initialized()) - return; - - cpuc = &__get_cpu_var(cpu_hw_counters); - if (cpuc->interrupts >= PERFMON_MAX_INTERRUPTS) { - /* - * Clear them before re-enabling irqs/NMIs again: - */ - cpuc->interrupts = 0; - perf_enable(); - } else { - cpuc->interrupts = 0; - } -} - void smp_perf_counter_interrupt(struct pt_regs *regs) { irq_enter(); diff --git a/include/linux/perf_counter.h b/include/linux/perf_counter.h index d3e85de9bf1..0c160be2078 100644 --- a/include/linux/perf_counter.h +++ b/include/linux/perf_counter.h @@ -570,7 +570,6 @@ extern int perf_counter_init_task(struct task_struct *child); extern void perf_counter_exit_task(struct task_struct *child); extern void perf_counter_do_pending(void); extern void perf_counter_print_debug(void); -extern void perf_counter_unthrottle(void); extern void __perf_disable(void); extern bool __perf_enable(void); extern void perf_disable(void); @@ -635,7 +634,6 @@ static inline int perf_counter_init_task(struct task_struct *child) { } static inline void perf_counter_exit_task(struct task_struct *child) { } static inline void perf_counter_do_pending(void) { } static inline void perf_counter_print_debug(void) { } -static inline void perf_counter_unthrottle(void) { } static inline void perf_disable(void) { } static inline void perf_enable(void) { } static inline int perf_counter_task_disable(void) { return -EINVAL; } -- cgit v1.2.3-70-g09d2 From 4efc0670baf4b14bc95502e54a83ccf639146125 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Tue, 28 Apr 2009 19:07:31 +0200 Subject: x86, mce: use 64bit machine check code on 32bit The 64bit machine check code is in many ways much better than the 32bit machine check code: it is more specification compliant, is cleaner, only has a single code base versus one per CPU, has better infrastructure for recovery, has a cleaner way to communicate with user space etc. etc. Use the 64bit code for 32bit too. This is the second attempt to do this. There was one a couple of years ago to unify this code for 32bit and 64bit. Back then this ran into some trouble with K7s and was reverted. I believe this time the K7 problems (and some others) are addressed. I went over the old handlers and was very careful to retain all quirks. But of course this needs a lot of testing on old systems. On newer 64bit capable systems I don't expect much problems because they have been already tested with the 64bit kernel. I made this a CONFIG for now that still allows to select the old machine check code. This is mostly to make testing easier, if someone runs into a problem we can ask them to try with the CONFIG switched. The new code is default y for more coverage. Once there is confidence the 64bit code works well on older hardware too the CONFIG_X86_OLD_MCE and the associated code can be easily removed. This causes a behaviour change for 32bit installations. They now have to install the mcelog package to be able to log corrected machine checks. The 64bit machine check code only handles CPUs which support the standard Intel machine check architecture described in the IA32 SDM. The 32bit code has special support for some older CPUs which have non standard machine check architectures, in particular WinChip C3 and Intel P5. I made those a separate CONFIG option and kept them for now. The WinChip variant could be probably removed without too much pain, it doesn't really do anything interesting. P5 is also disabled by default (like it was before) because many motherboards have it miswired, but according to Alan Cox a few embedded setups use that one. Forward ported/heavily changed version of old patch, original patch included review/fixes from Thomas Gleixner, Bert Wesarg. Signed-off-by: Andi Kleen Signed-off-by: H. Peter Anvin Signed-off-by: Hidetoshi Seto Signed-off-by: H. Peter Anvin --- arch/x86/Kconfig | 33 +++++++++++++++++++++++++++++++-- arch/x86/include/asm/entry_arch.h | 2 +- arch/x86/kernel/apic/apic.c | 4 ++-- arch/x86/kernel/apic/nmi.c | 2 +- arch/x86/kernel/cpu/mcheck/Makefile | 3 ++- arch/x86/kernel/cpu/mcheck/mce.c | 32 ++++++++++++++++++++++++++++---- arch/x86/kernel/cpu/mcheck/mce.h | 18 +++++++++++++++--- arch/x86/kernel/cpu/mcheck/p5.c | 5 +++++ arch/x86/kernel/irq.c | 4 ++-- arch/x86/kernel/irqinit_32.c | 2 +- arch/x86/kernel/signal.c | 4 ++-- arch/x86/kernel/traps.c | 4 ++-- 12 files changed, 92 insertions(+), 21 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index a6efe0a2e9a..c1c5ccd1937 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -789,6 +789,22 @@ config X86_MCE to disable it. MCE support simply ignores non-MCE processors like the 386 and 486, so nearly everyone can say Y here. +config X86_OLD_MCE + depends on X86_32 && X86_MCE + bool "Use legacy machine check code (will go away)" + default n + select X86_ANCIENT_MCE + ---help--- + Use the old i386 machine check code. This is merely intended for + testing in a transition period. Try this if you run into any machine + check related software problems, but report the problem to + linux-kernel. When in doubt say no. + +config X86_NEW_MCE + depends on X86_MCE + bool + default y if (!X86_OLD_MCE && X86_32) || X86_64 + config X86_MCE_INTEL def_bool y prompt "Intel MCE features" @@ -805,6 +821,15 @@ config X86_MCE_AMD Additional support for AMD specific MCE features such as the DRAM Error Threshold. +config X86_ANCIENT_MCE + def_bool n + depends on X86_32 + prompt "Support for old Pentium 5 / WinChip machine checks" + ---help--- + Include support for machine check handling on old Pentium 5 or WinChip + systems. These typically need to be enabled explicitely on the command + line. + config X86_MCE_THRESHOLD depends on X86_MCE_AMD || X86_MCE_INTEL bool @@ -812,7 +837,7 @@ config X86_MCE_THRESHOLD config X86_MCE_NONFATAL tristate "Check for non-fatal errors on AMD Athlon/Duron / Intel Pentium 4" - depends on X86_32 && X86_MCE + depends on X86_OLD_MCE ---help--- Enabling this feature starts a timer that triggers every 5 seconds which will look at the machine check registers to see if anything happened. @@ -825,11 +850,15 @@ config X86_MCE_NONFATAL config X86_MCE_P4THERMAL bool "check for P4 thermal throttling interrupt." - depends on X86_32 && X86_MCE && (X86_UP_APIC || SMP) + depends on X86_OLD_MCE && X86_MCE && (X86_UP_APIC || SMP) ---help--- Enabling this feature will cause a message to be printed when the P4 enters thermal throttling. +config X86_THERMAL_VECTOR + def_bool y + depends on X86_MCE_P4THERMAL || X86_MCE_INTEL + config VM86 bool "Enable VM86 support" if EMBEDDED default y diff --git a/arch/x86/include/asm/entry_arch.h b/arch/x86/include/asm/entry_arch.h index c2e6bedaf25..486c9e946f5 100644 --- a/arch/x86/include/asm/entry_arch.h +++ b/arch/x86/include/asm/entry_arch.h @@ -52,7 +52,7 @@ BUILD_INTERRUPT(spurious_interrupt,SPURIOUS_APIC_VECTOR) BUILD_INTERRUPT(perf_counter_interrupt, LOCAL_PERF_VECTOR) #endif -#ifdef CONFIG_X86_MCE_P4THERMAL +#ifdef CONFIG_X86_THERMAL_VECTOR BUILD_INTERRUPT(thermal_interrupt,THERMAL_APIC_VECTOR) #endif diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index f2870920f24..ad532289ef2 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -843,7 +843,7 @@ void clear_local_APIC(void) } /* lets not touch this if we didn't frob it */ -#if defined(CONFIG_X86_MCE_P4THERMAL) || defined(CONFIG_X86_MCE_INTEL) +#ifdef CONFIG_X86_THERMAL_VECTOR if (maxlvt >= 5) { v = apic_read(APIC_LVTTHMR); apic_write(APIC_LVTTHMR, v | APIC_LVT_MASKED); @@ -1962,7 +1962,7 @@ static int lapic_suspend(struct sys_device *dev, pm_message_t state) 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); -#if defined(CONFIG_X86_MCE_P4THERMAL) || defined(CONFIG_X86_MCE_INTEL) +#ifdef CONFIG_X86_THERMAL_VECTOR if (maxlvt >= 5) apic_pm_state.apic_thmr = apic_read(APIC_LVTTHMR); #endif diff --git a/arch/x86/kernel/apic/nmi.c b/arch/x86/kernel/apic/nmi.c index ce4fbfa315a..c4762276c17 100644 --- a/arch/x86/kernel/apic/nmi.c +++ b/arch/x86/kernel/apic/nmi.c @@ -66,7 +66,7 @@ static inline unsigned int get_nmi_count(int cpu) static inline int mce_in_progress(void) { -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) +#if defined(CONFIG_X86_NEW_MCE) return atomic_read(&mce_entry) > 0; #endif return 0; diff --git a/arch/x86/kernel/cpu/mcheck/Makefile b/arch/x86/kernel/cpu/mcheck/Makefile index 55f01b39a10..5f8b09425d3 100644 --- a/arch/x86/kernel/cpu/mcheck/Makefile +++ b/arch/x86/kernel/cpu/mcheck/Makefile @@ -1,6 +1,7 @@ obj-y = mce.o therm_throt.o -obj-$(CONFIG_X86_32) += k7.o p4.o p5.o p6.o winchip.o +obj-$(CONFIG_X86_OLD_MCE) += k7.o p4.o p6.o +obj-$(CONFIG_X86_ANCIENT_MCE) += winchip.o p5.o obj-$(CONFIG_X86_MCE_P4THERMAL) += mce_intel.o obj-$(CONFIG_X86_MCE_INTEL) += mce_intel_64.o mce_intel.o obj-$(CONFIG_X86_MCE_AMD) += mce_amd_64.o diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index f4d6841d2bd..e193de44ef1 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -52,7 +52,7 @@ void (*machine_check_vector)(struct pt_regs *, long error_code) = int mce_disabled; -#ifdef CONFIG_X86_64 +#ifdef CONFIG_X86_NEW_MCE #define MISC_MCELOG_MINOR 227 @@ -662,6 +662,21 @@ static void mce_cpu_quirks(struct cpuinfo_x86 *c) } } +static void __cpuinit mce_ancient_init(struct cpuinfo_x86 *c) +{ + if (c->x86 != 5) + return; + switch (c->x86_vendor) { + case X86_VENDOR_INTEL: + if (mce_p5_enabled()) + intel_p5_mcheck_init(c); + break; + case X86_VENDOR_CENTAUR: + winchip_mcheck_init(c); + break; + } +} + static void mce_cpu_features(struct cpuinfo_x86 *c) { switch (c->x86_vendor) { @@ -695,6 +710,11 @@ static void mce_init_timer(void) */ void __cpuinit mcheck_init(struct cpuinfo_x86 *c) { + if (mce_disabled) + return; + + mce_ancient_init(c); + if (!mce_available(c)) return; @@ -893,6 +913,10 @@ static struct miscdevice mce_log_device = { */ static int __init mcheck_enable(char *str) { + if (*str == 0) + enable_p5_mce(); + if (*str == '=') + str++; if (!strcmp(str, "off")) mce_disabled = 1; else if (!strcmp(str, "bootlog") || !strcmp(str, "nobootlog")) @@ -900,13 +924,13 @@ static int __init mcheck_enable(char *str) else if (isdigit(str[0])) get_option(&str, &tolerant); else { - printk(KERN_INFO "mce= argument %s ignored. Please use /sys\n", + printk(KERN_INFO "mce argument %s ignored. Please use /sys\n", str); return 0; } return 1; } -__setup("mce=", mcheck_enable); +__setup("mce", mcheck_enable); /* * Sysfs support @@ -1259,7 +1283,7 @@ static __init int mce_init_device(void) device_initcall(mce_init_device); -#else /* CONFIG_X86_32: */ +#else /* CONFIG_X86_OLD_MCE: */ int nr_mce_banks; EXPORT_SYMBOL_GPL(nr_mce_banks); /* non-fatal.o */ diff --git a/arch/x86/kernel/cpu/mcheck/mce.h b/arch/x86/kernel/cpu/mcheck/mce.h index 966ae3c5cb1..84a552b458c 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.h +++ b/arch/x86/kernel/cpu/mcheck/mce.h @@ -1,17 +1,29 @@ #include #include +#ifdef CONFIG_X86_OLD_MCE 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); +#endif +#ifdef CONFIG_X86_ANCIENT_MCE +void intel_p5_mcheck_init(struct cpuinfo_x86 *c); +void winchip_mcheck_init(struct cpuinfo_x86 *c); +extern int mce_p5_enable; +static inline int mce_p5_enabled(void) { return mce_p5_enable; } +static inline void enable_p5_mce(void) { mce_p5_enable = 1; } +#else +static inline void intel_p5_mcheck_init(struct cpuinfo_x86 *c) {} +static inline void winchip_mcheck_init(struct cpuinfo_x86 *c) {} +static inline int mce_p5_enabled(void) { return 0; } +static inline void enable_p5_mce(void) { } +#endif /* Call the installed machine check handler for this CPU setup. */ extern void (*machine_check_vector)(struct pt_regs *, long error_code); -#ifdef CONFIG_X86_32 +#ifdef CONFIG_X86_OLD_MCE extern int nr_mce_banks; diff --git a/arch/x86/kernel/cpu/mcheck/p5.c b/arch/x86/kernel/cpu/mcheck/p5.c index 8812f544183..015f481ab1b 100644 --- a/arch/x86/kernel/cpu/mcheck/p5.c +++ b/arch/x86/kernel/cpu/mcheck/p5.c @@ -14,6 +14,9 @@ #include "mce.h" +/* By default disabled */ +int mce_p5_enable; + /* Machine check handler for Pentium class Intel CPUs: */ static void pentium_machine_check(struct pt_regs *regs, long error_code) { @@ -44,9 +47,11 @@ void intel_p5_mcheck_init(struct cpuinfo_x86 *c) if (!cpu_has(c, X86_FEATURE_MCE)) return; +#ifdef CONFIG_X86_OLD_MCE /* Default P5 to off as its often misconnected: */ if (mce_disabled != -1) return; +#endif machine_check_vector = pentium_machine_check; /* Make sure the vector pointer is visible before we enable MCEs: */ diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c index c3fe010d74c..35eddc9ec99 100644 --- a/arch/x86/kernel/irq.c +++ b/arch/x86/kernel/irq.c @@ -89,7 +89,7 @@ static int show_other_interrupts(struct seq_file *p, int prec) for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_thermal_count); seq_printf(p, " Thermal event interrupts\n"); -# ifdef CONFIG_X86_64 +# ifdef CONFIG_X86_MCE_THRESHOLD seq_printf(p, "%*s: ", prec, "THR"); for_each_online_cpu(j) seq_printf(p, "%10u ", irq_stats(j)->irq_threshold_count); @@ -176,7 +176,7 @@ u64 arch_irq_stat_cpu(unsigned int cpu) #endif #ifdef CONFIG_X86_MCE sum += irq_stats(cpu)->irq_thermal_count; -# ifdef CONFIG_X86_64 +# ifdef CONFIG_X86_MCE_THRESHOLD sum += irq_stats(cpu)->irq_threshold_count; #endif #endif diff --git a/arch/x86/kernel/irqinit_32.c b/arch/x86/kernel/irqinit_32.c index 368b0a8836f..98846e03211 100644 --- a/arch/x86/kernel/irqinit_32.c +++ b/arch/x86/kernel/irqinit_32.c @@ -181,7 +181,7 @@ void __init native_init_IRQ(void) alloc_intr_gate(ERROR_APIC_VECTOR, error_interrupt); #endif -#if defined(CONFIG_X86_LOCAL_APIC) && defined(CONFIG_X86_MCE_P4THERMAL) +#ifdef CONFIG_X86_THERMAL_VECTOR /* thermal monitor LVT interrupt */ alloc_intr_gate(THERMAL_APIC_VECTOR, thermal_interrupt); #endif diff --git a/arch/x86/kernel/signal.c b/arch/x86/kernel/signal.c index 14425166b8e..d0851e3f77e 100644 --- a/arch/x86/kernel/signal.c +++ b/arch/x86/kernel/signal.c @@ -25,11 +25,11 @@ #include #include #include +#include #ifdef CONFIG_X86_64 #include #include -#include #endif /* CONFIG_X86_64 */ #include @@ -857,7 +857,7 @@ static void do_signal(struct pt_regs *regs) void do_notify_resume(struct pt_regs *regs, void *unused, __u32 thread_info_flags) { -#if defined(CONFIG_X86_64) && defined(CONFIG_X86_MCE) +#ifdef CONFIG_X86_NEW_MCE /* notify userspace of pending MCEs */ if (thread_info_flags & _TIF_MCE_NOTIFY) mce_notify_user(); diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index a1d288327ff..ad771f15bdd 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -798,7 +798,8 @@ unsigned long patch_espfix_desc(unsigned long uesp, unsigned long kesp) return new_kesp; } -#else +#endif + asmlinkage void __attribute__((weak)) smp_thermal_interrupt(void) { } @@ -806,7 +807,6 @@ asmlinkage void __attribute__((weak)) smp_thermal_interrupt(void) asmlinkage void __attribute__((weak)) mce_threshold_interrupt(void) { } -#endif /* * 'math_state_restore()' saves the current math information in the -- cgit v1.2.3-70-g09d2 From c323d95fa4dbe0b6bf6d59e24a0b7db067dd08a7 Mon Sep 17 00:00:00 2001 From: Yong Wang Date: Fri, 29 May 2009 13:28:35 +0800 Subject: perf_counter/x86: Always use NMI for performance-monitoring interrupt Always use NMI for performance-monitoring interrupt as there could be racy situations if we switch between irq and nmi mode frequently. Signed-off-by: Yong Wang LKML-Reference: <20090529052835.GA13657@ywang-moblin2.bj.intel.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/perf_counter.h | 4 ++-- arch/x86/kernel/apic/apic.c | 2 +- arch/x86/kernel/cpu/perf_counter.c | 19 +++++-------------- 3 files changed, 8 insertions(+), 17 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/perf_counter.h b/arch/x86/include/asm/perf_counter.h index d08dd52cb8f..876ed97147b 100644 --- a/arch/x86/include/asm/perf_counter.h +++ b/arch/x86/include/asm/perf_counter.h @@ -91,10 +91,10 @@ extern void set_perf_counter_pending(void); #ifdef CONFIG_PERF_COUNTERS extern void init_hw_perf_counters(void); -extern void perf_counters_lapic_init(int nmi); +extern void perf_counters_lapic_init(void); #else static inline void init_hw_perf_counters(void) { } -static inline void perf_counters_lapic_init(int nmi) { } +static inline void perf_counters_lapic_init(void) { } #endif #endif /* _ASM_X86_PERF_COUNTER_H */ diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index 89b63b5fad3..60df2efd7c8 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -1135,7 +1135,7 @@ void __cpuinit setup_local_APIC(void) apic_write(APIC_ESR, 0); } #endif - perf_counters_lapic_init(0); + perf_counters_lapic_init(); preempt_disable(); diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c index 2eeaa99add1..316b0c995f3 100644 --- a/arch/x86/kernel/cpu/perf_counter.c +++ b/arch/x86/kernel/cpu/perf_counter.c @@ -604,7 +604,7 @@ try_generic: hwc->counter_base = x86_pmu.perfctr; } - perf_counters_lapic_init(hwc->nmi); + perf_counters_lapic_init(); x86_pmu.disable(hwc, idx); @@ -863,24 +863,15 @@ void set_perf_counter_pending(void) apic->send_IPI_self(LOCAL_PENDING_VECTOR); } -void perf_counters_lapic_init(int nmi) +void perf_counters_lapic_init(void) { - u32 apic_val; - if (!x86_pmu_initialized()) return; /* - * Enable the performance counter vector in the APIC LVT: + * Always use NMI for PMU */ - apic_val = apic_read(APIC_LVTERR); - - apic_write(APIC_LVTERR, apic_val | APIC_LVT_MASKED); - if (nmi) - apic_write(APIC_LVTPC, APIC_DM_NMI); - else - apic_write(APIC_LVTPC, LOCAL_PERF_VECTOR); - apic_write(APIC_LVTERR, apic_val); + apic_write(APIC_LVTPC, APIC_DM_NMI); } static int __kprobes @@ -1054,7 +1045,7 @@ void __init init_hw_perf_counters(void) pr_info("... counter mask: %016Lx\n", perf_counter_mask); - perf_counters_lapic_init(0); + perf_counters_lapic_init(); register_die_notifier(&perf_counter_nmi_notifier); } -- cgit v1.2.3-70-g09d2 From 58f892e022e88438183c48661dcdc6a2997dab99 Mon Sep 17 00:00:00 2001 From: Naga Chumbalkar Date: Tue, 26 May 2009 21:48:07 +0000 Subject: x86: Print real IOAPIC version for x86-64 Fix the fact that the IOAPIC version number in the x86_64 code path always gets assigned to 0, instead of the correct value. Before the patch: (from "dmesg" output): ACPI: IOAPIC (id[0x08] address[0xfec00000] gsi_base[0]) IOAPIC[0]: apic_id 8, version 0, address 0xfec00000, GSI 0-23 <--- After the patch: ACPI: IOAPIC (id[0x08] address[0xfec00000] gsi_base[0]) IOAPIC[0]: apic_id 8, version 32, address 0xfec00000, GSI 0-23 <--- History: io_apic_get_version() was compiled out of the x86_64 code path in the commit f2c2cca3acef8b253a36381d9b469ad4fb08563a: Author: Andi Kleen Date: Tue Sep 26 10:52:37 2006 +0200 [PATCH] Remove APIC version/cpu capability mpparse checking/printing ACPI went to great trouble to get the APIC version and CPU capabilities of different CPUs before passing them to the mpparser. But all that data was used was to print it out. Actually it even faked some data based on the boot cpu, not on the actual CPU being booted. Remove all this code because it's not needed. Cc: len.brown@intel.com At the time, the IOAPIC version number was deliberately not printed in the x86_64 code path. However, after the x86 and x86_64 files were merged, the net result is that the IOAPIC version is printed incorrectly in the x86_64 code path. The patch below provides a fix. I have tested it with acpi, and with acpi=off, and did not see any problems. Signed-off-by: Naga Chumbalkar Acked-by: Yinghai Lu LKML-Reference: <20090416014230.4885.94926.sendpatchset@localhost.localdomain> Signed-off-by: Ingo Molnar ************************* --- arch/x86/kernel/acpi/boot.c | 5 +---- arch/x86/kernel/apic/io_apic.c | 2 +- 2 files changed, 2 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index 844e5e25213..631086159c5 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -985,11 +985,8 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base) set_fixmap_nocache(FIX_IO_APIC_BASE_0 + idx, address); mp_ioapics[idx].apicid = uniq_ioapic_id(id); -#ifdef CONFIG_X86_32 mp_ioapics[idx].apicver = io_apic_get_version(idx); -#else - mp_ioapics[idx].apicver = 0; -#endif + /* * Build basic GSI lookup table to facilitate gsi->io_apic lookups * and to prevent reprogramming of IOAPIC pins (PCI GSIs). diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index ac7f3b6ad58..f712f8ff403 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -4012,6 +4012,7 @@ int __init io_apic_get_unique_id(int ioapic, int apic_id) return apic_id; } +#endif int __init io_apic_get_version(int ioapic) { @@ -4024,7 +4025,6 @@ int __init io_apic_get_version(int ioapic) return reg_01.bits.version; } -#endif int acpi_get_override_irq(int bus_irq, int *trigger, int *polarity) { -- cgit v1.2.3-70-g09d2 From 3d58829b0510244596079c1d2f1762c53aef2e97 Mon Sep 17 00:00:00 2001 From: Jiri Slaby Date: Thu, 28 May 2009 09:54:47 +0200 Subject: x86, apic: Restore irqs on fail paths lapic_resume forgets to restore interrupts on fail paths. Fix that. Signed-off-by: Jiri Slaby Acked-by: Cyrill Gorcunov LKML-Reference: <1243497289-18591-1-git-send-email-jirislaby@gmail.com> Signed-off-by: Ingo Molnar Cc: H. Peter Anvin --- arch/x86/kernel/apic/apic.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index b0fd26442c4..e82488d3f0b 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -2027,7 +2027,7 @@ static int lapic_resume(struct sys_device *dev) unsigned int l, h; unsigned long flags; int maxlvt; - int ret; + int ret = 0; struct IO_APIC_route_entry **ioapic_entries = NULL; if (!apic_pm_state.active) @@ -2038,14 +2038,15 @@ static int lapic_resume(struct sys_device *dev) ioapic_entries = alloc_ioapic_entries(); if (!ioapic_entries) { WARN(1, "Alloc ioapic_entries in lapic resume failed."); - return -ENOMEM; + ret = -ENOMEM; + goto restore; } ret = save_IO_APIC_setup(ioapic_entries); if (ret) { WARN(1, "Saving IO-APIC state failed: %d\n", ret); free_ioapic_entries(ioapic_entries); - return ret; + goto restore; } mask_IO_APIC_setup(ioapic_entries); @@ -2097,10 +2098,10 @@ static int lapic_resume(struct sys_device *dev) restore_IO_APIC_setup(ioapic_entries); free_ioapic_entries(ioapic_entries); } - +restore: local_irq_restore(flags); - return 0; + return ret; } /* -- cgit v1.2.3-70-g09d2 From 103428e57be323c3c5545db8ad12667099bc6005 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Sun, 7 Jun 2009 16:48:40 +0400 Subject: x86, apic: Fix dummy apic read operation together with broken MP handling Ingo Molnar reported that read_apic is buggy novadays: [ 0.000000] Using APIC driver default [ 0.000000] SMP: Allowing 1 CPUs, 0 hotplug CPUs [ 0.000000] Local APIC disabled by BIOS -- you can enable it with "lapic" [ 0.000000] APIC: disable apic facility [ 0.000000] ------------[ cut here ]------------ [ 0.000000] WARNING: at arch/x86/kernel/apic/apic.c:254 native_apic_read_dummy+0x2d/0x3b() [ 0.000000] Hardware name: HP OmniBook PC Indeed we still rely on apic->read operation for SMP compiled kernel. And instead of disfigure the SMP code with #ifdef we allow to call apic->read. To capture any unexpected results we check for apic->read being called for sane reason via WARN_ON_ONCE but(!) instead of OR we should use AND logical operation (thanks Yinghai for spotting the root of the problem). Along with that we could be have bad MP table and we are to fix it that way no SMP started and no complains about BIOS bug if apic was just disabled via command line. Signed-off-by: Cyrill Gorcunov Cc: Yinghai Lu LKML-Reference: <20090607124840.GD4547@lenovo> Signed-off-by: Ingo Molnar --- arch/x86/kernel/apic/apic.c | 9 ++++++++- arch/x86/kernel/smpboot.c | 8 +++++--- 2 files changed, 13 insertions(+), 4 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index e82488d3f0b..a4c9cf0bf70 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -249,7 +249,7 @@ static void native_apic_write_dummy(u32 reg, u32 v) static u32 native_apic_read_dummy(u32 reg) { - WARN_ON_ONCE((cpu_has_apic || !disable_apic)); + WARN_ON_ONCE((cpu_has_apic && !disable_apic)); return 0; } @@ -1609,6 +1609,13 @@ void __init init_apic_mappings(void) new_apicid = read_apic_id(); if (boot_cpu_physical_apicid != new_apicid) { boot_cpu_physical_apicid = new_apicid; + /* + * yeah -- we lie about apic_version + * in case if apic was disabled via boot option + * but it's not a problem for SMP compiled kernel + * since smp_sanity_check is prepared for such a case + * and disable smp mode + */ apic_version[new_apicid] = GET_APIC_VERSION(apic_read(APIC_LVR)); } diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c index d2e8de95815..7c80007ea5f 100644 --- a/arch/x86/kernel/smpboot.c +++ b/arch/x86/kernel/smpboot.c @@ -992,10 +992,12 @@ static int __init smp_sanity_check(unsigned max_cpus) */ 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." + if (!disable_apic) { + pr_err("BIOS bug, local APIC #%d not detected!...\n", + boot_cpu_physical_apicid); + pr_err("... forcing use of dummy APIC emulation." "(tell your hw vendor)\n"); + } smpboot_clear_io_apic(); arch_disable_smp_support(); return -1; -- cgit v1.2.3-70-g09d2 From c4ed3f04ba9defe22aa729d1646f970f791c03d7 Mon Sep 17 00:00:00 2001 From: Jack Steiner Date: Mon, 8 Jun 2009 10:44:05 -0500 Subject: x86, UV: Fix macros for multiple coherency domains Fix bug in the SGI UV macros that support systems with multiple coherency domains. The macros used for referencing global MMR (chipset registers) are failing to correctly "or" the NASID (node identifier) bits that reside above M+N. These high bits are supplied automatically by the chipset for memory accesses coming from the processor socket. However, the bits must be present for references to the special global MMR space used to map chipset registers. (See uv_hub.h for more details ...) The bug results in references to invalid/incorrect nodes. Signed-off-by: Jack Steiner Cc: LKML-Reference: <20090608154405.GA16395@sgi.com> Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_hub.h | 6 ++++-- arch/x86/kernel/apic/x2apic_uv_x.c | 15 +++++++++------ 2 files changed, 13 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index d3a98ea1062..341070f7ad5 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -133,6 +133,7 @@ struct uv_scir_s { struct uv_hub_info_s { unsigned long global_mmr_base; unsigned long gpa_mask; + unsigned int gnode_extra; unsigned long gnode_upper; unsigned long lowmem_remap_top; unsigned long lowmem_remap_base; @@ -159,7 +160,8 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); * p - PNODE (local part of nsids, right shifted 1) */ #define UV_NASID_TO_PNODE(n) (((n) >> 1) & uv_hub_info->pnode_mask) -#define UV_PNODE_TO_NASID(p) (((p) << 1) | uv_hub_info->gnode_upper) +#define UV_PNODE_TO_GNODE(p) ((p) |uv_hub_info->gnode_extra) +#define UV_PNODE_TO_NASID(p) (UV_PNODE_TO_GNODE(p) << 1) #define UV_LOCAL_MMR_BASE 0xf4000000UL #define UV_GLOBAL_MMR32_BASE 0xf8000000UL @@ -173,7 +175,7 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); #define UV_GLOBAL_MMR32_PNODE_BITS(p) ((p) << (UV_GLOBAL_MMR32_PNODE_SHIFT)) #define UV_GLOBAL_MMR64_PNODE_BITS(p) \ - ((unsigned long)(p) << UV_GLOBAL_MMR64_PNODE_SHIFT) + ((unsigned long)(UV_PNODE_TO_GNODE(p)) << UV_GLOBAL_MMR64_PNODE_SHIFT) #define UV_APIC_PNODE_SHIFT 6 diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 2bda6935297..39f2af4b546 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -562,7 +562,7 @@ void __init uv_system_init(void) union uvh_node_id_u node_id; unsigned long gnode_upper, lowmem_redir_base, lowmem_redir_size; int bytes, nid, cpu, lcpu, pnode, blade, i, j, m_val, n_val; - int max_pnode = 0; + int gnode_extra, max_pnode = 0; unsigned long mmr_base, present, paddr; unsigned short pnode_mask; @@ -574,6 +574,13 @@ void __init uv_system_init(void) mmr_base = uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) & ~UV_MMR_ENABLE; + pnode_mask = (1 << n_val) - 1; + node_id.v = uv_read_local_mmr(UVH_NODE_ID); + gnode_extra = (node_id.s.node_id & ~((1 << n_val) - 1)) >> 1; + gnode_upper = ((unsigned long)gnode_extra << m_val); + printk(KERN_DEBUG "UV: N %d, M %d, gnode_upper 0x%lx, gnode_extra 0x%x\n", + n_val, m_val, gnode_upper, gnode_extra); + printk(KERN_DEBUG "UV: global MMR base 0x%lx\n", mmr_base); for(i = 0; i < UVH_NODE_PRESENT_TABLE_DEPTH; i++) @@ -607,11 +614,6 @@ void __init uv_system_init(void) } } - pnode_mask = (1 << n_val) - 1; - node_id.v = uv_read_local_mmr(UVH_NODE_ID); - gnode_upper = (((unsigned long)node_id.s.node_id) & - ~((1 << n_val) - 1)) << m_val; - uv_bios_init(); uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, &sn_region_size); @@ -634,6 +636,7 @@ void __init uv_system_init(void) uv_cpu_hub_info(cpu)->pnode_mask = pnode_mask; uv_cpu_hub_info(cpu)->gpa_mask = (1 << (m_val + n_val)) - 1; uv_cpu_hub_info(cpu)->gnode_upper = gnode_upper; + uv_cpu_hub_info(cpu)->gnode_extra = gnode_extra; uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base; uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id; uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu; -- cgit v1.2.3-70-g09d2 From 38c7fed2f5ffee17e1fa3e0f78b0e1bf43d52d13 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Mon, 25 May 2009 15:10:58 +0300 Subject: x86: remove some alloc_bootmem_cpumask_var calling Now that we set up the slab allocator earlier, we can get rid of some alloc_bootmem_cpumask_var() calls in boot code. Cc: Ingo Molnar Cc: Johannes Weiner Cc: Linus Torvalds Signed-off-by: Yinghai Lu Signed-off-by: Pekka Enberg --- arch/x86/kernel/apic/io_apic.c | 4 ++-- include/linux/irq.h | 18 +++++++----------- kernel/cpuset.c | 2 +- kernel/profile.c | 6 ------ lib/cpumask.c | 11 ++--------- 5 files changed, 12 insertions(+), 29 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 1946fac42ab..139201a562b 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -185,8 +185,8 @@ int __init arch_early_irq_init(void) for (i = 0; i < count; i++) { desc = irq_to_desc(i); desc->chip_data = &cfg[i]; - alloc_bootmem_cpumask_var(&cfg[i].domain); - alloc_bootmem_cpumask_var(&cfg[i].old_domain); + alloc_cpumask_var(&cfg[i].domain, GFP_NOWAIT); + alloc_cpumask_var(&cfg[i].old_domain, GFP_NOWAIT); if (i < NR_IRQS_LEGACY) cpumask_setall(cfg[i].domain); } diff --git a/include/linux/irq.h b/include/linux/irq.h index eedbb8e5e0c..1e50c34f006 100644 --- a/include/linux/irq.h +++ b/include/linux/irq.h @@ -430,23 +430,19 @@ extern int set_irq_msi(unsigned int irq, struct msi_desc *entry); * Returns true if successful (or not required). */ static inline bool alloc_desc_masks(struct irq_desc *desc, int node, - bool boot) + bool boot) { -#ifdef CONFIG_CPUMASK_OFFSTACK - if (boot) { - alloc_bootmem_cpumask_var(&desc->affinity); + gfp_t gfp = GFP_ATOMIC; -#ifdef CONFIG_GENERIC_PENDING_IRQ - alloc_bootmem_cpumask_var(&desc->pending_mask); -#endif - return true; - } + if (boot) + gfp = GFP_NOWAIT; - if (!alloc_cpumask_var_node(&desc->affinity, GFP_ATOMIC, node)) +#ifdef CONFIG_CPUMASK_OFFSTACK + if (!alloc_cpumask_var_node(&desc->affinity, gfp, node)) return false; #ifdef CONFIG_GENERIC_PENDING_IRQ - if (!alloc_cpumask_var_node(&desc->pending_mask, GFP_ATOMIC, node)) { + if (!alloc_cpumask_var_node(&desc->pending_mask, gfp, node)) { free_cpumask_var(desc->affinity); return false; } diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 026faccca86..d5a7e17474e 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -1857,7 +1857,7 @@ struct cgroup_subsys cpuset_subsys = { int __init cpuset_init_early(void) { - alloc_bootmem_cpumask_var(&top_cpuset.cpus_allowed); + alloc_cpumask_var(&top_cpuset.cpus_allowed, GFP_NOWAIT); top_cpuset.mems_generation = cpuset_mems_generation++; return 0; diff --git a/kernel/profile.c b/kernel/profile.c index 7724e0409ba..28cf26ad2d2 100644 --- a/kernel/profile.c +++ b/kernel/profile.c @@ -111,12 +111,6 @@ int __ref profile_init(void) /* only text is profiled */ prof_len = (_etext - _stext) >> prof_shift; buffer_bytes = prof_len*sizeof(atomic_t); - if (!slab_is_available()) { - prof_buffer = alloc_bootmem(buffer_bytes); - alloc_bootmem_cpumask_var(&prof_cpu_mask); - cpumask_copy(prof_cpu_mask, cpu_possible_mask); - return 0; - } if (!alloc_cpumask_var(&prof_cpu_mask, GFP_KERNEL)) return -ENOMEM; diff --git a/lib/cpumask.c b/lib/cpumask.c index eb23aaa0c7b..7bb4142a502 100644 --- a/lib/cpumask.c +++ b/lib/cpumask.c @@ -92,15 +92,8 @@ int cpumask_any_but(const struct cpumask *mask, unsigned int cpu) */ bool alloc_cpumask_var_node(cpumask_var_t *mask, gfp_t flags, int node) { - if (likely(slab_is_available())) - *mask = kmalloc_node(cpumask_size(), flags, node); - else { -#ifdef CONFIG_DEBUG_PER_CPU_MAPS - printk(KERN_ERR - "=> alloc_cpumask_var: kmalloc not available!\n"); -#endif - *mask = NULL; - } + *mask = kmalloc_node(cpumask_size(), flags, node); + #ifdef CONFIG_DEBUG_PER_CPU_MAPS if (!*mask) { printk(KERN_ERR "=> alloc_cpumask_var: failed!\n"); -- cgit v1.2.3-70-g09d2 From dad213aeb59718623fc59defeff95fe8c3feb8a0 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 28 May 2009 18:14:40 -0700 Subject: irq/cpumask: make memoryless node zero happy Don't hardcode to node zero for early boot IRQ setup memory allocations. [ penberg@cs.helsinki.fi: minor cleanups ] Cc: Ingo Molnar Cc: Johannes Weiner Cc: Linus Torvalds Signed-off-by: Yinghai Lu Signed-off-by: Pekka Enberg --- arch/x86/kernel/apic/io_apic.c | 6 ++++-- kernel/irq/handle.c | 9 +++++---- 2 files changed, 9 insertions(+), 6 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 139201a562b..94605e7f6a5 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -177,16 +177,18 @@ int __init arch_early_irq_init(void) struct irq_cfg *cfg; struct irq_desc *desc; int count; + int node; int i; cfg = irq_cfgx; count = ARRAY_SIZE(irq_cfgx); + node= cpu_to_node(boot_cpu_id); for (i = 0; i < count; i++) { desc = irq_to_desc(i); desc->chip_data = &cfg[i]; - alloc_cpumask_var(&cfg[i].domain, GFP_NOWAIT); - alloc_cpumask_var(&cfg[i].old_domain, GFP_NOWAIT); + alloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); + alloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); if (i < NR_IRQS_LEGACY) cpumask_setall(cfg[i].domain); } diff --git a/kernel/irq/handle.c b/kernel/irq/handle.c index a60018402f4..e161999b668 100644 --- a/kernel/irq/handle.c +++ b/kernel/irq/handle.c @@ -150,6 +150,7 @@ int __init early_irq_init(void) { struct irq_desc *desc; int legacy_count; + int node; int i; init_irq_default_affinity(); @@ -160,20 +161,20 @@ int __init early_irq_init(void) desc = irq_desc_legacy; legacy_count = ARRAY_SIZE(irq_desc_legacy); + node = first_online_node; /* allocate irq_desc_ptrs array based on nr_irqs */ irq_desc_ptrs = alloc_bootmem(nr_irqs * sizeof(void *)); /* allocate based on nr_cpu_ids */ - /* FIXME: invert kstat_irgs, and it'd be a per_cpu_alloc'd thing */ - kstat_irqs_legacy = alloc_bootmem(NR_IRQS_LEGACY * nr_cpu_ids * - sizeof(int)); + kstat_irqs_legacy = kzalloc_node(NR_IRQS_LEGACY * nr_cpu_ids * + sizeof(int), GFP_NOWAIT, node); for (i = 0; i < legacy_count; i++) { desc[i].irq = i; desc[i].kstat_irqs = kstat_irqs_legacy + i * nr_cpu_ids; lockdep_set_class(&desc[i].lock, &irq_desc_lock_class); - alloc_desc_masks(&desc[i], 0, true); + alloc_desc_masks(&desc[i], node, true); init_desc_masks(&desc[i]); irq_desc_ptrs[i] = desc + i; } -- cgit v1.2.3-70-g09d2 From 12274e96b42884f043dfaa87eb1e5e10281bfac3 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 11 Jun 2009 15:07:48 -0700 Subject: x86: use zalloc_cpumask_var in arch_early_irq_init So we make sure MAXSMP gets a cleared cpumask Signed-off-by: Yinghai Lu Signed-off-by: Linus Torvalds --- arch/x86/kernel/apic/io_apic.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel/apic') diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 94605e7f6a5..ef8d9290c7e 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -187,8 +187,8 @@ int __init arch_early_irq_init(void) for (i = 0; i < count; i++) { desc = irq_to_desc(i); desc->chip_data = &cfg[i]; - alloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); - alloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); + zalloc_cpumask_var_node(&cfg[i].domain, GFP_NOWAIT, node); + zalloc_cpumask_var_node(&cfg[i].old_domain, GFP_NOWAIT, node); if (i < NR_IRQS_LEGACY) cpumask_setall(cfg[i].domain); } -- cgit v1.2.3-70-g09d2