Re: [PATCH v2] perf: Rewrite core context handling

[Ravi Bangoria <ravi.bangoria@amd.com>]

On 11-Oct-22 11:17 PM, Peter Zijlstra wrote:
> On Tue, Oct 11, 2022 at 04:02:56PM +0200, Peter Zijlstra wrote:
>> On Tue, Oct 11, 2022 at 06:49:55PM +0530, Ravi Bangoria wrote:
>>> On 11-Oct-22 4:59 PM, Peter Zijlstra wrote:
>>>> On Sat, Oct 08, 2022 at 11:54:24AM +0530, Ravi Bangoria wrote:
>>>>
>>>>> +static void perf_event_swap_task_ctx_data(struct perf_event_context *prev_ctx,
>>>>> +					  struct perf_event_context *next_ctx)
>>>>> +{
>>>>> +	struct perf_event_pmu_context *prev_epc, *next_epc;
>>>>> +
>>>>> +	if (!prev_ctx->nr_task_data)
>>>>> +		return;
>>>>> +
>>>>> +	prev_epc = list_first_entry(&prev_ctx->pmu_ctx_list,
>>>>> +				    struct perf_event_pmu_context,
>>>>> +				    pmu_ctx_entry);
>>>>> +	next_epc = list_first_entry(&next_ctx->pmu_ctx_list,
>>>>> +				    struct perf_event_pmu_context,
>>>>> +				    pmu_ctx_entry);
>>>>> +
>>>>> +	while (&prev_epc->pmu_ctx_entry != &prev_ctx->pmu_ctx_list &&
>>>>> +	       &next_epc->pmu_ctx_entry != &next_ctx->pmu_ctx_list) {
>>>>> +
>>>>> +		WARN_ON_ONCE(prev_epc->pmu != next_epc->pmu);
>>>>> +
>>>>> +		/*
>>>>> +		 * PMU specific parts of task perf context can require
>>>>> +		 * additional synchronization. As an example of such
>>>>> +		 * synchronization see implementation details of Intel
>>>>> +		 * LBR call stack data profiling;
>>>>> +		 */
>>>>> +		if (prev_epc->pmu->swap_task_ctx)
>>>>> +			prev_epc->pmu->swap_task_ctx(prev_epc, next_epc);
>>>>> +		else
>>>>> +			swap(prev_epc->task_ctx_data, next_epc->task_ctx_data);
>>>>
>>>> Did I forget to advance the iterators here?
>>>
>>> Yeah. Seems so. I overlooked it too.
>>
>> OK; so I'm not slowly going crazy staring at this code ;-) Let me go add
>> it now then. :-)
>>
>> But first I gotta taxi the kids around for a bit, bbl.
> 
> OK, so I've been going over the perf_event_pmu_context life-time thing
> as well, there were a bunch of XXXs there and I'm not sure Im happy with
> things, but I'd also forgotten most of it.
> 
> Ideally epc works like it's a regular member of ctx -- locking wise that
> is, but I'm not sure we can make that stick -- see the ctx->mutex issues
> we have with put_ctx().
> 
> As such, I'm going to have to re-audit all the epc usage to see if
> pure ctx->lock is sufficient.
> 
> I did do make epc RCU freed, because pretty much everything is and that
> was easy enough to make happen -- it means we don't need to worry about
> that.
> 
> But I'm going cross-eyes from staring at this all day, so more tomorrow.
> The below is what I currently have.
> 
> ---
> --- a/include/linux/perf_event.h
> +++ b/include/linux/perf_event.h
> @@ -833,13 +833,13 @@ struct perf_event {
>   *           `--------[1:n]---------'     `-[n:1]-> pmu <-[1:n]-'
>   *
>   *
> - * XXX destroy epc when empty
> - *   refcount, !rcu
> + * epc lifetime is refcount based and RCU freed (similar to perf_event_context).
> + * epc locking is as if it were a member of perf_event_context; specifically:
>   *
> - * XXX epc locking
> + *   modification, both: ctx->mutex && ctx->lock
> + *   reading, either: ctx->mutex || ctx->lock
>   *
> - *   event->pmu_ctx            ctx->mutex && inactive
> - *   ctx->pmu_ctx_list         ctx->mutex && ctx->lock
> + * XXX except this isn't true ... see put_pmu_ctx().
>   *
>   */
>  struct perf_event_pmu_context {
> @@ -857,6 +857,7 @@ struct perf_event_pmu_context {
>  	unsigned int			nr_events;
>  
>  	atomic_t			refcount; /* event <-> epc */
> +	struct rcu_head			rcu_head;
>  
>  	void				*task_ctx_data; /* pmu specific data */
>  	/*
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -1727,6 +1727,10 @@ perf_event_groups_next(struct perf_event
>  	return NULL;
>  }
>  
> +#define perf_event_groups_for_cpu_pmu(event, groups, cpu, pmu)		\
> +	for (event = perf_event_groups_first(groups, cpu, pmu, NULL);	\
> +	     event; event = perf_event_groups_next(event, pmu))
> +
>  /*
>   * Iterate through the whole groups tree.
>   */
> @@ -3366,6 +3370,14 @@ static void perf_event_sync_stat(struct
>  	}
>  }
>  
> +#define list_for_each_entry_double(pos1, pos2, head1, head2, member)	\
> +	for (pos1 = list_first_entry(head1, typeof(*pos1), member),	\
> +	     pos2 = list_first_entry(head2, typeof(*pos2), member);	\
> +	     !list_entry_is_head(pos1, head1, member) &&		\
> +	     !list_entry_is_head(pos2, head2, member);			\
> +	     pos1 = list_next_entry(pos1, member),			\
> +	     pos2 = list_next_entry(pos2, member))
> +
>  static void perf_event_swap_task_ctx_data(struct perf_event_context *prev_ctx,
>  					  struct perf_event_context *next_ctx)
>  {
> @@ -3374,16 +3386,9 @@ static void perf_event_swap_task_ctx_dat
>  	if (!prev_ctx->nr_task_data)
>  		return;
>  
> -	prev_epc = list_first_entry(&prev_ctx->pmu_ctx_list,
> -				    struct perf_event_pmu_context,
> -				    pmu_ctx_entry);
> -	next_epc = list_first_entry(&next_ctx->pmu_ctx_list,
> -				    struct perf_event_pmu_context,
> -				    pmu_ctx_entry);
> -
> -	while (&prev_epc->pmu_ctx_entry != &prev_ctx->pmu_ctx_list &&
> -	       &next_epc->pmu_ctx_entry != &next_ctx->pmu_ctx_list) {
> -
> +	list_for_each_entry_double(prev_epc, next_epc,
> +				   &prev_ctx->pmu_ctx_list, &next_ctx->pmu_ctx_list,
> +				   pmu_ctx_entry) {

There are more places which can use list_for_each_entry_double().
I'll fix those.

>  		WARN_ON_ONCE(prev_epc->pmu != next_epc->pmu);
>  
>  		/*
> @@ -3706,7 +3711,6 @@ static noinline int visit_groups_merge(s
>  		perf_assert_pmu_disabled((*evt)->pmu_ctx->pmu);
>  	}
>  
> -
>  	min_heapify_all(&event_heap, &perf_min_heap);
>  
>  	while (event_heap.nr) {
> @@ -3845,7 +3849,6 @@ ctx_sched_in(struct perf_event_context *
>  		/* start ctx time */
>  		__update_context_time(ctx, false);
>  		perf_cgroup_set_timestamp(cpuctx);
> -		// XXX ctx->task =? task
>  		/*
>  		 * CPU-release for the below ->is_active store,
>  		 * see __load_acquire() in perf_event_time_now()
> @@ -4815,6 +4818,15 @@ find_get_pmu_context(struct pmu *pmu, st
>  
>  	__perf_init_event_pmu_context(new, pmu);
>  
> +	/*
> +	 * XXX
> +	 *
> +	 * lockdep_assert_held(&ctx->mutex);
> +	 *
> +	 * can't because perf_event_init_task() doesn't actually hold the
> +	 * child_ctx->mutex.
> +	 */
> +
>  	raw_spin_lock_irq(&ctx->lock);
>  	list_for_each_entry(epc, &ctx->pmu_ctx_list, pmu_ctx_entry) {
>  		if (epc->pmu == pmu) {
> @@ -4849,6 +4861,14 @@ static void get_pmu_ctx(struct perf_even
>  	WARN_ON_ONCE(!atomic_inc_not_zero(&epc->refcount));
>  }
>  
> +static void free_epc_rcu(struct rcu_head *head)
> +{
> +	struct perf_event_pmu_context *epc = container_of(head, typeof(*epc), rcu_head);
> +
> +	kfree(epc->task_ctx_data);
> +	kfree(epc);
> +}
> +
>  static void put_pmu_ctx(struct perf_event_pmu_context *epc)
>  {
>  	unsigned long flags;
> @@ -4859,7 +4879,14 @@ static void put_pmu_ctx(struct perf_even
>  	if (epc->ctx) {
>  		struct perf_event_context *ctx = epc->ctx;
>  
> -		// XXX ctx->mutex
> +		/*
> +		 * XXX
> +		 *
> +		 * lockdep_assert_held(&ctx->mutex);
> +		 *
> +		 * can't because of the call-site in _free_event()/put_event()
> +		 * which isn't always called under ctx->mutex.
> +		 */

Yes. I came across the same and could not figure out how to solve
this. So Just kept XXX as is.

>  
>  		WARN_ON_ONCE(list_empty(&epc->pmu_ctx_entry));
>  		raw_spin_lock_irqsave(&ctx->lock, flags);
> @@ -4874,17 +4901,15 @@ static void put_pmu_ctx(struct perf_even
>  	if (epc->embedded)
>  		return;
>  
> -	kfree(epc->task_ctx_data);
> -	kfree(epc);
> +	call_rcu(&epc->rcu_head, free_epc_rcu);
>  }
>  
>  static void perf_event_free_filter(struct perf_event *event);
>  
>  static void free_event_rcu(struct rcu_head *head)
>  {
> -	struct perf_event *event;
> +	struct perf_event *event = container_of(head, typeof(*event), rcu_head);
>  
> -	event = container_of(head, struct perf_event, rcu_head);
>  	if (event->ns)
>  		put_pid_ns(event->ns);
>  	perf_event_free_filter(event);
> @@ -12643,13 +12668,6 @@ perf_event_create_kernel_counter(struct
>  		goto err_alloc;
>  	}
>  
> -	pmu_ctx = find_get_pmu_context(pmu, ctx, event);
> -	if (IS_ERR(pmu_ctx)) {
> -		err = PTR_ERR(pmu_ctx);
> -		goto err_ctx;
> -	}
> -	event->pmu_ctx = pmu_ctx;
> -
>  	WARN_ON_ONCE(ctx->parent_ctx);
>  	mutex_lock(&ctx->mutex);
>  	if (ctx->task == TASK_TOMBSTONE) {
> @@ -12657,6 +12675,13 @@ perf_event_create_kernel_counter(struct
>  		goto err_unlock;
>  	}
>  
> +	pmu_ctx = find_get_pmu_context(pmu, ctx, event);
> +	if (IS_ERR(pmu_ctx)) {
> +		err = PTR_ERR(pmu_ctx);
> +		goto err_unlock;
> +	}
> +	event->pmu_ctx = pmu_ctx;

We should call find_get_pmu_context() with ctx->mutex held and thus
above perf_event_create_kernel_counter() change. Is my understanding
correct?

> +
>  	if (!task) {
>  		/*
>  		 * Check if the @cpu we're creating an event for is online.
> @@ -12668,13 +12693,13 @@ perf_event_create_kernel_counter(struct
>  			container_of(ctx, struct perf_cpu_context, ctx);
>  		if (!cpuctx->online) {
>  			err = -ENODEV;
> -			goto err_unlock;
> +			goto err_pmu_ctx;
>  		}
>  	}
>  
>  	if (!exclusive_event_installable(event, ctx)) {
>  		err = -EBUSY;
> -		goto err_unlock;
> +		goto err_pmu_ctx;
>  	}
>  
>  	perf_install_in_context(ctx, event, event->cpu);
> @@ -12683,9 +12708,10 @@ perf_event_create_kernel_counter(struct
>  
>  	return event;
>  
> +err_pmu_ctx:
> +	put_pmu_ctx(pmu_ctx);
>  err_unlock:
>  	mutex_unlock(&ctx->mutex);
> -err_ctx:
>  	perf_unpin_context(ctx);
>  	put_ctx(ctx);
>  err_alloc:
> @@ -12702,9 +12728,7 @@ static void __perf_pmu_remove(struct per
>  {
>  	struct perf_event *event, *sibling;
>  
> -	for (event = perf_event_groups_first(groups, cpu, pmu, NULL);
> -	     event; event = perf_event_groups_next(event, pmu)) {
> -
> +	perf_event_groups_for_cpu_pmu(event, groups, cpu, pmu) {
>  		perf_remove_from_context(event, 0);
>  		unaccount_event_cpu(event, cpu);
>  		put_pmu_ctx(event->pmu_ctx);
> @@ -12998,7 +13022,7 @@ void perf_event_free_task(struct task_st
>  	struct perf_event_context *ctx;
>  	struct perf_event *event, *tmp;
>  
> -	ctx = rcu_dereference(task->perf_event_ctxp);
> +	ctx = rcu_access_pointer(task->perf_event_ctxp);

We dereference ctx pointer but with mutex and lock held. And thus
rcu_access_pointer() is sufficient. Is my understanding correct?

Thanks,
Ravi