Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2eafeb6a · Linus Torvalds <[email protected]> · 2012-07-22 11:10:36 -07:00

Pull perf events changes from Ingo Molnar:

 "- kernel side:

   - Intel uncore PMU support for Nehalem and Sandy Bridge CPUs, we
     support both the events available via the MSR and via the PCI
     access space.

   - various uprobes cleanups and restructurings

   - PMU driver quirks by microcode version and required x86 microcode
     loader cleanups/robustization

   - various tracing robustness updates

   - static keys: remove obsolete static_branch()

  - tooling side:

   - GTK browser improvements

   - perf report browser: support screenshots to file

   - more automated tests

   - perf kvm improvements

   - perf bench refinements

   - build environment improvements

   - pipe mode improvements

   - libtraceevent updates, we have now hopefully merged most bits with
     the out of tree forked code base

  ... and many other goodies."

* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (138 commits)
  tracing: Check for allocation failure in __tracing_open()
  perf/x86: Fix intel_perfmon_event_mapformatting
  jump label: Remove static_branch()
  tracepoint: Use static_key_false(), since static_branch() is deprecated
  perf/x86: Uncore filter support for SandyBridge-EP
  perf/x86: Detect number of instances of uncore CBox
  perf/x86: Fix event constraint for SandyBridge-EP C-Box
  perf/x86: Use 0xff as pseudo code for fixed uncore event
  perf/x86: Save a few bytes in 'struct x86_pmu'
  perf/x86: Add a microcode revision check for SNB-PEBS
  perf/x86: Improve debug output in check_hw_exists()
  perf/x86/amd: Unify AMD's generic and family 15h pmus
  perf/x86: Move Intel specific code to intel_pmu_init()
  perf/x86: Rename Intel specific macros
  perf/x86: Fix USER/KERNEL tagging of samples
  perf tools: Split event symbols arrays to hw and sw parts
  perf tools: Split out PE_VALUE_SYM parsing token to SW and HW tokens
  perf tools: Add empty rule for new line in event syntax parsing
  perf test: Use ARRAY_SIZE in parse events tests
  tools lib traceevent: Cleanup realloc use
  ...

Commit: 2eafeb6a415864bc4c59df79151cf40f6ac74b9e
Parents: 16d286e6 · 6e0f17be

Merge commit. Changes below are compared with the first parent: 16d286e6.

Changes

109 files changed · +5782 −1706

File	Status	Add	Del	Content
arch/x86/include/asm/kvm_host.h	modified	+2	−2	View file
arch/x86/include/asm/msr.h	modified	+5	−41	View file
arch/x86/include/asm/paravirt.h	modified	+2	−39	View file
arch/x86/include/asm/paravirt_types.h	modified	+0	−2	View file
arch/x86/include/asm/perf_event.h	modified	+11	−11	View file
arch/x86/include/asm/uprobes.h	modified	+1	−1	View file
arch/x86/kernel/cpu/Makefile	modified	+3	−1	View file
arch/x86/kernel/cpu/amd.c	modified	+36	−3	View file
arch/x86/kernel/cpu/common.c	modified	+1	−1	View file
arch/x86/kernel/cpu/perf_event.c	modified	+43	−68	View file
arch/x86/kernel/cpu/perf_event.h	modified	+23	−3	View file
arch/x86/kernel/cpu/perf_event_amd.c	modified	+46	−57	View file
arch/x86/kernel/cpu/perf_event_intel.c	modified	+101	−21	View file
arch/x86/kernel/cpu/perf_event_intel_ds.c	modified	+6	−6	View file
arch/x86/kernel/cpu/perf_event_intel_uncore.c	added	+1850	−0	View file
arch/x86/kernel/cpu/perf_event_intel_uncore.h	added	+424	−0	View file
arch/x86/kernel/cpu/perf_event_p4.c	modified	+8	−8	View file
arch/x86/kernel/cpu/perf_event_p6.c	modified	+2	−2	View file
arch/x86/kernel/microcode_core.c	modified	+50	−14	View file
arch/x86/kernel/paravirt.c	modified	+0	−2	View file
arch/x86/kernel/process_64.c	modified	+2	−2	View file
arch/x86/kernel/uprobes.c	modified	+2	−1	View file
arch/x86/kvm/pmu.c	modified	+11	−11	View file
arch/x86/kvm/trace.h	modified	+1	−11	View file
arch/x86/lib/msr-reg-export.c	modified	+2	−2	View file
arch/x86/lib/msr-reg.S	modified	+5	−5	View file
arch/x86/oprofile/op_model_amd.c	modified	+2	−2	View file
arch/x86/vdso/vdso32-setup.c	modified	+3	−3	View file
arch/x86/xen/enlighten.c	modified	+0	−2	View file
drivers/oprofile/oprofile_perf.c	modified	+11	−12	View file
include/linux/ftrace_event.h	modified	+4	−1	View file
include/linux/jump_label.h	modified	+1	−16	View file
include/linux/pci_ids.h	modified	+11	−0	View file
include/linux/perf_event.h	modified	+3	−0	View file
include/linux/sched.h	modified	+0	−1	View file
include/linux/tracepoint.h	modified	+1	−1	View file
include/trace/ftrace.h	modified	+1	−0	View file
kernel/events/core.c	modified	+45	−4	View file
kernel/events/uprobes.c	modified	+228	−233	View file
kernel/trace/ftrace.c	modified	+2	−6	View file
kernel/trace/ring_buffer.c	modified	+4	−0	View file
kernel/trace/trace.c	modified	+24	−9	View file
kernel/trace/trace.h	modified	+8	−0	View file
kernel/trace/trace_functions_graph.c	modified	+1	−1	View file
kernel/trace/trace_output.c	modified	+1	−1	View file
tools/lib/traceevent/Makefile	modified	+10	−4	View file
tools/lib/traceevent/event-parse.c	modified	+295	−104	View file
tools/lib/traceevent/event-parse.h	modified	+7	−0	View file
tools/lib/traceevent/parse-filter.c	modified	+57	−29	View file
tools/perf/Documentation/perf-bench.txt	modified	+75	−3	View file
tools/perf/Documentation/perf-report.txt	modified	+1	−1	View file
tools/perf/Documentation/perf-top.txt	modified	+1	−1	View file
tools/perf/Makefile	modified	+7	−1	View file
tools/perf/bench/mem-memcpy.c	modified	+42	−42	View file
tools/perf/bench/mem-memset.c	modified	+44	−44	View file
tools/perf/builtin-bench.c	modified	+2	−2	View file
tools/perf/builtin-evlist.c	modified	+1	−1	View file
tools/perf/builtin-kmem.c	modified	+25	−12	View file
tools/perf/builtin-lock.c	modified	+2	−2	View file
tools/perf/builtin-record.c	modified	+2	−2	View file
tools/perf/builtin-report.c	modified	+9	−7	View file
tools/perf/builtin-sched.c	modified	+26	−12	View file
tools/perf/builtin-script.c	modified	+65	−41	View file
tools/perf/builtin-stat.c	modified	+6	−6	View file
tools/perf/builtin-test.c	modified	+1	−1	View file
tools/perf/builtin-top.c	modified	+6	−6	View file
tools/perf/config/feature-tests.mak	modified	+13	−0	View file
tools/perf/ui/browsers/annotate.c	modified	+3	−1	View file
tools/perf/ui/browsers/hists.c	modified	+199	−11	View file
tools/perf/ui/gtk/browser.c	modified	+66	−3	View file
tools/perf/ui/gtk/gtk.h	modified	+31	−0	View file
tools/perf/ui/gtk/setup.c	modified	+5	−0	View file
tools/perf/ui/gtk/util.c	added	+129	−0	View file
tools/perf/ui/tui/setup.c	modified	+6	−0	View file
tools/perf/ui/tui/util.c	added	+243	−0	View file
tools/perf/ui/util.c	modified	+56	−221	View file
tools/perf/ui/util.h	modified	+8	−1	View file
tools/perf/util/debug.c	modified	+1	−1	View file
tools/perf/util/debug.h	modified	+20	−3	View file
tools/perf/util/evlist.c	modified	+2	−2	View file
tools/perf/util/evlist.h	modified	+3	−0	View file
tools/perf/util/evsel.c	modified	+199	−28	View file
tools/perf/util/evsel.h	modified	+13	−2	View file
tools/perf/util/header.c	modified	+18	−15	View file
tools/perf/util/hist.h	modified	+1	−0	View file
tools/perf/util/include/linux/kernel.h	modified	+10	−0	View file
tools/perf/util/map.h	modified	+1	−1	View file
tools/perf/util/parse-events-test.c	modified	+165	−14	View file
tools/perf/util/parse-events.c	modified	+208	−262	View file
tools/perf/util/parse-events.h	modified	+13	−4	View file
tools/perf/util/parse-events.l	modified	+83	−51	View file
tools/perf/util/parse-events.y	modified	+63	−23	View file
tools/perf/util/pmu.c	modified	+168	−1	View file
tools/perf/util/pmu.h	modified	+10	−1	View file
tools/perf/util/scripting-engines/trace-event-perl.c	modified	+20	−12	View file
tools/perf/util/scripting-engines/trace-event-python.c	modified	+16	−9	View file
tools/perf/util/session.c	modified	+60	−5	View file
tools/perf/util/session.h	modified	+12	−2	View file
tools/perf/util/sort.c	modified	+49	−0	View file
tools/perf/util/sort.h	modified	+2	−0	View file
tools/perf/util/string.c	modified	+22	−0	View file
tools/perf/util/symbol.c	modified	+87	−7	View file
tools/perf/util/symbol.h	modified	+1	−0	View file
tools/perf/util/top.c	modified	+1	−1	View file
tools/perf/util/trace-event-parse.c	modified	+26	−32	View file
tools/perf/util/trace-event-read.c	modified	+50	−47	View file
tools/perf/util/trace-event-scripting.c	modified	+5	−2	View file
tools/perf/util/trace-event.h	modified	+21	−17	View file
tools/perf/util/util.h	modified	+2	−0	View file

Only the first 600 diff lines are shown. Download the complete patch.

arch/x86/include/asm/kvm_host.h

View file

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index db7c1f2..2da88c0 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -313,8 +313,8 @@ struct kvm_pmu {
 	u64 counter_bitmask[2];
 	u64 global_ctrl_mask;
 	u8 version;
-	struct kvm_pmc gp_counters[X86_PMC_MAX_GENERIC];
-	struct kvm_pmc fixed_counters[X86_PMC_MAX_FIXED];
+	struct kvm_pmc gp_counters[INTEL_PMC_MAX_GENERIC];
+	struct kvm_pmc fixed_counters[INTEL_PMC_MAX_FIXED];
 	struct irq_work irq_work;
 	u64 reprogram_pmi;
 };

arch/x86/include/asm/msr.h

View file

diff --git a/arch/x86/include/asm/msr.h b/arch/x86/include/asm/msr.h
index 084ef95..813ed10 100644
--- a/arch/x86/include/asm/msr.h
+++ b/arch/x86/include/asm/msr.h
@@ -115,8 +115,8 @@ notrace static inline int native_write_msr_safe(unsigned int msr,
 
 extern unsigned long long native_read_tsc(void);
 
-extern int native_rdmsr_safe_regs(u32 regs[8]);
-extern int native_wrmsr_safe_regs(u32 regs[8]);
+extern int rdmsr_safe_regs(u32 regs[8]);
+extern int wrmsr_safe_regs(u32 regs[8]);
 
 static __always_inline unsigned long long __native_read_tsc(void)
 {
@@ -187,43 +187,6 @@ static inline int rdmsrl_safe(unsigned msr, unsigned long long *p)
 	return err;
 }
 
-static inline int rdmsrl_amd_safe(unsigned msr, unsigned long long *p)
-{
-	u32 gprs[8] = { 0 };
-	int err;
-
-	gprs[1] = msr;
-	gprs[7] = 0x9c5a203a;
-
-	err = native_rdmsr_safe_regs(gprs);
-
-	*p = gprs[0] | ((u64)gprs[2] << 32);
-
-	return err;
-}
-
-static inline int wrmsrl_amd_safe(unsigned msr, unsigned long long val)
-{
-	u32 gprs[8] = { 0 };
-
-	gprs[0] = (u32)val;
-	gprs[1] = msr;
-	gprs[2] = val >> 32;
-	gprs[7] = 0x9c5a203a;
-
-	return native_wrmsr_safe_regs(gprs);
-}
-
-static inline int rdmsr_safe_regs(u32 regs[8])
-{
-	return native_rdmsr_safe_regs(regs);
-}
-
-static inline int wrmsr_safe_regs(u32 regs[8])
-{
-	return native_wrmsr_safe_regs(regs);
-}
-
 #define rdtscl(low)						\
 	((low) = (u32)__native_read_tsc())
 
@@ -237,6 +200,8 @@ do {							\
 	(high) = (u32)(_l >> 32);			\
 } while (0)
 
+#define rdpmcl(counter, val) ((val) = native_read_pmc(counter))
+
 #define rdtscp(low, high, aux)					\
 do {                                                            \
 	unsigned long long _val = native_read_tscp(&(aux));     \
@@ -248,8 +213,7 @@ do {                                                            \
 
 #endif	/* !CONFIG_PARAVIRT */
 
-
-#define checking_wrmsrl(msr, val) wrmsr_safe((msr), (u32)(val),		\
+#define wrmsrl_safe(msr, val) wrmsr_safe((msr), (u32)(val),		\
 					     (u32)((val) >> 32))
 
 #define write_tsc(val1, val2) wrmsr(MSR_IA32_TSC, (val1), (val2))

arch/x86/include/asm/paravirt.h

View file

diff --git a/arch/x86/include/asm/paravirt.h b/arch/x86/include/asm/paravirt.h
index 6cbbabf..0b47ddb 100644
--- a/arch/x86/include/asm/paravirt.h
+++ b/arch/x86/include/asm/paravirt.h
@@ -128,21 +128,11 @@ static inline u64 paravirt_read_msr(unsigned msr, int *err)
 	return PVOP_CALL2(u64, pv_cpu_ops.read_msr, msr, err);
 }
 
-static inline int paravirt_rdmsr_regs(u32 *regs)
-{
-	return PVOP_CALL1(int, pv_cpu_ops.rdmsr_regs, regs);
-}
-
 static inline int paravirt_write_msr(unsigned msr, unsigned low, unsigned high)
 {
 	return PVOP_CALL3(int, pv_cpu_ops.write_msr, msr, low, high);
 }
 
-static inline int paravirt_wrmsr_regs(u32 *regs)
-{
-	return PVOP_CALL1(int, pv_cpu_ops.wrmsr_regs, regs);
-}
-
 /* These should all do BUG_ON(_err), but our headers are too tangled. */
 #define rdmsr(msr, val1, val2)			\
 do {						\
@@ -176,9 +166,6 @@ do {						\
 	_err;					\
 })
 
-#define rdmsr_safe_regs(regs)	paravirt_rdmsr_regs(regs)
-#define wrmsr_safe_regs(regs)	paravirt_wrmsr_regs(regs)
-
 static inline int rdmsrl_safe(unsigned msr, unsigned long long *p)
 {
 	int err;
@@ -186,32 +173,6 @@ static inline int rdmsrl_safe(unsigned msr, unsigned long long *p)
 	*p = paravirt_read_msr(msr, &err);
 	return err;
 }
-static inline int rdmsrl_amd_safe(unsigned msr, unsigned long long *p)
-{
-	u32 gprs[8] = { 0 };
-	int err;
-
-	gprs[1] = msr;
-	gprs[7] = 0x9c5a203a;
-
-	err = paravirt_rdmsr_regs(gprs);
-
-	*p = gprs[0] | ((u64)gprs[2] << 32);
-
-	return err;
-}
-
-static inline int wrmsrl_amd_safe(unsigned msr, unsigned long long val)
-{
-	u32 gprs[8] = { 0 };
-
-	gprs[0] = (u32)val;
-	gprs[1] = msr;
-	gprs[2] = val >> 32;
-	gprs[7] = 0x9c5a203a;
-
-	return paravirt_wrmsr_regs(gprs);
-}
 
 static inline u64 paravirt_read_tsc(void)
 {
@@ -252,6 +213,8 @@ do {						\
 	high = _l >> 32;			\
 } while (0)
 
+#define rdpmcl(counter, val) ((val) = paravirt_read_pmc(counter))
+
 static inline unsigned long long paravirt_rdtscp(unsigned int *aux)
 {
 	return PVOP_CALL1(u64, pv_cpu_ops.read_tscp, aux);

arch/x86/include/asm/paravirt_types.h

View file

diff --git a/arch/x86/include/asm/paravirt_types.h b/arch/x86/include/asm/paravirt_types.h
index 8e8b9a4..8613cbb 100644
--- a/arch/x86/include/asm/paravirt_types.h
+++ b/arch/x86/include/asm/paravirt_types.h
@@ -153,9 +153,7 @@ struct pv_cpu_ops {
 	/* MSR, PMC and TSR operations.
 	   err = 0/-EFAULT.  wrmsr returns 0/-EFAULT. */
 	u64 (*read_msr)(unsigned int msr, int *err);
-	int (*rdmsr_regs)(u32 *regs);
 	int (*write_msr)(unsigned int msr, unsigned low, unsigned high);
-	int (*wrmsr_regs)(u32 *regs);
 
 	u64 (*read_tsc)(void);
 	u64 (*read_pmc)(int counter);

arch/x86/include/asm/perf_event.h

View file

diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h
index 588f52e..c78f14a 100644
--- a/arch/x86/include/asm/perf_event.h
+++ b/arch/x86/include/asm/perf_event.h
@@ -5,11 +5,10 @@
  * Performance event hw details:
  */
 
-#define X86_PMC_MAX_GENERIC				       32
-#define X86_PMC_MAX_FIXED					3
+#define INTEL_PMC_MAX_GENERIC				       32
+#define INTEL_PMC_MAX_FIXED					3
+#define INTEL_PMC_IDX_FIXED				       32
 
-#define X86_PMC_IDX_GENERIC				        0
-#define X86_PMC_IDX_FIXED				       32
 #define X86_PMC_IDX_MAX					       64
 
 #define MSR_ARCH_PERFMON_PERFCTR0			      0xc1
@@ -48,8 +47,7 @@
 	(X86_RAW_EVENT_MASK          |  \
 	 AMD64_EVENTSEL_EVENT)
 #define AMD64_NUM_COUNTERS				4
-#define AMD64_NUM_COUNTERS_F15H				6
-#define AMD64_NUM_COUNTERS_MAX				AMD64_NUM_COUNTERS_F15H
+#define AMD64_NUM_COUNTERS_CORE				6
 
 #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_SEL		0x3c
 #define ARCH_PERFMON_UNHALTED_CORE_CYCLES_UMASK		(0x00 << 8)
@@ -121,16 +119,16 @@ struct x86_pmu_capability {
 
 /* Instr_Retired.Any: */
 #define MSR_ARCH_PERFMON_FIXED_CTR0	0x309
-#define X86_PMC_IDX_FIXED_INSTRUCTIONS	(X86_PMC_IDX_FIXED + 0)
+#define INTEL_PMC_IDX_FIXED_INSTRUCTIONS	(INTEL_PMC_IDX_FIXED + 0)
 
 /* CPU_CLK_Unhalted.Core: */
 #define MSR_ARCH_PERFMON_FIXED_CTR1	0x30a
-#define X86_PMC_IDX_FIXED_CPU_CYCLES	(X86_PMC_IDX_FIXED + 1)
+#define INTEL_PMC_IDX_FIXED_CPU_CYCLES	(INTEL_PMC_IDX_FIXED + 1)
 
 /* CPU_CLK_Unhalted.Ref: */
 #define MSR_ARCH_PERFMON_FIXED_CTR2	0x30b
-#define X86_PMC_IDX_FIXED_REF_CYCLES	(X86_PMC_IDX_FIXED + 2)
-#define X86_PMC_MSK_FIXED_REF_CYCLES	(1ULL << X86_PMC_IDX_FIXED_REF_CYCLES)
+#define INTEL_PMC_IDX_FIXED_REF_CYCLES	(INTEL_PMC_IDX_FIXED + 2)
+#define INTEL_PMC_MSK_FIXED_REF_CYCLES	(1ULL << INTEL_PMC_IDX_FIXED_REF_CYCLES)
 
 /*
  * We model BTS tracing as another fixed-mode PMC.
@@ -139,7 +137,7 @@ struct x86_pmu_capability {
  * values are used by actual fixed events and higher values are used
  * to indicate other overflow conditions in the PERF_GLOBAL_STATUS msr.
  */
-#define X86_PMC_IDX_FIXED_BTS				(X86_PMC_IDX_FIXED + 16)
+#define INTEL_PMC_IDX_FIXED_BTS				(INTEL_PMC_IDX_FIXED + 16)
 
 /*
  * IBS cpuid feature detection
@@ -234,6 +232,7 @@ struct perf_guest_switch_msr {
 
 extern struct perf_guest_switch_msr *perf_guest_get_msrs(int *nr);
 extern void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap);
+extern void perf_check_microcode(void);
 #else
 static inline perf_guest_switch_msr *perf_guest_get_msrs(int *nr)
 {
@@ -247,6 +246,7 @@ static inline void perf_get_x86_pmu_capability(struct x86_pmu_capability *cap)
 }
 
 static inline void perf_events_lapic_init(void)	{ }
+static inline void perf_check_microcode(void) { }
 #endif
 
 #if defined(CONFIG_PERF_EVENTS) && defined(CONFIG_CPU_SUP_AMD)

arch/x86/include/asm/uprobes.h

View file

diff --git a/arch/x86/include/asm/uprobes.h b/arch/x86/include/asm/uprobes.h
index 1e9bed1..f3971bb 100644
--- a/arch/x86/include/asm/uprobes.h
+++ b/arch/x86/include/asm/uprobes.h
@@ -48,7 +48,7 @@ struct arch_uprobe_task {
 #endif
 };
 
-extern int  arch_uprobe_analyze_insn(struct arch_uprobe *aup, struct mm_struct *mm);
+extern int  arch_uprobe_analyze_insn(struct arch_uprobe *aup, struct mm_struct *mm, unsigned long addr);
 extern int  arch_uprobe_pre_xol(struct arch_uprobe *aup, struct pt_regs *regs);
 extern int  arch_uprobe_post_xol(struct arch_uprobe *aup, struct pt_regs *regs);
 extern bool arch_uprobe_xol_was_trapped(struct task_struct *tsk);

arch/x86/kernel/cpu/Makefile

View file

diff --git a/arch/x86/kernel/cpu/Makefile b/arch/x86/kernel/cpu/Makefile
index 6ab6aa2..bac4c38 100644
--- a/arch/x86/kernel/cpu/Makefile
+++ b/arch/x86/kernel/cpu/Makefile
@@ -32,7 +32,9 @@ obj-$(CONFIG_PERF_EVENTS)		+= perf_event.o
 
 ifdef CONFIG_PERF_EVENTS
 obj-$(CONFIG_CPU_SUP_AMD)		+= perf_event_amd.o
-obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_p6.o perf_event_p4.o perf_event_intel_lbr.o perf_event_intel_ds.o perf_event_intel.o
+obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_p6.o perf_event_p4.o
+obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_intel_lbr.o perf_event_intel_ds.o perf_event_intel.o
+obj-$(CONFIG_CPU_SUP_INTEL)		+= perf_event_intel_uncore.o
 endif
 
 obj-$(CONFIG_X86_MCE)			+= mcheck/

arch/x86/kernel/cpu/amd.c

View file

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index 146bb62..9d92e19 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -19,6 +19,39 @@
 
 #include "cpu.h"
 
+static inline int rdmsrl_amd_safe(unsigned msr, unsigned long long *p)
+{
+	struct cpuinfo_x86 *c = &cpu_data(smp_processor_id());
+	u32 gprs[8] = { 0 };
+	int err;
+
+	WARN_ONCE((c->x86 != 0xf), "%s should only be used on K8!\n", __func__);
+
+	gprs[1] = msr;
+	gprs[7] = 0x9c5a203a;
+
+	err = rdmsr_safe_regs(gprs);
+
+	*p = gprs[0] | ((u64)gprs[2] << 32);
+
+	return err;
+}
+
+static inline int wrmsrl_amd_safe(unsigned msr, unsigned long long val)
+{
+	struct cpuinfo_x86 *c = &cpu_data(smp_processor_id());
+	u32 gprs[8] = { 0 };
+
+	WARN_ONCE((c->x86 != 0xf), "%s should only be used on K8!\n", __func__);
+
+	gprs[0] = (u32)val;
+	gprs[1] = msr;
+	gprs[2] = val >> 32;
+	gprs[7] = 0x9c5a203a;
+
+	return wrmsr_safe_regs(gprs);
+}
+
 #ifdef CONFIG_X86_32
 /*
  *	B step AMD K6 before B 9730xxxx have hardware bugs that can cause
@@ -586,9 +619,9 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c)
 	    !cpu_has(c, X86_FEATURE_TOPOEXT)) {
 		u64 val;
 
-		if (!rdmsrl_amd_safe(0xc0011005, &val)) {
+		if (!rdmsrl_safe(0xc0011005, &val)) {
 			val |= 1ULL << 54;
-			wrmsrl_amd_safe(0xc0011005, val);
+			wrmsrl_safe(0xc0011005, val);
 			rdmsrl(0xc0011005, val);
 			if (val & (1ULL << 54)) {
 				set_cpu_cap(c, X86_FEATURE_TOPOEXT);
@@ -679,7 +712,7 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c)
 		err = rdmsrl_safe(MSR_AMD64_MCx_MASK(4), &mask);
 		if (err == 0) {
 			mask |= (1 << 10);
-			checking_wrmsrl(MSR_AMD64_MCx_MASK(4), mask);
+			wrmsrl_safe(MSR_AMD64_MCx_MASK(4), mask);
 		}
 	}
 

arch/x86/kernel/cpu/common.c

View file

diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 6b9333b..5bbc082 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -947,7 +947,7 @@ static void __cpuinit __print_cpu_msr(void)
 		index_max = msr_range_array[i].max;
 
 		for (index = index_min; index < index_max; index++) {
-			if (rdmsrl_amd_safe(index, &val))
+			if (rdmsrl_safe(index, &val))
 				continue;
 			printk(KERN_INFO " MSR%08x: %016llx\n", index, val);
 		}

arch/x86/kernel/cpu/perf_event.c

View file

diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
index c4706cf..29557aa 100644
--- a/arch/x86/kernel/cpu/perf_event.c
+++ b/arch/x86/kernel/cpu/perf_event.c
@@ -35,17 +35,6 @@
 
 #include "perf_event.h"
 
-#if 0
-#undef wrmsrl
-#define wrmsrl(msr, val) 					\
-do {								\
-	trace_printk("wrmsrl(%lx, %lx)\n", (unsigned long)(msr),\
-			(unsigned long)(val));			\
-	native_write_msr((msr), (u32)((u64)(val)), 		\
-			(u32)((u64)(val) >> 32));		\
-} while (0)
-#endif
-
 struct x86_pmu x86_pmu __read_mostly;
 
 DEFINE_PER_CPU(struct cpu_hw_events, cpu_hw_events) = {
@@ -74,7 +63,7 @@ u64 x86_perf_event_update(struct perf_event *event)
 	int idx = hwc->idx;
 	s64 delta;
 
-	if (idx == X86_PMC_IDX_FIXED_BTS)
+	if (idx == INTEL_PMC_IDX_FIXED_BTS)
 		return 0;
 
 	/*
@@ -86,7 +75,7 @@ u64 x86_perf_event_update(struct perf_event *event)
 	 */
 again:
 	prev_raw_count = local64_read(&hwc->prev_count);
-	rdmsrl(hwc->event_base, new_raw_count);
+	rdpmcl(hwc->event_base_rdpmc, new_raw_count);
 
 	if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
 					new_raw_count) != prev_raw_count)
@@ -189,7 +178,7 @@ static void release_pmc_hardware(void) {}
 
 static bool check_hw_exists(void)
 {
-	u64 val, val_new = 0;
+	u64 val, val_new = ~0;
 	int i, reg, ret = 0;
 
 	/*
@@ -222,8 +211,9 @@ static bool check_hw_exists(void)
 	 * that don't trap on the MSR access and always return 0s.
 	 */
 	val = 0xabcdUL;
-	ret = checking_wrmsrl(x86_pmu_event_addr(0), val);
-	ret |= rdmsrl_safe(x86_pmu_event_addr(0), &val_new);
+	reg = x86_pmu_event_addr(0);
+	ret = wrmsrl_safe(reg, val);
+	ret |= rdmsrl_safe(reg, &val_new);
 	if (ret || val != val_new)
 		goto msr_fail;
 
@@ -240,6 +230,7 @@ bios_fail:
 
 msr_fail:
 	printk(KERN_CONT "Broken PMU hardware detected, using software events only.\n");
+	printk(KERN_ERR "Failed to access perfctr msr (MSR %x is %Lx)\n", reg, val_new);
 
 	return false;
 }
@@ -388,7 +379,7 @@ int x86_pmu_hw_config(struct perf_event *event)
 		int precise = 0;
 
 		/* Support for constant skid */
-		if (x86_pmu.pebs_active) {
+		if (x86_pmu.pebs_active && !x86_pmu.pebs_broken) {
 			precise++;
 
 			/* Support for IP fixup */
@@ -637,8 +628,8 @@ static bool __perf_sched_find_counter(struct perf_sched *sched)
 	c = sched->constraints[sched->state.event];
 
 	/* Prefer fixed purpose counters */
-	if (x86_pmu.num_counters_fixed) {
-		idx = X86_PMC_IDX_FIXED;
+	if (c->idxmsk64 & (~0ULL << INTEL_PMC_IDX_FIXED)) {
+		idx = INTEL_PMC_IDX_FIXED;
 		for_each_set_bit_from(idx, c->idxmsk, X86_PMC_IDX_MAX) {
 			if (!__test_and_set_bit(idx, sched->state.used))
 				goto done;
@@ -646,7 +637,7 @@ static bool __perf_sched_find_counter(struct perf_sched *sched)
 	}
 	/* Grab the first unused counter starting with idx */
 	idx = sched->state.counter;
-	for_each_set_bit_from(idx, c->idxmsk, X86_PMC_IDX_FIXED) {
+	for_each_set_bit_from(idx, c->idxmsk, INTEL_PMC_IDX_FIXED) {
 		if (!__test_and_set_bit(idx, sched->state.used))
 			goto done;
 	}
@@ -704,8 +695,8 @@ static bool perf_sched_next_event(struct perf_sched *sched)
 /*
  * Assign a counter for each event.
  */
-static int perf_assign_events(struct event_constraint **constraints, int n,
-			      int wmin, int wmax, int *assign)
+int perf_assign_events(struct event_constraint **constraints, int n,
+			int wmin, int wmax, int *assign)
 {
 	struct perf_sched sched;
 
@@ -824,15 +815,17 @@ static inline void x86_assign_hw_event(struct perf_event *event,
 	hwc->last_cpu = smp_processor_id();
 	hwc->last_tag = ++cpuc->tags[i];
 
-	if (hwc->idx == X86_PMC_IDX_FIXED_BTS) {
+	if (hwc->idx == INTEL_PMC_IDX_FIXED_BTS) {
 		hwc->config_base = 0;
 		hwc->event_base	= 0;
-	} else if (hwc->idx >= X86_PMC_IDX_FIXED) {
+	} else if (hwc->idx >= INTEL_PMC_IDX_FIXED) {
 		hwc->config_base = MSR_ARCH_PERFMON_FIXED_CTR_CTRL;
-		hwc->event_base = MSR_ARCH_PERFMON_FIXED_CTR0 + (hwc->idx - X86_PMC_IDX_FIXED);
+		hwc->event_base = MSR_ARCH_PERFMON_FIXED_CTR0 + (hwc->idx - INTEL_PMC_IDX_FIXED);
+		hwc->event_base_rdpmc = (hwc->idx - INTEL_PMC_IDX_FIXED) | 1<<30;
 	} else {
 		hwc->config_base = x86_pmu_config_addr(hwc->idx);
 		hwc->event_base  = x86_pmu_event_addr(hwc->idx);
+		hwc->event_base_rdpmc = hwc->idx;
 	}
 }
 
@@ -930,7 +923,7 @@ int x86_perf_event_set_period(struct perf_event *event)
 	s64 period = hwc->sample_period;
 	int ret = 0, idx = hwc->idx;
 
-	if (idx == X86_PMC_IDX_FIXED_BTS)
+	if (idx == INTEL_PMC_IDX_FIXED_BTS)
 		return 0;
 
 	/*
@@ -1316,7 +1309,6 @@ static struct attribute_group x86_pmu_format_group = {
 static int __init init_hw_perf_events(void)
 {
 	struct x86_pmu_quirk *quirk;
-	struct event_constraint *c;
 	int err;
 
 	pr_info("Performance Events: ");
@@ -1347,21 +1339,8 @@ static int __init init_hw_perf_events(void)
 	for (quirk = x86_pmu.quirks; quirk; quirk = quirk->next)
 		quirk->func();
 
-	if (x86_pmu.num_counters > X86_PMC_MAX_GENERIC) {
-		WARN(1, KERN_ERR "hw perf events %d > max(%d), clipping!",
-		     x86_pmu.num_counters, X86_PMC_MAX_GENERIC);
-		x86_pmu.num_counters = X86_PMC_MAX_GENERIC;
-	}
-	x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1;
-
-	if (x86_pmu.num_counters_fixed > X86_PMC_MAX_FIXED) {
-		WARN(1, KERN_ERR "hw perf events fixed %d > max(%d), clipping!",
-		     x86_pmu.num_counters_fixed, X86_PMC_MAX_FIXED);
-		x86_pmu.num_counters_fixed = X86_PMC_MAX_FIXED;
-	}
-
-	x86_pmu.intel_ctrl |=
-		((1LL << x86_pmu.num_counters_fixed)-1) << X86_PMC_IDX_FIXED;
+	if (!x86_pmu.intel_ctrl)
+		x86_pmu.intel_ctrl = (1 << x86_pmu.num_counters) - 1;
 
 	perf_events_lapic_init();
 	register_nmi_handler(NMI_LOCAL, perf_event_nmi_handler, 0, "PMI");
@@ -1370,22 +1349,6 @@ static int __init init_hw_perf_events(void)
 		__EVENT_CONSTRAINT(0, (1ULL << x86_pmu.num_counters) - 1,
 				   0, x86_pmu.num_counters, 0);
 
-	if (x86_pmu.event_constraints) {
-		/*
-		 * event on fixed counter2 (REF_CYCLES) only works on this
-		 * counter, so do not extend mask to generic counters
-		 */
-		for_each_event_constraint(c, x86_pmu.event_constraints) {
-			if (c->cmask != X86_RAW_EVENT_MASK
-			    || c->idxmsk64 == X86_PMC_MSK_FIXED_REF_CYCLES) {
-				continue;
-			}
-
-			c->idxmsk64 |= (1ULL << x86_pmu.num_counters) - 1;
-			c->weight += x86_pmu.num_counters;
-		}
-	}
-
 	x86_pmu.attr_rdpmc = 1; /* enable userspace RDPMC usage by default */
 	x86_pmu_format_group.attrs = x86_pmu.format_attrs;
 
@@ -1620,8 +1583,8 @@ static int x86_pmu_event_idx(struct perf_event *event)
 	if (!x86_pmu.attr_rdpmc)
 		return 0;
 
-	if (x86_pmu.num_counters_fixed && idx >= X86_PMC_IDX_FIXED) {
-		idx -= X86_PMC_IDX_FIXED;
+	if (x86_pmu.num_counters_fixed && idx >= INTEL_PMC_IDX_FIXED) {
+		idx -= INTEL_PMC_IDX_FIXED;
 		idx |= 1 << 30;
 	}
 
@@ -1649,7 +1612,12 @@ static ssize_t set_attr_rdpmc(struct device *cdev,
 			      struct device_attribute *attr,
 			      const char *buf, size_t count)
 {
-	unsigned long val = simple_strtoul(buf, NULL, 0);
+	unsigned long val;
+	ssize_t ret;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret)
+		return ret;
 
 	if (!!val != !!x86_pmu.attr_rdpmc) {
 		x86_pmu.attr_rdpmc = !!val;
@@ -1682,13 +1650,20 @@ static void x86_pmu_flush_branch_stack(void)
 		x86_pmu.flush_branch_stack();
 }
 
+void perf_check_microcode(void)
+{
+	if (x86_pmu.check_microcode)
+		x86_pmu.check_microcode();
+}
+EXPORT_SYMBOL_GPL(perf_check_microcode);
+
 static struct pmu pmu = {
 	.pmu_enable		= x86_pmu_enable,
 	.pmu_disable		= x86_pmu_disable,
 