From b534816b552d35bbd3c60702139ed5c7da2f55c2 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Wed, 4 Feb 2009 18:33:38 -0800 Subject: x86: don't apply __supported_pte_mask to non-present ptes On an x86 system which doesn't support global mappings, __supported_pte_mask has _PAGE_GLOBAL clear, to make sure it never appears in the PTE. pfn_pte() and so on will enforce it with: static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot) { return __pte((((phys_addr_t)page_nr << PAGE_SHIFT) | pgprot_val(pgprot)) & __supported_pte_mask); } However, we overload _PAGE_GLOBAL with _PAGE_PROTNONE on non-present ptes to distinguish them from swap entries. However, applying __supported_pte_mask indiscriminately will clear the bit and corrupt the pte. I guess the best fix is to only apply __supported_pte_mask to present ptes. This seems like the right solution to me, as it means we can completely ignore the issue of overlaps between the present pte bits and the non-present pte-as-swap entry use of the bits. __supported_pte_mask contains the set of flags we support on the current hardware. We also use bits in the pte for things like logically present ptes with no permissions, and swap entries for swapped out pages. We should only apply __supported_pte_mask to present ptes, because otherwise we may destroy other information being stored in the ptes. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: H. Peter Anvin --- arch/x86/include/asm/pgtable.h | 26 ++++++++++++++++++++------ arch/x86/include/asm/xen/page.h | 2 +- 2 files changed, 21 insertions(+), 7 deletions(-) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index 06bbcbd66e9..4f5af8447d5 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -302,16 +302,30 @@ static inline pte_t pte_mkspecial(pte_t pte) extern pteval_t __supported_pte_mask; +/* + * Mask out unsupported bits in a present pgprot. Non-present pgprots + * can use those bits for other purposes, so leave them be. + */ +static inline pgprotval_t massage_pgprot(pgprot_t pgprot) +{ + pgprotval_t protval = pgprot_val(pgprot); + + if (protval & _PAGE_PRESENT) + protval &= __supported_pte_mask; + + return protval; +} + static inline pte_t pfn_pte(unsigned long page_nr, pgprot_t pgprot) { - return __pte((((phys_addr_t)page_nr << PAGE_SHIFT) | - pgprot_val(pgprot)) & __supported_pte_mask); + return __pte(((phys_addr_t)page_nr << PAGE_SHIFT) | + massage_pgprot(pgprot)); } static inline pmd_t pfn_pmd(unsigned long page_nr, pgprot_t pgprot) { - return __pmd((((phys_addr_t)page_nr << PAGE_SHIFT) | - pgprot_val(pgprot)) & __supported_pte_mask); + return __pmd(((phys_addr_t)page_nr << PAGE_SHIFT) | + massage_pgprot(pgprot)); } static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) @@ -323,7 +337,7 @@ static inline pte_t pte_modify(pte_t pte, pgprot_t newprot) * the newprot (if present): */ val &= _PAGE_CHG_MASK; - val |= pgprot_val(newprot) & (~_PAGE_CHG_MASK) & __supported_pte_mask; + val |= massage_pgprot(newprot) & ~_PAGE_CHG_MASK; return __pte(val); } @@ -339,7 +353,7 @@ static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) #define pte_pgprot(x) __pgprot(pte_flags(x) & PTE_FLAGS_MASK) -#define canon_pgprot(p) __pgprot(pgprot_val(p) & __supported_pte_mask) +#define canon_pgprot(p) __pgprot(massage_pgprot(p)) static inline int is_new_memtype_allowed(unsigned long flags, unsigned long new_flags) diff --git a/arch/x86/include/asm/xen/page.h b/arch/x86/include/asm/xen/page.h index 7ef617ef1df..4bd990ee43d 100644 --- a/arch/x86/include/asm/xen/page.h +++ b/arch/x86/include/asm/xen/page.h @@ -137,7 +137,7 @@ static inline pte_t mfn_pte(unsigned long page_nr, pgprot_t pgprot) pte_t pte; pte.pte = ((phys_addr_t)page_nr << PAGE_SHIFT) | - (pgprot_val(pgprot) & __supported_pte_mask); + massage_pgprot(pgprot); return pte; } -- cgit v1.2.3-70-g09d2 From e736ad548db152776de61d7a26805cfae77ce5ce Mon Sep 17 00:00:00 2001 From: "Pallipadi, Venkatesh" Date: Fri, 6 Feb 2009 16:52:05 -0800 Subject: x86: add clflush before monitor for Intel 7400 series For Intel 7400 series CPUs, the recommendation is to use a clflush on the monitored address just before monitor and mwait pair [1]. This clflush makes sure that there are no false wakeups from mwait when the monitored address was recently written to. [1] "MONITOR/MWAIT Recommendations for Intel Xeon Processor 7400 series" section in specification update document of 7400 series http://download.intel.com/design/xeon/specupdt/32033601.pdf Signed-off-by: Venkatesh Pallipadi Signed-off-by: Ingo Molnar --- arch/x86/include/asm/cpufeature.h | 1 + arch/x86/kernel/cpu/intel.c | 3 +++ arch/x86/kernel/process.c | 6 ++++++ 3 files changed, 10 insertions(+) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/cpufeature.h b/arch/x86/include/asm/cpufeature.h index ea408dcba51..7301e60dc4a 100644 --- a/arch/x86/include/asm/cpufeature.h +++ b/arch/x86/include/asm/cpufeature.h @@ -93,6 +93,7 @@ #define X86_FEATURE_XTOPOLOGY (3*32+22) /* cpu topology enum extensions */ #define X86_FEATURE_TSC_RELIABLE (3*32+23) /* TSC is known to be reliable */ #define X86_FEATURE_NONSTOP_TSC (3*32+24) /* TSC does not stop in C states */ +#define X86_FEATURE_CLFLUSH_MONITOR (3*32+25) /* "" clflush reqd with monitor */ /* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */ #define X86_FEATURE_XMM3 (4*32+ 0) /* "pni" SSE-3 */ diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 430e5c38a54..24ff26a38ad 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -291,6 +291,9 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) ds_init_intel(c); } + if (c->x86 == 6 && c->x86_model == 29 && cpu_has_clflush) + set_cpu_cap(c, X86_FEATURE_CLFLUSH_MONITOR); + #ifdef CONFIG_X86_64 if (c->x86 == 15) c->x86_cache_alignment = c->x86_clflush_size * 2; diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index e68bb9e3086..6d12f7e37f8 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -180,6 +180,9 @@ void mwait_idle_with_hints(unsigned long ax, unsigned long cx) trace_power_start(&it, POWER_CSTATE, (ax>>4)+1); if (!need_resched()) { + if (cpu_has(¤t_cpu_data, X86_FEATURE_CLFLUSH_MONITOR)) + clflush((void *)¤t_thread_info()->flags); + __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) @@ -194,6 +197,9 @@ static void mwait_idle(void) struct power_trace it; if (!need_resched()) { trace_power_start(&it, POWER_CSTATE, 1); + if (cpu_has(¤t_cpu_data, X86_FEATURE_CLFLUSH_MONITOR)) + clflush((void *)¤t_thread_info()->flags); + __monitor((void *)¤t_thread_info()->flags, 0, 0); smp_mb(); if (!need_resched()) -- cgit v1.2.3-70-g09d2 From 3f4a739c6accd651a11fcf3c7a20ec8147c42660 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Sun, 8 Feb 2009 16:18:03 -0800 Subject: x86: find nr_irqs_gsi with mp_ioapic_routing Impact: find right nr_irqs_gsi on some systems. One test-system has gap between gsi's: [ 0.000000] ACPI: IOAPIC (id[0x04] address[0xfec00000] gsi_base[0]) [ 0.000000] IOAPIC[0]: apic_id 4, version 0, address 0xfec00000, GSI 0-23 [ 0.000000] ACPI: IOAPIC (id[0x05] address[0xfeafd000] gsi_base[48]) [ 0.000000] IOAPIC[1]: apic_id 5, version 0, address 0xfeafd000, GSI 48-54 [ 0.000000] ACPI: IOAPIC (id[0x06] address[0xfeafc000] gsi_base[56]) [ 0.000000] IOAPIC[2]: apic_id 6, version 0, address 0xfeafc000, GSI 56-62 ... [ 0.000000] nr_irqs_gsi: 38 So nr_irqs_gsi is not right. some irq for MSI will overwrite with io_apic. need to get that with acpi_probe_gsi when acpi io_apic is used Signed-off-by: Yinghai Lu Signed-off-by: Ingo Molnar --- arch/x86/include/asm/mpspec.h | 6 ++++++ arch/x86/kernel/acpi/boot.c | 23 +++++++++++++++++++++++ arch/x86/kernel/io_apic.c | 20 +++++++++++++++----- 3 files changed, 44 insertions(+), 5 deletions(-) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/mpspec.h b/arch/x86/include/asm/mpspec.h index 62d14ce3cd0..bd22f2a3713 100644 --- a/arch/x86/include/asm/mpspec.h +++ b/arch/x86/include/asm/mpspec.h @@ -60,6 +60,7 @@ 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); +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, u32 gsi, int triggering, int polarity); @@ -71,6 +72,11 @@ mp_config_acpi_gsi(unsigned char number, unsigned int devfn, u8 pin, return 0; } #endif +#else /* !CONFIG_ACPI: */ +static inline int acpi_probe_gsi(void) +{ + return 0; +} #endif /* CONFIG_ACPI */ #define PHYSID_ARRAY_SIZE BITS_TO_LONGS(MAX_APICS) diff --git a/arch/x86/kernel/acpi/boot.c b/arch/x86/kernel/acpi/boot.c index d37593c2f43..7678f10c456 100644 --- a/arch/x86/kernel/acpi/boot.c +++ b/arch/x86/kernel/acpi/boot.c @@ -973,6 +973,29 @@ void __init mp_register_ioapic(int id, u32 address, u32 gsi_base) nr_ioapics++; } +int __init acpi_probe_gsi(void) +{ + int idx; + int gsi; + int max_gsi = 0; + + if (acpi_disabled) + return 0; + + if (!acpi_ioapic) + return 0; + + max_gsi = 0; + for (idx = 0; idx < nr_ioapics; idx++) { + gsi = mp_ioapic_routing[idx].gsi_end; + + if (gsi > max_gsi) + max_gsi = gsi; + } + + return max_gsi + 1; +} + static void assign_to_mp_irq(struct mp_config_intsrc *m, struct mp_config_intsrc *mp_irq) { diff --git a/arch/x86/kernel/io_apic.c b/arch/x86/kernel/io_apic.c index 9b0c480c383..bc7ac4da90d 100644 --- a/arch/x86/kernel/io_apic.c +++ b/arch/x86/kernel/io_apic.c @@ -3841,14 +3841,24 @@ int __init io_apic_get_redir_entries (int ioapic) void __init probe_nr_irqs_gsi(void) { - int idx; int nr = 0; - for (idx = 0; idx < nr_ioapics; idx++) - nr += io_apic_get_redir_entries(idx) + 1; - - if (nr > nr_irqs_gsi) + nr = acpi_probe_gsi(); + if (nr > nr_irqs_gsi) { nr_irqs_gsi = nr; + } else { + /* for acpi=off or acpi is not compiled in */ + int idx; + + nr = 0; + for (idx = 0; idx < nr_ioapics; idx++) + nr += io_apic_get_redir_entries(idx) + 1; + + if (nr > nr_irqs_gsi) + nr_irqs_gsi = nr; + } + + printk(KERN_DEBUG "nr_irqs_gsi: %d\n", nr_irqs_gsi); } /* -------------------------------------------------------------------------- -- cgit v1.2.3-70-g09d2 From 914c3d630b29b07d04908eab1b246812dadd5bd6 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 9 Feb 2009 22:17:39 +0900 Subject: x86: include correct %gs in a.out core dump Impact: dump the correct %gs into a.out core dump aout_dump_thread() read %gs but didn't include it in core dump. Fix it. Signed-off-by: Tejun Heo Signed-off-by: Ingo Molnar --- arch/x86/include/asm/a.out-core.h | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/a.out-core.h b/arch/x86/include/asm/a.out-core.h index 37822206083..3c601f8224b 100644 --- a/arch/x86/include/asm/a.out-core.h +++ b/arch/x86/include/asm/a.out-core.h @@ -23,8 +23,6 @@ */ static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump) { - u16 gs; - /* changed the size calculations - should hopefully work better. lbt */ dump->magic = CMAGIC; dump->start_code = 0; @@ -57,7 +55,7 @@ static inline void aout_dump_thread(struct pt_regs *regs, struct user *dump) dump->regs.ds = (u16)regs->ds; dump->regs.es = (u16)regs->es; dump->regs.fs = (u16)regs->fs; - savesegment(gs, gs); + savesegment(gs, dump->regs.gs); dump->regs.orig_ax = regs->orig_ax; dump->regs.ip = regs->ip; dump->regs.cs = (u16)regs->cs; -- cgit v1.2.3-70-g09d2 From ae6af41f5a4841f06eb92bc86ad020ad44ae2a30 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 9 Feb 2009 22:17:39 +0900 Subject: x86: math_emu info cleanup Impact: cleanup * Come on, struct info? s/struct info/struct math_emu_info/ * Use struct pt_regs and kernel_vm86_regs instead of defining its own register frame structure. Signed-off-by: Tejun Heo Signed-off-by: Ingo Molnar --- arch/x86/include/asm/math_emu.h | 29 +++++------------- arch/x86/include/asm/processor.h | 2 +- arch/x86/math-emu/fpu_entry.c | 2 +- arch/x86/math-emu/fpu_proto.h | 2 +- arch/x86/math-emu/fpu_system.h | 14 ++++----- arch/x86/math-emu/get_address.c | 63 +++++++++++++++++++--------------------- 6 files changed, 48 insertions(+), 64 deletions(-) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/math_emu.h b/arch/x86/include/asm/math_emu.h index 5a65b107ad5..302492c7795 100644 --- a/arch/x86/include/asm/math_emu.h +++ b/arch/x86/include/asm/math_emu.h @@ -1,31 +1,18 @@ #ifndef _ASM_X86_MATH_EMU_H #define _ASM_X86_MATH_EMU_H +#include +#include + /* This structure matches the layout of the data saved to the stack following a device-not-present interrupt, part of it saved automatically by the 80386/80486. */ -struct info { +struct math_emu_info { long ___orig_eip; - long ___ebx; - long ___ecx; - long ___edx; - long ___esi; - long ___edi; - long ___ebp; - long ___eax; - long ___ds; - long ___es; - long ___fs; - long ___orig_eax; - long ___eip; - long ___cs; - long ___eflags; - long ___esp; - long ___ss; - long ___vm86_es; /* This and the following only in vm86 mode */ - long ___vm86_ds; - long ___vm86_fs; - long ___vm86_gs; + union { + struct pt_regs regs; + struct kernel_vm86_regs vm86; + }; }; #endif /* _ASM_X86_MATH_EMU_H */ diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index 091cd8855f2..3bfd5235a9e 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -353,7 +353,7 @@ struct i387_soft_struct { u8 no_update; u8 rm; u8 alimit; - struct info *info; + struct math_emu_info *info; u32 entry_eip; }; diff --git a/arch/x86/math-emu/fpu_entry.c b/arch/x86/math-emu/fpu_entry.c index c7b06feb139..c268abe7225 100644 --- a/arch/x86/math-emu/fpu_entry.c +++ b/arch/x86/math-emu/fpu_entry.c @@ -659,7 +659,7 @@ static int valid_prefix(u_char *Byte, u_char __user **fpu_eip, } } -void math_abort(struct info *info, unsigned int signal) +void math_abort(struct math_emu_info *info, unsigned int signal) { FPU_EIP = FPU_ORIG_EIP; current->thread.trap_no = 16; diff --git a/arch/x86/math-emu/fpu_proto.h b/arch/x86/math-emu/fpu_proto.h index aa49b6a0d85..51bfbb61c5b 100644 --- a/arch/x86/math-emu/fpu_proto.h +++ b/arch/x86/math-emu/fpu_proto.h @@ -52,7 +52,7 @@ extern void fst_i_(void); extern void fstp_i(void); /* fpu_entry.c */ asmlinkage extern void math_emulate(long arg); -extern void math_abort(struct info *info, unsigned int signal); +extern void math_abort(struct math_emu_info *info, unsigned int signal); /* fpu_etc.c */ extern void FPU_etc(void); /* fpu_tags.c */ diff --git a/arch/x86/math-emu/fpu_system.h b/arch/x86/math-emu/fpu_system.h index 13488fa153e..6729c6a3134 100644 --- a/arch/x86/math-emu/fpu_system.h +++ b/arch/x86/math-emu/fpu_system.h @@ -18,7 +18,7 @@ /* This sets the pointer FPU_info to point to the argument part of the stack frame of math_emulate() */ -#define SETUP_DATA_AREA(arg) FPU_info = (struct info *) &arg +#define SETUP_DATA_AREA(arg) FPU_info = (struct math_emu_info *) &arg /* s is always from a cpu register, and the cpu does bounds checking * during register load --> no further bounds checks needed */ @@ -38,12 +38,12 @@ #define I387 (current->thread.xstate) #define FPU_info (I387->soft.info) -#define FPU_CS (*(unsigned short *) &(FPU_info->___cs)) -#define FPU_SS (*(unsigned short *) &(FPU_info->___ss)) -#define FPU_DS (*(unsigned short *) &(FPU_info->___ds)) -#define FPU_EAX (FPU_info->___eax) -#define FPU_EFLAGS (FPU_info->___eflags) -#define FPU_EIP (FPU_info->___eip) +#define FPU_CS (*(unsigned short *) &(FPU_info->regs.cs)) +#define FPU_SS (*(unsigned short *) &(FPU_info->regs.ss)) +#define FPU_DS (*(unsigned short *) &(FPU_info->regs.ds)) +#define FPU_EAX (FPU_info->regs.ax) +#define FPU_EFLAGS (FPU_info->regs.flags) +#define FPU_EIP (FPU_info->regs.ip) #define FPU_ORIG_EIP (FPU_info->___orig_eip) #define FPU_lookahead (I387->soft.lookahead) diff --git a/arch/x86/math-emu/get_address.c b/arch/x86/math-emu/get_address.c index d701e2b39e4..62daa7fcc44 100644 --- a/arch/x86/math-emu/get_address.c +++ b/arch/x86/math-emu/get_address.c @@ -29,42 +29,39 @@ #define FPU_WRITE_BIT 0x10 static int reg_offset[] = { - offsetof(struct info, ___eax), - offsetof(struct info, ___ecx), - offsetof(struct info, ___edx), - offsetof(struct info, ___ebx), - offsetof(struct info, ___esp), - offsetof(struct info, ___ebp), - offsetof(struct info, ___esi), - offsetof(struct info, ___edi) + offsetof(struct math_emu_info, regs.ax), + offsetof(struct math_emu_info, regs.cx), + offsetof(struct math_emu_info, regs.dx), + offsetof(struct math_emu_info, regs.bx), + offsetof(struct math_emu_info, regs.sp), + offsetof(struct math_emu_info, regs.bp), + offsetof(struct math_emu_info, regs.si), + offsetof(struct math_emu_info, regs.di) }; #define REG_(x) (*(long *)(reg_offset[(x)]+(u_char *) FPU_info)) static int reg_offset_vm86[] = { - offsetof(struct info, ___cs), - offsetof(struct info, ___vm86_ds), - offsetof(struct info, ___vm86_es), - offsetof(struct info, ___vm86_fs), - offsetof(struct info, ___vm86_gs), - offsetof(struct info, ___ss), - offsetof(struct info, ___vm86_ds) + offsetof(struct math_emu_info, regs.cs), + offsetof(struct math_emu_info, vm86.ds), + offsetof(struct math_emu_info, vm86.es), + offsetof(struct math_emu_info, vm86.fs), + offsetof(struct math_emu_info, vm86.gs), + offsetof(struct math_emu_info, regs.ss), + offsetof(struct math_emu_info, vm86.ds) }; #define VM86_REG_(x) (*(unsigned short *) \ (reg_offset_vm86[((unsigned)x)]+(u_char *) FPU_info)) -/* This dummy, gs is not saved on the stack. */ -#define ___GS ___ds - static int reg_offset_pm[] = { - offsetof(struct info, ___cs), - offsetof(struct info, ___ds), - offsetof(struct info, ___es), - offsetof(struct info, ___fs), - offsetof(struct info, ___GS), - offsetof(struct info, ___ss), - offsetof(struct info, ___ds) + offsetof(struct math_emu_info, regs.cs), + offsetof(struct math_emu_info, regs.ds), + offsetof(struct math_emu_info, regs.es), + offsetof(struct math_emu_info, regs.fs), + offsetof(struct math_emu_info, regs.ds), /* dummy, not saved on stack */ + offsetof(struct math_emu_info, regs.ss), + offsetof(struct math_emu_info, regs.ds) }; #define PM_REG_(x) (*(unsigned short *) \ @@ -349,34 +346,34 @@ void __user *FPU_get_address_16(u_char FPU_modrm, unsigned long *fpu_eip, } switch (rm) { case 0: - address += FPU_info->___ebx + FPU_info->___esi; + address += FPU_info->regs.bx + FPU_info->regs.si; break; case 1: - address += FPU_info->___ebx + FPU_info->___edi; + address += FPU_info->regs.bx + FPU_info->regs.di; break; case 2: - address += FPU_info->___ebp + FPU_info->___esi; + address += FPU_info->regs.bp + FPU_info->regs.si; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 3: - address += FPU_info->___ebp + FPU_info->___edi; + address += FPU_info->regs.bp + FPU_info->regs.di; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 4: - address += FPU_info->___esi; + address += FPU_info->regs.si; break; case 5: - address += FPU_info->___edi; + address += FPU_info->regs.di; break; case 6: - address += FPU_info->___ebp; + address += FPU_info->regs.bp; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 7: - address += FPU_info->___ebx; + address += FPU_info->regs.bx; break; } -- cgit v1.2.3-70-g09d2 From d315760ffa261c15ff92699ac6f514112543d7ca Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 9 Feb 2009 22:17:39 +0900 Subject: x86: fix math_emu register frame access do_device_not_available() is the handler for #NM and it declares that it takes a unsigned long and calls math_emu(), which takes a long argument and surprisingly expects the stack frame starting at the zero argument would match struct math_emu_info, which isn't true regardless of configuration in the current code. This patch makes do_device_not_available() take struct pt_regs like other exception handlers and initialize struct math_emu_info with pointer to it and pass pointer to the math_emu_info to math_emulate() like normal C functions do. This way, unless gcc makes a copy of struct pt_regs in do_device_not_available(), the register frame is correctly accessed regardless of kernel configuration or compiler used. This doesn't fix all math_emu problems but it at least gets it somewhat working. Signed-off-by: Tejun Heo Signed-off-by: Ingo Molnar --- arch/x86/include/asm/math_emu.h | 4 +-- arch/x86/include/asm/traps.h | 4 +-- arch/x86/kernel/traps.c | 15 ++++++---- arch/x86/math-emu/fpu_entry.c | 4 +-- arch/x86/math-emu/fpu_proto.h | 2 +- arch/x86/math-emu/fpu_system.h | 16 ++++------ arch/x86/math-emu/get_address.c | 66 ++++++++++++++++++++--------------------- 7 files changed, 55 insertions(+), 56 deletions(-) (limited to 'arch/x86/include/asm') diff --git a/arch/x86/include/asm/math_emu.h b/arch/x86/include/asm/math_emu.h index 302492c7795..031f6266f42 100644 --- a/arch/x86/include/asm/math_emu.h +++ b/arch/x86/include/asm/math_emu.h @@ -11,8 +11,8 @@ struct math_emu_info { long ___orig_eip; union { - struct pt_regs regs; - struct kernel_vm86_regs vm86; + struct pt_regs *regs; + struct kernel_vm86_regs *vm86; }; }; #endif /* _ASM_X86_MATH_EMU_H */ diff --git a/arch/x86/include/asm/traps.h b/arch/x86/include/asm/traps.h index 2ee0a3bceed..cf3bb053da0 100644 --- a/arch/x86/include/asm/traps.h +++ b/arch/x86/include/asm/traps.h @@ -41,7 +41,7 @@ dotraplinkage void do_int3(struct pt_regs *, long); dotraplinkage void do_overflow(struct pt_regs *, long); dotraplinkage void do_bounds(struct pt_regs *, long); dotraplinkage void do_invalid_op(struct pt_regs *, long); -dotraplinkage void do_device_not_available(struct pt_regs *, long); +dotraplinkage void do_device_not_available(struct pt_regs); dotraplinkage void do_coprocessor_segment_overrun(struct pt_regs *, long); dotraplinkage void do_invalid_TSS(struct pt_regs *, long); dotraplinkage void do_segment_not_present(struct pt_regs *, long); @@ -77,7 +77,7 @@ extern int panic_on_unrecovered_nmi; extern int kstack_depth_to_print; void math_error(void __user *); -asmlinkage void math_emulate(long); +void math_emulate(struct math_emu_info *); #ifdef CONFIG_X86_32 unsigned long patch_espfix_desc(unsigned long, unsigned long); #else diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 98c2d055284..7932338d7cb 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -896,7 +896,7 @@ asmlinkage void math_state_restore(void) EXPORT_SYMBOL_GPL(math_state_restore); #ifndef CONFIG_MATH_EMULATION -asmlinkage void math_emulate(long arg) +void math_emulate(struct math_emu_info *info) { printk(KERN_EMERG "math-emulation not enabled and no coprocessor found.\n"); @@ -906,16 +906,19 @@ asmlinkage void math_emulate(long arg) } #endif /* CONFIG_MATH_EMULATION */ -dotraplinkage void __kprobes -do_device_not_available(struct pt_regs *regs, long error) +dotraplinkage void __kprobes do_device_not_available(struct pt_regs regs) { #ifdef CONFIG_X86_32 if (read_cr0() & X86_CR0_EM) { - conditional_sti(regs); - math_emulate(0); + struct math_emu_info info = { }; + + conditional_sti(®s); + + info.regs = ®s; + math_emulate(&info); } else { math_state_restore(); /* interrupts still off */ - conditional_sti(regs); + conditional_sti(®s); } #else math_state_restore(); diff --git a/arch/x86/math-emu/fpu_entry.c b/arch/x86/math-emu/fpu_entry.c index c268abe7225..5d87f586f8d 100644 --- a/arch/x86/math-emu/fpu_entry.c +++ b/arch/x86/math-emu/fpu_entry.c @@ -131,7 +131,7 @@ u_char emulating = 0; static int valid_prefix(u_char *Byte, u_char __user ** fpu_eip, overrides * override); -asmlinkage void math_emulate(long arg) +void math_emulate(struct math_emu_info *info) { u_char FPU_modrm, byte1; unsigned short code; @@ -161,7 +161,7 @@ asmlinkage void math_emulate(long arg) RE_ENTRANT_CHECK_ON; #endif /* RE_ENTRANT_CHECKING */ - SETUP_DATA_AREA(arg); + FPU_info = info; FPU_ORIG_EIP = FPU_EIP; diff --git a/arch/x86/math-emu/fpu_proto.h b/arch/x86/math-emu/fpu_proto.h index 51bfbb61c5b..9779df436b7 100644 --- a/arch/x86/math-emu/fpu_proto.h +++ b/arch/x86/math-emu/fpu_proto.h @@ -51,7 +51,7 @@ extern void ffreep(void); extern void fst_i_(void); extern void fstp_i(void); /* fpu_entry.c */ -asmlinkage extern void math_emulate(long arg); +extern void math_emulate(struct math_emu_info *info); extern void math_abort(struct math_emu_info *info, unsigned int signal); /* fpu_etc.c */ extern void FPU_etc(void); diff --git a/arch/x86/math-emu/fpu_system.h b/arch/x86/math-emu/fpu_system.h index 6729c6a3134..50fa0ec2c8a 100644 --- a/arch/x86/math-emu/fpu_system.h +++ b/arch/x86/math-emu/fpu_system.h @@ -16,10 +16,6 @@ #include #include -/* This sets the pointer FPU_info to point to the argument part - of the stack frame of math_emulate() */ -#define SETUP_DATA_AREA(arg) FPU_info = (struct math_emu_info *) &arg - /* s is always from a cpu register, and the cpu does bounds checking * during register load --> no further bounds checks needed */ #define LDT_DESCRIPTOR(s) (((struct desc_struct *)current->mm->context.ldt)[(s) >> 3]) @@ -38,12 +34,12 @@ #define I387 (current->thread.xstate) #define FPU_info (I387->soft.info) -#define FPU_CS (*(unsigned short *) &(FPU_info->regs.cs)) -#define FPU_SS (*(unsigned short *) &(FPU_info->regs.ss)) -#define FPU_DS (*(unsigned short *) &(FPU_info->regs.ds)) -#define FPU_EAX (FPU_info->regs.ax) -#define FPU_EFLAGS (FPU_info->regs.flags) -#define FPU_EIP (FPU_info->regs.ip) +#define FPU_CS (*(unsigned short *) &(FPU_info->regs->cs)) +#define FPU_SS (*(unsigned short *) &(FPU_info->regs->ss)) +#define FPU_DS (*(unsigned short *) &(FPU_info->regs->ds)) +#define FPU_EAX (FPU_info->regs->ax) +#define FPU_EFLAGS (FPU_info->regs->flags) +#define FPU_EIP (FPU_info->regs->ip) #define FPU_ORIG_EIP (FPU_info->___orig_eip) #define FPU_lookahead (I387->soft.lookahead) diff --git a/arch/x86/math-emu/get_address.c b/arch/x86/math-emu/get_address.c index 62daa7fcc44..420b3b6e391 100644 --- a/arch/x86/math-emu/get_address.c +++ b/arch/x86/math-emu/get_address.c @@ -29,43 +29,43 @@ #define FPU_WRITE_BIT 0x10 static int reg_offset[] = { - offsetof(struct math_emu_info, regs.ax), - offsetof(struct math_emu_info, regs.cx), - offsetof(struct math_emu_info, regs.dx), - offsetof(struct math_emu_info, regs.bx), - offsetof(struct math_emu_info, regs.sp), - offsetof(struct math_emu_info, regs.bp), - offsetof(struct math_emu_info, regs.si), - offsetof(struct math_emu_info, regs.di) + offsetof(struct pt_regs, ax), + offsetof(struct pt_regs, cx), + offsetof(struct pt_regs, dx), + offsetof(struct pt_regs, bx), + offsetof(struct pt_regs, sp), + offsetof(struct pt_regs, bp), + offsetof(struct pt_regs, si), + offsetof(struct pt_regs, di) }; -#define REG_(x) (*(long *)(reg_offset[(x)]+(u_char *) FPU_info)) +#define REG_(x) (*(long *)(reg_offset[(x)] + (u_char *)FPU_info->regs)) static int reg_offset_vm86[] = { - offsetof(struct math_emu_info, regs.cs), - offsetof(struct math_emu_info, vm86.ds), - offsetof(struct math_emu_info, vm86.es), - offsetof(struct math_emu_info, vm86.fs), - offsetof(struct math_emu_info, vm86.gs), - offsetof(struct math_emu_info, regs.ss), - offsetof(struct math_emu_info, vm86.ds) + offsetof(struct pt_regs, cs), + offsetof(struct kernel_vm86_regs, ds), + offsetof(struct kernel_vm86_regs, es), + offsetof(struct kernel_vm86_regs, fs), + offsetof(struct kernel_vm86_regs, gs), + offsetof(struct pt_regs, ss), + offsetof(struct kernel_vm86_regs, ds) }; #define VM86_REG_(x) (*(unsigned short *) \ - (reg_offset_vm86[((unsigned)x)]+(u_char *) FPU_info)) + (reg_offset_vm86[((unsigned)x)] + (u_char *)FPU_info->regs)) static int reg_offset_pm[] = { - offsetof(struct math_emu_info, regs.cs), - offsetof(struct math_emu_info, regs.ds), - offsetof(struct math_emu_info, regs.es), - offsetof(struct math_emu_info, regs.fs), - offsetof(struct math_emu_info, regs.ds), /* dummy, not saved on stack */ - offsetof(struct math_emu_info, regs.ss), - offsetof(struct math_emu_info, regs.ds) + offsetof(struct pt_regs, cs), + offsetof(struct pt_regs, ds), + offsetof(struct pt_regs, es), + offsetof(struct pt_regs, fs), + offsetof(struct pt_regs, ds), /* dummy, not saved on stack */ + offsetof(struct pt_regs, ss), + offsetof(struct pt_regs, ds) }; #define PM_REG_(x) (*(unsigned short *) \ - (reg_offset_pm[((unsigned)x)]+(u_char *) FPU_info)) + (reg_offset_pm[((unsigned)x)] + (u_char *)FPU_info->regs)) /* Decode the SIB byte. This function assumes mod != 0 */ static int sib(int mod, unsigned long *fpu_eip) @@ -346,34 +346,34 @@ void __user *FPU_get_address_16(u_char FPU_modrm, unsigned long *fpu_eip, } switch (rm) { case 0: - address += FPU_info->regs.bx + FPU_info->regs.si; + address += FPU_info->regs->bx + FPU_info->regs->si; break; case 1: - address += FPU_info->regs.bx + FPU_info->regs.di; + address += FPU_info->regs->bx + FPU_info->regs->di; break; case 2: - address += FPU_info->regs.bp + FPU_info->regs.si; + address += FPU_info->regs->bp + FPU_info->regs->si; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 3: - address += FPU_info->regs.bp + FPU_info->regs.di; + address += FPU_info->regs->bp + FPU_info->regs->di; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 4: - address += FPU_info->regs.si; + address += FPU_info->regs->si; break; case 5: - address += FPU_info->regs.di; + address += FPU_info->regs->di; break; case 6: - address += FPU_info->regs.bp; + address += FPU_info->regs->bp; if (addr_modes.override.segment == PREFIX_DEFAULT) addr_modes.override.segment = PREFIX_SS_; break; case 7: - address += FPU_info->regs.bx; + address += FPU_info->regs->bx; break; } -- cgit v1.2.3-70-g09d2