public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
From: jbmoore <jbmoore61@gmail.com>
To: alexander.deucher@amd.com, christian.koenig@amd.com
Cc: "John B. Moore" <jbmoore61@gmail.com>, stable@vger.kernel.org
Subject: [PATCH 3/4] drm/amdgpu/gfx9: replace BUG_ON with WARN_ON_ONCE for KIQ 64-bit fence flag
Date: Sun, 26 Apr 2026 16:52:52 -0500	[thread overview]
Message-ID: <20260426215256.50722-4-jbmoore@nooks.dev> (raw)
In-Reply-To: <20260426215256.50722-1-jbmoore@nooks.dev>

From: "John B. Moore" <jbmoore61@gmail.com>

gfx_v9_0_ring_emit_fence_kiq() contains a BUG_ON() that fires when
the AMDGPU_FENCE_FLAG_64BIT flag is passed.  The KIQ (Kernel
Interface Queue) ring only allocates 32-bit writeback buffer
addresses for fence sequence numbers.  A 64-bit fence write would
overflow the allocated writeback slot, potentially corrupting
adjacent kernel memory.

Replace BUG_ON() with WARN_ON_ONCE() and mask off the unsupported
flag.  This prevents the kernel panic while still logging the
unexpected condition and falling back to a safe 32-bit fence write.

This is separated from the main gfx9 BUG_ON conversion patch
because it addresses a different security concern (potential buffer
overflow in kernel-managed writeback memory) rather than the address
alignment assertions in the ring emission paths.

Found by a custom amdgpu DRM ioctl fuzzer.

Fixes: b1023571479020e9 ("drm/amdgpu: implement GFX 9.0 support (v2)")
Signed-off-by: John B. Moore <jbmoore61@gmail.com>
Cc: stable@vger.kernel.org
---
 drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 47e81c33d..fb2a0f1af 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -5679,7 +5679,8 @@ static void gfx_v9_0_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr,
 	struct amdgpu_device *adev = ring->adev;
 
 	/* we only allocate 32bit for each seq wb address */
-	BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT);
+	if (WARN_ON_ONCE(flags & AMDGPU_FENCE_FLAG_64BIT))
+		flags &= ~AMDGPU_FENCE_FLAG_64BIT;
 
 	/* write fence seq to the "addr" */
 	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
-- 
2.43.0


  parent reply	other threads:[~2026-04-26 21:53 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20260426215256.50722-1-jbmoore@nooks.dev>
2026-04-26 21:52 ` [PATCH 1/4] drm/amdgpu/sdma4: replace BUG_ON with WARN_ON_ONCE in fence emission jbmoore
2026-04-26 21:52 ` [PATCH 2/4] drm/amdgpu/gfx9: replace BUG_ON/BUG with WARN_ON_ONCE in ring emission jbmoore
2026-04-26 21:52 ` jbmoore [this message]
2026-04-26 21:52 ` [PATCH 4/4] drm/amdgpu/vcn: prevent silent fence drop on 64-bit flag mismatch jbmoore

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=20260426215256.50722-4-jbmoore@nooks.dev \
    --to=jbmoore61@gmail.com \
    --cc=alexander.deucher@amd.com \
    --cc=christian.koenig@amd.com \
    --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