All of lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Rob Clark" <robin.clark@oss.qualcomm.com>
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [PATCH v6 09/16] drm/msm/gem: Remove useless locking in GEM import
Date: Fri, 17 Jul 2026 21:02:27 +0000	[thread overview]
Message-ID: <20260717210228.278A71F00A3A@smtp.kernel.org> (raw)
In-Reply-To: <20260717204841.26120-10-robin.clark@oss.qualcomm.com>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:

Pre-existing issues:
- [High] Pre-existing error path vulnerability in `msm_gem_import` causes a double-free of the exporter's `sgt`, NULL pointer dereference, and wild pointer dereferences during cleanup.
--

--- Patch [9]: [PATCH v6 09/16] drm/msm/gem: Remove useless locking in GEM import ---
commit 19246e79feb40c4e0066b56b3385fea83fcc3d75
Author: Rob Clark <robin.clark@oss.qualcomm.com>

drm/msm/gem: Remove useless locking in GEM import
    
This patch removes unnecessary locking during the GEM object import process
in msm_gem_import().

> diff --git a/drivers/gpu/drm/msm/msm_gem.c b/drivers/gpu/drm/msm/msm_gem.c
> index 2cb3ab04f1250..a358fb5ca6d73 100644
> --- a/drivers/gpu/drm/msm/msm_gem.c
> +++ b/drivers/gpu/drm/msm/msm_gem.c
> @@ -1302,23 +1302,18 @@ struct drm_gem_object *msm_gem_import(struct drm_device *dev,
>  	npages = size / PAGE_SIZE;
>  
>  	msm_obj = to_msm_bo(obj);
> -	msm_gem_lock(obj);
>  	msm_obj->sgt = sgt;
>  	msm_obj->pages = kvmalloc_objs(struct page *, npages);
>  	if (!msm_obj->pages) {
> -		msm_gem_unlock(obj);
>  		ret = -ENOMEM;
>  		goto fail;
>  	}
>  
>  	ret = drm_prime_sg_to_page_array(sgt, msm_obj->pages, npages);
>  	if (ret) {
> -		msm_gem_unlock(obj);
>  		goto fail;

[Severity: High]
This is a pre-existing issue, but does this error path lead to a NULL
pointer dereference and a double-free?

When msm_gem_import() jumps to fail, it calls drm_gem_object_put(obj).
At this point, obj->import_attach has not yet been set (it is assigned
later by the caller drm_gem_prime_import_dev() on success).

During cleanup, msm_gem_free_object() checks drm_gem_is_imported(obj).
Since import_attach is NULL, this evaluates to false, causing the cleanup
to fall through to the native path:

msm_gem_free_object() ->
    put_pages(obj)

In put_pages(), the code incorrectly frees the exporter-owned sgt (leading
to a double-free) and then calls drm_gem_put_pages().

Finally, drm_gem_put_pages() unconditionally accesses
file_inode(obj->filp)->i_mapping. Because obj->filp is NULL for imported
objects, this causes a NULL pointer dereference.

>  	}
>  
> -	msm_gem_unlock(obj);
> -
>  	drm_gem_lru_move_tail(&priv->lru.pinned, obj);
>  
>  	mutex_lock(&priv->obj_lock);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260717204841.26120-1-robin.clark@oss.qualcomm.com?part=9

  reply	other threads:[~2026-07-17 21:02 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-07-17 20:48 [PATCH v6 00/16] drm/msm: A couple lazy-vm fixes Rob Clark
2026-07-17 20:48 ` [PATCH v6 01/16] drm/msm: Fix barriers accessing ctx vm Rob Clark
2026-07-17 21:07   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 02/16] drm/msm: Rework queuelock Rob Clark
2026-07-17 21:00   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 03/16] drm/msm: Synchronize VM creation on ctxlock Rob Clark
2026-07-17 21:05   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 04/16] drm/msm: Synchronize set_sysprof " Rob Clark
2026-07-17 20:48 ` [PATCH v6 05/16] drm/msm: Move nr_cmds initialization Rob Clark
2026-07-17 21:06   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 06/16] drm/msm: Remove redundant SIZE_MAX check Rob Clark
2026-07-17 20:58   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 07/16] drm/msm/a6xx: Access VM directly in submit path Rob Clark
2026-07-17 21:03   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 08/16] drm/msm: Add helper to check for per-process pgtables VM Rob Clark
2026-07-17 20:48 ` [PATCH v6 09/16] drm/msm/gem: Remove useless locking in GEM import Rob Clark
2026-07-17 21:02   ` sashiko-bot [this message]
2026-07-17 20:48 ` [PATCH v6 10/16] drm/msm/gem: Extract bookkeeping init helper Rob Clark
2026-07-17 21:22   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 11/16] drm/msm/gem: Set resv before exposing obj Rob Clark
2026-07-17 21:06   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 12/16] drm/msm/gem: Validate lazy VM in GEM_NEW Rob Clark
2026-07-17 21:05   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 13/16] drm/msm: Allow lazy VM creation to fail Rob Clark
2026-07-17 20:48 ` [PATCH v6 14/16] drm/msm: Don't fallback to shared VM for VM_BIND Rob Clark
2026-07-17 21:07   ` sashiko-bot
2026-07-17 20:48 ` [PATCH v6 15/16] drm/msm: Fix per-process-pgtables check Rob Clark
2026-07-17 20:48 ` [PATCH v6 16/16] drm/msm: Fixup invalid overflow check Rob Clark

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=20260717210228.278A71F00A3A@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=robin.clark@oss.qualcomm.com \
    --cc=sashiko-reviews@lists.linux.dev \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.