Merge tag 'drm-xe-fixes-2025-07-17' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

Driver Changes:
- SR-IOV fixes for GT reset and TLB invalidation
- Fix memory copy direction during migration
- Fix alignment check on migration
- Fix MOCS and page fault init order to correctly
  account for topology

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

From: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://lore.kernel.org/r/6jworkgupwstm4v7aohbuzod3dyz4u7pyfhshr5ifgf2xisgj3@cm5em5yupjiu

Author: airlied

drivers/gpu/drm/xe/xe_gt.c

@@ -417,6 +417,8 @@ int xe_gt_init_early(struct xe_gt *gt)
 	if (err)
 		return err;
 
+	xe_mocs_init_early(gt);
+
 	return 0;
 }
 
@@ -630,17 +632,15 @@ int xe_gt_init(struct xe_gt *gt)
 	if (err)
 		return err;
 
-	err = xe_gt_pagefault_init(gt);
+	err = xe_gt_sysfs_init(gt);
 	if (err)
 		return err;
 
-	xe_mocs_init_early(gt);
-
-	err = xe_gt_sysfs_init(gt);
+	err = gt_fw_domain_init(gt);
 	if (err)
 		return err;
 
-	err = gt_fw_domain_init(gt);
+	err = xe_gt_pagefault_init(gt);
 	if (err)
 		return err;
 
@@ -839,6 +839,9 @@ static int gt_reset(struct xe_gt *gt)
 		goto err_out;
 	}
 
+	if (IS_SRIOV_PF(gt_to_xe(gt)))
+		xe_gt_sriov_pf_stop_prepare(gt);
+
 	xe_uc_gucrc_disable(&gt->uc);
 	xe_uc_stop_prepare(&gt->uc);
 	xe_gt_pagefault_reset(gt);

drivers/gpu/drm/xe/xe_gt_sriov_pf.c

@@ -172,6 +172,25 @@ void xe_gt_sriov_pf_sanitize_hw(struct xe_gt *gt, unsigned int vfid)
 	pf_clear_vf_scratch_regs(gt, vfid);
 }
 
+static void pf_cancel_restart(struct xe_gt *gt)
+{
+	xe_gt_assert(gt, IS_SRIOV_PF(gt_to_xe(gt)));
+
+	if (cancel_work_sync(&gt->sriov.pf.workers.restart))
+		xe_gt_sriov_dbg_verbose(gt, "pending restart canceled!\n");
+}
+
+/**
+ * xe_gt_sriov_pf_stop_prepare() - Prepare to stop SR-IOV support.
+ * @gt: the &xe_gt
+ *
+ * This function can only be called on the PF.
+ */
+void xe_gt_sriov_pf_stop_prepare(struct xe_gt *gt)
+{
+	pf_cancel_restart(gt);
+}
+
 static void pf_restart(struct xe_gt *gt)
 {
 	struct xe_device *xe = gt_to_xe(gt);

drivers/gpu/drm/xe/xe_gt_sriov_pf.h

@@ -13,6 +13,7 @@ int xe_gt_sriov_pf_init_early(struct xe_gt *gt);
 int xe_gt_sriov_pf_init(struct xe_gt *gt);
 void xe_gt_sriov_pf_init_hw(struct xe_gt *gt);
 void xe_gt_sriov_pf_sanitize_hw(struct xe_gt *gt, unsigned int vfid);
+void xe_gt_sriov_pf_stop_prepare(struct xe_gt *gt);
 void xe_gt_sriov_pf_restart(struct xe_gt *gt);
 #else
 static inline int xe_gt_sriov_pf_init_early(struct xe_gt *gt)
@@ -29,6 +30,10 @@ static inline void xe_gt_sriov_pf_init_hw(struct xe_gt *gt)
 {
 }
 
+static inline void xe_gt_sriov_pf_stop_prepare(struct xe_gt *gt)
+{
+}
+
 static inline void xe_gt_sriov_pf_restart(struct xe_gt *gt)
 {
 }

drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c

@@ -2364,6 +2364,21 @@ int xe_gt_sriov_pf_config_restore(struct xe_gt *gt, unsigned int vfid,
 	return err;
 }
 
+static int pf_push_self_config(struct xe_gt *gt)
+{
+	int err;
+
+	err = pf_push_full_vf_config(gt, PFID);
+	if (err) {
+		xe_gt_sriov_err(gt, "Failed to push self configuration (%pe)\n",
+				ERR_PTR(err));
+		return err;
+	}
+
+	xe_gt_sriov_dbg_verbose(gt, "self configuration completed\n");
+	return 0;
+}
+
 static void fini_config(void *arg)
 {
 	struct xe_gt *gt = arg;
@@ -2387,9 +2402,17 @@ static void fini_config(void *arg)
 int xe_gt_sriov_pf_config_init(struct xe_gt *gt)
 {
 	struct xe_device *xe = gt_to_xe(gt);
+	int err;
 
 	xe_gt_assert(gt, IS_SRIOV_PF(xe));
 
+	mutex_lock(xe_gt_sriov_pf_master_mutex(gt));
+	err = pf_push_self_config(gt);
+	mutex_unlock(xe_gt_sriov_pf_master_mutex(gt));
+
+	if (err)
+		return err;
+
 	return devm_add_action_or_reset(xe->drm.dev, fini_config, gt);
 }
 
@@ -2407,6 +2430,10 @@ void xe_gt_sriov_pf_config_restart(struct xe_gt *gt)
 	unsigned int n, total_vfs = xe_sriov_pf_get_totalvfs(gt_to_xe(gt));
 	unsigned int fail = 0, skip = 0;
 
+	mutex_lock(xe_gt_sriov_pf_master_mutex(gt));
+	pf_push_self_config(gt);
+	mutex_unlock(xe_gt_sriov_pf_master_mutex(gt));
+
 	for (n = 1; n <= total_vfs; n++) {
 		if (xe_gt_sriov_pf_config_is_empty(gt, n))
 			skip++;

drivers/gpu/drm/xe/xe_migrate.c

@@ -1817,8 +1817,8 @@ int xe_migrate_access_memory(struct xe_migrate *m, struct xe_bo *bo,
 	xe_bo_assert_held(bo);
 
 	/* Use bounce buffer for small access and unaligned access */
-	if (len & XE_CACHELINE_MASK ||
-	    ((uintptr_t)buf | offset) & XE_CACHELINE_MASK) {
+	if (!IS_ALIGNED(len, XE_CACHELINE_BYTES) ||
+	    !IS_ALIGNED((unsigned long)buf + offset, XE_CACHELINE_BYTES)) {
 		int buf_offset = 0;
 
 		/*
@@ -1848,7 +1848,7 @@ int xe_migrate_access_memory(struct xe_migrate *m, struct xe_bo *bo,
 				err = xe_migrate_access_memory(m, bo,
 							       offset & ~XE_CACHELINE_MASK,
 							       (void *)ptr,
-							       sizeof(bounce), 0);
+							       sizeof(bounce), write);
 				if (err)
 					return err;
 			} else {

drivers/gpu/drm/xe/xe_ring_ops.c

@@ -110,13 +110,14 @@ static int emit_bb_start(u64 batch_addr, u32 ppgtt_flag, u32 *dw, int i)
 	return i;
 }
 
-static int emit_flush_invalidate(u32 *dw, int i)
+static int emit_flush_invalidate(u32 addr, u32 val, u32 *dw, int i)
 {
 	dw[i++] = MI_FLUSH_DW | MI_INVALIDATE_TLB | MI_FLUSH_DW_OP_STOREDW |
-		  MI_FLUSH_IMM_DW | MI_FLUSH_DW_STORE_INDEX;
-	dw[i++] = LRC_PPHWSP_FLUSH_INVAL_SCRATCH_ADDR;
-	dw[i++] = 0;
+		  MI_FLUSH_IMM_DW;
+
+	dw[i++] = addr | MI_FLUSH_DW_USE_GTT;
 	dw[i++] = 0;
+	dw[i++] = val;
 
 	return i;
 }
@@ -397,23 +398,20 @@ static void __emit_job_gen12_render_compute(struct xe_sched_job *job,
 static void emit_migration_job_gen12(struct xe_sched_job *job,
 				     struct xe_lrc *lrc, u32 seqno)
 {
+	u32 saddr = xe_lrc_start_seqno_ggtt_addr(lrc);
 	u32 dw[MAX_JOB_SIZE_DW], i = 0;
 
 	i = emit_copy_timestamp(lrc, dw, i);
 
-	i = emit_store_imm_ggtt(xe_lrc_start_seqno_ggtt_addr(lrc),
-				seqno, dw, i);
+	i = emit_store_imm_ggtt(saddr, seqno, dw, i);
 
 	dw[i++] = MI_ARB_ON_OFF | MI_ARB_DISABLE; /* Enabled again below */
 
 	i = emit_bb_start(job->ptrs[0].batch_addr, BIT(8), dw, i);
 
-	if (!IS_SRIOV_VF(gt_to_xe(job->q->gt))) {
-		/* XXX: Do we need this? Leaving for now. */
-		dw[i++] = preparser_disable(true);
-		i = emit_flush_invalidate(dw, i);
-		dw[i++] = preparser_disable(false);
-	}
+	dw[i++] = preparser_disable(true);
+	i = emit_flush_invalidate(saddr, seqno, dw, i);
+	dw[i++] = preparser_disable(false);
 
 	i = emit_bb_start(job->ptrs[1].batch_addr, BIT(8), dw, i);