Commit e3be4266 authored by Linus Torvalds's avatar Linus Torvalds
Browse files

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

Pull perf fixes from Thomas Gleixner:
 "Another pile of fixes for perf:

   - Plug overflows and races in the core code

   - Sanitize the flow of the perf syscall so we error out before
     handling the more complex and hard to undo setups

   - Improve and fix Broadwell and Skylake hardware support

   - Revert a fix which broke what it tried to fix in perf tools

   - A couple of smaller fixes in various places of perf tools"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Fix copying of /proc/kcore
  perf intel-pt: Remove no_force_psb from documentation
  perf probe: Use existing routine to look for a kernel module by dso->short_name
  perf/x86: Change test_aperfmperf() and test_intel() to static
  tools lib traceevent: Fix string handling in heterogeneous arch environments
  perf record: Avoid infinite loop at buildid processing with no samples
  perf: Fix races in computing the header sizes
  perf: Fix u16 overflows
  perf: Restructure perf syscall point of no return
  perf/x86/intel: Fix Skylake FRONTEND MSR extrareg mask
  perf/x86/intel/pebs: Add PEBS frontend profiling for Skylake
  perf/x86/intel: Make the CYCLE_ACTIVITY.* constraint on Broadwell more specific
  perf tools: Bool functions shouldn't return -1
  tools build: Add test for presence of __get_cpuid() gcc builtin
  tools build: Add test for presence of numa_num_possible_cpus() in libnuma
  Revert "perf symbols: Fix mismatched declarations for elf_getphdrnum"
  perf stat: Fix per-pkg event reporting bug
parents 73f479b2 2530e399

arch/x86/include/asm/msr-index.h

+2 −0
Original line number Diff line number Diff line
@@ -141,6 +141,8 @@ 
#define DEBUGCTLMSR_BTS_OFF_USR		(1UL << 10)

#define DEBUGCTLMSR_FREEZE_LBRS_ON_PMI	(1UL << 11)



#define MSR_PEBS_FRONTEND		0x000003f7



#define MSR_IA32_POWER_CTL		0x000001fc



#define MSR_IA32_MC0_CTL		0x00000400

arch/x86/kernel/cpu/perf_event.h

+1 −0
Original line number Diff line number Diff line
@@ -47,6 +47,7 @@ enum extra_reg_type { 
	EXTRA_REG_RSP_1 = 1,	/* offcore_response_1 */

	EXTRA_REG_LBR   = 2,	/* lbr_select */

	EXTRA_REG_LDLAT = 3,	/* ld_lat_threshold */

	EXTRA_REG_FE    = 4,    /* fe_* */



	EXTRA_REG_MAX		/* number of entries needed */

};

arch/x86/kernel/cpu/perf_event_intel.c

+15 −2
Original line number Diff line number Diff line
@@ -205,6 +205,11 @@ static struct extra_reg intel_skl_extra_regs[] __read_mostly = { 
	INTEL_UEVENT_EXTRA_REG(0x01b7, MSR_OFFCORE_RSP_0, 0x3fffff8fffull, RSP_0),

	INTEL_UEVENT_EXTRA_REG(0x01bb, MSR_OFFCORE_RSP_1, 0x3fffff8fffull, RSP_1),

	INTEL_UEVENT_PEBS_LDLAT_EXTRA_REG(0x01cd),

	/*

	 * Note the low 8 bits eventsel code is not a continuous field, containing

	 * some #GPing bits. These are masked out.

	 */

	INTEL_UEVENT_EXTRA_REG(0x01c6, MSR_PEBS_FRONTEND, 0x7fff17, FE),

	EVENT_EXTRA_END

};
@@ -250,7 +255,7 @@ struct event_constraint intel_bdw_event_constraints[] = { 
	FIXED_EVENT_CONSTRAINT(0x003c, 1),	/* CPU_CLK_UNHALTED.CORE */

	FIXED_EVENT_CONSTRAINT(0x0300, 2),	/* CPU_CLK_UNHALTED.REF */

	INTEL_UEVENT_CONSTRAINT(0x148, 0x4),	/* L1D_PEND_MISS.PENDING */

	INTEL_EVENT_CONSTRAINT(0xa3, 0x4),	/* CYCLE_ACTIVITY.* */

	INTEL_UEVENT_CONSTRAINT(0x8a3, 0x4),	/* CYCLE_ACTIVITY.CYCLES_L1D_MISS */

	EVENT_CONSTRAINT_END

};
@@ -2891,6 +2896,8 @@ PMU_FORMAT_ATTR(offcore_rsp, "config1:0-63"); 


PMU_FORMAT_ATTR(ldlat, "config1:0-15");



PMU_FORMAT_ATTR(frontend, "config1:0-23");



static struct attribute *intel_arch3_formats_attr[] = {

	&format_attr_event.attr,

	&format_attr_umask.attr,
@@ -2907,6 +2914,11 @@ static struct attribute *intel_arch3_formats_attr[] = { 
	NULL,

};



static struct attribute *skl_format_attr[] = {

	&format_attr_frontend.attr,

	NULL,

};



static __initconst const struct x86_pmu core_pmu = {

	.name			= "core",

	.handle_irq		= x86_pmu_handle_irq,
@@ -3516,7 +3528,8 @@ __init int intel_pmu_init(void) 


		x86_pmu.hw_config = hsw_hw_config;

		x86_pmu.get_event_constraints = hsw_get_event_constraints;

		x86_pmu.cpu_events = hsw_events_attrs;

		x86_pmu.format_attrs = merge_attr(intel_arch3_formats_attr,

						  skl_format_attr);

		WARN_ON(!x86_pmu.format_attrs);

		x86_pmu.cpu_events = hsw_events_attrs;

		pr_cont("Skylake events, ");

arch/x86/kernel/cpu/perf_event_msr.c

+2 −2
Original line number Diff line number Diff line
@@ -10,12 +10,12 @@ enum perf_msr_id { 
	PERF_MSR_EVENT_MAX,

};



bool test_aperfmperf(int idx)

static bool test_aperfmperf(int idx)

{

	return boot_cpu_has(X86_FEATURE_APERFMPERF);

}



bool test_intel(int idx)

static bool test_intel(int idx)

{

	if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL ||

	    boot_cpu_data.x86 != 6)

kernel/events/core.c

+81 −33
Original line number Diff line number Diff line
@@ -1243,11 +1243,7 @@ static inline void perf_event__state_init(struct perf_event *event) 
					      PERF_EVENT_STATE_INACTIVE;

}



/*

 * Called at perf_event creation and when events are attached/detached from a

 * group.

 */

static void perf_event__read_size(struct perf_event *event)

static void __perf_event_read_size(struct perf_event *event, int nr_siblings)

{

	int entry = sizeof(u64); /* value */

	int size = 0;
@@ -1263,7 +1259,7 @@ static void perf_event__read_size(struct perf_event *event) 
		entry += sizeof(u64);



	if (event->attr.read_format & PERF_FORMAT_GROUP) {

		nr += event->group_leader->nr_siblings;

		nr += nr_siblings;

		size += sizeof(u64);

	}
@@ -1271,14 +1267,11 @@ static void perf_event__read_size(struct perf_event *event) 
	event->read_size = size;

}



static void perf_event__header_size(struct perf_event *event)

static void __perf_event_header_size(struct perf_event *event, u64 sample_type)

{

	struct perf_sample_data *data;

	u64 sample_type = event->attr.sample_type;

	u16 size = 0;



	perf_event__read_size(event);



	if (sample_type & PERF_SAMPLE_IP)

		size += sizeof(data->ip);
@@ -1303,6 +1296,17 @@ static void perf_event__header_size(struct perf_event *event) 
	event->header_size = size;

}



/*

 * Called at perf_event creation and when events are attached/detached from a

 * group.

 */

static void perf_event__header_size(struct perf_event *event)

{

	__perf_event_read_size(event,

			       event->group_leader->nr_siblings);

	__perf_event_header_size(event, event->attr.sample_type);

}



static void perf_event__id_header_size(struct perf_event *event)

{

	struct perf_sample_data *data;
@@ -1330,6 +1334,27 @@ static void perf_event__id_header_size(struct perf_event *event) 
	event->id_header_size = size;

}



static bool perf_event_validate_size(struct perf_event *event)

{

	/*

	 * The values computed here will be over-written when we actually

	 * attach the event.

	 */

	__perf_event_read_size(event, event->group_leader->nr_siblings + 1);

	__perf_event_header_size(event, event->attr.sample_type & ~PERF_SAMPLE_READ);

	perf_event__id_header_size(event);



	/*

	 * Sum the lot; should not exceed the 64k limit we have on records.

	 * Conservative limit to allow for callchains and other variable fields.

	 */

	if (event->read_size + event->header_size +

	    event->id_header_size + sizeof(struct perf_event_header) >= 16*1024)

		return false;



	return true;

}



static void perf_group_attach(struct perf_event *event)

{

	struct perf_event *group_leader = event->group_leader, *pos;
@@ -8297,13 +8322,35 @@ SYSCALL_DEFINE5(perf_event_open, 


	if (move_group) {

		gctx = group_leader->ctx;

		mutex_lock_double(&gctx->mutex, &ctx->mutex);

	} else {

		mutex_lock(&ctx->mutex);

	}



	if (!perf_event_validate_size(event)) {

		err = -E2BIG;

		goto err_locked;

	}



	/*

	 * Must be under the same ctx::mutex as perf_install_in_context(),

	 * because we need to serialize with concurrent event creation.

	 */

	if (!exclusive_event_installable(event, ctx)) {

		/* exclusive and group stuff are assumed mutually exclusive */

		WARN_ON_ONCE(move_group);



		err = -EBUSY;

		goto err_locked;

	}



	WARN_ON_ONCE(ctx->parent_ctx);



	if (move_group) {

		/*

		 * See perf_event_ctx_lock() for comments on the details

		 * of swizzling perf_event::ctx.

		 */

		mutex_lock_double(&gctx->mutex, &ctx->mutex);



		perf_remove_from_context(group_leader, false);



		list_for_each_entry(sibling, &group_leader->sibling_list,
@@ -8311,13 +8358,7 @@ SYSCALL_DEFINE5(perf_event_open, 
			perf_remove_from_context(sibling, false);

			put_ctx(gctx);

		}

	} else {

		mutex_lock(&ctx->mutex);

	}



	WARN_ON_ONCE(ctx->parent_ctx);



	if (move_group) {

		/*

		 * Wait for everybody to stop referencing the events through

		 * the old lists, before installing it on new lists.
@@ -8349,22 +8390,29 @@ SYSCALL_DEFINE5(perf_event_open, 
		perf_event__state_init(group_leader);

		perf_install_in_context(ctx, group_leader, group_leader->cpu);

		get_ctx(ctx);

	}



	if (!exclusive_event_installable(event, ctx)) {

		err = -EBUSY;

		mutex_unlock(&ctx->mutex);

		fput(event_file);

		goto err_context;

		/*

		 * Now that all events are installed in @ctx, nothing

		 * references @gctx anymore, so drop the last reference we have

		 * on it.

		 */

		put_ctx(gctx);

	}



	/*

	 * Precalculate sample_data sizes; do while holding ctx::mutex such

	 * that we're serialized against further additions and before

	 * perf_install_in_context() which is the point the event is active and

	 * can use these values.

	 */

	perf_event__header_size(event);

	perf_event__id_header_size(event);



	perf_install_in_context(ctx, event, event->cpu);

	perf_unpin_context(ctx);



	if (move_group) {

	if (move_group)

		mutex_unlock(&gctx->mutex);

		put_ctx(gctx);

	}

	mutex_unlock(&ctx->mutex);



	put_online_cpus();
@@ -8375,12 +8423,6 @@ SYSCALL_DEFINE5(perf_event_open, 
	list_add_tail(&event->owner_entry, &current->perf_event_list);

	mutex_unlock(&current->perf_event_mutex);



	/*

	 * Precalculate sample_data sizes

	 */

	perf_event__header_size(event);

	perf_event__id_header_size(event);



	/*

	 * Drop the reference on the group_event after placing the

	 * new event on the sibling_list. This ensures destruction
@@ -8391,6 +8433,12 @@ SYSCALL_DEFINE5(perf_event_open, 
	fd_install(event_fd, event_file);

	return event_fd;



err_locked:

	if (move_group)

		mutex_unlock(&gctx->mutex);

	mutex_unlock(&ctx->mutex);

/* err_file: */

	fput(event_file);

err_context:

	perf_unpin_context(ctx);

	put_ctx(ctx);

CRA Git | Maintained and supported by SUSTech CRA and CCSE