[tip: perf/urgent] perf: Fix list corruption in perf_cgroup_switch()

[tip-bot2 for Song Liu]

The following commit has been merged into the perf/urgent branch of tip:

Commit-ID:     5f4e5ce638e6a490b976ade4a40017b40abb2da0
Gitweb:        https://git.kernel.org/tip/5f4e5ce638e6a490b976ade4a40017b40abb2da0
Author:        Song Liu <song@xxxxxxxxxx>
AuthorDate:    Thu, 03 Feb 2022 16:40:57 -08:00
Committer:     Peter Zijlstra <peterz@xxxxxxxxxxxxx>
CommitterDate: Sun, 06 Feb 2022 22:37:27 +01:00

perf: Fix list corruption in perf_cgroup_switch()

There's list corruption on cgrp_cpuctx_list. This happens on the
following path:

  perf_cgroup_switch: list_for_each_entry(cgrp_cpuctx_list)
      cpu_ctx_sched_in
         ctx_sched_in
            ctx_pinned_sched_in
              merge_sched_in
                  perf_cgroup_event_disable: remove the event from the list

Use list_for_each_entry_safe() to allow removing an entry during
iteration.

Fixes: 058fe1c0440e ("perf/core: Make cgroup switch visit only cpuctxs with cgroup events")
Signed-off-by: Song Liu <song@xxxxxxxxxx>
Reviewed-by: Rik van Riel <riel@xxxxxxxxxxx>
Signed-off-by: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
Link: https://lkml.kernel.org/r/20220204004057.2961252-1-song@xxxxxxxxxx
---
 kernel/events/core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/events/core.c b/kernel/events/core.c
index 57c7197..6859229 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -839,7 +839,7 @@ static DEFINE_PER_CPU(struct list_head, cgrp_cpuctx_list);
  */
 static void perf_cgroup_switch(struct task_struct *task, int mode)
 {
-	struct perf_cpu_context *cpuctx;
+	struct perf_cpu_context *cpuctx, *tmp;
 	struct list_head *list;
 	unsigned long flags;
 
@@ -850,7 +850,7 @@ static void perf_cgroup_switch(struct task_struct *task, int mode)
 	local_irq_save(flags);
 
 	list = this_cpu_ptr(&cgrp_cpuctx_list);
-	list_for_each_entry(cpuctx, list, cgrp_cpuctx_entry) {
+	list_for_each_entry_safe(cpuctx, tmp, list, cgrp_cpuctx_entry) {
 		WARN_ON_ONCE(cpuctx->ctx.nr_cgroups == 0);
 
 		perf_ctx_lock(cpuctx, cpuctx->task_ctx);