From: "Michał Grzelak" <michal.grzelak@intel.com>
To: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: "Michał Grzelak" <michal.grzelak@intel.com>,
intel-xe@lists.freedesktop.org,
"Lucas De Marchi" <demarchi@kernel.org>,
"Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Subject: Re: [PATCH] drm/xe/configfs: free aux. array after storing the wa_bb
Date: Tue, 2 Dec 2025 15:17:08 +0100 (CET) [thread overview]
Message-ID: <fc771b86-9283-7b4c-ce13-44da4c9e96c2@intel.com> (raw)
In-Reply-To: <aS351nEwG5YWKC7H@intel.com>
[-- Attachment #1: Type: text/plain, Size: 1969 bytes --]
On Mon, 1 Dec 2025, Rodrigo Vivi wrote:
> On Sat, Nov 29, 2025 at 11:25:41PM +0100, Michał Grzelak wrote:
>> When commands issued via configfs are executed mid- or post GPU context
>> restoration, a temporary buffer is reallocated in order to store all of
>> commands of varying count. However, after content of the buffer has been
>> copied to the desired location it is never freed, which causes following
>> memory leak to appear when running igt@xe_configfs@ctx-restore-mid-bb:
>>
>> kmemleak_alloc+0x4a/0x90
>> __kmalloc_node_track_caller_noprof+0x645/0x9d0
>> krealloc_node_align_noprof+0x206/0x3c0
>> wa_bb_store+0xf4/0x2c0 [xe]
>> ctx_restore_mid_bb_store+0x21/0x30 [xe]
>> configfs_write_iter+0xe6/0x150
>> vfs_write+0x231/0x540
>> ksys_write+0x6d/0xf0
>> __x64_sys_write+0x19/0x30
>> x64_sys_call+0x79/0x2350
>> do_syscall_64+0x93/0x10f0
>> entry_SYSCALL_64_after_hwframe+0x76/0x7e
>>
>> Deallocate temporary array upon finishing the store of batch buffer
>> into wa_bb.
>>
>> Fixes: 39ac06f7006f ("drm/xe/configfs: Add post context restore bb")
>
> this hash is wrong:
>
> Fixes: 39ac06f70062 ("drm/xe/configfs: Add post context restore bb")
>
>> Signed-off-by: Michał Grzelak <michal.grzelak@intel.com>
>> ---
>> drivers/gpu/drm/xe/xe_configfs.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_configfs.c b/drivers/gpu/drm/xe/xe_configfs.c
>> index 9f6251b1008b..f9a64352138b 100644
>> --- a/drivers/gpu/drm/xe/xe_configfs.c
>> +++ b/drivers/gpu/drm/xe/xe_configfs.c
>> @@ -789,6 +789,8 @@ static ssize_t wa_bb_store(struct wa_bb wa_bb[static XE_ENGINE_CLASS_MAX],
>>
>> memcpy(wa_bb, tmp_wa_bb, sizeof(tmp_wa_bb));
>>
>> + kfree(tmp);
>
> hmm...
>
> do we need to change the kfree added in the xe_config_device_release()?
You're right, we don't need that. Sorry for the noise from this false
positive.
BR,
Michał
>
>> +
>> return len;
>> }
>>
>> --
>> 2.45.2
>>
>
prev parent reply other threads:[~2025-12-02 14:17 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-29 22:25 [PATCH] drm/xe/configfs: free aux. array after storing the wa_bb Michał Grzelak
2025-11-29 23:02 ` ✗ CI.checkpatch: warning for " Patchwork
2025-11-29 23:04 ` ✓ CI.KUnit: success " Patchwork
2025-11-29 23:37 ` ✓ Xe.CI.BAT: " Patchwork
2025-11-30 0:40 ` ✗ Xe.CI.Full: failure " Patchwork
2025-12-01 20:25 ` [PATCH] " Rodrigo Vivi
2025-12-02 14:17 ` Michał Grzelak [this message]
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=fc771b86-9283-7b4c-ce13-44da4c9e96c2@intel.com \
--to=michal.grzelak@intel.com \
--cc=demarchi@kernel.org \
--cc=intel-xe@lists.freedesktop.org \
--cc=rodrigo.vivi@intel.com \
--cc=thomas.hellstrom@linux.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