action.c   [plain text]


/*
 * Copyright (c) 2011 Apple Computer, Inc. All rights reserved.
 *
 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
 * 
 * This file contains Original Code and/or Modifications of Original Code
 * as defined in and that are subject to the Apple Public Source License
 * Version 2.0 (the 'License'). You may not use this file except in
 * compliance with the License. The rights granted to you under the License
 * may not be used to create, or enable the creation or redistribution of,
 * unlawful or unlicensed copies of an Apple operating system, or to
 * circumvent, violate, or enable the circumvention or violation of, any
 * terms of an Apple operating system software license agreement.
 * 
 * Please obtain a copy of the License at
 * http://www.opensource.apple.com/apsl/ and read it before using this file.
 * 
 * The Original Code and all software distributed under the License are
 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
 * Please see the License for the specific language governing rights and
 * limitations under the License.
 * 
 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
 */

/*
 * Called from a trigger. Actually takes the data from the different
 * modules and puts them in a buffer
 */

#include <mach/mach_types.h>
#include <machine/machine_routines.h>
#include <kern/kalloc.h>
#include <kern/debug.h> /* panic */
#include <kern/thread.h>
#include <sys/errno.h>
#include <sys/vm.h>

#include <kperf/action.h>
#include <kperf/ast.h>
#include <kperf/buffer.h>
#include <kperf/callstack.h>
#include <kperf/context.h>
#include <kperf/kdebug_trigger.h>
#include <kperf/kperf.h>
#include <kperf/kperf_kpc.h>
#include <kperf/kperf_timer.h>
#include <kperf/pet.h>
#include <kperf/sample.h>
#include <kperf/thread_samplers.h>

#define ACTION_MAX (32)

/* the list of different actions to take */
struct action
{
	uint32_t sample;
	uint32_t ucallstack_depth;
	uint32_t kcallstack_depth;
	uint32_t userdata;
	int pid_filter;
};

/* the list of actions */
static unsigned actionc = 0;
static struct action *actionv = NULL;

/* should emit tracepoint on context switch */
int kperf_kdebug_cswitch = 0;

static kern_return_t
kperf_sample_internal(struct kperf_sample *sbuf,
                      struct kperf_context *context,
                      unsigned sample_what, unsigned sample_flags,
                      unsigned actionid, uint32_t ucallstack_depth)
{
	int pended_ucallstack = 0;
	int pended_th_dispatch = 0;

	/* not much point continuing here, but what to do ? return
	 * Shutdown? cut a tracepoint and continue?
	 */
	if (sample_what == 0) {
		return SAMPLE_CONTINUE;
	}

	/* callstacks should be explicitly ignored */
	if (sample_flags & SAMPLE_FLAG_EMPTY_CALLSTACK) {
		sample_what &= ~(SAMPLER_KSTACK | SAMPLER_USTACK);
	}

	context->cur_thread->kperf_pet_gen = kperf_pet_gen;
	boolean_t is_kernel = (context->cur_pid == 0);

	if (actionid && actionid <= actionc) {
		sbuf->kcallstack.nframes = actionv[actionid - 1].kcallstack_depth;
	} else {
		sbuf->kcallstack.nframes = MAX_CALLSTACK_FRAMES;
	}

	if (ucallstack_depth) {
		sbuf->ucallstack.nframes = ucallstack_depth;
	} else {
		sbuf->ucallstack.nframes = MAX_CALLSTACK_FRAMES;
	}

	sbuf->kcallstack.flags = CALLSTACK_VALID;
	sbuf->ucallstack.flags = CALLSTACK_VALID;

	/* an event occurred. Sample everything and dump it in a
	 * buffer.
	 */

	/* collect data from samplers */
	if (sample_what & SAMPLER_TH_INFO) {
		kperf_thread_info_sample(&sbuf->th_info, context);

		/* See if we should drop idle thread samples */
		if (!(sample_flags & SAMPLE_FLAG_IDLE_THREADS)) {
			if (sbuf->th_info.kpthi_runmode & 0x40) {
				return SAMPLE_CONTINUE;
			}
		}
	}

	if (sample_what & SAMPLER_TH_SNAPSHOT) {
		kperf_thread_snapshot_sample(&(sbuf->th_snapshot), context);
	}
	if (sample_what & SAMPLER_TH_SCHEDULING) {
		kperf_thread_scheduling_sample(&(sbuf->th_scheduling), context);
	}
	if (sample_what & SAMPLER_KSTACK) {
		if (sample_flags & SAMPLE_FLAG_CONTINUATION) {
			kperf_continuation_sample(&(sbuf->kcallstack), context);
		/* outside of interrupt context, backtrace the current thread */
		} else if (sample_flags & SAMPLE_FLAG_NON_INTERRUPT) {
			kperf_backtrace_sample(&(sbuf->kcallstack), context);
		} else {
			kperf_kcallstack_sample(&(sbuf->kcallstack), context);
		}
	}
	if (sample_what & SAMPLER_TK_SNAPSHOT) {
		kperf_task_snapshot_sample(&(sbuf->tk_snapshot), context);
	}

	/* sensitive ones */
	if (!is_kernel) {
		if (sample_what & SAMPLER_MEMINFO) {
			kperf_meminfo_sample(&(sbuf->meminfo), context);
		}

		if (sample_flags & SAMPLE_FLAG_PEND_USER) {
			if (sample_what & SAMPLER_USTACK) {
				pended_ucallstack = kperf_ucallstack_pend(context, sbuf->ucallstack.nframes);
			}

			if (sample_what & SAMPLER_TH_DISPATCH) {
				pended_th_dispatch = kperf_thread_dispatch_pend(context);
			}
		} else {
			if (sample_what & SAMPLER_USTACK) {
				kperf_ucallstack_sample(&(sbuf->ucallstack), context);
			}

			if (sample_what & SAMPLER_TH_DISPATCH) {
				kperf_thread_dispatch_sample(&(sbuf->th_dispatch), context);
			}
		}
	}

	if (sample_what & SAMPLER_PMC_THREAD) {
		kperf_kpc_thread_sample(&(sbuf->kpcdata), sample_what);
	} else if (sample_what & SAMPLER_PMC_CPU) {
		kperf_kpc_cpu_sample(&(sbuf->kpcdata), sample_what);
	}

	/* lookup the user tag, if any */
	uint32_t userdata;
	if (actionid && (actionid <= actionc)) {
		userdata = actionv[actionid - 1].userdata;
	} else {
		userdata = actionid;
	}

	/* avoid logging if this sample only pended samples */
	if (sample_flags & SAMPLE_FLAG_PEND_USER &&
	    !(sample_what & ~(SAMPLER_USTACK | SAMPLER_TH_DISPATCH)))
	{
		return SAMPLE_CONTINUE;
	}

	/* stash the data into the buffer
	 * interrupts off to ensure we don't get split
	 */
	boolean_t enabled = ml_set_interrupts_enabled(FALSE);

	BUF_DATA(PERF_GEN_EVENT | DBG_FUNC_START, sample_what,
	         actionid, userdata, sample_flags);

	if (sample_what & SAMPLER_TH_INFO) {
		kperf_thread_info_log(&sbuf->th_info);
	}
	if (sample_what & SAMPLER_TH_SCHEDULING) {
		kperf_thread_scheduling_log(&(sbuf->th_scheduling));
	}
	if (sample_what & SAMPLER_TH_SNAPSHOT) {
		kperf_thread_snapshot_log(&(sbuf->th_snapshot));
	}
	if (sample_what & SAMPLER_KSTACK) {
		kperf_kcallstack_log(&sbuf->kcallstack);
	}
	if (sample_what & SAMPLER_TK_SNAPSHOT) {
		kperf_task_snapshot_log(&(sbuf->tk_snapshot));
	}

	/* dump user stuff */
	if (!is_kernel) {
		/* dump meminfo */
		if (sample_what & SAMPLER_MEMINFO) {
			kperf_meminfo_log(&(sbuf->meminfo));
		}

		if (sample_flags & SAMPLE_FLAG_PEND_USER) {
			if (pended_ucallstack) {
				BUF_INFO(PERF_CS_UPEND);
			}

			if (pended_th_dispatch) {
				BUF_INFO(PERF_TI_DISPPEND);
			}
		} else {
			if (sample_what & SAMPLER_USTACK) {
				kperf_ucallstack_log(&(sbuf->ucallstack));
			}

			if (sample_what & SAMPLER_TH_DISPATCH) {
				kperf_thread_dispatch_log(&(sbuf->th_dispatch));
			}
		}
	}

	if (sample_what & SAMPLER_PMC_THREAD) {
		kperf_kpc_thread_log(&(sbuf->kpcdata));
	} else if (sample_what & SAMPLER_PMC_CPU) {
		kperf_kpc_cpu_log(&(sbuf->kpcdata));
	}

	BUF_DATA(PERF_GEN_EVENT | DBG_FUNC_END, sample_what);

	/* intrs back on */
	ml_set_interrupts_enabled(enabled);

	return SAMPLE_CONTINUE;
}

/* Translate actionid into sample bits and take a sample */
kern_return_t
kperf_sample(struct kperf_sample *sbuf,
             struct kperf_context *context,
             unsigned actionid, unsigned sample_flags)
{
	/* work out what to sample, if anything */
	if ((actionid > actionc) || (actionid == 0)) {
		return SAMPLE_SHUTDOWN;
	}

	/* check the pid filter against the context's current pid.
	 * filter pid == -1 means any pid
	 */
	int pid_filter = actionv[actionid - 1].pid_filter;
	if ((pid_filter != -1) && (pid_filter != context->cur_pid)) {
		return SAMPLE_CONTINUE;
	}

	/* the samplers to run */
	unsigned int sample_what = actionv[actionid - 1].sample;

	/* do the actual sample operation */
	return kperf_sample_internal(sbuf, context, sample_what,
	                             sample_flags, actionid,
	                             actionv[actionid - 1].ucallstack_depth);
}

void
kperf_kdebug_handler(uint32_t debugid, uintptr_t *starting_fp)
{
	uint32_t sample_flags = SAMPLE_FLAG_PEND_USER;
	struct kperf_context ctx;
	struct kperf_sample *sample = NULL;
	kern_return_t kr = KERN_SUCCESS;
	int s;

	if (!kperf_kdebug_should_trigger(debugid)) {
		return;
	}

	BUF_VERB(PERF_KDBG_HNDLR | DBG_FUNC_START, debugid);

	ctx.cur_thread = current_thread();
	ctx.cur_pid = task_pid(get_threadtask(ctx.cur_thread));
	ctx.trigger_type = TRIGGER_TYPE_KDEBUG;
	ctx.trigger_id = 0;

	s = ml_set_interrupts_enabled(0);

	sample = kperf_intr_sample_buffer();

	if (!ml_at_interrupt_context()) {
		sample_flags |= SAMPLE_FLAG_NON_INTERRUPT;
		ctx.starting_fp = starting_fp;
	}

	kr = kperf_sample(sample, &ctx, kperf_kdebug_get_action(), sample_flags);

	ml_set_interrupts_enabled(s);
	BUF_VERB(PERF_KDBG_HNDLR | DBG_FUNC_END, kr);
}

/*
 * This function allocates >2.3KB of the stack.  Prevent the compiler from
 * inlining this function into ast_taken and ensure the stack memory is only
 * allocated for the kperf AST.
 */
__attribute__((noinline))
void
kperf_thread_ast_handler(thread_t thread)
{
	BUF_INFO(PERF_AST_HNDLR | DBG_FUNC_START, thread, kperf_get_thread_flags(thread));

	/* ~2KB of the stack for the sample since this is called from AST */
	struct kperf_sample sbuf;
	memset(&sbuf, 0, sizeof(struct kperf_sample));

	task_t task = get_threadtask(thread);

	/* make a context, take a sample */
	struct kperf_context ctx;
	ctx.cur_thread = thread;
	ctx.cur_pid = task_pid(task);

	/* decode the flags to determine what to sample */
	unsigned int sample_what = 0;
	uint32_t flags = kperf_get_thread_flags(thread);

	if (flags & T_KPERF_AST_DISPATCH) {
		sample_what |= SAMPLER_TH_DISPATCH;
	}
	if (flags & T_KPERF_AST_CALLSTACK) {
		sample_what |= SAMPLER_USTACK;
		sample_what |= SAMPLER_TH_INFO;
	}

	uint32_t ucallstack_depth = T_KPERF_GET_CALLSTACK_DEPTH(flags);

	int r = kperf_sample_internal(&sbuf, &ctx, sample_what, 0, 0, ucallstack_depth);

	BUF_INFO(PERF_AST_HNDLR | DBG_FUNC_END, r);
}

/* register AST bits */
int
kperf_ast_pend(thread_t thread, uint32_t set_flags)
{
	/* can only pend on the current thread */
	if (thread != current_thread()) {
		panic("pending to non-current thread");
	}

	/* get our current bits */
	uint32_t flags = kperf_get_thread_flags(thread);

	/* see if it's already been done or pended */
	if (!(flags & set_flags)) {
		/* set the bit on the thread */
		flags |= set_flags;
		kperf_set_thread_flags(thread, flags);

		/* set the actual AST */
		act_set_kperf(thread);
		return 1;
	}

	return 0;
}

void
kperf_ast_set_callstack_depth(thread_t thread, uint32_t depth)
{
	uint32_t ast_flags = kperf_get_thread_flags(thread);
	uint32_t existing_callstack_depth = T_KPERF_GET_CALLSTACK_DEPTH(ast_flags);

	if (existing_callstack_depth != depth) {
		ast_flags &= ~T_KPERF_SET_CALLSTACK_DEPTH(depth);
		ast_flags |= T_KPERF_SET_CALLSTACK_DEPTH(depth);

		kperf_set_thread_flags(thread, ast_flags);
	}
}

int
kperf_kdbg_cswitch_get(void)
{
	return kperf_kdebug_cswitch;
}

int
kperf_kdbg_cswitch_set(int newval)
{
	kperf_kdebug_cswitch = newval;
	kperf_on_cpu_update();

	return 0;
}

/*
 * Action configuration
 */
unsigned int
kperf_action_get_count(void)
{
	return actionc;
}

int
kperf_action_set_samplers(unsigned actionid, uint32_t samplers)
{
	if ((actionid > actionc) || (actionid == 0)) {
		return EINVAL;
	}

	/* disallow both CPU and thread counters to be sampled in the same
	 * action */
	if ((samplers & SAMPLER_PMC_THREAD) && (samplers & SAMPLER_PMC_CPU)) {
		return EINVAL;
	}

	actionv[actionid - 1].sample = samplers;

	return 0;
}

int
kperf_action_get_samplers(unsigned actionid, uint32_t *samplers_out)
{
	if ((actionid > actionc)) {
		return EINVAL;
	}

	if (actionid == 0) {
		*samplers_out = 0; /* "NULL" action */
	} else {
		*samplers_out = actionv[actionid - 1].sample;
	}

	return 0;
}

int
kperf_action_set_userdata(unsigned actionid, uint32_t userdata)
{
	if ((actionid > actionc) || (actionid == 0)) {
		return EINVAL;
	}

	actionv[actionid - 1].userdata = userdata;

	return 0;
}

int
kperf_action_get_userdata(unsigned actionid, uint32_t *userdata_out)
{
	if ((actionid > actionc)) {
		return EINVAL;
	}

	if (actionid == 0) {
		*userdata_out = 0; /* "NULL" action */
	} else {
		*userdata_out = actionv[actionid - 1].userdata;
	}

	return 0;
}

int
kperf_action_set_filter(unsigned actionid, int pid)
{
	if ((actionid > actionc) || (actionid == 0)) {
		return EINVAL;
	}

	actionv[actionid - 1].pid_filter = pid;

	return 0;
}

int
kperf_action_get_filter(unsigned actionid, int *pid_out)
{
	if ((actionid > actionc)) {
		return EINVAL;
	}

	if (actionid == 0) {
		*pid_out = -1; /* "NULL" action */
	} else {
		*pid_out = actionv[actionid - 1].pid_filter;
	}

	return 0;
}

void
kperf_action_reset(void)
{
	for (unsigned int i = 0; i < actionc; i++) {
		kperf_action_set_samplers(i + 1, 0);
		kperf_action_set_userdata(i + 1, 0);
		kperf_action_set_filter(i + 1, -1);
		kperf_action_set_ucallstack_depth(i + 1, MAX_CALLSTACK_FRAMES);
		kperf_action_set_kcallstack_depth(i + 1, MAX_CALLSTACK_FRAMES);
	}
}

int
kperf_action_set_count(unsigned count)
{
	struct action *new_actionv = NULL, *old_actionv = NULL;
	unsigned old_count;

	/* easy no-op */
	if (count == actionc) {
		return 0;
	}

	/* TODO: allow shrinking? */
	if (count < actionc) {
		return EINVAL;
	}

	/* cap it for good measure */
	if (count > ACTION_MAX) {
		return EINVAL;
	}

	/* creating the action arror for the first time. create a few
	 * more things, too.
	 */
	if (actionc == 0) {
		int r;
		if ((r = kperf_init())) {
			return r;
		}
	}

	/* create a new array */
	new_actionv = kalloc_tag(count * sizeof(*new_actionv), VM_KERN_MEMORY_DIAG);
	if (new_actionv == NULL) {
		return ENOMEM;
	}

	old_actionv = actionv;
	old_count = actionc;

	if (old_actionv != NULL) {
		memcpy(new_actionv, actionv, actionc * sizeof(*actionv));
	}

	memset(&(new_actionv[actionc]), 0, (count - old_count) * sizeof(*actionv));

	for (unsigned int i = old_count; i < count; i++) {
		new_actionv[i].pid_filter = -1;
		new_actionv[i].ucallstack_depth = MAX_CALLSTACK_FRAMES;
		new_actionv[i].kcallstack_depth = MAX_CALLSTACK_FRAMES;
	}

	actionv = new_actionv;
	actionc = count;

	if (old_actionv != NULL) {
		kfree(old_actionv, old_count * sizeof(*actionv));
	}

	return 0;
}

int
kperf_action_set_ucallstack_depth(unsigned action_id, uint32_t depth)
{
	if ((action_id > actionc) || (action_id == 0)) {
		return EINVAL;
	}

	if (depth > MAX_CALLSTACK_FRAMES) {
		return EINVAL;
	}

	actionv[action_id - 1].ucallstack_depth = depth;

	return 0;
}

int
kperf_action_set_kcallstack_depth(unsigned action_id, uint32_t depth)
{
	if ((action_id > actionc) || (action_id == 0)) {
		return EINVAL;
	}

	if (depth > MAX_CALLSTACK_FRAMES) {
		return EINVAL;
	}

	actionv[action_id - 1].kcallstack_depth = depth;

	return 0;
}

int
kperf_action_get_ucallstack_depth(unsigned action_id, uint32_t * depth_out)
{
	if ((action_id > actionc)) {
		return EINVAL;
	}

	assert(depth_out);

	if (action_id == 0) {
		*depth_out = MAX_CALLSTACK_FRAMES;
	} else {
		*depth_out = actionv[action_id - 1].ucallstack_depth;
	}

	return 0;
}

int
kperf_action_get_kcallstack_depth(unsigned action_id, uint32_t * depth_out)
{
	if ((action_id > actionc)) {
		return EINVAL;
	}

	assert(depth_out);

	if (action_id == 0) {
		*depth_out = MAX_CALLSTACK_FRAMES;
	} else {
		*depth_out = actionv[action_id - 1].kcallstack_depth;
	}

	return 0;
}