Re: [PATCH v5 2/2] perf/core: Disable context fast-swap for sigtrap events
From: Peter Zijlstra
Date: Fri Nov 20 2026 - 12:46:32 EST
On Fri, Nov 20, 2026 at 06:37:06PM +0100, Peter Zijlstra wrote:
> On Thu, Nov 19, 2026 at 09:51:01AM +0000, Tengda Wu wrote:
> > perf_sigtrap() contains:
> >
> > WARN_ON_ONCE(event->ctx->task != current)
> >
> > The warning assumes that an event's ctx->task always matches the task
> > that ends up running the pending task_work. That assumption can be
> > broken by the context fast-swap path.
> >
> > The failure sequence is:
> >
> > 1. A -> B switch, perf_event_context_sched_out() takes the fast path
> > and does WRITE_ONCE(ctx->task, next) / WRITE_ONCE(next_ctx->task, task).
> > An NMI then hits while the context is swapped, and
> > __perf_event_overflow() queues pending task_work on the *current*
> > task (A), even though the event now belongs to B's swapped-in
> > context view.
> >
> > 2. B -> A switch. Because A now has pending task_work,
> > nr_no_switch_fast is non-zero, so the slow path is taken and
> > ctx->task is *not* swapped back.
> >
> > 3. A returns to user mode, task_work runs perf_pending_task() ->
> > perf_sigtrap(), and event->ctx->task no longer equals current,
> > tripping the WARN_ON_ONCE().
> >
> > The context fast-swap path was already excluded for events that need
> > stable task association (inherit + sample read). sigtrap events have
> > the same requirement: the SIGTRAP is targeted at ctx->task and must not
> > be delivered after ctx->task has been swapped to another task.
>
> Oh gawd, let me guess, another one of those tracepoint events, right?
>
> Must be someone that doesn't respect perf_ctx_disable(). Fixing that
> will close this hole.
Here, does this work?
The clue is right in perf_event_context_sched_out():
perf_ctx_disable(ctx, 0);
/* PMIs are disabled; ctx->nr_no_switch_fast is stable. */
if (local_read(&ctx->nr_no_switch_fast) ||
local_read(&next_ctx->nr_no_switch_fast)) {
There must be no overflow after perf_ctx_disable().
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -1236,6 +1236,7 @@ extern const struct perf_event_attr *per
extern void perf_event_print_debug(void);
extern void perf_pmu_disable(struct pmu *pmu);
extern void perf_pmu_enable(struct pmu *pmu);
+extern bool perf_pmu_enabled(struct pmu *pmu);
extern void perf_sched_cb_dec(struct pmu *pmu);
extern void perf_sched_cb_inc(struct pmu *pmu);
extern int perf_event_task_disable(void);
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -1366,6 +1366,12 @@ void perf_pmu_enable(struct pmu *pmu)
pmu->pmu_enable(pmu);
}
+bool perf_pmu_enabled(struct pmu *pmu)
+{
+ int *count = &this_cpc(pmu)->pmu_disable_count;
+ return *count == 0;
+}
+
static void perf_assert_pmu_disabled(struct pmu *pmu)
{
int *count = &this_cpc(pmu)->pmu_disable_count;
@@ -11151,6 +11157,9 @@ static int perf_swevent_match(struct per
struct perf_sample_data *data,
struct pt_regs *regs)
{
+ if (!perf_pmu_enabled(event->pmu))
+ return 0;
+
if (event->attr.type != type)
return 0;
@@ -11531,6 +11540,9 @@ static int perf_tp_event_match(struct pe
struct perf_raw_record *raw,
struct pt_regs *regs)
{
+ if (!perf_pmu_enabled(event->pmu))
+ return 0;
+
if (event->hw.state & PERF_HES_STOPPED)
return 0;
/*