Merge tag 'drm-xe-fixes-2026-05-21' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

- SRIOV related fixes (Wajdeczko, Mohanram)
- Fix leak and double-free (Lin)
- Multi-cast register fixes (Gustavo)
- Multi-queue fix (Niranjana)

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

From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patch.msgid.link/ag9rR5VwCdkA0lzI@intel.com

11 files changed, 69 insertions, 34 deletions

diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h
index 9c88ca3ce768..353fe0bd49bf 100644
--- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h
+++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h
@@ -152,10 +152,11 @@
 
 #define XEHPG_INSTDONE_GEOM_SVGUNIT		XE_REG_MCR(0x666c)
 
-#define CACHE_MODE_1				XE_REG(0x7004, XE_REG_OPTION_MASKED)
+#define CACHE_MODE_1				XE_REG_MCR(0x7004, XE_REG_OPTION_MASKED)
 #define   MSAA_OPTIMIZATION_REDUC_DISABLE	REG_BIT(11)
 
 #define COMMON_SLICE_CHICKEN1			XE_REG(0x7010, XE_REG_OPTION_MASKED)
+#define XEHP_COMMON_SLICE_CHICKEN1		XE_REG_MCR(0x7010, XE_REG_OPTION_MASKED)
 #define   DISABLE_BOTTOM_CLIP_RECTANGLE_TEST	REG_BIT(14)
 
 #define HIZ_CHICKEN					XE_REG(0x7018, XE_REG_OPTION_MASKED)
@@ -178,6 +179,7 @@
 #define XEHPG_SC_INSTDONE_EXTRA2		XE_REG_MCR(0x7108)
 
 #define COMMON_SLICE_CHICKEN4			XE_REG(0x7300, XE_REG_OPTION_MASKED)
+#define XEHP_COMMON_SLICE_CHICKEN4		XE_REG_MCR(0x7300, XE_REG_OPTION_MASKED)
 #define   SBE_PUSH_CONSTANT_BEHIND_FIX_ENABLE	REG_BIT(12)
 #define   DISABLE_TDC_LOAD_BALANCING_CALC	REG_BIT(6)
 #define   HW_FILTERING				REG_BIT(5)
diff --git a/drivers/gpu/drm/xe/xe_gsc.c b/drivers/gpu/drm/xe/xe_gsc.c
index 0d13e357fb43..aab59dc647fb 100644
--- a/drivers/gpu/drm/xe/xe_gsc.c
+++ b/drivers/gpu/drm/xe/xe_gsc.c
@@ -482,8 +482,7 @@ int xe_gsc_init_post_hwconfig(struct xe_gsc *gsc)
 				 EXEC_QUEUE_FLAG_PERMANENT, 0);
 	if (IS_ERR(q)) {
 		xe_gt_err(gt, "Failed to create queue for GSC submission\n");
-		err = PTR_ERR(q);
-		goto out_bo;
+		return PTR_ERR(q);
 	}
 
 	wq = alloc_ordered_workqueue("gsc-ordered-wq", 0);
@@ -506,8 +505,6 @@ int xe_gsc_init_post_hwconfig(struct xe_gsc *gsc)
 
 out_q:
 	xe_exec_queue_put(q);
-out_bo:
-	xe_bo_unpin_map_no_vm(bo);
 	return err;
 }
 
diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.c
index 7d532bded02a..a85ba4435378 100644
--- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.c
+++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.c
@@ -114,8 +114,10 @@ int xe_gt_sriov_pf_monitor_process_guc2pf(struct xe_gt *gt, const u32 *msg, u32
  * VFs with no events are not printed.
  *
  * This function can only be called on PF.
+ *
+ * Return: always 0
  */
-void xe_gt_sriov_pf_monitor_print_events(struct xe_gt *gt, struct drm_printer *p)
+int xe_gt_sriov_pf_monitor_print_events(struct xe_gt *gt, struct drm_printer *p)
 {
 	unsigned int n, total_vfs = xe_gt_sriov_pf_get_totalvfs(gt);
 	const struct xe_gt_sriov_monitor *data;
@@ -144,4 +146,6 @@ void xe_gt_sriov_pf_monitor_print_events(struct xe_gt *gt, struct drm_printer *p
 #undef __format
 #undef __value
 	}
+
+	return 0;
 }
diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.h b/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.h
index 7ca9351a271b..0b8f088d3a16 100644
--- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.h
+++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_monitor.h
@@ -13,7 +13,7 @@ struct drm_printer;
 struct xe_gt;
 
 void xe_gt_sriov_pf_monitor_flr(struct xe_gt *gt, u32 vfid);
-void xe_gt_sriov_pf_monitor_print_events(struct xe_gt *gt, struct drm_printer *p);
+int xe_gt_sriov_pf_monitor_print_events(struct xe_gt *gt, struct drm_printer *p);
 
 #ifdef CONFIG_PCI_IOV
 int xe_gt_sriov_pf_monitor_process_guc2pf(struct xe_gt *gt, const u32 *msg, u32 len);
diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_vf.c b/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
index 8989c8e1be95..0cd9d77f3351 100644
--- a/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
+++ b/drivers/gpu/drm/xe/xe_gt_sriov_vf.c
@@ -1137,13 +1137,15 @@ void xe_gt_sriov_vf_write32(struct xe_gt *gt, struct xe_reg reg, u32 val)
 }
 
 /**
- * xe_gt_sriov_vf_print_config - Print VF self config.
+ * xe_gt_sriov_vf_print_config() - Print VF self config.
  * @gt: the &xe_gt
  * @p: the &drm_printer
  *
  * This function is for VF use only.
+ *
+ * Return: always 0.
  */
-void xe_gt_sriov_vf_print_config(struct xe_gt *gt, struct drm_printer *p)
+int xe_gt_sriov_vf_print_config(struct xe_gt *gt, struct drm_printer *p)
 {
 	struct xe_gt_sriov_vf_selfconfig *config = &gt->sriov.vf.self_config;
 	struct xe_device *xe = gt_to_xe(gt);
@@ -1170,16 +1172,20 @@ void xe_gt_sriov_vf_print_config(struct xe_gt *gt, struct drm_printer *p)
 
 	drm_printf(p, "GuC contexts:\t%u\n", config->num_ctxs);
 	drm_printf(p, "GuC doorbells:\t%u\n", config->num_dbs);
+
+	return 0;
 }
 
 /**
- * xe_gt_sriov_vf_print_runtime - Print VF's runtime regs received from PF.
+ * xe_gt_sriov_vf_print_runtime() - Print VF's runtime regs received from PF.
  * @gt: the &xe_gt
  * @p: the &drm_printer
  *
  * This function is for VF use only.
+ *
+ * Return: always 0.
  */
-void xe_gt_sriov_vf_print_runtime(struct xe_gt *gt, struct drm_printer *p)
+int xe_gt_sriov_vf_print_runtime(struct xe_gt *gt, struct drm_printer *p)
 {
 	struct vf_runtime_reg *vf_regs = gt->sriov.vf.runtime.regs;
 	unsigned int size = gt->sriov.vf.runtime.num_regs;
@@ -1188,16 +1194,20 @@ void xe_gt_sriov_vf_print_runtime(struct xe_gt *gt, struct drm_printer *p)
 
 	for (; size--; vf_regs++)
 		drm_printf(p, "%#x = %#x\n", vf_regs->offset, vf_regs->value);
+
+	return 0;
 }
 
 /**
- * xe_gt_sriov_vf_print_version - Print VF ABI versions.
+ * xe_gt_sriov_vf_print_version() - Print VF ABI versions.
  * @gt: the &xe_gt
  * @p: the &drm_printer
  *
  * This function is for VF use only.
+ *
+ * Return: always 0.
  */
-void xe_gt_sriov_vf_print_version(struct xe_gt *gt, struct drm_printer *p)
+int xe_gt_sriov_vf_print_version(struct xe_gt *gt, struct drm_printer *p)
 {
 	struct xe_device *xe = gt_to_xe(gt);
 	struct xe_uc_fw_version *guc_version = &gt->sriov.vf.guc_version;
@@ -1227,6 +1237,8 @@ void xe_gt_sriov_vf_print_version(struct xe_gt *gt, struct drm_printer *p)
 		   GUC_RELAY_VERSION_LATEST_MAJOR, GUC_RELAY_VERSION_LATEST_MINOR);
 	drm_printf(p, "\thandshake:\t%u.%u\n",
 		   pf_version->major, pf_version->minor);
+
+	return 0;
 }
 
 static bool vf_post_migration_shutdown(struct xe_gt *gt)
diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_vf.h b/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
index a6f7127521a5..79878f21b1da 100644
--- a/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
+++ b/drivers/gpu/drm/xe/xe_gt_sriov_vf.h
@@ -35,9 +35,9 @@ bool xe_gt_sriov_vf_sched_groups_enabled(struct xe_gt *gt);
 u32 xe_gt_sriov_vf_read32(struct xe_gt *gt, struct xe_reg reg);
 void xe_gt_sriov_vf_write32(struct xe_gt *gt, struct xe_reg reg, u32 val);
 
-void xe_gt_sriov_vf_print_config(struct xe_gt *gt, struct drm_printer *p);
-void xe_gt_sriov_vf_print_runtime(struct xe_gt *gt, struct drm_printer *p);
-void xe_gt_sriov_vf_print_version(struct xe_gt *gt, struct drm_printer *p);
+int xe_gt_sriov_vf_print_config(struct xe_gt *gt, struct drm_printer *p);
+int xe_gt_sriov_vf_print_runtime(struct xe_gt *gt, struct drm_printer *p);
+int xe_gt_sriov_vf_print_version(struct xe_gt *gt, struct drm_printer *p);
 
 int xe_gt_sriov_vf_wait_valid_ggtt(struct xe_gt *gt);
 int xe_vf_migration_fixups_complete_count(struct xe_gt *gt);
diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
index 10556156eaad..912182dc7704 100644
--- a/drivers/gpu/drm/xe/xe_guc_submit.c
+++ b/drivers/gpu/drm/xe/xe_guc_submit.c
@@ -1673,6 +1673,14 @@ static void guc_exec_queue_fini(struct xe_exec_queue *q)
 	struct xe_guc_exec_queue *ge = q->guc;
 	struct xe_guc *guc = exec_queue_to_guc(q);
 
+	if (xe_exec_queue_is_multi_queue_secondary(q)) {
+		struct xe_exec_queue_group *group = q->multi_queue.group;
+
+		mutex_lock(&group->list_lock);
+		list_del(&q->multi_queue.link);
+		mutex_unlock(&group->list_lock);
+	}
+
 	release_guc_id(guc, q);
 	xe_sched_entity_fini(&ge->entity);
 	xe_sched_fini(&ge->sched);
@@ -1694,14 +1702,6 @@ static void __guc_exec_queue_destroy_async(struct work_struct *w)
 	guard(xe_pm_runtime)(guc_to_xe(guc));
 	trace_xe_exec_queue_destroy(q);
 
-	if (xe_exec_queue_is_multi_queue_secondary(q)) {
-		struct xe_exec_queue_group *group = q->multi_queue.group;
-
-		mutex_lock(&group->list_lock);
-		list_del(&q->multi_queue.link);
-		mutex_unlock(&group->list_lock);
-	}
-
 	/* Confirm no work left behind accessing device structures */
 	cancel_delayed_work_sync(&ge->sched.base.work_tdr);
 
diff --git a/drivers/gpu/drm/xe/xe_memirq.c b/drivers/gpu/drm/xe/xe_memirq.c
index 811e07136efb..579af47edc61 100644
--- a/drivers/gpu/drm/xe/xe_memirq.c
+++ b/drivers/gpu/drm/xe/xe_memirq.c
@@ -427,13 +427,25 @@ static bool memirq_received(struct xe_memirq *memirq, struct iosys_map *vector,
 	return __memirq_received(memirq, vector, offset, name, true);
 }
 
+static void memirq_assume_received(struct xe_memirq *memirq, const char *source,
+				   u16 offset, const char *status)
+{
+	memirq_debug(memirq, "ASSUME %s %s(%u)\n", source, status, offset);
+}
+
 static void memirq_dispatch_engine(struct xe_memirq *memirq, struct iosys_map *status,
 				   struct xe_hw_engine *hwe)
 {
 	memirq_debug(memirq, "STATUS %s %*ph\n", hwe->name, 16, status->vaddr);
 
-	if (memirq_received(memirq, status, ilog2(GT_MI_USER_INTERRUPT), hwe->name))
-		xe_hw_engine_handle_irq(hwe, GT_MI_USER_INTERRUPT);
+	/*
+	 * The programming note says to assume that GT_MI_USER_INTERRUPT is always
+	 * set. Check and clear related status byte just for a debug.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_XE_DEBUG_MEMIRQ) &&
+	    !memirq_received(memirq, status, ilog2(GT_MI_USER_INTERRUPT), hwe->name))
+		memirq_assume_received(memirq, hwe->name, ilog2(GT_MI_USER_INTERRUPT), "USER");
+	xe_hw_engine_handle_irq(hwe, GT_MI_USER_INTERRUPT);
 }
 
 static void memirq_dispatch_guc(struct xe_memirq *memirq, struct iosys_map *status,
@@ -443,8 +455,14 @@ static void memirq_dispatch_guc(struct xe_memirq *memirq, struct iosys_map *stat
 
 	memirq_debug(memirq, "STATUS %s %*ph\n", name, 16, status->vaddr);
 
-	if (memirq_received(memirq, status, ilog2(GUC_INTR_GUC2HOST), name))
-		xe_guc_irq_handler(guc, GUC_INTR_GUC2HOST);
+	/*
+	 * The programming note says to assume that GUC_INTR_GUC2HOST is always
+	 * set. Check and clear related status byte just for a debug.
+	 */
+	if (IS_ENABLED(CONFIG_DRM_XE_DEBUG_MEMIRQ) &&
+	    !memirq_received(memirq, status, ilog2(GUC_INTR_GUC2HOST), name))
+		memirq_assume_received(memirq, name, ilog2(GUC_INTR_GUC2HOST), "GUC2HOST");
+	xe_guc_irq_handler(guc, GUC_INTR_GUC2HOST);
 
 	/*
 	 * This is a software interrupt that must be cleared after it's consumed
diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c
index 6337e671c97a..d908f4e03906 100644
--- a/drivers/gpu/drm/xe/xe_oa.c
+++ b/drivers/gpu/drm/xe/xe_oa.c
@@ -2032,8 +2032,10 @@ int xe_oa_stream_open_ioctl(struct drm_device *dev, u64 data, struct drm_file *f
 		if (XE_IOCTL_DBG(oa->xe, !param.exec_q))
 			return -ENOENT;
 
-		if (XE_IOCTL_DBG(oa->xe, param.exec_q->width > 1))
-			return -EOPNOTSUPP;
+		if (XE_IOCTL_DBG(oa->xe, param.exec_q->width > 1)) {
+			ret = -EOPNOTSUPP;
+			goto err_exec_q;
+		}
 	}
 
 	/*
diff --git a/drivers/gpu/drm/xe/xe_tuning.c b/drivers/gpu/drm/xe/xe_tuning.c
index 0b78ec2bc6a4..fcb6698abc6e 100644
--- a/drivers/gpu/drm/xe/xe_tuning.c
+++ b/drivers/gpu/drm/xe/xe_tuning.c
@@ -129,7 +129,7 @@ static const struct xe_rtp_entry_sr engine_tunings[] = {
 static const struct xe_rtp_entry_sr lrc_tunings[] = {
 	{ XE_RTP_NAME("Tuning: Windower HW Filtering"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(3000, 3599), ENGINE_CLASS(RENDER)),
-	  XE_RTP_ACTIONS(SET(COMMON_SLICE_CHICKEN4, HW_FILTERING))
+	  XE_RTP_ACTIONS(SET(XEHP_COMMON_SLICE_CHICKEN4, HW_FILTERING))
 	},
 
 	/* DG2 */
diff --git a/drivers/gpu/drm/xe/xe_wa.c b/drivers/gpu/drm/xe/xe_wa.c
index 4b1cbced06be..33df43d0bede 100644
--- a/drivers/gpu/drm/xe/xe_wa.c
+++ b/drivers/gpu/drm/xe/xe_wa.c
@@ -651,7 +651,7 @@ static const struct xe_rtp_entry_sr lrc_was[] = {
 	},
 	{ XE_RTP_NAME("18033852989"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(2001, 2004), ENGINE_CLASS(RENDER)),
-	  XE_RTP_ACTIONS(SET(COMMON_SLICE_CHICKEN1, DISABLE_BOTTOM_CLIP_RECTANGLE_TEST))
+	  XE_RTP_ACTIONS(SET(XEHP_COMMON_SLICE_CHICKEN1, DISABLE_BOTTOM_CLIP_RECTANGLE_TEST))
 	},
 	{ XE_RTP_NAME("15016589081"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(2001, 2004), ENGINE_CLASS(RENDER)),
@@ -754,7 +754,7 @@ static const struct xe_rtp_entry_sr lrc_was[] = {
 	},
 	{ XE_RTP_NAME("22021007897"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(2001, 2002), ENGINE_CLASS(RENDER)),
-	  XE_RTP_ACTIONS(SET(COMMON_SLICE_CHICKEN4, SBE_PUSH_CONSTANT_BEHIND_FIX_ENABLE))
+	  XE_RTP_ACTIONS(SET(XEHP_COMMON_SLICE_CHICKEN4, SBE_PUSH_CONSTANT_BEHIND_FIX_ENABLE))
 	},
 
 	/* Xe3_LPG */
@@ -770,7 +770,7 @@ static const struct xe_rtp_entry_sr lrc_was[] = {
 	},
 	{ XE_RTP_NAME("22021007897"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(3000, 3005), ENGINE_CLASS(RENDER)),
-	  XE_RTP_ACTIONS(SET(COMMON_SLICE_CHICKEN4, SBE_PUSH_CONSTANT_BEHIND_FIX_ENABLE))
+	  XE_RTP_ACTIONS(SET(XEHP_COMMON_SLICE_CHICKEN4, SBE_PUSH_CONSTANT_BEHIND_FIX_ENABLE))
 	},
 	{ XE_RTP_NAME("14024681466"),
 	  XE_RTP_RULES(GRAPHICS_VERSION_RANGE(3000, 3005), ENGINE_CLASS(RENDER)),