From: "tip-bot2 for Peter Zijlstra" <tip-bot2@linutronix.de>
To: linux-tip-commits@vger.kernel.org
Cc: "Peter Zijlstra (Intel)" <peterz@infradead.org>,
x86@kernel.org, linux-kernel@vger.kernel.org
Subject: [tip: perf/urgent] perf: Fix cpuctx refcounting
Date: Tue, 14 Nov 2023 21:56:49 -0000 [thread overview]
Message-ID: <169999900905.391.7191385772913382509.tip-bot2@tip-bot2> (raw)
In-Reply-To: <20230612093539.085862001@infradead.org>
The following commit has been merged into the perf/urgent branch of tip:
Commit-ID: cfaf1b99ac72d339c586d053885554dbe169d276
Gitweb: https://git.kernel.org/tip/cfaf1b99ac72d339c586d053885554dbe169d276
Author: Peter Zijlstra <peterz@infradead.org>
AuthorDate: Fri, 09 Jun 2023 12:34:46 +02:00
Committer: Peter Zijlstra <peterz@infradead.org>
CommitterDate: Tue, 14 Nov 2023 22:27:28 +01:00
perf: Fix cpuctx refcounting
Audit of the refcounting turned up that perf_pmu_migrate_context()
fails to migrate the ctx refcount.
Fixes: bd2756811766 ("perf: Rewrite core context handling")
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20230612093539.085862001@infradead.org
---
include/linux/perf_event.h | 13 ++++++++-----
kernel/events/core.c | 17 +++++++++++++++++
2 files changed, 25 insertions(+), 5 deletions(-)
diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h
index 4b93ff1..d2a15c0 100644
--- a/include/linux/perf_event.h
+++ b/include/linux/perf_event.h
@@ -843,11 +843,11 @@ struct perf_event {
};
/*
- * ,-----------------------[1:n]----------------------.
- * V V
- * perf_event_context <-[1:n]-> perf_event_pmu_context <--- perf_event
- * ^ ^ | |
- * `--------[1:n]---------' `-[n:1]-> pmu <-[1:n]-'
+ * ,-----------------------[1:n]------------------------.
+ * V V
+ * perf_event_context <-[1:n]-> perf_event_pmu_context <-[1:n]- perf_event
+ * | |
+ * `--[n:1]-> pmu <-[1:n]--'
*
*
* struct perf_event_pmu_context lifetime is refcount based and RCU freed
@@ -865,6 +865,9 @@ struct perf_event {
* ctx->mutex pinning the configuration. Since we hold a reference on
* group_leader (through the filedesc) it can't go away, therefore it's
* associated pmu_ctx must exist and cannot change due to ctx->mutex.
+ *
+ * perf_event holds a refcount on perf_event_context
+ * perf_event holds a refcount on perf_event_pmu_context
*/
struct perf_event_pmu_context {
struct pmu *pmu;
diff --git a/kernel/events/core.c b/kernel/events/core.c
index ce984aa..0825098 100644
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -4828,6 +4828,11 @@ find_get_pmu_context(struct pmu *pmu, struct perf_event_context *ctx,
void *task_ctx_data = NULL;
if (!ctx->task) {
+ /*
+ * perf_pmu_migrate_context() / __perf_pmu_install_event()
+ * relies on the fact that find_get_pmu_context() cannot fail
+ * for CPU contexts.
+ */
struct perf_cpu_pmu_context *cpc;
cpc = per_cpu_ptr(pmu->cpu_pmu_context, event->cpu);
@@ -12897,6 +12902,9 @@ static void __perf_pmu_install_event(struct pmu *pmu,
int cpu, struct perf_event *event)
{
struct perf_event_pmu_context *epc;
+ struct perf_event_context *old_ctx = event->ctx;
+
+ get_ctx(ctx); /* normally find_get_context() */
event->cpu = cpu;
epc = find_get_pmu_context(pmu, ctx, event);
@@ -12905,6 +12913,11 @@ static void __perf_pmu_install_event(struct pmu *pmu,
if (event->state >= PERF_EVENT_STATE_OFF)
event->state = PERF_EVENT_STATE_INACTIVE;
perf_install_in_context(ctx, event, cpu);
+
+ /*
+ * Now that event->ctx is updated and visible, put the old ctx.
+ */
+ put_ctx(old_ctx);
}
static void __perf_pmu_install(struct perf_event_context *ctx,
@@ -12943,6 +12956,10 @@ void perf_pmu_migrate_context(struct pmu *pmu, int src_cpu, int dst_cpu)
struct perf_event_context *src_ctx, *dst_ctx;
LIST_HEAD(events);
+ /*
+ * Since per-cpu context is persistent, no need to grab an extra
+ * reference.
+ */
src_ctx = &per_cpu_ptr(&perf_cpu_context, src_cpu)->ctx;
dst_ctx = &per_cpu_ptr(&perf_cpu_context, dst_cpu)->ctx;
next prev parent reply other threads:[~2023-11-14 21:56 UTC|newest]
Thread overview: 107+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-12 9:07 [PATCH v3 00/57] Scope-based Resource Management Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 01/57] dmaengine: ioat: Free up __cleanup() name Peter Zijlstra
2023-06-16 22:41 ` Dave Jiang
2023-06-12 9:07 ` [PATCH v3 02/57] apparmor: " Peter Zijlstra
2023-06-13 8:38 ` John Johansen
2023-06-13 10:28 ` John Johansen
2024-01-16 18:10 ` John Johansen
2023-06-12 9:07 ` [PATCH v3 03/57] locking: Introduce __cleanup() based infrastructure Peter Zijlstra
2023-06-13 10:55 ` Peter Zijlstra
2023-06-13 19:46 ` Kees Cook
2023-06-13 17:16 ` Miguel Ojeda
2023-08-14 7:24 ` Bartosz Golaszewski
2024-01-14 6:49 ` Dmitry Torokhov
2023-06-12 9:07 ` [PATCH v3 04/57] kbuild: Drop -Wdeclaration-after-statement Peter Zijlstra
2023-06-14 6:13 ` Masahiro Yamada
2023-06-14 8:25 ` Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 05/57] sched: Simplify get_nohz_timer_target() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 06/57] sched: Simplify sysctl_sched_uclamp_handler() Peter Zijlstra
2023-06-12 13:47 ` Dan Carpenter
2023-06-12 9:07 ` [PATCH v3 07/57] sched: Simplify: migrate_swap_stop() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 08/57] sched: Simplify wake_up_if_idle() Peter Zijlstra
2023-06-12 18:02 ` Boqun Feng
2023-06-12 20:02 ` Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 09/57] sched: Simplify ttwu() Peter Zijlstra
2023-06-12 13:51 ` Dan Carpenter
2023-06-12 14:08 ` Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 10/57] sched: Simplify sched_exec() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 11/57] sched: Simplify sched_tick_remote() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 12/57] sched: Simplify try_steal_cookie() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 13/57] sched: Simplify sched_core_cpu_{starting,deactivate}() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 14/57] sched: Simplify set_user_nice() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 15/57] sched: Simplify syscalls Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 16/57] sched: Simplify sched_{set,get}affinity() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 17/57] sched: Simplify yield_to() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 18/57] sched: Simplify sched_rr_get_interval() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 19/57] sched: Simplify sched_move_task() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 20/57] sched: Simplify tg_set_cfs_bandwidth() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 21/57] sched: Misc cleanups Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 22/57] perf: Fix cpuctx refcounting Peter Zijlstra
2023-11-14 21:56 ` tip-bot2 for Peter Zijlstra [this message]
2023-11-15 3:23 ` [tip: perf/urgent] perf/core: " tip-bot2 for Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 23/57] perf: Simplify perf_event_alloc() error path Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 24/57] perf: Simplify perf_pmu_register() " Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 25/57] perf: Simplify perf_fget_light() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 26/57] perf: Simplify event_function*() Peter Zijlstra
2023-06-12 14:46 ` Dan Carpenter
2023-06-12 14:58 ` Peter Zijlstra
2023-06-13 5:56 ` Namhyung Kim
2023-06-13 7:39 ` Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 27/57] perf: Simplify perf_cgroup_connect() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 28/57] perf; Simplify event_sched_in() Peter Zijlstra
2023-06-13 19:43 ` Kees Cook
2023-06-12 9:07 ` [PATCH v3 29/57] perf: Simplify: __perf_install_in_context() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 30/57] perf: Simplify: *perf_event_{dis,en}able*() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 31/57] perf: Simplify perf_event_modify_attr() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 32/57] perf: Simplify perf_event_context_sched_in() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 33/57] perf: Simplify perf_adjust_freq_unthr_context() Peter Zijlstra
2023-06-12 16:27 ` Linus Torvalds
2023-06-12 18:44 ` Peter Zijlstra
2023-06-12 18:55 ` Linus Torvalds
2023-06-12 20:05 ` Peter Zijlstra
2023-06-13 12:05 ` Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 34/57] perf: Simplify perf_event_*_on_exec() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 35/57] perf: Simplify *perf_event_read*() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 36/57] perf: Simplify find_get_pmu_context() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 37/57] perf: Simplify perf_read_group() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 38/57] perf: Simplify IOC_SET_OUTPUT Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 39/57] perf: Simplify perf_event_*_userpage() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 40/57] perf: Simplify perf_mmap_close()/perf_aux_sample_output() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 41/57] perf: Simplify __perf_event_output() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 42/57] perf: Simplify perf_iterate_sb() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 43/57] perf: Simplify perf_sw_event() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 44/57] perf: Simplify bpf_overflow_handler() Peter Zijlstra
2023-06-12 9:07 ` [PATCH v3 45/57] perf: Simplify perf_event_parse_addr_filter() Peter Zijlstra
2023-06-13 19:47 ` Kees Cook
2023-06-12 9:07 ` [PATCH v3 46/57] perf: Simplify pmu_dev_alloc() Peter Zijlstra
2023-06-12 9:44 ` Peter Zijlstra
2023-06-12 9:55 ` Greg KH
2023-06-12 12:18 ` Greg KH
2023-06-12 12:29 ` Paolo Bonzini
2023-06-12 12:58 ` Greg KH
2023-06-12 13:09 ` Greg KH
2023-11-15 9:28 ` [tip: perf/core] perf/core: Fix narrow startup race when creating the perf nr_addr_filters sysfs file tip-bot2 for Greg KH
2023-06-12 13:35 ` [PATCH v3 46/57] perf: Simplify pmu_dev_alloc() Greg KH
2023-06-12 14:13 ` Peter Zijlstra
2023-06-12 15:44 ` Greg KH
2023-06-13 7:34 ` Peter Zijlstra
2023-06-13 7:50 ` Greg KH
2023-06-13 10:50 ` Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 47/57] perf: Simplify perf_pmu_register() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 48/57] perf: Simplify perf_init_event() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 49/57] perf: Simplify perf_event_alloc() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 50/57] perf: Simplify sys_perf_event_open() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 51/57] perf: Simplify perf_event_create_kernel_counter() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 52/57] perf: Simplify perf_event_init_context() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 53/57] perf: Simplify perf_event_sysfs_init() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 54/57] perf: Misc cleanups Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 55/57] perf: Simplify find_get_context() Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 56/57] perf: Simplify perf_pmu_output_stop() Peter Zijlstra
2023-06-12 16:19 ` Linus Torvalds
2023-06-12 17:11 ` Sean Christopherson
2023-06-12 17:43 ` Linus Torvalds
2023-06-12 18:55 ` Peter Zijlstra
2023-06-12 9:08 ` [PATCH v3 57/57] perf: Simplify perf_install_in_context() Peter Zijlstra
2023-06-12 9:51 ` [PATCH v3 00/57] Scope-based Resource Management Peter Zijlstra
2023-06-12 16:37 ` Linus Torvalds
2025-06-14 9:20 ` [RFC] Expanding " Jemmy Wong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=169999900905.391.7191385772913382509.tip-bot2@tip-bot2 \
--to=tip-bot2@linutronix.de \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-tip-commits@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®