diff options
96 files changed, 5769 insertions, 4882 deletions
diff --git a/Documentation/kprobes.txt b/Documentation/kprobes.txt index 2f9115c0ae62..61c291cddf18 100644 --- a/Documentation/kprobes.txt +++ b/Documentation/kprobes.txt @@ -165,8 +165,8 @@ the user entry_handler invocation is also skipped. 1.4 How Does Jump Optimization Work? -If you configured your kernel with CONFIG_OPTPROBES=y (currently -this option is supported on x86/x86-64, non-preemptive kernel) and +If your kernel is built with CONFIG_OPTPROBES=y (currently this flag +is automatically set 'y' on x86/x86-64, non-preemptive kernel) and the "debug.kprobes_optimization" kernel parameter is set to 1 (see sysctl(8)), Kprobes tries to reduce probe-hit overhead by using a jump instruction instead of a breakpoint instruction at each probepoint. @@ -271,8 +271,6 @@ tweak the kernel's execution path, you need to suppress optimization, using one of the following techniques: - Specify an empty function for the kprobe's post_handler or break_handler. or -- Config CONFIG_OPTPROBES=n. - or - Execute 'sysctl -w debug.kprobes_optimization=n' 2. Architectures Supported @@ -307,10 +305,6 @@ it useful to "Compile the kernel with debug info" (CONFIG_DEBUG_INFO), so you can use "objdump -d -l vmlinux" to see the source-to-object code mapping. -If you want to reduce probing overhead, set "Kprobes jump optimization -support" (CONFIG_OPTPROBES) to "y". You can find this option under the -"Kprobes" line. - 4. API Reference The Kprobes API includes a "register" function and an "unregister" diff --git a/MAINTAINERS b/MAINTAINERS index 3d29fa389888..6c858e89c7d0 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4349,13 +4349,13 @@ M: Paul Mackerras <paulus@samba.org> M: Ingo Molnar <mingo@elte.hu> M: Arnaldo Carvalho de Melo <acme@redhat.com> S: Supported -F: kernel/perf_event.c +F: kernel/perf_event*.c F: include/linux/perf_event.h -F: arch/*/kernel/perf_event.c -F: arch/*/kernel/*/perf_event.c -F: arch/*/kernel/*/*/perf_event.c +F: arch/*/kernel/perf_event*.c +F: arch/*/kernel/*/perf_event*.c +F: arch/*/kernel/*/*/perf_event*.c F: arch/*/include/asm/perf_event.h -F: arch/*/lib/perf_event.c +F: arch/*/lib/perf_event*.c F: arch/*/kernel/perf_callchain.c F: tools/perf/ diff --git a/arch/Kconfig b/arch/Kconfig index e5eb1337a537..f06010fb4838 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -42,15 +42,10 @@ config KPROBES If in doubt, say "N". config OPTPROBES - bool "Kprobes jump optimization support (EXPERIMENTAL)" - default y - depends on KPROBES + def_bool y + depends on KPROBES && HAVE_OPTPROBES depends on !PREEMPT - depends on HAVE_OPTPROBES select KALLSYMS_ALL - help - This option will allow kprobes to optimize breakpoint to - a jump for reducing its overhead. config HAVE_EFFICIENT_UNALIGNED_ACCESS bool diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 0eacb1ffb421..7191b6eb16d8 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -58,6 +58,9 @@ config X86 select HAVE_ARCH_KMEMCHECK select HAVE_USER_RETURN_NOTIFIER +config INSTRUCTION_DECODER + def_bool (KPROBES || PERF_EVENTS) + config OUTPUT_FORMAT string default "elf32-i386" if X86_32 diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index a19829374e6a..918fbb1855cc 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -502,23 +502,3 @@ config CPU_SUP_UMC_32 CPU might render the kernel unbootable. If unsure, say N. - -config X86_DS - def_bool X86_PTRACE_BTS - depends on X86_DEBUGCTLMSR - select HAVE_HW_BRANCH_TRACER - -config X86_PTRACE_BTS - bool "Branch Trace Store" - default y - depends on X86_DEBUGCTLMSR - depends on BROKEN - ---help--- - This adds a ptrace interface to the hardware's branch trace store. - - Debuggers may use it to collect an execution trace of the debugged - application in order to answer the question 'how did I get here?'. - Debuggers may trace user mode as well as kernel mode. - - Say Y unless there is no application development on this machine - and you want to save a small amount of code size. diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index bc01e3ebfeb2..bd58c8abbfbd 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -174,15 +174,6 @@ config IOMMU_LEAK Add a simple leak tracer to the IOMMU code. This is useful when you are debugging a buggy device driver that leaks IOMMU mappings. -config X86_DS_SELFTEST - bool "DS selftest" - default y - depends on DEBUG_KERNEL - depends on X86_DS - ---help--- - Perform Debug Store selftests at boot time. - If in doubt, say "N". - config HAVE_MMIOTRACE_SUPPORT def_bool y diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index b4ac2cdcb64f..1fa03e04ae44 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -373,6 +373,7 @@ extern atomic_t init_deasserted; extern int wakeup_secondary_cpu_via_nmi(int apicid, unsigned long start_eip); #endif +#ifdef CONFIG_X86_LOCAL_APIC static inline u32 apic_read(u32 reg) { return apic->read(reg); @@ -403,10 +404,19 @@ static inline u32 safe_apic_wait_icr_idle(void) return apic->safe_wait_icr_idle(); } +#else /* CONFIG_X86_LOCAL_APIC */ + +static inline u32 apic_read(u32 reg) { return 0; } +static inline void apic_write(u32 reg, u32 val) { } +static inline u64 apic_icr_read(void) { return 0; } +static inline void apic_icr_write(u32 low, u32 high) { } +static inline void apic_wait_icr_idle(void) { } +static inline u32 safe_apic_wait_icr_idle(void) { return 0; } + +#endif /* CONFIG_X86_LOCAL_APIC */ static inline void ack_APIC_irq(void) { -#ifdef CONFIG_X86_LOCAL_APIC /* * ack_APIC_irq() actually gets compiled as a single instruction * ... yummie. @@ -414,7 +424,6 @@ static inline void ack_APIC_irq(void) /* Docs say use 0 for future compatibility */ apic_write(APIC_EOI, 0); -#endif } static inline unsigned default_get_apic_id(unsigned long x) diff --git a/arch/x86/include/asm/ds.h b/arch/x86/include/asm/ds.h deleted file mode 100644 index 70dac199b093..000000000000 --- a/arch/x86/include/asm/ds.h +++ /dev/null @@ -1,302 +0,0 @@ -/* - * Debug Store (DS) support - * - * This provides a low-level interface to the hardware's Debug Store - * feature that is used for branch trace store (BTS) and - * precise-event based sampling (PEBS). - * - * It manages: - * - DS and BTS hardware configuration - * - buffer overflow handling (to be done) - * - buffer access - * - * It does not do: - * - security checking (is the caller allowed to trace the task) - * - buffer allocation (memory accounting) - * - * - * Copyright (C) 2007-2009 Intel Corporation. - * Markus Metzger <markus.t.metzger@intel.com>, 2007-2009 - */ - -#ifndef _ASM_X86_DS_H -#define _ASM_X86_DS_H - - -#include <linux/types.h> -#include <linux/init.h> -#include <linux/err.h> - - -#ifdef CONFIG_X86_DS - -struct task_struct; -struct ds_context; -struct ds_tracer; -struct bts_tracer; -struct pebs_tracer; - -typedef void (*bts_ovfl_callback_t)(struct bts_tracer *); -typedef void (*pebs_ovfl_callback_t)(struct pebs_tracer *); - - -/* - * A list of features plus corresponding macros to talk about them in - * the ds_request function's flags parameter. - * - * We use the enum to index an array of corresponding control bits; - * we use the macro to index a flags bit-vector. - */ -enum ds_feature { - dsf_bts = 0, - dsf_bts_kernel, -#define BTS_KERNEL (1 << dsf_bts_kernel) - /* trace kernel-mode branches */ - - dsf_bts_user, -#define BTS_USER (1 << dsf_bts_user) - /* trace user-mode branches */ - - dsf_bts_overflow, - dsf_bts_max, - dsf_pebs = dsf_bts_max, - - dsf_pebs_max, - dsf_ctl_max = dsf_pebs_max, - dsf_bts_timestamps = dsf_ctl_max, -#define BTS_TIMESTAMPS (1 << dsf_bts_timestamps) - /* add timestamps into BTS trace */ - -#define BTS_USER_FLAGS (BTS_KERNEL | BTS_USER | BTS_TIMESTAMPS) -}; - - -/* - * Request BTS or PEBS - * - * Due to alignement constraints, the actual buffer may be slightly - * smaller than the requested or provided buffer. - * - * Returns a pointer to a tracer structure on success, or - * ERR_PTR(errcode) on failure. - * - * The interrupt threshold is independent from the overflow callback - * to allow users to use their own overflow interrupt handling mechanism. - * - * The function might sleep. - * - * task: the task to request recording for - * cpu: the cpu to request recording for - * base: the base pointer for the (non-pageable) buffer; - * size: the size of the provided buffer in bytes - * ovfl: pointer to a function to be called on buffer overflow; - * NULL if cyclic buffer requested - * th: the interrupt threshold in records from the end of the buffer; - * -1 if no interrupt threshold is requested. - * flags: a bit-mask of the above flags - */ -extern struct bts_tracer *ds_request_bts_task(struct task_struct *task, - void *base, size_t size, - bts_ovfl_callback_t ovfl, - size_t th, unsigned int flags); -extern struct bts_tracer *ds_request_bts_cpu(int cpu, void *base, size_t size, - bts_ovfl_callback_t ovfl, - size_t th, unsigned int flags); -extern struct pebs_tracer *ds_request_pebs_task(struct task_struct *task, - void *base, size_t size, - pebs_ovfl_callback_t ovfl, - size_t th, unsigned int flags); -extern struct pebs_tracer *ds_request_pebs_cpu(int cpu, - void *base, size_t size, - pebs_ovfl_callback_t ovfl, - size_t th, unsigned int flags); - -/* - * Release BTS or PEBS resources - * Suspend and resume BTS or PEBS tracing - * - * Must be called with irq's enabled. - * - * tracer: the tracer handle returned from ds_request_~() - */ -extern void ds_release_bts(struct bts_tracer *tracer); -extern void ds_suspend_bts(struct bts_tracer *tracer); -extern void ds_resume_bts(struct bts_tracer *tracer); -extern void ds_release_pebs(struct pebs_tracer *tracer); -extern void ds_suspend_pebs(struct pebs_tracer *tracer); -extern void ds_resume_pebs(struct pebs_tracer *tracer); - -/* - * Release BTS or PEBS resources - * Suspend and resume BTS or PEBS tracing - * - * Cpu tracers must call this on the traced cpu. - * Task tracers must call ds_release_~_noirq() for themselves. - * - * May be called with irq's disabled. - * - * Returns 0 if successful; - * -EPERM if the cpu tracer does not trace the current cpu. - * -EPERM if the task tracer does not trace itself. - * - * tracer: the tracer handle returned from ds_request_~() - */ -extern int ds_release_bts_noirq(struct bts_tracer *tracer); -extern int ds_suspend_bts_noirq(struct bts_tracer *tracer); -extern int ds_resume_bts_noirq(struct bts_tracer *tracer); -extern int ds_release_pebs_noirq(struct pebs_tracer *tracer); -extern int ds_suspend_pebs_noirq(struct pebs_tracer *tracer); -extern int ds_resume_pebs_noirq(struct pebs_tracer *tracer); - - -/* - * The raw DS buffer state as it is used for BTS and PEBS recording. - * - * This is the low-level, arch-dependent interface for working - * directly on the raw trace data. - */ -struct ds_trace { - /* the number of bts/pebs records */ - size_t n; - /* the size of a bts/pebs record in bytes */ - size_t size; - /* pointers into the raw buffer: - - to the first entry */ - void *begin; - /* - one beyond the last entry */ - void *end; - /* - one beyond the newest entry */ - void *top; - /* - the interrupt threshold */ - void *ith; - /* flags given on ds_request() */ - unsigned int flags; -}; - -/* - * An arch-independent view on branch trace data. - */ -enum bts_qualifier { - bts_invalid, -#define BTS_INVALID bts_invalid - - bts_branch, -#define BTS_BRANCH bts_branch - - bts_task_arrives, -#define BTS_TASK_ARRIVES bts_task_arrives - - bts_task_departs, -#define BTS_TASK_DEPARTS bts_task_departs - - bts_qual_bit_size = 4, - bts_qual_max = (1 << bts_qual_bit_size), -}; - -struct bts_struct { - __u64 qualifier; - union { - /* BTS_BRANCH */ - struct { - __u64 from; - __u64 to; - } lbr; - /* BTS_TASK_ARRIVES or BTS_TASK_DEPARTS */ - struct { - __u64 clock; - pid_t pid; - } event; - } variant; -}; - - -/* - * The BTS state. - * - * This gives access to the raw DS state and adds functions to provide - * an arch-independent view of the BTS data. - */ -struct bts_trace { - struct ds_trace ds; - - int (*read)(struct bts_tracer *tracer, const void *at, - struct bts_struct *out); - int (*write)(struct bts_tracer *tracer, const struct bts_struct *in); -}; - - -/* - * The PEBS state. - * - * This gives access to the raw DS state and the PEBS-specific counter - * reset value. - */ -struct pebs_trace { - struct ds_trace ds; - - /* the number of valid counters in the below array */ - unsigned int counters; - -#define MAX_PEBS_COUNTERS 4 - /* the counter reset value */ - unsigned long long counter_reset[MAX_PEBS_COUNTERS]; -}; - - -/* - * Read the BTS or PEBS trace. - * - * Returns a view on the trace collected for the parameter tracer. - * - * The view remains valid as long as the traced task is not running or - * the tracer is suspended. - * Writes into the trace buffer are not reflected. - * - * tracer: the tracer handle returned from ds_request_~() - */ -extern const struct bts_trace *ds_read_bts(struct bts_tracer *tracer); -extern const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer); - - -/* - * Reset the write pointer of the BTS/PEBS buffer. - * - * Returns 0 on success; -Eerrno on error - * - * tracer: the tracer handle returned from ds_request_~() - */ -extern int ds_reset_bts(struct bts_tracer *tracer); -extern int ds_reset_pebs(struct pebs_tracer *tracer); - -/* - * Set the PEBS counter reset value. - * - * Returns 0 on success; -Eerrno on error - * - * tracer: the tracer handle returned from ds_request_pebs() - * counter: the index of the counter - * value: the new counter reset value - */ -extern int ds_set_pebs_reset(struct pebs_tracer *tracer, - unsigned int counter, u64 value); - -/* - * Initialization - */ -struct cpuinfo_x86; -extern void __cpuinit ds_init_intel(struct cpuinfo_x86 *); - -/* - * Context switch work - */ -extern void ds_switch_to(struct task_struct *prev, struct task_struct *next); - -#else /* CONFIG_X86_DS */ - -struct cpuinfo_x86; -static inline void __cpuinit ds_init_intel(struct cpuinfo_x86 *ignored) {} -static inline void ds_switch_to(struct task_struct *prev, - struct task_struct *next) {} - -#endif /* CONFIG_X86_DS */ -#endif /* _ASM_X86_DS_H */ diff --git a/arch/x86/include/asm/insn.h b/arch/x86/include/asm/insn.h index 96c2e0ad04ca..88c765e16410 100644 --- a/arch/x86/include/asm/insn.h +++ b/arch/x86/include/asm/insn.h @@ -68,6 +68,8 @@ struct insn { const insn_byte_t *next_byte; }; +#define MAX_INSN_SIZE 16 + #define X86_MODRM_MOD(modrm) (((modrm) & 0xc0) >> 6) #define X86_MODRM_REG(modrm) (((modrm) & 0x38) >> 3) #define X86_MODRM_RM(modrm) ((modrm) & 0x07) diff --git a/arch/x86/include/asm/kprobes.h b/arch/x86/include/asm/kprobes.h index 4ffa345a8ccb..547882539157 100644 --- a/arch/x86/include/asm/kprobes.h +++ b/arch/x86/include/asm/kprobes.h @@ -24,6 +24,7 @@ #include <linux/types.h> #include <linux/ptrace.h> #include <linux/percpu.h> +#include <asm/insn.h> #define __ARCH_WANT_KPROBES_INSN_SLOT @@ -36,7 +37,6 @@ typedef u8 kprobe_opcode_t; #define RELATIVEJUMP_SIZE 5 #define RELATIVECALL_OPCODE 0xe8 #define RELATIVE_ADDR_SIZE 4 -#define MAX_INSN_SIZE 16 #define MAX_STACK_SIZE 64 #define MIN_STACK_SIZE(ADDR) \ (((MAX_STACK_SIZE) < (((unsigned long)current_thread_info()) + \ diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index 4604e6a54d36..bc473acfa7f9 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -71,11 +71,14 @@ #define MSR_IA32_LASTINTTOIP 0x000001de /* DEBUGCTLMSR bits (others vary by model): */ -#define _DEBUGCTLMSR_LBR 0 /* last branch recording */ -#define _DEBUGCTLMSR_BTF 1 /* single-step on branches */ - -#define DEBUGCTLMSR_LBR (1UL << _DEBUGCTLMSR_LBR) -#define DEBUGCTLMSR_BTF (1UL << _DEBUGCTLMSR_BTF) +#define DEBUGCTLMSR_LBR (1UL << 0) /* last branch recording */ +#define DEBUGCTLMSR_BTF (1UL << 1) /* single-step on branches */ +#define DEBUGCTLMSR_TR (1UL << 6) +#define DEBUGCTLMSR_BTS (1UL << 7) +#define DEBUGCTLMSR_BTINT (1UL << 8) +#define DEBUGCTLMSR_BTS_OFF_OS (1UL << 9) +#define DEBUGCTLMSR_BTS_OFF_USR (1UL << 10) +#define DEBUGCTLMSR_FREEZE_LBRS_ON_PMI (1UL << 11) #define MSR_IA32_MC0_CTL 0x00000400 #define MSR_IA32_MC0_STATUS 0x00000401 @@ -359,6 +362,8 @@ #define MSR_P4_U2L_ESCR0 0x000003b0 #define MSR_P4_U2L_ESCR1 0x000003b1 +#define MSR_P4_PEBS_MATRIX_VERT 0x000003f2 + /* Intel Core-based CPU performance counters */ #define MSR_CORE_PERF_FIXED_CTR0 0x00000309 #define MSR_CORE_PERF_FIXED_CTR1 0x0000030a diff --git a/arch/x86/include/asm/perf_event.h b/arch/x86/include/asm/perf_event.h index db6109a885a7..124dddd598f3 100644 --- a/arch/x86/include/asm/perf_event.h +++ b/arch/x86/include/asm/perf_event.h @@ -5,7 +5,7 @@ * Performance event hw details: */ -#define X86_PMC_MAX_GENERIC 8 +#define X86_PMC_MAX_GENERIC 32 #define X86_PMC_MAX_FIXED 3 #define X86_PMC_IDX_GENERIC 0 @@ -136,6 +136,25 @@ extern void perf_events_lapic_init(void); #define PERF_EVENT_INDEX_OFFSET 0 +/* + * Abuse bit 3 of the cpu eflags register to indicate proper PEBS IP fixups. + * This flag is otherwise unused and ABI specified to be 0, so nobody should + * care what we do with it. + */ +#define PERF_EFLAGS_EXACT (1UL << 3) + +#define perf_misc_flags(regs) \ +({ int misc = 0; \ + if (user_mode(regs)) \ + misc |= PERF_RECORD_MISC_USER; \ + else \ + misc |= PERF_RECORD_MISC_KERNEL; \ + if (regs->flags & PERF_EFLAGS_EXACT) \ + misc |= PERF_RECORD_MISC_EXACT; \ + misc; }) + +#define perf_instruction_pointer(regs) ((regs)->ip) + #else static inline void init_hw_perf_events(void) { } static inline void perf_events_lapic_init(void) { } diff --git a/arch/x86/include/asm/perf_event_p4.h b/arch/x86/include/asm/perf_event_p4.h new file mode 100644 index 000000000000..b05400a542ff --- /dev/null +++ b/arch/x86/include/asm/perf_event_p4.h @@ -0,0 +1,794 @@ +/* + * Netburst Perfomance Events (P4, old Xeon) + */ + +#ifndef PERF_EVENT_P4_H +#define PERF_EVENT_P4_H + +#include <linux/cpu.h> +#include <linux/bitops.h> + +/* + * NetBurst has perfomance MSRs shared between + * threads if HT is turned on, ie for both logical + * processors (mem: in turn in Atom with HT support + * perf-MSRs are not shared and every thread has its + * own perf-MSRs set) + */ +#define ARCH_P4_TOTAL_ESCR (46) +#define ARCH_P4_RESERVED_ESCR (2) /* IQ_ESCR(0,1) not always present */ +#define ARCH_P4_MAX_ESCR (ARCH_P4_TOTAL_ESCR - ARCH_P4_RESERVED_ESCR) +#define ARCH_P4_MAX_CCCR (18) +#define ARCH_P4_MAX_COUNTER (ARCH_P4_MAX_CCCR / 2) + +#define P4_ESCR_EVENT_MASK 0x7e000000U +#define P4_ESCR_EVENT_SHIFT 25 +#define P4_ESCR_EVENTMASK_MASK 0x01fffe00U +#define P4_ESCR_EVENTMASK_SHIFT 9 +#define P4_ESCR_TAG_MASK 0x000001e0U +#define P4_ESCR_TAG_SHIFT 5 +#define P4_ESCR_TAG_ENABLE 0x00000010U +#define P4_ESCR_T0_OS 0x00000008U +#define P4_ESCR_T0_USR 0x00000004U +#define P4_ESCR_T1_OS 0x00000002U +#define P4_ESCR_T1_USR 0x00000001U + +#define P4_ESCR_EVENT(v) ((v) << P4_ESCR_EVENT_SHIFT) +#define P4_ESCR_EMASK(v) ((v) << P4_ESCR_EVENTMASK_SHIFT) +#define P4_ESCR_TAG(v) ((v) << P4_ESCR_TAG_SHIFT) + +/* Non HT mask */ +#define P4_ESCR_MASK \ + (P4_ESCR_EVENT_MASK | \ + P4_ESCR_EVENTMASK_MASK | \ + P4_ESCR_TAG_MASK | \ + P4_ESCR_TAG_ENABLE | \ + P4_ESCR_T0_OS | \ + P4_ESCR_T0_USR) + +/* HT mask */ +#define P4_ESCR_MASK_HT \ + (P4_ESCR_MASK | P4_ESCR_T1_OS | P4_ESCR_T1_USR) + +#define P4_CCCR_OVF 0x80000000U +#define P4_CCCR_CASCADE 0x40000000U +#define P4_CCCR_OVF_PMI_T0 0x04000000U +#define P4_CCCR_OVF_PMI_T1 0x08000000U +#define P4_CCCR_FORCE_OVF 0x02000000U +#define P4_CCCR_EDGE 0x01000000U +#define P4_CCCR_THRESHOLD_MASK 0x00f00000U +#define P4_CCCR_THRESHOLD_SHIFT 20 +#define P4_CCCR_COMPLEMENT 0x00080000U +#define P4_CCCR_COMPARE 0x00040000U +#define P4_CCCR_ESCR_SELECT_MASK 0x0000e000U +#define P4_CCCR_ESCR_SELECT_SHIFT 13 +#define P4_CCCR_ENABLE 0x00001000U +#define P4_CCCR_THREAD_SINGLE 0x00010000U +#define P4_CCCR_THREAD_BOTH 0x00020000U +#define P4_CCCR_THREAD_ANY 0x00030000U +#define P4_CCCR_RESERVED 0x00000fffU + +#define P4_CCCR_THRESHOLD(v) ((v) << P4_CCCR_THRESHOLD_SHIFT) +#define P4_CCCR_ESEL(v) ((v) << P4_CCCR_ESCR_SELECT_SHIFT) + +/* Custom bits in reerved CCCR area */ +#define P4_CCCR_CACHE_OPS_MASK 0x0000003fU + + +/* Non HT mask */ +#define P4_CCCR_MASK \ + (P4_CCCR_OVF | \ + P4_CCCR_CASCADE | \ + P4_CCCR_OVF_PMI_T0 | \ + P4_CCCR_FORCE_OVF | \ + P4_CCCR_EDGE | \ + P4_CCCR_THRESHOLD_MASK | \ + P4_CCCR_COMPLEMENT | \ + P4_CCCR_COMPARE | \ + P4_CCCR_ESCR_SELECT_MASK | \ + P4_CCCR_ENABLE) + +/* HT mask */ +#define P4_CCCR_MASK_HT (P4_CCCR_MASK | P4_CCCR_THREAD_ANY) + +#define P4_GEN_ESCR_EMASK(class, name, bit) \ + class##__##name = ((1 << bit) << P4_ESCR_EVENTMASK_SHIFT) +#define P4_ESCR_EMASK_BIT(class, name) class##__##name + +/* + * config field is 64bit width and consists of + * HT << 63 | ESCR << 32 | CCCR + * where HT is HyperThreading bit (since ESCR + * has it reserved we may use it for own purpose) + * + * note that this is NOT the addresses of respective + * ESCR and CCCR but rather an only packed value should + * be unpacked and written to a proper addresses + * + * the base idea is to pack as much info as + * possible + */ +#define p4_config_pack_escr(v) (((u64)(v)) << 32) +#define p4_config_pack_cccr(v) (((u64)(v)) & 0xffffffffULL) +#define p4_config_unpack_escr(v) (((u64)(v)) >> 32) +#define p4_config_unpack_cccr(v) (((u64)(v)) & 0xffffffffULL) + +#define p4_config_unpack_emask(v) \ + ({ \ + u32 t = p4_config_unpack_escr((v)); \ + t = t & P4_ESCR_EVENTMASK_MASK; \ + t = t >> P4_ESCR_EVENTMASK_SHIFT; \ + t; \ + }) + +#define p4_config_unpack_event(v) \ + ({ \ + u32 t = p4_config_unpack_escr((v)); \ + t = t & P4_ESCR_EVENT_MASK; \ + t = t >> P4_ESCR_EVENT_SHIFT; \ + t; \ + }) + +#define p4_config_unpack_cache_event(v) (((u64)(v)) & P4_CCCR_CACHE_OPS_MASK) + +#define P4_CONFIG_HT_SHIFT 63 +#define P4_CONFIG_HT (1ULL << P4_CONFIG_HT_SHIFT) + +static inline bool p4_is_event_cascaded(u64 config) +{ + u32 cccr = p4_config_unpack_cccr(config); + return !!(cccr & P4_CCCR_CASCADE); +} + +static inline int p4_ht_config_thread(u64 config) +{ + return !!(config & P4_CONFIG_HT); +} + +static inline u64 p4_set_ht_bit(u64 config) +{ + return config | P4_CONFIG_HT; +} + +static inline u64 p4_clear_ht_bit(u64 config) +{ + return config & ~P4_CONFIG_HT; +} + +static inline int p4_ht_active(void) +{ +#ifdef CONFIG_SMP + return smp_num_siblings > 1; +#endif + return 0; +} + +static inline int p4_ht_thread(int cpu) +{ +#ifdef CONFIG_SMP + if (smp_num_siblings == 2) + return cpu != cpumask_first(__get_cpu_var(cpu_sibling_map)); +#endif + return 0; +} + +static inline int p4_should_swap_ts(u64 config, int cpu) +{ + return p4_ht_config_thread(config) ^ p4_ht_thread(cpu); +} + +static inline u32 p4_default_cccr_conf(int cpu) +{ + /* + * Note that P4_CCCR_THREAD_ANY is "required" on + * non-HT machines (on HT machines we count TS events + * regardless the state of second logical processor + */ + u32 cccr = P4_CCCR_THREAD_ANY; + + if (!p4_ht_thread(cpu)) + cccr |= P4_CCCR_OVF_PMI_T0; + else + cccr |= P4_CCCR_OVF_PMI_T1; + + return cccr; +} + +static inline u32 p4_default_escr_conf(int cpu, int exclude_os, int exclude_usr) +{ + u32 escr = 0; + + if (!p4_ht_thread(cpu)) { + if (!exclude_os) + escr |= P4_ESCR_T0_OS; + if (!exclude_usr) + escr |= P4_ESCR_T0_USR; + } else { + if (!exclude_os) + escr |= P4_ESCR_T1_OS; + if (!exclude_usr) + escr |= P4_ESCR_T1_USR; + } + + return escr; +} + +enum P4_EVENTS { + P4_EVENT_TC_DELIVER_MODE, + P4_EVENT_BPU_FETCH_REQUEST, + P4_EVENT_ITLB_REFERENCE, + P4_EVENT_MEMORY_CANCEL, + P4_EVENT_MEMORY_COMPLETE, + P4_EVENT_LOAD_PORT_REPLAY, + P4_EVENT_STORE_PORT_REPLAY, + P4_EVENT_MOB_LOAD_REPLAY, + P4_EVENT_PAGE_WALK_TYPE, + P4_EVENT_BSQ_CACHE_REFERENCE, + P4_EVENT_IOQ_ALLOCATION, + P4_EVENT_IOQ_ACTIVE_ENTRIES, + P4_EVENT_FSB_DATA_ACTIVITY, + P4_EVENT_BSQ_ALLOCATION, + P4_EVENT_BSQ_ACTIVE_ENTRIES, + P4_EVENT_SSE_INPUT_ASSIST, + P4_EVENT_PACKED_SP_UOP, + P4_EVENT_PACKED_DP_UOP, + P4_EVENT_SCALAR_SP_UOP, + P4_EVENT_SCALAR_DP_UOP, + P4_EVENT_64BIT_MMX_UOP, + P4_EVENT_128BIT_MMX_UOP, + P4_EVENT_X87_FP_UOP, + P4_EVENT_TC_MISC, + P4_EVENT_GLOBAL_POWER_EVENTS, + P4_EVENT_TC_MS_XFER, + P4_EVENT_UOP_QUEUE_WRITES, + P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, + P4_EVENT_RETIRED_BRANCH_TYPE, + P4_EVENT_RESOURCE_STALL, + P4_EVENT_WC_BUFFER, + P4_EVENT_B2B_CYCLES, + P4_EVENT_BNR, + P4_EVENT_SNOOP, + P4_EVENT_RESPONSE, + P4_EVENT_FRONT_END_EVENT, + P4_EVENT_EXECUTION_EVENT, + P4_EVENT_REPLAY_EVENT, + P4_EVENT_INSTR_RETIRED, + P4_EVENT_UOPS_RETIRED, + P4_EVENT_UOP_TYPE, + P4_EVENT_BRANCH_RETIRED, + P4_EVENT_MISPRED_BRANCH_RETIRED, + P4_EVENT_X87_ASSIST, + P4_EVENT_MACHINE_CLEAR, + P4_EVENT_INSTR_COMPLETED, +}; + +#define P4_OPCODE(event) event##_OPCODE +#define P4_OPCODE_ESEL(opcode) ((opcode & 0x00ff) >> 0) +#define P4_OPCODE_EVNT(opcode) ((opcode & 0xff00) >> 8) +#define P4_OPCODE_PACK(event, sel) (((event) << 8) | sel) + +/* + * Comments below the event represent ESCR restriction + * for this event and counter index per ESCR + * + * MSR_P4_IQ_ESCR0 and MSR_P4_IQ_ESCR1 are available only on early + * processor builds (family 0FH, models 01H-02H). These MSRs + * are not available on later versions, so that we don't use + * them completely + * + * Also note that CCCR1 do not have P4_CCCR_ENABLE bit properly + * working so that we should not use this CCCR and respective + * counter as result + */ +enum P4_EVENT_OPCODES { + P4_OPCODE(P4_EVENT_TC_DELIVER_MODE) = P4_OPCODE_PACK(0x01, 0x01), + /* + * MSR_P4_TC_ESCR0: 4, 5 + * MSR_P4_TC_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_BPU_FETCH_REQUEST) = P4_OPCODE_PACK(0x03, 0x00), + /* + * MSR_P4_BPU_ESCR0: 0, 1 + * MSR_P4_BPU_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_ITLB_REFERENCE) = P4_OPCODE_PACK(0x18, 0x03), + /* + * MSR_P4_ITLB_ESCR0: 0, 1 + * MSR_P4_ITLB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_MEMORY_CANCEL) = P4_OPCODE_PACK(0x02, 0x05), + /* + * MSR_P4_DAC_ESCR0: 8, 9 + * MSR_P4_DAC_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_MEMORY_COMPLETE) = P4_OPCODE_PACK(0x08, 0x02), + /* + * MSR_P4_SAAT_ESCR0: 8, 9 + * MSR_P4_SAAT_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_LOAD_PORT_REPLAY) = P4_OPCODE_PACK(0x04, 0x02), + /* + * MSR_P4_SAAT_ESCR0: 8, 9 + * MSR_P4_SAAT_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_STORE_PORT_REPLAY) = P4_OPCODE_PACK(0x05, 0x02), + /* + * MSR_P4_SAAT_ESCR0: 8, 9 + * MSR_P4_SAAT_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_MOB_LOAD_REPLAY) = P4_OPCODE_PACK(0x03, 0x02), + /* + * MSR_P4_MOB_ESCR0: 0, 1 + * MSR_P4_MOB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_PAGE_WALK_TYPE) = P4_OPCODE_PACK(0x01, 0x04), + /* + * MSR_P4_PMH_ESCR0: 0, 1 + * MSR_P4_PMH_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_BSQ_CACHE_REFERENCE) = P4_OPCODE_PACK(0x0c, 0x07), + /* + * MSR_P4_BSU_ESCR0: 0, 1 + * MSR_P4_BSU_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_IOQ_ALLOCATION) = P4_OPCODE_PACK(0x03, 0x06), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_IOQ_ACTIVE_ENTRIES) = P4_OPCODE_PACK(0x1a, 0x06), + /* + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_FSB_DATA_ACTIVITY) = P4_OPCODE_PACK(0x17, 0x06), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_BSQ_ALLOCATION) = P4_OPCODE_PACK(0x05, 0x07), + /* + * MSR_P4_BSU_ESCR0: 0, 1 + */ + + P4_OPCODE(P4_EVENT_BSQ_ACTIVE_ENTRIES) = P4_OPCODE_PACK(0x06, 0x07), + /* + * NOTE: no ESCR name in docs, it's guessed + * MSR_P4_BSU_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_SSE_INPUT_ASSIST) = P4_OPCODE_PACK(0x34, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_PACKED_SP_UOP) = P4_OPCODE_PACK(0x08, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_PACKED_DP_UOP) = P4_OPCODE_PACK(0x0c, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_SCALAR_SP_UOP) = P4_OPCODE_PACK(0x0a, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_SCALAR_DP_UOP) = P4_OPCODE_PACK(0x0e, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_64BIT_MMX_UOP) = P4_OPCODE_PACK(0x02, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_128BIT_MMX_UOP) = P4_OPCODE_PACK(0x1a, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_X87_FP_UOP) = P4_OPCODE_PACK(0x04, 0x01), + /* + * MSR_P4_FIRM_ESCR0: 8, 9 + * MSR_P4_FIRM_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_TC_MISC) = P4_OPCODE_PACK(0x06, 0x01), + /* + * MSR_P4_TC_ESCR0: 4, 5 + * MSR_P4_TC_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_GLOBAL_POWER_EVENTS) = P4_OPCODE_PACK(0x13, 0x06), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_TC_MS_XFER) = P4_OPCODE_PACK(0x05, 0x00), + /* + * MSR_P4_MS_ESCR0: 4, 5 + * MSR_P4_MS_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_UOP_QUEUE_WRITES) = P4_OPCODE_PACK(0x09, 0x00), + /* + * MSR_P4_MS_ESCR0: 4, 5 + * MSR_P4_MS_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE) = P4_OPCODE_PACK(0x05, 0x02), + /* + * MSR_P4_TBPU_ESCR0: 4, 5 + * MSR_P4_TBPU_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_RETIRED_BRANCH_TYPE) = P4_OPCODE_PACK(0x04, 0x02), + /* + * MSR_P4_TBPU_ESCR0: 4, 5 + * MSR_P4_TBPU_ESCR1: 6, 7 + */ + + P4_OPCODE(P4_EVENT_RESOURCE_STALL) = P4_OPCODE_PACK(0x01, 0x01), + /* + * MSR_P4_ALF_ESCR0: 12, 13, 16 + * MSR_P4_ALF_ESCR1: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_WC_BUFFER) = P4_OPCODE_PACK(0x05, 0x05), + /* + * MSR_P4_DAC_ESCR0: 8, 9 + * MSR_P4_DAC_ESCR1: 10, 11 + */ + + P4_OPCODE(P4_EVENT_B2B_CYCLES) = P4_OPCODE_PACK(0x16, 0x03), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_BNR) = P4_OPCODE_PACK(0x08, 0x03), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_SNOOP) = P4_OPCODE_PACK(0x06, 0x03), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_RESPONSE) = P4_OPCODE_PACK(0x04, 0x03), + /* + * MSR_P4_FSB_ESCR0: 0, 1 + * MSR_P4_FSB_ESCR1: 2, 3 + */ + + P4_OPCODE(P4_EVENT_FRONT_END_EVENT) = P4_OPCODE_PACK(0x08, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_EXECUTION_EVENT) = P4_OPCODE_PACK(0x0c, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_REPLAY_EVENT) = P4_OPCODE_PACK(0x09, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_INSTR_RETIRED) = P4_OPCODE_PACK(0x02, 0x04), + /* + * MSR_P4_CRU_ESCR0: 12, 13, 16 + * MSR_P4_CRU_ESCR1: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_UOPS_RETIRED) = P4_OPCODE_PACK(0x01, 0x04), + /* + * MSR_P4_CRU_ESCR0: 12, 13, 16 + * MSR_P4_CRU_ESCR1: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_UOP_TYPE) = P4_OPCODE_PACK(0x02, 0x02), + /* + * MSR_P4_RAT_ESCR0: 12, 13, 16 + * MSR_P4_RAT_ESCR1: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_BRANCH_RETIRED) = P4_OPCODE_PACK(0x06, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_MISPRED_BRANCH_RETIRED) = P4_OPCODE_PACK(0x03, 0x04), + /* + * MSR_P4_CRU_ESCR0: 12, 13, 16 + * MSR_P4_CRU_ESCR1: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_X87_ASSIST) = P4_OPCODE_PACK(0x03, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_MACHINE_CLEAR) = P4_OPCODE_PACK(0x02, 0x05), + /* + * MSR_P4_CRU_ESCR2: 12, 13, 16 + * MSR_P4_CRU_ESCR3: 14, 15, 17 + */ + + P4_OPCODE(P4_EVENT_INSTR_COMPLETED) = P4_OPCODE_PACK(0x07, 0x04), + /* + * MSR_P4_CRU_ESCR0: 12, 13, 16 + * MSR_P4_CRU_ESCR1: 14, 15, 17 + */ +}; + +/* + * a caller should use P4_ESCR_EMASK_NAME helper to + * pick the EventMask needed, for example + * + * P4_ESCR_EMASK_NAME(P4_EVENT_TC_DELIVER_MODE, DD) + */ +enum P4_ESCR_EMASKS { + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, DD, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, DB, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, DI, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, BD, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, BB, 4), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, BI, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_TC_DELIVER_MODE, ID, 6), + + P4_GEN_ESCR_EMASK(P4_EVENT_BPU_FETCH_REQUEST, TCMISS, 0), + + P4_GEN_ESCR_EMASK(P4_EVENT_ITLB_REFERENCE, HIT, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_ITLB_REFERENCE, MISS, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_ITLB_REFERENCE, HIT_UK, 2), + + P4_GEN_ESCR_EMASK(P4_EVENT_MEMORY_CANCEL, ST_RB_FULL, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_MEMORY_CANCEL, 64K_CONF, 3), + + P4_GEN_ESCR_EMASK(P4_EVENT_MEMORY_COMPLETE, LSC, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_MEMORY_COMPLETE, SSC, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_LOAD_PORT_REPLAY, SPLIT_LD, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_STORE_PORT_REPLAY, SPLIT_ST, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_MOB_LOAD_REPLAY, NO_STA, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_MOB_LOAD_REPLAY, NO_STD, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_MOB_LOAD_REPLAY, PARTIAL_DATA, 4), + P4_GEN_ESCR_EMASK(P4_EVENT_MOB_LOAD_REPLAY, UNALGN_ADDR, 5), + + P4_GEN_ESCR_EMASK(P4_EVENT_PAGE_WALK_TYPE, DTMISS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_PAGE_WALK_TYPE, ITMISS, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITE, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITM, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITS, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITE, 4), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITM, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_MISS, 8), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_MISS, 9), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_CACHE_REFERENCE, WR_2ndL_MISS, 10), + + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, DEFAULT, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, ALL_READ, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, ALL_WRITE, 6), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, MEM_UC, 7), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, MEM_WC, 8), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, MEM_WT, 9), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, MEM_WP, 10), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, MEM_WB, 11), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, OWN, 13), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, OTHER, 14), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ALLOCATION, PREFETCH, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, DEFAULT, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, ALL_READ, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, ALL_WRITE, 6), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_UC, 7), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WC, 8), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WT, 9), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WP, 10), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, MEM_WB, 11), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, OWN, 13), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, OTHER, 14), + P4_GEN_ESCR_EMASK(P4_EVENT_IOQ_ACTIVE_ENTRIES, PREFETCH, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_DRV, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OWN, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OTHER, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_DRV, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_OWN, 4), + P4_GEN_ESCR_EMASK(P4_EVENT_FSB_DATA_ACTIVITY, DBSY_OTHER, 5), + + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_TYPE0, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_TYPE1, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_LEN0, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_LEN1, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_IO_TYPE, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_LOCK_TYPE, 6), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_CACHE_TYPE, 7), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_SPLIT_TYPE, 8), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_DEM_TYPE, 9), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, REQ_ORD_TYPE, 10), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE0, 11), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE1, 12), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ALLOCATION, MEM_TYPE2, 13), + + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_TYPE0, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_TYPE1, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LEN0, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LEN1, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_IO_TYPE, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_LOCK_TYPE, 6), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_CACHE_TYPE, 7), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_SPLIT_TYPE, 8), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_DEM_TYPE, 9), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, REQ_ORD_TYPE, 10), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE0, 11), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE1, 12), + P4_GEN_ESCR_EMASK(P4_EVENT_BSQ_ACTIVE_ENTRIES, MEM_TYPE2, 13), + + P4_GEN_ESCR_EMASK(P4_EVENT_SSE_INPUT_ASSIST, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_PACKED_SP_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_PACKED_DP_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_SCALAR_SP_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_SCALAR_DP_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_64BIT_MMX_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_128BIT_MMX_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_X87_FP_UOP, ALL, 15), + + P4_GEN_ESCR_EMASK(P4_EVENT_TC_MISC, FLUSH, 4), + + P4_GEN_ESCR_EMASK(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING, 0), + + P4_GEN_ESCR_EMASK(P4_EVENT_TC_MS_XFER, CISC, 0), + + P4_GEN_ESCR_EMASK(P4_EVENT_UOP_QUEUE_WRITES, FROM_TC_BUILD, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_UOP_QUEUE_WRITES, FROM_TC_DELIVER, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_UOP_QUEUE_WRITES, FROM_ROM, 2), + + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, CONDITIONAL, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, CALL, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, RETURN, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE, INDIRECT, 4), + + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_BRANCH_TYPE, CONDITIONAL, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_BRANCH_TYPE, CALL, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_BRANCH_TYPE, RETURN, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_RETIRED_BRANCH_TYPE, INDIRECT, 4), + + P4_GEN_ESCR_EMASK(P4_EVENT_RESOURCE_STALL, SBFULL, 5), + + P4_GEN_ESCR_EMASK(P4_EVENT_WC_BUFFER, WCB_EVICTS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_WC_BUFFER, WCB_FULL_EVICTS, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_FRONT_END_EVENT, NBOGUS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_FRONT_END_EVENT, BOGUS, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, NBOGUS0, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, NBOGUS1, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, NBOGUS2, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, NBOGUS3, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, BOGUS0, 4), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, BOGUS1, 5), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, BOGUS2, 6), + P4_GEN_ESCR_EMASK(P4_EVENT_EXECUTION_EVENT, BOGUS3, 7), + + P4_GEN_ESCR_EMASK(P4_EVENT_REPLAY_EVENT, NBOGUS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_REPLAY_EVENT, BOGUS, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_RETIRED, NBOGUSNTAG, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_RETIRED, NBOGUSTAG, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_RETIRED, BOGUSNTAG, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_RETIRED, BOGUSTAG, 3), + + P4_GEN_ESCR_EMASK(P4_EVENT_UOPS_RETIRED, NBOGUS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_UOPS_RETIRED, BOGUS, 1), + + P4_GEN_ESCR_EMASK(P4_EVENT_UOP_TYPE, TAGLOADS, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_UOP_TYPE, TAGSTORES, 2), + + P4_GEN_ESCR_EMASK(P4_EVENT_BRANCH_RETIRED, MMNP, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_BRANCH_RETIRED, MMNM, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_BRANCH_RETIRED, MMTP, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_BRANCH_RETIRED, MMTM, 3), + + P4_GEN_ESCR_EMASK(P4_EVENT_MISPRED_BRANCH_RETIRED, NBOGUS, 0), + + P4_GEN_ESCR_EMASK(P4_EVENT_X87_ASSIST, FPSU, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_X87_ASSIST, FPSO, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_X87_ASSIST, POAO, 2), + P4_GEN_ESCR_EMASK(P4_EVENT_X87_ASSIST, POAU, 3), + P4_GEN_ESCR_EMASK(P4_EVENT_X87_ASSIST, PREA, 4), + + P4_GEN_ESCR_EMASK(P4_EVENT_MACHINE_CLEAR, CLEAR, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_MACHINE_CLEAR, MOCLEAR, 1), + P4_GEN_ESCR_EMASK(P4_EVENT_MACHINE_CLEAR, SMCLEAR, 2), + + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_COMPLETED, NBOGUS, 0), + P4_GEN_ESCR_EMASK(P4_EVENT_INSTR_COMPLETED, BOGUS, 1), +}; + +/* P4 PEBS: stale for a while */ +#define P4_PEBS_METRIC_MASK 0x00001fffU +#define P4_PEBS_UOB_TAG 0x01000000U +#define P4_PEBS_ENABLE 0x02000000U + +/* Replay metrics for MSR_IA32_PEBS_ENABLE and MSR_P4_PEBS_MATRIX_VERT */ +#define P4_PEBS__1stl_cache_load_miss_retired 0x3000001 +#define P4_PEBS__2ndl_cache_load_miss_retired 0x3000002 +#define P4_PEBS__dtlb_load_miss_retired 0x3000004 +#define P4_PEBS__dtlb_store_miss_retired 0x3000004 +#define P4_PEBS__dtlb_all_miss_retired 0x3000004 +#define P4_PEBS__tagged_mispred_branch 0x3018000 +#define P4_PEBS__mob_load_replay_retired 0x3000200 +#define P4_PEBS__split_load_retired 0x3000400 +#define P4_PEBS__split_store_retired 0x3000400 + +#define P4_VERT__1stl_cache_load_miss_retired 0x0000001 +#define P4_VERT__2ndl_cache_load_miss_retired 0x0000001 +#define P4_VERT__dtlb_load_miss_retired 0x0000001 +#define P4_VERT__dtlb_store_miss_retired 0x0000002 +#define P4_VERT__dtlb_all_miss_retired 0x0000003 +#define P4_VERT__tagged_mispred_branch 0x0000010 +#define P4_VERT__mob_load_replay_retired 0x0000001 +#define P4_VERT__split_load_retired 0x0000001 +#define P4_VERT__split_store_retired 0x0000002 + +enum P4_CACHE_EVENTS { + P4_CACHE__NONE, + + P4_CACHE__1stl_cache_load_miss_retired, + P4_CACHE__2ndl_cache_load_miss_retired, + P4_CACHE__dtlb_load_miss_retired, + P4_CACHE__dtlb_store_miss_retired, + P4_CACHE__itlb_reference_hit, + P4_CACHE__itlb_reference_miss, + + P4_CACHE__MAX +}; + +#endif /* PERF_EVENT_P4_H */ diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h index b753ea59703a..32428b410b55 100644 --- a/arch/x86/include/asm/processor.h +++ b/arch/x86/include/asm/processor.h @@ -21,7 +21,6 @@ struct mm_struct; #include <asm/msr.h> #include <asm/desc_defs.h> #include <asm/nops.h> -#include <asm/ds.h> #include <linux/personality.h> #include <linux/cpumask.h> @@ -29,6 +28,7 @@ struct mm_struct; #include <linux/threads.h> #include <linux/math64.h> #include <linux/init.h> +#include <linux/err.h> #define HBP_NUM 4 /* @@ -473,10 +473,6 @@ struct thread_struct { unsigned long iopl; /* Max allowed port in the bitmap, in bytes: */ unsigned io_bitmap_max; -/* MSR_IA32_DEBUGCTLMSR value to switch in if TIF_DEBUGCTLMSR is set. */ - unsigned long debugctlmsr; - /* Debug Store context; see asm/ds.h */ - struct ds_context *ds_ctx; }; static inline unsigned long native_get_debugreg(int regno) @@ -803,7 +799,7 @@ extern void cpu_init(void); static inline unsigned long get_debugctlmsr(void) { - unsigned long debugctlmsr = 0; + unsigned long debugctlmsr = 0; #ifndef CONFIG_X86_DEBUGCTLMSR if (boot_cpu_data.x86 < 6) @@ -811,21 +807,6 @@ static inline unsigned long get_debugctlmsr(void) #endif rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctlmsr); - return debugctlmsr; -} - -static inline unsigned long get_debugctlmsr_on_cpu(int cpu) -{ - u64 debugctlmsr = 0; - u32 val1, val2; - -#ifndef CONFIG_X86_DEBUGCTLMSR - if (boot_cpu_data.x86 < 6) - return 0; -#endif - rdmsr_on_cpu(cpu, MSR_IA32_DEBUGCTLMSR, &val1, &val2); - debugctlmsr = val1 | ((u64)val2 << 32); - return debugctlmsr; } @@ -838,18 +819,6 @@ static inline void update_debugctlmsr(unsigned long debugctlmsr) wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctlmsr); } -static inline void update_debugctlmsr_on_cpu(int cpu, - unsigned long debugctlmsr) -{ -#ifndef CONFIG_X86_DEBUGCTLMSR - if (boot_cpu_data.x86 < 6) - return; -#endif - wrmsr_on_cpu(cpu, MSR_IA32_DEBUGCTLMSR, - (u32)((u64)debugctlmsr), - (u32)((u64)debugctlmsr >> 32)); -} - /* * from system description table in BIOS. Mostly for MCA use, but * others may find it useful: diff --git a/arch/x86/include/asm/ptrace-abi.h b/arch/x86/include/asm/ptrace-abi.h index 86723035a515..52b098a6eebb 100644 --- a/arch/x86/include/asm/ptrace-abi.h +++ b/arch/x86/include/asm/ptrace-abi.h @@ -82,61 +82,6 @@ #ifndef __ASSEMBLY__ #include <linux/types.h> - -/* configuration/status structure used in PTRACE_BTS_CONFIG and - PTRACE_BTS_STATUS commands. -*/ -struct ptrace_bts_config { - /* requested or actual size of BTS buffer in bytes */ - __u32 size; - /* bitmask of below flags */ - __u32 flags; - /* buffer overflow signal */ - __u32 signal; - /* actual size of bts_struct in bytes */ - __u32 bts_size; -}; -#endif /* __ASSEMBLY__ */ - -#define PTRACE_BTS_O_TRACE 0x1 /* branch trace */ -#define PTRACE_BTS_O_SCHED 0x2 /* scheduling events w/ jiffies */ -#define PTRACE_BTS_O_SIGNAL 0x4 /* send SIG<signal> on buffer overflow - instead of wrapping around */ -#define PTRACE_BTS_O_ALLOC 0x8 /* (re)allocate buffer */ - -#define PTRACE_BTS_CONFIG 40 -/* Configure branch trace recording. - ADDR points to a struct ptrace_bts_config. - DATA gives the size of that buffer. - A new buffer is allocated, if requested in the flags. - An overflow signal may only be requested for new buffers. - Returns the number of bytes read. -*/ -#define PTRACE_BTS_STATUS 41 -/* Return the current configuration in a struct ptrace_bts_config - pointed to by ADDR; DATA gives the size of that buffer. - Returns the number of bytes written. -*/ -#define PTRACE_BTS_SIZE 42 -/* Return the number of available BTS records for draining. - DATA and ADDR are ignored. -*/ -#define PTRACE_BTS_GET 43 -/* Get a single BTS record. - DATA defines the index into the BTS array, where 0 is the newest - entry, and higher indices refer to older entries. - ADDR is pointing to struct bts_struct (see asm/ds.h). -*/ -#define PTRACE_BTS_CLEAR 44 -/* Clear the BTS buffer. - DATA and ADDR are ignored. -*/ -#define PTRACE_BTS_DRAIN 45 -/* Read all available BTS records and clear the buffer. - ADDR points to an array of struct bts_struct. - DATA gives the size of that buffer. - BTS records are read from oldest to newest. - Returns number of BTS records drained. -*/ +#endif #endif /* _ASM_X86_PTRACE_ABI_H */ diff --git a/arch/x86/include/asm/ptrace.h b/arch/x86/include/asm/ptrace.h index 69a686a7dff0..78cd1ea94500 100644 --- a/arch/x86/include/asm/ptrace.h +++ b/arch/x86/include/asm/ptrace.h @@ -289,12 +289,6 @@ extern int do_get_thread_area(struct task_struct *p, int idx, extern int do_set_thread_area(struct task_struct *p, int idx, struct user_desc __user *info, int can_allocate); -#ifdef CONFIG_X86_PTRACE_BTS -extern void ptrace_bts_untrace(struct task_struct *tsk); - -#define arch_ptrace_untrace(tsk) ptrace_bts_untrace(tsk) -#endif /* CONFIG_X86_PTRACE_BTS */ - #endif /* __KERNEL__ */ #endif /* !__ASSEMBLY__ */ diff --git a/arch/x86/include/asm/thread_info.h b/arch/x86/include/asm/thread_info.h index e0d28901e969..d017ed5502e2 100644 --- a/arch/x86/include/asm/thread_info.h +++ b/arch/x86/include/asm/thread_info.h @@ -92,8 +92,7 @@ struct thread_info { #define TIF_IO_BITMAP 22 /* uses I/O bitmap */ #define TIF_FREEZE 23 /* is freezing for suspend */ #define TIF_FORCED_TF 24 /* true if TF in eflags artificially */ -#define TIF_DEBUGCTLMSR 25 /* uses thread_struct.debugctlmsr */ -#define TIF_DS_AREA_MSR 26 /* uses thread_struct.ds_area_msr */ +#define TIF_BLOCKSTEP 25 /* set when we want DEBUGCTLMSR_BTF */ #define TIF_LAZY_MMU_UPDATES 27 /* task is updating the mmu lazily */ #define TIF_SYSCALL_TRACEPOINT 28 /* syscall tracepoint instrumentation */ @@ -115,8 +114,7 @@ struct thread_info { #define _TIF_IO_BITMAP (1 << TIF_IO_BITMAP) #define _TIF_FREEZE (1 << TIF_FREEZE) #define _TIF_FORCED_TF (1 << TIF_FORCED_TF) -#define _TIF_DEBUGCTLMSR (1 << TIF_DEBUGCTLMSR) -#define _TIF_DS_AREA_MSR (1 << TIF_DS_AREA_MSR) +#define _TIF_BLOCKSTEP (1 << TIF_BLOCKSTEP) #define _TIF_LAZY_MMU_UPDATES (1 << TIF_LAZY_MMU_UPDATES) #define _TIF_SYSCALL_TRACEPOINT (1 << TIF_SYSCALL_TRACEPOINT) @@ -147,7 +145,7 @@ struct thread_info { /* flags to check in __switch_to() */ #define _TIF_WORK_CTXSW \ - (_TIF_IO_BITMAP|_TIF_DEBUGCTLMSR|_TIF_DS_AREA_MSR|_TIF_NOTSC) + (_TIF_IO_BITMAP|_TIF_NOTSC|_TIF_BLOCKSTEP) #define _TIF_WORK_CTXSW_PREV (_TIF_WORK_CTXSW|_TIF_USER_RETURN_NOTIFY) #define _TIF_WORK_CTXSW_NEXT (_TIF_WORK_CTXSW|_TIF_DEBUG) diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 4c58352209e0..e77b22083721 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -47,8 +47,6 @@ obj-$(CONFIG_X86_TRAMPOLINE) += trampoline.o obj-y += process.o obj-y += i387.o xsave.o obj-y += ptrace.o -obj-$(CONFIG_X86_DS) += ds.o -obj-$(CONFIG_X86_DS_SELFTEST) += ds_selftest.o obj-$(CONFIG_X86_32) += tls.o obj-$(CONFIG_IA32_EMULATION) += tls.o obj-y += step.o diff --git a/arch/x86/kernel/cpu/intel.c b/arch/x86/kernel/cpu/intel.c index 7e1cca13af35..d72377c41c76 100644 --- a/arch/x86/kernel/cpu/intel.c +++ b/arch/x86/kernel/cpu/intel.c @@ -12,7 +12,6 @@ #include <asm/processor.h> #include <asm/pgtable.h> #include <asm/msr.h> -#include <asm/ds.h> #include <asm/bugs.h> #include <asm/cpu.h> @@ -367,7 +366,6 @@ static void __cpuinit init_intel(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_BTS); if (!(l1 & (1<<12))) set_cpu_cap(c, X86_FEATURE_PEBS); - ds_init_intel(c); } if (c->x86 == 6 && c->x86_model == 29 && cpu_has_clflush) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index 53ea4cf1a878..b53435661813 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -30,46 +30,51 @@ #include <asm/nmi.h> #include <asm/compat.h> -static u64 perf_event_mask __read_mostly; +#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 -/* The maximal number of PEBS events: */ -#define MAX_PEBS_EVENTS 4 +/* + * best effort, GUP based copy_from_user() that assumes IRQ or NMI context + */ +static unsigned long +copy_from_user_nmi(void *to, const void __user *from, unsigned long n) +{ + unsigned long offset, addr = (unsigned long)from; + int type = in_nmi() ? KM_NMI : KM_IRQ0; + unsigned long size, len = 0; + struct page *page; + void *map; + int ret; -/* The size of a BTS record in bytes: */ -#define BTS_RECORD_SIZE 24 + do { + ret = __get_user_pages_fast(addr, 1, 0, &page); + if (!ret) + break; -/* The size of a per-cpu BTS buffer in bytes: */ -#define BTS_BUFFER_SIZE (BTS_RECORD_SIZE * 2048) + offset = addr & (PAGE_SIZE - 1); + size = min(PAGE_SIZE - offset, n - len); -/* The BTS overflow threshold in bytes from the end of the buffer: */ -#define BTS_OVFL_TH (BTS_RECORD_SIZE * 128) + map = kmap_atomic(page, type); + memcpy(to, map+offset, size); + kunmap_atomic(map, type); + put_page(page); + len += size; + to += size; + addr += size; -/* - * Bits in the debugctlmsr controlling branch tracing. - */ -#define X86_DEBUGCTL_TR (1 << 6) -#define X86_DEBUGCTL_BTS (1 << 7) -#define X86_DEBUGCTL_BTINT (1 << 8) -#define X86_DEBUGCTL_BTS_OFF_OS (1 << 9) -#define X86_DEBUGCTL_BTS_OFF_USR (1 << 10) + } while (len < n); -/* - * A debug store configuration. - * - * We only support architectures that use 64bit fields. - */ -struct debug_store { - u64 bts_buffer_base; - u64 bts_index; - u64 bts_absolute_maximum; - u64 bts_interrupt_threshold; - u64 pebs_buffer_base; - u64 pebs_index; - u64 pebs_absolute_maximum; - u64 pebs_interrupt_threshold; - u64 pebs_event_reset[MAX_PEBS_EVENTS]; -}; + return len; +} struct event_constraint { union { @@ -88,18 +93,39 @@ struct amd_nb { struct event_constraint event_constraints[X86_PMC_IDX_MAX]; }; +#define MAX_LBR_ENTRIES 16 + struct cpu_hw_events { + /* + * Generic x86 PMC bits + */ struct perf_event *events[X86_PMC_IDX_MAX]; /* in counter order */ unsigned long active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; - unsigned long interrupts; int enabled; - struct debug_store *ds; int n_events; int n_added; int assign[X86_PMC_IDX_MAX]; /* event to counter assignment */ u64 tags[X86_PMC_IDX_MAX]; struct perf_event *event_list[X86_PMC_IDX_MAX]; /* in enabled order */ + + /* + * Intel DebugStore bits + */ + struct debug_store *ds; + u64 pebs_enabled; + + /* + * Intel LBR bits + */ + int lbr_users; + void *lbr_context; + struct perf_branch_stack lbr_stack; + struct perf_branch_entry lbr_entries[MAX_LBR_ENTRIES]; + + /* + * AMD specific bits + */ struct amd_nb *amd_nb; }; @@ -113,29 +139,57 @@ struct cpu_hw_events { #define EVENT_CONSTRAINT(c, n, m) \ __EVENT_CONSTRAINT(c, n, m, HWEIGHT(n)) +/* + * Constraint on the Event code. + */ #define INTEL_EVENT_CONSTRAINT(c, n) \ EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVTSEL_MASK) +/* + * Constraint on the Event code + UMask + fixed-mask + */ #define FIXED_EVENT_CONSTRAINT(c, n) \ EVENT_CONSTRAINT(c, (1ULL << (32+n)), INTEL_ARCH_FIXED_MASK) +/* + * Constraint on the Event code + UMask + */ +#define PEBS_EVENT_CONSTRAINT(c, n) \ + EVENT_CONSTRAINT(c, n, INTEL_ARCH_EVENT_MASK) + #define EVENT_CONSTRAINT_END \ EVENT_CONSTRAINT(0, 0, 0) #define for_each_event_constraint(e, c) \ for ((e) = (c); (e)->cmask; (e)++) +union perf_capabilities { + struct { + u64 lbr_format : 6; + u64 pebs_trap : 1; + u64 pebs_arch_reg : 1; + u64 pebs_format : 4; + u64 smm_freeze : 1; + }; + u64 capabilities; +}; + /* * struct x86_pmu - generic x86 pmu */ struct x86_pmu { + /* + * Generic x86 PMC bits + */ const char *name; int version; int (*handle_irq)(struct pt_regs *); void (*disable_all)(void); - void (*enable_all)(void); + void (*enable_all)(int added); void (*enable)(struct perf_event *); void (*disable)(struct perf_event *); + int (*hw_config)(struct perf_event_attr *attr, struct hw_perf_event *hwc); + int (*schedule_events)(struct cpu_hw_events *cpuc, int n, int *assign); unsigned eventsel; unsigned perfctr; u64 (*event_map)(int); @@ -147,10 +201,6 @@ struct x86_pmu { u64 event_mask; int apic; u64 max_period; - u64 intel_ctrl; - void (*enable_bts)(u64 config); - void (*disable_bts)(void); - struct event_constraint * (*get_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); @@ -158,11 +208,32 @@ struct x86_pmu { void (*put_event_constraints)(struct cpu_hw_events *cpuc, struct perf_event *event); struct event_constraint *event_constraints; + void (*quirks)(void); int (*cpu_prepare)(int cpu); void (*cpu_starting)(int cpu); void (*cpu_dying)(int cpu); void (*cpu_dead)(int cpu); + + /* + * Intel Arch Perfmon v2+ + */ + u64 intel_ctrl; + union perf_capabilities intel_cap; + + /* + * Intel DebugStore bits + */ + int bts, pebs; + int pebs_record_size; + void (*drain_pebs)(struct pt_regs *regs); + struct event_constraint *pebs_constraints; + + /* + * Intel LBR + */ + unsigned long lbr_tos, lbr_from, lbr_to; /* MSR base regs */ + int lbr_nr; /* hardware stack size */ }; static struct x86_pmu x86_pmu __read_mostly; @@ -240,9 +311,10 @@ again: static atomic_t active_events; static DEFINE_MUTEX(pmc_reserve_mutex); +#ifdef CONFIG_X86_LOCAL_APIC + static bool reserve_pmc_hardware(void) { -#ifdef CONFIG_X86_LOCAL_APIC int i; if (nmi_watchdog == NMI_LOCAL_APIC) @@ -257,11 +329,9 @@ static bool reserve_pmc_hardware(void) if (!reserve_evntsel_nmi(x86_pmu.eventsel + i)) goto eventsel_fail; } -#endif return true; -#ifdef CONFIG_X86_LOCAL_APIC eventsel_fail: for (i--; i >= 0; i--) release_evntsel_nmi(x86_pmu.eventsel + i); @@ -276,12 +346,10 @@ perfctr_fail: enable_lapic_nmi_watchdog(); return false; -#endif } static void release_pmc_hardware(void) { -#ifdef CONFIG_X86_LOCAL_APIC int i; for (i = 0; i < x86_pmu.num_events; i++) { @@ -291,113 +359,23 @@ static void release_pmc_hardware(void) if (nmi_watchdog == NMI_LOCAL_APIC) enable_lapic_nmi_watchdog(); -#endif } -static inline bool bts_available(void) -{ - return x86_pmu.enable_bts != NULL; -} - -static void init_debug_store_on_cpu(int cpu) -{ - struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds; - - if (!ds) - return; - - wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, - (u32)((u64)(unsigned long)ds), - (u32)((u64)(unsigned long)ds >> 32)); -} - -static void fini_debug_store_on_cpu(int cpu) -{ - if (!per_cpu(cpu_hw_events, cpu).ds) - return; - - wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, 0, 0); -} - -static void release_bts_hardware(void) -{ - int cpu; - - if (!bts_available()) - return; - - get_online_cpus(); - - for_each_online_cpu(cpu) - fini_debug_store_on_cpu(cpu); - - for_each_possible_cpu(cpu) { - struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds; - - if (!ds) - continue; - - per_cpu(cpu_hw_events, cpu).ds = NULL; - - kfree((void *)(unsigned long)ds->bts_buffer_base); - kfree(ds); - } - - put_online_cpus(); -} - -static int reserve_bts_hardware(void) -{ - int cpu, err = 0; - - if (!bts_available()) - return 0; - - get_online_cpus(); - - for_each_possible_cpu(cpu) { - struct debug_store *ds; - void *buffer; - - err = -ENOMEM; - buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL); - if (unlikely(!buffer)) - break; - - ds = kzalloc(sizeof(*ds), GFP_KERNEL); - if (unlikely(!ds)) { - kfree(buffer); - break; - } - - ds->bts_buffer_base = (u64)(unsigned long)buffer; - ds->bts_index = ds->bts_buffer_base; - ds->bts_absolute_maximum = - ds->bts_buffer_base + BTS_BUFFER_SIZE; - ds->bts_interrupt_threshold = - ds->bts_absolute_maximum - BTS_OVFL_TH; +#else - per_cpu(cpu_hw_events, cpu).ds = ds; - err = 0; - } +static bool reserve_pmc_hardware(void) { return true; } +static void release_pmc_hardware(void) {} - if (err) - release_bts_hardware(); - else { - for_each_online_cpu(cpu) - init_debug_store_on_cpu(cpu); - } - - put_online_cpus(); +#endif - return err; -} +static int reserve_ds_buffers(void); +static void release_ds_buffers(void); static void hw_perf_event_destroy(struct perf_event *event) { if (atomic_dec_and_mutex_lock(&active_events, &pmc_reserve_mutex)) { release_pmc_hardware(); - release_bts_hardware(); + release_ds_buffers(); mutex_unlock(&pmc_reserve_mutex); } } @@ -440,6 +418,25 @@ set_ext_hw_attr(struct hw_perf_event *hwc, struct perf_event_attr *attr) return 0; } +static int x86_hw_config(struct perf_event_attr *attr, struct hw_perf_event *hwc) +{ + /* + * Generate PMC IRQs: + * (keep 'enabled' bit clear for now) + */ + hwc->config = ARCH_PERFMON_EVENTSEL_INT; + + /* + * Count user and OS events unless requested not to + */ + if (!attr->exclude_user) + hwc->config |= ARCH_PERFMON_EVENTSEL_USR; + if (!attr->exclude_kernel) + hwc->config |= ARCH_PERFMON_EVENTSEL_OS; + + return 0; +} + /* * Setup the hardware configuration for a given attr_type */ @@ -459,8 +456,11 @@ static int __hw_perf_event_init(struct perf_event *event) if (atomic_read(&active_events) == 0) { if (!reserve_pmc_hardware()) err = -EBUSY; - else - err = reserve_bts_hardware(); + else { + err = reserve_ds_buffers(); + if (err) + release_pmc_hardware(); + } } if (!err) atomic_inc(&active_events); @@ -471,23 +471,14 @@ static int __hw_perf_event_init(struct perf_event *event) event->destroy = hw_perf_event_destroy; - /* - * Generate PMC IRQs: - * (keep 'enabled' bit clear for now) - */ - hwc->config = ARCH_PERFMON_EVENTSEL_INT; - hwc->idx = -1; hwc->last_cpu = -1; hwc->last_tag = ~0ULL; - /* - * Count user and OS events unless requested not to. - */ - if (!attr->exclude_user) - hwc->config |= ARCH_PERFMON_EVENTSEL_USR; - if (!attr->exclude_kernel) - hwc->config |= ARCH_PERFMON_EVENTSEL_OS; + /* Processor specifics */ + err = x86_pmu.hw_config(attr, hwc); + if (err) + return err; if (!hwc->sample_period) { hwc->sample_period = x86_pmu.max_period; @@ -538,11 +529,11 @@ static int __hw_perf_event_init(struct perf_event *event) if ((attr->config == PERF_COUNT_HW_BRANCH_INSTRUCTIONS) && (hwc->sample_period == 1)) { /* BTS is not supported by this architecture. */ - if (!bts_available()) + if (!x86_pmu.bts) return -EOPNOTSUPP; /* BTS is currently only allowed for user-mode. */ - if (hwc->config & ARCH_PERFMON_EVENTSEL_OS) + if (!attr->exclude_kernel) return -EOPNOTSUPP; } @@ -586,7 +577,7 @@ void hw_perf_disable(void) x86_pmu.disable_all(); } -static void x86_pmu_enable_all(void) +static void x86_pmu_enable_all(int added) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); int idx; @@ -794,7 +785,7 @@ void hw_perf_enable(void) struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); struct perf_event *event; struct hw_perf_event *hwc; - int i; + int i, added = cpuc->n_added; if (!x86_pmu_initialized()) return; @@ -846,19 +837,20 @@ void hw_perf_enable(void) cpuc->enabled = 1; barrier(); - x86_pmu.enable_all(); + x86_pmu.enable_all(added); } static inline void __x86_pmu_enable_event(struct hw_perf_event *hwc) { - (void)checking_wrmsrl(hwc->config_base + hwc->idx, + wrmsrl(hwc->config_base + hwc->idx, hwc->config | ARCH_PERFMON_EVENTSEL_ENABLE); } static inline void x86_pmu_disable_event(struct perf_event *event) { struct hw_perf_event *hwc = &event->hw; - (void)checking_wrmsrl(hwc->config_base + hwc->idx, hwc->config); + + wrmsrl(hwc->config_base + hwc->idx, hwc->config); } static DEFINE_PER_CPU(u64 [X86_PMC_IDX_MAX], pmc_prev_left); @@ -873,7 +865,7 @@ x86_perf_event_set_period(struct perf_event *event) struct hw_perf_event *hwc = &event->hw; s64 left = atomic64_read(&hwc->period_left); s64 period = hwc->sample_period; - int err, ret = 0, idx = hwc->idx; + int ret = 0, idx = hwc->idx; if (idx == X86_PMC_IDX_FIXED_BTS) return 0; @@ -911,8 +903,8 @@ x86_perf_event_set_period(struct perf_event *event) */ atomic64_set(&hwc->prev_count, (u64)-left); - err = checking_wrmsrl(hwc->event_base + idx, - (u64)(-left) & x86_pmu.event_mask); + wrmsrl(hwc->event_base + idx, + (u64)(-left) & x86_pmu.event_mask); perf_event_update_userpage(event); @@ -949,7 +941,7 @@ static int x86_pmu_enable(struct perf_event *event) if (n < 0) return n; - ret = x86_schedule_events(cpuc, n, assign); + ret = x86_pmu.schedule_events(cpuc, n, assign); if (ret) return ret; /* @@ -990,6 +982,7 @@ static void x86_pmu_unthrottle(struct perf_event *event) void perf_event_print_debug(void) { u64 ctrl, status, overflow, pmc_ctrl, pmc_count, prev_left, fixed; + u64 pebs; struct cpu_hw_events *cpuc; unsigned long flags; int cpu, idx; @@ -1007,14 +1000,16 @@ void perf_event_print_debug(void) rdmsrl(MSR_CORE_PERF_GLOBAL_STATUS, status); rdmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, overflow); rdmsrl(MSR_ARCH_PERFMON_FIXED_CTR_CTRL, fixed); + rdmsrl(MSR_IA32_PEBS_ENABLE, pebs); pr_info("\n"); pr_info("CPU#%d: ctrl: %016llx\n", cpu, ctrl); pr_info("CPU#%d: status: %016llx\n", cpu, status); pr_info("CPU#%d: overflow: %016llx\n", cpu, overflow); pr_info("CPU#%d: fixed: %016llx\n", cpu, fixed); + pr_info("CPU#%d: pebs: %016llx\n", cpu, pebs); } - pr_info("CPU#%d: active: %016llx\n", cpu, *(u64 *)cpuc->active_mask); + pr_info("CPU#%d: active: %016llx\n", cpu, *(u64 *)cpuc->active_mask); for (idx = 0; idx < x86_pmu.num_events; idx++) { rdmsrl(x86_pmu.eventsel + idx, pmc_ctrl); @@ -1145,7 +1140,6 @@ void set_perf_event_pending(void) void perf_events_lapic_init(void) { -#ifdef CONFIG_X86_LOCAL_APIC if (!x86_pmu.apic || !x86_pmu_initialized()) return; @@ -1153,7 +1147,6 @@ void perf_events_lapic_init(void) * Always use NMI for PMU */ apic_write(APIC_LVTPC, APIC_DM_NMI); -#endif } static int __kprobes @@ -1177,9 +1170,7 @@ perf_event_nmi_handler(struct notifier_block *self, regs = args->regs; -#ifdef CONFIG_X86_LOCAL_APIC apic_write(APIC_LVTPC, APIC_DM_NMI); -#endif /* * Can't rely on the handled return value to say it was our NMI, two * events could trigger 'simultaneously' raising two back-to-back NMIs. @@ -1273,12 +1264,15 @@ int hw_perf_group_sched_in(struct perf_event *leader, int assign[X86_PMC_IDX_MAX]; int n0, n1, ret; + if (!x86_pmu_initialized()) + return 0; + /* n0 = total number of events */ n0 = collect_events(cpuc, leader, true); if (n0 < 0) return n0; - ret = x86_schedule_events(cpuc, n0, assign); + ret = x86_pmu.schedule_events(cpuc, n0, assign); if (ret) return ret; @@ -1328,6 +1322,9 @@ undo: #include "perf_event_amd.c" #include "perf_event_p6.c" +#include "perf_event_p4.c" +#include "perf_event_intel_lbr.c" +#include "perf_event_intel_ds.c" #include "perf_event_intel.c" static int __cpuinit @@ -1401,12 +1398,15 @@ void __init init_hw_perf_events(void) pr_cont("%s PMU driver.\n", x86_pmu.name); + if (x86_pmu.quirks) + x86_pmu.quirks(); + if (x86_pmu.num_events > X86_PMC_MAX_GENERIC) { WARN(1, KERN_ERR "hw perf events %d > max(%d), clipping!", x86_pmu.num_events, X86_PMC_MAX_GENERIC); x86_pmu.num_events = X86_PMC_MAX_GENERIC; } - perf_event_mask = (1 << x86_pmu.num_events) - 1; + x86_pmu.intel_ctrl = (1 << x86_pmu.num_events) - 1; perf_max_events = x86_pmu.num_events; if (x86_pmu.num_events_fixed > X86_PMC_MAX_FIXED) { @@ -1415,9 +1415,8 @@ void __init init_hw_perf_events(void) x86_pmu.num_events_fixed = X86_PMC_MAX_FIXED; } - perf_event_mask |= + x86_pmu.intel_ctrl |= ((1LL << x86_pmu.num_events_fixed)-1) << X86_PMC_IDX_FIXED; - x86_pmu.intel_ctrl = perf_event_mask; perf_events_lapic_init(); register_die_notifier(&perf_event_nmi_notifier); @@ -1442,7 +1441,7 @@ void __init init_hw_perf_events(void) pr_info("... value mask: %016Lx\n", x86_pmu.event_mask); pr_info("... max period: %016Lx\n", x86_pmu.max_period); pr_info("... fixed-purpose events: %d\n", x86_pmu.num_events_fixed); - pr_info("... event mask: %016Lx\n", perf_event_mask); + pr_info("... event mask: %016Lx\n", x86_pmu.intel_ctrl); perf_cpu_notifier(x86_pmu_notifier); } @@ -1462,6 +1461,32 @@ static const struct pmu pmu = { }; /* + * validate that we can schedule this event + */ +static int validate_event(struct perf_event *event) +{ + struct cpu_hw_events *fake_cpuc; + struct event_constraint *c; + int ret = 0; + + fake_cpuc = kmalloc(sizeof(*fake_cpuc), GFP_KERNEL | __GFP_ZERO); + if (!fake_cpuc) + return -ENOMEM; + + c = x86_pmu.get_event_constraints(fake_cpuc, event); + + if (!c || !c->weight) + ret = -ENOSPC; + + if (x86_pmu.put_event_constraints) + x86_pmu.put_event_constraints(fake_cpuc, event); + + kfree(fake_cpuc); + + return ret; +} + +/* * validate a single event group * * validation include: @@ -1501,7 +1526,7 @@ static int validate_group(struct perf_event *event) fake_cpuc->n_events = n; - ret = x86_schedule_events(fake_cpuc, n, NULL); + ret = x86_pmu.schedule_events(fake_cpuc, n, NULL); out_free: kfree(fake_cpuc); @@ -1526,6 +1551,8 @@ const struct pmu *hw_perf_event_init(struct perf_event *event) if (event->group_leader != event) err = validate_group(event); + else + err = validate_event(event); event->pmu = tmp; } @@ -1596,41 +1623,6 @@ perf_callchain_kernel(struct pt_regs *regs, struct perf_callchain_entry *entry) dump_trace(NULL, regs, NULL, regs->bp, &backtrace_ops, entry); } -/* - * best effort, GUP based copy_from_user() that assumes IRQ or NMI context - */ -static unsigned long -copy_from_user_nmi(void *to, const void __user *from, unsigned long n) -{ - unsigned long offset, addr = (unsigned long)from; - int type = in_nmi() ? KM_NMI : KM_IRQ0; - unsigned long size, len = 0; - struct page *page; - void *map; - int ret; - - do { - ret = __get_user_pages_fast(addr, 1, 0, &page); - if (!ret) - break; - - offset = addr & (PAGE_SIZE - 1); - size = min(PAGE_SIZE - offset, n - len); - - map = kmap_atomic(page, type); - memcpy(to, map+offset, size); - kunmap_atomic(map, type); - put_page(page); - - len += size; - to += size; - addr += size; - - } while (len < n); - - return len; -} - #ifdef CONFIG_COMPAT static inline int perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) diff --git a/arch/x86/kernel/cpu/perf_event_amd.c b/arch/x86/kernel/cpu/perf_event_amd.c index db6f7d4056e1..285623bc3cc8 100644 --- a/arch/x86/kernel/cpu/perf_event_amd.c +++ b/arch/x86/kernel/cpu/perf_event_amd.c @@ -378,6 +378,8 @@ static __initconst struct x86_pmu amd_pmu = { .enable_all = x86_pmu_enable_all, .enable = x86_pmu_enable_event, .disable = x86_pmu_disable_event, + .hw_config = x86_hw_config, + .schedule_events = x86_schedule_events, .eventsel = MSR_K7_EVNTSEL0, .perfctr = MSR_K7_PERFCTR0, .event_map = amd_pmu_event_map, diff --git a/arch/x86/kernel/cpu/perf_event_intel.c b/arch/x86/kernel/cpu/perf_event_intel.c index 84bfde64a337..676aac27aca4 100644 --- a/arch/x86/kernel/cpu/perf_event_intel.c +++ b/arch/x86/kernel/cpu/perf_event_intel.c @@ -470,42 +470,6 @@ static u64 intel_pmu_raw_event(u64 hw_event) return hw_event & CORE_EVNTSEL_MASK; } -static void intel_pmu_enable_bts(u64 config) -{ - unsigned long debugctlmsr; - - debugctlmsr = get_debugctlmsr(); - - debugctlmsr |= X86_DEBUGCTL_TR; - debugctlmsr |= X86_DEBUGCTL_BTS; - debugctlmsr |= X86_DEBUGCTL_BTINT; - - if (!(config & ARCH_PERFMON_EVENTSEL_OS)) - debugctlmsr |= X86_DEBUGCTL_BTS_OFF_OS; - - if (!(config & ARCH_PERFMON_EVENTSEL_USR)) - debugctlmsr |= X86_DEBUGCTL_BTS_OFF_USR; - - update_debugctlmsr(debugctlmsr); -} - -static void intel_pmu_disable_bts(void) -{ - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - unsigned long debugctlmsr; - - if (!cpuc->ds) - return; - - debugctlmsr = get_debugctlmsr(); - - debugctlmsr &= - ~(X86_DEBUGCTL_TR | X86_DEBUGCTL_BTS | X86_DEBUGCTL_BTINT | - X86_DEBUGCTL_BTS_OFF_OS | X86_DEBUGCTL_BTS_OFF_USR); - - update_debugctlmsr(debugctlmsr); -} - static void intel_pmu_disable_all(void) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); @@ -514,12 +478,17 @@ static void intel_pmu_disable_all(void) if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) intel_pmu_disable_bts(); + + intel_pmu_pebs_disable_all(); + intel_pmu_lbr_disable_all(); } -static void intel_pmu_enable_all(void) +static void intel_pmu_enable_all(int added) { struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + intel_pmu_pebs_enable_all(); + intel_pmu_lbr_enable_all(); wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, x86_pmu.intel_ctrl); if (test_bit(X86_PMC_IDX_FIXED_BTS, cpuc->active_mask)) { @@ -533,6 +502,40 @@ static void intel_pmu_enable_all(void) } } +/* + * Workaround for: + * Intel Errata AAK100 (model 26) + * Intel Errata AAP53 (model 30) + * + * These chips need to be 'reset' when adding counters by programming + * the magic three (non counting) events 0x4300D2, 0x4300B1 and 0x4300B5 + * either in sequence on the same PMC or on different PMCs. + */ +static void intel_pmu_nhm_enable_all(int added) +{ + if (added) { + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int i; + + wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 0, 0x4300D2); + wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 1, 0x4300B1); + wrmsrl(MSR_ARCH_PERFMON_EVENTSEL0 + 2, 0x4300B5); + + wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0x3); + wrmsrl(MSR_CORE_PERF_GLOBAL_CTRL, 0x0); + + for (i = 0; i < 3; i++) { + struct perf_event *event = cpuc->events[i]; + + if (!event) + continue; + + __x86_pmu_enable_event(&event->hw); + } + } + intel_pmu_enable_all(added); +} + static inline u64 intel_pmu_get_status(void) { u64 status; @@ -547,8 +550,7 @@ static inline void intel_pmu_ack_status(u64 ack) wrmsrl(MSR_CORE_PERF_GLOBAL_OVF_CTRL, ack); } -static inline void -intel_pmu_disable_fixed(struct hw_perf_event *hwc) +static void intel_pmu_disable_fixed(struct hw_perf_event *hwc) { int idx = hwc->idx - X86_PMC_IDX_FIXED; u64 ctrl_val, mask; @@ -557,71 +559,10 @@ intel_pmu_disable_fixed(struct hw_perf_event *hwc) rdmsrl(hwc->config_base, ctrl_val); ctrl_val &= ~mask; - (void)checking_wrmsrl(hwc->config_base, ctrl_val); -} - -static void intel_pmu_drain_bts_buffer(void) -{ - struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); - struct debug_store *ds = cpuc->ds; - struct bts_record { - u64 from; - u64 to; - u64 flags; - }; - struct perf_event *event = cpuc->events[X86_PMC_IDX_FIXED_BTS]; - struct bts_record *at, *top; - struct perf_output_handle handle; - struct perf_event_header header; - struct perf_sample_data data; - struct pt_regs regs; - - if (!event) - return; - - if (!ds) - return; - - at = (struct bts_record *)(unsigned long)ds->bts_buffer_base; - top = (struct bts_record *)(unsigned long)ds->bts_index; - - if (top <= at) - return; - - ds->bts_index = ds->bts_buffer_base; - - perf_sample_data_init(&data, 0); - - data.period = event->hw.last_period; - regs.ip = 0; - - /* - * Prepare a generic sample, i.e. fill in the invariant fields. - * We will overwrite the from and to address before we output - * the sample. - */ - perf_prepare_sample(&header, &data, event, ®s); - - if (perf_output_begin(&handle, event, - header.size * (top - at), 1, 1)) - return; - - for (; at < top; at++) { - data.ip = at->from; - data.addr = at->to; - - perf_output_sample(&handle, &header, &data, event); - } - - perf_output_end(&handle); - - /* There's new data available. */ - event->hw.interrupts++; - event->pending_kill = POLL_IN; + wrmsrl(hwc->config_base, ctrl_val); } -static inline void -intel_pmu_disable_event(struct perf_event *event) +static void intel_pmu_disable_event(struct perf_event *event) { struct hw_perf_event *hwc = &event->hw; @@ -637,14 +578,15 @@ intel_pmu_disable_event(struct perf_event *event) } x86_pmu_disable_event(event); + + if (unlikely(event->attr.precise)) + intel_pmu_pebs_disable(event); } -static inline void -intel_pmu_enable_fixed(struct hw_perf_event *hwc) +static void intel_pmu_enable_fixed(struct hw_perf_event *hwc) { int idx = hwc->idx - X86_PMC_IDX_FIXED; u64 ctrl_val, bits, mask; - int err; /* * Enable IRQ generation (0x8), @@ -669,7 +611,7 @@ intel_pmu_enable_fixed(struct hw_perf_event *hwc) rdmsrl(hwc->config_base, ctrl_val); ctrl_val &= ~mask; ctrl_val |= bits; - err = checking_wrmsrl(hwc->config_base, ctrl_val); + wrmsrl(hwc->config_base, ctrl_val); } static void intel_pmu_enable_event(struct perf_event *event) @@ -689,6 +631,9 @@ static void intel_pmu_enable_event(struct perf_event *event) return; } + if (unlikely(event->attr.precise)) + intel_pmu_pebs_enable(event); + __x86_pmu_enable_event(hwc); } @@ -747,7 +692,7 @@ static int intel_pmu_handle_irq(struct pt_regs *regs) intel_pmu_drain_bts_buffer(); status = intel_pmu_get_status(); if (!status) { - intel_pmu_enable_all(); + intel_pmu_enable_all(0); return 0; } @@ -762,6 +707,15 @@ again: inc_irq_stat(apic_perf_irqs); ack = status; + + intel_pmu_lbr_read(); + + /* + * PEBS overflow sets bit 62 in the global status register + */ + if (__test_and_clear_bit(62, (unsigned long *)&status)) + x86_pmu.drain_pebs(regs); + for_each_set_bit(bit, (unsigned long *)&status, X86_PMC_IDX_MAX) { struct perf_event *event = cpuc->events[bit]; @@ -787,26 +741,22 @@ again: goto again; done: - intel_pmu_enable_all(); + intel_pmu_enable_all(0); return 1; } -static struct event_constraint bts_constraint = - EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0); - static struct event_constraint * -intel_special_constraints(struct perf_event *event) +intel_bts_constraints(struct perf_event *event) { - unsigned int hw_event; - - hw_event = event->hw.config & INTEL_ARCH_EVENT_MASK; + struct hw_perf_event *hwc = &event->hw; + unsigned int hw_event, bts_event; - if (unlikely((hw_event == - x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS)) && - (event->hw.sample_period == 1))) { + hw_event = hwc->config & INTEL_ARCH_EVENT_MASK; + bts_event = x86_pmu.event_map(PERF_COUNT_HW_BRANCH_INSTRUCTIONS); + if (unlikely(hw_event == bts_event && hwc->sample_period == 1)) return &bts_constraint; - } + return NULL; } @@ -815,7 +765,11 @@ intel_get_event_constraints(struct cpu_hw_events *cpuc, struct perf_event *event { struct event_constraint *c; - c = intel_special_constraints(event); + c = intel_bts_constraints(event); + if (c) + return c; + + c = intel_pebs_constraints(event); if (c) return c; @@ -829,6 +783,8 @@ static __initconst struct x86_pmu core_pmu = { .enable_all = x86_pmu_enable_all, .enable = x86_pmu_enable_event, .disable = x86_pmu_disable_event, + .hw_config = x86_hw_config, + .schedule_events = x86_schedule_events, .eventsel = MSR_ARCH_PERFMON_EVENTSEL0, .perfctr = MSR_ARCH_PERFMON_PERFCTR0, .event_map = intel_pmu_event_map, @@ -845,6 +801,20 @@ static __initconst struct x86_pmu core_pmu = { .event_constraints = intel_core_event_constraints, }; +static void intel_pmu_cpu_starting(int cpu) +{ + init_debug_store_on_cpu(cpu); + /* + * Deal with CPUs that don't clear their LBRs on power-up. + */ + intel_pmu_lbr_reset(); +} + +static void intel_pmu_cpu_dying(int cpu) +{ + fini_debug_store_on_cpu(cpu); +} + static __initconst struct x86_pmu intel_pmu = { .name = "Intel", .handle_irq = intel_pmu_handle_irq, @@ -852,6 +822,8 @@ static __initconst struct x86_pmu intel_pmu = { .enable_all = intel_pmu_enable_all, .enable = intel_pmu_enable_event, .disable = intel_pmu_disable_event, + .hw_config = x86_hw_config, + .schedule_events = x86_schedule_events, .eventsel = MSR_ARCH_PERFMON_EVENTSEL0, .perfctr = MSR_ARCH_PERFMON_PERFCTR0, .event_map = intel_pmu_event_map, @@ -864,14 +836,38 @@ static __initconst struct x86_pmu intel_pmu = { * the generic event period: */ .max_period = (1ULL << 31) - 1, - .enable_bts = intel_pmu_enable_bts, - .disable_bts = intel_pmu_disable_bts, .get_event_constraints = intel_get_event_constraints, - .cpu_starting = init_debug_store_on_cpu, - .cpu_dying = fini_debug_store_on_cpu, + .cpu_starting = intel_pmu_cpu_starting, + .cpu_dying = intel_pmu_cpu_dying, }; +static void intel_clovertown_quirks(void) +{ + /* + * PEBS is unreliable due to: + * + * AJ67 - PEBS may experience CPL leaks + * AJ68 - PEBS PMI may be delayed by one event + * AJ69 - GLOBAL_STATUS[62] will only be set when DEBUGCTL[12] + * AJ106 - FREEZE_LBRS_ON_PMI doesn't work in combination with PEBS + * + * AJ67 could be worked around by restricting the OS/USR flags. + * AJ69 could be worked around by setting PMU_FREEZE_ON_PMI. + * + * AJ106 could possibly be worked around by not allowing LBR + * usage from PEBS, including the fixup. + * AJ68 could possibly be worked around by always programming + * a pebs_event_reset[0] value and coping with the lost events. + * + * But taken together it might just make sense to not enable PEBS on + * these chips. + */ + printk(KERN_WARNING "PEBS disabled due to CPU errata.\n"); + x86_pmu.pebs = 0; + x86_pmu.pebs_constraints = NULL; +} + static __init int intel_pmu_init(void) { union cpuid10_edx edx; @@ -881,12 +877,13 @@ static __init int intel_pmu_init(void) int version; if (!cpu_has(&boot_cpu_data, X86_FEATURE_ARCH_PERFMON)) { - /* check for P6 processor family */ - if (boot_cpu_data.x86 == 6) { - return p6_pmu_init(); - } else { + switch (boot_cpu_data.x86) { + case 0x6: + return p6_pmu_init(); + case 0xf: + return p4_pmu_init(); + } return -ENODEV; - } } /* @@ -916,6 +913,18 @@ static __init int intel_pmu_init(void) x86_pmu.num_events_fixed = max((int)edx.split.num_events_fixed, 3); /* + * v2 and above have a perf capabilities MSR + */ + if (version > 1) { + u64 capabilities; + + rdmsrl(MSR_IA32_PERF_CAPABILITIES, capabilities); + x86_pmu.intel_cap.capabilities = capabilities; + } + + intel_ds_init(); + + /* * Install the hw-cache-events table: */ switch (boot_cpu_data.x86_model) { @@ -924,12 +933,15 @@ static __init int intel_pmu_init(void) break; case 15: /* original 65 nm celeron/pentium/core2/xeon, "Merom"/"Conroe" */ + x86_pmu.quirks = intel_clovertown_quirks; case 22: /* single-core 65 nm celeron/core2solo "Merom-L"/"Conroe-L" */ case 23: /* current 45 nm celeron/core2/xeon "Penryn"/"Wolfdale" */ case 29: /* six-core 45 nm xeon "Dunnington" */ memcpy(hw_cache_event_ids, core2_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + intel_pmu_lbr_init_core(); + x86_pmu.event_constraints = intel_core2_event_constraints; pr_cont("Core2 events, "); break; @@ -939,13 +951,19 @@ static __init int intel_pmu_init(void) memcpy(hw_cache_event_ids, nehalem_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + intel_pmu_lbr_init_nhm(); + x86_pmu.event_constraints = intel_nehalem_event_constraints; - pr_cont("Nehalem/Corei7 events, "); + x86_pmu.enable_all = intel_pmu_nhm_enable_all; + pr_cont("Nehalem events, "); break; + case 28: /* Atom */ memcpy(hw_cache_event_ids, atom_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + intel_pmu_lbr_init_atom(); + x86_pmu.event_constraints = intel_gen_event_constraints; pr_cont("Atom events, "); break; @@ -955,6 +973,8 @@ static __init int intel_pmu_init(void) memcpy(hw_cache_event_ids, westmere_hw_cache_event_ids, sizeof(hw_cache_event_ids)); + intel_pmu_lbr_init_nhm(); + x86_pmu.event_constraints = intel_westmere_event_constraints; pr_cont("Westmere events, "); break; diff --git a/arch/x86/kernel/cpu/perf_event_intel_ds.c b/arch/x86/kernel/cpu/perf_event_intel_ds.c new file mode 100644 index 000000000000..2fea3622af7f --- /dev/null +++ b/arch/x86/kernel/cpu/perf_event_intel_ds.c @@ -0,0 +1,664 @@ +#ifdef CONFIG_CPU_SUP_INTEL + +/* The maximal number of PEBS events: */ +#define MAX_PEBS_EVENTS 4 + +/* The size of a BTS record in bytes: */ +#define BTS_RECORD_SIZE 24 + +#define BTS_BUFFER_SIZE (PAGE_SIZE << 4) +#define PEBS_BUFFER_SIZE PAGE_SIZE + +/* + * pebs_record_32 for p4 and core not supported + +struct pebs_record_32 { + u32 flags, ip; + u32 ax, bc, cx, dx; + u32 si, di, bp, sp; +}; + + */ + +struct pebs_record_core { + u64 flags, ip; + u64 ax, bx, cx, dx; + u64 si, di, bp, sp; + u64 r8, r9, r10, r11; + u64 r12, r13, r14, r15; +}; + +struct pebs_record_nhm { + u64 flags, ip; + u64 ax, bx, cx, dx; + u64 si, di, bp, sp; + u64 r8, r9, r10, r11; + u64 r12, r13, r14, r15; + u64 status, dla, dse, lat; +}; + +/* + * A debug store configuration. + * + * We only support architectures that use 64bit fields. + */ +struct debug_store { + u64 bts_buffer_base; + u64 bts_index; + u64 bts_absolute_maximum; + u64 bts_interrupt_threshold; + u64 pebs_buffer_base; + u64 pebs_index; + u64 pebs_absolute_maximum; + u64 pebs_interrupt_threshold; + u64 pebs_event_reset[MAX_PEBS_EVENTS]; +}; + +static void init_debug_store_on_cpu(int cpu) +{ + struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds; + + if (!ds) + return; + + wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, + (u32)((u64)(unsigned long)ds), + (u32)((u64)(unsigned long)ds >> 32)); +} + +static void fini_debug_store_on_cpu(int cpu) +{ + if (!per_cpu(cpu_hw_events, cpu).ds) + return; + + wrmsr_on_cpu(cpu, MSR_IA32_DS_AREA, 0, 0); +} + +static void release_ds_buffers(void) +{ + int cpu; + + if (!x86_pmu.bts && !x86_pmu.pebs) + return; + + get_online_cpus(); + + for_each_online_cpu(cpu) + fini_debug_store_on_cpu(cpu); + + for_each_possible_cpu(cpu) { + struct debug_store *ds = per_cpu(cpu_hw_events, cpu).ds; + + if (!ds) + continue; + + per_cpu(cpu_hw_events, cpu).ds = NULL; + + kfree((void *)(unsigned long)ds->pebs_buffer_base); + kfree((void *)(unsigned long)ds->bts_buffer_base); + kfree(ds); + } + + put_online_cpus(); +} + +static int reserve_ds_buffers(void) +{ + int cpu, err = 0; + + if (!x86_pmu.bts && !x86_pmu.pebs) + return 0; + + get_online_cpus(); + + for_each_possible_cpu(cpu) { + struct debug_store *ds; + void *buffer; + int max, thresh; + + err = -ENOMEM; + ds = kzalloc(sizeof(*ds), GFP_KERNEL); + if (unlikely(!ds)) + break; + per_cpu(cpu_hw_events, cpu).ds = ds; + + if (x86_pmu.bts) { + buffer = kzalloc(BTS_BUFFER_SIZE, GFP_KERNEL); + if (unlikely(!buffer)) + break; + + max = BTS_BUFFER_SIZE / BTS_RECORD_SIZE; + thresh = max / 16; + + ds->bts_buffer_base = (u64)(unsigned long)buffer; + ds->bts_index = ds->bts_buffer_base; + ds->bts_absolute_maximum = ds->bts_buffer_base + + max * BTS_RECORD_SIZE; + ds->bts_interrupt_threshold = ds->bts_absolute_maximum - + thresh * BTS_RECORD_SIZE; + } + + if (x86_pmu.pebs) { + buffer = kzalloc(PEBS_BUFFER_SIZE, GFP_KERNEL); + if (unlikely(!buffer)) + break; + + max = PEBS_BUFFER_SIZE / x86_pmu.pebs_record_size; + + ds->pebs_buffer_base = (u64)(unsigned long)buffer; + ds->pebs_index = ds->pebs_buffer_base; + ds->pebs_absolute_maximum = ds->pebs_buffer_base + + max * x86_pmu.pebs_record_size; + /* + * Always use single record PEBS + */ + ds->pebs_interrupt_threshold = ds->pebs_buffer_base + + x86_pmu.pebs_record_size; + } + + err = 0; + } + + if (err) + release_ds_buffers(); + else { + for_each_online_cpu(cpu) + init_debug_store_on_cpu(cpu); + } + + put_online_cpus(); + + return err; +} + +/* + * BTS + */ + +static struct event_constraint bts_constraint = + EVENT_CONSTRAINT(0, 1ULL << X86_PMC_IDX_FIXED_BTS, 0); + +static void intel_pmu_enable_bts(u64 config) +{ + unsigned long debugctlmsr; + + debugctlmsr = get_debugctlmsr(); + + debugctlmsr |= DEBUGCTLMSR_TR; + debugctlmsr |= DEBUGCTLMSR_BTS; + debugctlmsr |= DEBUGCTLMSR_BTINT; + + if (!(config & ARCH_PERFMON_EVENTSEL_OS)) + debugctlmsr |= DEBUGCTLMSR_BTS_OFF_OS; + + if (!(config & ARCH_PERFMON_EVENTSEL_USR)) + debugctlmsr |= DEBUGCTLMSR_BTS_OFF_USR; + + update_debugctlmsr(debugctlmsr); +} + +static void intel_pmu_disable_bts(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + unsigned long debugctlmsr; + + if (!cpuc->ds) + return; + + debugctlmsr = get_debugctlmsr(); + + debugctlmsr &= + ~(DEBUGCTLMSR_TR | DEBUGCTLMSR_BTS | DEBUGCTLMSR_BTINT | + DEBUGCTLMSR_BTS_OFF_OS | DEBUGCTLMSR_BTS_OFF_USR); + + update_debugctlmsr(debugctlmsr); +} + +static void intel_pmu_drain_bts_buffer(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct debug_store *ds = cpuc->ds; + struct bts_record { + u64 from; + u64 to; + u64 flags; + }; + struct perf_event *event = cpuc->events[X86_PMC_IDX_FIXED_BTS]; + struct bts_record *at, *top; + struct perf_output_handle handle; + struct perf_event_header header; + struct perf_sample_data data; + struct pt_regs regs; + + if (!event) + return; + + if (!ds) + return; + + at = (struct bts_record *)(unsigned long)ds->bts_buffer_base; + top = (struct bts_record *)(unsigned long)ds->bts_index; + + if (top <= at) + return; + + ds->bts_index = ds->bts_buffer_base; + + perf_sample_data_init(&data, 0); + data.period = event->hw.last_period; + regs.ip = 0; + + /* + * Prepare a generic sample, i.e. fill in the invariant fields. + * We will overwrite the from and to address before we output + * the sample. + */ + perf_prepare_sample(&header, &data, event, ®s); + + if (perf_output_begin(&handle, event, header.size * (top - at), 1, 1)) + return; + + for (; at < top; at++) { + data.ip = at->from; + data.addr = at->to; + + perf_output_sample(&handle, &header, &data, event); + } + + perf_output_end(&handle); + + /* There's new data available. */ + event->hw.interrupts++; + event->pending_kill = POLL_IN; +} + +/* + * PEBS + */ + +static struct event_constraint intel_core_pebs_events[] = { + PEBS_EVENT_CONSTRAINT(0x00c0, 0x1), /* INSTR_RETIRED.ANY */ + PEBS_EVENT_CONSTRAINT(0xfec1, 0x1), /* X87_OPS_RETIRED.ANY */ + PEBS_EVENT_CONSTRAINT(0x00c5, 0x1), /* BR_INST_RETIRED.MISPRED */ + PEBS_EVENT_CONSTRAINT(0x1fc7, 0x1), /* SIMD_INST_RETURED.ANY */ + PEBS_EVENT_CONSTRAINT(0x01cb, 0x1), /* MEM_LOAD_RETIRED.L1D_MISS */ + PEBS_EVENT_CONSTRAINT(0x02cb, 0x1), /* MEM_LOAD_RETIRED.L1D_LINE_MISS */ + PEBS_EVENT_CONSTRAINT(0x04cb, 0x1), /* MEM_LOAD_RETIRED.L2_MISS */ + PEBS_EVENT_CONSTRAINT(0x08cb, 0x1), /* MEM_LOAD_RETIRED.L2_LINE_MISS */ + PEBS_EVENT_CONSTRAINT(0x10cb, 0x1), /* MEM_LOAD_RETIRED.DTLB_MISS */ + EVENT_CONSTRAINT_END +}; + +static struct event_constraint intel_nehalem_pebs_events[] = { + PEBS_EVENT_CONSTRAINT(0x00c0, 0xf), /* INSTR_RETIRED.ANY */ + PEBS_EVENT_CONSTRAINT(0xfec1, 0xf), /* X87_OPS_RETIRED.ANY */ + PEBS_EVENT_CONSTRAINT(0x00c5, 0xf), /* BR_INST_RETIRED.MISPRED */ + PEBS_EVENT_CONSTRAINT(0x1fc7, 0xf), /* SIMD_INST_RETURED.ANY */ + PEBS_EVENT_CONSTRAINT(0x01cb, 0xf), /* MEM_LOAD_RETIRED.L1D_MISS */ + PEBS_EVENT_CONSTRAINT(0x02cb, 0xf), /* MEM_LOAD_RETIRED.L1D_LINE_MISS */ + PEBS_EVENT_CONSTRAINT(0x04cb, 0xf), /* MEM_LOAD_RETIRED.L2_MISS */ + PEBS_EVENT_CONSTRAINT(0x08cb, 0xf), /* MEM_LOAD_RETIRED.L2_LINE_MISS */ + PEBS_EVENT_CONSTRAINT(0x10cb, 0xf), /* MEM_LOAD_RETIRED.DTLB_MISS */ + EVENT_CONSTRAINT_END +}; + +static struct event_constraint * +intel_pebs_constraints(struct perf_event *event) +{ + struct event_constraint *c; + + if (!event->attr.precise) + return NULL; + + if (x86_pmu.pebs_constraints) { + for_each_event_constraint(c, x86_pmu.pebs_constraints) { + if ((event->hw.config & c->cmask) == c->code) + return c; + } + } + + return &emptyconstraint; +} + +static void intel_pmu_pebs_enable(struct perf_event *event) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct hw_perf_event *hwc = &event->hw; + + hwc->config &= ~ARCH_PERFMON_EVENTSEL_INT; + + cpuc->pebs_enabled |= 1ULL << hwc->idx; + WARN_ON_ONCE(cpuc->enabled); + + if (x86_pmu.intel_cap.pebs_trap) + intel_pmu_lbr_enable(event); +} + +static void intel_pmu_pebs_disable(struct perf_event *event) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct hw_perf_event *hwc = &event->hw; + + cpuc->pebs_enabled &= ~(1ULL << hwc->idx); + if (cpuc->enabled) + wrmsrl(MSR_IA32_PEBS_ENABLE, cpuc->pebs_enabled); + + hwc->config |= ARCH_PERFMON_EVENTSEL_INT; + + if (x86_pmu.intel_cap.pebs_trap) + intel_pmu_lbr_disable(event); +} + +static void intel_pmu_pebs_enable_all(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (cpuc->pebs_enabled) + wrmsrl(MSR_IA32_PEBS_ENABLE, cpuc->pebs_enabled); +} + +static void intel_pmu_pebs_disable_all(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (cpuc->pebs_enabled) + wrmsrl(MSR_IA32_PEBS_ENABLE, 0); +} + +#include <asm/insn.h> + +static inline bool kernel_ip(unsigned long ip) +{ +#ifdef CONFIG_X86_32 + return ip > PAGE_OFFSET; +#else + return (long)ip < 0; +#endif +} + +static int intel_pmu_pebs_fixup_ip(struct pt_regs *regs) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + unsigned long from = cpuc->lbr_entries[0].from; + unsigned long old_to, to = cpuc->lbr_entries[0].to; + unsigned long ip = regs->ip; + + /* + * We don't need to fixup if the PEBS assist is fault like + */ + if (!x86_pmu.intel_cap.pebs_trap) + return 1; + + /* + * No LBR entry, no basic block, no rewinding + */ + if (!cpuc->lbr_stack.nr || !from || !to) + return 0; + + /* + * Basic blocks should never cross user/kernel boundaries + */ + if (kernel_ip(ip) != kernel_ip(to)) + return 0; + + /* + * unsigned math, either ip is before the start (impossible) or + * the basic block is larger than 1 page (sanity) + */ + if ((ip - to) > PAGE_SIZE) + return 0; + + /* + * We sampled a branch insn, rewind using the LBR stack + */ + if (ip == to) { + regs->ip = from; + return 1; + } + + do { + struct insn insn; + u8 buf[MAX_INSN_SIZE]; + void *kaddr; + + old_to = to; + if (!kernel_ip(ip)) { + int bytes, size = MAX_INSN_SIZE; + + bytes = copy_from_user_nmi(buf, (void __user *)to, size); + if (bytes != size) + return 0; + + kaddr = buf; + } else + kaddr = (void *)to; + + kernel_insn_init(&insn, kaddr); + insn_get_length(&insn); + to += insn.length; + } while (to < ip); + + if (to == ip) { + regs->ip = old_to; + return 1; + } + + /* + * Even though we decoded the basic block, the instruction stream + * never matched the given IP, either the TO or the IP got corrupted. + */ + return 0; +} + +static int intel_pmu_save_and_restart(struct perf_event *event); + +static void intel_pmu_drain_pebs_core(struct pt_regs *iregs) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct debug_store *ds = cpuc->ds; + struct perf_event *event = cpuc->events[0]; /* PMC0 only */ + struct pebs_record_core *at, *top; + struct perf_sample_data data; + struct perf_raw_record raw; + struct pt_regs regs; + int n; + + if (!ds || !x86_pmu.pebs) + return; + + at = (struct pebs_record_core *)(unsigned long)ds->pebs_buffer_base; + top = (struct pebs_record_core *)(unsigned long)ds->pebs_index; + + /* + * Whatever else happens, drain the thing + */ + ds->pebs_index = ds->pebs_buffer_base; + + if (!test_bit(0, cpuc->active_mask)) + return; + + WARN_ON_ONCE(!event); + + if (!event->attr.precise) + return; + + n = top - at; + if (n <= 0) + return; + + if (!intel_pmu_save_and_restart(event)) + return; + + /* + * Should not happen, we program the threshold at 1 and do not + * set a reset value. + */ + WARN_ON_ONCE(n > 1); + at += n - 1; + + perf_sample_data_init(&data, 0); + data.period = event->hw.last_period; + + if (event->attr.sample_type & PERF_SAMPLE_RAW) { + raw.size = x86_pmu.pebs_record_size; + raw.data = at; + data.raw = &raw; + } + + /* + * We use the interrupt regs as a base because the PEBS record + * does not contain a full regs set, specifically it seems to + * lack segment descriptors, which get used by things like + * user_mode(). + * + * In the simple case fix up only the IP and BP,SP regs, for + * PERF_SAMPLE_IP and PERF_SAMPLE_CALLCHAIN to function properly. + * A possible PERF_SAMPLE_REGS will have to transfer all regs. + */ + regs = *iregs; + regs.ip = at->ip; + regs.bp = at->bp; + regs.sp = at->sp; + + if (intel_pmu_pebs_fixup_ip(®s)) + regs.flags |= PERF_EFLAGS_EXACT; + else + regs.flags &= ~PERF_EFLAGS_EXACT; + + if (perf_event_overflow(event, 1, &data, ®s)) + x86_pmu_stop(event); +} + +static void intel_pmu_drain_pebs_nhm(struct pt_regs *iregs) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + struct debug_store *ds = cpuc->ds; + struct pebs_record_nhm *at, *top; + struct perf_sample_data data; + struct perf_event *event = NULL; + struct perf_raw_record raw; + struct pt_regs regs; + u64 status = 0; + int bit, n; + + if (!ds || !x86_pmu.pebs) + return; + + at = (struct pebs_record_nhm *)(unsigned long)ds->pebs_buffer_base; + top = (struct pebs_record_nhm *)(unsigned long)ds->pebs_index; + + ds->pebs_index = ds->pebs_buffer_base; + + n = top - at; + if (n <= 0) + return; + + /* + * Should not happen, we program the threshold at 1 and do not + * set a reset value. + */ + WARN_ON_ONCE(n > MAX_PEBS_EVENTS); + + for ( ; at < top; at++) { + for_each_bit(bit, (unsigned long *)&at->status, MAX_PEBS_EVENTS) { + event = cpuc->events[bit]; + if (!test_bit(bit, cpuc->active_mask)) + continue; + + WARN_ON_ONCE(!event); + + if (!event->attr.precise) + continue; + + if (__test_and_set_bit(bit, (unsigned long *)&status)) + continue; + + break; + } + + if (!event || bit >= MAX_PEBS_EVENTS) + continue; + + if (!intel_pmu_save_and_restart(event)) + continue; + + perf_sample_data_init(&data, 0); + data.period = event->hw.last_period; + + if (event->attr.sample_type & PERF_SAMPLE_RAW) { + raw.size = x86_pmu.pebs_record_size; + raw.data = at; + data.raw = &raw; + } + + /* + * See the comment in intel_pmu_drain_pebs_core() + */ + regs = *iregs; + regs.ip = at->ip; + regs.bp = at->bp; + regs.sp = at->sp; + + if (intel_pmu_pebs_fixup_ip(®s)) + regs.flags |= PERF_EFLAGS_EXACT; + else + regs.flags &= ~PERF_EFLAGS_EXACT; + + if (perf_event_overflow(event, 1, &data, ®s)) + x86_pmu_stop(event); + } +} + +/* + * BTS, PEBS probe and setup + */ + +static void intel_ds_init(void) +{ + /* + * No support for 32bit formats + */ + if (!boot_cpu_has(X86_FEATURE_DTES64)) + return; + + x86_pmu.bts = boot_cpu_has(X86_FEATURE_BTS); + x86_pmu.pebs = boot_cpu_has(X86_FEATURE_PEBS); + if (x86_pmu.pebs) { + char pebs_type = x86_pmu.intel_cap.pebs_trap ? '+' : '-'; + int format = x86_pmu.intel_cap.pebs_format; + + switch (format) { + case 0: + printk(KERN_CONT "PEBS fmt0%c, ", pebs_type); + x86_pmu.pebs_record_size = sizeof(struct pebs_record_core); + x86_pmu.drain_pebs = intel_pmu_drain_pebs_core; + x86_pmu.pebs_constraints = intel_core_pebs_events; + break; + + case 1: + printk(KERN_CONT "PEBS fmt1%c, ", pebs_type); + x86_pmu.pebs_record_size = sizeof(struct pebs_record_nhm); + x86_pmu.drain_pebs = intel_pmu_drain_pebs_nhm; + x86_pmu.pebs_constraints = intel_nehalem_pebs_events; + break; + + default: + printk(KERN_CONT "no PEBS fmt%d%c, ", format, pebs_type); + x86_pmu.pebs = 0; + break; + } + } +} + +#else /* CONFIG_CPU_SUP_INTEL */ + +static int reserve_ds_buffers(void) +{ + return 0; +} + +static void release_ds_buffers(void) +{ +} + +#endif /* CONFIG_CPU_SUP_INTEL */ diff --git a/arch/x86/kernel/cpu/perf_event_intel_lbr.c b/arch/x86/kernel/cpu/perf_event_intel_lbr.c new file mode 100644 index 000000000000..d202c1bece1a --- /dev/null +++ b/arch/x86/kernel/cpu/perf_event_intel_lbr.c @@ -0,0 +1,218 @@ +#ifdef CONFIG_CPU_SUP_INTEL + +enum { + LBR_FORMAT_32 = 0x00, + LBR_FORMAT_LIP = 0x01, + LBR_FORMAT_EIP = 0x02, + LBR_FORMAT_EIP_FLAGS = 0x03, +}; + +/* + * We only support LBR implementations that have FREEZE_LBRS_ON_PMI + * otherwise it becomes near impossible to get a reliable stack. + */ + +static void __intel_pmu_lbr_enable(void) +{ + u64 debugctl; + + rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); + debugctl |= (DEBUGCTLMSR_LBR | DEBUGCTLMSR_FREEZE_LBRS_ON_PMI); + wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); +} + +static void __intel_pmu_lbr_disable(void) +{ + u64 debugctl; + + rdmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); + debugctl &= ~(DEBUGCTLMSR_LBR | DEBUGCTLMSR_FREEZE_LBRS_ON_PMI); + wrmsrl(MSR_IA32_DEBUGCTLMSR, debugctl); +} + +static void intel_pmu_lbr_reset_32(void) +{ + int i; + + for (i = 0; i < x86_pmu.lbr_nr; i++) + wrmsrl(x86_pmu.lbr_from + i, 0); +} + +static void intel_pmu_lbr_reset_64(void) +{ + int i; + + for (i = 0; i < x86_pmu.lbr_nr; i++) { + wrmsrl(x86_pmu.lbr_from + i, 0); + wrmsrl(x86_pmu.lbr_to + i, 0); + } +} + +static void intel_pmu_lbr_reset(void) +{ + if (!x86_pmu.lbr_nr) + return; + + if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_32) + intel_pmu_lbr_reset_32(); + else + intel_pmu_lbr_reset_64(); +} + +static void intel_pmu_lbr_enable(struct perf_event *event) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (!x86_pmu.lbr_nr) + return; + + WARN_ON_ONCE(cpuc->enabled); + + /* + * Reset the LBR stack if we changed task context to + * avoid data leaks. + */ + + if (event->ctx->task && cpuc->lbr_context != event->ctx) { + intel_pmu_lbr_reset(); + cpuc->lbr_context = event->ctx; + } + + cpuc->lbr_users++; +} + +static void intel_pmu_lbr_disable(struct perf_event *event) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (!x86_pmu.lbr_nr) + return; + + cpuc->lbr_users--; + WARN_ON_ONCE(cpuc->lbr_users < 0); + + if (cpuc->enabled && !cpuc->lbr_users) + __intel_pmu_lbr_disable(); +} + +static void intel_pmu_lbr_enable_all(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (cpuc->lbr_users) + __intel_pmu_lbr_enable(); +} + +static void intel_pmu_lbr_disable_all(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (cpuc->lbr_users) + __intel_pmu_lbr_disable(); +} + +static inline u64 intel_pmu_lbr_tos(void) +{ + u64 tos; + + rdmsrl(x86_pmu.lbr_tos, tos); + + return tos; +} + +static void intel_pmu_lbr_read_32(struct cpu_hw_events *cpuc) +{ + unsigned long mask = x86_pmu.lbr_nr - 1; + u64 tos = intel_pmu_lbr_tos(); + int i; + + for (i = 0; i < x86_pmu.lbr_nr; i++) { + unsigned long lbr_idx = (tos - i) & mask; + union { + struct { + u32 from; + u32 to; + }; + u64 lbr; + } msr_lastbranch; + + rdmsrl(x86_pmu.lbr_from + lbr_idx, msr_lastbranch.lbr); + + cpuc->lbr_entries[i].from = msr_lastbranch.from; + cpuc->lbr_entries[i].to = msr_lastbranch.to; + cpuc->lbr_entries[i].flags = 0; + } + cpuc->lbr_stack.nr = i; +} + +#define LBR_FROM_FLAG_MISPRED (1ULL << 63) + +/* + * Due to lack of segmentation in Linux the effective address (offset) + * is the same as the linear address, allowing us to merge the LIP and EIP + * LBR formats. + */ +static void intel_pmu_lbr_read_64(struct cpu_hw_events *cpuc) +{ + unsigned long mask = x86_pmu.lbr_nr - 1; + int lbr_format = x86_pmu.intel_cap.lbr_format; + u64 tos = intel_pmu_lbr_tos(); + int i; + + for (i = 0; i < x86_pmu.lbr_nr; i++) { + unsigned long lbr_idx = (tos - i) & mask; + u64 from, to, flags = 0; + + rdmsrl(x86_pmu.lbr_from + lbr_idx, from); + rdmsrl(x86_pmu.lbr_to + lbr_idx, to); + + if (lbr_format == LBR_FORMAT_EIP_FLAGS) { + flags = !!(from & LBR_FROM_FLAG_MISPRED); + from = (u64)((((s64)from) << 1) >> 1); + } + + cpuc->lbr_entries[i].from = from; + cpuc->lbr_entries[i].to = to; + cpuc->lbr_entries[i].flags = flags; + } + cpuc->lbr_stack.nr = i; +} + +static void intel_pmu_lbr_read(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + + if (!cpuc->lbr_users) + return; + + if (x86_pmu.intel_cap.lbr_format == LBR_FORMAT_32) + intel_pmu_lbr_read_32(cpuc); + else + intel_pmu_lbr_read_64(cpuc); +} + +static void intel_pmu_lbr_init_core(void) +{ + x86_pmu.lbr_nr = 4; + x86_pmu.lbr_tos = 0x01c9; + x86_pmu.lbr_from = 0x40; + x86_pmu.lbr_to = 0x60; +} + +static void intel_pmu_lbr_init_nhm(void) +{ + x86_pmu.lbr_nr = 16; + x86_pmu.lbr_tos = 0x01c9; + x86_pmu.lbr_from = 0x680; + x86_pmu.lbr_to = 0x6c0; +} + +static void intel_pmu_lbr_init_atom(void) +{ + x86_pmu.lbr_nr = 8; + x86_pmu.lbr_tos = 0x01c9; + x86_pmu.lbr_from = 0x40; + x86_pmu.lbr_to = 0x60; +} + +#endif /* CONFIG_CPU_SUP_INTEL */ diff --git a/arch/x86/kernel/cpu/perf_event_p4.c b/arch/x86/kernel/cpu/perf_event_p4.c new file mode 100644 index 000000000000..0d1be36cbe9e --- /dev/null +++ b/arch/x86/kernel/cpu/perf_event_p4.c @@ -0,0 +1,830 @@ +/* + * Netburst Perfomance Events (P4, old Xeon) + * + * Copyright (C) 2010 Parallels, Inc., Cyrill Gorcunov <gorcunov@openvz.org> + * Copyright (C) 2010 Intel Corporation, Lin Ming <ming.m.lin@intel.com> + * + * For licencing details see kernel-base/COPYING + */ + +#ifdef CONFIG_CPU_SUP_INTEL + +#include <asm/perf_event_p4.h> + +#define P4_CNTR_LIMIT 3 +/* + * array indices: 0,1 - HT threads, used with HT enabled cpu + */ +struct p4_event_bind { + unsigned int opcode; /* Event code and ESCR selector */ + unsigned int escr_msr[2]; /* ESCR MSR for this event */ + unsigned char cntr[2][P4_CNTR_LIMIT]; /* counter index (offset), -1 on abscence */ +}; + +struct p4_cache_event_bind { + unsigned int metric_pebs; + unsigned int metric_vert; +}; + +#define P4_GEN_CACHE_EVENT_BIND(name) \ + [P4_CACHE__##name] = { \ + .metric_pebs = P4_PEBS__##name, \ + .metric_vert = P4_VERT__##name, \ + } + +static struct p4_cache_event_bind p4_cache_event_bind_map[] = { + P4_GEN_CACHE_EVENT_BIND(1stl_cache_load_miss_retired), + P4_GEN_CACHE_EVENT_BIND(2ndl_cache_load_miss_retired), + P4_GEN_CACHE_EVENT_BIND(dtlb_load_miss_retired), + P4_GEN_CACHE_EVENT_BIND(dtlb_store_miss_retired), +}; + +/* + * Note that we don't use CCCR1 here, there is an + * exception for P4_BSQ_ALLOCATION but we just have + * no workaround + * + * consider this binding as resources which particular + * event may borrow, it doesn't contain EventMask, + * Tags and friends -- they are left to a caller + */ +static struct p4_event_bind p4_event_bind_map[] = { + [P4_EVENT_TC_DELIVER_MODE] = { + .opcode = P4_OPCODE(P4_EVENT_TC_DELIVER_MODE), + .escr_msr = { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_BPU_FETCH_REQUEST] = { + .opcode = P4_OPCODE(P4_EVENT_BPU_FETCH_REQUEST), + .escr_msr = { MSR_P4_BPU_ESCR0, MSR_P4_BPU_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_ITLB_REFERENCE] = { + .opcode = P4_OPCODE(P4_EVENT_ITLB_REFERENCE), + .escr_msr = { MSR_P4_ITLB_ESCR0, MSR_P4_ITLB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_MEMORY_CANCEL] = { + .opcode = P4_OPCODE(P4_EVENT_MEMORY_CANCEL), + .escr_msr = { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_MEMORY_COMPLETE] = { + .opcode = P4_OPCODE(P4_EVENT_MEMORY_COMPLETE), + .escr_msr = { MSR_P4_SAAT_ESCR0 , MSR_P4_SAAT_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_LOAD_PORT_REPLAY] = { + .opcode = P4_OPCODE(P4_EVENT_LOAD_PORT_REPLAY), + .escr_msr = { MSR_P4_SAAT_ESCR0, MSR_P4_SAAT_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_STORE_PORT_REPLAY] = { + .opcode = P4_OPCODE(P4_EVENT_STORE_PORT_REPLAY), + .escr_msr = { MSR_P4_SAAT_ESCR0 , MSR_P4_SAAT_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_MOB_LOAD_REPLAY] = { + .opcode = P4_OPCODE(P4_EVENT_MOB_LOAD_REPLAY), + .escr_msr = { MSR_P4_MOB_ESCR0, MSR_P4_MOB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_PAGE_WALK_TYPE] = { + .opcode = P4_OPCODE(P4_EVENT_PAGE_WALK_TYPE), + .escr_msr = { MSR_P4_PMH_ESCR0, MSR_P4_PMH_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_BSQ_CACHE_REFERENCE] = { + .opcode = P4_OPCODE(P4_EVENT_BSQ_CACHE_REFERENCE), + .escr_msr = { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_IOQ_ALLOCATION] = { + .opcode = P4_OPCODE(P4_EVENT_IOQ_ALLOCATION), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_IOQ_ACTIVE_ENTRIES] = { /* shared ESCR */ + .opcode = P4_OPCODE(P4_EVENT_IOQ_ACTIVE_ENTRIES), + .escr_msr = { MSR_P4_FSB_ESCR1, MSR_P4_FSB_ESCR1 }, + .cntr = { {2, -1, -1}, {3, -1, -1} }, + }, + [P4_EVENT_FSB_DATA_ACTIVITY] = { + .opcode = P4_OPCODE(P4_EVENT_FSB_DATA_ACTIVITY), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_BSQ_ALLOCATION] = { /* shared ESCR, broken CCCR1 */ + .opcode = P4_OPCODE(P4_EVENT_BSQ_ALLOCATION), + .escr_msr = { MSR_P4_BSU_ESCR0, MSR_P4_BSU_ESCR0 }, + .cntr = { {0, -1, -1}, {1, -1, -1} }, + }, + [P4_EVENT_BSQ_ACTIVE_ENTRIES] = { /* shared ESCR */ + .opcode = P4_OPCODE(P4_EVENT_BSQ_ACTIVE_ENTRIES), + .escr_msr = { MSR_P4_BSU_ESCR1 , MSR_P4_BSU_ESCR1 }, + .cntr = { {2, -1, -1}, {3, -1, -1} }, + }, + [P4_EVENT_SSE_INPUT_ASSIST] = { + .opcode = P4_OPCODE(P4_EVENT_SSE_INPUT_ASSIST), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_PACKED_SP_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_PACKED_SP_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_PACKED_DP_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_PACKED_DP_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_SCALAR_SP_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_SCALAR_SP_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_SCALAR_DP_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_SCALAR_DP_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_64BIT_MMX_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_64BIT_MMX_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_128BIT_MMX_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_128BIT_MMX_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_X87_FP_UOP] = { + .opcode = P4_OPCODE(P4_EVENT_X87_FP_UOP), + .escr_msr = { MSR_P4_FIRM_ESCR0, MSR_P4_FIRM_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_TC_MISC] = { + .opcode = P4_OPCODE(P4_EVENT_TC_MISC), + .escr_msr = { MSR_P4_TC_ESCR0, MSR_P4_TC_ESCR1 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_GLOBAL_POWER_EVENTS] = { + .opcode = P4_OPCODE(P4_EVENT_GLOBAL_POWER_EVENTS), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_TC_MS_XFER] = { + .opcode = P4_OPCODE(P4_EVENT_TC_MS_XFER), + .escr_msr = { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_UOP_QUEUE_WRITES] = { + .opcode = P4_OPCODE(P4_EVENT_UOP_QUEUE_WRITES), + .escr_msr = { MSR_P4_MS_ESCR0, MSR_P4_MS_ESCR1 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE] = { + .opcode = P4_OPCODE(P4_EVENT_RETIRED_MISPRED_BRANCH_TYPE), + .escr_msr = { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR0 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_RETIRED_BRANCH_TYPE] = { + .opcode = P4_OPCODE(P4_EVENT_RETIRED_BRANCH_TYPE), + .escr_msr = { MSR_P4_TBPU_ESCR0 , MSR_P4_TBPU_ESCR1 }, + .cntr = { {4, 5, -1}, {6, 7, -1} }, + }, + [P4_EVENT_RESOURCE_STALL] = { + .opcode = P4_OPCODE(P4_EVENT_RESOURCE_STALL), + .escr_msr = { MSR_P4_ALF_ESCR0, MSR_P4_ALF_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_WC_BUFFER] = { + .opcode = P4_OPCODE(P4_EVENT_WC_BUFFER), + .escr_msr = { MSR_P4_DAC_ESCR0, MSR_P4_DAC_ESCR1 }, + .cntr = { {8, 9, -1}, {10, 11, -1} }, + }, + [P4_EVENT_B2B_CYCLES] = { + .opcode = P4_OPCODE(P4_EVENT_B2B_CYCLES), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_BNR] = { + .opcode = P4_OPCODE(P4_EVENT_BNR), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_SNOOP] = { + .opcode = P4_OPCODE(P4_EVENT_SNOOP), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_RESPONSE] = { + .opcode = P4_OPCODE(P4_EVENT_RESPONSE), + .escr_msr = { MSR_P4_FSB_ESCR0, MSR_P4_FSB_ESCR1 }, + .cntr = { {0, -1, -1}, {2, -1, -1} }, + }, + [P4_EVENT_FRONT_END_EVENT] = { + .opcode = P4_OPCODE(P4_EVENT_FRONT_END_EVENT), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_EXECUTION_EVENT] = { + .opcode = P4_OPCODE(P4_EVENT_EXECUTION_EVENT), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_REPLAY_EVENT] = { + .opcode = P4_OPCODE(P4_EVENT_REPLAY_EVENT), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_INSTR_RETIRED] = { + .opcode = P4_OPCODE(P4_EVENT_INSTR_RETIRED), + .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_UOPS_RETIRED] = { + .opcode = P4_OPCODE(P4_EVENT_UOPS_RETIRED), + .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_UOP_TYPE] = { + .opcode = P4_OPCODE(P4_EVENT_UOP_TYPE), + .escr_msr = { MSR_P4_RAT_ESCR0, MSR_P4_RAT_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_BRANCH_RETIRED] = { + .opcode = P4_OPCODE(P4_EVENT_BRANCH_RETIRED), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_MISPRED_BRANCH_RETIRED] = { + .opcode = P4_OPCODE(P4_EVENT_MISPRED_BRANCH_RETIRED), + .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_X87_ASSIST] = { + .opcode = P4_OPCODE(P4_EVENT_X87_ASSIST), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_MACHINE_CLEAR] = { + .opcode = P4_OPCODE(P4_EVENT_MACHINE_CLEAR), + .escr_msr = { MSR_P4_CRU_ESCR2, MSR_P4_CRU_ESCR3 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, + [P4_EVENT_INSTR_COMPLETED] = { + .opcode = P4_OPCODE(P4_EVENT_INSTR_COMPLETED), + .escr_msr = { MSR_P4_CRU_ESCR0, MSR_P4_CRU_ESCR1 }, + .cntr = { {12, 13, 16}, {14, 15, 17} }, + }, +}; + +#define P4_GEN_CACHE_EVENT(event, bit, cache_event) \ + p4_config_pack_escr(P4_ESCR_EVENT(event) | \ + P4_ESCR_EMASK_BIT(event, bit)) | \ + p4_config_pack_cccr(cache_event | \ + P4_CCCR_ESEL(P4_OPCODE_ESEL(P4_OPCODE(event)))) + +static __initconst u64 p4_hw_cache_event_ids + [PERF_COUNT_HW_CACHE_MAX] + [PERF_COUNT_HW_CACHE_OP_MAX] + [PERF_COUNT_HW_CACHE_RESULT_MAX] = +{ + [ C(L1D ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS, + P4_CACHE__1stl_cache_load_miss_retired), + }, + }, + [ C(LL ) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS, + P4_CACHE__2ndl_cache_load_miss_retired), + }, +}, + [ C(DTLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS, + P4_CACHE__dtlb_load_miss_retired), + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = 0x0, + [ C(RESULT_MISS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_REPLAY_EVENT, NBOGUS, + P4_CACHE__dtlb_store_miss_retired), + }, + }, + [ C(ITLB) ] = { + [ C(OP_READ) ] = { + [ C(RESULT_ACCESS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_ITLB_REFERENCE, HIT, + P4_CACHE__itlb_reference_hit), + [ C(RESULT_MISS) ] = P4_GEN_CACHE_EVENT(P4_EVENT_ITLB_REFERENCE, MISS, + P4_CACHE__itlb_reference_miss), + }, + [ C(OP_WRITE) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + [ C(OP_PREFETCH) ] = { + [ C(RESULT_ACCESS) ] = -1, + [ C(RESULT_MISS) ] = -1, + }, + }, +}; + +static u64 p4_general_events[PERF_COUNT_HW_MAX] = { + /* non-halted CPU clocks */ + [PERF_COUNT_HW_CPU_CYCLES] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_GLOBAL_POWER_EVENTS) | + P4_ESCR_EMASK_BIT(P4_EVENT_GLOBAL_POWER_EVENTS, RUNNING)), + + /* + * retired instructions + * in a sake of simplicity we don't use the FSB tagging + */ + [PERF_COUNT_HW_INSTRUCTIONS] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_INSTR_RETIRED) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, NBOGUSNTAG) | + P4_ESCR_EMASK_BIT(P4_EVENT_INSTR_RETIRED, BOGUSNTAG)), + + /* cache hits */ + [PERF_COUNT_HW_CACHE_REFERENCES] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_BSQ_CACHE_REFERENCE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_HITM) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_HITM)), + + /* cache misses */ + [PERF_COUNT_HW_CACHE_MISSES] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_BSQ_CACHE_REFERENCE) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_2ndL_MISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, RD_3rdL_MISS) | + P4_ESCR_EMASK_BIT(P4_EVENT_BSQ_CACHE_REFERENCE, WR_2ndL_MISS)), + + /* branch instructions retired */ + [PERF_COUNT_HW_BRANCH_INSTRUCTIONS] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_RETIRED_BRANCH_TYPE) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CONDITIONAL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, CALL) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, RETURN) | + P4_ESCR_EMASK_BIT(P4_EVENT_RETIRED_BRANCH_TYPE, INDIRECT)), + + /* mispredicted branches retired */ + [PERF_COUNT_HW_BRANCH_MISSES] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_MISPRED_BRANCH_RETIRED) | + P4_ESCR_EMASK_BIT(P4_EVENT_MISPRED_BRANCH_RETIRED, NBOGUS)), + + /* bus ready clocks (cpu is driving #DRDY_DRV\#DRDY_OWN): */ + [PERF_COUNT_HW_BUS_CYCLES] = + p4_config_pack_escr(P4_ESCR_EVENT(P4_EVENT_FSB_DATA_ACTIVITY) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_DRV) | + P4_ESCR_EMASK_BIT(P4_EVENT_FSB_DATA_ACTIVITY, DRDY_OWN)) | + p4_config_pack_cccr(P4_CCCR_EDGE | P4_CCCR_COMPARE), +}; + +static struct p4_event_bind *p4_config_get_bind(u64 config) +{ + unsigned int evnt = p4_config_unpack_event(config); + struct p4_event_bind *bind = NULL; + + if (evnt < ARRAY_SIZE(p4_event_bind_map)) + bind = &p4_event_bind_map[evnt]; + + return bind; +} + +static u64 p4_pmu_event_map(int hw_event) +{ + struct p4_event_bind *bind; + unsigned int esel; + u64 config; + + if (hw_event > ARRAY_SIZE(p4_general_events)) { + printk_once(KERN_ERR "P4 PMU: Bad index: %i\n", hw_event); + return 0; + } + + config = p4_general_events[hw_event]; + bind = p4_config_get_bind(config); + esel = P4_OPCODE_ESEL(bind->opcode); + config |= p4_config_pack_cccr(P4_CCCR_ESEL(esel)); + + return config; +} + +/* + * We don't control raw events so it's up to the caller + * to pass sane values (and we don't count the thread number + * on HT machine but allow HT-compatible specifics to be + * passed on) + */ +static u64 p4_pmu_raw_event(u64 hw_event) +{ + return hw_event & + (p4_config_pack_escr(P4_ESCR_MASK_HT) | + p4_config_pack_cccr(P4_CCCR_MASK_HT)); +} + +static int p4_hw_config(struct perf_event_attr *attr, struct hw_perf_event *hwc) +{ + int cpu = raw_smp_processor_id(); + u32 escr, cccr; + + /* + * the reason we use cpu that early is that: if we get scheduled + * first time on the same cpu -- we will not need swap thread + * specific flags in config (and will save some cpu cycles) + */ + + cccr = p4_default_cccr_conf(cpu); + escr = p4_default_escr_conf(cpu, attr->exclude_kernel, attr->exclude_user); + hwc->config = p4_config_pack_escr(escr) | p4_config_pack_cccr(cccr); + + if (p4_ht_active() && p4_ht_thread(cpu)) + hwc->config = p4_set_ht_bit(hwc->config); + + return 0; +} + +static inline void p4_pmu_clear_cccr_ovf(struct hw_perf_event *hwc) +{ + unsigned long dummy; + + rdmsrl(hwc->config_base + hwc->idx, dummy); + if (dummy & P4_CCCR_OVF) { + (void)checking_wrmsrl(hwc->config_base + hwc->idx, + ((u64)dummy) & ~P4_CCCR_OVF); + } +} + +static inline void p4_pmu_disable_event(struct perf_event *event) +{ + struct hw_perf_event *hwc = &event->hw; + + /* + * If event gets disabled while counter is in overflowed + * state we need to clear P4_CCCR_OVF, otherwise interrupt get + * asserted again and again + */ + (void)checking_wrmsrl(hwc->config_base + hwc->idx, + (u64)(p4_config_unpack_cccr(hwc->config)) & + ~P4_CCCR_ENABLE & ~P4_CCCR_OVF & ~P4_CCCR_RESERVED); +} + +static void p4_pmu_disable_all(void) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int idx; + + for (idx = 0; idx < x86_pmu.num_events; idx++) { + struct perf_event *event = cpuc->events[idx]; + if (!test_bit(idx, cpuc->active_mask)) + continue; + p4_pmu_disable_event(event); + } +} + +static void p4_pmu_enable_event(struct perf_event *event) +{ + struct hw_perf_event *hwc = &event->hw; + int thread = p4_ht_config_thread(hwc->config); + u64 escr_conf = p4_config_unpack_escr(p4_clear_ht_bit(hwc->config)); + unsigned int idx = p4_config_unpack_event(hwc->config); + unsigned int idx_cache = p4_config_unpack_cache_event(hwc->config); + struct p4_event_bind *bind; + struct p4_cache_event_bind *bind_cache; + u64 escr_addr, cccr; + + bind = &p4_event_bind_map[idx]; + escr_addr = (u64)bind->escr_msr[thread]; + + /* + * - we dont support cascaded counters yet + * - and counter 1 is broken (erratum) + */ + WARN_ON_ONCE(p4_is_event_cascaded(hwc->config)); + WARN_ON_ONCE(hwc->idx == 1); + + /* we need a real Event value */ + escr_conf &= ~P4_ESCR_EVENT_MASK; + escr_conf |= P4_ESCR_EVENT(P4_OPCODE_EVNT(bind->opcode)); + + cccr = p4_config_unpack_cccr(hwc->config); + + /* + * it could be Cache event so that we need to + * set metrics into additional MSRs + */ + BUILD_BUG_ON(P4_CACHE__MAX > P4_CCCR_CACHE_OPS_MASK); + if (idx_cache > P4_CACHE__NONE && + idx_cache < ARRAY_SIZE(p4_cache_event_bind_map)) { + bind_cache = &p4_cache_event_bind_map[idx_cache]; + (void)checking_wrmsrl(MSR_IA32_PEBS_ENABLE, (u64)bind_cache->metric_pebs); + (void)checking_wrmsrl(MSR_P4_PEBS_MATRIX_VERT, (u64)bind_cache->metric_vert); + } + + (void)checking_wrmsrl(escr_addr, escr_conf); + (void)checking_wrmsrl(hwc->config_base + hwc->idx, + (cccr & ~P4_CCCR_RESERVED) | P4_CCCR_ENABLE); +} + +static void p4_pmu_enable_all(int added) +{ + struct cpu_hw_events *cpuc = &__get_cpu_var(cpu_hw_events); + int idx; + + for (idx = 0; idx < x86_pmu.num_events; idx++) { + struct perf_event *event = cpuc->events[idx]; + if (!test_bit(idx, cpuc->active_mask)) + continue; + p4_pmu_enable_event(event); + } +} + +static int p4_pmu_handle_irq(struct pt_regs *regs) +{ + struct perf_sample_data data; + struct cpu_hw_events *cpuc; + struct perf_event *event; + struct hw_perf_event *hwc; + int idx, handled = 0; + u64 val; + + data.addr = 0; + data.raw = NULL; + + cpuc = &__get_cpu_var(cpu_hw_events); + + for (idx = 0; idx < x86_pmu.num_events; idx++) { + + if (!test_bit(idx, cpuc->active_mask)) + continue; + + event = cpuc->events[idx]; + hwc = &event->hw; + + WARN_ON_ONCE(hwc->idx != idx); + + /* + * FIXME: Redundant call, actually not needed + * but just to check if we're screwed + */ + p4_pmu_clear_cccr_ovf(hwc); + + val = x86_perf_event_update(event); + if (val & (1ULL << (x86_pmu.event_bits - 1))) + continue; + + /* + * event overflow + */ + handled = 1; + data.period = event->hw.last_period; + + if (!x86_perf_event_set_period(event)) + continue; + if (perf_event_overflow(event, 1, &data, regs)) + p4_pmu_disable_event(event); + } + + if (handled) { + /* p4 quirk: unmask it again */ + apic_write(APIC_LVTPC, apic_read(APIC_LVTPC) & ~APIC_LVT_MASKED); + inc_irq_stat(apic_perf_irqs); + } + + return handled; +} + +/* + * swap thread specific fields according to a thread + * we are going to run on + */ +static void p4_pmu_swap_config_ts(struct hw_perf_event *hwc, int cpu) +{ + u32 escr, cccr; + + /* + * we either lucky and continue on same cpu or no HT support + */ + if (!p4_should_swap_ts(hwc->config, cpu)) + return; + + /* + * the event is migrated from an another logical + * cpu, so we need to swap thread specific flags + */ + + escr = p4_config_unpack_escr(hwc->config); + cccr = p4_config_unpack_cccr(hwc->config); + + if (p4_ht_thread(cpu)) { + cccr &= ~P4_CCCR_OVF_PMI_T0; + cccr |= P4_CCCR_OVF_PMI_T1; + if (escr & P4_ESCR_T0_OS) { + escr &= ~P4_ESCR_T0_OS; + escr |= P4_ESCR_T1_OS; + } + if (escr & P4_ESCR_T0_USR) { + escr &= ~P4_ESCR_T0_USR; + escr |= P4_ESCR_T1_USR; + } + hwc->config = p4_config_pack_escr(escr); + hwc->config |= p4_config_pack_cccr(cccr); + hwc->config |= P4_CONFIG_HT; + } else { + cccr &= ~P4_CCCR_OVF_PMI_T1; + cccr |= P4_CCCR_OVF_PMI_T0; + if (escr & P4_ESCR_T1_OS) { + escr &= ~P4_ESCR_T1_OS; + escr |= P4_ESCR_T0_OS; + } + if (escr & P4_ESCR_T1_USR) { + escr &= ~P4_ESCR_T1_USR; + escr |= P4_ESCR_T0_USR; + } + hwc->config = p4_config_pack_escr(escr); + hwc->config |= p4_config_pack_cccr(cccr); + hwc->config &= ~P4_CONFIG_HT; + } +} + +/* ESCRs are not sequential in memory so we need a map */ +static const unsigned int p4_escr_map[ARCH_P4_TOTAL_ESCR] = { + MSR_P4_ALF_ESCR0, /* 0 */ + MSR_P4_ALF_ESCR1, /* 1 */ + MSR_P4_BPU_ESCR0, /* 2 */ + MSR_P4_BPU_ESCR1, /* 3 */ + MSR_P4_BSU_ESCR0, /* 4 */ + MSR_P4_BSU_ESCR1, /* 5 */ + MSR_P4_CRU_ESCR0, /* 6 */ + MSR_P4_CRU_ESCR1, /* 7 */ + MSR_P4_CRU_ESCR2, /* 8 */ + MSR_P4_CRU_ESCR3, /* 9 */ + MSR_P4_CRU_ESCR4, /* 10 */ + MSR_P4_CRU_ESCR5, /* 11 */ + MSR_P4_DAC_ESCR0, /* 12 */ + MSR_P4_DAC_ESCR1, /* 13 */ + MSR_P4_FIRM_ESCR0, /* 14 */ + MSR_P4_FIRM_ESCR1, /* 15 */ + MSR_P4_FLAME_ESCR0, /* 16 */ + MSR_P4_FLAME_ESCR1, /* 17 */ + MSR_P4_FSB_ESCR0, /* 18 */ + MSR_P4_FSB_ESCR1, /* 19 */ + MSR_P4_IQ_ESCR0, /* 20 */ + MSR_P4_IQ_ESCR1, /* 21 */ + MSR_P4_IS_ESCR0, /* 22 */ + MSR_P4_IS_ESCR1, /* 23 */ + MSR_P4_ITLB_ESCR0, /* 24 */ + MSR_P4_ITLB_ESCR1, /* 25 */ + MSR_P4_IX_ESCR0, /* 26 */ + MSR_P4_IX_ESCR1, /* 27 */ + MSR_P4_MOB_ESCR0, /* 28 */ + MSR_P4_MOB_ESCR1, /* 29 */ + MSR_P4_MS_ESCR0, /* 30 */ + MSR_P4_MS_ESCR1, /* 31 */ + MSR_P4_PMH_ESCR0, /* 32 */ + MSR_P4_PMH_ESCR1, /* 33 */ + MSR_P4_RAT_ESCR0, /* 34 */ + MSR_P4_RAT_ESCR1, /* 35 */ + MSR_P4_SAAT_ESCR0, /* 36 */ + MSR_P4_SAAT_ESCR1, /* 37 */ + MSR_P4_SSU_ESCR0, /* 38 */ + MSR_P4_SSU_ESCR1, /* 39 */ + MSR_P4_TBPU_ESCR0, /* 40 */ + MSR_P4_TBPU_ESCR1, /* 41 */ + MSR_P4_TC_ESCR0, /* 42 */ + MSR_P4_TC_ESCR1, /* 43 */ + MSR_P4_U2L_ESCR0, /* 44 */ + MSR_P4_U2L_ESCR1, /* 45 */ +}; + +static int p4_get_escr_idx(unsigned int addr) +{ + unsigned int i; + + for (i = 0; i < ARRAY_SIZE(p4_escr_map); i++) { + if (addr == p4_escr_map[i]) + return i; + } + + return -1; +} + +static int p4_next_cntr(int thread, unsigned long *used_mask, + struct p4_event_bind *bind) +{ + int i = 0, j; + + for (i = 0; i < P4_CNTR_LIMIT; i++) { + j = bind->cntr[thread][i++]; + if (j == -1 || !test_bit(j, used_mask)) + return j; + } + + return -1; +} + +static int p4_pmu_schedule_events(struct cpu_hw_events *cpuc, int n, int *assign) +{ + unsigned long used_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)]; + unsigned long escr_mask[BITS_TO_LONGS(ARCH_P4_TOTAL_ESCR)]; + int cpu = raw_smp_processor_id(); + struct hw_perf_event *hwc; + struct p4_event_bind *bind; + unsigned int i, thread, num; + int cntr_idx, escr_idx; + + bitmap_zero(used_mask, X86_PMC_IDX_MAX); + bitmap_zero(escr_mask, ARCH_P4_TOTAL_ESCR); + + for (i = 0, num = n; i < n; i++, num--) { + + hwc = &cpuc->event_list[i]->hw; + thread = p4_ht_thread(cpu); + bind = p4_config_get_bind(hwc->config); + escr_idx = p4_get_escr_idx(bind->escr_msr[thread]); + + if (hwc->idx != -1 && !p4_should_swap_ts(hwc->config, cpu)) { + cntr_idx = hwc->idx; + if (assign) + assign[i] = hwc->idx; + goto reserve; + } + + cntr_idx = p4_next_cntr(thread, used_mask, bind); + if (cntr_idx == -1 || test_bit(escr_idx, escr_mask)) + goto done; + + p4_pmu_swap_config_ts(hwc, cpu); + if (assign) + assign[i] = cntr_idx; +reserve: + set_bit(cntr_idx, used_mask); + set_bit(escr_idx, escr_mask); + } + +done: + return num ? -ENOSPC : 0; +} + +static __initconst struct x86_pmu p4_pmu = { + .name = "Netburst P4/Xeon", + .handle_irq = p4_pmu_handle_irq, + .disable_all = p4_pmu_disable_all, + .enable_all = p4_pmu_enable_all, + .enable = p4_pmu_enable_event, + .disable = p4_pmu_disable_event, + .eventsel = MSR_P4_BPU_CCCR0, + .perfctr = MSR_P4_BPU_PERFCTR0, + .event_map = p4_pmu_event_map, + .raw_event = p4_pmu_raw_event, + .max_events = ARRAY_SIZE(p4_general_events), + .get_event_constraints = x86_get_event_constraints, + /* + * IF HT disabled we may need to use all + * ARCH_P4_MAX_CCCR counters simulaneously + * though leave it restricted at moment assuming + * HT is on + */ + .num_events = ARCH_P4_MAX_CCCR, + .apic = 1, + .event_bits = 40, + .event_mask = (1ULL << 40) - 1, + .max_period = (1ULL << 39) - 1, + .hw_config = p4_hw_config, + .schedule_events = p4_pmu_schedule_events, +}; + +static __init int p4_pmu_init(void) +{ + unsigned int low, high; + + /* If we get stripped -- indexig fails */ + BUILD_BUG_ON(ARCH_P4_MAX_CCCR > X86_PMC_MAX_GENERIC); + + rdmsr(MSR_IA32_MISC_ENABLE, low, high); + if (!(low & (1 << 7))) { + pr_cont("unsupported Netburst CPU model %d ", + boot_cpu_data.x86_model); + return -ENODEV; + } + + memcpy(hw_cache_event_ids, p4_hw_cache_event_ids, + sizeof(hw_cache_event_ids)); + + pr_cont("Netburst events, "); + + x86_pmu = p4_pmu; + + return 0; +} + +#endif /* CONFIG_CPU_SUP_INTEL */ diff --git a/arch/x86/kernel/cpu/perf_event_p6.c b/arch/x86/kernel/cpu/perf_event_p6.c index a330485d14da..877182c850df 100644 --- a/arch/x86/kernel/cpu/perf_event_p6.c +++ b/arch/x86/kernel/cpu/perf_event_p6.c @@ -66,7 +66,7 @@ static void p6_pmu_disable_all(void) wrmsrl(MSR_P6_EVNTSEL0, val); } -static void p6_pmu_enable_all(void) +static void p6_pmu_enable_all(int added) { unsigned long val; @@ -109,6 +109,8 @@ static __initconst struct x86_pmu p6_pmu = { .enable_all = p6_pmu_enable_all, .enable = p6_pmu_enable_event, .disable = p6_pmu_disable_event, + .hw_config = x86_hw_config, + .schedule_events = x86_schedule_events, .eventsel = MSR_P6_EVNTSEL0, .perfctr = MSR_P6_PERFCTR0, .event_map = p6_pmu_event_map, diff --git a/arch/x86/kernel/ds.c b/arch/x86/kernel/ds.c deleted file mode 100644 index 1c47390dd0e5..000000000000 --- a/arch/x86/kernel/ds.c +++ /dev/null @@ -1,1437 +0,0 @@ -/* - * Debug Store support - * - * This provides a low-level interface to the hardware's Debug Store - * feature that is used for branch trace store (BTS) and - * precise-event based sampling (PEBS). - * - * It manages: - * - DS and BTS hardware configuration - * - buffer overflow handling (to be done) - * - buffer access - * - * It does not do: - * - security checking (is the caller allowed to trace the task) - * - buffer allocation (memory accounting) - * - * - * Copyright (C) 2007-2009 Intel Corporation. - * Markus Metzger <markus.t.metzger@intel.com>, 2007-2009 - */ - -#include <linux/kernel.h> -#include <linux/string.h> -#include <linux/errno.h> -#include <linux/sched.h> -#include <linux/slab.h> -#include <linux/mm.h> -#include <linux/trace_clock.h> - -#include <asm/ds.h> - -#include "ds_selftest.h" - -/* - * The configuration for a particular DS hardware implementation: - */ -struct ds_configuration { - /* The name of the configuration: */ - const char *name; - - /* The size of pointer-typed fields in DS, BTS, and PEBS: */ - unsigned char sizeof_ptr_field; - - /* The size of a BTS/PEBS record in bytes: */ - unsigned char sizeof_rec[2]; - - /* The number of pebs counter reset values in the DS structure. */ - unsigned char nr_counter_reset; - - /* Control bit-masks indexed by enum ds_feature: */ - unsigned long ctl[dsf_ctl_max]; -}; -static struct ds_configuration ds_cfg __read_mostly; - - -/* Maximal size of a DS configuration: */ -#define MAX_SIZEOF_DS 0x80 - -/* Maximal size of a BTS record: */ -#define MAX_SIZEOF_BTS (3 * 8) - -/* BTS and PEBS buffer alignment: */ -#define DS_ALIGNMENT (1 << 3) - -/* Number of buffer pointers in DS: */ -#define NUM_DS_PTR_FIELDS 8 - -/* Size of a pebs reset value in DS: */ -#define PEBS_RESET_FIELD_SIZE 8 - -/* Mask of control bits in the DS MSR register: */ -#define BTS_CONTROL \ - ( ds_cfg.ctl[dsf_bts] | \ - ds_cfg.ctl[dsf_bts_kernel] | \ - ds_cfg.ctl[dsf_bts_user] | \ - ds_cfg.ctl[dsf_bts_overflow] ) - -/* - * A BTS or PEBS tracer. - * - * This holds the configuration of the tracer and serves as a handle - * to identify tracers. - */ -struct ds_tracer { - /* The DS context (partially) owned by this tracer. */ - struct ds_context *context; - /* The buffer provided on ds_request() and its size in bytes. */ - void *buffer; - size_t size; -}; - -struct bts_tracer { - /* The common DS part: */ - struct ds_tracer ds; - - /* The trace including the DS configuration: */ - struct bts_trace trace; - - /* Buffer overflow notification function: */ - bts_ovfl_callback_t ovfl; - - /* Active flags affecting trace collection. */ - unsigned int flags; -}; - -struct pebs_tracer { - /* The common DS part: */ - struct ds_tracer ds; - - /* The trace including the DS configuration: */ - struct pebs_trace trace; - - /* Buffer overflow notification function: */ - pebs_ovfl_callback_t ovfl; -}; - -/* - * Debug Store (DS) save area configuration (see Intel64 and IA32 - * Architectures Software Developer's Manual, section 18.5) - * - * The DS configuration consists of the following fields; different - * architetures vary in the size of those fields. - * - * - double-word aligned base linear address of the BTS buffer - * - write pointer into the BTS buffer - * - end linear address of the BTS buffer (one byte beyond the end of - * the buffer) - * - interrupt pointer into BTS buffer - * (interrupt occurs when write pointer passes interrupt pointer) - * - double-word aligned base linear address of the PEBS buffer - * - write pointer into the PEBS buffer - * - end linear address of the PEBS buffer (one byte beyond the end of - * the buffer) - * - interrupt pointer into PEBS buffer - * (interrupt occurs when write pointer passes interrupt pointer) - * - value to which counter is reset following counter overflow - * - * Later architectures use 64bit pointers throughout, whereas earlier - * architectures use 32bit pointers in 32bit mode. - * - * - * We compute the base address for the first 8 fields based on: - * - the field size stored in the DS configuration - * - the relative field position - * - an offset giving the start of the respective region - * - * This offset is further used to index various arrays holding - * information for BTS and PEBS at the respective index. - * - * On later 32bit processors, we only access the lower 32bit of the - * 64bit pointer fields. The upper halves will be zeroed out. - */ - -enum ds_field { - ds_buffer_base = 0, - ds_index, - ds_absolute_maximum, - ds_interrupt_threshold, -}; - -enum ds_qualifier { - ds_bts = 0, - ds_pebs -}; - -static inline unsigned long -ds_get(const unsigned char *base, enum ds_qualifier qual, enum ds_field field) -{ - base += (ds_cfg.sizeof_ptr_field * (field + (4 * qual))); - return *(unsigned long *)base; -} - -static inline void -ds_set(unsigned char *base, enum ds_qualifier qual, enum ds_field field, - unsigned long value) -{ - base += (ds_cfg.sizeof_ptr_field * (field + (4 * qual))); - (*(unsigned long *)base) = value; -} - - -/* - * Locking is done only for allocating BTS or PEBS resources. - */ -static DEFINE_SPINLOCK(ds_lock); - -/* - * We either support (system-wide) per-cpu or per-thread allocation. - * We distinguish the two based on the task_struct pointer, where a - * NULL pointer indicates per-cpu allocation for the current cpu. - * - * Allocations are use-counted. As soon as resources are allocated, - * further allocations must be of the same type (per-cpu or - * per-thread). We model this by counting allocations (i.e. the number - * of tracers of a certain type) for one type negatively: - * =0 no tracers - * >0 number of per-thread tracers - * <0 number of per-cpu tracers - * - * Tracers essentially gives the number of ds contexts for a certain - * type of allocation. - */ -static atomic_t tracers = ATOMIC_INIT(0); - -static inline int get_tracer(struct task_struct *task) -{ - int error; - - spin_lock_irq(&ds_lock); - - if (task) { - error = -EPERM; - if (atomic_read(&tracers) < 0) - goto out; - atomic_inc(&tracers); - } else { - error = -EPERM; - if (atomic_read(&tracers) > 0) - goto out; - atomic_dec(&tracers); - } - - error = 0; -out: - spin_unlock_irq(&ds_lock); - return error; -} - -static inline void put_tracer(struct task_struct *task) -{ - if (task) - atomic_dec(&tracers); - else - atomic_inc(&tracers); -} - -/* - * The DS context is either attached to a thread or to a cpu: - * - in the former case, the thread_struct contains a pointer to the - * attached context. - * - in the latter case, we use a static array of per-cpu context - * pointers. - * - * Contexts are use-counted. They are allocated on first access and - * deallocated when the last user puts the context. - */ -struct ds_context { - /* The DS configuration; goes into MSR_IA32_DS_AREA: */ - unsigned char ds[MAX_SIZEOF_DS]; - - /* The owner of the BTS and PEBS configuration, respectively: */ - struct bts_tracer *bts_master; - struct pebs_tracer *pebs_master; - - /* Use count: */ - unsigned long count; - - /* Pointer to the context pointer field: */ - struct ds_context **this; - - /* The traced task; NULL for cpu tracing: */ - struct task_struct *task; - - /* The traced cpu; only valid if task is NULL: */ - int cpu; -}; - -static DEFINE_PER_CPU(struct ds_context *, cpu_ds_context); - - -static struct ds_context *ds_get_context(struct task_struct *task, int cpu) -{ - struct ds_context **p_context = - (task ? &task->thread.ds_ctx : &per_cpu(cpu_ds_context, cpu)); - struct ds_context *context = NULL; - struct ds_context *new_context = NULL; - - /* Chances are small that we already have a context. */ - new_context = kzalloc(sizeof(*new_context), GFP_KERNEL); - if (!new_context) - return NULL; - - spin_lock_irq(&ds_lock); - - context = *p_context; - if (likely(!context)) { - context = new_context; - - context->this = p_context; - context->task = task; - context->cpu = cpu; - context->count = 0; - - *p_context = context; - } - - context->count++; - - spin_unlock_irq(&ds_lock); - - if (context != new_context) - kfree(new_context); - - return context; -} - -static void ds_put_context(struct ds_context *context) -{ - struct task_struct *task; - unsigned long irq; - - if (!context) - return; - - spin_lock_irqsave(&ds_lock, irq); - - if (--context->count) { - spin_unlock_irqrestore(&ds_lock, irq); - return; - } - - *(context->this) = NULL; - - task = context->task; - - if (task) - clear_tsk_thread_flag(task, TIF_DS_AREA_MSR); - - /* - * We leave the (now dangling) pointer to the DS configuration in - * the DS_AREA msr. This is as good or as bad as replacing it with - * NULL - the hardware would crash if we enabled tracing. - * - * This saves us some problems with having to write an msr on a - * different cpu while preventing others from doing the same for the - * next context for that same cpu. - */ - - spin_unlock_irqrestore(&ds_lock, irq); - - /* The context might still be in use for context switching. */ - if (task && (task != current)) - wait_task_context_switch(task); - - kfree(context); -} - -static void ds_install_ds_area(struct ds_context *context) -{ - unsigned long ds; - - ds = (unsigned long)context->ds; - - /* - * There is a race between the bts master and the pebs master. - * - * The thread/cpu access is synchronized via get/put_cpu() for - * task tracing and via wrmsr_on_cpu for cpu tracing. - * - * If bts and pebs are collected for the same task or same cpu, - * the same confiuration is written twice. - */ - if (context->task) { - get_cpu(); - if (context->task == current) - wrmsrl(MSR_IA32_DS_AREA, ds); - set_tsk_thread_flag(context->task, TIF_DS_AREA_MSR); - put_cpu(); - } else - wrmsr_on_cpu(context->cpu, MSR_IA32_DS_AREA, - (u32)((u64)ds), (u32)((u64)ds >> 32)); -} - -/* - * Call the tracer's callback on a buffer overflow. - * - * context: the ds context - * qual: the buffer type - */ -static void ds_overflow(struct ds_context *context, enum ds_qualifier qual) -{ - switch (qual) { - case ds_bts: - if (context->bts_master && - context->bts_master->ovfl) - context->bts_master->ovfl(context->bts_master); - break; - case ds_pebs: - if (context->pebs_master && - context->pebs_master->ovfl) - context->pebs_master->ovfl(context->pebs_master); - break; - } -} - - -/* - * Write raw data into the BTS or PEBS buffer. - * - * The remainder of any partially written record is zeroed out. - * - * context: the DS context - * qual: the buffer type - * record: the data to write - * size: the size of the data - */ -static int ds_write(struct ds_context *context, enum ds_qualifier qual, - const void *record, size_t size) -{ - int bytes_written = 0; - - if (!record) - return -EINVAL; - - while (size) { - unsigned long base, index, end, write_end, int_th; - unsigned long write_size, adj_write_size; - - /* - * Write as much as possible without producing an - * overflow interrupt. - * - * Interrupt_threshold must either be - * - bigger than absolute_maximum or - * - point to a record between buffer_base and absolute_maximum - * - * Index points to a valid record. - */ - base = ds_get(context->ds, qual, ds_buffer_base); - index = ds_get(context->ds, qual, ds_index); - end = ds_get(context->ds, qual, ds_absolute_maximum); - int_th = ds_get(context->ds, qual, ds_interrupt_threshold); - - write_end = min(end, int_th); - - /* - * If we are already beyond the interrupt threshold, - * we fill the entire buffer. - */ - if (write_end <= index) - write_end = end; - - if (write_end <= index) - break; - - write_size = min((unsigned long) size, write_end - index); - memcpy((void *)index, record, write_size); - - record = (const char *)record + write_size; - size -= write_size; - bytes_written += write_size; - - adj_write_size = write_size / ds_cfg.sizeof_rec[qual]; - adj_write_size *= ds_cfg.sizeof_rec[qual]; - - /* Zero out trailing bytes. */ - memset((char *)index + write_size, 0, - adj_write_size - write_size); - index += adj_write_size; - - if (index >= end) - index = base; - ds_set(context->ds, qual, ds_index, index); - - if (index >= int_th) - ds_overflow(context, qual); - } - - return bytes_written; -} - - -/* - * Branch Trace Store (BTS) uses the following format. Different - * architectures vary in the size of those fields. - * - source linear address - * - destination linear address - * - flags - * - * Later architectures use 64bit pointers throughout, whereas earlier - * architectures use 32bit pointers in 32bit mode. - * - * We compute the base address for the fields based on: - * - the field size stored in the DS configuration - * - the relative field position - * - * In order to store additional information in the BTS buffer, we use - * a special source address to indicate that the record requires - * special interpretation. - * - * Netburst indicated via a bit in the flags field whether the branch - * was predicted; this is ignored. - * - * We use two levels of abstraction: - * - the raw data level defined here - * - an arch-independent level defined in ds.h - */ - -enum bts_field { - bts_from, - bts_to, - bts_flags, - - bts_qual = bts_from, - bts_clock = bts_to, - bts_pid = bts_flags, - - bts_qual_mask = (bts_qual_max - 1), - bts_escape = ((unsigned long)-1 & ~bts_qual_mask) -}; - -static inline unsigned long bts_get(const char *base, unsigned long field) -{ - base += (ds_cfg.sizeof_ptr_field * field); - return *(unsigned long *)base; -} - -static inline void bts_set(char *base, unsigned long field, unsigned long val) -{ - base += (ds_cfg.sizeof_ptr_field * field); - (*(unsigned long *)base) = val; -} - - -/* - * The raw BTS data is architecture dependent. - * - * For higher-level users, we give an arch-independent view. - * - ds.h defines struct bts_struct - * - bts_read translates one raw bts record into a bts_struct - * - bts_write translates one bts_struct into the raw format and - * writes it into the top of the parameter tracer's buffer. - * - * return: bytes read/written on success; -Eerrno, otherwise - */ -static int -bts_read(struct bts_tracer *tracer, const void *at, struct bts_struct *out) -{ - if (!tracer) - return -EINVAL; - - if (at < tracer->trace.ds.begin) - return -EINVAL; - - if (tracer->trace.ds.end < (at + tracer->trace.ds.size)) - return -EINVAL; - - memset(out, 0, sizeof(*out)); - if ((bts_get(at, bts_qual) & ~bts_qual_mask) == bts_escape) { - out->qualifier = (bts_get(at, bts_qual) & bts_qual_mask); - out->variant.event.clock = bts_get(at, bts_clock); - out->variant.event.pid = bts_get(at, bts_pid); - } else { - out->qualifier = bts_branch; - out->variant.lbr.from = bts_get(at, bts_from); - out->variant.lbr.to = bts_get(at, bts_to); - - if (!out->variant.lbr.from && !out->variant.lbr.to) - out->qualifier = bts_invalid; - } - - return ds_cfg.sizeof_rec[ds_bts]; -} - -static int bts_write(struct bts_tracer *tracer, const struct bts_struct *in) -{ - unsigned char raw[MAX_SIZEOF_BTS]; - - if (!tracer) - return -EINVAL; - - if (MAX_SIZEOF_BTS < ds_cfg.sizeof_rec[ds_bts]) - return -EOVERFLOW; - - switch (in->qualifier) { - case bts_invalid: - bts_set(raw, bts_from, 0); - bts_set(raw, bts_to, 0); - bts_set(raw, bts_flags, 0); - break; - case bts_branch: - bts_set(raw, bts_from, in->variant.lbr.from); - bts_set(raw, bts_to, in->variant.lbr.to); - bts_set(raw, bts_flags, 0); - break; - case bts_task_arrives: - case bts_task_departs: - bts_set(raw, bts_qual, (bts_escape | in->qualifier)); - bts_set(raw, bts_clock, in->variant.event.clock); - bts_set(raw, bts_pid, in->variant.event.pid); - break; - default: - return -EINVAL; - } - - return ds_write(tracer->ds.context, ds_bts, raw, - ds_cfg.sizeof_rec[ds_bts]); -} - - -static void ds_write_config(struct ds_context *context, - struct ds_trace *cfg, enum ds_qualifier qual) -{ - unsigned char *ds = context->ds; - - ds_set(ds, qual, ds_buffer_base, (unsigned long)cfg->begin); - ds_set(ds, qual, ds_index, (unsigned long)cfg->top); - ds_set(ds, qual, ds_absolute_maximum, (unsigned long)cfg->end); - ds_set(ds, qual, ds_interrupt_threshold, (unsigned long)cfg->ith); -} - -static void ds_read_config(struct ds_context *context, - struct ds_trace *cfg, enum ds_qualifier qual) -{ - unsigned char *ds = context->ds; - - cfg->begin = (void *)ds_get(ds, qual, ds_buffer_base); - cfg->top = (void *)ds_get(ds, qual, ds_index); - cfg->end = (void *)ds_get(ds, qual, ds_absolute_maximum); - cfg->ith = (void *)ds_get(ds, qual, ds_interrupt_threshold); -} - -static void ds_init_ds_trace(struct ds_trace *trace, enum ds_qualifier qual, - void *base, size_t size, size_t ith, - unsigned int flags) { - unsigned long buffer, adj; - - /* - * Adjust the buffer address and size to meet alignment - * constraints: - * - buffer is double-word aligned - * - size is multiple of record size - * - * We checked the size at the very beginning; we have enough - * space to do the adjustment. - */ - buffer = (unsigned long)base; - - adj = ALIGN(buffer, DS_ALIGNMENT) - buffer; - buffer += adj; - size -= adj; - - trace->n = size / ds_cfg.sizeof_rec[qual]; - trace->size = ds_cfg.sizeof_rec[qual]; - - size = (trace->n * trace->size); - - trace->begin = (void *)buffer; - trace->top = trace->begin; - trace->end = (void *)(buffer + size); - /* - * The value for 'no threshold' is -1, which will set the - * threshold outside of the buffer, just like we want it. - */ - ith *= ds_cfg.sizeof_rec[qual]; - trace->ith = (void *)(buffer + size - ith); - - trace->flags = flags; -} - - -static int ds_request(struct ds_tracer *tracer, struct ds_trace *trace, - enum ds_qualifier qual, struct task_struct *task, - int cpu, void *base, size_t size, size_t th) -{ - struct ds_context *context; - int error; - size_t req_size; - - error = -EOPNOTSUPP; - if (!ds_cfg.sizeof_rec[qual]) - goto out; - - error = -EINVAL; - if (!base) - goto out; - - req_size = ds_cfg.sizeof_rec[qual]; - /* We might need space for alignment adjustments. */ - if (!IS_ALIGNED((unsigned long)base, DS_ALIGNMENT)) - req_size += DS_ALIGNMENT; - - error = -EINVAL; - if (size < req_size) - goto out; - - if (th != (size_t)-1) { - th *= ds_cfg.sizeof_rec[qual]; - - error = -EINVAL; - if (size <= th) - goto out; - } - - tracer->buffer = base; - tracer->size = size; - - error = -ENOMEM; - context = ds_get_context(task, cpu); - if (!context) - goto out; - tracer->context = context; - - /* - * Defer any tracer-specific initialization work for the context until - * context ownership has been clarified. - */ - - error = 0; - out: - return error; -} - -static struct bts_tracer *ds_request_bts(struct task_struct *task, int cpu, - void *base, size_t size, - bts_ovfl_callback_t ovfl, size_t th, - unsigned int flags) -{ - struct bts_tracer *tracer; - int error; - - /* Buffer overflow notification is not yet implemented. */ - error = -EOPNOTSUPP; - if (ovfl) - goto out; - - error = get_tracer(task); - if (error < 0) - goto out; - - error = -ENOMEM; - tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); - if (!tracer) - goto out_put_tracer; - tracer->ovfl = ovfl; - - /* Do some more error checking and acquire a tracing context. */ - error = ds_request(&tracer->ds, &tracer->trace.ds, - ds_bts, task, cpu, base, size, th); - if (error < 0) - goto out_tracer; - - /* Claim the bts part of the tracing context we acquired above. */ - spin_lock_irq(&ds_lock); - - error = -EPERM; - if (tracer->ds.context->bts_master) - goto out_unlock; - tracer->ds.context->bts_master = tracer; - - spin_unlock_irq(&ds_lock); - - /* - * Now that we own the bts part of the context, let's complete the - * initialization for that part. - */ - ds_init_ds_trace(&tracer->trace.ds, ds_bts, base, size, th, flags); - ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_bts); - ds_install_ds_area(tracer->ds.context); - - tracer->trace.read = bts_read; - tracer->trace.write = bts_write; - - /* Start tracing. */ - ds_resume_bts(tracer); - - return tracer; - - out_unlock: - spin_unlock_irq(&ds_lock); - ds_put_context(tracer->ds.context); - out_tracer: - kfree(tracer); - out_put_tracer: - put_tracer(task); - out: - return ERR_PTR(error); -} - -struct bts_tracer *ds_request_bts_task(struct task_struct *task, - void *base, size_t size, - bts_ovfl_callback_t ovfl, - size_t th, unsigned int flags) -{ - return ds_request_bts(task, 0, base, size, ovfl, th, flags); -} - -struct bts_tracer *ds_request_bts_cpu(int cpu, void *base, size_t size, - bts_ovfl_callback_t ovfl, - size_t th, unsigned int flags) -{ - return ds_request_bts(NULL, cpu, base, size, ovfl, th, flags); -} - -static struct pebs_tracer *ds_request_pebs(struct task_struct *task, int cpu, - void *base, size_t size, - pebs_ovfl_callback_t ovfl, size_t th, - unsigned int flags) -{ - struct pebs_tracer *tracer; - int error; - - /* Buffer overflow notification is not yet implemented. */ - error = -EOPNOTSUPP; - if (ovfl) - goto out; - - error = get_tracer(task); - if (error < 0) - goto out; - - error = -ENOMEM; - tracer = kzalloc(sizeof(*tracer), GFP_KERNEL); - if (!tracer) - goto out_put_tracer; - tracer->ovfl = ovfl; - - /* Do some more error checking and acquire a tracing context. */ - error = ds_request(&tracer->ds, &tracer->trace.ds, - ds_pebs, task, cpu, base, size, th); - if (error < 0) - goto out_tracer; - - /* Claim the pebs part of the tracing context we acquired above. */ - spin_lock_irq(&ds_lock); - - error = -EPERM; - if (tracer->ds.context->pebs_master) - goto out_unlock; - tracer->ds.context->pebs_master = tracer; - - spin_unlock_irq(&ds_lock); - - /* - * Now that we own the pebs part of the context, let's complete the - * initialization for that part. - */ - ds_init_ds_trace(&tracer->trace.ds, ds_pebs, base, size, th, flags); - ds_write_config(tracer->ds.context, &tracer->trace.ds, ds_pebs); - ds_install_ds_area(tracer->ds.context); - - /* Start tracing. */ - ds_resume_pebs(tracer); - - return tracer; - - out_unlock: - spin_unlock_irq(&ds_lock); - ds_put_context(tracer->ds.context); - out_tracer: - kfree(tracer); - out_put_tracer: - put_tracer(task); - out: - return ERR_PTR(error); -} - -struct pebs_tracer *ds_request_pebs_task(struct task_struct *task, - void *base, size_t size, - pebs_ovfl_callback_t ovfl, - size_t th, unsigned int flags) -{ - return ds_request_pebs(task, 0, base, size, ovfl, th, flags); -} - -struct pebs_tracer *ds_request_pebs_cpu(int cpu, void *base, size_t size, - pebs_ovfl_callback_t ovfl, - size_t th, unsigned int flags) -{ - return ds_request_pebs(NULL, cpu, base, size, ovfl, th, flags); -} - -static void ds_free_bts(struct bts_tracer *tracer) -{ - struct task_struct *task; - - task = tracer->ds.context->task; - - WARN_ON_ONCE(tracer->ds.context->bts_master != tracer); - tracer->ds.context->bts_master = NULL; - - /* Make sure tracing stopped and the tracer is not in use. */ - if (task && (task != current)) - wait_task_context_switch(task); - - ds_put_context(tracer->ds.context); - put_tracer(task); - - kfree(tracer); -} - -void ds_release_bts(struct bts_tracer *tracer) -{ - might_sleep(); - - if (!tracer) - return; - - ds_suspend_bts(tracer); - ds_free_bts(tracer); -} - -int ds_release_bts_noirq(struct bts_tracer *tracer) -{ - struct task_struct *task; - unsigned long irq; - int error; - - if (!tracer) - return 0; - - task = tracer->ds.context->task; - - local_irq_save(irq); - - error = -EPERM; - if (!task && - (tracer->ds.context->cpu != smp_processor_id())) - goto out; - - error = -EPERM; - if (task && (task != current)) - goto out; - - ds_suspend_bts_noirq(tracer); - ds_free_bts(tracer); - - error = 0; - out: - local_irq_restore(irq); - return error; -} - -static void update_task_debugctlmsr(struct task_struct *task, - unsigned long debugctlmsr) -{ - task->thread.debugctlmsr = debugctlmsr; - - get_cpu(); - if (task == current) - update_debugctlmsr(debugctlmsr); - put_cpu(); -} - -void ds_suspend_bts(struct bts_tracer *tracer) -{ - struct task_struct *task; - unsigned long debugctlmsr; - int cpu; - - if (!tracer) - return; - - tracer->flags = 0; - - task = tracer->ds.context->task; - cpu = tracer->ds.context->cpu; - - WARN_ON(!task && irqs_disabled()); - - debugctlmsr = (task ? - task->thread.debugctlmsr : - get_debugctlmsr_on_cpu(cpu)); - debugctlmsr &= ~BTS_CONTROL; - - if (task) - update_task_debugctlmsr(task, debugctlmsr); - else - update_debugctlmsr_on_cpu(cpu, debugctlmsr); -} - -int ds_suspend_bts_noirq(struct bts_tracer *tracer) -{ - struct task_struct *task; - unsigned long debugctlmsr, irq; - int cpu, error = 0; - - if (!tracer) - return 0; - - tracer->flags = 0; - - task = tracer->ds.context->task; - cpu = tracer->ds.context->cpu; - - local_irq_save(irq); - - error = -EPERM; - if (!task && (cpu != smp_processor_id())) - goto out; - - debugctlmsr = (task ? - task->thread.debugctlmsr : - get_debugctlmsr()); - debugctlmsr &= ~BTS_CONTROL; - - if (task) - update_task_debugctlmsr(task, debugctlmsr); - else - update_debugctlmsr(debugctlmsr); - - error = 0; - out: - local_irq_restore(irq); - return error; -} - -static unsigned long ds_bts_control(struct bts_tracer *tracer) -{ - unsigned long control; - - control = ds_cfg.ctl[dsf_bts]; - if (!(tracer->trace.ds.flags & BTS_KERNEL)) - control |= ds_cfg.ctl[dsf_bts_kernel]; - if (!(tracer->trace.ds.flags & BTS_USER)) - control |= ds_cfg.ctl[dsf_bts_user]; - - return control; -} - -void ds_resume_bts(struct bts_tracer *tracer) -{ - struct task_struct *task; - unsigned long debugctlmsr; - int cpu; - - if (!tracer) - return; - - tracer->flags = tracer->trace.ds.flags; - - task = tracer->ds.context->task; - cpu = tracer->ds.context->cpu; - - WARN_ON(!task && irqs_disabled()); - - debugctlmsr = (task ? - task->thread.debugctlmsr : - get_debugctlmsr_on_cpu(cpu)); - debugctlmsr |= ds_bts_control(tracer); - - if (task) - update_task_debugctlmsr(task, debugctlmsr); - else - update_debugctlmsr_on_cpu(cpu, debugctlmsr); -} - -int ds_resume_bts_noirq(struct bts_tracer *tracer) -{ - struct task_struct *task; - unsigned long debugctlmsr, irq; - int cpu, error = 0; - - if (!tracer) - return 0; - - tracer->flags = tracer->trace.ds.flags; - - task = tracer->ds.context->task; - cpu = tracer->ds.context->cpu; - - local_irq_save(irq); - - error = -EPERM; - if (!task && (cpu != smp_processor_id())) - goto out; - - debugctlmsr = (task ? - task->thread.debugctlmsr : - get_debugctlmsr()); - debugctlmsr |= ds_bts_control(tracer); - - if (task) - update_task_debugctlmsr(task, debugctlmsr); - else - update_debugctlmsr(debugctlmsr); - - error = 0; - out: - local_irq_restore(irq); - return error; -} - -static void ds_free_pebs(struct pebs_tracer *tracer) -{ - struct task_struct *task; - - task = tracer->ds.context->task; - - WARN_ON_ONCE(tracer->ds.context->pebs_master != tracer); - tracer->ds.context->pebs_master = NULL; - - ds_put_context(tracer->ds.context); - put_tracer(task); - - kfree(tracer); -} - -void ds_release_pebs(struct pebs_tracer *tracer) -{ - might_sleep(); - - if (!tracer) - return; - - ds_suspend_pebs(tracer); - ds_free_pebs(tracer); -} - -int ds_release_pebs_noirq(struct pebs_tracer *tracer) -{ - struct task_struct *task; - unsigned long irq; - int error; - - if (!tracer) - return 0; - - task = tracer->ds.context->task; - - local_irq_save(irq); - - error = -EPERM; - if (!task && - (tracer->ds.context->cpu != smp_processor_id())) - goto out; - - error = -EPERM; - if (task && (task != current)) - goto out; - - ds_suspend_pebs_noirq(tracer); - ds_free_pebs(tracer); - - error = 0; - out: - local_irq_restore(irq); - return error; -} - -void ds_suspend_pebs(struct pebs_tracer *tracer) -{ - -} - -int ds_suspend_pebs_noirq(struct pebs_tracer *tracer) -{ - return 0; -} - -void ds_resume_pebs(struct pebs_tracer *tracer) -{ - -} - -int ds_resume_pebs_noirq(struct pebs_tracer *tracer) -{ - return 0; -} - -const struct bts_trace *ds_read_bts(struct bts_tracer *tracer) -{ - if (!tracer) - return NULL; - - ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_bts); - return &tracer->trace; -} - -const struct pebs_trace *ds_read_pebs(struct pebs_tracer *tracer) -{ - if (!tracer) - return NULL; - - ds_read_config(tracer->ds.context, &tracer->trace.ds, ds_pebs); - - tracer->trace.counters = ds_cfg.nr_counter_reset; - memcpy(tracer->trace.counter_reset, - tracer->ds.context->ds + - (NUM_DS_PTR_FIELDS * ds_cfg.sizeof_ptr_field), - ds_cfg.nr_counter_reset * PEBS_RESET_FIELD_SIZE); - - return &tracer->trace; -} - -int ds_reset_bts(struct bts_tracer *tracer) -{ - if (!tracer) - return -EINVAL; - - tracer->trace.ds.top = tracer->trace.ds.begin; - - ds_set(tracer->ds.context->ds, ds_bts, ds_index, - (unsigned long)tracer->trace.ds.top); - - return 0; -} - -int ds_reset_pebs(struct pebs_tracer *tracer) -{ - if (!tracer) - return -EINVAL; - - tracer->trace.ds.top = tracer->trace.ds.begin; - - ds_set(tracer->ds.context->ds, ds_pebs, ds_index, - (unsigned long)tracer->trace.ds.top); - - return 0; -} - -int ds_set_pebs_reset(struct pebs_tracer *tracer, - unsigned int counter, u64 value) -{ - if (!tracer) - return -EINVAL; - - if (ds_cfg.nr_counter_reset < counter) - return -EINVAL; - - *(u64 *)(tracer->ds.context->ds + - (NUM_DS_PTR_FIELDS * ds_cfg.sizeof_ptr_field) + - (counter * PEBS_RESET_FIELD_SIZE)) = value; - - return 0; -} - -static const struct ds_configuration ds_cfg_netburst = { - .name = "Netburst", - .ctl[dsf_bts] = (1 << 2) | (1 << 3), - .ctl[dsf_bts_kernel] = (1 << 5), - .ctl[dsf_bts_user] = (1 << 6), - .nr_counter_reset = 1, -}; -static const struct ds_configuration ds_cfg_pentium_m = { - .name = "Pentium M", - .ctl[dsf_bts] = (1 << 6) | (1 << 7), - .nr_counter_reset = 1, -}; -static const struct ds_configuration ds_cfg_core2_atom = { - .name = "Core 2/Atom", - .ctl[dsf_bts] = (1 << 6) | (1 << 7), - .ctl[dsf_bts_kernel] = (1 << 9), - .ctl[dsf_bts_user] = (1 << 10), - .nr_counter_reset = 1, -}; -static const struct ds_configuration ds_cfg_core_i7 = { - .name = "Core i7", - .ctl[dsf_bts] = (1 << 6) | (1 << 7), - .ctl[dsf_bts_kernel] = (1 << 9), - .ctl[dsf_bts_user] = (1 << 10), - .nr_counter_reset = 4, -}; - -static void -ds_configure(const struct ds_configuration *cfg, - struct cpuinfo_x86 *cpu) -{ - unsigned long nr_pebs_fields = 0; - - printk(KERN_INFO "[ds] using %s configuration\n", cfg->name); - -#ifdef __i386__ - nr_pebs_fields = 10; -#else - nr_pebs_fields = 18; -#endif - - /* - * Starting with version 2, architectural performance - * monitoring supports a format specifier. - */ - if ((cpuid_eax(0xa) & 0xff) > 1) { - unsigned long perf_capabilities, format; - - rdmsrl(MSR_IA32_PERF_CAPABILITIES, perf_capabilities); - - format = (perf_capabilities >> 8) & 0xf; - - switch (format) { - case 0: - nr_pebs_fields = 18; - break; - case 1: - nr_pebs_fields = 22; - break; - default: - printk(KERN_INFO - "[ds] unknown PEBS format: %lu\n", format); - nr_pebs_fields = 0; - break; - } - } - - memset(&ds_cfg, 0, sizeof(ds_cfg)); - ds_cfg = *cfg; - - ds_cfg.sizeof_ptr_field = - (cpu_has(cpu, X86_FEATURE_DTES64) ? 8 : 4); - - ds_cfg.sizeof_rec[ds_bts] = ds_cfg.sizeof_ptr_field * 3; - ds_cfg.sizeof_rec[ds_pebs] = ds_cfg.sizeof_ptr_field * nr_pebs_fields; - - if (!cpu_has(cpu, X86_FEATURE_BTS)) { - ds_cfg.sizeof_rec[ds_bts] = 0; - printk(KERN_INFO "[ds] bts not available\n"); - } - if (!cpu_has(cpu, X86_FEATURE_PEBS)) { - ds_cfg.sizeof_rec[ds_pebs] = 0; - printk(KERN_INFO "[ds] pebs not available\n"); - } - - printk(KERN_INFO "[ds] sizes: address: %u bit, ", - 8 * ds_cfg.sizeof_ptr_field); - printk("bts/pebs record: %u/%u bytes\n", - ds_cfg.sizeof_rec[ds_bts], ds_cfg.sizeof_rec[ds_pebs]); - - WARN_ON_ONCE(MAX_PEBS_COUNTERS < ds_cfg.nr_counter_reset); -} - -void __cpuinit ds_init_intel(struct cpuinfo_x86 *c) -{ - /* Only configure the first cpu. Others are identical. */ - if (ds_cfg.name) - return; - - switch (c->x86) { - case 0x6: - switch (c->x86_model) { - case 0x9: - case 0xd: /* Pentium M */ - ds_configure(&ds_cfg_pentium_m, c); - break; - case 0xf: - case 0x17: /* Core2 */ - case 0x1c: /* Atom */ - ds_configure(&ds_cfg_core2_atom, c); - break; - case 0x1a: /* Core i7 */ - ds_configure(&ds_cfg_core_i7, c); - break; - default: - /* Sorry, don't know about them. */ - break; - } - break; - case 0xf: - switch (c->x86_model) { - case 0x0: - case 0x1: - case 0x2: /* Netburst */ - ds_configure(&ds_cfg_netburst, c); - break; - default: - /* Sorry, don't know about them. */ - break; - } - break; - default: - /* Sorry, don't know about them. */ - break; - } -} - -static inline void ds_take_timestamp(struct ds_context *context, - enum bts_qualifier qualifier, - struct task_struct *task) -{ - struct bts_tracer *tracer = context->bts_master; - struct bts_struct ts; - - /* Prevent compilers from reading the tracer pointer twice. */ - barrier(); - - if (!tracer || !(tracer->flags & BTS_TIMESTAMPS)) - return; - - memset(&ts, 0, sizeof(ts)); - ts.qualifier = qualifier; - ts.variant.event.clock = trace_clock_global(); - ts.variant.event.pid = task->pid; - - bts_write(tracer, &ts); -} - -/* - * Change the DS configuration from tracing prev to tracing next. - */ -void ds_switch_to(struct task_struct *prev, struct task_struct *next) -{ - struct ds_context *prev_ctx = prev->thread.ds_ctx; - struct ds_context *next_ctx = next->thread.ds_ctx; - unsigned long debugctlmsr = next->thread.debugctlmsr; - - /* Make sure all data is read before we start. */ - barrier(); - - if (prev_ctx) { - update_debugctlmsr(0); - - ds_take_timestamp(prev_ctx, bts_task_departs, prev); - } - - if (next_ctx) { - ds_take_timestamp(next_ctx, bts_task_arrives, next); - - wrmsrl(MSR_IA32_DS_AREA, (unsigned long)next_ctx->ds); - } - - update_debugctlmsr(debugctlmsr); -} - -static __init int ds_selftest(void) -{ - if (ds_cfg.sizeof_rec[ds_bts]) { - int error; - - error = ds_selftest_bts(); - if (error) { - WARN(1, "[ds] selftest failed. disabling bts.\n"); - ds_cfg.sizeof_rec[ds_bts] = 0; - } - } - - if (ds_cfg.sizeof_rec[ds_pebs]) { - int error; - - error = ds_selftest_pebs(); - if (error) { - WARN(1, "[ds] selftest failed. disabling pebs.\n"); - ds_cfg.sizeof_rec[ds_pebs] = 0; - } - } - - return 0; -} -device_initcall(ds_selftest); diff --git a/arch/x86/kernel/ds_selftest.c b/arch/x86/kernel/ds_selftest.c deleted file mode 100644 index 6bc7c199ab99..000000000000 --- a/arch/x86/kernel/ds_selftest.c +++ /dev/null @@ -1,408 +0,0 @@ -/* - * Debug Store support - selftest - * - * - * Copyright (C) 2009 Intel Corporation. - * Markus Metzger <markus.t.metzger@intel.com>, 2009 - */ - -#include "ds_selftest.h" - -#include <linux/kernel.h> -#include <linux/string.h> -#include <linux/smp.h> -#include <linux/cpu.h> - -#include <asm/ds.h> - - -#define BUFFER_SIZE 521 /* Intentionally chose an odd size. */ -#define SMALL_BUFFER_SIZE 24 /* A single bts entry. */ - -struct ds_selftest_bts_conf { - struct bts_tracer *tracer; - int error; - int (*suspend)(struct bts_tracer *); - int (*resume)(struct bts_tracer *); -}; - -static int ds_selftest_bts_consistency(const struct bts_trace *trace) -{ - int error = 0; - - if (!trace) { - printk(KERN_CONT "failed to access trace..."); - /* Bail out. Other tests are pointless. */ - return -1; - } - - if (!trace->read) { - printk(KERN_CONT "bts read not available..."); - error = -1; - } - - /* Do some sanity checks on the trace configuration. */ - if (!trace->ds.n) { - printk(KERN_CONT "empty bts buffer..."); - error = -1; - } - if (!trace->ds.size) { - printk(KERN_CONT "bad bts trace setup..."); - error = -1; - } - if (trace->ds.end != - (char *)trace->ds.begin + (trace->ds.n * trace->ds.size)) { - printk(KERN_CONT "bad bts buffer setup..."); - error = -1; - } - /* - * We allow top in [begin; end], since its not clear when the - * overflow adjustment happens: after the increment or before the - * write. - */ - if ((trace->ds.top < trace->ds.begin) || - (trace->ds.end < trace->ds.top)) { - printk(KERN_CONT "bts top out of bounds..."); - error = -1; - } - - return error; -} - -static int ds_selftest_bts_read(struct bts_tracer *tracer, - const struct bts_trace *trace, - const void *from, const void *to) -{ - const unsigned char *at; - - /* - * Check a few things which do not belong to this test. - * They should be covered by other tests. - */ - if (!trace) - return -1; - - if (!trace->read) - return -1; - - if (to < from) - return -1; - - if (from < trace->ds.begin) - return -1; - - if (trace->ds.end < to) - return -1; - - if (!trace->ds.size) - return -1; - - /* Now to the test itself. */ - for (at = from; (void *)at < to; at += trace->ds.size) { - struct bts_struct bts; - unsigned long index; - int error; - - if (((void *)at - trace->ds.begin) % trace->ds.size) { - printk(KERN_CONT - "read from non-integer index..."); - return -1; - } - index = ((void *)at - trace->ds.begin) / trace->ds.size; - - memset(&bts, 0, sizeof(bts)); - error = trace->read(tracer, at, &bts); - if (error < 0) { - printk(KERN_CONT - "error reading bts trace at [%lu] (0x%p)...", - index, at); - return error; - } - - switch (bts.qualifier) { - case BTS_BRANCH: - break; - default: - printk(KERN_CONT - "unexpected bts entry %llu at [%lu] (0x%p)...", - bts.qualifier, index, at); - return -1; - } - } - - return 0; -} - -static void ds_selftest_bts_cpu(void *arg) -{ - struct ds_selftest_bts_conf *conf = arg; - const struct bts_trace *trace; - void *top; - - if (IS_ERR(conf->tracer)) { - conf->error = PTR_ERR(conf->tracer); - conf->tracer = NULL; - - printk(KERN_CONT - "initialization failed (err: %d)...", conf->error); - return; - } - - /* We should meanwhile have enough trace. */ - conf->error = conf->suspend(conf->tracer); - if (conf->error < 0) - return; - - /* Let's see if we can access the trace. */ - trace = ds_read_bts(conf->tracer); - - conf->error = ds_selftest_bts_consistency(trace); - if (conf->error < 0) - return; - - /* If everything went well, we should have a few trace entries. */ - if (trace->ds.top == trace->ds.begin) { - /* - * It is possible but highly unlikely that we got a - * buffer overflow and end up at exactly the same - * position we started from. - * Let's issue a warning, but continue. - */ - printk(KERN_CONT "no trace/overflow..."); - } - - /* Let's try to read the trace we collected. */ - conf->error = - ds_selftest_bts_read(conf->tracer, trace, - trace->ds.begin, trace->ds.top); - if (conf->error < 0) - return; - - /* - * Let's read the trace again. - * Since we suspended tracing, we should get the same result. - */ - top = trace->ds.top; - - trace = ds_read_bts(conf->tracer); - conf->error = ds_selftest_bts_consistency(trace); - if (conf->error < 0) - return; - - if (top != trace->ds.top) { - printk(KERN_CONT "suspend not working..."); - conf->error = -1; - return; - } - - /* Let's collect some more trace - see if resume is working. */ - conf->error = conf->resume(conf->tracer); - if (conf->error < 0) - return; - - conf->error = conf->suspend(conf->tracer); - if (conf->error < 0) - return; - - trace = ds_read_bts(conf->tracer); - - conf->error = ds_selftest_bts_consistency(trace); - if (conf->error < 0) - return; - - if (trace->ds.top == top) { - /* - * It is possible but highly unlikely that we got a - * buffer overflow and end up at exactly the same - * position we started from. - * Let's issue a warning and check the full trace. - */ - printk(KERN_CONT - "no resume progress/overflow..."); - - conf->error = - ds_selftest_bts_read(conf->tracer, trace, - trace->ds.begin, trace->ds.end); - } else if (trace->ds.top < top) { - /* - * We had a buffer overflow - the entire buffer should - * contain trace records. - */ - conf->error = - ds_selftest_bts_read(conf->tracer, trace, - trace->ds.begin, trace->ds.end); - } else { - /* - * It is quite likely that the buffer did not overflow. - * Let's just check the delta trace. - */ - conf->error = - ds_selftest_bts_read(conf->tracer, trace, top, - trace->ds.top); - } - if (conf->error < 0) - return; - - conf->error = 0; -} - -static int ds_suspend_bts_wrap(struct bts_tracer *tracer) -{ - ds_suspend_bts(tracer); - return 0; -} - -static int ds_resume_bts_wrap(struct bts_tracer *tracer) -{ - ds_resume_bts(tracer); - return 0; -} - -static void ds_release_bts_noirq_wrap(void *tracer) -{ - (void)ds_release_bts_noirq(tracer); -} - -static int ds_selftest_bts_bad_release_noirq(int cpu, - struct bts_tracer *tracer) -{ - int error = -EPERM; - - /* Try to release the tracer on the wrong cpu. */ - get_cpu(); - if (cpu != smp_processor_id()) { - error = ds_release_bts_noirq(tracer); - if (error != -EPERM) - printk(KERN_CONT "release on wrong cpu..."); - } - put_cpu(); - - return error ? 0 : -1; -} - -static int ds_selftest_bts_bad_request_cpu(int cpu, void *buffer) -{ - struct bts_tracer *tracer; - int error; - - /* Try to request cpu tracing while task tracing is active. */ - tracer = ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE, NULL, - (size_t)-1, BTS_KERNEL); - error = PTR_ERR(tracer); - if (!IS_ERR(tracer)) { - ds_release_bts(tracer); - error = 0; - } - - if (error != -EPERM) - printk(KERN_CONT "cpu/task tracing overlap..."); - - return error ? 0 : -1; -} - -static int ds_selftest_bts_bad_request_task(void *buffer) -{ - struct bts_tracer *tracer; - int error; - - /* Try to request cpu tracing while task tracing is active. */ - tracer = ds_request_bts_task(current, buffer, BUFFER_SIZE, NULL, - (size_t)-1, BTS_KERNEL); - error = PTR_ERR(tracer); - if (!IS_ERR(tracer)) { - error = 0; - ds_release_bts(tracer); - } - - if (error != -EPERM) - printk(KERN_CONT "task/cpu tracing overlap..."); - - return error ? 0 : -1; -} - -int ds_selftest_bts(void) -{ - struct ds_selftest_bts_conf conf; - unsigned char buffer[BUFFER_SIZE], *small_buffer; - unsigned long irq; - int cpu; - - printk(KERN_INFO "[ds] bts selftest..."); - conf.error = 0; - - small_buffer = (unsigned char *)ALIGN((unsigned long)buffer, 8) + 8; - - get_online_cpus(); - for_each_online_cpu(cpu) { - conf.suspend = ds_suspend_bts_wrap; - conf.resume = ds_resume_bts_wrap; - conf.tracer = - ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE, - NULL, (size_t)-1, BTS_KERNEL); - ds_selftest_bts_cpu(&conf); - if (conf.error >= 0) - conf.error = ds_selftest_bts_bad_request_task(buffer); - ds_release_bts(conf.tracer); - if (conf.error < 0) - goto out; - - conf.suspend = ds_suspend_bts_noirq; - conf.resume = ds_resume_bts_noirq; - conf.tracer = - ds_request_bts_cpu(cpu, buffer, BUFFER_SIZE, - NULL, (size_t)-1, BTS_KERNEL); - smp_call_function_single(cpu, ds_selftest_bts_cpu, &conf, 1); - if (conf.error >= 0) { - conf.error = - ds_selftest_bts_bad_release_noirq(cpu, - conf.tracer); - /* We must not release the tracer twice. */ - if (conf.error < 0) - conf.tracer = NULL; - } - if (conf.error >= 0) - conf.error = ds_selftest_bts_bad_request_task(buffer); - smp_call_function_single(cpu, ds_release_bts_noirq_wrap, - conf.tracer, 1); - if (conf.error < 0) - goto out; - } - - conf.suspend = ds_suspend_bts_wrap; - conf.resume = ds_resume_bts_wrap; - conf.tracer = - ds_request_bts_task(current, buffer, BUFFER_SIZE, - NULL, (size_t)-1, BTS_KERNEL); - ds_selftest_bts_cpu(&conf); - if (conf.error >= 0) - conf.error = ds_selftest_bts_bad_request_cpu(0, buffer); - ds_release_bts(conf.tracer); - if (conf.error < 0) - goto out; - - conf.suspend = ds_suspend_bts_noirq; - conf.resume = ds_resume_bts_noirq; - conf.tracer = - ds_request_bts_task(current, small_buffer, SMALL_BUFFER_SIZE, - NULL, (size_t)-1, BTS_KERNEL); - local_irq_save(irq); - ds_selftest_bts_cpu(&conf); - if (conf.error >= 0) - conf.error = ds_selftest_bts_bad_request_cpu(0, buffer); - ds_release_bts_noirq(conf.tracer); - local_irq_restore(irq); - if (conf.error < 0) - goto out; - - conf.error = 0; - out: - put_online_cpus(); - printk(KERN_CONT "%s.\n", (conf.error ? "failed" : "passed")); - - return conf.error; -} - -int ds_selftest_pebs(void) -{ - return 0; -} diff --git a/arch/x86/kernel/ds_selftest.h b/arch/x86/kernel/ds_selftest.h deleted file mode 100644 index 2ba8745c6663..000000000000 --- a/arch/x86/kernel/ds_selftest.h +++ /dev/null @@ -1,15 +0,0 @@ -/* - * Debug Store support - selftest - * - * - * Copyright (C) 2009 Intel Corporation. - * Markus Metzger <markus.t.metzger@intel.com>, 2009 - */ - -#ifdef CONFIG_X86_DS_SELFTEST -extern int ds_selftest_bts(void); -extern int ds_selftest_pebs(void); -#else -static inline int ds_selftest_bts(void) { return 0; } -static inline int ds_selftest_pebs(void) { return 0; } -#endif diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c index 6d817554780a..c89a386930b7 100644 --- a/arch/x86/kernel/dumpstack.c +++ b/arch/x86/kernel/dumpstack.c @@ -224,11 +224,6 @@ unsigned __kprobes long oops_begin(void) int cpu; unsigned long flags; - /* notify the hw-branch tracer so it may disable tracing and - add the last trace to the trace buffer - - the earlier this happens, the more useful the trace. */ - trace_hw_branch_oops(); - oops_enter(); /* racy, but better than risking deadlock. */ diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index b43bbaebe2c0..f2f56c0967b6 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -422,14 +422,22 @@ static void __kprobes set_current_kprobe(struct kprobe *p, struct pt_regs *regs, static void __kprobes clear_btf(void) { - if (test_thread_flag(TIF_DEBUGCTLMSR)) - update_debugctlmsr(0); + if (test_thread_flag(TIF_BLOCKSTEP)) { + unsigned long debugctl = get_debugctlmsr(); + + debugctl &= ~DEBUGCTLMSR_BTF; + update_debugctlmsr(debugctl); + } } static void __kprobes restore_btf(void) { - if (test_thread_flag(TIF_DEBUGCTLMSR)) - update_debugctlmsr(current->thread.debugctlmsr); + if (test_thread_flag(TIF_BLOCKSTEP)) { + unsigned long debugctl = get_debugctlmsr(); + + debugctl |= DEBUGCTLMSR_BTF; + update_debugctlmsr(debugctl); + } } void __kprobes arch_prepare_kretprobe(struct kretprobe_instance *ri, diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 28ad9f4d8b94..eccdb57094e3 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -20,7 +20,6 @@ #include <asm/idle.h> #include <asm/uaccess.h> #include <asm/i387.h> -#include <asm/ds.h> #include <asm/debugreg.h> unsigned long idle_halt; @@ -50,8 +49,6 @@ void free_thread_xstate(struct task_struct *tsk) kmem_cache_free(task_xstate_cachep, tsk->thread.xstate); tsk->thread.xstate = NULL; } - - WARN(tsk->thread.ds_ctx, "leaking DS context\n"); } void free_thread_info(struct thread_info *ti) @@ -198,11 +195,16 @@ void __switch_to_xtra(struct task_struct *prev_p, struct task_struct *next_p, prev = &prev_p->thread; next = &next_p->thread; - if (test_tsk_thread_flag(next_p, TIF_DS_AREA_MSR) || - test_tsk_thread_flag(prev_p, TIF_DS_AREA_MSR)) - ds_switch_to(prev_p, next_p); - else if (next->debugctlmsr != prev->debugctlmsr) - update_debugctlmsr(next->debugctlmsr); + if (test_tsk_thread_flag(prev_p, TIF_BLOCKSTEP) ^ + test_tsk_thread_flag(next_p, TIF_BLOCKSTEP)) { + unsigned long debugctl = get_debugctlmsr(); + + debugctl &= ~DEBUGCTLMSR_BTF; + if (test_tsk_thread_flag(next_p, TIF_BLOCKSTEP)) + debugctl |= DEBUGCTLMSR_BTF; + + update_debugctlmsr(debugctl); + } if (test_tsk_thread_flag(prev_p, TIF_NOTSC) ^ test_tsk_thread_flag(next_p, TIF_NOTSC)) { diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index f6c62667e30c..75090c589b7a 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -55,7 +55,6 @@ #include <asm/cpu.h> #include <asm/idle.h> #include <asm/syscalls.h> -#include <asm/ds.h> #include <asm/debugreg.h> asmlinkage void ret_from_fork(void) __asm__("ret_from_fork"); @@ -238,13 +237,6 @@ int copy_thread(unsigned long clone_flags, unsigned long sp, kfree(p->thread.io_bitmap_ptr); p->thread.io_bitmap_max = 0; } - - clear_tsk_thread_flag(p, TIF_DS_AREA_MSR); - p->thread.ds_ctx = NULL; - - clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); - p->thread.debugctlmsr = 0; - return err; } diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index dc9690b4c4cc..cc4258f2beb5 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -49,7 +49,6 @@ #include <asm/ia32.h> #include <asm/idle.h> #include <asm/syscalls.h> -#include <asm/ds.h> #include <asm/debugreg.h> asmlinkage extern void ret_from_fork(void); @@ -313,13 +312,6 @@ int copy_thread(unsigned long clone_flags, unsigned long sp, if (err) goto out; } - - clear_tsk_thread_flag(p, TIF_DS_AREA_MSR); - p->thread.ds_ctx = NULL; - - clear_tsk_thread_flag(p, TIF_DEBUGCTLMSR); - p->thread.debugctlmsr = 0; - err = 0; out: if (err && p->thread.io_bitmap_ptr) { diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c index a503b1fd04e5..f2fd3b80e565 100644 --- a/arch/x86/kernel/ptrace.c +++ b/arch/x86/kernel/ptrace.c @@ -2,9 +2,6 @@ /* * Pentium III FXSR, SSE support * Gareth Hughes <gareth@valinux.com>, May 2000 - * - * BTS tracing - * Markus Metzger <markus.t.metzger@intel.com>, Dec 2007 */ #include <linux/kernel.h> @@ -21,7 +18,6 @@ #include <linux/audit.h> #include <linux/seccomp.h> #include <linux/signal.h> -#include <linux/workqueue.h> #include <linux/perf_event.h> #include <linux/hw_breakpoint.h> @@ -35,7 +31,6 @@ #include <asm/desc.h> #include <asm/prctl.h> #include <asm/proto.h> -#include <asm/ds.h> #include <asm/hw_breakpoint.h> #include "tls.h" @@ -788,342 +783,6 @@ static int ioperm_get(struct task_struct *target, 0, IO_BITMAP_BYTES); } -#ifdef CONFIG_X86_PTRACE_BTS -/* - * A branch trace store context. - * - * Contexts may only be installed by ptrace_bts_config() and only for - * ptraced tasks. - * - * Contexts are destroyed when the tracee is detached from the tracer. - * The actual destruction work requires interrupts enabled, so the - * work is deferred and will be scheduled during __ptrace_unlink(). - * - * Contexts hold an additional task_struct reference on the traced - * task, as well as a reference on the tracer's mm. - * - * Ptrace already holds a task_struct for the duration of ptrace operations, - * but since destruction is deferred, it may be executed after both - * tracer and tracee exited. - */ -struct bts_context { - /* The branch trace handle. */ - struct bts_tracer *tracer; - - /* The buffer used to store the branch trace and its size. */ - void *buffer; - unsigned int size; - - /* The mm that paid for the above buffer. */ - struct mm_struct *mm; - - /* The task this context belongs to. */ - struct task_struct *task; - - /* The signal to send on a bts buffer overflow. */ - unsigned int bts_ovfl_signal; - - /* The work struct to destroy a context. */ - struct work_struct work; -}; - -static int alloc_bts_buffer(struct bts_context *context, unsigned int size) -{ - void *buffer = NULL; - int err = -ENOMEM; - - err = account_locked_memory(current->mm, current->signal->rlim, size); - if (err < 0) - return err; - - buffer = kzalloc(size, GFP_KERNEL); - if (!buffer) - goto out_refund; - - context->buffer = buffer; - context->size = size; - context->mm = get_task_mm(current); - - return 0; - - out_refund: - refund_locked_memory(current->mm, size); - return err; -} - -static inline void free_bts_buffer(struct bts_context *context) -{ - if (!context->buffer) - return; - - kfree(context->buffer); - context->buffer = NULL; - - refund_locked_memory(context->mm, context->size); - context->size = 0; - - mmput(context->mm); - context->mm = NULL; -} - -static void free_bts_context_work(struct work_struct *w) -{ - struct bts_context *context; - - context = container_of(w, struct bts_context, work); - - ds_release_bts(context->tracer); - put_task_struct(context->task); - free_bts_buffer(context); - kfree(context); -} - -static inline void free_bts_context(struct bts_context *context) -{ - INIT_WORK(&context->work, free_bts_context_work); - schedule_work(&context->work); -} - -static inline struct bts_context *alloc_bts_context(struct task_struct *task) -{ - struct bts_context *context = kzalloc(sizeof(*context), GFP_KERNEL); - if (context) { - context->task = task; - task->bts = context; - - get_task_struct(task); - } - - return context; -} - -static int ptrace_bts_read_record(struct task_struct *child, size_t index, - struct bts_struct __user *out) -{ - struct bts_context *context; - const struct bts_trace *trace; - struct bts_struct bts; - const unsigned char *at; - int error; - - context = child->bts; - if (!context) - return -ESRCH; - - trace = ds_read_bts(context->tracer); - if (!trace) - return -ESRCH; - - at = trace->ds.top - ((index + 1) * trace->ds.size); - if ((void *)at < trace->ds.begin) - at += (trace->ds.n * trace->ds.size); - - if (!trace->read) - return -EOPNOTSUPP; - - error = trace->read(context->tracer, at, &bts); - if (error < 0) - return error; - - if (copy_to_user(out, &bts, sizeof(bts))) - return -EFAULT; - - return sizeof(bts); -} - -static int ptrace_bts_drain(struct task_struct *child, - long size, - struct bts_struct __user *out) -{ - struct bts_context *context; - const struct bts_trace *trace; - const unsigned char *at; - int error, drained = 0; - - context = child->bts; - if (!context) - return -ESRCH; - - trace = ds_read_bts(context->tracer); - if (!trace) - return -ESRCH; - - if (!trace->read) - return -EOPNOTSUPP; - - if (size < (trace->ds.top - trace->ds.begin)) - return -EIO; - - for (at = trace->ds.begin; (void *)at < trace->ds.top; - out++, drained++, at += trace->ds.size) { - struct bts_struct bts; - - error = trace->read(context->tracer, at, &bts); - if (error < 0) - return error; - - if (copy_to_user(out, &bts, sizeof(bts))) - return -EFAULT; - } - - memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); - - error = ds_reset_bts(context->tracer); - if (error < 0) - return error; - - return drained; -} - -static int ptrace_bts_config(struct task_struct *child, - long cfg_size, - const struct ptrace_bts_config __user *ucfg) -{ - struct bts_context *context; - struct ptrace_bts_config cfg; - unsigned int flags = 0; - - if (cfg_size < sizeof(cfg)) - return -EIO; - - if (copy_from_user(&cfg, ucfg, sizeof(cfg))) - return -EFAULT; - - context = child->bts; - if (!context) - context = alloc_bts_context(child); - if (!context) - return -ENOMEM; - - if (cfg.flags & PTRACE_BTS_O_SIGNAL) { - if (!cfg.signal) - return -EINVAL; - - return -EOPNOTSUPP; - context->bts_ovfl_signal = cfg.signal; - } - - ds_release_bts(context->tracer); - context->tracer = NULL; - - if ((cfg.flags & PTRACE_BTS_O_ALLOC) && (cfg.size != context->size)) { - int err; - - free_bts_buffer(context); - if (!cfg.size) - return 0; - - err = alloc_bts_buffer(context, cfg.size); - if (err < 0) - return err; - } - - if (cfg.flags & PTRACE_BTS_O_TRACE) - flags |= BTS_USER; - - if (cfg.flags & PTRACE_BTS_O_SCHED) - flags |= BTS_TIMESTAMPS; - - context->tracer = - ds_request_bts_task(child, context->buffer, context->size, - NULL, (size_t)-1, flags); - if (unlikely(IS_ERR(context->tracer))) { - int error = PTR_ERR(context->tracer); - - free_bts_buffer(context); - context->tracer = NULL; - return error; - } - - return sizeof(cfg); -} - -static int ptrace_bts_status(struct task_struct *child, - long cfg_size, - struct ptrace_bts_config __user *ucfg) -{ - struct bts_context *context; - const struct bts_trace *trace; - struct ptrace_bts_config cfg; - - context = child->bts; - if (!context) - return -ESRCH; - - if (cfg_size < sizeof(cfg)) - return -EIO; - - trace = ds_read_bts(context->tracer); - if (!trace) - return -ESRCH; - - memset(&cfg, 0, sizeof(cfg)); - cfg.size = trace->ds.end - trace->ds.begin; - cfg.signal = context->bts_ovfl_signal; - cfg.bts_size = sizeof(struct bts_struct); - - if (cfg.signal) - cfg.flags |= PTRACE_BTS_O_SIGNAL; - - if (trace->ds.flags & BTS_USER) - cfg.flags |= PTRACE_BTS_O_TRACE; - - if (trace->ds.flags & BTS_TIMESTAMPS) - cfg.flags |= PTRACE_BTS_O_SCHED; - - if (copy_to_user(ucfg, &cfg, sizeof(cfg))) - return -EFAULT; - - return sizeof(cfg); -} - -static int ptrace_bts_clear(struct task_struct *child) -{ - struct bts_context *context; - const struct bts_trace *trace; - - context = child->bts; - if (!context) - return -ESRCH; - - trace = ds_read_bts(context->tracer); - if (!trace) - return -ESRCH; - - memset(trace->ds.begin, 0, trace->ds.n * trace->ds.size); - - return ds_reset_bts(context->tracer); -} - -static int ptrace_bts_size(struct task_struct *child) -{ - struct bts_context *context; - const struct bts_trace *trace; - - context = child->bts; - if (!context) - return -ESRCH; - - trace = ds_read_bts(context->tracer); - if (!trace) - return -ESRCH; - - return (trace->ds.top - trace->ds.begin) / trace->ds.size; -} - -/* - * Called from __ptrace_unlink() after the child has been moved back - * to its original parent. - */ -void ptrace_bts_untrace(struct task_struct *child) -{ - if (unlikely(child->bts)) { - free_bts_context(child->bts); - child->bts = NULL; - } -} -#endif /* CONFIG_X86_PTRACE_BTS */ - /* * Called by kernel/ptrace.c when detaching.. * @@ -1251,39 +910,6 @@ long arch_ptrace(struct task_struct *child, long request, long addr, long data) break; #endif - /* - * These bits need more cooking - not enabled yet: - */ -#ifdef CONFIG_X86_PTRACE_BTS - case PTRACE_BTS_CONFIG: - ret = ptrace_bts_config - (child, data, (struct ptrace_bts_config __user *)addr); - break; - - case PTRACE_BTS_STATUS: - ret = ptrace_bts_status - (child, data, (struct ptrace_bts_config __user *)addr); - break; - - case PTRACE_BTS_SIZE: - ret = ptrace_bts_size(child); - break; - - case PTRACE_BTS_GET: - ret = ptrace_bts_read_record - (child, data, (struct bts_struct __user *) addr); - break; - - case PTRACE_BTS_CLEAR: - ret = ptrace_bts_clear(child); - break; - - case PTRACE_BTS_DRAIN: - ret = ptrace_bts_drain - (child, data, (struct bts_struct __user *) addr); - break; -#endif /* CONFIG_X86_PTRACE_BTS */ - default: ret = ptrace_request(child, request, addr, data); break; @@ -1543,14 +1169,6 @@ long compat_arch_ptrace(struct task_struct *child, compat_long_t request, case PTRACE_GET_THREAD_AREA: case PTRACE_SET_THREAD_AREA: -#ifdef CONFIG_X86_PTRACE_BTS - case PTRACE_BTS_CONFIG: - case PTRACE_BTS_STATUS: - case PTRACE_BTS_SIZE: - case PTRACE_BTS_GET: - case PTRACE_BTS_CLEAR: - case PTRACE_BTS_DRAIN: -#endif /* CONFIG_X86_PTRACE_BTS */ return arch_ptrace(child, request, addr, data); default: diff --git a/arch/x86/kernel/step.c b/arch/x86/kernel/step.c index 3149032ff107..58de45ee08b6 100644 --- a/arch/x86/kernel/step.c +++ b/arch/x86/kernel/step.c @@ -158,22 +158,6 @@ static int enable_single_step(struct task_struct *child) } /* - * Install this value in MSR_IA32_DEBUGCTLMSR whenever child is running. - */ -static void write_debugctlmsr(struct task_struct *child, unsigned long val) -{ - if (child->thread.debugctlmsr == val) - return; - - child->thread.debugctlmsr = val; - - if (child != current) - return; - - update_debugctlmsr(val); -} - -/* * Enable single or block step. */ static void enable_step(struct task_struct *child, bool block) @@ -186,15 +170,17 @@ static void enable_step(struct task_struct *child, bool block) * that uses user-mode single stepping itself. */ if (enable_single_step(child) && block) { - set_tsk_thread_flag(child, TIF_DEBUGCTLMSR); - write_debugctlmsr(child, - child->thread.debugctlmsr | DEBUGCTLMSR_BTF); - } else { - write_debugctlmsr(child, - child->thread.debugctlmsr & ~DEBUGCTLMSR_BTF); - - if (!child->thread.debugctlmsr) - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + unsigned long debugctl = get_debugctlmsr(); + + debugctl |= DEBUGCTLMSR_BTF; + update_debugctlmsr(debugctl); + set_tsk_thread_flag(child, TIF_BLOCKSTEP); + } else if (test_tsk_thread_flag(child, TIF_BLOCKSTEP)) { + unsigned long debugctl = get_debugctlmsr(); + + debugctl &= ~DEBUGCTLMSR_BTF; + update_debugctlmsr(debugctl); + clear_tsk_thread_flag(child, TIF_BLOCKSTEP); } } @@ -213,11 +199,13 @@ void user_disable_single_step(struct task_struct *child) /* * Make sure block stepping (BTF) is disabled. */ - write_debugctlmsr(child, - child->thread.debugctlmsr & ~DEBUGCTLMSR_BTF); + if (test_tsk_thread_flag(child, TIF_BLOCKSTEP)) { + unsigned long debugctl = get_debugctlmsr(); - if (!child->thread.debugctlmsr) - clear_tsk_thread_flag(child, TIF_DEBUGCTLMSR); + debugctl &= ~DEBUGCTLMSR_BTF; + update_debugctlmsr(debugctl); + clear_tsk_thread_flag(child, TIF_BLOCKSTEP); + } /* Always clear TIF_SINGLESTEP... */ clear_tsk_thread_flag(child, TIF_SINGLESTEP); diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c index 1168e4454188..36f1bd9f8e76 100644 --- a/arch/x86/kernel/traps.c +++ b/arch/x86/kernel/traps.c @@ -543,11 +543,11 @@ dotraplinkage void __kprobes do_debug(struct pt_regs *regs, long error_code) /* DR6 may or may not be cleared by the CPU */ set_debugreg(0, 6); + /* * The processor cleared BTF, so don't mark that we need it set. */ - clear_tsk_thread_flag(tsk, TIF_DEBUGCTLMSR); - tsk->thread.debugctlmsr = 0; + clear_tsk_thread_flag(tsk, TIF_BLOCKSTEP); /* Store the virtualized DR6 value */ tsk->thread.debugreg6 = dr6; diff --git a/arch/x86/lib/Makefile b/arch/x86/lib/Makefile index 419386c24b82..cbaf8f2b83df 100644 --- a/arch/x86/lib/Makefile +++ b/arch/x86/lib/Makefile @@ -20,7 +20,7 @@ lib-y := delay.o lib-y += thunk_$(BITS).o lib-y += usercopy_$(BITS).o getuser.o putuser.o lib-y += memcpy_$(BITS).o -lib-$(CONFIG_KPROBES) += insn.o inat.o +lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o obj-y += msr.o msr-reg.o msr-reg-export.o diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h index 01e6adea07ec..cc12b3c556b3 100644 --- a/include/linux/ftrace.h +++ b/include/linux/ftrace.h @@ -504,18 +504,6 @@ extern int ftrace_dump_on_oops; #define INIT_TRACE_RECURSION #endif -#ifdef CONFIG_HW_BRANCH_TRACER - -void trace_hw_branch(u64 from, u64 to); -void trace_hw_branch_oops(void); - -#else /* CONFIG_HW_BRANCH_TRACER */ - -static inline void trace_hw_branch(u64 from, u64 to) {} -static inline void trace_hw_branch_oops(void) {} - -#endif /* CONFIG_HW_BRANCH_TRACER */ - #ifdef CONFIG_FTRACE_SYSCALLS unsigned long arch_syscall_addr(int nr); diff --git a/include/linux/mm.h b/include/linux/mm.h index e70f21beb4b4..c8442b655111 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -19,7 +19,6 @@ struct anon_vma; struct file_ra_state; struct user_struct; struct writeback_control; -struct rlimit; #ifndef CONFIG_DISCONTIGMEM /* Don't use mapnrs, do it properly */ extern unsigned long max_mapnr; @@ -1449,9 +1448,6 @@ int vmemmap_populate_basepages(struct page *start_page, int vmemmap_populate(struct page *start_page, unsigned long pages, int node); void vmemmap_populate_print_last(void); -extern int account_locked_memory(struct mm_struct *mm, struct rlimit *rlim, - size_t size); -extern void refund_locked_memory(struct mm_struct *mm, size_t size); enum mf_flags { MF_COUNT_INCREASED = 1 << 0, diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index c8e375440403..6e96cc8225d4 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -203,8 +203,9 @@ struct perf_event_attr { enable_on_exec : 1, /* next exec enables */ task : 1, /* trace fork/exit */ watermark : 1, /* wakeup_watermark */ + precise : 1, /* OoO invariant counter */ - __reserved_1 : 49; + __reserved_1 : 48; union { __u32 wakeup_events; /* wakeup every n events */ @@ -293,6 +294,12 @@ struct perf_event_mmap_page { #define PERF_RECORD_MISC_USER (2 << 0) #define PERF_RECORD_MISC_HYPERVISOR (3 << 0) +#define PERF_RECORD_MISC_EXACT (1 << 14) +/* + * Reserve the last bit to indicate some extended misc field + */ +#define PERF_RECORD_MISC_EXT_RESERVED (1 << 15) + struct perf_event_header { __u32 type; __u16 misc; @@ -468,6 +475,17 @@ struct perf_raw_record { void *data; }; +struct perf_branch_entry { + __u64 from; + __u64 to; + __u64 flags; +}; + +struct perf_branch_stack { + __u64 nr; + struct perf_branch_entry entries[0]; +}; + struct task_struct; /** diff --git a/include/linux/ptrace.h b/include/linux/ptrace.h index e1fb60729979..4272521e29e9 100644 --- a/include/linux/ptrace.h +++ b/include/linux/ptrace.h @@ -345,18 +345,6 @@ static inline void user_single_step_siginfo(struct task_struct *tsk, #define arch_ptrace_stop(code, info) do { } while (0) #endif -#ifndef arch_ptrace_untrace -/* - * Do machine-specific work before untracing child. - * - * This is called for a normal detach as well as from ptrace_exit() - * when the tracing task dies. - * - * Called with write_lock(&tasklist_lock) held. - */ -#define arch_ptrace_untrace(task) do { } while (0) -#endif - extern int task_current_syscall(struct task_struct *target, long *callno, unsigned long args[6], unsigned int maxargs, unsigned long *sp, unsigned long *pc); diff --git a/include/linux/sched.h b/include/linux/sched.h index dad7f668ebf7..e0447c64af6a 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -99,7 +99,6 @@ struct futex_pi_state; struct robust_list_head; struct bio_list; struct fs_struct; -struct bts_context; struct perf_event_context; /* @@ -1272,12 +1271,6 @@ struct task_struct { struct list_head ptraced; struct list_head ptrace_entry; - /* - * This is the tracer handle for the ptrace BTS extension. - * This field actually belongs to the ptracer task. - */ - struct bts_context *bts; - /* PID/PID hash table linkage. */ struct pid_link pids[PIDTYPE_MAX]; struct list_head thread_group; @@ -2123,10 +2116,8 @@ extern void set_task_comm(struct task_struct *tsk, char *from); extern char *get_task_comm(char *to, struct task_struct *tsk); #ifdef CONFIG_SMP -extern void wait_task_context_switch(struct task_struct *p); extern unsigned long wait_task_inactive(struct task_struct *, long match_state); #else -static inline void wait_task_context_switch(struct task_struct *p) {} static inline unsigned long wait_task_inactive(struct task_struct *p, long match_state) { diff --git a/kernel/fork.c b/kernel/fork.c index 4799c5f0e6d0..d67f1dbfbe03 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1108,9 +1108,6 @@ static struct task_struct *copy_process(unsigned long clone_flags, p->memcg_batch.do_batch = 0; p->memcg_batch.memcg = NULL; #endif - - p->bts = NULL; - p->stack_start = stack_start; /* Perform scheduler related setup. Assign this task to a CPU. */ diff --git a/kernel/perf_event.c b/kernel/perf_event.c index 96aae13c7960..074ba2e036c8 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -1366,6 +1366,8 @@ void perf_event_task_sched_in(struct task_struct *task) if (cpuctx->task_ctx == ctx) return; + perf_disable(); + /* * We want to keep the following priority order: * cpu pinned (that don't need to move), task pinned, @@ -1378,6 +1380,8 @@ void perf_event_task_sched_in(struct task_struct *task) ctx_sched_in(ctx, cpuctx, EVENT_FLEXIBLE); cpuctx->task_ctx = ctx; + + perf_enable(); } #define MAX_INTERRUPTS (~0ULL) diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 42ad8ae729a0..9fb51237b18c 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -76,7 +76,6 @@ void __ptrace_unlink(struct task_struct *child) child->parent = child->real_parent; list_del_init(&child->ptrace_entry); - arch_ptrace_untrace(child); if (task_is_traced(child)) ptrace_untrace(child); } diff --git a/kernel/sched.c b/kernel/sched.c index 49d2fa7b687a..1038ca163890 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -2077,49 +2077,6 @@ migrate_task(struct task_struct *p, int dest_cpu, struct migration_req *req) } /* - * wait_task_context_switch - wait for a thread to complete at least one - * context switch. - * - * @p must not be current. - */ -void wait_task_context_switch(struct task_struct *p) -{ - unsigned long nvcsw, nivcsw, flags; - int running; - struct rq *rq; - - nvcsw = p->nvcsw; - nivcsw = p->nivcsw; - for (;;) { - /* - * The runqueue is assigned before the actual context - * switch. We need to take the runqueue lock. - * - * We could check initially without the lock but it is - * very likely that we need to take the lock in every - * iteration. - */ - rq = task_rq_lock(p, &flags); - running = task_running(rq, p); - task_rq_unlock(rq, &flags); - - if (likely(!running)) - break; - /* - * The switch count is incremented before the actual - * context switch. We thus wait for two switches to be - * sure at least one completed. - */ - if ((p->nvcsw - nvcsw) > 1) - break; - if ((p->nivcsw - nivcsw) > 1) - break; - - cpu_relax(); - } -} - -/* * wait_task_inactive - wait for a thread to unschedule. * * If @match_state is nonzero, it's the @p->state value just checked and diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig index 13e13d428cd3..8b1797c4545b 100644 --- a/kernel/trace/Kconfig +++ b/kernel/trace/Kconfig @@ -44,9 +44,6 @@ config HAVE_FTRACE_MCOUNT_RECORD help See Documentation/trace/ftrace-design.txt -config HAVE_HW_BRANCH_TRACER - bool - config HAVE_SYSCALL_TRACEPOINTS bool help @@ -374,14 +371,6 @@ config STACK_TRACER Say N if unsure. -config HW_BRANCH_TRACER - depends on HAVE_HW_BRANCH_TRACER - bool "Trace hw branches" - select GENERIC_TRACER - help - This tracer records all branches on the system in a circular - buffer, giving access to the last N branches for each cpu. - config KMEMTRACE bool "Trace SLAB allocations" select GENERIC_TRACER diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile index 78edc6490038..ffb1a5b0550e 100644 --- a/kernel/trace/Makefile +++ b/kernel/trace/Makefile @@ -41,7 +41,6 @@ obj-$(CONFIG_MMIOTRACE) += trace_mmiotrace.o obj-$(CONFIG_BOOT_TRACER) += trace_boot.o obj-$(CONFIG_FUNCTION_GRAPH_TRACER) += trace_functions_graph.o obj-$(CONFIG_TRACE_BRANCH_PROFILING) += trace_branch.o -obj-$(CONFIG_HW_BRANCH_TRACER) += trace_hw_branches.o obj-$(CONFIG_KMEMTRACE) += kmemtrace.o obj-$(CONFIG_WORKQUEUE_TRACER) += trace_workqueue.o obj-$(CONFIG_BLK_DEV_IO_TRACE) += blktrace.o diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h index 2825ef2c0b15..bec2c973ff0c 100644 --- a/kernel/trace/trace.h +++ b/kernel/trace/trace.h @@ -34,7 +34,6 @@ enum trace_type { TRACE_GRAPH_RET, TRACE_GRAPH_ENT, TRACE_USER_STACK, - TRACE_HW_BRANCHES, TRACE_KMEM_ALLOC, TRACE_KMEM_FREE, TRACE_BLK, @@ -229,7 +228,6 @@ extern void __ftrace_bad_type(void); TRACE_GRAPH_ENT); \ IF_ASSIGN(var, ent, struct ftrace_graph_ret_entry, \ TRACE_GRAPH_RET); \ - IF_ASSIGN(var, ent, struct hw_branch_entry, TRACE_HW_BRANCHES);\ IF_ASSIGN(var, ent, struct kmemtrace_alloc_entry, \ TRACE_KMEM_ALLOC); \ IF_ASSIGN(var, ent, struct kmemtrace_free_entry, \ @@ -467,8 +465,6 @@ extern int trace_selftest_startup_sysprof(struct tracer *trace, struct trace_array *tr); extern int trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr); -extern int trace_selftest_startup_hw_branches(struct tracer *trace, - struct trace_array *tr); extern int trace_selftest_startup_ksym(struct tracer *trace, struct trace_array *tr); #endif /* CONFIG_FTRACE_STARTUP_TEST */ diff --git a/kernel/trace/trace_entries.h b/kernel/trace/trace_entries.h index c16a08f399df..dc008c1240da 100644 --- a/kernel/trace/trace_entries.h +++ b/kernel/trace/trace_entries.h @@ -318,18 +318,6 @@ FTRACE_ENTRY(branch, trace_branch, __entry->func, __entry->file, __entry->correct) ); -FTRACE_ENTRY(hw_branch, hw_branch_entry, - - TRACE_HW_BRANCHES, - - F_STRUCT( - __field( u64, from ) - __field( u64, to ) - ), - - F_printk("from: %llx to: %llx", __entry->from, __entry->to) -); - FTRACE_ENTRY(kmem_alloc, kmemtrace_alloc_entry, TRACE_KMEM_ALLOC, diff --git a/kernel/trace/trace_hw_branches.c b/kernel/trace/trace_hw_branches.c deleted file mode 100644 index 7b97000745f5..000000000000 --- a/kernel/trace/trace_hw_branches.c +++ /dev/null @@ -1,312 +0,0 @@ -/* - * h/w branch tracer for x86 based on BTS - * - * Copyright (C) 2008-2009 Intel Corporation. - * Markus Metzger <markus.t.metzger@gmail.com>, 2008-2009 - */ -#include <linux/kallsyms.h> -#include <linux/debugfs.h> -#include <linux/ftrace.h> -#include <linux/module.h> -#include <linux/cpu.h> -#include <linux/smp.h> -#include <linux/fs.h> - -#include <asm/ds.h> - -#include "trace_output.h" -#include "trace.h" - - -#define BTS_BUFFER_SIZE (1 << 13) - -static DEFINE_PER_CPU(struct bts_tracer *, hwb_tracer); -static DEFINE_PER_CPU(unsigned char[BTS_BUFFER_SIZE], hwb_buffer); - -#define this_tracer per_cpu(hwb_tracer, smp_processor_id()) - -static int trace_hw_branches_enabled __read_mostly; -static int trace_hw_branches_suspended __read_mostly; -static struct trace_array *hw_branch_trace __read_mostly; - - -static void bts_trace_init_cpu(int cpu) -{ - per_cpu(hwb_tracer, cpu) = - ds_request_bts_cpu(cpu, per_cpu(hwb_buffer, cpu), - BTS_BUFFER_SIZE, NULL, (size_t)-1, - BTS_KERNEL); - - if (IS_ERR(per_cpu(hwb_tracer, cpu))) - per_cpu(hwb_tracer, cpu) = NULL; -} - -static int bts_trace_init(struct trace_array *tr) -{ - int cpu; - - hw_branch_trace = tr; - trace_hw_branches_enabled = 0; - - get_online_cpus(); - for_each_online_cpu(cpu) { - bts_trace_init_cpu(cpu); - - if (likely(per_cpu(hwb_tracer, cpu))) - trace_hw_branches_enabled = 1; - } - trace_hw_branches_suspended = 0; - put_online_cpus(); - - /* If we could not enable tracing on a single cpu, we fail. */ - return trace_hw_branches_enabled ? 0 : -EOPNOTSUPP; -} - -static void bts_trace_reset(struct trace_array *tr) -{ - int cpu; - - get_online_cpus(); - for_each_online_cpu(cpu) { - if (likely(per_cpu(hwb_tracer, cpu))) { - ds_release_bts(per_cpu(hwb_tracer, cpu)); - per_cpu(hwb_tracer, cpu) = NULL; - } - } - trace_hw_branches_enabled = 0; - trace_hw_branches_suspended = 0; - put_online_cpus(); -} - -static void bts_trace_start(struct trace_array *tr) -{ - int cpu; - - get_online_cpus(); - for_each_online_cpu(cpu) - if (likely(per_cpu(hwb_tracer, cpu))) - ds_resume_bts(per_cpu(hwb_tracer, cpu)); - trace_hw_branches_suspended = 0; - put_online_cpus(); -} - -static void bts_trace_stop(struct trace_array *tr) -{ - int cpu; - - get_online_cpus(); - for_each_online_cpu(cpu) - if (likely(per_cpu(hwb_tracer, cpu))) - ds_suspend_bts(per_cpu(hwb_tracer, cpu)); - trace_hw_branches_suspended = 1; - put_online_cpus(); -} - -static int __cpuinit bts_hotcpu_handler(struct notifier_block *nfb, - unsigned long action, void *hcpu) -{ - int cpu = (long)hcpu; - - switch (action) { - case CPU_ONLINE: - case CPU_DOWN_FAILED: - /* The notification is sent with interrupts enabled. */ - if (trace_hw_branches_enabled) { - bts_trace_init_cpu(cpu); - - if (trace_hw_branches_suspended && - likely(per_cpu(hwb_tracer, cpu))) - ds_suspend_bts(per_cpu(hwb_tracer, cpu)); - } - break; - - case CPU_DOWN_PREPARE: - /* The notification is sent with interrupts enabled. */ - if (likely(per_cpu(hwb_tracer, cpu))) { - ds_release_bts(per_cpu(hwb_tracer, cpu)); - per_cpu(hwb_tracer, cpu) = NULL; - } - } - - return NOTIFY_DONE; -} - -static struct notifier_block bts_hotcpu_notifier __cpuinitdata = { - .notifier_call = bts_hotcpu_handler -}; - -static void bts_trace_print_header(struct seq_file *m) -{ - seq_puts(m, "# CPU# TO <- FROM\n"); -} - -static enum print_line_t bts_trace_print_line(struct trace_iterator *iter) -{ - unsigned long symflags = TRACE_ITER_SYM_OFFSET; - struct trace_entry *entry = iter->ent; - struct trace_seq *seq = &iter->seq; - struct hw_branch_entry *it; - - trace_assign_type(it, entry); - - if (entry->type == TRACE_HW_BRANCHES) { - if (trace_seq_printf(seq, "%4d ", iter->cpu) && - seq_print_ip_sym(seq, it->to, symflags) && - trace_seq_printf(seq, "\t <- ") && - seq_print_ip_sym(seq, it->from, symflags) && - trace_seq_printf(seq, "\n")) - return TRACE_TYPE_HANDLED; - return TRACE_TYPE_PARTIAL_LINE; - } - return TRACE_TYPE_UNHANDLED; -} - -void trace_hw_branch(u64 from, u64 to) -{ - struct ftrace_event_call *call = &event_hw_branch; - struct trace_array *tr = hw_branch_trace; - struct ring_buffer_event *event; - struct ring_buffer *buf; - struct hw_branch_entry *entry; - unsigned long irq1; - int cpu; - - if (unlikely(!tr)) - return; - - if (unlikely(!trace_hw_branches_enabled)) - return; - - local_irq_save(irq1); - cpu = raw_smp_processor_id(); - if (atomic_inc_return(&tr->data[cpu]->disabled) != 1) - goto out; - - buf = tr->buffer; - event = trace_buffer_lock_reserve(buf, TRACE_HW_BRANCHES, - sizeof(*entry), 0, 0); - if (!event) - goto out; - entry = ring_buffer_event_data(event); - tracing_generic_entry_update(&entry->ent, 0, from); - entry->ent.type = TRACE_HW_BRANCHES; - entry->from = from; - entry->to = to; - if (!filter_check_discard(call, entry, buf, event)) - trace_buffer_unlock_commit(buf, event, 0, 0); - - out: - atomic_dec(&tr->data[cpu]->disabled); - local_irq_restore(irq1); -} - -static void trace_bts_at(const struct bts_trace *trace, void *at) -{ - struct bts_struct bts; - int err = 0; - - WARN_ON_ONCE(!trace->read); - if (!trace->read) - return; - - err = trace->read(this_tracer, at, &bts); - if (err < 0) - return; - - switch (bts.qualifier) { - case BTS_BRANCH: - trace_hw_branch(bts.variant.lbr.from, bts.variant.lbr.to); - break; - } -} - -/* - * Collect the trace on the current cpu and write it into the ftrace buffer. - * - * pre: tracing must be suspended on the current cpu - */ -static void trace_bts_cpu(void *arg) -{ - struct trace_array *tr = (struct trace_array *)arg; - const struct bts_trace *trace; - unsigned char *at; - - if (unlikely(!tr)) - return; - - if (unlikely(atomic_read(&tr->data[raw_smp_processor_id()]->disabled))) - return; - - if (unlikely(!this_tracer)) - return; - - trace = ds_read_bts(this_tracer); - if (!trace) - return; - - for (at = trace->ds.top; (void *)at < trace->ds.end; - at += trace->ds.size) - trace_bts_at(trace, at); - - for (at = trace->ds.begin; (void *)at < trace->ds.top; - at += trace->ds.size) - trace_bts_at(trace, at); -} - -static void trace_bts_prepare(struct trace_iterator *iter) -{ - int cpu; - - get_online_cpus(); - for_each_online_cpu(cpu) - if (likely(per_cpu(hwb_tracer, cpu))) - ds_suspend_bts(per_cpu(hwb_tracer, cpu)); - /* - * We need to collect the trace on the respective cpu since ftrace - * implicitly adds the record for the current cpu. - * Once that is more flexible, we could collect the data from any cpu. - */ - on_each_cpu(trace_bts_cpu, iter->tr, 1); - - for_each_online_cpu(cpu) - if (likely(per_cpu(hwb_tracer, cpu))) - ds_resume_bts(per_cpu(hwb_tracer, cpu)); - put_online_cpus(); -} - -static void trace_bts_close(struct trace_iterator *iter) -{ - tracing_reset_online_cpus(iter->tr); -} - -void trace_hw_branch_oops(void) -{ - if (this_tracer) { - ds_suspend_bts_noirq(this_tracer); - trace_bts_cpu(hw_branch_trace); - ds_resume_bts_noirq(this_tracer); - } -} - -struct tracer bts_tracer __read_mostly = -{ - .name = "hw-branch-tracer", - .init = bts_trace_init, - .reset = bts_trace_reset, - .print_header = bts_trace_print_header, - .print_line = bts_trace_print_line, - .start = bts_trace_start, - .stop = bts_trace_stop, - .open = trace_bts_prepare, - .close = trace_bts_close, -#ifdef CONFIG_FTRACE_SELFTEST - .selftest = trace_selftest_startup_hw_branches, -#endif /* CONFIG_FTRACE_SELFTEST */ -}; - -__init static int init_bts_trace(void) -{ - register_hotcpu_notifier(&bts_hotcpu_notifier); - return register_tracer(&bts_tracer); -} -device_initcall(init_bts_trace); diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index 280fea470d67..a7084e7c0427 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -16,7 +16,6 @@ static inline int trace_valid_entry(struct trace_entry *entry) case TRACE_BRANCH: case TRACE_GRAPH_ENT: case TRACE_GRAPH_RET: - case TRACE_HW_BRANCHES: case TRACE_KSYM: return 1; } @@ -754,62 +753,6 @@ trace_selftest_startup_branch(struct tracer *trace, struct trace_array *tr) } #endif /* CONFIG_BRANCH_TRACER */ -#ifdef CONFIG_HW_BRANCH_TRACER -int -trace_selftest_startup_hw_branches(struct tracer *trace, - struct trace_array *tr) -{ - struct trace_iterator *iter; - struct tracer tracer; - unsigned long count; - int ret; - - if (!trace->open) { - printk(KERN_CONT "missing open function..."); - return -1; - } - - ret = tracer_init(trace, tr); - if (ret) { - warn_failed_init_tracer(trace, ret); - return ret; - } - - /* - * The hw-branch tracer needs to collect the trace from the various - * cpu trace buffers - before tracing is stopped. - */ - iter = kzalloc(sizeof(*iter), GFP_KERNEL); - if (!iter) - return -ENOMEM; - - memcpy(&tracer, trace, sizeof(tracer)); - - iter->trace = &tracer; - iter->tr = tr; - iter->pos = -1; - mutex_init(&iter->mutex); - - trace->open(iter); - - mutex_destroy(&iter->mutex); - kfree(iter); - - tracing_stop(); - - ret = trace_test_buffer(tr, &count); - trace->reset(tr); - tracing_start(); - - if (!ret && !count) { - printk(KERN_CONT "no entries found.."); - ret = -1; - } - - return ret; -} -#endif /* CONFIG_HW_BRANCH_TRACER */ - #ifdef CONFIG_KSYM_TRACER static int ksym_selftest_dummy; diff --git a/mm/mlock.c b/mm/mlock.c index 8f4e2dfceec1..3f82720e0515 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -607,44 +607,3 @@ void user_shm_unlock(size_t size, struct user_struct *user) spin_unlock(&shmlock_user_lock); free_uid(user); } - -int account_locked_memory(struct mm_struct *mm, struct rlimit *rlim, - size_t size) -{ - unsigned long lim, vm, pgsz; - int error = -ENOMEM; - - pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; - - down_write(&mm->mmap_sem); - - lim = ACCESS_ONCE(rlim[RLIMIT_AS].rlim_cur) >> PAGE_SHIFT; - vm = mm->total_vm + pgsz; - if (lim < vm) - goto out; - - lim = ACCESS_ONCE(rlim[RLIMIT_MEMLOCK].rlim_cur) >> PAGE_SHIFT; - vm = mm->locked_vm + pgsz; - if (lim < vm) - goto out; - - mm->total_vm += pgsz; - mm->locked_vm += pgsz; - - error = 0; - out: - up_write(&mm->mmap_sem); - return error; -} - -void refund_locked_memory(struct mm_struct *mm, size_t size) -{ - unsigned long pgsz = PAGE_ALIGN(size) >> PAGE_SHIFT; - - down_write(&mm->mmap_sem); - - mm->total_vm -= pgsz; - mm->locked_vm -= pgsz; - - up_write(&mm->mmap_sem); -} diff --git a/tools/perf/Documentation/perf-probe.txt b/tools/perf/Documentation/perf-probe.txt index 34202b1be0bb..bb671b346774 100644 --- a/tools/perf/Documentation/perf-probe.txt +++ b/tools/perf/Documentation/perf-probe.txt @@ -57,6 +57,11 @@ OPTIONS --force:: Forcibly add events with existing name. +-n:: +--dry-run:: + Dry run. With this option, --add and --del doesn't execute actual + adding and removal operations. + PROBE SYNTAX ------------ Probe points are defined by following syntax. @@ -80,7 +85,7 @@ LINE SYNTAX ----------- Line range is descripted by following syntax. - "FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]" + "FUNC[:RLN[+NUM|-RLN2]]|SRC:ALN[+NUM|-ALN2]" FUNC specifies the function name of showing lines. 'RLN' is the start line number from function entry line, and 'RLN2' is the end line number. As same as diff --git a/tools/perf/Makefile b/tools/perf/Makefile index 8a8f52db7e38..690364577611 100644 --- a/tools/perf/Makefile +++ b/tools/perf/Makefile @@ -150,6 +150,8 @@ all:: # Define LDFLAGS=-static to build a static binary. # # Define EXTRA_CFLAGS=-m64 or EXTRA_CFLAGS=-m32 as appropriate for cross-builds. +# +# Define NO_DWARF if you do not want debug-info analysis feature at all. PERF-VERSION-FILE: .FORCE-PERF-VERSION-FILE @$(SHELL_PATH) util/PERF-VERSION-GEN @@ -506,12 +508,21 @@ endif ifneq ($(shell sh -c "(echo '\#include <dwarf.h>'; echo '\#include <libdw.h>'; echo 'int main(void) { Dwarf *dbg; dbg = dwarf_begin(0, DWARF_C_READ); return (long)dbg; }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -I/usr/include/elfutils -ldw -lelf -o $(BITBUCKET) $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) msg := $(warning No libdw.h found or old libdw.h found, disables dwarf support. Please install elfutils-devel/elfutils-dev); - BASIC_CFLAGS += -DNO_DWARF_SUPPORT else - BASIC_CFLAGS += -I/usr/include/elfutils +ifndef NO_DWARF + BASIC_CFLAGS += -I/usr/include/elfutils -DDWARF_SUPPORT EXTLIBS += -lelf -ldw LIB_OBJS += util/probe-finder.o endif +endif + +ifneq ($(shell sh -c "(echo '\#include <newt.h>'; echo 'int main(void) { newtInit(); newtCls(); return newtFinished(); }') | $(CC) -x c - $(ALL_CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -lnewt -o $(BITBUCKET) $(ALL_LDFLAGS) $(EXTLIBS) "$(QUIET_STDERR)" && echo y"), y) + msg := $(warning newt not found, disables TUI support. Please install newt-devel or libnewt-dev); + BASIC_CFLAGS += -DNO_NEWT_SUPPORT +else + EXTLIBS += -lnewt + LIB_OBJS += util/newt.o +endif ifndef NO_LIBPERL PERL_EMBED_LDOPTS = `perl -MExtUtils::Embed -e ldopts 2>/dev/null` diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c index 6ad7148451c5..887e8e04a6f9 100644 --- a/tools/perf/builtin-annotate.c +++ b/tools/perf/builtin-annotate.c @@ -69,13 +69,13 @@ static int sym__alloc_hist(struct symbol *self) static int annotate__hist_hit(struct hist_entry *he, u64 ip) { unsigned int sym_size, offset; - struct symbol *sym = he->sym; + struct symbol *sym = he->ms.sym; struct sym_priv *priv; struct sym_hist *h; he->count++; - if (!sym || !he->map) + if (!sym || !he->ms.map) return 0; priv = symbol__priv(sym); @@ -85,7 +85,7 @@ static int annotate__hist_hit(struct hist_entry *he, u64 ip) sym_size = sym->end - sym->start; offset = ip - sym->start; - pr_debug3("%s: ip=%#Lx\n", __func__, he->map->unmap_ip(he->map, ip)); + pr_debug3("%s: ip=%#Lx\n", __func__, he->ms.map->unmap_ip(he->ms.map, ip)); if (offset >= sym_size) return 0; @@ -94,8 +94,8 @@ static int annotate__hist_hit(struct hist_entry *he, u64 ip) h->sum++; h->ip[offset]++; - pr_debug3("%#Lx %s: count++ [ip: %#Lx, %#Lx] => %Ld\n", he->sym->start, - he->sym->name, ip, ip - he->sym->start, h->ip[offset]); + pr_debug3("%#Lx %s: count++ [ip: %#Lx, %#Lx] => %Ld\n", he->ms.sym->start, + he->ms.sym->name, ip, ip - he->ms.sym->start, h->ip[offset]); return 0; } @@ -187,7 +187,7 @@ static struct objdump_line *objdump__get_next_ip_line(struct list_head *head, static int parse_line(FILE *file, struct hist_entry *he, struct list_head *head) { - struct symbol *sym = he->sym; + struct symbol *sym = he->ms.sym; struct objdump_line *objdump_line; char *line = NULL, *tmp, *tmp2; size_t line_len; @@ -226,7 +226,7 @@ static int parse_line(FILE *file, struct hist_entry *he, } if (line_ip != -1) { - u64 start = map__rip_2objdump(he->map, sym->start); + u64 start = map__rip_2objdump(he->ms.map, sym->start); offset = line_ip - start; } @@ -244,7 +244,7 @@ static int objdump_line__print(struct objdump_line *self, struct list_head *head, struct hist_entry *he, u64 len) { - struct symbol *sym = he->sym; + struct symbol *sym = he->ms.sym; static const char *prev_line; static const char *prev_color; @@ -327,7 +327,7 @@ static void insert_source_line(struct sym_ext *sym_ext) static void free_source_line(struct hist_entry *he, int len) { - struct sym_priv *priv = symbol__priv(he->sym); + struct sym_priv *priv = symbol__priv(he->ms.sym); struct sym_ext *sym_ext = priv->ext; int i; @@ -346,7 +346,7 @@ static void free_source_line(struct hist_entry *he, int len) static void get_source_line(struct hist_entry *he, int len, const char *filename) { - struct symbol *sym = he->sym; + struct symbol *sym = he->ms.sym; u64 start; int i; char cmd[PATH_MAX * 2]; @@ -361,7 +361,7 @@ get_source_line(struct hist_entry *he, int len, const char *filename) if (!priv->ext) return; - start = he->map->unmap_ip(he->map, sym->start); + start = he->ms.map->unmap_ip(he->ms.map, sym->start); for (i = 0; i < len; i++) { char *path = NULL; @@ -425,7 +425,7 @@ static void print_summary(const char *filename) static void hist_entry__print_hits(struct hist_entry *self) { - struct symbol *sym = self->sym; + struct symbol *sym = self->ms.sym; struct sym_priv *priv = symbol__priv(sym); struct sym_hist *h = priv->hist; u64 len = sym->end - sym->start, offset; @@ -439,9 +439,9 @@ static void hist_entry__print_hits(struct hist_entry *self) static void annotate_sym(struct hist_entry *he) { - struct map *map = he->map; + struct map *map = he->ms.map; struct dso *dso = map->dso; - struct symbol *sym = he->sym; + struct symbol *sym = he->ms.sym; const char *filename = dso->long_name, *d_filename; u64 len; char command[PATH_MAX*2]; @@ -452,6 +452,16 @@ static void annotate_sym(struct hist_entry *he) if (!filename) return; + if (dso->origin == DSO__ORIG_KERNEL) { + if (dso->annotate_warned) + return; + dso->annotate_warned = 1; + pr_err("Can't annotate %s: No vmlinux file was found in the " + "path:\n", sym->name); + vmlinux_path__fprintf(stderr); + return; + } + pr_debug("%s: filename=%s, sym=%s, start=%#Lx, end=%#Lx\n", __func__, filename, sym->name, map->unmap_ip(map, sym->start), map->unmap_ip(map, sym->end)); @@ -516,17 +526,17 @@ static void perf_session__find_annotations(struct perf_session *self) struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node); struct sym_priv *priv; - if (he->sym == NULL) + if (he->ms.sym == NULL) continue; - priv = symbol__priv(he->sym); + priv = symbol__priv(he->ms.sym); if (priv->hist == NULL) continue; annotate_sym(he); /* * Since we have a hist_entry per IP for the same symbol, free - * he->sym->hist to signal we already processed this symbol. + * he->ms.sym->hist to signal we already processed this symbol. */ free(priv->hist); priv->hist = NULL; @@ -581,6 +591,8 @@ static const char * const annotate_usage[] = { static const struct option options[] = { OPT_STRING('i', "input", &input_name, "file", "input file name"), + OPT_STRING('d', "dsos", &symbol_conf.dso_list_str, "dso[,dso...]", + "only consider symbols in these dsos"), OPT_STRING('s', "symbol", &sym_hist_filter, "symbol", "symbol to annotate"), OPT_BOOLEAN('f', "force", &force, "don't complain, do it"), diff --git a/tools/perf/builtin-probe.c b/tools/perf/builtin-probe.c index 152d6c9b1fa4..b3ba25a910fa 100644 --- a/tools/perf/builtin-probe.c +++ b/tools/perf/builtin-probe.c @@ -36,11 +36,9 @@ #include "builtin.h" #include "util/util.h" #include "util/strlist.h" -#include "util/event.h" +#include "util/symbol.h" #include "util/debug.h" #include "util/debugfs.h" -#include "util/symbol.h" -#include "util/thread.h" #include "util/parse-options.h" #include "util/parse-events.h" /* For debugfs_path */ #include "util/probe-finder.h" @@ -50,32 +48,29 @@ /* Session management structure */ static struct { - bool need_dwarf; bool list_events; bool force_add; bool show_lines; - int nr_probe; - struct probe_point probes[MAX_PROBES]; + int nevents; + struct perf_probe_event events[MAX_PROBES]; struct strlist *dellist; - struct map_groups kmap_groups; - struct map *kmaps[MAP__NR_TYPES]; struct line_range line_range; -} session; +} params; /* Parse an event definition. Note that any error must die. */ static void parse_probe_event(const char *str) { - struct probe_point *pp = &session.probes[session.nr_probe]; + struct perf_probe_event *pev = ¶ms.events[params.nevents]; - pr_debug("probe-definition(%d): %s\n", session.nr_probe, str); - if (++session.nr_probe == MAX_PROBES) + pr_debug("probe-definition(%d): %s\n", params.nevents, str); + if (++params.nevents == MAX_PROBES) die("Too many probes (> %d) are specified.", MAX_PROBES); - /* Parse perf-probe event into probe_point */ - parse_perf_probe_event(str, pp, &session.need_dwarf); + /* Parse a perf-probe command into event */ + parse_perf_probe_command(str, pev); - pr_debug("%d arguments\n", pp->nr_args); + pr_debug("%d arguments\n", pev->nargs); } static void parse_probe_event_argv(int argc, const char **argv) @@ -87,9 +82,7 @@ static void parse_probe_event_argv(int argc, const char **argv) len = 0; for (i = 0; i < argc; i++) len += strlen(argv[i]) + 1; - buf = zalloc(len + 1); - if (!buf) - die("Failed to allocate memory for binding arguments."); + buf = xzalloc(len + 1); len = 0; for (i = 0; i < argc; i++) len += sprintf(&buf[len], "%s ", argv[i]); @@ -109,43 +102,21 @@ static int opt_del_probe_event(const struct option *opt __used, const char *str, int unset __used) { if (str) { - if (!session.dellist) - session.dellist = strlist__new(true, NULL); - strlist__add(session.dellist, str); + if (!params.dellist) + params.dellist = strlist__new(true, NULL); + strlist__add(params.dellist, str); } return 0; } -/* Currently just checking function name from symbol map */ -static void evaluate_probe_point(struct probe_point *pp) -{ - struct symbol *sym; - sym = map__find_symbol_by_name(session.kmaps[MAP__FUNCTION], - pp->function, NULL); - if (!sym) - die("Kernel symbol \'%s\' not found - probe not added.", - pp->function); -} - -#ifndef NO_DWARF_SUPPORT -static int open_vmlinux(void) -{ - if (map__load(session.kmaps[MAP__FUNCTION], NULL) < 0) { - pr_debug("Failed to load kernel map.\n"); - return -EINVAL; - } - pr_debug("Try to open %s\n", - session.kmaps[MAP__FUNCTION]->dso->long_name); - return open(session.kmaps[MAP__FUNCTION]->dso->long_name, O_RDONLY); -} - +#ifdef DWARF_SUPPORT static int opt_show_lines(const struct option *opt __used, const char *str, int unset __used) { if (str) - parse_line_range_desc(str, &session.line_range); - INIT_LIST_HEAD(&session.line_range.line_list); - session.show_lines = true; + parse_line_range_desc(str, ¶ms.line_range); + INIT_LIST_HEAD(¶ms.line_range.line_list); + params.show_lines = true; return 0; } #endif @@ -155,7 +126,7 @@ static const char * const probe_usage[] = { "perf probe [<options>] --add 'PROBEDEF' [--add 'PROBEDEF' ...]", "perf probe [<options>] --del '[GROUP:]EVENT' ...", "perf probe --list", -#ifndef NO_DWARF_SUPPORT +#ifdef DWARF_SUPPORT "perf probe --line 'LINEDESC'", #endif NULL @@ -164,20 +135,16 @@ static const char * const probe_usage[] = { static const struct option options[] = { OPT_BOOLEAN('v', "verbose", &verbose, "be more verbose (show parsed arguments, etc)"), -#ifndef NO_DWARF_SUPPORT - OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, - "file", "vmlinux pathname"), -#endif - OPT_BOOLEAN('l', "list", &session.list_events, + OPT_BOOLEAN('l', "list", ¶ms.list_events, "list up current probe events"), OPT_CALLBACK('d', "del", NULL, "[GROUP:]EVENT", "delete a probe event.", opt_del_probe_event), OPT_CALLBACK('a', "add", NULL, -#ifdef NO_DWARF_SUPPORT - "[EVENT=]FUNC[+OFF|%return] [ARG ...]", -#else +#ifdef DWARF_SUPPORT "[EVENT=]FUNC[@SRC][+OFF|%return|:RL|;PT]|SRC:AL|SRC;PT" " [ARG ...]", +#else + "[EVENT=]FUNC[+OFF|%return] [ARG ...]", #endif "probe point definition, where\n" "\t\tGROUP:\tGroup name (optional)\n" @@ -185,52 +152,32 @@ static const struct option options[] = { "\t\tFUNC:\tFunction name\n" "\t\tOFF:\tOffset from function entry (in byte)\n" "\t\t%return:\tPut the probe at function return\n" -#ifdef NO_DWARF_SUPPORT - "\t\tARG:\tProbe argument (only \n" -#else +#ifdef DWARF_SUPPORT "\t\tSRC:\tSource code path\n" "\t\tRL:\tRelative line number from function entry.\n" "\t\tAL:\tAbsolute line number in file.\n" "\t\tPT:\tLazy expression of line code.\n" "\t\tARG:\tProbe argument (local variable name or\n" -#endif "\t\t\tkprobe-tracer argument format.)\n", +#else + "\t\tARG:\tProbe argument (kprobe-tracer argument format.)\n", +#endif opt_add_probe_event), - OPT_BOOLEAN('f', "force", &session.force_add, "forcibly add events" + OPT_BOOLEAN('f', "force", ¶ms.force_add, "forcibly add events" " with existing name"), -#ifndef NO_DWARF_SUPPORT +#ifdef DWARF_SUPPORT OPT_CALLBACK('L', "line", NULL, - "FUNC[:RLN[+NUM|:RLN2]]|SRC:ALN[+NUM|:ALN2]", + "FUNC[:RLN[+NUM|-RLN2]]|SRC:ALN[+NUM|-ALN2]", "Show source code lines.", opt_show_lines), + OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name, + "file", "vmlinux pathname"), #endif + OPT__DRY_RUN(&probe_event_dry_run), OPT_END() }; -/* Initialize symbol maps for vmlinux */ -static void init_vmlinux(void) -{ - symbol_conf.sort_by_name = true; - if (symbol_conf.vmlinux_name == NULL) - symbol_conf.try_vmlinux_path = true; - else - pr_debug("Use vmlinux: %s\n", symbol_conf.vmlinux_name); - if (symbol__init() < 0) - die("Failed to init symbol map."); - - map_groups__init(&session.kmap_groups); - if (map_groups__create_kernel_maps(&session.kmap_groups, - session.kmaps) < 0) - die("Failed to create kernel maps."); -} - int cmd_probe(int argc, const char **argv, const char *prefix __used) { - int i, ret; -#ifndef NO_DWARF_SUPPORT - int fd; -#endif - struct probe_point *pp; - argc = parse_options(argc, argv, options, probe_usage, PARSE_OPT_STOP_AT_NON_OPTION); if (argc > 0) { @@ -241,20 +188,20 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) parse_probe_event_argv(argc, argv); } - if ((!session.nr_probe && !session.dellist && !session.list_events && - !session.show_lines)) + if ((!params.nevents && !params.dellist && !params.list_events && + !params.show_lines)) usage_with_options(probe_usage, options); if (debugfs_valid_mountpoint(debugfs_path) < 0) die("Failed to find debugfs path."); - if (session.list_events) { - if (session.nr_probe != 0 || session.dellist) { + if (params.list_events) { + if (params.nevents != 0 || params.dellist) { pr_warning(" Error: Don't use --list with" " --add/--del.\n"); usage_with_options(probe_usage, options); } - if (session.show_lines) { + if (params.show_lines) { pr_warning(" Error: Don't use --list with --line.\n"); usage_with_options(probe_usage, options); } @@ -262,99 +209,27 @@ int cmd_probe(int argc, const char **argv, const char *prefix __used) return 0; } -#ifndef NO_DWARF_SUPPORT - if (session.show_lines) { - if (session.nr_probe != 0 || session.dellist) { +#ifdef DWARF_SUPPORT + if (params.show_lines) { + if (params.nevents != 0 || params.dellist) { pr_warning(" Error: Don't use --line with" " --add/--del.\n"); usage_with_options(probe_usage, options); } - init_vmlinux(); - fd = open_vmlinux(); - if (fd < 0) - die("Could not open debuginfo file."); - ret = find_line_range(fd, &session.line_range); - if (ret <= 0) - die("Source line is not found.\n"); - close(fd); - show_line_range(&session.line_range); + + show_line_range(¶ms.line_range); return 0; } #endif - if (session.dellist) { - del_trace_kprobe_events(session.dellist); - strlist__delete(session.dellist); - if (session.nr_probe == 0) + if (params.dellist) { + del_perf_probe_events(params.dellist); + strlist__delete(params.dellist); + if (params.nevents == 0) return 0; } - /* Add probes */ - init_vmlinux(); - - if (session.need_dwarf) -#ifdef NO_DWARF_SUPPORT - die("Debuginfo-analysis is not supported"); -#else /* !NO_DWARF_SUPPORT */ - pr_debug("Some probes require debuginfo.\n"); - - fd = open_vmlinux(); - if (fd < 0) { - if (session.need_dwarf) - die("Could not open debuginfo file."); - - pr_debug("Could not open vmlinux/module file." - " Try to use symbols.\n"); - goto end_dwarf; - } - - /* Searching probe points */ - for (i = 0; i < session.nr_probe; i++) { - pp = &session.probes[i]; - if (pp->found) - continue; - - lseek(fd, SEEK_SET, 0); - ret = find_probe_point(fd, pp); - if (ret > 0) - continue; - if (ret == 0) { /* No error but failed to find probe point. */ - synthesize_perf_probe_point(pp); - die("Probe point '%s' not found. - probe not added.", - pp->probes[0]); - } - /* Error path */ - if (session.need_dwarf) { - if (ret == -ENOENT) - pr_warning("No dwarf info found in the vmlinux - please rebuild with CONFIG_DEBUG_INFO=y.\n"); - die("Could not analyze debuginfo."); - } - pr_debug("An error occurred in debuginfo analysis." - " Try to use symbols.\n"); - break; - } - close(fd); - -end_dwarf: -#endif /* !NO_DWARF_SUPPORT */ - - /* Synthesize probes without dwarf */ - for (i = 0; i < session.nr_probe; i++) { - pp = &session.probes[i]; - if (pp->found) /* This probe is already found. */ - continue; - - evaluate_probe_point(pp); - ret = synthesize_trace_kprobe_event(pp); - if (ret == -E2BIG) - die("probe point definition becomes too long."); - else if (ret < 0) - die("Failed to synthesize a probe point."); - } - - /* Settng up probe points */ - add_trace_kprobe_events(session.probes, session.nr_probe, - session.force_add); + add_perf_probe_events(params.events, params.nevents, params.force_add); return 0; } diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c index 3b8b6387c47c..60ecdd3dd26d 100644 --- a/tools/perf/builtin-record.c +++ b/tools/perf/builtin-record.c @@ -27,7 +27,7 @@ #include <unistd.h> #include <sched.h> -static int fd[MAX_NR_CPUS][MAX_COUNTERS]; +static int *fd[MAX_NR_CPUS][MAX_COUNTERS]; static long default_interval = 0; @@ -43,6 +43,9 @@ static int raw_samples = 0; static int system_wide = 0; static int profile_cpu = -1; static pid_t target_pid = -1; +static pid_t target_tid = -1; +static pid_t *all_tids = NULL; +static int thread_num = 0; static pid_t child_pid = -1; static int inherit = 1; static int force = 0; @@ -60,7 +63,7 @@ static struct timeval this_read; static u64 bytes_written = 0; -static struct pollfd event_array[MAX_NR_CPUS * MAX_COUNTERS]; +static struct pollfd *event_array; static int nr_poll = 0; static int nr_cpu = 0; @@ -77,7 +80,7 @@ struct mmap_data { unsigned int prev; }; -static struct mmap_data mmap_array[MAX_NR_CPUS][MAX_COUNTERS]; +static struct mmap_data *mmap_array[MAX_NR_CPUS][MAX_COUNTERS]; static unsigned long mmap_read_head(struct mmap_data *md) { @@ -225,12 +228,13 @@ static struct perf_header_attr *get_header_attr(struct perf_event_attr *a, int n return h_attr; } -static void create_counter(int counter, int cpu, pid_t pid) +static void create_counter(int counter, int cpu) { char *filter = filters[counter]; struct perf_event_attr *attr = attrs + counter; struct perf_header_attr *h_attr; int track = !counter; /* only the first counter needs these */ + int thread_index; int ret; struct { u64 count; @@ -275,118 +279,129 @@ static void create_counter(int counter, int cpu, pid_t pid) attr->mmap = track; attr->comm = track; attr->inherit = inherit; - attr->disabled = 1; + if (target_pid == -1 && !system_wide) { + attr->disabled = 1; + attr->enable_on_exec = 1; + } + for (thread_index = 0; thread_index < thread_num; thread_index++) { try_again: - fd[nr_cpu][counter] = sys_perf_event_open(attr, pid, cpu, group_fd, 0); - - if (fd[nr_cpu][counter] < 0) { - int err = errno; - - if (err == EPERM || err == EACCES) - die("Permission error - are you root?\n"); - else if (err == ENODEV && profile_cpu != -1) - die("No such device - did you specify an out-of-range profile CPU?\n"); + fd[nr_cpu][counter][thread_index] = sys_perf_event_open(attr, + all_tids[thread_index], cpu, group_fd, 0); + + if (fd[nr_cpu][counter][thread_index] < 0) { + int err = errno; + + if (err == EPERM || err == EACCES) + die("Permission error - are you root?\n" + "\t Consider tweaking" + " /proc/sys/kernel/perf_event_paranoid.\n"); + else if (err == ENODEV && profile_cpu != -1) { + die("No such device - did you specify" + " an out-of-range profile CPU?\n"); + } - /* - * If it's cycles then fall back to hrtimer - * based cpu-clock-tick sw counter, which - * is always available even if no PMU support: - */ - if (attr->type == PERF_TYPE_HARDWARE - && attr->config == PERF_COUNT_HW_CPU_CYCLES) { - - if (verbose) - warning(" ... trying to fall back to cpu-clock-ticks\n"); - attr->type = PERF_TYPE_SOFTWARE; - attr->config = PERF_COUNT_SW_CPU_CLOCK; - goto try_again; - } - printf("\n"); - error("perfcounter syscall returned with %d (%s)\n", - fd[nr_cpu][counter], strerror(err)); + /* + * If it's cycles then fall back to hrtimer + * based cpu-clock-tick sw counter, which + * is always available even if no PMU support: + */ + if (attr->type == PERF_TYPE_HARDWARE + && attr->config == PERF_COUNT_HW_CPU_CYCLES) { + + if (verbose) + warning(" ... trying to fall back to cpu-clock-ticks\n"); + attr->type = PERF_TYPE_SOFTWARE; + attr->config = PERF_COUNT_SW_CPU_CLOCK; + goto try_again; + } + printf("\n"); + error("perfcounter syscall returned with %d (%s)\n", + fd[nr_cpu][counter][thread_index], strerror(err)); #if defined(__i386__) || defined(__x86_64__) - if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP) - die("No hardware sampling interrupt available. No APIC? If so then you can boot the kernel with the \"lapic\" boot parameter to force-enable it.\n"); + if (attr->type == PERF_TYPE_HARDWARE && err == EOPNOTSUPP) + die("No hardware sampling interrupt available." + " No APIC? If so then you can boot the kernel" + " with the \"lapic\" boot parameter to" + " force-enable it.\n"); #endif - die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); - exit(-1); - } + die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); + exit(-1); + } - h_attr = get_header_attr(attr, counter); - if (h_attr == NULL) - die("nomem\n"); + h_attr = get_header_attr(attr, counter); + if (h_attr == NULL) + die("nomem\n"); - if (!file_new) { - if (memcmp(&h_attr->attr, attr, sizeof(*attr))) { - fprintf(stderr, "incompatible append\n"); - exit(-1); + if (!file_new) { + if (memcmp(&h_attr->attr, attr, sizeof(*attr))) { + fprintf(stderr, "incompatible append\n"); + exit(-1); + } } - } - if (read(fd[nr_cpu][counter], &read_data, sizeof(read_data)) == -1) { - perror("Unable to read perf file descriptor\n"); - exit(-1); - } + if (read(fd[nr_cpu][counter][thread_index], &read_data, sizeof(read_data)) == -1) { + perror("Unable to read perf file descriptor\n"); + exit(-1); + } - if (perf_header_attr__add_id(h_attr, read_data.id) < 0) { - pr_warning("Not enough memory to add id\n"); - exit(-1); - } + if (perf_header_attr__add_id(h_attr, read_data.id) < 0) { + pr_warning("Not enough memory to add id\n"); + exit(-1); + } - assert(fd[nr_cpu][counter] >= 0); - fcntl(fd[nr_cpu][counter], F_SETFL, O_NONBLOCK); + assert(fd[nr_cpu][counter][thread_index] >= 0); + fcntl(fd[nr_cpu][counter][thread_index], F_SETFL, O_NONBLOCK); - /* - * First counter acts as the group leader: - */ - if (group && group_fd == -1) - group_fd = fd[nr_cpu][counter]; - if (multiplex && multiplex_fd == -1) - multiplex_fd = fd[nr_cpu][counter]; + /* + * First counter acts as the group leader: + */ + if (group && group_fd == -1) + group_fd = fd[nr_cpu][counter][thread_index]; + if (multiplex && multiplex_fd == -1) + multiplex_fd = fd[nr_cpu][counter][thread_index]; - if (multiplex && fd[nr_cpu][counter] != multiplex_fd) { + if (multiplex && fd[nr_cpu][counter][thread_index] != multiplex_fd) { - ret = ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_SET_OUTPUT, multiplex_fd); - assert(ret != -1); - } else { - event_array[nr_poll].fd = fd[nr_cpu][counter]; - event_array[nr_poll].events = POLLIN; - nr_poll++; - - mmap_array[nr_cpu][counter].counter = counter; - mmap_array[nr_cpu][counter].prev = 0; - mmap_array[nr_cpu][counter].mask = mmap_pages*page_size - 1; - mmap_array[nr_cpu][counter].base = mmap(NULL, (mmap_pages+1)*page_size, - PROT_READ|PROT_WRITE, MAP_SHARED, fd[nr_cpu][counter], 0); - if (mmap_array[nr_cpu][counter].base == MAP_FAILED) { - error("failed to mmap with %d (%s)\n", errno, strerror(errno)); - exit(-1); + ret = ioctl(fd[nr_cpu][counter][thread_index], PERF_EVENT_IOC_SET_OUTPUT, multiplex_fd); + assert(ret != -1); + } else { + event_array[nr_poll].fd = fd[nr_cpu][counter][thread_index]; + event_array[nr_poll].events = POLLIN; + nr_poll++; + + mmap_array[nr_cpu][counter][thread_index].counter = counter; + mmap_array[nr_cpu][counter][thread_index].prev = 0; + mmap_array[nr_cpu][counter][thread_index].mask = mmap_pages*page_size - 1; + mmap_array[nr_cpu][counter][thread_index].base = mmap(NULL, (mmap_pages+1)*page_size, + PROT_READ|PROT_WRITE, MAP_SHARED, fd[nr_cpu][counter][thread_index], 0); + if (mmap_array[nr_cpu][counter][thread_index].base == MAP_FAILED) { + error("failed to mmap with %d (%s)\n", errno, strerror(errno)); + exit(-1); + } } - } - if (filter != NULL) { - ret = ioctl(fd[nr_cpu][counter], - PERF_EVENT_IOC_SET_FILTER, filter); - if (ret) { - error("failed to set filter with %d (%s)\n", errno, - strerror(errno)); - exit(-1); + if (filter != NULL) { + ret = ioctl(fd[nr_cpu][counter][thread_index], + PERF_EVENT_IOC_SET_FILTER, filter); + if (ret) { + error("failed to set filter with %d (%s)\n", errno, + strerror(errno)); + exit(-1); + } } } - - ioctl(fd[nr_cpu][counter], PERF_EVENT_IOC_ENABLE); } -static void open_counters(int cpu, pid_t pid) +static void open_counters(int cpu) { int counter; group_fd = -1; for (counter = 0; counter < nr_counters; counter++) - create_counter(counter, cpu, pid); + create_counter(counter, cpu); nr_cpu++; } @@ -421,7 +436,7 @@ static int __cmd_record(int argc, const char **argv) int err; unsigned long waking = 0; int child_ready_pipe[2], go_pipe[2]; - const bool forks = target_pid == -1 && argc > 0; + const bool forks = argc > 0; char buf; page_size = sysconf(_SC_PAGE_SIZE); @@ -492,13 +507,13 @@ static int __cmd_record(int argc, const char **argv) atexit(atexit_header); if (forks) { - pid = fork(); + child_pid = fork(); if (pid < 0) { perror("failed to fork"); exit(-1); } - if (!pid) { + if (!child_pid) { close(child_ready_pipe[0]); close(go_pipe[1]); fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC); @@ -527,10 +542,8 @@ static int __cmd_record(int argc, const char **argv) exit(-1); } - child_pid = pid; - - if (!system_wide) - target_pid = pid; + if (!system_wide && target_tid == -1 && target_pid == -1) + all_tids[0] = child_pid; close(child_ready_pipe[1]); close(go_pipe[0]); @@ -544,13 +557,12 @@ static int __cmd_record(int argc, const char **argv) close(child_ready_pipe[0]); } - if ((!system_wide && !inherit) || profile_cpu != -1) { - open_counters(profile_cpu, target_pid); + open_counters(profile_cpu); } else { nr_cpus = read_cpu_map(); for (i = 0; i < nr_cpus; i++) - open_counters(cpumap[i], target_pid); + open_counters(cpumap[i]); } if (file_new) { @@ -575,7 +587,7 @@ static int __cmd_record(int argc, const char **argv) } if (!system_wide && profile_cpu == -1) - event__synthesize_thread(target_pid, process_synthesized_event, + event__synthesize_thread(target_tid, process_synthesized_event, session); else event__synthesize_threads(process_synthesized_event, session); @@ -598,11 +610,16 @@ static int __cmd_record(int argc, const char **argv) for (;;) { int hits = samples; + int thread; for (i = 0; i < nr_cpu; i++) { for (counter = 0; counter < nr_counters; counter++) { - if (mmap_array[i][counter].base) - mmap_read(&mmap_array[i][counter]); + for (thread = 0; + thread < thread_num; thread++) { + if (mmap_array[i][counter][thread].base) + mmap_read(&mmap_array[i][counter][thread]); + } + } } @@ -615,8 +632,15 @@ static int __cmd_record(int argc, const char **argv) if (done) { for (i = 0; i < nr_cpu; i++) { - for (counter = 0; counter < nr_counters; counter++) - ioctl(fd[i][counter], PERF_EVENT_IOC_DISABLE); + for (counter = 0; + counter < nr_counters; + counter++) { + for (thread = 0; + thread < thread_num; + thread++) + ioctl(fd[i][counter][thread], + PERF_EVENT_IOC_DISABLE); + } } } } @@ -648,7 +672,9 @@ static const struct option options[] = { OPT_CALLBACK(0, "filter", NULL, "filter", "event filter", parse_filter), OPT_INTEGER('p', "pid", &target_pid, - "record events on existing pid"), + "record events on existing process id"), + OPT_INTEGER('t', "tid", &target_tid, + "record events on existing thread id"), OPT_INTEGER('r', "realtime", &realtime_prio, "collect data with this RT SCHED_FIFO priority"), OPT_BOOLEAN('R', "raw-samples", &raw_samples, @@ -689,10 +715,12 @@ static const struct option options[] = { int cmd_record(int argc, const char **argv, const char *prefix __used) { int counter; + int i,j; argc = parse_options(argc, argv, options, record_usage, PARSE_OPT_STOP_AT_NON_OPTION); - if (!argc && target_pid == -1 && !system_wide && profile_cpu == -1) + if (!argc && target_pid == -1 && target_tid == -1 && + !system_wide && profile_cpu == -1) usage_with_options(record_usage, options); symbol__init(); @@ -703,6 +731,37 @@ int cmd_record(int argc, const char **argv, const char *prefix __used) attrs[0].config = PERF_COUNT_HW_CPU_CYCLES; } + if (target_pid != -1) { + target_tid = target_pid; + thread_num = find_all_tid(target_pid, &all_tids); + if (thread_num <= 0) { + fprintf(stderr, "Can't find all threads of pid %d\n", + target_pid); + usage_with_options(record_usage, options); + } + } else { + all_tids=malloc(sizeof(pid_t)); + if (!all_tids) + return -ENOMEM; + + all_tids[0] = target_tid; + thread_num = 1; + } + + for (i = 0; i < MAX_NR_CPUS; i++) { + for (j = 0; j < MAX_COUNTERS; j++) { + fd[i][j] = malloc(sizeof(int)*thread_num); + mmap_array[i][j] = zalloc( + sizeof(struct mmap_data)*thread_num); + if (!fd[i][j] || !mmap_array[i][j]) + return -ENOMEM; + } + } + event_array = malloc( + sizeof(struct pollfd)*MAX_NR_CPUS*MAX_COUNTERS*thread_num); + if (!event_array) + return -ENOMEM; + /* * User specified count overrides default frequency. */ diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c index f815de25d0fc..6ab16980dd66 100644 --- a/tools/perf/builtin-report.c +++ b/tools/perf/builtin-report.c @@ -81,8 +81,10 @@ static int perf_session__add_hist_entry(struct perf_session *self, struct addr_location *al, struct sample_data *data) { - struct symbol **syms = NULL, *parent = NULL; + struct map_symbol *syms = NULL; + struct symbol *parent = NULL; bool hit; + int err; struct hist_entry *he; struct event_stat_id *stats; struct perf_event_attr *attr; @@ -109,8 +111,11 @@ static int perf_session__add_hist_entry(struct perf_session *self, if (symbol_conf.use_callchain) { if (!hit) callchain_init(&he->callchain); - append_chain(&he->callchain, data->callchain, syms); + err = append_chain(&he->callchain, data->callchain, syms); free(syms); + + if (err) + return err; } return 0; @@ -267,6 +272,7 @@ static int __cmd_report(void) int ret = -EINVAL; struct perf_session *session; struct rb_node *next; + const char *help = "For a higher level overview, try: perf report --sort comm,dso"; session = perf_session__new(input_name, O_RDONLY, force); if (session == NULL) @@ -301,30 +307,38 @@ static int __cmd_report(void) stats = rb_entry(next, struct event_stat_id, rb_node); perf_session__collapse_resort(&stats->hists); perf_session__output_resort(&stats->hists, stats->stats.total); - if (rb_first(&session->stats_by_id) == - rb_last(&session->stats_by_id)) - fprintf(stdout, "# Samples: %Ld\n#\n", - stats->stats.total); - else - fprintf(stdout, "# Samples: %Ld %s\n#\n", - stats->stats.total, - __event_name(stats->type, stats->config)); - perf_session__fprintf_hists(&stats->hists, NULL, false, stdout, + if (use_browser) + perf_session__browse_hists(&stats->hists, + stats->stats.total, help); + else { + if (rb_first(&session->stats_by_id) == + rb_last(&session->stats_by_id)) + fprintf(stdout, "# Samples: %Ld\n#\n", + stats->stats.total); + else + fprintf(stdout, "# Samples: %Ld %s\n#\n", + stats->stats.total, + __event_name(stats->type, stats->config)); + + perf_session__fprintf_hists(&stats->hists, NULL, false, stdout, stats->stats.total); - fprintf(stdout, "\n\n"); + fprintf(stdout, "\n\n"); + } + next = rb_next(&stats->rb_node); } - if (sort_order == default_sort_order && - parent_pattern == default_parent_pattern) - fprintf(stdout, "#\n# (For a higher level overview, try: perf report --sort comm,dso)\n#\n"); + if (!use_browser && sort_order == default_sort_order && + parent_pattern == default_parent_pattern) { + fprintf(stdout, "#\n# (%s)\n#\n", help); - if (show_threads) { - bool raw_printing_style = !strcmp(pretty_printing_style, "raw"); - perf_read_values_display(stdout, &show_threads_values, - raw_printing_style); - perf_read_values_destroy(&show_threads_values); + if (show_threads) { + bool style = !strcmp(pretty_printing_style, "raw"); + perf_read_values_display(stdout, &show_threads_values, + style); + perf_read_values_destroy(&show_threads_values); + } } out_delete: perf_session__delete(session); @@ -447,7 +461,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __used) { argc = parse_options(argc, argv, options, report_usage, 0); - setup_pager(); + setup_browser(); if (symbol__init() < 0) return -1; diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c index 95db31cff6fd..1036ca739e6f 100644 --- a/tools/perf/builtin-stat.c +++ b/tools/perf/builtin-stat.c @@ -46,6 +46,7 @@ #include "util/debug.h" #include "util/header.h" #include "util/cpumap.h" +#include "util/thread.h" #include <sys/prctl.h> #include <math.h> @@ -74,10 +75,13 @@ static int run_count = 1; static int inherit = 1; static int scale = 1; static pid_t target_pid = -1; +static pid_t target_tid = -1; +static pid_t *all_tids = NULL; +static int thread_num = 0; static pid_t child_pid = -1; static int null_run = 0; -static int fd[MAX_NR_CPUS][MAX_COUNTERS]; +static int *fd[MAX_NR_CPUS][MAX_COUNTERS]; static int event_scaled[MAX_COUNTERS]; @@ -140,9 +144,11 @@ struct stats runtime_branches_stats; #define ERR_PERF_OPEN \ "Error: counter %d, sys_perf_event_open() syscall returned with %d (%s)\n" -static void create_perf_stat_counter(int counter, int pid) +static int create_perf_stat_counter(int counter) { struct perf_event_attr *attr = attrs + counter; + int thread; + int ncreated = 0; if (scale) attr->read_format = PERF_FORMAT_TOTAL_TIME_ENABLED | @@ -152,21 +158,33 @@ static void create_perf_stat_counter(int counter, int pid) unsigned int cpu; for (cpu = 0; cpu < nr_cpus; cpu++) { - fd[cpu][counter] = sys_perf_event_open(attr, -1, cpumap[cpu], -1, 0); - if (fd[cpu][counter] < 0 && verbose) - fprintf(stderr, ERR_PERF_OPEN, counter, - fd[cpu][counter], strerror(errno)); + fd[cpu][counter][0] = sys_perf_event_open(attr, + -1, cpumap[cpu], -1, 0); + if (fd[cpu][counter][0] < 0) + pr_debug(ERR_PERF_OPEN, counter, + fd[cpu][counter][0], strerror(errno)); + else + ++ncreated; } } else { attr->inherit = inherit; - attr->disabled = 1; - attr->enable_on_exec = 1; - - fd[0][counter] = sys_perf_event_open(attr, pid, -1, -1, 0); - if (fd[0][counter] < 0 && verbose) - fprintf(stderr, ERR_PERF_OPEN, counter, - fd[0][counter], strerror(errno)); + if (target_pid == -1) { + attr->disabled = 1; + attr->enable_on_exec = 1; + } + for (thread = 0; thread < thread_num; thread++) { + fd[0][counter][thread] = sys_perf_event_open(attr, + all_tids[thread], -1, -1, 0); + if (fd[0][counter][thread] < 0) + pr_debug(ERR_PERF_OPEN, counter, + fd[0][counter][thread], + strerror(errno)); + else + ++ncreated; + } } + + return ncreated; } /* @@ -190,25 +208,28 @@ static void read_counter(int counter) unsigned int cpu; size_t res, nv; int scaled; - int i; + int i, thread; count[0] = count[1] = count[2] = 0; nv = scale ? 3 : 1; for (cpu = 0; cpu < nr_cpus; cpu++) { - if (fd[cpu][counter] < 0) - continue; - - res = read(fd[cpu][counter], single_count, nv * sizeof(u64)); - assert(res == nv * sizeof(u64)); - - close(fd[cpu][counter]); - fd[cpu][counter] = -1; - - count[0] += single_count[0]; - if (scale) { - count[1] += single_count[1]; - count[2] += single_count[2]; + for (thread = 0; thread < thread_num; thread++) { + if (fd[cpu][counter][thread] < 0) + continue; + + res = read(fd[cpu][counter][thread], + single_count, nv * sizeof(u64)); + assert(res == nv * sizeof(u64)); + + close(fd[cpu][counter][thread]); + fd[cpu][counter][thread] = -1; + + count[0] += single_count[0]; + if (scale) { + count[1] += single_count[1]; + count[2] += single_count[2]; + } } } @@ -250,10 +271,9 @@ static int run_perf_stat(int argc __used, const char **argv) { unsigned long long t0, t1; int status = 0; - int counter; - int pid = target_pid; + int counter, ncreated = 0; int child_ready_pipe[2], go_pipe[2]; - const bool forks = (target_pid == -1 && argc > 0); + const bool forks = (argc > 0); char buf; if (!system_wide) @@ -265,10 +285,10 @@ static int run_perf_stat(int argc __used, const char **argv) } if (forks) { - if ((pid = fork()) < 0) + if ((child_pid = fork()) < 0) perror("failed to fork"); - if (!pid) { + if (!child_pid) { close(child_ready_pipe[0]); close(go_pipe[1]); fcntl(go_pipe[0], F_SETFD, FD_CLOEXEC); @@ -297,7 +317,8 @@ static int run_perf_stat(int argc __used, const char **argv) exit(-1); } - child_pid = pid; + if (target_tid == -1 && target_pid == -1 && !system_wide) + all_tids[0] = child_pid; /* * Wait for the child to be ready to exec. @@ -310,7 +331,16 @@ static int run_perf_stat(int argc __used, const char **argv) } for (counter = 0; counter < nr_counters; counter++) - create_perf_stat_counter(counter, pid); + ncreated += create_perf_stat_counter(counter); + + if (ncreated == 0) { + pr_err("No permission to collect %sstats.\n" + "Consider tweaking /proc/sys/kernel/perf_event_paranoid.\n", + system_wide ? "system-wide " : ""); + if (child_pid != -1) + kill(child_pid, SIGTERM); + return -1; + } /* * Enable counters and exec the command: @@ -321,7 +351,7 @@ static int run_perf_stat(int argc __used, const char **argv) close(go_pipe[1]); wait(&status); } else { - while(!done); + while(!done) sleep(1); } t1 = rdclock(); @@ -429,12 +459,14 @@ static void print_stat(int argc, const char **argv) fprintf(stderr, "\n"); fprintf(stderr, " Performance counter stats for "); - if(target_pid == -1) { + if(target_pid == -1 && target_tid == -1) { fprintf(stderr, "\'%s", argv[0]); for (i = 1; i < argc; i++) fprintf(stderr, " %s", argv[i]); - }else - fprintf(stderr, "task pid \'%d", target_pid); + } else if (target_pid != -1) + fprintf(stderr, "process id \'%d", target_pid); + else + fprintf(stderr, "thread id \'%d", target_tid); fprintf(stderr, "\'"); if (run_count > 1) @@ -459,7 +491,7 @@ static volatile int signr = -1; static void skip_signal(int signo) { - if(target_pid != -1) + if(child_pid == -1) done = 1; signr = signo; @@ -489,7 +521,9 @@ static const struct option options[] = { OPT_BOOLEAN('i', "inherit", &inherit, "child tasks inherit counters"), OPT_INTEGER('p', "pid", &target_pid, - "stat events on existing pid"), + "stat events on existing process id"), + OPT_INTEGER('t', "tid", &target_tid, + "stat events on existing thread id"), OPT_BOOLEAN('a', "all-cpus", &system_wide, "system-wide collection from all CPUs"), OPT_BOOLEAN('c', "scale", &scale, @@ -506,10 +540,11 @@ static const struct option options[] = { int cmd_stat(int argc, const char **argv, const char *prefix __used) { int status; + int i,j; argc = parse_options(argc, argv, options, stat_usage, PARSE_OPT_STOP_AT_NON_OPTION); - if (!argc && target_pid == -1) + if (!argc && target_pid == -1 && target_tid == -1) usage_with_options(stat_usage, options); if (run_count <= 0) usage_with_options(stat_usage, options); @@ -525,6 +560,31 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) else nr_cpus = 1; + if (target_pid != -1) { + target_tid = target_pid; + thread_num = find_all_tid(target_pid, &all_tids); + if (thread_num <= 0) { + fprintf(stderr, "Can't find all threads of pid %d\n", + target_pid); + usage_with_options(stat_usage, options); + } + } else { + all_tids=malloc(sizeof(pid_t)); + if (!all_tids) + return -ENOMEM; + + all_tids[0] = target_tid; + thread_num = 1; + } + + for (i = 0; i < MAX_NR_CPUS; i++) { + for (j = 0; j < MAX_COUNTERS; j++) { + fd[i][j] = malloc(sizeof(int)*thread_num); + if (!fd[i][j]) + return -ENOMEM; + } + } + /* * We dont want to block the signals - that would cause * child tasks to inherit that and Ctrl-C would not work. @@ -543,7 +603,8 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used) status = run_perf_stat(argc, argv); } - print_stat(argc, argv); + if (status != -1) + print_stat(argc, argv); return status; } diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c index 1f529321607e..4abdd9b646b3 100644 --- a/tools/perf/builtin-top.c +++ b/tools/perf/builtin-top.c @@ -55,7 +55,7 @@ #include <linux/unistd.h> #include <linux/types.h> -static int fd[MAX_NR_CPUS][MAX_COUNTERS]; +static int *fd[MAX_NR_CPUS][MAX_COUNTERS]; static int system_wide = 0; @@ -65,6 +65,9 @@ static int count_filter = 5; static int print_entries; static int target_pid = -1; +static int target_tid = -1; +static pid_t *all_tids = NULL; +static int thread_num = 0; static int inherit = 0; static int profile_cpu = -1; static int nr_cpus = 0; @@ -133,7 +136,7 @@ static inline struct symbol *sym_entry__symbol(struct sym_entry *self) return ((void *)self) + symbol_conf.priv_size; } -static void get_term_dimensions(struct winsize *ws) +void get_term_dimensions(struct winsize *ws) { char *s = getenv("LINES"); @@ -169,7 +172,7 @@ static void sig_winch_handler(int sig __used) update_print_entries(&winsize); } -static void parse_source(struct sym_entry *syme) +static int parse_source(struct sym_entry *syme) { struct symbol *sym; struct sym_entry_source *source; @@ -180,12 +183,21 @@ static void parse_source(struct sym_entry *syme) u64 len; if (!syme) - return; + return -1; + + sym = sym_entry__symbol(syme); + map = syme->map; + + /* + * We can't annotate with just /proc/kallsyms + */ + if (map->dso->origin == DSO__ORIG_KERNEL) + return -1; if (syme->src == NULL) { syme->src = zalloc(sizeof(*source)); if (syme->src == NULL) - return; + return -1; pthread_mutex_init(&syme->src->lock, NULL); } @@ -195,9 +207,6 @@ static void parse_source(struct sym_entry *syme) pthread_mutex_lock(&source->lock); goto out_assign; } - - sym = sym_entry__symbol(syme); - map = syme->map; path = map->dso->long_name; len = sym->end - sym->start; @@ -209,7 +218,7 @@ static void parse_source(struct sym_entry *syme) file = popen(command, "r"); if (!file) - return; + return -1; pthread_mutex_lock(&source->lock); source->lines_tail = &source->lines; @@ -245,6 +254,7 @@ static void parse_source(struct sym_entry *syme) out_assign: sym_filter_entry = syme; pthread_mutex_unlock(&source->lock); + return 0; } static void __zero_source_counters(struct sym_entry *syme) @@ -411,6 +421,7 @@ static double sym_weight(const struct sym_entry *sym) static long samples; static long userspace_samples; +static long exact_samples; static const char CONSOLE_CLEAR[] = "[H[2J"; static void __list_insert_active_sym(struct sym_entry *syme) @@ -451,6 +462,7 @@ static void print_sym_table(void) int counter, snap = !display_weighted ? sym_counter : 0; float samples_per_sec = samples/delay_secs; float ksamples_per_sec = (samples-userspace_samples)/delay_secs; + float esamples_percent = (100.0*exact_samples)/samples; float sum_ksamples = 0.0; struct sym_entry *syme, *n; struct rb_root tmp = RB_ROOT; @@ -458,7 +470,7 @@ static void print_sym_table(void) int sym_width = 0, dso_width = 0, dso_short_width = 0; const int win_width = winsize.ws_col - 1; - samples = userspace_samples = 0; + samples = userspace_samples = exact_samples = 0; /* Sort the active symbols */ pthread_mutex_lock(&active_symbols_lock); @@ -489,9 +501,10 @@ static void print_sym_table(void) puts(CONSOLE_CLEAR); printf("%-*.*s\n", win_width, win_width, graph_dotted_line); - printf( " PerfTop:%8.0f irqs/sec kernel:%4.1f%% [", + printf( " PerfTop:%8.0f irqs/sec kernel:%4.1f%% exact: %4.1f%% [", samples_per_sec, - 100.0 - (100.0*((samples_per_sec-ksamples_per_sec)/samples_per_sec))); + 100.0 - (100.0*((samples_per_sec-ksamples_per_sec)/samples_per_sec)), + esamples_percent); if (nr_counters == 1 || !display_weighted) { printf("%Ld", (u64)attrs[0].sample_period); @@ -514,13 +527,15 @@ static void print_sym_table(void) if (target_pid != -1) printf(" (target_pid: %d", target_pid); + else if (target_tid != -1) + printf(" (target_tid: %d", target_tid); else printf(" (all"); if (profile_cpu != -1) printf(", cpu: %d)\n", profile_cpu); else { - if (target_pid != -1) + if (target_tid != -1) printf(")\n"); else printf(", %d CPUs)\n", nr_cpus); @@ -960,6 +975,9 @@ static void event__process_sample(const event_t *self, return; } + if (self->header.misc & PERF_RECORD_MISC_EXACT) + exact_samples++; + if (event__preprocess_sample(self, session, &al, symbol_filter) < 0 || al.filtered) return; @@ -990,7 +1008,17 @@ static void event__process_sample(const event_t *self, if (sym_filter_entry_sched) { sym_filter_entry = sym_filter_entry_sched; sym_filter_entry_sched = NULL; - parse_source(sym_filter_entry); + if (parse_source(sym_filter_entry) < 0) { + struct symbol *sym = sym_entry__symbol(sym_filter_entry); + + pr_err("Can't annotate %s", sym->name); + if (sym_filter_entry->map->dso->origin == DSO__ORIG_KERNEL) { + pr_err(": No vmlinux file was found in the path:\n"); + vmlinux_path__fprintf(stderr); + } else + pr_err(".\n"); + exit(1); + } } syme = symbol__priv(al.sym); @@ -1106,16 +1134,21 @@ static void perf_session__mmap_read_counter(struct perf_session *self, md->prev = old; } -static struct pollfd event_array[MAX_NR_CPUS * MAX_COUNTERS]; -static struct mmap_data mmap_array[MAX_NR_CPUS][MAX_COUNTERS]; +static struct pollfd *event_array; +static struct mmap_data *mmap_array[MAX_NR_CPUS][MAX_COUNTERS]; static void perf_session__mmap_read(struct perf_session *self) { - int i, counter; + int i, counter, thread_index; for (i = 0; i < nr_cpus; i++) { for (counter = 0; counter < nr_counters; counter++) - perf_session__mmap_read_counter(self, &mmap_array[i][counter]); + for (thread_index = 0; + thread_index < thread_num; + thread_index++) { + perf_session__mmap_read_counter(self, + &mmap_array[i][counter][thread_index]); + } } } @@ -1126,9 +1159,10 @@ static void start_counter(int i, int counter) { struct perf_event_attr *attr; int cpu; + int thread_index; cpu = profile_cpu; - if (target_pid == -1 && profile_cpu == -1) + if (target_tid == -1 && profile_cpu == -1) cpu = cpumap[i]; attr = attrs + counter; @@ -1144,55 +1178,58 @@ static void start_counter(int i, int counter) attr->inherit = (cpu < 0) && inherit; attr->mmap = 1; + for (thread_index = 0; thread_index < thread_num; thread_index++) { try_again: - fd[i][counter] = sys_perf_event_open(attr, target_pid, cpu, group_fd, 0); - - if (fd[i][counter] < 0) { - int err = errno; + fd[i][counter][thread_index] = sys_perf_event_open(attr, + all_tids[thread_index], cpu, group_fd, 0); + + if (fd[i][counter][thread_index] < 0) { + int err = errno; + + if (err == EPERM || err == EACCES) + die("No permission - are you root?\n"); + /* + * If it's cycles then fall back to hrtimer + * based cpu-clock-tick sw counter, which + * is always available even if no PMU support: + */ + if (attr->type == PERF_TYPE_HARDWARE + && attr->config == PERF_COUNT_HW_CPU_CYCLES) { + + if (verbose) + warning(" ... trying to fall back to cpu-clock-ticks\n"); + + attr->type = PERF_TYPE_SOFTWARE; + attr->config = PERF_COUNT_SW_CPU_CLOCK; + goto try_again; + } + printf("\n"); + error("perfcounter syscall returned with %d (%s)\n", + fd[i][counter][thread_index], strerror(err)); + die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); + exit(-1); + } + assert(fd[i][counter][thread_index] >= 0); + fcntl(fd[i][counter][thread_index], F_SETFL, O_NONBLOCK); - if (err == EPERM || err == EACCES) - die("No permission - are you root?\n"); /* - * If it's cycles then fall back to hrtimer - * based cpu-clock-tick sw counter, which - * is always available even if no PMU support: + * First counter acts as the group leader: */ - if (attr->type == PERF_TYPE_HARDWARE - && attr->config == PERF_COUNT_HW_CPU_CYCLES) { - - if (verbose) - warning(" ... trying to fall back to cpu-clock-ticks\n"); - - attr->type = PERF_TYPE_SOFTWARE; - attr->config = PERF_COUNT_SW_CPU_CLOCK; - goto try_again; - } - printf("\n"); - error("perfcounter syscall returned with %d (%s)\n", - fd[i][counter], strerror(err)); - die("No CONFIG_PERF_EVENTS=y kernel support configured?\n"); - exit(-1); + if (group && group_fd == -1) + group_fd = fd[i][counter][thread_index]; + + event_array[nr_poll].fd = fd[i][counter][thread_index]; + event_array[nr_poll].events = POLLIN; + nr_poll++; + + mmap_array[i][counter][thread_index].counter = counter; + mmap_array[i][counter][thread_index].prev = 0; + mmap_array[i][counter][thread_index].mask = mmap_pages*page_size - 1; + mmap_array[i][counter][thread_index].base = mmap(NULL, (mmap_pages+1)*page_size, + PROT_READ, MAP_SHARED, fd[i][counter][thread_index], 0); + if (mmap_array[i][counter][thread_index].base == MAP_FAILED) + die("failed to mmap with %d (%s)\n", errno, strerror(errno)); } - assert(fd[i][counter] >= 0); - fcntl(fd[i][counter], F_SETFL, O_NONBLOCK); - - /* - * First counter acts as the group leader: - */ - if (group && group_fd == -1) - group_fd = fd[i][counter]; - - event_array[nr_poll].fd = fd[i][counter]; - event_array[nr_poll].events = POLLIN; - nr_poll++; - - mmap_array[i][counter].counter = counter; - mmap_array[i][counter].prev = 0; - mmap_array[i][counter].mask = mmap_pages*page_size - 1; - mmap_array[i][counter].base = mmap(NULL, (mmap_pages+1)*page_size, - PROT_READ, MAP_SHARED, fd[i][counter], 0); - if (mmap_array[i][counter].base == MAP_FAILED) - die("failed to mmap with %d (%s)\n", errno, strerror(errno)); } static int __cmd_top(void) @@ -1208,8 +1245,8 @@ static int __cmd_top(void) if (session == NULL) return -ENOMEM; - if (target_pid != -1) - event__synthesize_thread(target_pid, event__process, session); + if (target_tid != -1) + event__synthesize_thread(target_tid, event__process, session); else event__synthesize_threads(event__process, session); @@ -1220,7 +1257,7 @@ static int __cmd_top(void) } /* Wait for a minimal set of events before starting the snapshot */ - poll(event_array, nr_poll, 100); + poll(&event_array[0], nr_poll, 100); perf_session__mmap_read(session); @@ -1263,7 +1300,9 @@ static const struct option options[] = { OPT_INTEGER('c', "count", &default_interval, "event period to sample"), OPT_INTEGER('p', "pid", &target_pid, - "profile events on existing pid"), + "profile events on existing process id"), + OPT_INTEGER('t', "tid", &target_tid, + "profile events on existing thread id"), OPT_BOOLEAN('a', "all-cpus", &system_wide, "system-wide collection from all CPUs"), OPT_INTEGER('C', "CPU", &profile_cpu, @@ -1304,6 +1343,7 @@ static const struct option options[] = { int cmd_top(int argc, const char **argv, const char *prefix __used) { int counter; + int i,j; page_size = sysconf(_SC_PAGE_SIZE); @@ -1311,8 +1351,39 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) if (argc) usage_with_options(top_usage, options); + if (target_pid != -1) { + target_tid = target_pid; + thread_num = find_all_tid(target_pid, &all_tids); + if (thread_num <= 0) { + fprintf(stderr, "Can't find all threads of pid %d\n", + target_pid); + usage_with_options(top_usage, options); + } + } else { + all_tids=malloc(sizeof(pid_t)); + if (!all_tids) + return -ENOMEM; + + all_tids[0] = target_tid; + thread_num = 1; + } + + for (i = 0; i < MAX_NR_CPUS; i++) { + for (j = 0; j < MAX_COUNTERS; j++) { + fd[i][j] = malloc(sizeof(int)*thread_num); + mmap_array[i][j] = zalloc( + sizeof(struct mmap_data)*thread_num); + if (!fd[i][j] || !mmap_array[i][j]) + return -ENOMEM; + } + } + event_array = malloc( + sizeof(struct pollfd)*MAX_NR_CPUS*MAX_COUNTERS*thread_num); + if (!event_array) + return -ENOMEM; + /* CPU and PID are mutually exclusive */ - if (target_pid != -1 && profile_cpu != -1) { + if (target_tid > 0 && profile_cpu != -1) { printf("WARNING: PID switch overriding CPU\n"); sleep(1); profile_cpu = -1; @@ -1353,7 +1424,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __used) attrs[counter].sample_period = default_interval; } - if (target_pid != -1 || profile_cpu != -1) + if (target_tid != -1 || profile_cpu != -1) nr_cpus = 1; else nr_cpus = read_cpu_map(); diff --git a/tools/perf/perf-archive.sh b/tools/perf/perf-archive.sh index 910468e6e01c..2e7a4f417e20 100644 --- a/tools/perf/perf-archive.sh +++ b/tools/perf/perf-archive.sh @@ -30,4 +30,7 @@ done tar cfj $PERF_DATA.tar.bz2 -C $DEBUGDIR -T $MANIFEST rm -f $MANIFEST $BUILDIDS +echo -e "Now please run:\n" +echo -e "$ tar xvf $PERF_DATA.tar.bz2 -C ~/.debug\n" +echo "wherever you need to run 'perf report' on." exit 0 diff --git a/tools/perf/perf.c b/tools/perf/perf.c index cd32c200cdb3..2826e702986e 100644 --- a/tools/perf/perf.c +++ b/tools/perf/perf.c @@ -16,6 +16,8 @@ #include "util/string.h" #include "util/debugfs.h" +bool use_browser; + const char perf_usage_string[] = "perf [--version] [--help] COMMAND [ARGS]"; @@ -262,6 +264,8 @@ static int run_builtin(struct cmd_struct *p, int argc, const char **argv) set_debugfs_path(); status = p->fn(argc, argv, prefix); + exit_browser(status); + if (status) return status & 0xff; diff --git a/tools/perf/perf.h b/tools/perf/perf.h index 6fb379bc1d1f..ec212748d651 100644 --- a/tools/perf/perf.h +++ b/tools/perf/perf.h @@ -1,6 +1,10 @@ #ifndef _PERF_PERF_H #define _PERF_PERF_H +struct winsize; + +void get_term_dimensions(struct winsize *ws); + #if defined(__i386__) #include "../../arch/x86/include/asm/unistd.h" #define rmb() asm volatile("lock; addl $0,0(%%esp)" ::: "memory") @@ -102,8 +106,6 @@ static inline unsigned long long rdclock(void) #define __user #define asmlinkage -#define __used __attribute__((__unused__)) - #define unlikely(x) __builtin_expect(!!(x), 0) #define min(x, y) ({ \ typeof(x) _min1 = (x); \ diff --git a/tools/perf/util/cache.h b/tools/perf/util/cache.h index 918eb376abe3..4b9aab7f0405 100644 --- a/tools/perf/util/cache.h +++ b/tools/perf/util/cache.h @@ -1,6 +1,7 @@ #ifndef __PERF_CACHE_H #define __PERF_CACHE_H +#include <stdbool.h> #include "util.h" #include "strbuf.h" #include "../perf.h" @@ -69,6 +70,19 @@ extern const char *pager_program; extern int pager_in_use(void); extern int pager_use_color; +extern bool use_browser; + +#ifdef NO_NEWT_SUPPORT +static inline void setup_browser(void) +{ + setup_pager(); +} +static inline void exit_browser(bool wait_for_ok __used) {} +#else +void setup_browser(void); +void exit_browser(bool wait_for_ok); +#endif + extern const char *editor_program; extern const char *excludes_file; diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c index b3b71258272a..db628af6d20d 100644 --- a/tools/perf/util/callchain.c +++ b/tools/perf/util/callchain.c @@ -1,5 +1,5 @@ /* - * Copyright (C) 2009, Frederic Weisbecker <fweisbec@gmail.com> + * Copyright (C) 2009-2010, Frederic Weisbecker <fweisbec@gmail.com> * * Handle the callchains from the stream in an ad-hoc radix tree and then * sort them in an rbtree. @@ -183,12 +183,23 @@ create_child(struct callchain_node *parent, bool inherit_children) return new; } + +struct resolved_ip { + u64 ip; + struct map_symbol ms; +}; + +struct resolved_chain { + u64 nr; + struct resolved_ip ips[0]; +}; + + /* * Fill the node with callchain values */ static void -fill_node(struct callchain_node *node, struct ip_callchain *chain, - int start, struct symbol **syms) +fill_node(struct callchain_node *node, struct resolved_chain *chain, int start) { unsigned int i; @@ -200,8 +211,8 @@ fill_node(struct callchain_node *node, struct ip_callchain *chain, perror("not enough memory for the code path tree"); return; } - call->ip = chain->ips[i]; - call->sym = syms[i]; + call->ip = chain->ips[i].ip; + call->ms = chain->ips[i].ms; list_add_tail(&call->list, &node->val); } node->val_nr = chain->nr - start; @@ -210,13 +221,13 @@ fill_node(struct callchain_node *node, struct ip_callchain *chain, } static void -add_child(struct callchain_node *parent, struct ip_callchain *chain, - int start, struct symbol **syms) +add_child(struct callchain_node *parent, struct resolved_chain *chain, + int start) { struct callchain_node *new; new = create_child(parent, false); - fill_node(new, chain, start, syms); + fill_node(new, chain, start); new->children_hit = 0; new->hit = 1; @@ -228,9 +239,8 @@ add_child(struct callchain_node *parent, struct ip_callchain *chain, * Then create another child to host the given callchain of new branch */ static void -split_add_child(struct callchain_node *parent, struct ip_callchain *chain, - struct callchain_list *to_split, int idx_parents, int idx_local, - struct symbol **syms) +split_add_child(struct callchain_node *parent, struct resolved_chain *chain, + struct callchain_list *to_split, int idx_parents, int idx_local) { struct callchain_node *new; struct list_head *old_tail; @@ -257,7 +267,7 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, /* create a new child for the new branch if any */ if (idx_total < chain->nr) { parent->hit = 0; - add_child(parent, chain, idx_total, syms); + add_child(parent, chain, idx_total); parent->children_hit++; } else { parent->hit = 1; @@ -265,32 +275,33 @@ split_add_child(struct callchain_node *parent, struct ip_callchain *chain, } static int -__append_chain(struct callchain_node *root, struct ip_callchain *chain, - unsigned int start, struct symbol **syms); +__append_chain(struct callchain_node *root, struct resolved_chain *chain, + unsigned int start); static void -__append_chain_children(struct callchain_node *root, struct ip_callchain *chain, - struct symbol **syms, unsigned int start) +__append_chain_children(struct callchain_node *root, + struct resolved_chain *chain, + unsigned int start) { struct callchain_node *rnode; /* lookup in childrens */ chain_for_each_child(rnode, root) { - unsigned int ret = __append_chain(rnode, chain, start, syms); + unsigned int ret = __append_chain(rnode, chain, start); if (!ret) goto inc_children_hit; } /* nothing in children, add to the current node */ - add_child(root, chain, start, syms); + add_child(root, chain, start); inc_children_hit: root->children_hit++; } static int -__append_chain(struct callchain_node *root, struct ip_callchain *chain, - unsigned int start, struct symbol **syms) +__append_chain(struct callchain_node *root, struct resolved_chain *chain, + unsigned int start) { struct callchain_list *cnode; unsigned int i = start; @@ -302,13 +313,19 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, * anywhere inside a function. */ list_for_each_entry(cnode, &root->val, list) { + struct symbol *sym; + if (i == chain->nr) break; - if (cnode->sym && syms[i]) { - if (cnode->sym->start != syms[i]->start) + + sym = chain->ips[i].ms.sym; + + if (cnode->ms.sym && sym) { + if (cnode->ms.sym->start != sym->start) break; - } else if (cnode->ip != chain->ips[i]) + } else if (cnode->ip != chain->ips[i].ip) break; + if (!found) found = true; i++; @@ -320,7 +337,7 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, /* we match only a part of the node. Split it and add the new chain */ if (i - start < root->val_nr) { - split_add_child(root, chain, cnode, start, i - start, syms); + split_add_child(root, chain, cnode, start, i - start); return 0; } @@ -331,15 +348,50 @@ __append_chain(struct callchain_node *root, struct ip_callchain *chain, } /* We match the node and still have a part remaining */ - __append_chain_children(root, chain, syms, i); + __append_chain_children(root, chain, i); return 0; } -void append_chain(struct callchain_node *root, struct ip_callchain *chain, - struct symbol **syms) +static void filter_context(struct ip_callchain *old, struct resolved_chain *new, + struct map_symbol *syms) +{ + int i, j = 0; + + for (i = 0; i < (int)old->nr; i++) { + if (old->ips[i] >= PERF_CONTEXT_MAX) + continue; + + new->ips[j].ip = old->ips[i]; + new->ips[j].ms = syms[i]; + j++; + } + + new->nr = j; +} + + +int append_chain(struct callchain_node *root, struct ip_callchain *chain, + struct map_symbol *syms) { + struct resolved_chain *filtered; + if (!chain->nr) - return; - __append_chain_children(root, chain, syms, 0); + return 0; + + filtered = malloc(sizeof(*filtered) + + chain->nr * sizeof(struct resolved_ip)); + if (!filtered) + return -ENOMEM; + + filter_context(chain, filtered, syms); + + if (!filtered->nr) + goto end; + + __append_chain_children(root, filtered, 0); +end: + free(filtered); + + return 0; } diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h index ad4626de4c2b..8a7e8bbd0fda 100644 --- a/tools/perf/util/callchain.h +++ b/tools/perf/util/callchain.h @@ -39,7 +39,7 @@ struct callchain_param { struct callchain_list { u64 ip; - struct symbol *sym; + struct map_symbol ms; struct list_head list; }; @@ -56,6 +56,6 @@ static inline u64 cumul_hits(struct callchain_node *node) } int register_callchain_param(struct callchain_param *param); -void append_chain(struct callchain_node *root, struct ip_callchain *chain, - struct symbol **syms); +int append_chain(struct callchain_node *root, struct ip_callchain *chain, + struct map_symbol *syms); #endif /* __PERF_CALLCHAIN_H */ diff --git a/tools/perf/util/color.c b/tools/perf/util/color.c index e88bca55a599..9da01914e0af 100644 --- a/tools/perf/util/color.c +++ b/tools/perf/util/color.c @@ -203,7 +203,10 @@ int color_fprintf(FILE *fp, const char *color, const char *fmt, ...) int r; va_start(args, fmt); - r = color_vfprintf(fp, color, fmt, args); + if (use_browser) + r = vfprintf(fp, fmt, args); + else + r = color_vfprintf(fp, color, fmt, args); va_end(args); return r; } diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c index 0905600c3851..033d66db863a 100644 --- a/tools/perf/util/debug.c +++ b/tools/perf/util/debug.c @@ -6,6 +6,7 @@ #include <stdarg.h> #include <stdio.h> +#include "cache.h" #include "color.h" #include "event.h" #include "debug.h" @@ -21,7 +22,10 @@ int eprintf(int level, const char *fmt, ...) if (verbose >= level) { va_start(args, fmt); - ret = vfprintf(stderr, fmt, args); + if (use_browser) + ret = browser__show_help(fmt, args); + else + ret = vfprintf(stderr, fmt, args); va_end(args); } diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h index c6c24c522dea..0172edf3f153 100644 --- a/tools/perf/util/debug.h +++ b/tools/perf/util/debug.h @@ -7,9 +7,16 @@ extern int verbose; extern int dump_trace; -int eprintf(int level, - const char *fmt, ...) __attribute__((format(printf, 2, 3))); int dump_printf(const char *fmt, ...) __attribute__((format(printf, 1, 2))); void trace_event(event_t *event); +#ifdef NO_NEWT_SUPPORT +static inline int browser__show_help(const char *format __used, va_list ap __used) +{ + return 0; +} +#else +int browser__show_help(const char *format, va_list ap); +#endif + #endif /* __PERF_DEBUG_H */ diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c index 705ec63548b4..052eaeccc202 100644 --- a/tools/perf/util/event.c +++ b/tools/perf/util/event.c @@ -393,7 +393,8 @@ int event__process_mmap(event_t *self, struct perf_session *session) } thread = perf_session__findnew(session, self->mmap.pid); - map = map__new(&self->mmap, MAP__FUNCTION, + map = map__new(self->mmap.start, self->mmap.len, self->mmap.pgoff, + self->mmap.pid, self->mmap.filename, MAP__FUNCTION, session->cwd, session->cwdlen); if (thread == NULL || map == NULL) @@ -513,24 +514,32 @@ int event__preprocess_sample(const event_t *self, struct perf_session *session, dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid); - thread__find_addr_location(thread, session, cpumode, MAP__FUNCTION, - self->ip.ip, al, filter); + thread__find_addr_map(thread, session, cpumode, MAP__FUNCTION, + self->ip.ip, al); dump_printf(" ...... dso: %s\n", al->map ? al->map->dso->long_name : al->level == 'H' ? "[hypervisor]" : "<not found>"); - /* - * We have to do this here as we may have a dso with no symbol hit that - * has a name longer than the ones with symbols sampled. - */ - if (al->map && !sort_dso.elide && !al->map->dso->slen_calculated) - dso__calc_col_width(al->map->dso); - - if (symbol_conf.dso_list && - (!al->map || !al->map->dso || - !(strlist__has_entry(symbol_conf.dso_list, al->map->dso->short_name) || - (al->map->dso->short_name != al->map->dso->long_name && - strlist__has_entry(symbol_conf.dso_list, al->map->dso->long_name))))) - goto out_filtered; + al->sym = NULL; + + if (al->map) { + if (symbol_conf.dso_list && + (!al->map || !al->map->dso || + !(strlist__has_entry(symbol_conf.dso_list, + al->map->dso->short_name) || + (al->map->dso->short_name != al->map->dso->long_name && + strlist__has_entry(symbol_conf.dso_list, + al->map->dso->long_name))))) + goto out_filtered; + /* + * We have to do this here as we may have a dso with no symbol + * hit that has a name longer than the ones with symbols + * sampled. + */ + if (!sort_dso.elide && !al->map->dso->slen_calculated) + dso__calc_col_width(al->map->dso); + + al->sym = map__find_symbol(al->map, al->addr, filter); + } if (symbol_conf.sym_list && al->sym && !strlist__has_entry(symbol_conf.sym_list, al->sym->name)) diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c index 2be33c7dbf03..09e09e78cb62 100644 --- a/tools/perf/util/hist.c +++ b/tools/perf/util/hist.c @@ -22,8 +22,10 @@ struct hist_entry *__perf_session__add_hist_entry(struct rb_root *hists, struct hist_entry *he; struct hist_entry entry = { .thread = al->thread, - .map = al->map, - .sym = al->sym, + .ms = { + .map = al->map, + .sym = al->sym, + }, .ip = al->addr, .level = al->level, .count = count, @@ -258,8 +260,8 @@ static size_t ipchain__fprintf_graph(FILE *fp, struct callchain_list *chain, } else ret += fprintf(fp, "%s", " "); } - if (chain->sym) - ret += fprintf(fp, "%s\n", chain->sym->name); + if (chain->ms.sym) + ret += fprintf(fp, "%s\n", chain->ms.sym->name); else ret += fprintf(fp, "%p\n", (void *)(long)chain->ip); @@ -278,7 +280,7 @@ static void init_rem_hits(void) } strcpy(rem_sq_bracket->name, "[...]"); - rem_hits.sym = rem_sq_bracket; + rem_hits.ms.sym = rem_sq_bracket; } static size_t __callchain__fprintf_graph(FILE *fp, struct callchain_node *self, @@ -328,8 +330,6 @@ static size_t __callchain__fprintf_graph(FILE *fp, struct callchain_node *self, left_margin); i = 0; list_for_each_entry(chain, &child->val, list) { - if (chain->ip >= PERF_CONTEXT_MAX) - continue; ret += ipchain__fprintf_graph(fp, chain, depth, new_depth_mask, i++, new_total, @@ -368,9 +368,6 @@ static size_t callchain__fprintf_graph(FILE *fp, struct callchain_node *self, int ret = 0; list_for_each_entry(chain, &self->val, list) { - if (chain->ip >= PERF_CONTEXT_MAX) - continue; - if (!i++ && sort__first_dimension == SORT_SYM) continue; @@ -385,8 +382,8 @@ static size_t callchain__fprintf_graph(FILE *fp, struct callchain_node *self, } else ret += callchain__fprintf_left_margin(fp, left_margin); - if (chain->sym) - ret += fprintf(fp, " %s\n", chain->sym->name); + if (chain->ms.sym) + ret += fprintf(fp, " %s\n", chain->ms.sym->name); else ret += fprintf(fp, " %p\n", (void *)(long)chain->ip); } @@ -411,8 +408,8 @@ static size_t callchain__fprintf_flat(FILE *fp, struct callchain_node *self, list_for_each_entry(chain, &self->val, list) { if (chain->ip >= PERF_CONTEXT_MAX) continue; - if (chain->sym) - ret += fprintf(fp, " %s\n", chain->sym->name); + if (chain->ms.sym) + ret += fprintf(fp, " %s\n", chain->ms.sym->name); else ret += fprintf(fp, " %p\n", (void *)(long)chain->ip); @@ -455,11 +452,11 @@ static size_t hist_entry_callchain__fprintf(FILE *fp, struct hist_entry *self, return ret; } -static size_t hist_entry__fprintf(struct hist_entry *self, - struct perf_session *pair_session, - bool show_displacement, - long displacement, FILE *fp, - u64 session_total) +size_t hist_entry__fprintf(struct hist_entry *self, + struct perf_session *pair_session, + bool show_displacement, + long displacement, FILE *fp, + u64 session_total) { struct sort_entry *se; u64 count, total; @@ -485,9 +482,9 @@ static size_t hist_entry__fprintf(struct hist_entry *self, if (symbol_conf.show_nr_samples) { if (sep) - fprintf(fp, "%c%lld", *sep, count); + ret += fprintf(fp, "%c%lld", *sep, count); else - fprintf(fp, "%11lld", count); + ret += fprintf(fp, "%11lld", count); } if (pair_session) { @@ -518,9 +515,9 @@ static size_t hist_entry__fprintf(struct hist_entry *self, snprintf(bf, sizeof(bf), " "); if (sep) - fprintf(fp, "%c%s", *sep, bf); + ret += fprintf(fp, "%c%s", *sep, bf); else - fprintf(fp, "%6.6s", bf); + ret += fprintf(fp, "%6.6s", bf); } } @@ -528,27 +525,27 @@ static size_t hist_entry__fprintf(struct hist_entry *self, if (se->elide) continue; - fprintf(fp, "%s", sep ?: " "); + ret += fprintf(fp, "%s", sep ?: " "); ret += se->print(fp, self, se->width ? *se->width : 0); } - ret += fprintf(fp, "\n"); - - if (symbol_conf.use_callchain) { - int left_margin = 0; + return ret + fprintf(fp, "\n"); +} - if (sort__first_dimension == SORT_COMM) { - se = list_first_entry(&hist_entry__sort_list, typeof(*se), - list); - left_margin = se->width ? *se->width : 0; - left_margin -= thread__comm_len(self->thread); - } +static size_t hist_entry__fprintf_callchain(struct hist_entry *self, FILE *fp, + u64 session_total) +{ + int left_margin = 0; - hist_entry_callchain__fprintf(fp, self, session_total, - left_margin); + if (sort__first_dimension == SORT_COMM) { + struct sort_entry *se = list_first_entry(&hist_entry__sort_list, + typeof(*se), list); + left_margin = se->width ? *se->width : 0; + left_margin -= thread__comm_len(self->thread); } - return ret; + return hist_entry_callchain__fprintf(fp, self, session_total, + left_margin); } size_t perf_session__fprintf_hists(struct rb_root *hists, @@ -655,7 +652,11 @@ print_entries: } ret += hist_entry__fprintf(h, pair, show_displacement, displacement, fp, session_total); - if (h->map == NULL && verbose > 1) { + + if (symbol_conf.use_callchain) + ret += hist_entry__fprintf_callchain(h, fp, session_total); + + if (h->ms.map == NULL && verbose > 1) { __map_groups__fprintf_maps(&h->thread->mg, MAP__FUNCTION, fp); fprintf(fp, "%.10s end\n", graph_dotted_line); diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h index 16f360cce5bf..fe366ce5db45 100644 --- a/tools/perf/util/hist.h +++ b/tools/perf/util/hist.h @@ -18,6 +18,11 @@ struct hist_entry *__perf_session__add_hist_entry(struct rb_root *hists, u64 count, bool *hit); extern int64_t hist_entry__cmp(struct hist_entry *, struct hist_entry *); extern int64_t hist_entry__collapse(struct hist_entry *, struct hist_entry *); +size_t hist_entry__fprintf(struct hist_entry *self, + struct perf_session *pair_session, + bool show_displacement, + long displacement, FILE *fp, + u64 session_total); void hist_entry__free(struct hist_entry *); void perf_session__output_resort(struct rb_root *hists, u64 total_samples); diff --git a/tools/perf/util/include/linux/compiler.h b/tools/perf/util/include/linux/compiler.h index dfb0713ed47f..791f9dd27ebf 100644 --- a/tools/perf/util/include/linux/compiler.h +++ b/tools/perf/util/include/linux/compiler.h @@ -7,4 +7,6 @@ #define __user #define __attribute_const__ +#define __used __attribute__((__unused__)) + #endif diff --git a/tools/perf/util/include/linux/kernel.h b/tools/perf/util/include/linux/kernel.h index f2611655ab51..388ab1bfd114 100644 --- a/tools/perf/util/include/linux/kernel.h +++ b/tools/perf/util/include/linux/kernel.h @@ -85,16 +85,19 @@ simple_strtoul(const char *nptr, char **endptr, int base) return strtoul(nptr, endptr, base); } +int eprintf(int level, + const char *fmt, ...) __attribute__((format(printf, 2, 3))); + #ifndef pr_fmt #define pr_fmt(fmt) fmt #endif #define pr_err(fmt, ...) \ - do { fprintf(stderr, pr_fmt(fmt), ##__VA_ARGS__); } while (0) + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) #define pr_warning(fmt, ...) \ - do { fprintf(stderr, pr_fmt(fmt), ##__VA_ARGS__); } while (0) + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) #define pr_info(fmt, ...) \ - do { fprintf(stderr, pr_fmt(fmt), ##__VA_ARGS__); } while (0) + eprintf(0, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debug(fmt, ...) \ eprintf(1, pr_fmt(fmt), ##__VA_ARGS__) #define pr_debugN(n, fmt, ...) \ diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c index e509cd59c67d..9f2963f9ee9a 100644 --- a/tools/perf/util/map.c +++ b/tools/perf/util/map.c @@ -1,9 +1,9 @@ -#include "event.h" #include "symbol.h" +#include <limits.h> #include <stdlib.h> #include <string.h> #include <stdio.h> -#include "debug.h" +#include "map.h" const char *map_type__name[MAP__NR_TYPES] = { [MAP__FUNCTION] = "Functions", @@ -38,13 +38,12 @@ void map__init(struct map *self, enum map_type type, RB_CLEAR_NODE(&self->rb_node); } -struct map *map__new(struct mmap_event *event, enum map_type type, - char *cwd, int cwdlen) +struct map *map__new(u64 start, u64 len, u64 pgoff, u32 pid, char *filename, + enum map_type type, char *cwd, int cwdlen) { struct map *self = malloc(sizeof(*self)); if (self != NULL) { - const char *filename = event->filename; char newfilename[PATH_MAX]; struct dso *dso; int anon; @@ -62,7 +61,7 @@ struct map *map__new(struct mmap_event *event, enum map_type type, anon = is_anon_memory(filename); if (anon) { - snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", event->pid); + snprintf(newfilename, sizeof(newfilename), "/tmp/perf-%d.map", pid); filename = newfilename; } @@ -70,8 +69,7 @@ struct map *map__new(struct mmap_event *event, enum map_type type, if (dso == NULL) goto out_delete; - map__init(self, type, event->start, event->start + event->len, - event->pgoff, dso); + map__init(self, type, start, start + len, pgoff, dso); if (anon) { set_identity: @@ -235,3 +233,84 @@ u64 map__objdump_2ip(struct map *map, u64 addr) map->unmap_ip(map, addr); /* RIP -> IP */ return ip; } + +struct symbol *map_groups__find_symbol(struct map_groups *self, + enum map_type type, u64 addr, + symbol_filter_t filter) +{ + struct map *map = map_groups__find(self, type, addr); + + if (map != NULL) + return map__find_symbol(map, map->map_ip(map, addr), filter); + + return NULL; +} + +static u64 map__reloc_map_ip(struct map *map, u64 ip) +{ + return ip + (s64)map->pgoff; +} + +static u64 map__reloc_unmap_ip(struct map *map, u64 ip) +{ + return ip - (s64)map->pgoff; +} + +void map__reloc_vmlinux(struct map *self) +{ + struct kmap *kmap = map__kmap(self); + s64 reloc; + + if (!kmap->ref_reloc_sym || !kmap->ref_reloc_sym->unrelocated_addr) + return; + + reloc = (kmap->ref_reloc_sym->unrelocated_addr - + kmap->ref_reloc_sym->addr); + + if (!reloc) + return; + + self->map_ip = map__reloc_map_ip; + self->unmap_ip = map__reloc_unmap_ip; + self->pgoff = reloc; +} + +void maps__insert(struct rb_root *maps, struct map *map) +{ + struct rb_node **p = &maps->rb_node; + struct rb_node *parent = NULL; + const u64 ip = map->start; + struct map *m; + + while (*p != NULL) { + parent = *p; + m = rb_entry(parent, struct map, rb_node); + if (ip < m->start) + p = &(*p)->rb_left; + else + p = &(*p)->rb_right; + } + + rb_link_node(&map->rb_node, parent, p); + rb_insert_color(&map->rb_node, maps); +} + +struct map *maps__find(struct rb_root *maps, u64 ip) +{ + struct rb_node **p = &maps->rb_node; + struct rb_node *parent = NULL; + struct map *m; + + while (*p != NULL) { + parent = *p; + m = rb_entry(parent, struct map, rb_node); + if (ip < m->start) + p = &(*p)->rb_left; + else if (ip > m->end) + p = &(*p)->rb_right; + else + return m; + } + + return NULL; +} diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h index b756368076c6..6a703fa74707 100644 --- a/tools/perf/util/map.h +++ b/tools/perf/util/map.h @@ -4,7 +4,8 @@ #include <linux/compiler.h> #include <linux/list.h> #include <linux/rbtree.h> -#include <linux/types.h> +#include <stdio.h> +#include "types.h" enum map_type { MAP__FUNCTION = 0, @@ -68,14 +69,13 @@ u64 map__rip_2objdump(struct map *map, u64 rip); u64 map__objdump_2ip(struct map *map, u64 addr); struct symbol; -struct mmap_event; typedef int (*symbol_filter_t)(struct map *map, struct symbol *sym); void map__init(struct map *self, enum map_type type, u64 start, u64 end, u64 pgoff, struct dso *dso); -struct map *map__new(struct mmap_event *event, enum map_type, - char *cwd, int cwdlen); +struct map *map__new(u64 start, u64 len, u64 pgoff, u32 pid, char *filename, + enum map_type type, char *cwd, int cwdlen); void map__delete(struct map *self); struct map *map__clone(struct map *self); int map__overlap(struct map *l, struct map *r); @@ -91,4 +91,48 @@ void map__fixup_end(struct map *self); void map__reloc_vmlinux(struct map *self); +struct map_groups { + struct rb_root maps[MAP__NR_TYPES]; + struct list_head removed_maps[MAP__NR_TYPES]; +}; + +size_t __map_groups__fprintf_maps(struct map_groups *self, + enum map_type type, FILE *fp); +void maps__insert(struct rb_root *maps, struct map *map); +struct map *maps__find(struct rb_root *maps, u64 addr); +void map_groups__init(struct map_groups *self); +size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp); + +static inline void map_groups__insert(struct map_groups *self, struct map *map) +{ + maps__insert(&self->maps[map->type], map); +} + +static inline struct map *map_groups__find(struct map_groups *self, + enum map_type type, u64 addr) +{ + return maps__find(&self->maps[type], addr); +} + +struct symbol *map_groups__find_symbol(struct map_groups *self, + enum map_type type, u64 addr, + symbol_filter_t filter); + +static inline struct symbol *map_groups__find_function(struct map_groups *self, + u64 addr, + symbol_filter_t filter) +{ + return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); +} + +struct map *map_groups__find_by_name(struct map_groups *self, + enum map_type type, const char *name); +int __map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES], + struct dso *kernel); +int map_groups__create_kernel_maps(struct map_groups *self, + struct map *vmlinux_maps[MAP__NR_TYPES]); +struct map *map_groups__new_module(struct map_groups *self, u64 start, + const char *filename); + #endif /* __PERF_MAP_H */ diff --git a/tools/perf/util/newt.c b/tools/perf/util/newt.c new file mode 100644 index 000000000000..e99bcc8d1939 --- /dev/null +++ b/tools/perf/util/newt.c @@ -0,0 +1,526 @@ +#define _GNU_SOURCE +#include <stdio.h> +#undef _GNU_SOURCE + +#include <stdlib.h> +#include <newt.h> +#include <sys/ttydefaults.h> + +#include "cache.h" +#include "hist.h" +#include "session.h" +#include "sort.h" +#include "symbol.h" + +static void newt_form__set_exit_keys(newtComponent self) +{ + newtFormAddHotKey(self, NEWT_KEY_ESCAPE); + newtFormAddHotKey(self, 'Q'); + newtFormAddHotKey(self, 'q'); + newtFormAddHotKey(self, CTRL('c')); +} + +static newtComponent newt_form__new(void) +{ + newtComponent self = newtForm(NULL, NULL, 0); + if (self) + newt_form__set_exit_keys(self); + return self; +} + +static int popup_menu(int argc, const char *argv[]) +{ + struct newtExitStruct es; + int i, rc = -1, max_len = 5; + newtComponent listbox, form = newt_form__new(); + + if (form == NULL) + return -1; + + listbox = newtListbox(0, 0, argc, NEWT_FLAG_RETURNEXIT); + if (listbox == NULL) + goto out_destroy_form; + + newtFormAddComponents(form, listbox, NULL); + + for (i = 0; i < argc; ++i) { + int len = strlen(argv[i]); + if (len > max_len) + max_len = len; + if (newtListboxAddEntry(listbox, argv[i], (void *)(long)i)) + goto out_destroy_form; + } + + newtCenteredWindow(max_len, argc, NULL); + newtFormRun(form, &es); + rc = newtListboxGetCurrent(listbox) - NULL; + if (es.reason == NEWT_EXIT_HOTKEY) + rc = -1; + newtPopWindow(); +out_destroy_form: + newtFormDestroy(form); + return rc; +} + +static bool dialog_yesno(const char *msg) +{ + /* newtWinChoice should really be accepting const char pointers... */ + char yes[] = "Yes", no[] = "No"; + return newtWinChoice(NULL, no, yes, (char *)msg) == 2; +} + +/* + * When debugging newt problems it was useful to be able to "unroll" + * the calls to newtCheckBoxTreeAdd{Array,Item}, so that we can generate + * a source file with the sequence of calls to these methods, to then + * tweak the arrays to get the intended results, so I'm keeping this code + * here, may be useful again in the future. + */ +#undef NEWT_DEBUG + +static void newt_checkbox_tree__add(newtComponent tree, const char *str, + void *priv, int *indexes) +{ +#ifdef NEWT_DEBUG + /* Print the newtCheckboxTreeAddArray to tinker with its index arrays */ + int i = 0, len = 40 - strlen(str); + + fprintf(stderr, + "\tnewtCheckboxTreeAddItem(tree, %*.*s\"%s\", (void *)%p, 0, ", + len, len, " ", str, priv); + while (indexes[i] != NEWT_ARG_LAST) { + if (indexes[i] != NEWT_ARG_APPEND) + fprintf(stderr, " %d,", indexes[i]); + else + fprintf(stderr, " %s,", "NEWT_ARG_APPEND"); + ++i; + } + fprintf(stderr, " %s", " NEWT_ARG_LAST);\n"); + fflush(stderr); +#endif + newtCheckboxTreeAddArray(tree, str, priv, 0, indexes); +} + +static char *callchain_list__sym_name(struct callchain_list *self, + char *bf, size_t bfsize) +{ + if (self->ms.sym) + return self->ms.sym->name; + + snprintf(bf, bfsize, "%#Lx", self->ip); + return bf; +} + +static void __callchain__append_graph_browser(struct callchain_node *self, + newtComponent tree, u64 total, + int *indexes, int depth) +{ + struct rb_node *node; + u64 new_total, remaining; + int idx = 0; + + if (callchain_param.mode == CHAIN_GRAPH_REL) + new_total = self->children_hit; + else + new_total = total; + + remaining = new_total; + node = rb_first(&self->rb_root); + while (node) { + struct callchain_node *child = rb_entry(node, struct callchain_node, rb_node); + struct rb_node *next = rb_next(node); + u64 cumul = cumul_hits(child); + struct callchain_list *chain; + int first = true, printed = 0; + int chain_idx = -1; + remaining -= cumul; + + indexes[depth] = NEWT_ARG_APPEND; + indexes[depth + 1] = NEWT_ARG_LAST; + + list_for_each_entry(chain, &child->val, list) { + char ipstr[BITS_PER_LONG / 4 + 1], + *alloc_str = NULL; + const char *str = callchain_list__sym_name(chain, ipstr, sizeof(ipstr)); + + if (first) { + double percent = cumul * 100.0 / new_total; + + first = false; + if (asprintf(&alloc_str, "%2.2f%% %s", percent, str) < 0) + str = "Not enough memory!"; + else + str = alloc_str; + } else { + indexes[depth] = idx; + indexes[depth + 1] = NEWT_ARG_APPEND; + indexes[depth + 2] = NEWT_ARG_LAST; + ++chain_idx; + } + newt_checkbox_tree__add(tree, str, &chain->ms, indexes); + free(alloc_str); + ++printed; + } + + indexes[depth] = idx; + if (chain_idx != -1) + indexes[depth + 1] = chain_idx; + if (printed != 0) + ++idx; + __callchain__append_graph_browser(child, tree, new_total, indexes, + depth + (chain_idx != -1 ? 2 : 1)); + node = next; + } +} + +static void callchain__append_graph_browser(struct callchain_node *self, + newtComponent tree, u64 total, + int *indexes, int parent_idx) +{ + struct callchain_list *chain; + int i = 0; + + indexes[1] = NEWT_ARG_APPEND; + indexes[2] = NEWT_ARG_LAST; + + list_for_each_entry(chain, &self->val, list) { + char ipstr[BITS_PER_LONG / 4 + 1], *str; + + if (chain->ip >= PERF_CONTEXT_MAX) + continue; + + if (!i++ && sort__first_dimension == SORT_SYM) + continue; + + str = callchain_list__sym_name(chain, ipstr, sizeof(ipstr)); + newt_checkbox_tree__add(tree, str, &chain->ms, indexes); + } + + indexes[1] = parent_idx; + indexes[2] = NEWT_ARG_APPEND; + indexes[3] = NEWT_ARG_LAST; + __callchain__append_graph_browser(self, tree, total, indexes, 2); +} + +static void hist_entry__append_callchain_browser(struct hist_entry *self, + newtComponent tree, u64 total, int parent_idx) +{ + struct rb_node *rb_node; + int indexes[1024] = { [0] = parent_idx, }; + int idx = 0; + struct callchain_node *chain; + + rb_node = rb_first(&self->sorted_chain); + while (rb_node) { + chain = rb_entry(rb_node, struct callchain_node, rb_node); + switch (callchain_param.mode) { + case CHAIN_FLAT: + break; + case CHAIN_GRAPH_ABS: /* falldown */ + case CHAIN_GRAPH_REL: + callchain__append_graph_browser(chain, tree, total, indexes, idx++); + break; + case CHAIN_NONE: + default: + break; + } + rb_node = rb_next(rb_node); + } +} + +/* + * FIXME: get lib/string.c linked with perf somehow + */ +static char *skip_spaces(const char *str) +{ + while (isspace(*str)) + ++str; + return (char *)str; +} + +static char *strim(char *s) +{ + size_t size; + char *end; + + s = skip_spaces(s); + size = strlen(s); + if (!size) + return s; + + end = s + size - 1; + while (end >= s && isspace(*end)) + end--; + *(end + 1) = '\0'; + + return s; +} + +static size_t hist_entry__append_browser(struct hist_entry *self, + newtComponent tree, u64 total) +{ + char bf[1024], *s; + FILE *fp; + + if (symbol_conf.exclude_other && !self->parent) + return 0; + + fp = fmemopen(bf, sizeof(bf), "w"); + if (fp == NULL) + return 0; + + hist_entry__fprintf(self, NULL, false, 0, fp, total); + fclose(fp); + + /* + * FIXME: We shouldn't need to trim, as the printing routines shouldn't + * add spaces it in the first place, the stdio output routines should + * call a __snprintf method instead of the current __print (that + * actually is a __fprintf) one, but get the raw string and _then_ add + * the newline, as this is a detail of stdio printing, not needed in + * other UIs, e.g. newt. + */ + s = strim(bf); + + if (symbol_conf.use_callchain) { + int indexes[2]; + + indexes[0] = NEWT_ARG_APPEND; + indexes[1] = NEWT_ARG_LAST; + newt_checkbox_tree__add(tree, s, &self->ms, indexes); + } else + newtListboxAppendEntry(tree, s, &self->ms); + + return strlen(s); +} + +static void map_symbol__annotate_browser(const struct map_symbol *self) +{ + FILE *fp; + int cols, rows; + newtComponent form, tree; + struct newtExitStruct es; + char *str; + size_t line_len, max_line_len = 0; + size_t max_usable_width; + char *line = NULL; + + if (self->sym == NULL) + return; + + if (asprintf(&str, "perf annotate -d \"%s\" %s 2>&1 | expand", + self->map->dso->name, self->sym->name) < 0) + return; + + fp = popen(str, "r"); + if (fp == NULL) + goto out_free_str; + + newtPushHelpLine("Press ESC to exit"); + newtGetScreenSize(&cols, &rows); + tree = newtListbox(0, 0, rows - 5, NEWT_FLAG_SCROLL); + + while (!feof(fp)) { + if (getline(&line, &line_len, fp) < 0 || !line_len) + break; + while (line_len != 0 && isspace(line[line_len - 1])) + line[--line_len] = '\0'; + + if (line_len > max_line_len) + max_line_len = line_len; + newtListboxAppendEntry(tree, line, NULL); + } + fclose(fp); + free(line); + + max_usable_width = cols - 22; + if (max_line_len > max_usable_width) + max_line_len = max_usable_width; + + newtListboxSetWidth(tree, max_line_len); + + newtCenteredWindow(max_line_len + 2, rows - 5, self->sym->name); + form = newt_form__new(); + newtFormAddComponents(form, tree, NULL); + + newtFormRun(form, &es); + newtFormDestroy(form); + newtPopWindow(); + newtPopHelpLine(); +out_free_str: + free(str); +} + +static const void *newt__symbol_tree_get_current(newtComponent self) +{ + if (symbol_conf.use_callchain) + return newtCheckboxTreeGetCurrent(self); + return newtListboxGetCurrent(self); +} + +static void perf_session__selection(newtComponent self, void *data) +{ + const struct map_symbol **symbol_ptr = data; + *symbol_ptr = newt__symbol_tree_get_current(self); +} + +void perf_session__browse_hists(struct rb_root *hists, u64 session_total, + const char *helpline) +{ + struct sort_entry *se; + struct rb_node *nd; + char seq[] = "."; + unsigned int width; + char *col_width = symbol_conf.col_width_list_str; + int rows, cols, idx; + int max_len = 0; + char str[1024]; + newtComponent form, tree; + struct newtExitStruct es; + const struct map_symbol *selection; + + snprintf(str, sizeof(str), "Samples: %Ld", session_total); + newtDrawRootText(0, 0, str); + newtPushHelpLine(helpline); + + newtGetScreenSize(&cols, &rows); + + if (symbol_conf.use_callchain) + tree = newtCheckboxTreeMulti(0, 0, rows - 5, seq, + NEWT_FLAG_SCROLL); + else + tree = newtListbox(0, 0, rows - 5, (NEWT_FLAG_SCROLL | + NEWT_FLAG_RETURNEXIT)); + + newtComponentAddCallback(tree, perf_session__selection, &selection); + + list_for_each_entry(se, &hist_entry__sort_list, list) { + if (se->elide) + continue; + width = strlen(se->header); + if (se->width) { + if (symbol_conf.col_width_list_str) { + if (col_width) { + *se->width = atoi(col_width); + col_width = strchr(col_width, ','); + if (col_width) + ++col_width; + } + } + *se->width = max(*se->width, width); + } + } + + idx = 0; + for (nd = rb_first(hists); nd; nd = rb_next(nd)) { + struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node); + int len = hist_entry__append_browser(h, tree, session_total); + if (len > max_len) + max_len = len; + if (symbol_conf.use_callchain) + hist_entry__append_callchain_browser(h, tree, session_total, idx++); + } + + if (max_len > cols) + max_len = cols - 3; + + if (!symbol_conf.use_callchain) + newtListboxSetWidth(tree, max_len); + + newtCenteredWindow(max_len + (symbol_conf.use_callchain ? 5 : 0), + rows - 5, "Report"); + form = newt_form__new(); + newtFormAddHotKey(form, 'A'); + newtFormAddHotKey(form, 'a'); + newtFormAddHotKey(form, NEWT_KEY_RIGHT); + newtFormAddComponents(form, tree, NULL); + selection = newt__symbol_tree_get_current(tree); + + while (1) { + char annotate[512]; + const char *options[2]; + int nr_options = 0, choice = 0; + + newtFormRun(form, &es); + if (es.reason == NEWT_EXIT_HOTKEY) { + if (toupper(es.u.key) == 'A') + goto do_annotate; + if (es.u.key == NEWT_KEY_ESCAPE || + toupper(es.u.key) == 'Q' || + es.u.key == CTRL('c')) { + if (dialog_yesno("Do you really want to exit?")) + break; + else + continue; + } + } + + if (selection->sym != NULL) { + snprintf(annotate, sizeof(annotate), + "Annotate %s", selection->sym->name); + options[nr_options++] = annotate; + } + + options[nr_options++] = "Exit"; + choice = popup_menu(nr_options, options); + if (choice == nr_options - 1) + break; +do_annotate: + if (selection->sym != NULL && choice >= 0) { + if (selection->map->dso->origin == DSO__ORIG_KERNEL) { + newtPopHelpLine(); + newtPushHelpLine("No vmlinux file found, can't " + "annotate with just a " + "kallsyms file"); + continue; + } + map_symbol__annotate_browser(selection); + } + } + + newtFormDestroy(form); + newtPopWindow(); +} + +static char browser__last_msg[1024]; + +int browser__show_help(const char *format, va_list ap) +{ + int ret; + static int backlog; + + ret = vsnprintf(browser__last_msg + backlog, + sizeof(browser__last_msg) - backlog, format, ap); + backlog += ret; + + if (browser__last_msg[backlog - 1] == '\n') { + newtPopHelpLine(); + newtPushHelpLine(browser__last_msg); + newtRefresh(); + backlog = 0; + } + + return ret; +} + +void setup_browser(void) +{ + if (!isatty(1)) + return; + + use_browser = true; + newtInit(); + newtCls(); + newtPushHelpLine(" "); +} + +void exit_browser(bool wait_for_ok) +{ + if (use_browser) { + if (wait_for_ok) { + char title[] = "Fatal Error", ok[] = "Ok"; + newtWinMessage(title, ok, browser__last_msg); + } + newtFinished(); + } +} diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c index 05d0c5c2030c..435781e0c205 100644 --- a/tools/perf/util/parse-events.c +++ b/tools/perf/util/parse-events.c @@ -5,6 +5,7 @@ #include "parse-events.h" #include "exec_cmd.h" #include "string.h" +#include "symbol.h" #include "cache.h" #include "header.h" #include "debugfs.h" @@ -656,6 +657,10 @@ parse_raw_event(const char **strp, struct perf_event_attr *attr) return EVT_FAILED; n = hex2u64(str + 1, &config); if (n > 0) { + if (str[n+1] == 'p') { + attr->precise = 1; + n++; + } *strp = str + n + 1; attr->type = PERF_TYPE_RAW; attr->config = config; diff --git a/tools/perf/util/parse-options.c b/tools/perf/util/parse-options.c index efebd5b476b3..79dfa0c34b3c 100644 --- a/tools/perf/util/parse-options.c +++ b/tools/perf/util/parse-options.c @@ -500,6 +500,7 @@ int usage_with_options_internal(const char * const *usagestr, void usage_with_options(const char * const *usagestr, const struct option *opts) { + exit_browser(false); usage_with_options_internal(usagestr, opts, 0); exit(129); } diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c index 7c004b6ef24f..3fc0be741b8e 100644 --- a/tools/perf/util/probe-event.c +++ b/tools/perf/util/probe-event.c @@ -33,19 +33,26 @@ #include <limits.h> #undef _GNU_SOURCE +#include "util.h" #include "event.h" #include "string.h" #include "strlist.h" #include "debug.h" #include "cache.h" #include "color.h" -#include "parse-events.h" /* For debugfs_path */ +#include "symbol.h" +#include "thread.h" +#include "trace-event.h" /* For __unused */ +#include "parse-events.h" /* For debugfs_path */ #include "probe-event.h" +#include "probe-finder.h" #define MAX_CMDLEN 256 #define MAX_PROBE_ARGS 128 #define PERFPROBE_GROUP "probe" +bool probe_event_dry_run; /* Dry run flag */ + #define semantic_error(msg ...) die("Semantic error :" msg) /* If there is no space to write, returns -E2BIG. */ @@ -64,6 +71,208 @@ static int e_snprintf(char *str, size_t size, const char *format, ...) return ret; } +static char *synthesize_perf_probe_point(struct perf_probe_point *pp); +static struct map_groups kmap_groups; +static struct map *kmaps[MAP__NR_TYPES]; + +/* Initialize symbol maps and path of vmlinux */ +static void init_vmlinux(void) +{ + symbol_conf.sort_by_name = true; + if (symbol_conf.vmlinux_name == NULL) + symbol_conf.try_vmlinux_path = true; + else + pr_debug("Use vmlinux: %s\n", symbol_conf.vmlinux_name); + if (symbol__init() < 0) + die("Failed to init symbol map."); + + map_groups__init(&kmap_groups); + if (map_groups__create_kernel_maps(&kmap_groups, kmaps) < 0) + die("Failed to create kernel maps."); +} + +#ifdef DWARF_SUPPORT +static int open_vmlinux(void) +{ + if (map__load(kmaps[MAP__FUNCTION], NULL) < 0) { + pr_debug("Failed to load kernel map.\n"); + return -EINVAL; + } + pr_debug("Try to open %s\n", kmaps[MAP__FUNCTION]->dso->long_name); + return open(kmaps[MAP__FUNCTION]->dso->long_name, O_RDONLY); +} + +static void convert_to_perf_probe_point(struct kprobe_trace_point *tp, + struct perf_probe_point *pp) +{ + struct symbol *sym; + int fd, ret = 0; + + sym = map__find_symbol_by_name(kmaps[MAP__FUNCTION], + tp->symbol, NULL); + if (sym) { + fd = open_vmlinux(); + ret = find_perf_probe_point(fd, sym->start + tp->offset, pp); + close(fd); + } + if (ret <= 0) { + pp->function = xstrdup(tp->symbol); + pp->offset = tp->offset; + } + pp->retprobe = tp->retprobe; +} + +/* Try to find perf_probe_event with debuginfo */ +static int try_to_find_kprobe_trace_events(struct perf_probe_event *pev, + struct kprobe_trace_event **tevs) +{ + bool need_dwarf = perf_probe_event_need_dwarf(pev); + int fd, ntevs; + + fd = open_vmlinux(); + if (fd < 0) { + if (need_dwarf) + die("Could not open debuginfo file."); + + pr_debug("Could not open vmlinux. Try to use symbols.\n"); + return 0; + } + + /* Searching trace events corresponding to probe event */ + ntevs = find_kprobe_trace_events(fd, pev, tevs); + close(fd); + + if (ntevs > 0) /* Succeeded to find trace events */ + return ntevs; + + if (ntevs == 0) /* No error but failed to find probe point. */ + die("Probe point '%s' not found. - probe not added.", + synthesize_perf_probe_point(&pev->point)); + + /* Error path */ + if (need_dwarf) { + if (ntevs == -ENOENT) + pr_warning("No dwarf info found in the vmlinux - " + "please rebuild with CONFIG_DEBUG_INFO=y.\n"); + die("Could not analyze debuginfo."); + } + pr_debug("An error occurred in debuginfo analysis." + " Try to use symbols.\n"); + return 0; + +} + +#define LINEBUF_SIZE 256 +#define NR_ADDITIONAL_LINES 2 + +static void show_one_line(FILE *fp, unsigned int l, bool skip, bool show_num) +{ + char buf[LINEBUF_SIZE]; + const char *color = PERF_COLOR_BLUE; + + if (fgets(buf, LINEBUF_SIZE, fp) == NULL) + goto error; + if (!skip) { + if (show_num) + fprintf(stdout, "%7u %s", l, buf); + else + color_fprintf(stdout, color, " %s", buf); + } + + while (strlen(buf) == LINEBUF_SIZE - 1 && + buf[LINEBUF_SIZE - 2] != '\n') { + if (fgets(buf, LINEBUF_SIZE, fp) == NULL) + goto error; + if (!skip) { + if (show_num) + fprintf(stdout, "%s", buf); + else + color_fprintf(stdout, color, "%s", buf); + } + } + return; +error: + if (feof(fp)) + die("Source file is shorter than expected."); + else + die("File read error: %s", strerror(errno)); +} + +/* + * Show line-range always requires debuginfo to find source file and + * line number. + */ +void show_line_range(struct line_range *lr) +{ + unsigned int l = 1; + struct line_node *ln; + FILE *fp; + int fd, ret; + + /* Search a line range */ + init_vmlinux(); + fd = open_vmlinux(); + if (fd < 0) + die("Could not open debuginfo file."); + ret = find_line_range(fd, lr); + if (ret <= 0) + die("Source line is not found.\n"); + close(fd); + + setup_pager(); + + if (lr->function) + fprintf(stdout, "<%s:%d>\n", lr->function, + lr->start - lr->offset); + else + fprintf(stdout, "<%s:%d>\n", lr->file, lr->start); + + fp = fopen(lr->path, "r"); + if (fp == NULL) + die("Failed to open %s: %s", lr->path, strerror(errno)); + /* Skip to starting line number */ + while (l < lr->start) + show_one_line(fp, l++, true, false); + + list_for_each_entry(ln, &lr->line_list, list) { + while (ln->line > l) + show_one_line(fp, (l++) - lr->offset, false, false); + show_one_line(fp, (l++) - lr->offset, false, true); + } + + if (lr->end == INT_MAX) + lr->end = l + NR_ADDITIONAL_LINES; + while (l < lr->end && !feof(fp)) + show_one_line(fp, (l++) - lr->offset, false, false); + + fclose(fp); +} + +#else /* !DWARF_SUPPORT */ + +static void convert_to_perf_probe_point(struct kprobe_trace_point *tp, + struct perf_probe_point *pp) +{ + pp->function = xstrdup(tp->symbol); + pp->offset = tp->offset; + pp->retprobe = tp->retprobe; +} + +static int try_to_find_kprobe_trace_events(struct perf_probe_event *pev, + struct kprobe_trace_event **tevs __unused) +{ + if (perf_probe_event_need_dwarf(pev)) + die("Debuginfo-analysis is not supported"); + return 0; +} + +void show_line_range(struct line_range *lr __unused) +{ + die("Debuginfo-analysis is not supported"); +} + +#endif + void parse_line_range_desc(const char *arg, struct line_range *lr) { const char *ptr; @@ -90,9 +299,9 @@ void parse_line_range_desc(const char *arg, struct line_range *lr) if (*tmp != '\0') semantic_error("Tailing with invalid character '%d'.", *tmp); - tmp = strndup(arg, (ptr - arg)); + tmp = xstrndup(arg, (ptr - arg)); } else - tmp = strdup(arg); + tmp = xstrdup(arg); if (strchr(tmp, '.')) lr->file = tmp; @@ -113,8 +322,9 @@ static bool check_event_name(const char *name) } /* Parse probepoint definition. */ -static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) +static void parse_perf_probe_point(char *arg, struct perf_probe_event *pev) { + struct perf_probe_point *pp = &pev->point; char *ptr, *tmp; char c, nc = 0; /* @@ -135,7 +345,8 @@ static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) if (!check_event_name(arg)) semantic_error("%s is bad for event name -it must " "follow C symbol-naming rule.", arg); - pp->event = strdup(arg); + pev->event = xstrdup(arg); + pev->group = NULL; arg = tmp; } @@ -147,17 +358,16 @@ static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) /* Check arg is function or file and copy it */ if (strchr(arg, '.')) /* File */ - pp->file = strdup(arg); + pp->file = xstrdup(arg); else /* Function */ - pp->function = strdup(arg); - DIE_IF(pp->file == NULL && pp->function == NULL); + pp->function = xstrdup(arg); /* Parse other options */ while (ptr) { arg = ptr; c = nc; if (c == ';') { /* Lazy pattern must be the last part */ - pp->lazy_line = strdup(arg); + pp->lazy_line = xstrdup(arg); break; } ptr = strpbrk(arg, ";:+@%"); @@ -181,8 +391,7 @@ static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) case '@': /* File name */ if (pp->file) semantic_error("SRC@SRC is not allowed."); - pp->file = strdup(arg); - DIE_IF(pp->file == NULL); + pp->file = xstrdup(arg); break; case '%': /* Probe places */ if (strcmp(arg, "return") == 0) { @@ -220,59 +429,108 @@ static void parse_perf_probe_probepoint(char *arg, struct probe_point *pp) semantic_error("Offset/Line/Lazy pattern can't be used with " "return probe."); - pr_debug("symbol:%s file:%s line:%d offset:%d return:%d lazy:%s\n", + pr_debug("symbol:%s file:%s line:%d offset:%lu return:%d lazy:%s\n", pp->function, pp->file, pp->line, pp->offset, pp->retprobe, pp->lazy_line); } -/* Parse perf-probe event definition */ -void parse_perf_probe_event(const char *str, struct probe_point *pp, - bool *need_dwarf) +/* Parse perf-probe event argument */ +static void parse_perf_probe_arg(const char *str, struct perf_probe_arg *arg) +{ + const char *tmp; + struct perf_probe_arg_field **fieldp; + + pr_debug("parsing arg: %s into ", str); + + tmp = strpbrk(str, "-."); + if (!is_c_varname(str) || !tmp) { + /* A variable, register, symbol or special value */ + arg->name = xstrdup(str); + pr_debug("%s\n", arg->name); + return; + } + + /* Structure fields */ + arg->name = xstrndup(str, tmp - str); + pr_debug("%s, ", arg->name); + fieldp = &arg->field; + + do { + *fieldp = xzalloc(sizeof(struct perf_probe_arg_field)); + if (*tmp == '.') { + str = tmp + 1; + (*fieldp)->ref = false; + } else if (tmp[1] == '>') { + str = tmp + 2; + (*fieldp)->ref = true; + } else + semantic_error("Argument parse error: %s", str); + + tmp = strpbrk(str, "-."); + if (tmp) { + (*fieldp)->name = xstrndup(str, tmp - str); + pr_debug("%s(%d), ", (*fieldp)->name, (*fieldp)->ref); + fieldp = &(*fieldp)->next; + } + } while (tmp); + (*fieldp)->name = xstrdup(str); + pr_debug("%s(%d)\n", (*fieldp)->name, (*fieldp)->ref); +} + +/* Parse perf-probe event command */ +void parse_perf_probe_command(const char *cmd, struct perf_probe_event *pev) { char **argv; int argc, i; - *need_dwarf = false; - - argv = argv_split(str, &argc); + argv = argv_split(cmd, &argc); if (!argv) die("argv_split failed."); if (argc > MAX_PROBE_ARGS + 1) semantic_error("Too many arguments"); /* Parse probe point */ - parse_perf_probe_probepoint(argv[0], pp); - if (pp->file || pp->line || pp->lazy_line) - *need_dwarf = true; + parse_perf_probe_point(argv[0], pev); /* Copy arguments and ensure return probe has no C argument */ - pp->nr_args = argc - 1; - pp->args = zalloc(sizeof(char *) * pp->nr_args); - for (i = 0; i < pp->nr_args; i++) { - pp->args[i] = strdup(argv[i + 1]); - if (!pp->args[i]) - die("Failed to copy argument."); - if (is_c_varname(pp->args[i])) { - if (pp->retprobe) - semantic_error("You can't specify local" - " variable for kretprobe"); - *need_dwarf = true; - } + pev->nargs = argc - 1; + pev->args = xzalloc(sizeof(struct perf_probe_arg) * pev->nargs); + for (i = 0; i < pev->nargs; i++) { + parse_perf_probe_arg(argv[i + 1], &pev->args[i]); + if (is_c_varname(pev->args[i].name) && pev->point.retprobe) + semantic_error("You can't specify local variable for" + " kretprobe"); } argv_free(argv); } +/* Return true if this perf_probe_event requires debuginfo */ +bool perf_probe_event_need_dwarf(struct perf_probe_event *pev) +{ + int i; + + if (pev->point.file || pev->point.line || pev->point.lazy_line) + return true; + + for (i = 0; i < pev->nargs; i++) + if (is_c_varname(pev->args[i].name)) + return true; + + return false; +} + /* Parse kprobe_events event into struct probe_point */ -void parse_trace_kprobe_event(const char *str, struct probe_point *pp) +void parse_kprobe_trace_command(const char *cmd, struct kprobe_trace_event *tev) { + struct kprobe_trace_point *tp = &tev->point; char pr; char *p; int ret, i, argc; char **argv; - pr_debug("Parsing kprobe_events: %s\n", str); - argv = argv_split(str, &argc); + pr_debug("Parsing kprobe_events: %s\n", cmd); + argv = argv_split(cmd, &argc); if (!argv) die("argv_split failed."); if (argc < 2) @@ -280,132 +538,314 @@ void parse_trace_kprobe_event(const char *str, struct probe_point *pp) /* Scan event and group name. */ ret = sscanf(argv[0], "%c:%a[^/ \t]/%a[^ \t]", - &pr, (float *)(void *)&pp->group, - (float *)(void *)&pp->event); + &pr, (float *)(void *)&tev->group, + (float *)(void *)&tev->event); if (ret != 3) semantic_error("Failed to parse event name: %s", argv[0]); - pr_debug("Group:%s Event:%s probe:%c\n", pp->group, pp->event, pr); + pr_debug("Group:%s Event:%s probe:%c\n", tev->group, tev->event, pr); - pp->retprobe = (pr == 'r'); + tp->retprobe = (pr == 'r'); /* Scan function name and offset */ - ret = sscanf(argv[1], "%a[^+]+%d", (float *)(void *)&pp->function, - &pp->offset); + ret = sscanf(argv[1], "%a[^+]+%lu", (float *)(void *)&tp->symbol, + &tp->offset); if (ret == 1) - pp->offset = 0; - - /* kprobe_events doesn't have this information */ - pp->line = 0; - pp->file = NULL; + tp->offset = 0; - pp->nr_args = argc - 2; - pp->args = zalloc(sizeof(char *) * pp->nr_args); - for (i = 0; i < pp->nr_args; i++) { + tev->nargs = argc - 2; + tev->args = xzalloc(sizeof(struct kprobe_trace_arg) * tev->nargs); + for (i = 0; i < tev->nargs; i++) { p = strchr(argv[i + 2], '='); if (p) /* We don't need which register is assigned. */ - *p = '\0'; - pp->args[i] = strdup(argv[i + 2]); - if (!pp->args[i]) - die("Failed to copy argument."); + *p++ = '\0'; + else + p = argv[i + 2]; + tev->args[i].name = xstrdup(argv[i + 2]); + /* TODO: parse regs and offset */ + tev->args[i].value = xstrdup(p); } argv_free(argv); } -/* Synthesize only probe point (not argument) */ -int synthesize_perf_probe_point(struct probe_point *pp) +/* Compose only probe arg */ +int synthesize_perf_probe_arg(struct perf_probe_arg *pa, char *buf, size_t len) { - char *buf; - char offs[64] = "", line[64] = ""; + struct perf_probe_arg_field *field = pa->field; int ret; + char *tmp = buf; - pp->probes[0] = buf = zalloc(MAX_CMDLEN); - pp->found = 1; - if (!buf) - die("Failed to allocate memory by zalloc."); + ret = e_snprintf(tmp, len, "%s", pa->name); + if (ret <= 0) + goto error; + tmp += ret; + len -= ret; + + while (field) { + ret = e_snprintf(tmp, len, "%s%s", field->ref ? "->" : ".", + field->name); + if (ret <= 0) + goto error; + tmp += ret; + len -= ret; + field = field->next; + } + return tmp - buf; +error: + die("Failed to synthesize perf probe argument: %s", strerror(-ret)); +} + +/* Compose only probe point (not argument) */ +static char *synthesize_perf_probe_point(struct perf_probe_point *pp) +{ + char *buf, *tmp; + char offs[32] = "", line[32] = "", file[32] = ""; + int ret, len; + + buf = xzalloc(MAX_CMDLEN); if (pp->offset) { - ret = e_snprintf(offs, 64, "+%d", pp->offset); + ret = e_snprintf(offs, 32, "+%lu", pp->offset); if (ret <= 0) goto error; } if (pp->line) { - ret = e_snprintf(line, 64, ":%d", pp->line); + ret = e_snprintf(line, 32, ":%d", pp->line); + if (ret <= 0) + goto error; + } + if (pp->file) { + len = strlen(pp->file) - 32; + if (len < 0) + len = 0; + tmp = strchr(pp->file + len, '/'); + if (!tmp) + tmp = pp->file + len - 1; + ret = e_snprintf(file, 32, "@%s", tmp + 1); if (ret <= 0) goto error; } if (pp->function) - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s", pp->function, - offs, pp->retprobe ? "%return" : "", line); + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s%s%s%s", pp->function, + offs, pp->retprobe ? "%return" : "", line, + file); else - ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", pp->file, line); - if (ret <= 0) { + ret = e_snprintf(buf, MAX_CMDLEN, "%s%s", file, line); + if (ret <= 0) + goto error; + + return buf; error: - free(pp->probes[0]); - pp->probes[0] = NULL; - pp->found = 0; - } - return ret; + die("Failed to synthesize perf probe point: %s", strerror(-ret)); } -int synthesize_perf_probe_event(struct probe_point *pp) +#if 0 +char *synthesize_perf_probe_command(struct perf_probe_event *pev) { char *buf; int i, len, ret; - len = synthesize_perf_probe_point(pp); - if (len < 0) - return 0; + buf = synthesize_perf_probe_point(&pev->point); + if (!buf) + return NULL; - buf = pp->probes[0]; - for (i = 0; i < pp->nr_args; i++) { + len = strlen(buf); + for (i = 0; i < pev->nargs; i++) { ret = e_snprintf(&buf[len], MAX_CMDLEN - len, " %s", - pp->args[i]); - if (ret <= 0) - goto error; + pev->args[i].name); + if (ret <= 0) { + free(buf); + return NULL; + } len += ret; } - pp->found = 1; - return pp->found; -error: - free(pp->probes[0]); - pp->probes[0] = NULL; + return buf; +} +#endif - return ret; +static int __synthesize_kprobe_trace_arg_ref(struct kprobe_trace_arg_ref *ref, + char **buf, size_t *buflen, + int depth) +{ + int ret; + if (ref->next) { + depth = __synthesize_kprobe_trace_arg_ref(ref->next, buf, + buflen, depth + 1); + if (depth < 0) + goto out; + } + + ret = e_snprintf(*buf, *buflen, "%+ld(", ref->offset); + if (ret < 0) + depth = ret; + else { + *buf += ret; + *buflen -= ret; + } +out: + return depth; + +} + +static int synthesize_kprobe_trace_arg(struct kprobe_trace_arg *arg, + char *buf, size_t buflen) +{ + int ret, depth = 0; + char *tmp = buf; + + /* Argument name or separator */ + if (arg->name) + ret = e_snprintf(buf, buflen, " %s=", arg->name); + else + ret = e_snprintf(buf, buflen, " "); + if (ret < 0) + return ret; + buf += ret; + buflen -= ret; + + /* Dereferencing arguments */ + if (arg->ref) { + depth = __synthesize_kprobe_trace_arg_ref(arg->ref, &buf, + &buflen, 1); + if (depth < 0) + return depth; + } + + /* Print argument value */ + ret = e_snprintf(buf, buflen, "%s", arg->value); + if (ret < 0) + return ret; + buf += ret; + buflen -= ret; + + /* Closing */ + while (depth--) { + ret = e_snprintf(buf, buflen, ")"); + if (ret < 0) + return ret; + buf += ret; + buflen -= ret; + } + + return buf - tmp; } -int synthesize_trace_kprobe_event(struct probe_point *pp) +char *synthesize_kprobe_trace_command(struct kprobe_trace_event *tev) { + struct kprobe_trace_point *tp = &tev->point; char *buf; int i, len, ret; - pp->probes[0] = buf = zalloc(MAX_CMDLEN); - if (!buf) - die("Failed to allocate memory by zalloc."); - ret = e_snprintf(buf, MAX_CMDLEN, "%s+%d", pp->function, pp->offset); - if (ret <= 0) + buf = xzalloc(MAX_CMDLEN); + len = e_snprintf(buf, MAX_CMDLEN, "%c:%s/%s %s+%lu", + tp->retprobe ? 'r' : 'p', + tev->group, tev->event, + tp->symbol, tp->offset); + if (len <= 0) goto error; - len = ret; - for (i = 0; i < pp->nr_args; i++) { - ret = e_snprintf(&buf[len], MAX_CMDLEN - len, " %s", - pp->args[i]); + for (i = 0; i < tev->nargs; i++) { + ret = synthesize_kprobe_trace_arg(&tev->args[i], buf + len, + MAX_CMDLEN - len); if (ret <= 0) goto error; len += ret; } - pp->found = 1; - return pp->found; + return buf; error: - free(pp->probes[0]); - pp->probes[0] = NULL; + free(buf); + return NULL; +} - return ret; +void convert_to_perf_probe_event(struct kprobe_trace_event *tev, + struct perf_probe_event *pev) +{ + char buf[64]; + int i; + + /* Convert event/group name */ + pev->event = xstrdup(tev->event); + pev->group = xstrdup(tev->group); + + /* Convert trace_point to probe_point */ + convert_to_perf_probe_point(&tev->point, &pev->point); + + /* Convert trace_arg to probe_arg */ + pev->nargs = tev->nargs; + pev->args = xzalloc(sizeof(struct perf_probe_arg) * pev->nargs); + for (i = 0; i < tev->nargs; i++) + if (tev->args[i].name) + pev->args[i].name = xstrdup(tev->args[i].name); + else { + synthesize_kprobe_trace_arg(&tev->args[i], buf, 64); + pev->args[i].name = xstrdup(buf); + } +} + +void clear_perf_probe_event(struct perf_probe_event *pev) +{ + struct perf_probe_point *pp = &pev->point; + struct perf_probe_arg_field *field, *next; + int i; + + if (pev->event) + free(pev->event); + if (pev->group) + free(pev->group); + if (pp->file) + free(pp->file); + if (pp->function) + free(pp->function); + if (pp->lazy_line) + free(pp->lazy_line); + for (i = 0; i < pev->nargs; i++) { + if (pev->args[i].name) + free(pev->args[i].name); + field = pev->args[i].field; + while (field) { + next = field->next; + if (field->name) + free(field->name); + free(field); + field = next; + } + } + if (pev->args) + free(pev->args); + memset(pev, 0, sizeof(*pev)); } -static int open_kprobe_events(int flags, int mode) +void clear_kprobe_trace_event(struct kprobe_trace_event *tev) +{ + struct kprobe_trace_arg_ref *ref, *next; + int i; + + if (tev->event) + free(tev->event); + if (tev->group) + free(tev->group); + if (tev->point.symbol) + free(tev->point.symbol); + for (i = 0; i < tev->nargs; i++) { + if (tev->args[i].name) + free(tev->args[i].name); + if (tev->args[i].value) + free(tev->args[i].value); + ref = tev->args[i].ref; + while (ref) { + next = ref->next; + free(ref); + ref = next; + } + } + if (tev->args) + free(tev->args); + memset(tev, 0, sizeof(*tev)); +} + +static int open_kprobe_events(bool readwrite) { char buf[PATH_MAX]; int ret; @@ -414,7 +854,11 @@ static int open_kprobe_events(int flags, int mode) if (ret < 0) die("Failed to make kprobe_events path."); - ret = open(buf, flags, mode); + if (readwrite && !probe_event_dry_run) + ret = open(buf, O_RDWR, O_APPEND); + else + ret = open(buf, O_RDONLY, 0); + if (ret < 0) { if (errno == ENOENT) die("kprobe_events file does not exist -" @@ -427,7 +871,7 @@ static int open_kprobe_events(int flags, int mode) } /* Get raw string list of current kprobe_events */ -static struct strlist *get_trace_kprobe_event_rawlist(int fd) +static struct strlist *get_kprobe_trace_command_rawlist(int fd) { int ret, idx; FILE *fp; @@ -455,99 +899,85 @@ static struct strlist *get_trace_kprobe_event_rawlist(int fd) return sl; } -/* Free and zero clear probe_point */ -static void clear_probe_point(struct probe_point *pp) -{ - int i; - - if (pp->event) - free(pp->event); - if (pp->group) - free(pp->group); - if (pp->function) - free(pp->function); - if (pp->file) - free(pp->file); - if (pp->lazy_line) - free(pp->lazy_line); - for (i = 0; i < pp->nr_args; i++) - free(pp->args[i]); - if (pp->args) - free(pp->args); - for (i = 0; i < pp->found; i++) - free(pp->probes[i]); - memset(pp, 0, sizeof(*pp)); -} - /* Show an event */ -static void show_perf_probe_event(const char *event, const char *place, - struct probe_point *pp) +static void show_perf_probe_event(struct perf_probe_event *pev) { int i, ret; char buf[128]; + char *place; - ret = e_snprintf(buf, 128, "%s:%s", pp->group, event); + /* Synthesize only event probe point */ + place = synthesize_perf_probe_point(&pev->point); + + ret = e_snprintf(buf, 128, "%s:%s", pev->group, pev->event); if (ret < 0) die("Failed to copy event: %s", strerror(-ret)); - printf(" %-40s (on %s", buf, place); + printf(" %-20s (on %s", buf, place); - if (pp->nr_args > 0) { + if (pev->nargs > 0) { printf(" with"); - for (i = 0; i < pp->nr_args; i++) - printf(" %s", pp->args[i]); + for (i = 0; i < pev->nargs; i++) { + synthesize_perf_probe_arg(&pev->args[i], buf, 128); + printf(" %s", buf); + } } printf(")\n"); + free(place); } /* List up current perf-probe events */ void show_perf_probe_events(void) { int fd; - struct probe_point pp; + struct kprobe_trace_event tev; + struct perf_probe_event pev; struct strlist *rawlist; struct str_node *ent; setup_pager(); - memset(&pp, 0, sizeof(pp)); + init_vmlinux(); + + memset(&tev, 0, sizeof(tev)); + memset(&pev, 0, sizeof(pev)); - fd = open_kprobe_events(O_RDONLY, 0); - rawlist = get_trace_kprobe_event_rawlist(fd); + fd = open_kprobe_events(false); + rawlist = get_kprobe_trace_command_rawlist(fd); close(fd); strlist__for_each(ent, rawlist) { - parse_trace_kprobe_event(ent->s, &pp); - /* Synthesize only event probe point */ - synthesize_perf_probe_point(&pp); + parse_kprobe_trace_command(ent->s, &tev); + convert_to_perf_probe_event(&tev, &pev); /* Show an event */ - show_perf_probe_event(pp.event, pp.probes[0], &pp); - clear_probe_point(&pp); + show_perf_probe_event(&pev); + clear_perf_probe_event(&pev); + clear_kprobe_trace_event(&tev); } strlist__delete(rawlist); } /* Get current perf-probe event names */ -static struct strlist *get_perf_event_names(int fd, bool include_group) +static struct strlist *get_kprobe_trace_event_names(int fd, bool include_group) { char buf[128]; struct strlist *sl, *rawlist; struct str_node *ent; - struct probe_point pp; + struct kprobe_trace_event tev; - memset(&pp, 0, sizeof(pp)); - rawlist = get_trace_kprobe_event_rawlist(fd); + memset(&tev, 0, sizeof(tev)); + rawlist = get_kprobe_trace_command_rawlist(fd); sl = strlist__new(true, NULL); strlist__for_each(ent, rawlist) { - parse_trace_kprobe_event(ent->s, &pp); + parse_kprobe_trace_command(ent->s, &tev); if (include_group) { - if (e_snprintf(buf, 128, "%s:%s", pp.group, - pp.event) < 0) + if (e_snprintf(buf, 128, "%s:%s", tev.group, + tev.event) < 0) die("Failed to copy group:event name."); strlist__add(sl, buf); } else - strlist__add(sl, pp.event); - clear_probe_point(&pp); + strlist__add(sl, tev.event); + clear_kprobe_trace_event(&tev); } strlist__delete(rawlist); @@ -555,14 +985,18 @@ static struct strlist *get_perf_event_names(int fd, bool include_group) return sl; } -static void write_trace_kprobe_event(int fd, const char *buf) +static void write_kprobe_trace_event(int fd, struct kprobe_trace_event *tev) { int ret; + char *buf = synthesize_kprobe_trace_command(tev); pr_debug("Writing event: %s\n", buf); - ret = write(fd, buf, strlen(buf)); - if (ret <= 0) - die("Failed to write event: %s", strerror(errno)); + if (!probe_event_dry_run) { + ret = write(fd, buf, strlen(buf)); + if (ret <= 0) + die("Failed to write event: %s", strerror(errno)); + } + free(buf); } static void get_new_event_name(char *buf, size_t len, const char *base, @@ -595,65 +1029,145 @@ static void get_new_event_name(char *buf, size_t len, const char *base, die("Too many events are on the same function."); } -void add_trace_kprobe_events(struct probe_point *probes, int nr_probes, - bool force_add) +static void __add_kprobe_trace_events(struct perf_probe_event *pev, + struct kprobe_trace_event *tevs, + int ntevs, bool allow_suffix) { - int i, j, fd; - struct probe_point *pp; - char buf[MAX_CMDLEN]; - char event[64]; + int i, fd; + struct kprobe_trace_event *tev = NULL; + char buf[64]; + const char *event, *group; struct strlist *namelist; - bool allow_suffix; - fd = open_kprobe_events(O_RDWR, O_APPEND); + fd = open_kprobe_events(true); /* Get current event names */ - namelist = get_perf_event_names(fd, false); - - for (j = 0; j < nr_probes; j++) { - pp = probes + j; - if (!pp->event) - pp->event = strdup(pp->function); - if (!pp->group) - pp->group = strdup(PERFPROBE_GROUP); - DIE_IF(!pp->event || !pp->group); - /* If force_add is true, suffix search is allowed */ - allow_suffix = force_add; - for (i = 0; i < pp->found; i++) { - /* Get an unused new event name */ - get_new_event_name(event, 64, pp->event, namelist, - allow_suffix); - snprintf(buf, MAX_CMDLEN, "%c:%s/%s %s\n", - pp->retprobe ? 'r' : 'p', - pp->group, event, - pp->probes[i]); - write_trace_kprobe_event(fd, buf); - printf("Added new event:\n"); - /* Get the first parameter (probe-point) */ - sscanf(pp->probes[i], "%s", buf); - show_perf_probe_event(event, buf, pp); - /* Add added event name to namelist */ - strlist__add(namelist, event); - /* - * Probes after the first probe which comes from same - * user input are always allowed to add suffix, because - * there might be several addresses corresponding to - * one code line. - */ - allow_suffix = true; - } + namelist = get_kprobe_trace_event_names(fd, false); + + printf("Add new event%s\n", (ntevs > 1) ? "s:" : ":"); + for (i = 0; i < ntevs; i++) { + tev = &tevs[i]; + if (pev->event) + event = pev->event; + else + if (pev->point.function) + event = pev->point.function; + else + event = tev->point.symbol; + if (pev->group) + group = pev->group; + else + group = PERFPROBE_GROUP; + + /* Get an unused new event name */ + get_new_event_name(buf, 64, event, namelist, allow_suffix); + event = buf; + + tev->event = xstrdup(event); + tev->group = xstrdup(group); + write_kprobe_trace_event(fd, tev); + /* Add added event name to namelist */ + strlist__add(namelist, event); + + /* Trick here - save current event/group */ + event = pev->event; + group = pev->group; + pev->event = tev->event; + pev->group = tev->group; + show_perf_probe_event(pev); + /* Trick here - restore current event/group */ + pev->event = (char *)event; + pev->group = (char *)group; + + /* + * Probes after the first probe which comes from same + * user input are always allowed to add suffix, because + * there might be several addresses corresponding to + * one code line. + */ + allow_suffix = true; } /* Show how to use the event. */ printf("\nYou can now use it on all perf tools, such as:\n\n"); - printf("\tperf record -e %s:%s -a sleep 1\n\n", PERFPROBE_GROUP, event); + printf("\tperf record -e %s:%s -a sleep 1\n\n", tev->group, tev->event); strlist__delete(namelist); close(fd); } +static int convert_to_kprobe_trace_events(struct perf_probe_event *pev, + struct kprobe_trace_event **tevs) +{ + struct symbol *sym; + int ntevs = 0, i; + struct kprobe_trace_event *tev; + + /* Convert perf_probe_event with debuginfo */ + ntevs = try_to_find_kprobe_trace_events(pev, tevs); + if (ntevs > 0) + return ntevs; + + /* Allocate trace event buffer */ + ntevs = 1; + tev = *tevs = xzalloc(sizeof(struct kprobe_trace_event)); + + /* Copy parameters */ + tev->point.symbol = xstrdup(pev->point.function); + tev->point.offset = pev->point.offset; + tev->nargs = pev->nargs; + if (tev->nargs) { + tev->args = xzalloc(sizeof(struct kprobe_trace_arg) + * tev->nargs); + for (i = 0; i < tev->nargs; i++) + tev->args[i].value = xstrdup(pev->args[i].name); + } + + /* Currently just checking function name from symbol map */ + sym = map__find_symbol_by_name(kmaps[MAP__FUNCTION], + tev->point.symbol, NULL); + if (!sym) + die("Kernel symbol \'%s\' not found - probe not added.", + tev->point.symbol); + + return ntevs; +} + +struct __event_package { + struct perf_probe_event *pev; + struct kprobe_trace_event *tevs; + int ntevs; +}; + +void add_perf_probe_events(struct perf_probe_event *pevs, int npevs, + bool force_add) +{ + int i; + struct __event_package *pkgs; + + pkgs = xzalloc(sizeof(struct __event_package) * npevs); + + /* Init vmlinux path */ + init_vmlinux(); + + /* Loop 1: convert all events */ + for (i = 0; i < npevs; i++) { + pkgs[i].pev = &pevs[i]; + /* Convert with or without debuginfo */ + pkgs[i].ntevs = convert_to_kprobe_trace_events(pkgs[i].pev, + &pkgs[i].tevs); + } + + /* Loop 2: add all events */ + for (i = 0; i < npevs; i++) + __add_kprobe_trace_events(pkgs[i].pev, pkgs[i].tevs, + pkgs[i].ntevs, force_add); + /* TODO: cleanup all trace events? */ +} + static void __del_trace_kprobe_event(int fd, struct str_node *ent) { char *p; char buf[128]; + int ret; /* Convert from perf-probe event to trace-kprobe event */ if (e_snprintf(buf, 128, "-:%s", ent->s) < 0) @@ -663,7 +1177,10 @@ static void __del_trace_kprobe_event(int fd, struct str_node *ent) die("Internal error: %s should have ':' but not.", ent->s); *p = '/'; - write_trace_kprobe_event(fd, buf); + pr_debug("Writing event: %s\n", buf); + ret = write(fd, buf, strlen(buf)); + if (ret <= 0) + die("Failed to write event: %s", strerror(errno)); printf("Remove event: %s\n", ent->s); } @@ -696,7 +1213,7 @@ static void del_trace_kprobe_event(int fd, const char *group, pr_info("Info: event \"%s\" does not exist, could not remove it.\n", buf); } -void del_trace_kprobe_events(struct strlist *dellist) +void del_perf_probe_events(struct strlist *dellist) { int fd; const char *group, *event; @@ -704,14 +1221,12 @@ void del_trace_kprobe_events(struct strlist *dellist) struct str_node *ent; struct strlist *namelist; - fd = open_kprobe_events(O_RDWR, O_APPEND); + fd = open_kprobe_events(true); /* Get current event names */ - namelist = get_perf_event_names(fd, true); + namelist = get_kprobe_trace_event_names(fd, true); strlist__for_each(ent, dellist) { - str = strdup(ent->s); - if (!str) - die("Failed to copy event."); + str = xstrdup(ent->s); pr_debug("Parsing: %s\n", str); p = strchr(str, ':'); if (p) { @@ -730,73 +1245,3 @@ void del_trace_kprobe_events(struct strlist *dellist) close(fd); } -#define LINEBUF_SIZE 256 -#define NR_ADDITIONAL_LINES 2 - -static void show_one_line(FILE *fp, unsigned int l, bool skip, bool show_num) -{ - char buf[LINEBUF_SIZE]; - const char *color = PERF_COLOR_BLUE; - - if (fgets(buf, LINEBUF_SIZE, fp) == NULL) - goto error; - if (!skip) { - if (show_num) - fprintf(stdout, "%7u %s", l, buf); - else - color_fprintf(stdout, color, " %s", buf); - } - - while (strlen(buf) == LINEBUF_SIZE - 1 && - buf[LINEBUF_SIZE - 2] != '\n') { - if (fgets(buf, LINEBUF_SIZE, fp) == NULL) - goto error; - if (!skip) { - if (show_num) - fprintf(stdout, "%s", buf); - else - color_fprintf(stdout, color, "%s", buf); - } - } - return; -error: - if (feof(fp)) - die("Source file is shorter than expected."); - else - die("File read error: %s", strerror(errno)); -} - -void show_line_range(struct line_range *lr) -{ - unsigned int l = 1; - struct line_node *ln; - FILE *fp; - - setup_pager(); - - if (lr->function) - fprintf(stdout, "<%s:%d>\n", lr->function, - lr->start - lr->offset); - else - fprintf(stdout, "<%s:%d>\n", lr->file, lr->start); - - fp = fopen(lr->path, "r"); - if (fp == NULL) - die("Failed to open %s: %s", lr->path, strerror(errno)); - /* Skip to starting line number */ - while (l < lr->start) - show_one_line(fp, l++, true, false); - - list_for_each_entry(ln, &lr->line_list, list) { - while (ln->line > l) - show_one_line(fp, (l++) - lr->offset, false, false); - show_one_line(fp, (l++) - lr->offset, false, true); - } - - if (lr->end == INT_MAX) - lr->end = l + NR_ADDITIONAL_LINES; - while (l < lr->end && !feof(fp)) - show_one_line(fp, (l++) - lr->offset, false, false); - - fclose(fp); -} diff --git a/tools/perf/util/probe-event.h b/tools/perf/util/probe-event.h index 711287d4baea..cd308b0a4d96 100644 --- a/tools/perf/util/probe-event.h +++ b/tools/perf/util/probe-event.h @@ -2,22 +2,123 @@ #define _PROBE_EVENT_H #include <stdbool.h> -#include "probe-finder.h" #include "strlist.h" -extern void parse_line_range_desc(const char *arg, struct line_range *lr); -extern void parse_perf_probe_event(const char *str, struct probe_point *pp, - bool *need_dwarf); -extern int synthesize_perf_probe_point(struct probe_point *pp); -extern int synthesize_perf_probe_event(struct probe_point *pp); -extern void parse_trace_kprobe_event(const char *str, struct probe_point *pp); -extern int synthesize_trace_kprobe_event(struct probe_point *pp); -extern void add_trace_kprobe_events(struct probe_point *probes, int nr_probes, - bool force_add); -extern void del_trace_kprobe_events(struct strlist *dellist); +extern bool probe_event_dry_run; + +/* kprobe-tracer tracing point */ +struct kprobe_trace_point { + char *symbol; /* Base symbol */ + unsigned long offset; /* Offset from symbol */ + bool retprobe; /* Return probe flag */ +}; + +/* kprobe-tracer tracing argument referencing offset */ +struct kprobe_trace_arg_ref { + struct kprobe_trace_arg_ref *next; /* Next reference */ + long offset; /* Offset value */ +}; + +/* kprobe-tracer tracing argument */ +struct kprobe_trace_arg { + char *name; /* Argument name */ + char *value; /* Base value */ + struct kprobe_trace_arg_ref *ref; /* Referencing offset */ +}; + +/* kprobe-tracer tracing event (point + arg) */ +struct kprobe_trace_event { + char *event; /* Event name */ + char *group; /* Group name */ + struct kprobe_trace_point point; /* Trace point */ + int nargs; /* Number of args */ + struct kprobe_trace_arg *args; /* Arguments */ +}; + +/* Perf probe probing point */ +struct perf_probe_point { + char *file; /* File path */ + char *function; /* Function name */ + int line; /* Line number */ + char *lazy_line; /* Lazy matching pattern */ + unsigned long offset; /* Offset from function entry */ + bool retprobe; /* Return probe flag */ +}; + +/* Perf probe probing argument field chain */ +struct perf_probe_arg_field { + struct perf_probe_arg_field *next; /* Next field */ + char *name; /* Name of the field */ + bool ref; /* Referencing flag */ +}; + +/* Perf probe probing argument */ +struct perf_probe_arg { + char *name; /* Argument name */ + struct perf_probe_arg_field *field; /* Structure fields */ +}; + +/* Perf probe probing event (point + arg) */ +struct perf_probe_event { + char *event; /* Event name */ + char *group; /* Group name */ + struct perf_probe_point point; /* Probe point */ + int nargs; /* Number of arguments */ + struct perf_probe_arg *args; /* Arguments */ +}; + + +/* Line number container */ +struct line_node { + struct list_head list; + unsigned int line; +}; + +/* Line range */ +struct line_range { + char *file; /* File name */ + char *function; /* Function name */ + unsigned int start; /* Start line number */ + unsigned int end; /* End line number */ + int offset; /* Start line offset */ + char *path; /* Real path name */ + struct list_head line_list; /* Visible lines */ +}; + +/* Command string to events */ +extern void parse_perf_probe_command(const char *cmd, + struct perf_probe_event *pev); +extern void parse_kprobe_trace_command(const char *cmd, + struct kprobe_trace_event *tev); + +/* Events to command string */ +extern char *synthesize_perf_probe_command(struct perf_probe_event *pev); +extern char *synthesize_kprobe_trace_command(struct kprobe_trace_event *tev); +extern int synthesize_perf_probe_arg(struct perf_probe_arg *pa, char *buf, + size_t len); + +/* Check the perf_probe_event needs debuginfo */ +extern bool perf_probe_event_need_dwarf(struct perf_probe_event *pev); + +/* Convert from kprobe_trace_event to perf_probe_event */ +extern void convert_to_perf_probe_event(struct kprobe_trace_event *tev, + struct perf_probe_event *pev); + +/* Release event contents */ +extern void clear_perf_probe_event(struct perf_probe_event *pev); +extern void clear_kprobe_trace_event(struct kprobe_trace_event *tev); + +/* Command string to line-range */ +extern void parse_line_range_desc(const char *cmd, struct line_range *lr); + + +extern void add_perf_probe_events(struct perf_probe_event *pevs, int ntevs, + bool force_add); +extern void del_perf_probe_events(struct strlist *dellist); extern void show_perf_probe_events(void); extern void show_line_range(struct line_range *lr); + /* Maximum index number of event-name postfix */ #define MAX_EVENT_INDEX 1024 diff --git a/tools/perf/util/probe-finder.c b/tools/perf/util/probe-finder.c index c171a243d05b..a8513772df08 100644 --- a/tools/perf/util/probe-finder.c +++ b/tools/perf/util/probe-finder.c @@ -125,8 +125,7 @@ static void line_list__add_line(struct list_head *head, unsigned int line) p = head; found: pr_debug("line list: add a line %u\n", line); - ln = zalloc(sizeof(struct line_node)); - DIE_IF(ln == NULL); + ln = xzalloc(sizeof(struct line_node)); ln->line = line; INIT_LIST_HEAD(&ln->list); list_add(&ln->list, p); @@ -184,9 +183,89 @@ static const char *cu_find_realpath(Dwarf_Die *cu_die, const char *fname) if (strtailcmp(src, fname) == 0) break; } + if (i == nfiles) + return NULL; return src; } +/* Compare diename and tname */ +static bool die_compare_name(Dwarf_Die *dw_die, const char *tname) +{ + const char *name; + name = dwarf_diename(dw_die); + DIE_IF(name == NULL); + return strcmp(tname, name); +} + +/* Get entry pc(or low pc, 1st entry of ranges) of the die */ +static Dwarf_Addr die_get_entrypc(Dwarf_Die *dw_die) +{ + Dwarf_Addr epc; + int ret; + + ret = dwarf_entrypc(dw_die, &epc); + DIE_IF(ret == -1); + return epc; +} + +/* Get type die, but skip qualifiers and typedef */ +static Dwarf_Die *die_get_real_type(Dwarf_Die *vr_die, Dwarf_Die *die_mem) +{ + Dwarf_Attribute attr; + int tag; + + do { + if (dwarf_attr(vr_die, DW_AT_type, &attr) == NULL || + dwarf_formref_die(&attr, die_mem) == NULL) + return NULL; + + tag = dwarf_tag(die_mem); + vr_die = die_mem; + } while (tag == DW_TAG_const_type || + tag == DW_TAG_restrict_type || + tag == DW_TAG_volatile_type || + tag == DW_TAG_shared_type || + tag == DW_TAG_typedef); + + return die_mem; +} + +/* Return values for die_find callbacks */ +enum { + DIE_FIND_CB_FOUND = 0, /* End of Search */ + DIE_FIND_CB_CHILD = 1, /* Search only children */ + DIE_FIND_CB_SIBLING = 2, /* Search only siblings */ + DIE_FIND_CB_CONTINUE = 3, /* Search children and siblings */ +}; + +/* Search a child die */ +static Dwarf_Die *die_find_child(Dwarf_Die *rt_die, + int (*callback)(Dwarf_Die *, void *), + void *data, Dwarf_Die *die_mem) +{ + Dwarf_Die child_die; + int ret; + + ret = dwarf_child(rt_die, die_mem); + if (ret != 0) + return NULL; + + do { + ret = callback(die_mem, data); + if (ret == DIE_FIND_CB_FOUND) + return die_mem; + + if ((ret & DIE_FIND_CB_CHILD) && + die_find_child(die_mem, callback, data, &child_die)) { + memcpy(die_mem, &child_die, sizeof(Dwarf_Die)); + return die_mem; + } + } while ((ret & DIE_FIND_CB_SIBLING) && + dwarf_siblingof(die_mem, die_mem) == 0); + + return NULL; +} + struct __addr_die_search_param { Dwarf_Addr addr; Dwarf_Die *die_mem; @@ -205,8 +284,8 @@ static int __die_search_func_cb(Dwarf_Die *fn_die, void *data) } /* Search a real subprogram including this line, */ -static Dwarf_Die *die_get_real_subprogram(Dwarf_Die *cu_die, Dwarf_Addr addr, - Dwarf_Die *die_mem) +static Dwarf_Die *die_find_real_subprogram(Dwarf_Die *cu_die, Dwarf_Addr addr, + Dwarf_Die *die_mem) { struct __addr_die_search_param ad; ad.addr = addr; @@ -218,77 +297,64 @@ static Dwarf_Die *die_get_real_subprogram(Dwarf_Die *cu_die, Dwarf_Addr addr, return die_mem; } -/* Similar to dwarf_getfuncs, but returns inlined_subroutine if exists. */ -static Dwarf_Die *die_get_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, - Dwarf_Die *die_mem) +/* die_find callback for inline function search */ +static int __die_find_inline_cb(Dwarf_Die *die_mem, void *data) { - Dwarf_Die child_die; - int ret; - - ret = dwarf_child(sp_die, die_mem); - if (ret != 0) - return NULL; + Dwarf_Addr *addr = data; - do { - if (dwarf_tag(die_mem) == DW_TAG_inlined_subroutine && - dwarf_haspc(die_mem, addr)) - return die_mem; + if (dwarf_tag(die_mem) == DW_TAG_inlined_subroutine && + dwarf_haspc(die_mem, *addr)) + return DIE_FIND_CB_FOUND; - if (die_get_inlinefunc(die_mem, addr, &child_die)) { - memcpy(die_mem, &child_die, sizeof(Dwarf_Die)); - return die_mem; - } - } while (dwarf_siblingof(die_mem, die_mem) == 0); - - return NULL; + return DIE_FIND_CB_CONTINUE; } -/* Compare diename and tname */ -static bool die_compare_name(Dwarf_Die *dw_die, const char *tname) +/* Similar to dwarf_getfuncs, but returns inlined_subroutine if exists. */ +static Dwarf_Die *die_find_inlinefunc(Dwarf_Die *sp_die, Dwarf_Addr addr, + Dwarf_Die *die_mem) { - const char *name; - name = dwarf_diename(dw_die); - DIE_IF(name == NULL); - return strcmp(tname, name); + return die_find_child(sp_die, __die_find_inline_cb, &addr, die_mem); } -/* Get entry pc(or low pc, 1st entry of ranges) of the die */ -static Dwarf_Addr die_get_entrypc(Dwarf_Die *dw_die) +static int __die_find_variable_cb(Dwarf_Die *die_mem, void *data) { - Dwarf_Addr epc; - int ret; + const char *name = data; + int tag; - ret = dwarf_entrypc(dw_die, &epc); - DIE_IF(ret == -1); - return epc; + tag = dwarf_tag(die_mem); + if ((tag == DW_TAG_formal_parameter || + tag == DW_TAG_variable) && + (die_compare_name(die_mem, name) == 0)) + return DIE_FIND_CB_FOUND; + + return DIE_FIND_CB_CONTINUE; } -/* Get a variable die */ +/* Find a variable called 'name' */ static Dwarf_Die *die_find_variable(Dwarf_Die *sp_die, const char *name, Dwarf_Die *die_mem) { - Dwarf_Die child_die; - int tag; - int ret; + return die_find_child(sp_die, __die_find_variable_cb, (void *)name, + die_mem); +} - ret = dwarf_child(sp_die, die_mem); - if (ret != 0) - return NULL; +static int __die_find_member_cb(Dwarf_Die *die_mem, void *data) +{ + const char *name = data; - do { - tag = dwarf_tag(die_mem); - if ((tag == DW_TAG_formal_parameter || - tag == DW_TAG_variable) && - (die_compare_name(die_mem, name) == 0)) - return die_mem; + if ((dwarf_tag(die_mem) == DW_TAG_member) && + (die_compare_name(die_mem, name) == 0)) + return DIE_FIND_CB_FOUND; - if (die_find_variable(die_mem, name, &child_die)) { - memcpy(die_mem, &child_die, sizeof(Dwarf_Die)); - return die_mem; - } - } while (dwarf_siblingof(die_mem, die_mem) == 0); + return DIE_FIND_CB_SIBLING; +} - return NULL; +/* Find a member called 'name' */ +static Dwarf_Die *die_find_member(Dwarf_Die *st_die, const char *name, + Dwarf_Die *die_mem) +{ + return die_find_child(st_die, __die_find_member_cb, (void *)name, + die_mem); } /* @@ -296,19 +362,20 @@ static Dwarf_Die *die_find_variable(Dwarf_Die *sp_die, const char *name, */ /* Show a location */ -static void show_location(Dwarf_Op *op, struct probe_finder *pf) +static void convert_location(Dwarf_Op *op, struct probe_finder *pf) { unsigned int regn; Dwarf_Word offs = 0; - int deref = 0, ret; + bool ref = false; const char *regs; + struct kprobe_trace_arg *tvar = pf->tvar; /* TODO: support CFA */ /* If this is based on frame buffer, set the offset */ if (op->atom == DW_OP_fbreg) { if (pf->fb_ops == NULL) die("The attribute of frame base is not supported.\n"); - deref = 1; + ref = true; offs = op->number; op = &pf->fb_ops[0]; } @@ -316,13 +383,13 @@ static void show_location(Dwarf_Op *op, struct probe_finder *pf) if (op->atom >= DW_OP_breg0 && op->atom <= DW_OP_breg31) { regn = op->atom - DW_OP_breg0; offs += op->number; - deref = 1; + ref = true; } else if (op->atom >= DW_OP_reg0 && op->atom <= DW_OP_reg31) { regn = op->atom - DW_OP_reg0; } else if (op->atom == DW_OP_bregx) { regn = op->number; offs += op->number2; - deref = 1; + ref = true; } else if (op->atom == DW_OP_regx) { regn = op->number; } else @@ -332,17 +399,75 @@ static void show_location(Dwarf_Op *op, struct probe_finder *pf) if (!regs) die("%u exceeds max register number.", regn); - if (deref) - ret = snprintf(pf->buf, pf->len, " %s=%+jd(%s)", - pf->var, (intmax_t)offs, regs); - else - ret = snprintf(pf->buf, pf->len, " %s=%s", pf->var, regs); - DIE_IF(ret < 0); - DIE_IF(ret >= pf->len); + tvar->value = xstrdup(regs); + if (ref) { + tvar->ref = xzalloc(sizeof(struct kprobe_trace_arg_ref)); + tvar->ref->offset = (long)offs; + } +} + +static void convert_variable_fields(Dwarf_Die *vr_die, const char *varname, + struct perf_probe_arg_field *field, + struct kprobe_trace_arg_ref **ref_ptr) +{ + struct kprobe_trace_arg_ref *ref = *ref_ptr; + Dwarf_Attribute attr; + Dwarf_Die member; + Dwarf_Die type; + Dwarf_Word offs; + + pr_debug("converting %s in %s\n", field->name, varname); + if (die_get_real_type(vr_die, &type) == NULL) + die("Failed to get a type information of %s.", varname); + + /* Check the pointer and dereference */ + if (dwarf_tag(&type) == DW_TAG_pointer_type) { + if (!field->ref) + die("Semantic error: %s must be referred by '->'", + field->name); + /* Get the type pointed by this pointer */ + if (die_get_real_type(&type, &type) == NULL) + die("Failed to get a type information of %s.", varname); + + /* Verify it is a data structure */ + if (dwarf_tag(&type) != DW_TAG_structure_type) + die("%s is not a data structure.", varname); + + ref = xzalloc(sizeof(struct kprobe_trace_arg_ref)); + if (*ref_ptr) + (*ref_ptr)->next = ref; + else + *ref_ptr = ref; + } else { + /* Verify it is a data structure */ + if (dwarf_tag(&type) != DW_TAG_structure_type) + die("%s is not a data structure.", varname); + + if (field->ref) + die("Semantic error: %s must be referred by '.'", + field->name); + if (!ref) + die("Structure on a register is not supported yet."); + } + + if (die_find_member(&type, field->name, &member) == NULL) + die("%s(tyep:%s) has no member %s.", varname, + dwarf_diename(&type), field->name); + + /* Get the offset of the field */ + if (dwarf_attr(&member, DW_AT_data_member_location, &attr) == NULL || + dwarf_formudata(&attr, &offs) != 0) + die("Failed to get the offset of %s.", field->name); + ref->offset += (long)offs; + + /* Converting next field */ + if (field->next) + convert_variable_fields(&member, field->name, field->next, + &ref); } /* Show a variables in kprobe event format */ -static void show_variable(Dwarf_Die *vr_die, struct probe_finder *pf) +static void convert_variable(Dwarf_Die *vr_die, struct probe_finder *pf) { Dwarf_Attribute attr; Dwarf_Op *expr; @@ -356,83 +481,77 @@ static void show_variable(Dwarf_Die *vr_die, struct probe_finder *pf) if (ret <= 0 || nexpr == 0) goto error; - show_location(expr, pf); + convert_location(expr, pf); + + if (pf->pvar->field) + convert_variable_fields(vr_die, pf->pvar->name, + pf->pvar->field, &pf->tvar->ref); /* *expr will be cached in libdw. Don't free it. */ return ; error: /* TODO: Support const_value */ die("Failed to find the location of %s at this address.\n" - " Perhaps, it has been optimized out.", pf->var); + " Perhaps, it has been optimized out.", pf->pvar->name); } /* Find a variable in a subprogram die */ static void find_variable(Dwarf_Die *sp_die, struct probe_finder *pf) { - int ret; Dwarf_Die vr_die; + char buf[128]; /* TODO: Support struct members and arrays */ - if (!is_c_varname(pf->var)) { - /* Output raw parameters */ - ret = snprintf(pf->buf, pf->len, " %s", pf->var); - DIE_IF(ret < 0); - DIE_IF(ret >= pf->len); - return ; + if (!is_c_varname(pf->pvar->name)) { + /* Copy raw parameters */ + pf->tvar->value = xstrdup(pf->pvar->name); + } else { + synthesize_perf_probe_arg(pf->pvar, buf, 128); + pf->tvar->name = xstrdup(buf); + pr_debug("Searching '%s' variable in context.\n", + pf->pvar->name); + /* Search child die for local variables and parameters. */ + if (!die_find_variable(sp_die, pf->pvar->name, &vr_die)) + die("Failed to find '%s' in this function.", + pf->pvar->name); + convert_variable(&vr_die, pf); } - - pr_debug("Searching '%s' variable in context.\n", pf->var); - /* Search child die for local variables and parameters. */ - if (!die_find_variable(sp_die, pf->var, &vr_die)) - die("Failed to find '%s' in this function.", pf->var); - - show_variable(&vr_die, pf); } /* Show a probe point to output buffer */ -static void show_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf) +static void convert_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf) { - struct probe_point *pp = pf->pp; + struct kprobe_trace_event *tev; Dwarf_Addr eaddr; Dwarf_Die die_mem; const char *name; - char tmp[MAX_PROBE_BUFFER]; - int ret, i, len; + int ret, i; Dwarf_Attribute fb_attr; size_t nops; + if (pf->ntevs == MAX_PROBES) + die("Too many( > %d) probe point found.\n", MAX_PROBES); + tev = &pf->tevs[pf->ntevs++]; + /* If no real subprogram, find a real one */ if (!sp_die || dwarf_tag(sp_die) != DW_TAG_subprogram) { - sp_die = die_get_real_subprogram(&pf->cu_die, + sp_die = die_find_real_subprogram(&pf->cu_die, pf->addr, &die_mem); if (!sp_die) die("Probe point is not found in subprograms."); } - /* Output name of probe point */ + /* Copy the name of probe point */ name = dwarf_diename(sp_die); if (name) { dwarf_entrypc(sp_die, &eaddr); - ret = snprintf(tmp, MAX_PROBE_BUFFER, "%s+%lu", name, - (unsigned long)(pf->addr - eaddr)); - /* Copy the function name if possible */ - if (!pp->function) { - pp->function = strdup(name); - pp->offset = (size_t)(pf->addr - eaddr); - } - } else { + tev->point.symbol = xstrdup(name); + tev->point.offset = (unsigned long)(pf->addr - eaddr); + } else /* This function has no name. */ - ret = snprintf(tmp, MAX_PROBE_BUFFER, "0x%jx", - (uintmax_t)pf->addr); - if (!pp->function) { - /* TODO: Use _stext */ - pp->function = strdup(""); - pp->offset = (size_t)pf->addr; - } - } - DIE_IF(ret < 0); - DIE_IF(ret >= MAX_PROBE_BUFFER); - len = ret; - pr_debug("Probe point found: %s\n", tmp); + tev->point.offset = (unsigned long)pf->addr; + + pr_debug("Probe point found: %s+%lu\n", tev->point.symbol, + tev->point.offset); /* Get the frame base attribute/ops */ dwarf_attr(sp_die, DW_AT_frame_base, &fb_attr); @@ -442,22 +561,16 @@ static void show_probe_point(Dwarf_Die *sp_die, struct probe_finder *pf) /* Find each argument */ /* TODO: use dwarf_cfi_addrframe */ - for (i = 0; i < pp->nr_args; i++) { - pf->var = pp->args[i]; - pf->buf = &tmp[len]; - pf->len = MAX_PROBE_BUFFER - len; + tev->nargs = pf->pev->nargs; + tev->args = xzalloc(sizeof(struct kprobe_trace_arg) * tev->nargs); + for (i = 0; i < pf->pev->nargs; i++) { + pf->pvar = &pf->pev->args[i]; + pf->tvar = &tev->args[i]; find_variable(sp_die, pf); - len += strlen(pf->buf); } /* *pf->fb_ops will be cached in libdw. Don't free it. */ pf->fb_ops = NULL; - - if (pp->found == MAX_PROBES) - die("Too many( > %d) probe point found.\n", MAX_PROBES); - - pp->probes[pp->found] = strdup(tmp); - pp->found++; } /* Find probe point from its line number */ @@ -489,7 +602,7 @@ static void find_probe_point_by_line(struct probe_finder *pf) (int)i, lineno, (uintmax_t)addr); pf->addr = addr; - show_probe_point(NULL, pf); + convert_probe_point(NULL, pf); /* Continuing, because target line might be inlined. */ } } @@ -506,8 +619,7 @@ static int find_lazy_match_lines(struct list_head *head, if (fd < 0) die("failed to open %s", fname); DIE_IF(fstat(fd, &st) < 0); - fbuf = malloc(st.st_size + 2); - DIE_IF(fbuf == NULL); + fbuf = xmalloc(st.st_size + 2); DIE_IF(read(fd, fbuf, st.st_size) < 0); close(fd); fbuf[st.st_size] = '\n'; /* Dummy line */ @@ -541,7 +653,7 @@ static void find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) if (list_empty(&pf->lcache)) { /* Matching lazy line pattern */ ret = find_lazy_match_lines(&pf->lcache, pf->fname, - pf->pp->lazy_line); + pf->pev->point.lazy_line); if (ret <= 0) die("No matched lines found in %s.", pf->fname); } @@ -566,7 +678,7 @@ static void find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) if (!dwarf_haspc(sp_die, addr)) continue; /* Address filtering 2: No child include addr? */ - if (die_get_inlinefunc(sp_die, addr, &die_mem)) + if (die_find_inlinefunc(sp_die, addr, &die_mem)) continue; } @@ -574,7 +686,7 @@ static void find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) (int)i, lineno, (unsigned long long)addr); pf->addr = addr; - show_probe_point(sp_die, pf); + convert_probe_point(sp_die, pf); /* Continuing, because target line might be inlined. */ } /* TODO: deallocate lines, but how? */ @@ -583,7 +695,7 @@ static void find_probe_point_lazy(Dwarf_Die *sp_die, struct probe_finder *pf) static int probe_point_inline_cb(Dwarf_Die *in_die, void *data) { struct probe_finder *pf = (struct probe_finder *)data; - struct probe_point *pp = pf->pp; + struct perf_probe_point *pp = &pf->pev->point; if (pp->lazy_line) find_probe_point_lazy(in_die, pf); @@ -594,7 +706,7 @@ static int probe_point_inline_cb(Dwarf_Die *in_die, void *data) pr_debug("found inline addr: 0x%jx\n", (uintmax_t)pf->addr); - show_probe_point(in_die, pf); + convert_probe_point(in_die, pf); } return DWARF_CB_OK; @@ -604,7 +716,7 @@ static int probe_point_inline_cb(Dwarf_Die *in_die, void *data) static int probe_point_search_cb(Dwarf_Die *sp_die, void *data) { struct probe_finder *pf = (struct probe_finder *)data; - struct probe_point *pp = pf->pp; + struct perf_probe_point *pp = &pf->pev->point; /* Check tag and diename */ if (dwarf_tag(sp_die) != DW_TAG_subprogram || @@ -624,7 +736,7 @@ static int probe_point_search_cb(Dwarf_Die *sp_die, void *data) pf->addr = die_get_entrypc(sp_die); pf->addr += pp->offset; /* TODO: Check the address in this function */ - show_probe_point(sp_die, pf); + convert_probe_point(sp_die, pf); } } else /* Inlined function: search instances */ @@ -638,20 +750,25 @@ static void find_probe_point_by_func(struct probe_finder *pf) dwarf_getfuncs(&pf->cu_die, probe_point_search_cb, pf, 0); } -/* Find a probe point */ -int find_probe_point(int fd, struct probe_point *pp) +/* Find kprobe_trace_events specified by perf_probe_event from debuginfo */ +int find_kprobe_trace_events(int fd, struct perf_probe_event *pev, + struct kprobe_trace_event **tevs) { - struct probe_finder pf = {.pp = pp}; + struct probe_finder pf = {.pev = pev}; + struct perf_probe_point *pp = &pev->point; Dwarf_Off off, noff; size_t cuhl; Dwarf_Die *diep; Dwarf *dbg; + pf.tevs = xzalloc(sizeof(struct kprobe_trace_event) * MAX_PROBES); + *tevs = pf.tevs; + pf.ntevs = 0; + dbg = dwarf_begin(fd, DWARF_C_READ); if (!dbg) return -ENOENT; - pp->found = 0; off = 0; line_list__init(&pf.lcache); /* Loop on CUs (Compilation Unit) */ @@ -682,9 +799,81 @@ int find_probe_point(int fd, struct probe_point *pp) line_list__free(&pf.lcache); dwarf_end(dbg); - return pp->found; + return pf.ntevs; +} + +/* Reverse search */ +int find_perf_probe_point(int fd, unsigned long addr, + struct perf_probe_point *ppt) +{ + Dwarf_Die cudie, spdie, indie; + Dwarf *dbg; + Dwarf_Line *line; + Dwarf_Addr laddr, eaddr; + const char *tmp; + int lineno, ret = 0; + + dbg = dwarf_begin(fd, DWARF_C_READ); + if (!dbg) + return -ENOENT; + + /* Find cu die */ + if (!dwarf_addrdie(dbg, (Dwarf_Addr)addr, &cudie)) { + ret = -EINVAL; + goto end; + } + + /* Find a corresponding line */ + line = dwarf_getsrc_die(&cudie, (Dwarf_Addr)addr); + if (line) { + dwarf_lineaddr(line, &laddr); + if ((Dwarf_Addr)addr == laddr) { + dwarf_lineno(line, &lineno); + ppt->line = lineno; + + tmp = dwarf_linesrc(line, NULL, NULL); + DIE_IF(!tmp); + ppt->file = xstrdup(tmp); + ret = 1; + } + } + + /* Find a corresponding function */ + if (die_find_real_subprogram(&cudie, (Dwarf_Addr)addr, &spdie)) { + tmp = dwarf_diename(&spdie); + if (!tmp) + goto end; + + dwarf_entrypc(&spdie, &eaddr); + if (!lineno) { + /* We don't have a line number, let's use offset */ + ppt->function = xstrdup(tmp); + ppt->offset = addr - (unsigned long)eaddr; + ret = 1; + goto end; + } + if (die_find_inlinefunc(&spdie, (Dwarf_Addr)addr, &indie)) { + /* addr in an inline function */ + tmp = dwarf_diename(&indie); + if (!tmp) + goto end; + dwarf_decl_line(&indie, &lineno); + } else { + if (eaddr == addr) /* No offset: function entry */ + lineno = ppt->line; + else + dwarf_decl_line(&spdie, &lineno); + } + ppt->function = xstrdup(tmp); + ppt->line -= lineno; /* Make a relative line number */ + } + +end: + dwarf_end(dbg); + return ret; } + /* Find line range from its line number */ static void find_line_range_by_line(Dwarf_Die *sp_die, struct line_finder *lf) { @@ -716,7 +905,7 @@ static void find_line_range_by_line(Dwarf_Die *sp_die, struct line_finder *lf) continue; /* Address filtering 2: No child include addr? */ - if (die_get_inlinefunc(sp_die, addr, &die_mem)) + if (die_find_inlinefunc(sp_die, addr, &die_mem)) continue; } @@ -727,7 +916,7 @@ static void find_line_range_by_line(Dwarf_Die *sp_die, struct line_finder *lf) /* Copy real path */ if (!lf->lr->path) - lf->lr->path = strdup(src); + lf->lr->path = xstrdup(src); line_list__add_line(&lf->lr->line_list, (unsigned int)lineno); } /* Update status */ diff --git a/tools/perf/util/probe-finder.h b/tools/perf/util/probe-finder.h index 21f7354397b4..3564f22954f1 100644 --- a/tools/perf/util/probe-finder.h +++ b/tools/perf/util/probe-finder.h @@ -3,6 +3,7 @@ #include <stdbool.h> #include "util.h" +#include "probe-event.h" #define MAX_PATH_LEN 256 #define MAX_PROBE_BUFFER 1024 @@ -14,67 +15,36 @@ static inline int is_c_varname(const char *name) return isalpha(name[0]) || name[0] == '_'; } -struct probe_point { - char *event; /* Event name */ - char *group; /* Event group */ +#ifdef DWARF_SUPPORT +/* Find kprobe_trace_events specified by perf_probe_event from debuginfo */ +extern int find_kprobe_trace_events(int fd, struct perf_probe_event *pev, + struct kprobe_trace_event **tevs); - /* Inputs */ - char *file; /* File name */ - int line; /* Line number */ - char *lazy_line; /* Lazy line pattern */ +/* Find a perf_probe_point from debuginfo */ +extern int find_perf_probe_point(int fd, unsigned long addr, + struct perf_probe_point *ppt); - char *function; /* Function name */ - int offset; /* Offset bytes */ - - int nr_args; /* Number of arguments */ - char **args; /* Arguments */ - - int retprobe; /* Return probe */ - - /* Output */ - int found; /* Number of found probe points */ - char *probes[MAX_PROBES]; /* Output buffers (will be allocated)*/ -}; - -/* Line number container */ -struct line_node { - struct list_head list; - unsigned int line; -}; - -/* Line range */ -struct line_range { - char *file; /* File name */ - char *function; /* Function name */ - unsigned int start; /* Start line number */ - unsigned int end; /* End line number */ - int offset; /* Start line offset */ - char *path; /* Real path name */ - struct list_head line_list; /* Visible lines */ -}; - -#ifndef NO_DWARF_SUPPORT -extern int find_probe_point(int fd, struct probe_point *pp); extern int find_line_range(int fd, struct line_range *lr); #include <dwarf.h> #include <libdw.h> struct probe_finder { - struct probe_point *pp; /* Target probe point */ + struct perf_probe_event *pev; /* Target probe event */ + int ntevs; /* number of trace events */ + struct kprobe_trace_event *tevs; /* Result trace events */ /* For function searching */ Dwarf_Addr addr; /* Address */ - const char *fname; /* File name */ + const char *fname; /* Real file name */ int lno; /* Line number */ Dwarf_Die cu_die; /* Current CU */ + struct list_head lcache; /* Line cache for lazy match */ /* For variable searching */ Dwarf_Op *fb_ops; /* Frame base attribute */ - const char *var; /* Current variable name */ - char *buf; /* Current output buffer */ - int len; /* Length of output buffer */ - struct list_head lcache; /* Line cache for lazy match */ + struct perf_probe_arg *pvar; /* Current target variable */ + struct kprobe_trace_arg *tvar; /* Current result variable */ }; struct line_finder { @@ -87,6 +57,6 @@ struct line_finder { int found; }; -#endif /* NO_DWARF_SUPPORT */ +#endif /* DWARF_SUPPORT */ #endif /*_PROBE_FINDER_H */ diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c index eed1cb889008..76b4ac689df9 100644 --- a/tools/perf/util/session.c +++ b/tools/perf/util/session.c @@ -117,13 +117,13 @@ static bool symbol__match_parent_regex(struct symbol *sym) return 0; } -struct symbol **perf_session__resolve_callchain(struct perf_session *self, - struct thread *thread, - struct ip_callchain *chain, - struct symbol **parent) +struct map_symbol *perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent) { u8 cpumode = PERF_RECORD_MISC_USER; - struct symbol **syms = NULL; + struct map_symbol *syms = NULL; unsigned int i; if (symbol_conf.use_callchain) { @@ -160,7 +160,8 @@ struct symbol **perf_session__resolve_callchain(struct perf_session *self, *parent = al.sym; if (!symbol_conf.use_callchain) break; - syms[i] = al.sym; + syms[i].map = al.map; + syms[i].sym = al.sym; } } @@ -543,32 +544,3 @@ int perf_session__set_kallsyms_ref_reloc_sym(struct perf_session *self, return 0; } - -static u64 map__reloc_map_ip(struct map *map, u64 ip) -{ - return ip + (s64)map->pgoff; -} - -static u64 map__reloc_unmap_ip(struct map *map, u64 ip) -{ - return ip - (s64)map->pgoff; -} - -void map__reloc_vmlinux(struct map *self) -{ - struct kmap *kmap = map__kmap(self); - s64 reloc; - - if (!kmap->ref_reloc_sym || !kmap->ref_reloc_sym->unrelocated_addr) - return; - - reloc = (kmap->ref_reloc_sym->unrelocated_addr - - kmap->ref_reloc_sym->addr); - - if (!reloc) - return; - - self->map_ip = map__reloc_map_ip; - self->unmap_ip = map__reloc_unmap_ip; - self->pgoff = reloc; -} diff --git a/tools/perf/util/session.h b/tools/perf/util/session.h index 5c33417eebb3..631f8157fc17 100644 --- a/tools/perf/util/session.h +++ b/tools/perf/util/session.h @@ -57,10 +57,10 @@ int __perf_session__process_events(struct perf_session *self, int perf_session__process_events(struct perf_session *self, struct perf_event_ops *event_ops); -struct symbol **perf_session__resolve_callchain(struct perf_session *self, - struct thread *thread, - struct ip_callchain *chain, - struct symbol **parent); +struct map_symbol *perf_session__resolve_callchain(struct perf_session *self, + struct thread *thread, + struct ip_callchain *chain, + struct symbol **parent); bool perf_session__has_traces(struct perf_session *self, const char *msg); @@ -86,4 +86,13 @@ static inline struct map * { return map_groups__new_module(&self->kmaps, start, filename); } + +#ifdef NO_NEWT_SUPPORT +static inline void perf_session__browse_hists(struct rb_root *hists __used, + u64 session_total __used, + const char *helpline __used) {} +#else +void perf_session__browse_hists(struct rb_root *hists, u64 session_total, + const char *helpline); +#endif #endif /* __PERF_SESSION_H */ diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c index cb0f327de9e8..9b80c13cae46 100644 --- a/tools/perf/util/sort.c +++ b/tools/perf/util/sort.c @@ -131,8 +131,8 @@ sort__comm_print(FILE *fp, struct hist_entry *self, unsigned int width) int64_t sort__dso_cmp(struct hist_entry *left, struct hist_entry *right) { - struct dso *dso_l = left->map ? left->map->dso : NULL; - struct dso *dso_r = right->map ? right->map->dso : NULL; + struct dso *dso_l = left->ms.map ? left->ms.map->dso : NULL; + struct dso *dso_r = right->ms.map ? right->ms.map->dso : NULL; const char *dso_name_l, *dso_name_r; if (!dso_l || !dso_r) @@ -152,9 +152,9 @@ sort__dso_cmp(struct hist_entry *left, struct hist_entry *right) size_t sort__dso_print(FILE *fp, struct hist_entry *self, unsigned int width) { - if (self->map && self->map->dso) { - const char *dso_name = !verbose ? self->map->dso->short_name : - self->map->dso->long_name; + if (self->ms.map && self->ms.map->dso) { + const char *dso_name = !verbose ? self->ms.map->dso->short_name : + self->ms.map->dso->long_name; return repsep_fprintf(fp, "%-*s", width, dso_name); } @@ -168,11 +168,11 @@ sort__sym_cmp(struct hist_entry *left, struct hist_entry *right) { u64 ip_l, ip_r; - if (left->sym == right->sym) + if (left->ms.sym == right->ms.sym) return 0; - ip_l = left->sym ? left->sym->start : left->ip; - ip_r = right->sym ? right->sym->start : right->ip; + ip_l = left->ms.sym ? left->ms.sym->start : left->ip; + ip_r = right->ms.sym ? right->ms.sym->start : right->ip; return (int64_t)(ip_r - ip_l); } @@ -184,13 +184,13 @@ sort__sym_print(FILE *fp, struct hist_entry *self, unsigned int width __used) size_t ret = 0; if (verbose) { - char o = self->map ? dso__symtab_origin(self->map->dso) : '!'; + char o = self->ms.map ? dso__symtab_origin(self->ms.map->dso) : '!'; ret += repsep_fprintf(fp, "%#018llx %c ", (u64)self->ip, o); } ret += repsep_fprintf(fp, "[%c] ", self->level); - if (self->sym) - ret += repsep_fprintf(fp, "%s", self->sym->name); + if (self->ms.sym) + ret += repsep_fprintf(fp, "%s", self->ms.sym->name); else ret += repsep_fprintf(fp, "%#016llx", (u64)self->ip); diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h index 753f9ea99fb0..598568696f97 100644 --- a/tools/perf/util/sort.h +++ b/tools/perf/util/sort.h @@ -45,8 +45,7 @@ struct hist_entry { struct rb_node rb_node; u64 count; struct thread *thread; - struct map *map; - struct symbol *sym; + struct map_symbol ms; u64 ip; char level; struct symbol *parent; diff --git a/tools/perf/util/string.c b/tools/perf/util/string.c index a175949ed216..d4389242cfd7 100644 --- a/tools/perf/util/string.c +++ b/tools/perf/util/string.c @@ -1,49 +1,6 @@ #include "string.h" #include "util.h" -static int hex(char ch) -{ - if ((ch >= '0') && (ch <= '9')) - return ch - '0'; - if ((ch >= 'a') && (ch <= 'f')) - return ch - 'a' + 10; - if ((ch >= 'A') && (ch <= 'F')) - return ch - 'A' + 10; - return -1; -} - -/* - * While we find nice hex chars, build a long_val. - * Return number of chars processed. - */ -int hex2u64(const char *ptr, u64 *long_val) -{ - const char *p = ptr; - *long_val = 0; - - while (*p) { - const int hex_val = hex(*p); - - if (hex_val < 0) - break; - - *long_val = (*long_val << 4) | hex_val; - p++; - } - - return p - ptr; -} - -char *strxfrchar(char *s, char from, char to) -{ - char *p = s; - - while ((p = strchr(p, from)) != NULL) - *p++ = to; - - return s; -} - #define K 1024LL /* * perf_atoll() diff --git a/tools/perf/util/string.h b/tools/perf/util/string.h index 542e44de3719..700582416664 100644 --- a/tools/perf/util/string.h +++ b/tools/perf/util/string.h @@ -4,8 +4,6 @@ #include <stdbool.h> #include "types.h" -int hex2u64(const char *ptr, u64 *val); -char *strxfrchar(char *s, char from, char to); s64 perf_atoll(const char *str); char **argv_split(const char *str, int *argcp); void argv_free(char **argv); diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c index c458c4a371d1..f3d4151e46a1 100644 --- a/tools/perf/util/symbol.c +++ b/tools/perf/util/symbol.c @@ -1,13 +1,19 @@ -#include "util.h" -#include "../perf.h" -#include "sort.h" -#include "string.h" +#define _GNU_SOURCE +#include <ctype.h> +#include <dirent.h> +#include <errno.h> +#include <libgen.h> +#include <stdlib.h> +#include <stdio.h> +#include <string.h> +#include <sys/types.h> +#include <sys/stat.h> +#include <sys/param.h> +#include <fcntl.h> +#include <unistd.h> #include "symbol.h" -#include "thread.h" +#include "strlist.h" -#include "debug.h" - -#include <asm/bug.h> #include <libelf.h> #include <gelf.h> #include <elf.h> @@ -18,18 +24,6 @@ #define NT_GNU_BUILD_ID 3 #endif -enum dso_origin { - DSO__ORIG_KERNEL = 0, - DSO__ORIG_JAVA_JIT, - DSO__ORIG_BUILD_ID_CACHE, - DSO__ORIG_FEDORA, - DSO__ORIG_UBUNTU, - DSO__ORIG_BUILDID, - DSO__ORIG_DSO, - DSO__ORIG_KMODULE, - DSO__ORIG_NOT_FOUND, -}; - static void dsos__add(struct list_head *head, struct dso *dso); static struct map *map__new2(u64 start, struct dso *dso, enum map_type type); static int dso__load_kernel_sym(struct dso *self, struct map *map, @@ -126,8 +120,8 @@ static void map_groups__fixup_end(struct map_groups *self) static struct symbol *symbol__new(u64 start, u64 len, const char *name) { size_t namelen = strlen(name) + 1; - struct symbol *self = zalloc(symbol_conf.priv_size + - sizeof(*self) + namelen); + struct symbol *self = calloc(1, (symbol_conf.priv_size + + sizeof(*self) + namelen)); if (self == NULL) return NULL; @@ -178,7 +172,7 @@ static void dso__set_basename(struct dso *self) struct dso *dso__new(const char *name) { - struct dso *self = zalloc(sizeof(*self) + strlen(name) + 1); + struct dso *self = calloc(1, sizeof(*self) + strlen(name) + 1); if (self != NULL) { int i; @@ -870,8 +864,8 @@ out_close: if (err == 0) return nr; out: - pr_warning("%s: problems reading %s PLT info.\n", - __func__, self->long_name); + pr_debug("%s: problems reading %s PLT info.\n", + __func__, self->long_name); return 0; } @@ -1025,7 +1019,7 @@ static int dso__load_sym(struct dso *self, struct map *map, const char *name, } curr_map->map_ip = identity__map_ip; curr_map->unmap_ip = identity__map_ip; - curr_dso->origin = DSO__ORIG_KERNEL; + curr_dso->origin = self->origin; map_groups__insert(kmap->kmaps, curr_map); dsos__add(&dsos__kernel, curr_dso); dso__set_loaded(curr_dso, map->type); @@ -1394,13 +1388,13 @@ static int dso__kernel_module_get_build_id(struct dso *self) return 0; } -static int map_groups__set_modules_path_dir(struct map_groups *self, char *dirname) +static int map_groups__set_modules_path_dir(struct map_groups *self, char *dir_name) { struct dirent *dent; - DIR *dir = opendir(dirname); + DIR *dir = opendir(dir_name); if (!dir) { - pr_debug("%s: cannot open %s dir\n", __func__, dirname); + pr_debug("%s: cannot open %s dir\n", __func__, dir_name); return -1; } @@ -1413,7 +1407,7 @@ static int map_groups__set_modules_path_dir(struct map_groups *self, char *dirna continue; snprintf(path, sizeof(path), "%s/%s", - dirname, dent->d_name); + dir_name, dent->d_name); if (map_groups__set_modules_path_dir(self, path) < 0) goto failure; } else { @@ -1433,7 +1427,7 @@ static int map_groups__set_modules_path_dir(struct map_groups *self, char *dirna continue; snprintf(path, sizeof(path), "%s/%s", - dirname, dent->d_name); + dir_name, dent->d_name); long_name = strdup(path); if (long_name == NULL) @@ -1470,8 +1464,8 @@ static int map_groups__set_modules_path(struct map_groups *self) */ static struct map *map__new2(u64 start, struct dso *dso, enum map_type type) { - struct map *self = zalloc(sizeof(*self) + - (dso->kernel ? sizeof(struct kmap) : 0)); + struct map *self = calloc(1, (sizeof(*self) + + (dso->kernel ? sizeof(struct kmap) : 0))); if (self != NULL) { /* * ->end will be filled after we load all the symbols @@ -1895,6 +1889,17 @@ out_fail: return -1; } +size_t vmlinux_path__fprintf(FILE *fp) +{ + int i; + size_t printed = 0; + + for (i = 0; i < vmlinux_path__nr_entries; ++i) + printed += fprintf(fp, "[%d] %s\n", i, vmlinux_path[i]); + + return printed; +} + static int setup_list(struct strlist **list, const char *list_str, const char *list_name) { @@ -1964,3 +1969,46 @@ int map_groups__create_kernel_maps(struct map_groups *self, map_groups__fixup_end(self); return 0; } + +static int hex(char ch) +{ + if ((ch >= '0') && (ch <= '9')) + return ch - '0'; + if ((ch >= 'a') && (ch <= 'f')) + return ch - 'a' + 10; + if ((ch >= 'A') && (ch <= 'F')) + return ch - 'A' + 10; + return -1; +} + +/* + * While we find nice hex chars, build a long_val. + * Return number of chars processed. + */ +int hex2u64(const char *ptr, u64 *long_val) +{ + const char *p = ptr; + *long_val = 0; + + while (*p) { + const int hex_val = hex(*p); + + if (hex_val < 0) + break; + + *long_val = (*long_val << 4) | hex_val; + p++; + } + + return p - ptr; +} + +char *strxfrchar(char *s, char from, char to) +{ + char *p = s; + + while ((p = strchr(p, from)) != NULL) + *p++ = to; + + return s; +} diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h index f30a37428919..757fae3f5ee0 100644 --- a/tools/perf/util/symbol.h +++ b/tools/perf/util/symbol.h @@ -3,10 +3,11 @@ #include <linux/types.h> #include <stdbool.h> -#include "types.h" +#include <stdint.h> +#include "map.h" #include <linux/list.h> #include <linux/rbtree.h> -#include "event.h" +#include <stdio.h> #define DEBUG_CACHE_DIR ".debug" @@ -29,6 +30,9 @@ static inline char *bfd_demangle(void __used *v, const char __used *c, #endif #endif +int hex2u64(const char *ptr, u64 *val); +char *strxfrchar(char *s, char from, char to); + /* * libelf 0.8.x and earlier do not support ELF_C_READ_MMAP; * for newer versions we can use mmap to reduce memory usage: @@ -44,6 +48,8 @@ static inline char *bfd_demangle(void __used *v, const char __used *c, #define DMGL_ANSI (1 << 1) /* Include const, volatile, etc */ #endif +#define BUILD_ID_SIZE 20 + struct symbol { struct rb_node rb_node; u64 start; @@ -88,6 +94,11 @@ struct ref_reloc_sym { u64 unrelocated_addr; }; +struct map_symbol { + struct map *map; + struct symbol *sym; +}; + struct addr_location { struct thread *thread; struct map *map; @@ -106,6 +117,7 @@ struct dso { u8 has_build_id:1; u8 kernel:1; u8 hit:1; + u8 annotate_warned:1; unsigned char origin; u8 sorted_by_name; u8 loaded; @@ -150,6 +162,19 @@ size_t dsos__fprintf_buildid(FILE *fp, bool with_hits); size_t dso__fprintf_buildid(struct dso *self, FILE *fp); size_t dso__fprintf(struct dso *self, enum map_type type, FILE *fp); + +enum dso_origin { + DSO__ORIG_KERNEL = 0, + DSO__ORIG_JAVA_JIT, + DSO__ORIG_BUILD_ID_CACHE, + DSO__ORIG_FEDORA, + DSO__ORIG_UBUNTU, + DSO__ORIG_BUILDID, + DSO__ORIG_DSO, + DSO__ORIG_KMODULE, + DSO__ORIG_NOT_FOUND, +}; + char dso__symtab_origin(const struct dso *self); void dso__set_long_name(struct dso *self, char *name); void dso__set_build_id(struct dso *self, void *build_id); @@ -169,4 +194,6 @@ int kallsyms__parse(const char *filename, void *arg, int symbol__init(void); bool symbol_type__is_a(char symbol_type, enum map_type map_type); +size_t vmlinux_path__fprintf(FILE *fp); + #endif /* __PERF_SYMBOL */ diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c index fa968312ee7d..9bbe27d75306 100644 --- a/tools/perf/util/thread.c +++ b/tools/perf/util/thread.c @@ -7,6 +7,37 @@ #include "util.h" #include "debug.h" +int find_all_tid(int pid, pid_t ** all_tid) +{ + char name[256]; + int items; + struct dirent **namelist = NULL; + int ret = 0; + int i; + + sprintf(name, "/proc/%d/task", pid); + items = scandir(name, &namelist, NULL, NULL); + if (items <= 0) + return -ENOENT; + *all_tid = malloc(sizeof(pid_t) * items); + if (!*all_tid) { + ret = -ENOMEM; + goto failure; + } + + for (i = 0; i < items; i++) + (*all_tid)[i] = atoi(namelist[i]->d_name); + + ret = items; + +failure: + for (i=0; i<items; i++) + free(namelist[i]); + free(namelist); + + return ret; +} + void map_groups__init(struct map_groups *self) { int i; @@ -241,46 +272,6 @@ static int map_groups__fixup_overlappings(struct map_groups *self, return 0; } -void maps__insert(struct rb_root *maps, struct map *map) -{ - struct rb_node **p = &maps->rb_node; - struct rb_node *parent = NULL; - const u64 ip = map->start; - struct map *m; - - while (*p != NULL) { - parent = *p; - m = rb_entry(parent, struct map, rb_node); - if (ip < m->start) - p = &(*p)->rb_left; - else - p = &(*p)->rb_right; - } - - rb_link_node(&map->rb_node, parent, p); - rb_insert_color(&map->rb_node, maps); -} - -struct map *maps__find(struct rb_root *maps, u64 ip) -{ - struct rb_node **p = &maps->rb_node; - struct rb_node *parent = NULL; - struct map *m; - - while (*p != NULL) { - parent = *p; - m = rb_entry(parent, struct map, rb_node); - if (ip < m->start) - p = &(*p)->rb_left; - else if (ip > m->end) - p = &(*p)->rb_right; - else - return m; - } - - return NULL; -} - void thread__insert_map(struct thread *self, struct map *map) { map_groups__fixup_overlappings(&self->mg, map); @@ -336,15 +327,3 @@ size_t perf_session__fprintf(struct perf_session *self, FILE *fp) return ret; } - -struct symbol *map_groups__find_symbol(struct map_groups *self, - enum map_type type, u64 addr, - symbol_filter_t filter) -{ - struct map *map = map_groups__find(self, type, addr); - - if (map != NULL) - return map__find_symbol(map, map->map_ip(map, addr), filter); - - return NULL; -} diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h index dcf70303e58e..9c488fcadec9 100644 --- a/tools/perf/util/thread.h +++ b/tools/perf/util/thread.h @@ -5,14 +5,6 @@ #include <unistd.h> #include "symbol.h" -struct map_groups { - struct rb_root maps[MAP__NR_TYPES]; - struct list_head removed_maps[MAP__NR_TYPES]; -}; - -size_t __map_groups__fprintf_maps(struct map_groups *self, - enum map_type type, FILE *fp); - struct thread { struct rb_node rb_node; struct map_groups mg; @@ -23,29 +15,16 @@ struct thread { int comm_len; }; -void map_groups__init(struct map_groups *self); +struct perf_session; + +int find_all_tid(int pid, pid_t ** all_tid); int thread__set_comm(struct thread *self, const char *comm); int thread__comm_len(struct thread *self); struct thread *perf_session__findnew(struct perf_session *self, pid_t pid); void thread__insert_map(struct thread *self, struct map *map); int thread__fork(struct thread *self, struct thread *parent); -size_t map_groups__fprintf_maps(struct map_groups *self, FILE *fp); size_t perf_session__fprintf(struct perf_session *self, FILE *fp); -void maps__insert(struct rb_root *maps, struct map *map); -struct map *maps__find(struct rb_root *maps, u64 addr); - -static inline void map_groups__insert(struct map_groups *self, struct map *map) -{ - maps__insert(&self->maps[map->type], map); -} - -static inline struct map *map_groups__find(struct map_groups *self, - enum map_type type, u64 addr) -{ - return maps__find(&self->maps[type], addr); -} - static inline struct map *thread__find_map(struct thread *self, enum map_type type, u64 addr) { @@ -62,26 +41,4 @@ void thread__find_addr_location(struct thread *self, enum map_type type, u64 addr, struct addr_location *al, symbol_filter_t filter); -struct symbol *map_groups__find_symbol(struct map_groups *self, - enum map_type type, u64 addr, - symbol_filter_t filter); - -static inline struct symbol *map_groups__find_function(struct map_groups *self, - u64 addr, - symbol_filter_t filter) -{ - return map_groups__find_symbol(self, MAP__FUNCTION, addr, filter); -} - -struct map *map_groups__find_by_name(struct map_groups *self, - enum map_type type, const char *name); - -int __map_groups__create_kernel_maps(struct map_groups *self, - struct map *vmlinux_maps[MAP__NR_TYPES], - struct dso *kernel); -int map_groups__create_kernel_maps(struct map_groups *self, - struct map *vmlinux_maps[MAP__NR_TYPES]); - -struct map *map_groups__new_module(struct map_groups *self, u64 start, - const char *filename); #endif /* __PERF_THREAD_H */ diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h index 0f5b2a6f1080..52701087ce04 100644 --- a/tools/perf/util/util.h +++ b/tools/perf/util/util.h @@ -295,6 +295,13 @@ extern void *xmemdupz(const void *data, size_t len); extern char *xstrndup(const char *str, size_t len); extern void *xrealloc(void *ptr, size_t size) __attribute__((weak)); +static inline void *xzalloc(size_t size) +{ + void *buf = xmalloc(size); + + return memset(buf, 0, size); +} + static inline void *zalloc(size_t size) { return calloc(1, size); @@ -309,6 +316,7 @@ static inline int has_extension(const char *filename, const char *ext) { size_t len = strlen(filename); size_t extlen = strlen(ext); + return len > extlen && !memcmp(filename + len - extlen, ext, extlen); } @@ -322,6 +330,7 @@ static inline int has_extension(const char *filename, const char *ext) #undef isalnum #undef tolower #undef toupper + extern unsigned char sane_ctype[256]; #define GIT_SPACE 0x01 #define GIT_DIGIT 0x02 |