Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Imre Deak <imre.deak@intel.com>
To: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org, Lyude Paul <lyude@redhat.com>
Subject: [PATCH 4/7] drm/dp_mst: Fix down request message timeout handling
Date: Tue,  3 Dec 2024 18:02:20 +0200	[thread overview]
Message-ID: <20241203160223.2926014-5-imre.deak@intel.com> (raw)
In-Reply-To: <20241203160223.2926014-1-imre.deak@intel.com>

If receiving a reply for an MST down request message times out, the
thread receiving the reply in drm_dp_mst_handle_down_rep() could try to
dereference the drm_dp_sideband_msg_tx txmsg request message after the
thread waiting for the reply - calling drm_dp_mst_wait_tx_reply() - has
timed out and freed txmsg, hence leading to a use-after-free in
drm_dp_mst_handle_down_rep().

Prevent the above by holding the drm_dp_mst_topology_mgr::qlock in
drm_dp_mst_handle_down_rep() for the whole duration txmsg is looked up
from the request list and dereferenced.

Cc: Lyude Paul <lyude@redhat.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
---
 drivers/gpu/drm/display/drm_dp_mst_topology.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/display/drm_dp_mst_topology.c b/drivers/gpu/drm/display/drm_dp_mst_topology.c
index 6ec8680998d5a..95742d82510a7 100644
--- a/drivers/gpu/drm/display/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/display/drm_dp_mst_topology.c
@@ -3984,9 +3984,9 @@ static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
 
 	/* find the message */
 	mutex_lock(&mgr->qlock);
+
 	txmsg = list_first_entry_or_null(&mgr->tx_msg_downq,
 					 struct drm_dp_sideband_msg_tx, next);
-	mutex_unlock(&mgr->qlock);
 
 	/* Were we actually expecting a response, and from this mstb? */
 	if (!txmsg || txmsg->dst != mstb) {
@@ -3995,6 +3995,9 @@ static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
 		hdr = &msg->initial_hdr;
 		drm_dbg_kms(mgr->dev, "Got MST reply with no msg %p %d %d %02x %02x\n",
 			    mstb, hdr->seqno, hdr->lct, hdr->rad[0], msg->msg[0]);
+
+		mutex_unlock(&mgr->qlock);
+
 		goto out_clear_reply;
 	}
 
@@ -4013,9 +4016,9 @@ static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
 			    txmsg->reply.u.nak.nak_data);
 	}
 
-	mutex_lock(&mgr->qlock);
 	txmsg->state = DRM_DP_SIDEBAND_TX_RX;
 	list_del(&txmsg->next);
+
 	mutex_unlock(&mgr->qlock);
 
 	wake_up_all(&mgr->tx_waitq);
-- 
2.44.2


  parent reply	other threads:[~2024-12-03 16:02 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-03 16:02 [PATCH 0/7] drm/dp_mst: Fix a few side-band message handling issues Imre Deak
2024-12-03 16:02 ` [PATCH 1/7] drm/dp_mst: Fix resetting msg rx state after topology removal Imre Deak
2024-12-03 16:02 ` [PATCH 2/7] drm/dp_mst: Verify request type in the corresponding down message reply Imre Deak
2024-12-03 16:02 ` [PATCH 3/7] drm/dp_mst: Simplify error path in drm_dp_mst_handle_down_rep() Imre Deak
2024-12-03 16:02 ` Imre Deak [this message]
2024-12-03 17:46   ` [PATCH v2 4/7] drm/dp_mst: Fix down request message timeout handling Imre Deak
2024-12-03 16:02 ` [PATCH 5/7] drm/dp_mst: Ensure mst_primary pointer is valid in drm_dp_mst_handle_up_req() Imre Deak
2024-12-04 13:20   ` [PATCH v2 " Imre Deak
2024-12-03 16:02 ` [PATCH 6/7] drm/dp_mst: Reset message rx state after OOM " Imre Deak
2024-12-03 16:02 ` [PATCH 7/7] drm/dp_mst: Use reset_msg_rx_state() instead of open coding it Imre Deak
2024-12-03 16:34 ` ✗ Fi.CI.CHECKPATCH: warning for drm/dp_mst: Fix a few side-band message handling issues Patchwork
2024-12-03 16:34 ` ✗ Fi.CI.SPARSE: " Patchwork
2024-12-03 16:50 ` ✗ i915.CI.BAT: failure " Patchwork
2024-12-03 17:07   ` Imre Deak
2024-12-04  6:09     ` Ravali, JupallyX
2024-12-03 18:03 ` [PATCH 0/7] " Lyude Paul
2024-12-03 18:28 ` ✗ Fi.CI.CHECKPATCH: warning for drm/dp_mst: Fix a few side-band message handling issues (rev2) Patchwork
2024-12-03 18:28 ` ✗ Fi.CI.SPARSE: " Patchwork
2024-12-03 18:44 ` ✗ i915.CI.BAT: failure " Patchwork
2024-12-04  6:05 ` ✓ i915.CI.BAT: success " Patchwork
2024-12-04  7:17 ` ✗ i915.CI.Full: failure " Patchwork
2024-12-04 14:07 ` ✗ Fi.CI.CHECKPATCH: warning for drm/dp_mst: Fix a few side-band message handling issues (rev3) Patchwork
2024-12-04 14:07 ` ✗ Fi.CI.SPARSE: " Patchwork
2024-12-04 14:22 ` ✓ i915.CI.BAT: success " Patchwork
2024-12-04 15:30 ` ✓ i915.CI.Full: " Patchwork
2024-12-05 15:16   ` Imre Deak

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=20241203160223.2926014-5-imre.deak@intel.com \
    --to=imre.deak@intel.com \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=lyude@redhat.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