Merge tag 'drm-xe-fixes-2024-10-10' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

Driver Changes:
- Fix error checking with xa_store() (Matthe Auld)
- Fix missing freq restore on GSC load error (Vinay)
- Fix wedged_mode file permission (Matt Roper)
- Fix use-after-free in ct communication (Matthew Auld)

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/jri65tmv3bjbhqhxs5smv45nazssxzhtwphojem4uufwtjuliy@gsdhlh6kzsdy

4 files changed, 33 insertions, 26 deletions

diff --git a/drivers/gpu/drm/xe/xe_debugfs.c b/drivers/gpu/drm/xe/xe_debugfs.c
index 668615c6b172..fe4319eb13fd 100644
--- a/drivers/gpu/drm/xe/xe_debugfs.c
+++ b/drivers/gpu/drm/xe/xe_debugfs.c
@@ -187,7 +187,7 @@ void xe_debugfs_register(struct xe_device *xe)
 	debugfs_create_file("forcewake_all", 0400, root, xe,
 			    &forcewake_all_fops);
 
-	debugfs_create_file("wedged_mode", 0400, root, xe,
+	debugfs_create_file("wedged_mode", 0600, root, xe,
 			    &wedged_mode_fops);
 
 	for (mem_type = XE_PL_VRAM0; mem_type <= XE_PL_VRAM1; ++mem_type) {
diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
index f0dc2bf24c7b..ea65cf59372c 100644
--- a/drivers/gpu/drm/xe/xe_gt.c
+++ b/drivers/gpu/drm/xe/xe_gt.c
@@ -874,7 +874,9 @@ int xe_gt_sanitize_freq(struct xe_gt *gt)
 	int ret = 0;
 
 	if ((!xe_uc_fw_is_available(&gt->uc.gsc.fw) ||
-	     xe_uc_fw_is_loaded(&gt->uc.gsc.fw)) && XE_WA(gt, 22019338487))
+	     xe_uc_fw_is_loaded(&gt->uc.gsc.fw) ||
+	     xe_uc_fw_is_in_error_state(&gt->uc.gsc.fw)) &&
+	    XE_WA(gt, 22019338487))
 		ret = xe_guc_pc_restore_stashed_freq(&gt->uc.guc.pc);
 
 	return ret;
diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c
index f24dd5223926..d16eb9ab49fb 100644
--- a/drivers/gpu/drm/xe/xe_guc_ct.c
+++ b/drivers/gpu/drm/xe/xe_guc_ct.c
@@ -667,16 +667,12 @@ static int __guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action,
 		num_g2h = 1;
 
 		if (g2h_fence_needs_alloc(g2h_fence)) {
-			void *ptr;
-
 			g2h_fence->seqno = next_ct_seqno(ct, true);
-			ptr = xa_store(&ct->fence_lookup,
-				       g2h_fence->seqno,
-				       g2h_fence, GFP_ATOMIC);
-			if (IS_ERR(ptr)) {
-				ret = PTR_ERR(ptr);
+			ret = xa_err(xa_store(&ct->fence_lookup,
+					      g2h_fence->seqno, g2h_fence,
+					      GFP_ATOMIC));
+			if (ret)
 				goto out;
-			}
 		}
 
 		seqno = g2h_fence->seqno;
@@ -879,14 +875,11 @@ retry:
 retry_same_fence:
 	ret = guc_ct_send(ct, action, len, 0, 0, &g2h_fence);
 	if (unlikely(ret == -ENOMEM)) {
-		void *ptr;
-
 		/* Retry allocation /w GFP_KERNEL */
-		ptr = xa_store(&ct->fence_lookup,
-			       g2h_fence.seqno,
-			       &g2h_fence, GFP_KERNEL);
-		if (IS_ERR(ptr))
-			return PTR_ERR(ptr);
+		ret = xa_err(xa_store(&ct->fence_lookup, g2h_fence.seqno,
+				      &g2h_fence, GFP_KERNEL));
+		if (ret)
+			return ret;
 
 		goto retry_same_fence;
 	} else if (unlikely(ret)) {
@@ -903,16 +896,26 @@ retry_same_fence:
 	}
 
 	ret = wait_event_timeout(ct->g2h_fence_wq, g2h_fence.done, HZ);
+
+	/*
+	 * Ensure we serialize with completion side to prevent UAF with fence going out of scope on
+	 * the stack, since we have no clue if it will fire after the timeout before we can erase
+	 * from the xa. Also we have some dependent loads and stores below for which we need the
+	 * correct ordering, and we lack the needed barriers.
+	 */
+	mutex_lock(&ct->lock);
 	if (!ret) {
-		xe_gt_err(gt, "Timed out wait for G2H, fence %u, action %04x",
-			  g2h_fence.seqno, action[0]);
+		xe_gt_err(gt, "Timed out wait for G2H, fence %u, action %04x, done %s",
+			  g2h_fence.seqno, action[0], str_yes_no(g2h_fence.done));
 		xa_erase_irq(&ct->fence_lookup, g2h_fence.seqno);
+		mutex_unlock(&ct->lock);
 		return -ETIME;
 	}
 
 	if (g2h_fence.retry) {
 		xe_gt_dbg(gt, "H2G action %#x retrying: reason %#x\n",
 			  action[0], g2h_fence.reason);
+		mutex_unlock(&ct->lock);
 		goto retry;
 	}
 	if (g2h_fence.fail) {
@@ -921,7 +924,12 @@ retry_same_fence:
 		ret = -EIO;
 	}
 
-	return ret > 0 ? response_buffer ? g2h_fence.response_len : g2h_fence.response_data : ret;
+	if (ret > 0)
+		ret = response_buffer ? g2h_fence.response_len : g2h_fence.response_data;
+
+	mutex_unlock(&ct->lock);
+
+	return ret;
 }
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
index bf3836614aca..63495007f336 100644
--- a/drivers/gpu/drm/xe/xe_guc_submit.c
+++ b/drivers/gpu/drm/xe/xe_guc_submit.c
@@ -320,7 +320,6 @@ static void __release_guc_id(struct xe_guc *guc, struct xe_exec_queue *q, u32 xa
 static int alloc_guc_id(struct xe_guc *guc, struct xe_exec_queue *q)
 {
 	int ret;
-	void *ptr;
 	int i;
 
 	/*
@@ -340,12 +339,10 @@ static int alloc_guc_id(struct xe_guc *guc, struct xe_exec_queue *q)
 	q->guc->id = ret;
 
 	for (i = 0; i < q->width; ++i) {
-		ptr = xa_store(&guc->submission_state.exec_queue_lookup,
-			       q->guc->id + i, q, GFP_NOWAIT);
-		if (IS_ERR(ptr)) {
-			ret = PTR_ERR(ptr);
+		ret = xa_err(xa_store(&guc->submission_state.exec_queue_lookup,
+				      q->guc->id + i, q, GFP_NOWAIT));
+		if (ret)
 			goto err_release;
-		}
 	}
 
 	return 0;