From 60e019eb37a8d989031ad47ae9810453536f3127 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Mon, 29 Apr 2013 16:04:20 +0200 Subject: x86: Get rid of ->hard_math and all the FPU asm fu Reimplement FPU detection code in C and drop old, not-so-recommended detection method in asm. Move all the relevant stuff into i387.c where it conceptually belongs. Finally drop cpuinfo_x86.hard_math. [ hpa: huge thanks to Borislav for taking my original concept patch and productizing it ] [ Boris, note to self: do not use static_cpu_has before alternatives! ] Signed-off-by: H. Peter Anvin Link: http://lkml.kernel.org/r/1367244262-29511-2-git-send-email-bp@alien8.de Link: http://lkml.kernel.org/r/1365436666-9837-2-git-send-email-bp@alien8.de Signed-off-by: Borislav Petkov Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/common.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 22018f70a671..d4dd99350e9d 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -711,10 +711,9 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c) return; cpu_detect(c); - get_cpu_vendor(c); - get_cpu_cap(c); + fpu_detect(c); if (this_cpu->c_early_init) this_cpu->c_early_init(c); -- cgit v1.2.3 From c3b83598c1eeb1507603b461f5843ec2a49e3033 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Sun, 9 Jun 2013 12:07:30 +0200 Subject: x86, cpu: Add a synthetic, always true, cpu feature This will be used in alternatives later as an always-replace flag. Signed-off-by: Borislav Petkov Link: http://lkml.kernel.org/r/1370772454-6106-2-git-send-email-bp@alien8.de Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/common.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index d4dd99350e9d..d388ce1f3e1b 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -723,6 +723,8 @@ static void __init early_identify_cpu(struct cpuinfo_x86 *c) if (this_cpu->c_bsp_init) this_cpu->c_bsp_init(c); + + setup_force_cpu_cap(X86_FEATURE_ALWAYS); } void __init early_cpu_init(void) -- cgit v1.2.3 From 5700f743b597951743da9c7d891d3989aac0486e Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Sun, 9 Jun 2013 12:07:32 +0200 Subject: x86: Sanity-check static_cpu_has usage static_cpu_has may be used only after alternatives have run. Before that it always returns false if constant folding with __builtin_constant_p() doesn't happen. And you don't want that. This patch is the result of me debugging an issue where I overzealously put static_cpu_has in code which executed before alternatives have run and had to spend some time with scratching head and cursing at the monitor. So add a jump to a warning which screams loudly when we use this function too early. The alternatives patch that check away in conjunction with patching the rest of the kernel image. [ hpa: factored this into its own configuration option. If we want to have an overarching option, it should be an option which selects other options, not as a group option in the source code. ] Signed-off-by: Borislav Petkov Link: http://lkml.kernel.org/r/1370772454-6106-4-git-send-email-bp@alien8.de Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/common.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index d388ce1f3e1b..59adaa17e220 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1364,3 +1364,11 @@ void __cpuinit cpu_init(void) fpu_init(); } #endif + +#ifdef CONFIG_X86_DEBUG_STATIC_CPU_HAS +void warn_pre_alternatives(void) +{ + WARN(1, "You're using static_cpu_has before alternatives have run!\n"); +} +EXPORT_SYMBOL_GPL(warn_pre_alternatives); +#endif -- cgit v1.2.3 From 4a90a99c4f8002edaa6be11bd756872ebf3f3d97 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Sun, 9 Jun 2013 12:07:33 +0200 Subject: x86: Add a static_cpu_has_safe variant We want to use this in early code where alternatives might not have run yet and for that case we fall back to the dynamic boot_cpu_has. For that, force a 5-byte jump since the compiler could be generating differently sized jumps for each label. Signed-off-by: Borislav Petkov Link: http://lkml.kernel.org/r/1370772454-6106-5-git-send-email-bp@alien8.de Signed-off-by: H. Peter Anvin --- arch/x86/kernel/cpu/common.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 59adaa17e220..a4a07c0acb1f 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1372,3 +1372,9 @@ void warn_pre_alternatives(void) } EXPORT_SYMBOL_GPL(warn_pre_alternatives); #endif + +inline bool __static_cpu_has_safe(u16 bit) +{ + return boot_cpu_has(bit); +} +EXPORT_SYMBOL_GPL(__static_cpu_has_safe); -- cgit v1.2.3 From 629f4f9d59a27d8e58aa612e886e6a9a63ea7aeb Mon Sep 17 00:00:00 2001 From: Seiji Aguchi Date: Thu, 20 Jun 2013 11:45:44 -0400 Subject: x86: Rename variables for debugging Rename variables for debugging to describe meaning of them precisely. Also, introduce a generic way to switch IDT by checking a current state, debug on/off. Signed-off-by: Seiji Aguchi Link: http://lkml.kernel.org/r/51C323A8.7050905@hds.com Signed-off-by: H. Peter Anvin Cc: Steven Rostedt --- arch/x86/kernel/cpu/common.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 22018f70a671..8f6a0f909d6f 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1071,8 +1071,8 @@ __setup("clearcpuid=", setup_disablecpuid); #ifdef CONFIG_X86_64 struct desc_ptr idt_descr = { NR_VECTORS * 16 - 1, (unsigned long) idt_table }; -struct desc_ptr nmi_idt_descr = { NR_VECTORS * 16 - 1, - (unsigned long) nmi_idt_table }; +struct desc_ptr debug_idt_descr = { NR_VECTORS * 16 - 1, + (unsigned long) debug_idt_table }; DEFINE_PER_CPU_FIRST(union irq_stack_union, irq_stack_union) __aligned(PAGE_SIZE); @@ -1148,20 +1148,20 @@ int is_debug_stack(unsigned long addr) addr > (__get_cpu_var(debug_stack_addr) - DEBUG_STKSZ)); } -static DEFINE_PER_CPU(u32, debug_stack_use_ctr); +DEFINE_PER_CPU(u32, debug_idt_ctr); void debug_stack_set_zero(void) { - this_cpu_inc(debug_stack_use_ctr); - load_idt((const struct desc_ptr *)&nmi_idt_descr); + this_cpu_inc(debug_idt_ctr); + load_current_idt(); } void debug_stack_reset(void) { - if (WARN_ON(!this_cpu_read(debug_stack_use_ctr))) + if (WARN_ON(!this_cpu_read(debug_idt_ctr))) return; - if (this_cpu_dec_return(debug_stack_use_ctr) == 0) - load_idt((const struct desc_ptr *)&idt_descr); + if (this_cpu_dec_return(debug_idt_ctr) == 0) + load_current_idt(); } #else /* CONFIG_X86_64 */ -- cgit v1.2.3 From cf910e83ae23692fdeefc7e506e504c4c468d38a Mon Sep 17 00:00:00 2001 From: Seiji Aguchi Date: Thu, 20 Jun 2013 11:46:53 -0400 Subject: x86, trace: Add irq vector tracepoints [Purpose of this patch] As Vaibhav explained in the thread below, tracepoints for irq vectors are useful. http://www.spinics.net/lists/mm-commits/msg85707.html The current interrupt traces from irq_handler_entry and irq_handler_exit provide when an interrupt is handled. They provide good data about when the system has switched to kernel space and how it affects the currently running processes. There are some IRQ vectors which trigger the system into kernel space, which are not handled in generic IRQ handlers. Tracing such events gives us the information about IRQ interaction with other system events. The trace also tells where the system is spending its time. We want to know which cores are handling interrupts and how they are affecting other processes in the system. Also, the trace provides information about when the cores are idle and which interrupts are changing that state. On the other hand, my usecase is tracing just local timer event and getting a value of instruction pointer. I suggested to add an argument local timer event to get instruction pointer before. But there is another way to get it with external module like systemtap. So, I don't need to add any argument to irq vector tracepoints now. [Patch Description] Vaibhav's patch shared a trace point ,irq_vector_entry/irq_vector_exit, in all events. But there is an above use case to trace specific irq_vector rather than tracing all events. In this case, we are concerned about overhead due to unwanted events. So, add following tracepoints instead of introducing irq_vector_entry/exit. so that we can enable them independently. - local_timer_vector - reschedule_vector - call_function_vector - call_function_single_vector - irq_work_entry_vector - error_apic_vector - thermal_apic_vector - threshold_apic_vector - spurious_apic_vector - x86_platform_ipi_vector Also, introduce a logic switching IDT at enabling/disabling time so that a time penalty makes a zero when tracepoints are disabled. Detailed explanations are as follows. - Create trace irq handlers with entering_irq()/exiting_irq(). - Create a new IDT, trace_idt_table, at boot time by adding a logic to _set_gate(). It is just a copy of original idt table. - Register the new handlers for tracpoints to the new IDT by introducing macros to alloc_intr_gate() called at registering time of irq_vector handlers. - Add checking, whether irq vector tracing is on/off, into load_current_idt(). This has to be done below debug checking for these reasons. - Switching to debug IDT may be kicked while tracing is enabled. - On the other hands, switching to trace IDT is kicked only when debugging is disabled. In addition, the new IDT is created only when CONFIG_TRACING is enabled to avoid being used for other purposes. Signed-off-by: Seiji Aguchi Link: http://lkml.kernel.org/r/51C323ED.5050708@hds.com Signed-off-by: H. Peter Anvin Cc: Steven Rostedt --- arch/x86/kernel/cpu/common.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 8f6a0f909d6f..d9c8c0947ec2 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -1257,7 +1257,7 @@ void __cpuinit cpu_init(void) switch_to_new_gdt(cpu); loadsegment(fs, 0); - load_idt((const struct desc_ptr *)&idt_descr); + load_current_idt(); memset(me->thread.tls_array, 0, GDT_ENTRY_TLS_ENTRIES * 8); syscall_init(); @@ -1334,7 +1334,7 @@ void __cpuinit cpu_init(void) if (cpu_has_vme || cpu_has_tsc || cpu_has_de) clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); - load_idt(&idt_descr); + load_current_idt(); switch_to_new_gdt(cpu); /* -- cgit v1.2.3 From 148f9bb87745ed45f7a11b2cbd3bc0f017d5d257 Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Tue, 18 Jun 2013 18:23:59 -0400 Subject: x86: delete __cpuinit usage from all x86 files The __cpuinit type of throwaway sections might have made sense some time ago when RAM was more constrained, but now the savings do not offset the cost and complications. For example, the fix in commit 5e427ec2d0 ("x86: Fix bit corruption at CPU resume time") is a good example of the nasty type of bugs that can be created with improper use of the various __init prefixes. After a discussion on LKML[1] it was decided that cpuinit should go the way of devinit and be phased out. Once all the users are gone, we can then finally remove the macros themselves from linux/init.h. Note that some harmless section mismatch warnings may result, since notify_cpu_starting() and cpu_up() are arch independent (kernel/cpu.c) are flagged as __cpuinit -- so if we remove the __cpuinit from arch specific callers, we will also get section mismatch warnings. As an intermediate step, we intend to turn the linux/init.h cpuinit content into no-ops as early as possible, since that will get rid of these warnings. In any case, they are temporary and harmless. This removes all the arch/x86 uses of the __cpuinit macros from all C files. x86 only had the one __CPUINIT used in assembly files, and it wasn't paired off with a .previous or a __FINIT, so we can delete it directly w/o any corresponding additional change there. [1] https://lkml.org/lkml/2013/5/20/589 Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: x86@kernel.org Acked-by: Ingo Molnar Acked-by: Thomas Gleixner Acked-by: H. Peter Anvin Signed-off-by: Paul Gortmaker --- arch/x86/kernel/cpu/common.c | 64 ++++++++++++++++++++++---------------------- 1 file changed, 32 insertions(+), 32 deletions(-) (limited to 'arch/x86/kernel/cpu/common.c') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 548bd039784e..25eb2747b063 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -63,7 +63,7 @@ void __init setup_cpu_local_masks(void) alloc_bootmem_cpumask_var(&cpu_sibling_setup_mask); } -static void __cpuinit default_init(struct cpuinfo_x86 *c) +static void default_init(struct cpuinfo_x86 *c) { #ifdef CONFIG_X86_64 cpu_detect_cache_sizes(c); @@ -80,13 +80,13 @@ static void __cpuinit default_init(struct cpuinfo_x86 *c) #endif } -static const struct cpu_dev __cpuinitconst default_cpu = { +static const struct cpu_dev default_cpu = { .c_init = default_init, .c_vendor = "Unknown", .c_x86_vendor = X86_VENDOR_UNKNOWN, }; -static const struct cpu_dev *this_cpu __cpuinitdata = &default_cpu; +static const struct cpu_dev *this_cpu = &default_cpu; DEFINE_PER_CPU_PAGE_ALIGNED(struct gdt_page, gdt_page) = { .gdt = { #ifdef CONFIG_X86_64 @@ -160,8 +160,8 @@ static int __init x86_xsaveopt_setup(char *s) __setup("noxsaveopt", x86_xsaveopt_setup); #ifdef CONFIG_X86_32 -static int cachesize_override __cpuinitdata = -1; -static int disable_x86_serial_nr __cpuinitdata = 1; +static int cachesize_override = -1; +static int disable_x86_serial_nr = 1; static int __init cachesize_setup(char *str) { @@ -215,12 +215,12 @@ static inline int flag_is_changeable_p(u32 flag) } /* Probe for the CPUID instruction */ -int __cpuinit have_cpuid_p(void) +int have_cpuid_p(void) { return flag_is_changeable_p(X86_EFLAGS_ID); } -static void __cpuinit squash_the_stupid_serial_number(struct cpuinfo_x86 *c) +static void squash_the_stupid_serial_number(struct cpuinfo_x86 *c) { unsigned long lo, hi; @@ -298,7 +298,7 @@ struct cpuid_dependent_feature { u32 level; }; -static const struct cpuid_dependent_feature __cpuinitconst +static const struct cpuid_dependent_feature cpuid_dependent_features[] = { { X86_FEATURE_MWAIT, 0x00000005 }, { X86_FEATURE_DCA, 0x00000009 }, @@ -306,7 +306,7 @@ cpuid_dependent_features[] = { { 0, 0 } }; -static void __cpuinit filter_cpuid_features(struct cpuinfo_x86 *c, bool warn) +static void filter_cpuid_features(struct cpuinfo_x86 *c, bool warn) { const struct cpuid_dependent_feature *df; @@ -344,7 +344,7 @@ static void __cpuinit filter_cpuid_features(struct cpuinfo_x86 *c, bool warn) */ /* Look up CPU names by table lookup. */ -static const char *__cpuinit table_lookup_model(struct cpuinfo_x86 *c) +static const char *table_lookup_model(struct cpuinfo_x86 *c) { const struct cpu_model_info *info; @@ -364,8 +364,8 @@ static const char *__cpuinit table_lookup_model(struct cpuinfo_x86 *c) return NULL; /* Not found */ } -__u32 cpu_caps_cleared[NCAPINTS] __cpuinitdata; -__u32 cpu_caps_set[NCAPINTS] __cpuinitdata; +__u32 cpu_caps_cleared[NCAPINTS]; +__u32 cpu_caps_set[NCAPINTS]; void load_percpu_segment(int cpu) { @@ -394,9 +394,9 @@ void switch_to_new_gdt(int cpu) load_percpu_segment(cpu); } -static const struct cpu_dev *__cpuinitdata cpu_devs[X86_VENDOR_NUM] = {}; +static const struct cpu_dev *cpu_devs[X86_VENDOR_NUM] = {}; -static void __cpuinit get_model_name(struct cpuinfo_x86 *c) +static void get_model_name(struct cpuinfo_x86 *c) { unsigned int *v; char *p, *q; @@ -425,7 +425,7 @@ static void __cpuinit get_model_name(struct cpuinfo_x86 *c) } } -void __cpuinit cpu_detect_cache_sizes(struct cpuinfo_x86 *c) +void cpu_detect_cache_sizes(struct cpuinfo_x86 *c) { unsigned int n, dummy, ebx, ecx, edx, l2size; @@ -479,7 +479,7 @@ u16 __read_mostly tlb_lld_4m[NR_INFO]; */ s8 __read_mostly tlb_flushall_shift = -1; -void __cpuinit cpu_detect_tlb(struct cpuinfo_x86 *c) +void cpu_detect_tlb(struct cpuinfo_x86 *c) { if (this_cpu->c_detect_tlb) this_cpu->c_detect_tlb(c); @@ -493,7 +493,7 @@ void __cpuinit cpu_detect_tlb(struct cpuinfo_x86 *c) tlb_flushall_shift); } -void __cpuinit detect_ht(struct cpuinfo_x86 *c) +void detect_ht(struct cpuinfo_x86 *c) { #ifdef CONFIG_X86_HT u32 eax, ebx, ecx, edx; @@ -544,7 +544,7 @@ out: #endif } -static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) +static void get_cpu_vendor(struct cpuinfo_x86 *c) { char *v = c->x86_vendor_id; int i; @@ -571,7 +571,7 @@ static void __cpuinit get_cpu_vendor(struct cpuinfo_x86 *c) this_cpu = &default_cpu; } -void __cpuinit cpu_detect(struct cpuinfo_x86 *c) +void cpu_detect(struct cpuinfo_x86 *c) { /* Get vendor name */ cpuid(0x00000000, (unsigned int *)&c->cpuid_level, @@ -601,7 +601,7 @@ void __cpuinit cpu_detect(struct cpuinfo_x86 *c) } } -void __cpuinit get_cpu_cap(struct cpuinfo_x86 *c) +void get_cpu_cap(struct cpuinfo_x86 *c) { u32 tfms, xlvl; u32 ebx; @@ -652,7 +652,7 @@ void __cpuinit get_cpu_cap(struct cpuinfo_x86 *c) init_scattered_cpuid_features(c); } -static void __cpuinit identify_cpu_without_cpuid(struct cpuinfo_x86 *c) +static void identify_cpu_without_cpuid(struct cpuinfo_x86 *c) { #ifdef CONFIG_X86_32 int i; @@ -769,7 +769,7 @@ void __init early_cpu_init(void) * unless we can find a reliable way to detect all the broken cases. * Enable it explicitly on 64-bit for non-constant inputs of cpu_has(). */ -static void __cpuinit detect_nopl(struct cpuinfo_x86 *c) +static void detect_nopl(struct cpuinfo_x86 *c) { #ifdef CONFIG_X86_32 clear_cpu_cap(c, X86_FEATURE_NOPL); @@ -778,7 +778,7 @@ static void __cpuinit detect_nopl(struct cpuinfo_x86 *c) #endif } -static void __cpuinit generic_identify(struct cpuinfo_x86 *c) +static void generic_identify(struct cpuinfo_x86 *c) { c->extended_cpuid_level = 0; @@ -815,7 +815,7 @@ static void __cpuinit generic_identify(struct cpuinfo_x86 *c) /* * This does the hard work of actually picking apart the CPU stuff... */ -static void __cpuinit identify_cpu(struct cpuinfo_x86 *c) +static void identify_cpu(struct cpuinfo_x86 *c) { int i; @@ -960,7 +960,7 @@ void __init identify_boot_cpu(void) cpu_detect_tlb(&boot_cpu_data); } -void __cpuinit identify_secondary_cpu(struct cpuinfo_x86 *c) +void identify_secondary_cpu(struct cpuinfo_x86 *c) { BUG_ON(c == &boot_cpu_data); identify_cpu(c); @@ -975,14 +975,14 @@ struct msr_range { unsigned max; }; -static const struct msr_range msr_range_array[] __cpuinitconst = { +static const struct msr_range msr_range_array[] = { { 0x00000000, 0x00000418}, { 0xc0000000, 0xc000040b}, { 0xc0010000, 0xc0010142}, { 0xc0011000, 0xc001103b}, }; -static void __cpuinit __print_cpu_msr(void) +static void __print_cpu_msr(void) { unsigned index_min, index_max; unsigned index; @@ -1001,7 +1001,7 @@ static void __cpuinit __print_cpu_msr(void) } } -static int show_msr __cpuinitdata; +static int show_msr; static __init int setup_show_msr(char *arg) { @@ -1022,7 +1022,7 @@ static __init int setup_noclflush(char *arg) } __setup("noclflush", setup_noclflush); -void __cpuinit print_cpu_info(struct cpuinfo_x86 *c) +void print_cpu_info(struct cpuinfo_x86 *c) { const char *vendor = NULL; @@ -1051,7 +1051,7 @@ void __cpuinit print_cpu_info(struct cpuinfo_x86 *c) print_cpu_msr(c); } -void __cpuinit print_cpu_msr(struct cpuinfo_x86 *c) +void print_cpu_msr(struct cpuinfo_x86 *c) { if (c->cpu_index < show_msr) __print_cpu_msr(); @@ -1216,7 +1216,7 @@ static void dbg_restore_debug_regs(void) */ #ifdef CONFIG_X86_64 -void __cpuinit cpu_init(void) +void cpu_init(void) { struct orig_ist *oist; struct task_struct *me; @@ -1315,7 +1315,7 @@ void __cpuinit cpu_init(void) #else -void __cpuinit cpu_init(void) +void cpu_init(void) { int cpu = smp_processor_id(); struct task_struct *curr = current; -- cgit v1.2.3