From: sashiko-bot@kernel.org
To: "Umesh Nerlige Ramappa" <umesh.nerlige.ramappa@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 2/2] drm/i915/perf: Remove sysfs group on OA config idr_alloc() failure
Date: Tue, 08 Sep 2026 23:26:01 +0000 [thread overview]
Message-ID: <20260908232601.DBAE11F00A3A@smtp.kernel.org> (raw)
In-Reply-To: <20260908231511.1217426-6-umesh.nerlige.ramappa@intel.com>
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [Critical] Concurrent execution of i915_perf_add_config_ioctl and i915_perf_unregister can lead to a Use-After-Free of perf->metrics_kobj.
--
commit e80590029db9506ad40df9f792adcff0ddb3c4d7
Author: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
drm/i915/perf: Remove sysfs group on OA config idr_alloc() failure
This commit fixes a resource leak where an OA config's sysfs group was not
removed if idr_alloc() failed. It adds a new remove_sysfs error path to
clean up the sysfs group.
> diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c
> index 96823dcd3fae2..abda7b1c49569 100644
> --- a/drivers/gpu/drm/i915/i915_perf.c
> +++ b/drivers/gpu/drm/i915/i915_perf.c
[ ... ]
> @@ -4738,6 +4738,8 @@ int i915_perf_add_config_ioctl(struct drm_device *dev, void *data,
>
> return id;
>
> +remove_sysfs:
> + sysfs_remove_group(perf->metrics_kobj, &oa_config->sysfs_metric);
[Severity: Critical]
This is a pre-existing issue, but can this sequence lead to a Use-After-Free
of perf->metrics_kobj during a concurrent hardware hot-unplug?
If a user-space process repeatedly calls the DRM_IOCTL_I915_PERF_ADD_CONFIG
ioctl while the driver is being unbound:
CPU1 (Hot-unplug/Unbind)
i915_perf_unregister()
kobject_put(perf->metrics_kobj);
perf->metrics_kobj = NULL;
CPU2 (ioctl)
i915_perf_add_config_ioctl()
// passes early NULL check for perf->metrics_kobj before CPU1 clears it
Then CPU2 could eventually reach this newly added error path (or the existing
create_dynamic_oa_sysfs_entry() call) and pass the freed perf->metrics_kobj
pointer to sysfs_remove_group(), causing a Use-After-Free.
Since i915_driver_unregister() calls i915_perf_unregister() before calling
drm_dev_unplug(), the DRM core hasn't yet blocked incoming ioctls.
Does this lockless teardown of the kobject require synchronization or holding
a reference during the ioctl?
> sysfs_err:
> mutex_unlock(&perf->metrics_lock);
> reg_err:
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260908231511.1217426-4-umesh.nerlige.ramappa@intel.com?part=2
next prev parent reply other threads:[~2026-09-08 23:26 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-08 23:15 [PATCH 0/2] Minor fixes to OA in i915 Umesh Nerlige Ramappa
2026-09-08 23:15 ` [PATCH 1/2] drm/i915/perf: Enable the stream before installing its fd Umesh Nerlige Ramappa
2026-09-08 23:29 ` sashiko-bot
2026-09-08 23:15 ` [PATCH 2/2] drm/i915/perf: Remove sysfs group on OA config idr_alloc() failure Umesh Nerlige Ramappa
2026-09-08 23:26 ` sashiko-bot [this message]
2026-09-09 0:19 ` ✓ i915.CI.BAT: success for Minor fixes to OA in i915 Patchwork
2026-09-09 17:10 ` ✓ i915.CI.Full: " Patchwork
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=20260908232601.DBAE11F00A3A@smtp.kernel.org \
--to=sashiko-bot@kernel.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=sashiko-reviews@lists.linux.dev \
--cc=umesh.nerlige.ramappa@intel.com \
/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