Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Sat, 8 Oct 2022 14:41:47 +0200
From:   Peter Zijlstra <peterz@infradead.org>
To:     Marco Elver <elver@google.com>
Cc:     Ingo Molnar <mingo@redhat.com>,
        Arnaldo Carvalho de Melo <acme@kernel.org>,
        Mark Rutland <mark.rutland@arm.com>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Jiri Olsa <jolsa@kernel.org>,
        Namhyung Kim <namhyung@kernel.org>,
        linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
        kasan-dev@googlegroups.com, Dmitry Vyukov <dvyukov@google.com>
Subject: Re: [PATCH] perf: Fix missing SIGTRAPs
Message-ID: <Y0FwC1yo1pcyL9J/@hirez.programming.kicks-ass.net>
References: <20220927121322.1236730-1-elver@google.com>
 <Yz7ZLaT4jW3Y9EYS@hirez.programming.kicks-ass.net>
 <Yz7fWw8duIOezSW1@elver.google.com>
 <Yz78MMMJ74tBw0gu@hirez.programming.kicks-ass.net>
 <Yz/zXpF1yLshrJm/@elver.google.com>
 <Y0Ak/D05KhJeKaed@hirez.programming.kicks-ass.net>
 <Y0AwaxcJNOWhMKXP@elver.google.com>
 <Y0BQYxewPB/6KWLz@elver.google.com>
 <Y0E3uG7jOywn7vy3@elver.google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y0E3uG7jOywn7vy3@elver.google.com>
Precedence: bulk

On Sat, Oct 08, 2022 at 10:41:28AM +0200, Marco Elver wrote:
> The below patch to the sigtrap_threads test can repro the issue (when
> run lots of them concurrently again). It also illustrates the original
> problem we're trying to solve, where the event never gets rearmed again
> and the test times out (doesn't happen with the almost-working fix).

Excellent, that helps. Also, I'm an idiot ;-)

The below seems to fix it for me.

---
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -3441,7 +3448,8 @@ static void perf_event_context_sched_out
 			perf_pmu_disable(pmu);
 
 			/* PMIs are disabled; ctx->nr_pending is stable. */
-			if (local_read(&ctx->nr_pending)) {
+			if (local_read(&ctx->nr_pending) ||
+			    local_read(&next_ctx->nr_pending)) {
 				/*
 				 * Must not swap out ctx when there's pending
 				 * events that rely on the ctx->task relation.