Merge tag 'drm-xe-next-fixes-2026-08-13' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-next

Couple drm-ras fixes, a probe failure clean-up fix, a GT freq
boundaries fixes for BMG/CRI and a Media engines/slice fix.

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

From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patch.msgid.link/an4NGzsYN9MOTFII@intel.com

5 files changed, 32 insertions, 15 deletions

diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 7007b6113760..71ce153737ab 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -1141,7 +1141,18 @@ int xe_device_probe(struct xe_device *xe)
 	if (err)
 		goto err_unregister_display;
 
-	return devm_add_action_or_reset(xe->drm.dev, xe_device_sanitize, xe);
+	/*
+	 * Process and log any errors detected by hardware. Possible results can
+	 * include declaring the device as wedged, which must be done only after
+	 * xe_device_wedged_fini() is registered.
+	 */
+	xe_ras_process_errors(xe);
+
+	err = devm_add_action_or_reset(xe->drm.dev, xe_device_sanitize, xe);
+	if (err)
+		goto err_unregister_display;
+
+	return 0;
 
 err_unregister_display:
 	xe_display_unregister(xe);
diff --git a/drivers/gpu/drm/xe/xe_drm_ras.c b/drivers/gpu/drm/xe/xe_drm_ras.c
index 7937d8ba0ed9..984fa67b9015 100644
--- a/drivers/gpu/drm/xe/xe_drm_ras.c
+++ b/drivers/gpu/drm/xe/xe_drm_ras.c
@@ -199,6 +199,9 @@ int xe_drm_ras_init(struct xe_device *xe)
 	struct drm_ras_node *node;
 	int err;
 
+	if (!xe->info.has_drm_ras)
+		return 0;
+
 	node = drmm_kcalloc(&xe->drm, DRM_XE_RAS_ERR_SEV_MAX, sizeof(*node), GFP_KERNEL);
 	if (!node)
 		return -ENOMEM;
diff --git a/drivers/gpu/drm/xe/xe_gt_mcr.c b/drivers/gpu/drm/xe/xe_gt_mcr.c
index a97b236dab7c..1949bf7d370a 100644
--- a/drivers/gpu/drm/xe/xe_gt_mcr.c
+++ b/drivers/gpu/drm/xe/xe_gt_mcr.c
@@ -443,11 +443,16 @@ static void init_steering_dss(struct xe_gt *gt)
 
 static void init_steering_oaddrm(struct xe_gt *gt)
 {
+	u64 hwe_mask = XE_HW_ENGINE_VCS0 | XE_HW_ENGINE_VECS0;
+
+	/* TODO: Add 'VD per SCMI' and 'VE per SCMI' values into 'struct xe_media_desc' */
+	if (MEDIA_VERx100(gt_to_xe(gt)) >= 3500)
+		hwe_mask |= XE_HW_ENGINE_VCS1 | XE_HW_ENGINE_VECS1;
 	/*
 	 * First instance is only terminated if the entire first media slice
-	 * is absent (i.e., no VCS0 or VECS0).
+	 * is absent (i.e., no engines in hwe_mask).
 	 */
-	if (gt->info.engine_mask & (XE_HW_ENGINE_VCS0 | XE_HW_ENGINE_VECS0))
+	if (gt->info.engine_mask & hwe_mask)
 		gt->steering[OADDRM].group_target = 0;
 	else
 		gt->steering[OADDRM].group_target = 1;
diff --git a/drivers/gpu/drm/xe/xe_guc_pc.c b/drivers/gpu/drm/xe/xe_guc_pc.c
index 59f2fa79ad42..7cf8f4858598 100644
--- a/drivers/gpu/drm/xe/xe_guc_pc.c
+++ b/drivers/gpu/drm/xe/xe_guc_pc.c
@@ -911,6 +911,7 @@ static bool pc_needs_min_freq_change(struct xe_guc_pc *pc)
 static int pc_adjust_freq_bounds(struct xe_guc_pc *pc)
 {
 	int ret;
+	u32 min_freq;
 
 	lockdep_assert_held(&pc->freq_lock);
 
@@ -933,8 +934,14 @@ static int pc_adjust_freq_bounds(struct xe_guc_pc *pc)
 	 * Same thing happens for Server platforms where min is listed as
 	 * RPMax
 	 */
-	if (pc_get_min_freq(pc) > pc->rp0_freq)
+	min_freq = pc_get_min_freq(pc);
+	if (min_freq > pc->rp0_freq) {
 		ret = pc_set_min_freq(pc, pc->rp0_freq);
+		if (ret)
+			goto out;
+
+		min_freq = pc->rp0_freq;
+	}
 
 	/*
 	 * Setting GT RP min frequency to 1.2GHz by default for
@@ -947,8 +954,8 @@ static int pc_adjust_freq_bounds(struct xe_guc_pc *pc)
 	 * we aren't expecting high power output across board
 	 *
 	 */
-	if (pc_needs_min_freq_change(pc))
-		ret = pc_set_min_freq(pc, max(BMG_MIN_FREQ, pc_get_min_freq(pc)));
+	if (pc_needs_min_freq_change(pc) && min_freq < BMG_MIN_FREQ)
+		ret = pc_set_min_freq(pc, BMG_MIN_FREQ);
 
 out:
 	return ret;
diff --git a/drivers/gpu/drm/xe/xe_ras.c b/drivers/gpu/drm/xe/xe_ras.c
index a31e06b8aa67..d98ff9453f60 100644
--- a/drivers/gpu/drm/xe/xe_ras.c
+++ b/drivers/gpu/drm/xe/xe_ras.c
@@ -681,9 +681,6 @@ void xe_ras_init(struct xe_device *xe)
 {
 	int ret;
 
-	if (!xe->info.has_drm_ras)
-		return;
-
 	xe_drm_ras_init(xe);
 
 	if (!xe->info.has_sysctrl)
@@ -692,12 +689,6 @@ void xe_ras_init(struct xe_device *xe)
 	if (IS_ENABLED(CONFIG_PCIEAER))
 		ras_usp_aer_init(xe);
 
-	/*
-	 * During probe, process and log any errors detected by firmware while the driver was not
-	 * loaded. Critical errors such as Punit and CSC are reported through Pcode init failure,
-	 * causing the driver to enter survivability mode.
-	 */
-	xe_ras_process_errors(xe);
 	ret = devm_device_add_group(xe->drm.dev, &gpu_health_group);
 	if (ret)
 		xe_err(xe, "Failed to create GPU health sysfs, err=%d\n", ret);