From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
patches@lists.linux.dev,
"Peter Zijlstra (Intel)" <peterz@infradead.org>,
Ravi Bangoria <ravi.bangoria@amd.com>,
Sasha Levin <sashal@kernel.org>
Subject: [PATCH 6.6 06/87] perf: Unify perf_event_free_task() / perf_event_exit_task_context()
Date: Tue, 25 Aug 2026 15:25:29 +0200 [thread overview]
Message-ID: <20260825132542.066931268@linuxfoundation.org> (raw)
In-Reply-To: <20260825132541.813800447@linuxfoundation.org>
6.6-stable review patch. If anyone has any objections, please let me know.
------------------
From: Peter Zijlstra <peterz@infradead.org>
[ Upstream commit 90661365021a6d0d7f3a2c5046ebe33e4df53b92 ]
Both perf_event_free_task() and perf_event_exit_task_context() are
very similar, except perf_event_exit_task_context() is a little more
generic / makes less assumptions.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ravi Bangoria <ravi.bangoria@amd.com>
Link: https://lkml.kernel.org/r/20250307193723.274039710@infradead.org
Stable-dep-of: 42c5ca1f0a28 ("perf/core: Fix group leader use-after-free after sibling detach")
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
kernel/events/core.c | 96 +++++++++++++--------------------------------------
1 file changed, 26 insertions(+), 70 deletions(-)
--- a/kernel/events/core.c
+++ b/kernel/events/core.c
@@ -2393,6 +2393,7 @@ __perf_remove_from_context(struct perf_e
perf_cgroup_event_disable(event, ctx);
perf_event_set_state(event, min(event->state, state));
+
if (flags & DETACH_GROUP)
perf_group_detach(event);
if (flags & DETACH_CHILD)
@@ -13276,13 +13277,11 @@ perf_event_exit_event(struct perf_event
perf_event_wakeup(event);
}
-static void perf_event_exit_task_context(struct task_struct *child)
+static void perf_event_exit_task_context(struct task_struct *child, bool exit)
{
struct perf_event_context *child_ctx, *clone_ctx = NULL;
struct perf_event *child_event, *next;
- WARN_ON_ONCE(child != current);
-
child_ctx = perf_pin_task_context(child);
if (!child_ctx)
return;
@@ -13305,7 +13304,8 @@ static void perf_event_exit_task_context
* in.
*/
raw_spin_lock_irq(&child_ctx->lock);
- task_ctx_sched_out(child_ctx, EVENT_ALL);
+ if (exit)
+ task_ctx_sched_out(child_ctx, EVENT_ALL);
/*
* Now that the context is inactive, destroy the task <-> ctx relation
@@ -13314,7 +13314,7 @@ static void perf_event_exit_task_context
RCU_INIT_POINTER(child->perf_event_ctxp, NULL);
put_ctx(child_ctx); /* cannot be last */
WRITE_ONCE(child_ctx->task, TASK_TOMBSTONE);
- put_task_struct(current); /* cannot be last */
+ put_task_struct(child); /* cannot be last */
clone_ctx = unclone_ctx(child_ctx);
raw_spin_unlock_irq(&child_ctx->lock);
@@ -13327,13 +13327,31 @@ static void perf_event_exit_task_context
* won't get any samples after PERF_RECORD_EXIT. We can however still
* get a few PERF_RECORD_READ events.
*/
- perf_event_task(child, child_ctx, 0);
+ if (exit)
+ perf_event_task(child, child_ctx, 0);
list_for_each_entry_safe(child_event, next, &child_ctx->event_list, event_entry)
perf_event_exit_event(child_event, child_ctx, 0);
mutex_unlock(&child_ctx->mutex);
+ if (!exit) {
+ /*
+ * perf_event_release_kernel() could still have a reference on
+ * this context. In that case we must wait for these events to
+ * have been freed (in particular all their references to this
+ * task must've been dropped).
+ *
+ * Without this copy_process() will unconditionally free this
+ * task (irrespective of its reference count) and
+ * _free_event()'s put_task_struct(event->hw.target) will be a
+ * use-after-free.
+ *
+ * Wait for all events to drop their context reference.
+ */
+ wait_var_event(&child_ctx->refcount,
+ refcount_read(&child_ctx->refcount) == 1);
+ }
put_ctx(child_ctx);
}
@@ -13361,7 +13379,7 @@ void perf_event_exit_task(struct task_st
}
mutex_unlock(&child->perf_event_mutex);
- perf_event_exit_task_context(child);
+ perf_event_exit_task_context(child, true);
/*
* The perf_event_exit_task_context calls perf_event_task
@@ -13372,27 +13390,6 @@ void perf_event_exit_task(struct task_st
perf_event_task(child, NULL, 0);
}
-static void perf_free_event(struct perf_event *event,
- struct perf_event_context *ctx)
-{
- struct perf_event *parent = event->parent;
-
- if (WARN_ON_ONCE(!parent))
- return;
-
- mutex_lock(&parent->child_mutex);
- list_del_init(&event->child_list);
- mutex_unlock(&parent->child_mutex);
-
- put_event(parent);
-
- raw_spin_lock_irq(&ctx->lock);
- perf_group_detach(event);
- list_del_event(event, ctx);
- raw_spin_unlock_irq(&ctx->lock);
- free_event(event);
-}
-
/*
* Free a context as created by inheritance by perf_event_init_task() below,
* used by fork() in case of fail.
@@ -13402,48 +13399,7 @@ static void perf_free_event(struct perf_
*/
void perf_event_free_task(struct task_struct *task)
{
- struct perf_event_context *ctx;
- struct perf_event *event, *tmp;
-
- ctx = rcu_access_pointer(task->perf_event_ctxp);
- if (!ctx)
- return;
-
- mutex_lock(&ctx->mutex);
- raw_spin_lock_irq(&ctx->lock);
- /*
- * Destroy the task <-> ctx relation and mark the context dead.
- *
- * This is important because even though the task hasn't been
- * exposed yet the context has been (through child_list).
- */
- RCU_INIT_POINTER(task->perf_event_ctxp, NULL);
- WRITE_ONCE(ctx->task, TASK_TOMBSTONE);
- put_task_struct(task); /* cannot be last */
- raw_spin_unlock_irq(&ctx->lock);
-
-
- list_for_each_entry_safe(event, tmp, &ctx->event_list, event_entry)
- perf_free_event(event, ctx);
-
- mutex_unlock(&ctx->mutex);
-
- /*
- * perf_event_release_kernel() could've stolen some of our
- * child events and still have them on its free_list. In that
- * case we must wait for these events to have been freed (in
- * particular all their references to this task must've been
- * dropped).
- *
- * Without this copy_process() will unconditionally free this
- * task (irrespective of its reference count) and
- * _free_event()'s put_task_struct(event->hw.target) will be a
- * use-after-free.
- *
- * Wait for all events to drop their context reference.
- */
- wait_var_event(&ctx->refcount, refcount_read(&ctx->refcount) == 1);
- put_ctx(ctx); /* must be last */
+ perf_event_exit_task_context(task, false);
}
void perf_event_delayed_put(struct task_struct *task)
next prev parent reply other threads:[~2026-08-25 13:47 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-25 13:25 [PATCH 6.6 00/87] 6.6.154-rc1 review Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 01/87] PCI: host-generic: Fix NULL pointer dereference on 32-bit CAM systems Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 02/87] Bluetooth: RFCOMM: take rfcomm_mutex for the deferred setup accept Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 03/87] rndis_host: add overflow check in rndis_rx_fixup() Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 04/87] NTB: ntb_netdev: Preserve RX queue depth on allocation failure Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 05/87] serial: amba-pl011: synchronize DMA teardown Greg Kroah-Hartman
2026-08-25 13:25 ` Greg Kroah-Hartman [this message]
2026-08-25 13:25 ` [PATCH 6.6 07/87] perf/core: Fix group leader use-after-free after sibling detach Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 08/87] serial: qcom-geni: fix TX DMA buffer flush Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 09/87] serial: sc16is7xx: fix copy-paste errors in EFR_SWFLOWx_BIT constants Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 10/87] serial: sc16is7xx: convert bitmask definitions to use BIT() macro Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 11/87] serial: sc16is7xx: rename EFR mutex with generic name Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 12/87] serial: sc16is7xx: use guards for simple mutex locks Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 13/87] serial: sc16is7xx: enable THRI before filling TX FIFO Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 14/87] inet: frags: add inet_frag_putn() helper Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 15/87] ipv4: frags: remove ipq_put() Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 16/87] inet: frags: change inet_frag_kill() to defer refcount updates Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 17/87] inet: frags: save a pair of atomic operations in reassembly Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 18/87] inet: frags: publish queues before arming timer Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 19/87] mmc: atmel-mci: Fix use-after-free in atmci_remove due to race condition Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 20/87] xfs: dont use a xfs_log_iovec for ri_buf in log recovery Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 21/87] xfs: bounds-check buffer log items dirty bitmap Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 22/87] ALSA: dummy: Check card index validity at probe Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 23/87] ocfs2: fix missing metadata reservation for large xattrs Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 24/87] null_blk: fix UBSAN shift-out-of-bounds when zone_size is 0 or overflows Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 25/87] kcov: fix data corruption and race conditions on PREEMPT_RT Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 26/87] ext4: stop retrying saturated xattr cache entries Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 27/87] ext4: clear error before retrying inode xattr space fallback Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 28/87] xfs: validate attr entry pointer before field access Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 29/87] gpio: ml-ioh: use raw_spinlock_t for the register lock Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 30/87] s390/vfio_ccw: Free all memory if cp_init() fails Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 31/87] s390/vfio_ccw: Ensure first IDAW remains constant Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 32/87] s390/vfio_ccw: Calculate idal length based on idaw type Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 33/87] s390/vfio_ccw: Implement a crw lock Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 34/87] drm/amd/amdgpu: disable ASPM in some situations Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 35/87] drm/amd/display: Fix BT2020 YCbCr limited/full range input Greg Kroah-Hartman
2026-08-25 13:25 ` [PATCH 6.6 36/87] drm/amd/display: fix BT.2020 YCbCr limited output CSC matrix Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 37/87] drm/amdgpu: check ASPM on the dGPU host link Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 38/87] nfc: digital: clamp SENSF_RES length to the destination buffer Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 39/87] nfc: fdp: bound the device-reported read length and fix an skb leak Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 40/87] nfc: microread: validate target discovery payload lengths Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 41/87] nfc: llcp: bound the connect_sn TLV walk to the skb Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 42/87] nfc: llcp: fix OOB read and u8 offset wrap in TLV parsers Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 43/87] nfc: llcp: reject PDUs shorter than the LLCP header Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 44/87] nfc: pn533: purge fragmented skbs during cleanup Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 45/87] nfc: st21nfca: validate ATR_REQ length against the received frame Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 46/87] nfc: nci: fix out-of-bounds write in nci_target_auto_activated() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 47/87] nfc: nci: fix uninit-value in the RF discover/activated NTF handlers Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 48/87] nfc: nci: free destination parameters when closing a connection Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 49/87] ndisc: ndisc_send_redirect() cleanup Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 50/87] Input: byd - synchronize timer deletion before freeing private data Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 51/87] libceph: fix OOB read in decode_watchers() via missing bounds check Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 52/87] ipv4: reject undersized MTUs in ip_do_fragment() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 53/87] ipv6: fix use-after-free in ip6_finish_output2() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 54/87] nvmet-auth: zero the AUTH_RECEIVE response buffer Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 55/87] nvmet-fc: fix invalid free in LS IOD error path Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 56/87] nvmet-tcp: Do not WARN on remotely-controlled oversized SGL allocations Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 57/87] ASoC: sof: pcm: use snd_pcm_direction_name() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 58/87] ASoC: SOF: Relocate and rework functionality for PCM stream freeing Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 59/87] ASoC: SOF: pcm: Move period/buffer configuration print after platform open Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 60/87] ASoC: SOF: pcm: Add snd_sof_pcm specific wrappers for dev_dbg() and dev_err() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 61/87] ASoC: SOF: ipc4-pcm: Continue the pipeline trigger in case of IPC timeout Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 62/87] ASoC: codecs: lpass-tx-macro: Fix enum kcontrol accesses Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 63/87] mptcp: pm: fix data race in add_addr timer callback Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 64/87] HID: magicmouse: fix battery reporting for Bluetooth Magic Trackpad USB-C Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 65/87] HID: magicmouse: do not keep a stale msc->input if no input is claimed Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 66/87] HID: magicmouse: Prevent out-of-bounds (OOB) read during DOUBLE_REPORT_ID Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 67/87] HID: core: fix OOB read of field->usage in hid_set_field() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 68/87] xfrm: fix sk_dst_cache double-free in xfrm_user_policy() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 69/87] Bluetooth: hci_sync: Fix advertising data UAFs Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 70/87] Input: atkbd - skip deactivate for HONOR FMB-Ps internal keyboard Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 71/87] Input: atkbd - skip deactivate for HONOR ZQC-P Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 72/87] mptcp: pm: fix memory leak from alloc-during-teardown race Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 73/87] selinux: fix style issues in security/selinux/ss/policydb.h Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 74/87] selinux: fix style issues in security/selinux/ss/policydb.c Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 75/87] selinux: use known type instead of void pointer Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 76/87] selinux: avoid unnecessary indirection in struct level_datum Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 77/87] selinux: make more use of str_read() when loading the policy Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 78/87] selinux: use u16 for security classes Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 79/87] selinux: more strict policy parsing Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 80/87] selinux: reject a permission value exceeding the class permission count Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 81/87] selinux: require a classs permission values to cover its " Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 82/87] HID: nintendo: fix out-of-bounds read in joycon_ctlr_read_handler() Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 83/87] HID: core: fix number/pointer type confusion on long items Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 84/87] HID: sensor: custom: Fix use-after-free in enable_sensor Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 85/87] HID: hyperv: validate initial device info bounds Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 86/87] Bluetooth: hci_event: fix LE list UAF on reset Greg Kroah-Hartman
2026-08-25 13:26 ` [PATCH 6.6 87/87] Bluetooth: hci_event: validate LE Set CIG Parameters response Greg Kroah-Hartman
2026-08-25 19:32 ` [PATCH 6.6 00/87] 6.6.154-rc1 review Pavel Machek
2026-08-25 21:59 ` Florian Fainelli
2026-08-26 0:12 ` Shuah Khan
2026-08-26 6:22 ` Ron Economos
2026-08-26 9:46 ` Wentao Guan
2026-08-26 10:06 ` Barry K. Nathan
2026-08-26 10:32 ` Brett A C Sheffield
2026-08-26 11:53 ` Miguel Ojeda
2026-08-26 12:04 ` Peter Schneider
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=20260825132542.066931268@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=patches@lists.linux.dev \
--cc=peterz@infradead.org \
--cc=ravi.bangoria@amd.com \
--cc=sashal@kernel.org \
--cc=stable@vger.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