dri-devel Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Maíra Canal" <mcanal@igalia.com>
Cc: dri-devel@lists.freedesktop.org, ojeda@kernel.org, gary@garyguo.net
Subject: Re: [PATCH 1/6] drm/shmem: Add a drm_gem_shmem_sync() helper
Date: Thu, 19 Nov 2026 19:55:20 +0000	[thread overview]
Message-ID: <sashiko-outbox-223986@kernel.org> (raw)
In-Reply-To: <20261119-pan-drop-dma-buf-hooks-v1-1-6c6f1629e7f9@igalia.com>

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [High] drm/shmem: unlocked sgt traversal in drm_gem_shmem_sync()
  causes UAF
- [Medium] drm/shmem: drm_gem_shmem_sync() populates pages on purged or
  unbacked BOs
--

commit 9fdf5667225c80a2c62b9b65dbf8c174b87f854f
Author: Faith Ekstrand <faith.ekstrand@collabora.com>

drm/shmem: Add a drm_gem_shmem_sync() helper

This commit adds a drm_gem_shmem_sync() helper to flush or invalidate CPU
caches for a range of a shmem GEM object by walking its scatter-gather table
and calling dma_sync_single_for_device() and dma_sync_single_for_cpu().

> diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c
> index d654a22a30ded..72933ed283731 100644
> --- a/drivers/gpu/drm/drm_gem_shmem_helper.c
> +++ b/drivers/gpu/drm/drm_gem_shmem_helper.c
> @@ -790,6 +790,99 @@ int drm_gem_shmem_mmap(struct drm_gem_shmem_object *shmem, struct vm_area_struct
>  }
>  EXPORT_SYMBOL_GPL(drm_gem_shmem_mmap);
>  
> +/**
> + * drm_gem_shmem_sync - Sync CPU-mapped data
> + * @shmem: shmem GEM object
> + * @offset: Offset into the GEM object
> + * @size: Size of the area to sync
> + * @type: Type of synchronization
> + *
> + * Returns:
> + * 0 on success or a negative error code on failure.
> + */
> +int drm_gem_shmem_sync(struct drm_gem_shmem_object *shmem, size_t offset,
> +		       size_t size, enum drm_gem_shmem_sync_type type)
> +{
> +	const struct drm_device *dev = shmem->base.dev;
> +	struct sg_table *sgt;
> +	struct scatterlist *sgl;
> +	unsigned int count;
> +
> +	/* Make sure the range is in bounds. */
> +	if (offset + size < offset || offset + size > shmem->base.size)
> +		return -EINVAL;
> +
> +	/* Disallow CPU-cache maintenance on imported buffers. */
> +	if (drm_gem_is_imported(&shmem->base))
> +		return -EINVAL;
> +
> +	switch (type) {
> +	case DRM_GEM_SHMEM_SYNC_CPU_CACHE_FLUSH:
> +	case DRM_GEM_SHMEM_SYNC_CPU_CACHE_FLUSH_AND_INVALIDATE:
> +		break;
> +
> +	default:
> +		return -EINVAL;
> +	}
> +
> +	/* Don't bother if it's WC-mapped */
> +	if (shmem->map_wc)
> +		return 0;
> +
> +	/* Nothing to do if the size is zero. */
> +	if (size == 0)
> +		return 0;
> +
> +	sgt = drm_gem_shmem_get_pages_sgt(shmem);
> +	if (IS_ERR(sgt))
> +		return PTR_ERR(sgt);

[Severity: Medium]
Should drm_gem_shmem_sync() populate backing pages and a scatter-gather
table when shmem->pages is NULL or the object has been purged?

Calling drm_gem_shmem_get_pages_sgt() on an unbacked object allocates and
DMA-maps pages when there is no CPU-cached data to sync.  If the object was
already purged (shmem->madv < 0), drm_gem_shmem_get_pages_sgt() allocates a
new shmem->sgt and pages (or takes another reference on truncated pages
kept alive by an open VMA), and because drm_gem_shmem_is_purgeable()
requires shmem->madv > 0, the newly populated pages and sgt can never be
purged again until the GEM object is freed.

> +
> +	for_each_sgtable_dma_sg(sgt, sgl, count) {
> +		if (size == 0)
> +			break;
> +
> +		dma_addr_t paddr = sg_dma_address(sgl);
> +		size_t len = sg_dma_len(sgl);
> +
> +		if (len <= offset) {
> +			offset -= len;
> +			continue;
> +		}
> +
> +		paddr += offset;
> +		len -= offset;
> +		len = min_t(size_t, len, size);
> +		size -= len;
> +		offset = 0;
[ ... ]
> +		dma_sync_single_for_device(dev->dev, paddr, len, DMA_TO_DEVICE);
> +		if (type == DRM_GEM_SHMEM_SYNC_CPU_CACHE_FLUSH_AND_INVALIDATE)
> +			dma_sync_single_for_cpu(dev->dev, paddr, len, DMA_FROM_DEVICE);
> +	}

[Severity: High]
Can sgt be unmapped and freed while drm_gem_shmem_sync() is walking it?

drm_gem_shmem_get_pages_sgt() unlocks shmem->base.resv before returning
sgt, and does not increment shmem->pages_pin_count.  If the buffer is
marked purgeable (shmem->madv > 0, pages_pin_count == 0), a concurrent
shrinker can purge it while the loop runs:

CPU0                                  CPU1
drm_gem_shmem_sync()
  drm_gem_shmem_get_pages_sgt()
    dma_resv_lock_interruptible(...)
    drm_gem_shmem_get_pages_sgt_locked()
    dma_resv_unlock(...)
                                      ...
                                        drm_gem_shmem_purge_locked()
                                          dma_unmap_sgtable(..., shmem->sgt, ...)
                                          sg_free_table(shmem->sgt)
                                          kfree(shmem->sgt)
  for_each_sgtable_dma_sg(sgt, ...)
    sg_dma_address(sgl)  <-- use-after-free

Holding shmem->base.resv across drm_gem_shmem_get_pages_sgt_locked() and
the scatterlist walk would prevent concurrent purging or sgt teardown.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261119-pan-drop-dma-buf-hooks-v1-0-6c6f1629e7f9@igalia.com?part=1

  reply	other threads:[~2026-11-19 19:55 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-19 19:38 [PATCH 0/6] drm/panfrost,panthor: Refuse to export WB_MMAP BOs Maíra Canal
2026-11-19 19:38 ` [PATCH 1/6] drm/shmem: Add a drm_gem_shmem_sync() helper Maíra Canal
2026-11-19 19:55   ` sashiko-bot [this message]
2026-11-19 19:38 ` [PATCH 2/6] drm/panfrost: Refuse to export WB_MMAP BOs Maíra Canal
2026-11-19 19:38 ` [PATCH 3/6] drm/panfrost: Drop custom dma_buf_ops Maíra Canal
2026-11-19 19:38 ` [PATCH 4/6] drm/panfrost: Use drm_gem_shmem_sync() Maíra Canal
2026-11-19 19:38 ` [PATCH 5/6] drm/panthor: Refuse to export WB_MMAP BOs Maíra Canal
2026-11-19 19:38 ` [PATCH 6/6] drm/panthor: Drop custom dma_buf_ops Maíra Canal

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=sashiko-outbox-223986@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=gary@garyguo.net \
    --cc=mcanal@igalia.com \
    --cc=ojeda@kernel.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox