From 375074cc736ab1d89a708c0a8d7baa4a70d5d476 Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Fri, 24 Oct 2014 15:58:07 -0700 Subject: x86: Clean up cr4 manipulation CR4 manipulation was split, seemingly at random, between direct (write_cr4) and using a helper (set/clear_in_cr4). Unfortunately, the set_in_cr4 and clear_in_cr4 helpers also poke at the boot code, which only a small subset of users actually wanted. This patch replaces all cr4 access in functions that don't leave cr4 exactly the way they found it with new helpers cr4_set_bits, cr4_clear_bits, and cr4_set_bits_and_update_boot. Signed-off-by: Andy Lutomirski Reviewed-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Cc: Andrea Arcangeli Cc: Vince Weaver Cc: "hillf.zj" Cc: Valdis Kletnieks Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Kees Cook Cc: Linus Torvalds Link: http://lkml.kernel.org/r/495a10bdc9e67016b8fd3945700d46cfd5c12c2f.1414190806.git.luto@amacapital.net Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/common.c | 10 +++++----- arch/x86/kernel/cpu/mcheck/mce.c | 3 ++- arch/x86/kernel/cpu/mcheck/p5.c | 3 ++- arch/x86/kernel/cpu/mcheck/winchip.c | 3 ++- arch/x86/kernel/cpu/perf_event.c | 7 ++++--- arch/x86/kernel/i387.c | 3 ++- arch/x86/kernel/process.c | 5 +++-- arch/x86/kernel/xsave.c | 3 ++- 8 files changed, 22 insertions(+), 15 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index c6049650c093..9d8fc49f0922 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -278,7 +278,7 @@ __setup("nosmep", setup_disable_smep); static __always_inline void setup_smep(struct cpuinfo_x86 *c) { if (cpu_has(c, X86_FEATURE_SMEP)) - set_in_cr4(X86_CR4_SMEP); + cr4_set_bits(X86_CR4_SMEP); } static __init int setup_disable_smap(char *arg) @@ -298,9 +298,9 @@ static __always_inline void setup_smap(struct cpuinfo_x86 *c) if (cpu_has(c, X86_FEATURE_SMAP)) { #ifdef CONFIG_X86_SMAP - set_in_cr4(X86_CR4_SMAP); + cr4_set_bits(X86_CR4_SMAP); #else - clear_in_cr4(X86_CR4_SMAP); + cr4_clear_bits(X86_CR4_SMAP); #endif } } @@ -1312,7 +1312,7 @@ void cpu_init(void) pr_debug("Initializing CPU#%d\n", cpu); - clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); + cr4_clear_bits(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); /* * Initialize the per-CPU GDT with the boot GDT, @@ -1393,7 +1393,7 @@ void cpu_init(void) printk(KERN_INFO "Initializing CPU#%d\n", cpu); if (cpu_feature_enabled(X86_FEATURE_VME) || cpu_has_tsc || cpu_has_de) - clear_in_cr4(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); + cr4_clear_bits(X86_CR4_VME|X86_CR4_PVI|X86_CR4_TSD|X86_CR4_DE); load_current_idt(); switch_to_new_gdt(cpu); diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c index d23179900755..15ad3ed1a3cd 100644 --- a/arch/x86/kernel/cpu/mcheck/mce.c +++ b/arch/x86/kernel/cpu/mcheck/mce.c @@ -44,6 +44,7 @@ #include #include +#include #include #include @@ -1449,7 +1450,7 @@ static void __mcheck_cpu_init_generic(void) bitmap_fill(all_banks, MAX_NR_BANKS); machine_check_poll(MCP_UC | m_fl, &all_banks); - set_in_cr4(X86_CR4_MCE); + cr4_set_bits(X86_CR4_MCE); rdmsrl(MSR_IA32_MCG_CAP, cap); if (cap & MCG_CTL_P) diff --git a/arch/x86/kernel/cpu/mcheck/p5.c b/arch/x86/kernel/cpu/mcheck/p5.c index ec2663a708e4..737b0ad4e61a 100644 --- a/arch/x86/kernel/cpu/mcheck/p5.c +++ b/arch/x86/kernel/cpu/mcheck/p5.c @@ -9,6 +9,7 @@ #include #include +#include #include #include @@ -65,7 +66,7 @@ void intel_p5_mcheck_init(struct cpuinfo_x86 *c) "Intel old style machine check architecture supported.\n"); /* Enable MCE: */ - set_in_cr4(X86_CR4_MCE); + cr4_set_bits(X86_CR4_MCE); printk(KERN_INFO "Intel old style machine check reporting enabled on CPU#%d.\n", smp_processor_id()); diff --git a/arch/x86/kernel/cpu/mcheck/winchip.c b/arch/x86/kernel/cpu/mcheck/winchip.c index bd5d46a32210..44f138296fbe 100644 --- a/arch/x86/kernel/cpu/mcheck/winchip.c +++ b/arch/x86/kernel/cpu/mcheck/winchip.c @@ -8,6 +8,7 @@ #include #include +#include #include #include @@ -36,7 +37,7 @@ void winchip_mcheck_init(struct cpuinfo_x86 *c) lo &= ~(1<<4); /* Enable MCE */ wrmsr(MSR_IDT_FCR1, lo, hi); - set_in_cr4(X86_CR4_MCE); + cr4_set_bits(X86_CR4_MCE); printk(KERN_INFO "Winchip machine check reporting enabled on CPU#0.\n"); diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 143e5f5dc855..6b5acd5f4a34 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include #include @@ -1328,7 +1329,7 @@ x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu) case CPU_STARTING: if (x86_pmu.attr_rdpmc) - set_in_cr4(X86_CR4_PCE); + cr4_set_bits(X86_CR4_PCE); if (x86_pmu.cpu_starting) x86_pmu.cpu_starting(cpu); break; @@ -1834,9 +1835,9 @@ static void change_rdpmc(void *info) bool enable = !!(unsigned long)info; if (enable) - set_in_cr4(X86_CR4_PCE); + cr4_set_bits(X86_CR4_PCE); else - clear_in_cr4(X86_CR4_PCE); + cr4_clear_bits(X86_CR4_PCE); } static ssize_t set_attr_rdpmc(struct device *cdev, diff --git a/arch/x86/kernel/i387.c b/arch/x86/kernel/i387.c index a9a4229f6161..87727b03196d 100644 --- a/arch/x86/kernel/i387.c +++ b/arch/x86/kernel/i387.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include #include @@ -180,7 +181,7 @@ void fpu_init(void) if (cpu_has_xmm) cr4_mask |= X86_CR4_OSXMMEXCPT; if (cr4_mask) - set_in_cr4(cr4_mask); + cr4_set_bits(cr4_mask); cr0 = read_cr0(); cr0 &= ~(X86_CR0_TS|X86_CR0_EM); /* clear TS and EM */ diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index e127ddaa2d5a..046e2d620bbe 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -28,6 +28,7 @@ #include #include #include +#include /* * per-CPU TSS segments. Threads are completely 'soft' on Linux, @@ -141,7 +142,7 @@ void flush_thread(void) static void hard_disable_TSC(void) { - write_cr4(read_cr4() | X86_CR4_TSD); + cr4_set_bits(X86_CR4_TSD); } void disable_TSC(void) @@ -158,7 +159,7 @@ void disable_TSC(void) static void hard_enable_TSC(void) { - write_cr4(read_cr4() & ~X86_CR4_TSD); + cr4_clear_bits(X86_CR4_TSD); } static void enable_TSC(void) diff --git a/arch/x86/kernel/xsave.c b/arch/x86/kernel/xsave.c index 0de1fae2bdf0..34f66e58a896 100644 --- a/arch/x86/kernel/xsave.c +++ b/arch/x86/kernel/xsave.c @@ -12,6 +12,7 @@ #include #include #include +#include #include /* @@ -453,7 +454,7 @@ static void prepare_fx_sw_frame(void) */ static inline void xstate_enable(void) { - set_in_cr4(X86_CR4_OSXSAVE); + cr4_set_bits(X86_CR4_OSXSAVE); xsetbv(XCR_XFEATURE_ENABLED_MASK, pcntxt_mask); } -- cgit v1.2.1 From 1e02ce4cccdcb9688386e5b8d2c9fa4660b45389 Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Fri, 24 Oct 2014 15:58:08 -0700 Subject: x86: Store a per-cpu shadow copy of CR4 Context switches and TLB flushes can change individual bits of CR4. CR4 reads take several cycles, so store a shadow copy of CR4 in a per-cpu variable. To avoid wasting a cache line, I added the CR4 shadow to cpu_tlbstate, which is already touched in switch_mm. The heaviest users of the cr4 shadow will be switch_mm and __switch_to_xtra, and __switch_to_xtra is called shortly after switch_mm during context switch, so the cacheline is likely to be hot. Signed-off-by: Andy Lutomirski Reviewed-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Cc: Kees Cook Cc: Andrea Arcangeli Cc: Vince Weaver Cc: "hillf.zj" Cc: Valdis Kletnieks Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Linus Torvalds Link: http://lkml.kernel.org/r/3a54dd3353fffbf84804398e00dfdc5b7c1afd7d.1414190806.git.luto@amacapital.net Signed-off-by: Ingo Molnar --- arch/x86/kernel/acpi/sleep.c | 2 +- arch/x86/kernel/cpu/common.c | 7 +++++++ arch/x86/kernel/cpu/mtrr/cyrix.c | 6 +++--- arch/x86/kernel/cpu/mtrr/generic.c | 6 +++--- arch/x86/kernel/head32.c | 1 + arch/x86/kernel/head64.c | 2 ++ arch/x86/kernel/process_32.c | 2 +- arch/x86/kernel/process_64.c | 2 +- arch/x86/kernel/setup.c | 2 +- 9 files changed, 20 insertions(+), 10 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c index 31368207837c..d1daead5fcdd 100644 --- a/arch/x86/kernel/acpi/sleep.c +++ b/arch/x86/kernel/acpi/sleep.c @@ -78,7 +78,7 @@ int x86_acpi_suspend_lowlevel(void) header->pmode_cr0 = read_cr0(); if (__this_cpu_read(cpu_info.cpuid_level) >= 0) { - header->pmode_cr4 = read_cr4(); + header->pmode_cr4 = __read_cr4(); header->pmode_behavior |= (1 << WAKEUP_BEHAVIOR_RESTORE_CR4); } if (!rdmsr_safe(MSR_IA32_MISC_ENABLE, diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 9d8fc49f0922..07f2fc3c13a4 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -19,6 +19,7 @@ #include #include #include +#include #include #include #include @@ -1293,6 +1294,12 @@ void cpu_init(void) wait_for_master_cpu(cpu); + /* + * Initialize the CR4 shadow before doing anything that could + * try to read it. + */ + cr4_init_shadow(); + /* * Load microcode on this cpu if a valid microcode is available. * This is early microcode loading procedure. diff --git a/arch/x86/kernel/cpu/mtrr/cyrix.c b/arch/x86/kernel/cpu/mtrr/cyrix.c index 9e451b0876b5..f8c81ba0b465 100644 --- a/arch/x86/kernel/cpu/mtrr/cyrix.c +++ b/arch/x86/kernel/cpu/mtrr/cyrix.c @@ -138,8 +138,8 @@ static void prepare_set(void) /* Save value of CR4 and clear Page Global Enable (bit 7) */ if (cpu_has_pge) { - cr4 = read_cr4(); - write_cr4(cr4 & ~X86_CR4_PGE); + cr4 = __read_cr4(); + __write_cr4(cr4 & ~X86_CR4_PGE); } /* @@ -171,7 +171,7 @@ static void post_set(void) /* Restore value of CR4 */ if (cpu_has_pge) - write_cr4(cr4); + __write_cr4(cr4); } static void cyrix_set_arr(unsigned int reg, unsigned long base, diff --git a/arch/x86/kernel/cpu/mtrr/generic.c b/arch/x86/kernel/cpu/mtrr/generic.c index 0e25a1bc5ab5..7d74f7b3c6ba 100644 --- a/arch/x86/kernel/cpu/mtrr/generic.c +++ b/arch/x86/kernel/cpu/mtrr/generic.c @@ -678,8 +678,8 @@ static void prepare_set(void) __acquires(set_atomicity_lock) /* Save value of CR4 and clear Page Global Enable (bit 7) */ if (cpu_has_pge) { - cr4 = read_cr4(); - write_cr4(cr4 & ~X86_CR4_PGE); + cr4 = __read_cr4(); + __write_cr4(cr4 & ~X86_CR4_PGE); } /* Flush all TLBs via a mov %cr3, %reg; mov %reg, %cr3 */ @@ -708,7 +708,7 @@ static void post_set(void) __releases(set_atomicity_lock) /* Restore value of CR4 */ if (cpu_has_pge) - write_cr4(cr4); + __write_cr4(cr4); raw_spin_unlock(&set_atomicity_lock); } diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c index d6c1b9836995..2911ef3a9f1c 100644 --- a/arch/x86/kernel/head32.c +++ b/arch/x86/kernel/head32.c @@ -31,6 +31,7 @@ static void __init i386_default_early_setup(void) asmlinkage __visible void __init i386_start_kernel(void) { + cr4_init_shadow(); sanitize_boot_params(&boot_params); /* Call the subarch specific early setup function */ diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c index eda1a865641e..3b241f0ca005 100644 --- a/arch/x86/kernel/head64.c +++ b/arch/x86/kernel/head64.c @@ -155,6 +155,8 @@ asmlinkage __visible void __init x86_64_start_kernel(char * real_mode_data) (__START_KERNEL & PGDIR_MASK))); BUILD_BUG_ON(__fix_to_virt(__end_of_fixed_addresses) <= MODULES_END); + cr4_init_shadow(); + /* Kill off the identity-map trampoline */ reset_early_page_tables(); diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 8f3ebfe710d0..603c4f99cb5a 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -101,7 +101,7 @@ void __show_regs(struct pt_regs *regs, int all) cr0 = read_cr0(); cr2 = read_cr2(); cr3 = read_cr3(); - cr4 = read_cr4_safe(); + cr4 = __read_cr4_safe(); printk(KERN_DEFAULT "CR0: %08lx CR2: %08lx CR3: %08lx CR4: %08lx\n", cr0, cr2, cr3, cr4); diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 5a2c02913af3..67fcc43577d2 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -93,7 +93,7 @@ void __show_regs(struct pt_regs *regs, int all) cr0 = read_cr0(); cr2 = read_cr2(); cr3 = read_cr3(); - cr4 = read_cr4(); + cr4 = __read_cr4(); printk(KERN_DEFAULT "FS: %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n", fs, fsindex, gs, gsindex, shadowgs); diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index ab4734e5411d..04e6c62f1a93 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -1178,7 +1178,7 @@ void __init setup_arch(char **cmdline_p) if (boot_cpu_data.cpuid_level >= 0) { /* A CPU has %cr4 if and only if it has CPUID */ - mmu_cr4_features = read_cr4(); + mmu_cr4_features = __read_cr4(); if (trampoline_cr4_features) *trampoline_cr4_features = mmu_cr4_features; } -- cgit v1.2.1 From c1317ec2b906442930318d9d6e51425c5a69e9cb Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Fri, 24 Oct 2014 15:58:11 -0700 Subject: perf: Pass the event to arch_perf_update_userpage() Signed-off-by: Andy Lutomirski Signed-off-by: Peter Zijlstra (Intel) Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Kees Cook Cc: Andrea Arcangeli Cc: Vince Weaver Cc: "hillf.zj" Cc: Valdis Kletnieks Cc: Linus Torvalds Link: http://lkml.kernel.org/r/0fea9a7fac3c1eea86cb0a5954184e74f4213666.1414190806.git.luto@amacapital.net Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 6b5acd5f4a34..73e84a348de1 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -1915,7 +1915,8 @@ static struct pmu pmu = { .flush_branch_stack = x86_pmu_flush_branch_stack, }; -void arch_perf_update_userpage(struct perf_event_mmap_page *userpg, u64 now) +void arch_perf_update_userpage(struct perf_event *event, + struct perf_event_mmap_page *userpg, u64 now) { struct cyc2ns_data *data; -- cgit v1.2.1 From 7911d3f7af14a614617e38245fedf98a724e46a9 Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Fri, 24 Oct 2014 15:58:12 -0700 Subject: perf/x86: Only allow rdpmc if a perf_event is mapped We currently allow any process to use rdpmc. This significantly weakens the protection offered by PR_TSC_DISABLED, and it could be helpful to users attempting to exploit timing attacks. Since we can't enable access to individual counters, use a very coarse heuristic to limit access to rdpmc: allow access only when a perf_event is mmapped. This protects seccomp sandboxes. There is plenty of room to further tighen these restrictions. For example, this allows rdpmc for any x86_pmu event, but it's only useful for self-monitoring tasks. As a side effect, cap_user_rdpmc will now be false for AMD uncore events. This isn't a real regression, since .event_idx is disabled for these events anyway for the time being. Whenever that gets re-added, the cap_user_rdpmc code can be adjusted or refactored accordingly. Signed-off-by: Andy Lutomirski Signed-off-by: Peter Zijlstra (Intel) Cc: Arnaldo Carvalho de Melo Cc: Kees Cook Cc: Andrea Arcangeli Cc: Vince Weaver Cc: "hillf.zj" Cc: Valdis Kletnieks Cc: Linus Torvalds Link: http://lkml.kernel.org/r/a2bdb3cf3a1d70c26980d7c6dddfbaa69f3182bf.1414190806.git.luto@amacapital.net Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 57 ++++++++++++++++++++++++++-------------- arch/x86/kernel/cpu/perf_event.h | 2 ++ 2 files changed, 40 insertions(+), 19 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 73e84a348de1..bec5cff7dc80 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -31,6 +31,7 @@ #include #include #include +#include #include #include #include @@ -1328,8 +1329,6 @@ x86_pmu_notifier(struct notifier_block *self, unsigned long action, void *hcpu) break; case CPU_STARTING: - if (x86_pmu.attr_rdpmc) - cr4_set_bits(X86_CR4_PCE); if (x86_pmu.cpu_starting) x86_pmu.cpu_starting(cpu); break; @@ -1805,14 +1804,44 @@ static int x86_pmu_event_init(struct perf_event *event) event->destroy(event); } + if (ACCESS_ONCE(x86_pmu.attr_rdpmc)) + event->hw.flags |= PERF_X86_EVENT_RDPMC_ALLOWED; + return err; } +static void refresh_pce(void *ignored) +{ + if (current->mm) + load_mm_cr4(current->mm); +} + +static void x86_pmu_event_mapped(struct perf_event *event) +{ + if (!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED)) + return; + + if (atomic_inc_return(¤t->mm->context.perf_rdpmc_allowed) == 1) + on_each_cpu_mask(mm_cpumask(current->mm), refresh_pce, NULL, 1); +} + +static void x86_pmu_event_unmapped(struct perf_event *event) +{ + if (!current->mm) + return; + + if (!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED)) + return; + + if (atomic_dec_and_test(¤t->mm->context.perf_rdpmc_allowed)) + on_each_cpu_mask(mm_cpumask(current->mm), refresh_pce, NULL, 1); +} + static int x86_pmu_event_idx(struct perf_event *event) { int idx = event->hw.idx; - if (!x86_pmu.attr_rdpmc) + if (!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED)) return 0; if (x86_pmu.num_counters_fixed && idx >= INTEL_PMC_IDX_FIXED) { @@ -1830,16 +1859,6 @@ static ssize_t get_attr_rdpmc(struct device *cdev, return snprintf(buf, 40, "%d\n", x86_pmu.attr_rdpmc); } -static void change_rdpmc(void *info) -{ - bool enable = !!(unsigned long)info; - - if (enable) - cr4_set_bits(X86_CR4_PCE); - else - cr4_clear_bits(X86_CR4_PCE); -} - static ssize_t set_attr_rdpmc(struct device *cdev, struct device_attribute *attr, const char *buf, size_t count) @@ -1854,11 +1873,7 @@ static ssize_t set_attr_rdpmc(struct device *cdev, if (x86_pmu.attr_rdpmc_broken) return -ENOTSUPP; - if (!!val != !!x86_pmu.attr_rdpmc) { - x86_pmu.attr_rdpmc = !!val; - on_each_cpu(change_rdpmc, (void *)val, 1); - } - + x86_pmu.attr_rdpmc = !!val; return count; } @@ -1901,6 +1916,9 @@ static struct pmu pmu = { .event_init = x86_pmu_event_init, + .event_mapped = x86_pmu_event_mapped, + .event_unmapped = x86_pmu_event_unmapped, + .add = x86_pmu_add, .del = x86_pmu_del, .start = x86_pmu_start, @@ -1922,7 +1940,8 @@ void arch_perf_update_userpage(struct perf_event *event, userpg->cap_user_time = 0; userpg->cap_user_time_zero = 0; - userpg->cap_user_rdpmc = x86_pmu.attr_rdpmc; + userpg->cap_user_rdpmc = + !!(event->hw.flags & PERF_X86_EVENT_RDPMC_ALLOWED); userpg->pmc_width = x86_pmu.cntval_bits; if (!sched_clock_stable()) diff --git a/arch/x86/kernel/cpu/perf_event.h b/arch/x86/kernel/cpu/perf_event.h index 4e6cdb0ddc70..df525d2be1e8 100644 --- a/arch/x86/kernel/cpu/perf_event.h +++ b/arch/x86/kernel/cpu/perf_event.h @@ -71,6 +71,8 @@ struct event_constraint { #define PERF_X86_EVENT_COMMITTED 0x8 /* event passed commit_txn */ #define PERF_X86_EVENT_PEBS_LD_HSW 0x10 /* haswell style datala, load */ #define PERF_X86_EVENT_PEBS_NA_HSW 0x20 /* haswell style datala, unknown */ +#define PERF_X86_EVENT_RDPMC_ALLOWED 0x40 /* grant rdpmc permission */ + struct amd_nb { int nb_id; /* NorthBridge id */ -- cgit v1.2.1 From a66734297f78707ce39d756b656bfae861d53f62 Mon Sep 17 00:00:00 2001 From: Andy Lutomirski Date: Fri, 24 Oct 2014 15:58:13 -0700 Subject: perf/x86: Add /sys/devices/cpu/rdpmc=2 to allow rdpmc for all tasks While perfmon2 is a sufficiently evil library (it pokes MSRs directly) that breaking it is fair game, it's still useful, so we might as well try to support it. This allows users to write 2 to /sys/devices/cpu/rdpmc to disable all rdpmc protection so that hack like perfmon2 can continue to work. At some point, if perf_event becomes fast enough to replace perfmon2, then this can go. Signed-off-by: Andy Lutomirski Signed-off-by: Peter Zijlstra (Intel) Cc: Paul Mackerras Cc: Arnaldo Carvalho de Melo Cc: Kees Cook Cc: Andrea Arcangeli Cc: Vince Weaver Cc: "hillf.zj" Cc: Valdis Kletnieks Cc: Linus Torvalds Link: http://lkml.kernel.org/r/caac3c1c707dcca48ecbc35f4def21495856f479.1414190806.git.luto@amacapital.net Signed-off-by: Ingo Molnar --- arch/x86/kernel/cpu/perf_event.c | 21 ++++++++++++++++++++- 1 file changed, 20 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index bec5cff7dc80..b71a7f86d68a 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -45,6 +45,8 @@ DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = { .enabled = 1, }; +struct static_key rdpmc_always_available = STATIC_KEY_INIT_FALSE; + u64 __read_mostly hw_cache_event_ids [PERF_COUNT_HW_CACHE_MAX] [PERF_COUNT_HW_CACHE_OP_MAX] @@ -1870,10 +1872,27 @@ static ssize_t set_attr_rdpmc(struct device *cdev, if (ret) return ret; + if (val > 2) + return -EINVAL; + if (x86_pmu.attr_rdpmc_broken) return -ENOTSUPP; - x86_pmu.attr_rdpmc = !!val; + if ((val == 2) != (x86_pmu.attr_rdpmc == 2)) { + /* + * Changing into or out of always available, aka + * perf-event-bypassing mode. This path is extremely slow, + * but only root can trigger it, so it's okay. + */ + if (val == 2) + static_key_slow_inc(&rdpmc_always_available); + else + static_key_slow_dec(&rdpmc_always_available); + on_each_cpu(refresh_pce, NULL, 1); + } + + x86_pmu.attr_rdpmc = val; + return count; } -- cgit v1.2.1