- Release guc ids before cancelling work (Tejas)

- Fix new warnings around userptr (Thomas)
 - Temporaritly disable D3Cold on BMG (Rodrigo)
 - Retry and wait longer for GuC PC to start (Rodrigo)
 - Remove redundant check in xe_vm_create_ioctl (Xin)
 -----BEGIN PGP SIGNATURE-----
 
 iQEzBAABCAAdFiEEbSBwaO7dZQkcLOKj+mJfZA7rE8oFAmfTB3EACgkQ+mJfZA7r
 E8pkoAgAui5ItatSInI77C15ajI1Mn4kVVu95Pus+rz8pj4NDpWPVwo9dIySjp+h
 BIwDFLzpi9aEAYKT81JmXp6KDtqQDv+pyBzw/hVdMT47hP5nOJax1j6uAa7gAaXB
 KxvMmHP2p4N917n6oUg8tOM0rPrY/GjXRkPETjD3snJ1gP0kqj4xDOWupkU51fec
 Mk+PI+o67fnDFTMZziHNS0BS3MdXGi6uEldNAzm2s6xgoIGwIv2Gl2Sm8uTNxEqY
 olo03ssQGO3NKh78J/laoNEendcKG+aarwt3h7WQhAgx0P1asjn1F2pFX0nNKpyJ
 Y1SRMu+g6lzozo3I1tCB3BaIJ/1XEQ==
 =1WN7
 -----END PGP SIGNATURE-----

Merge tag 'drm-xe-fixes-2025-03-13' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes

- Release guc ids before cancelling work (Tejas)
- Fix new warnings around userptr (Thomas)
- Temporaritly disable D3Cold on BMG (Rodrigo)
- Retry and wait longer for GuC PC to start (Rodrigo)
- Remove redundant check in xe_vm_create_ioctl (Xin)

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

From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/Z9MJWeIlZPuvXZ_G@intel.com

drivers/gpu/drm/xe/xe_guc_pc.c

@@ -6,6 +6,7 @@
 #include "xe_guc_pc.h"
 
 #include <linux/delay.h>
+#include <linux/ktime.h>
 
 #include <drm/drm_managed.h>
 #include <generated/xe_wa_oob.h>
@@ -19,6 +20,7 @@
 #include "xe_gt.h"
 #include "xe_gt_idle.h"
 #include "xe_gt_printk.h"
+#include "xe_gt_throttle.h"
 #include "xe_gt_types.h"
 #include "xe_guc.h"
 #include "xe_guc_ct.h"
@@ -49,6 +51,9 @@
 #define LNL_MERT_FREQ_CAP	800
 #define BMG_MERT_FREQ_CAP	2133
 
+#define SLPC_RESET_TIMEOUT_MS 5 /* roughly 5ms, but no need for precision */
+#define SLPC_RESET_EXTENDED_TIMEOUT_MS 1000 /* To be used only at pc_start */
+
 /**
  * DOC: GuC Power Conservation (PC)
  *
@@ -113,9 +118,10 @@ static struct iosys_map *pc_to_maps(struct xe_guc_pc *pc)
 	 FIELD_PREP(HOST2GUC_PC_SLPC_REQUEST_MSG_1_EVENT_ARGC, count))
 
 static int wait_for_pc_state(struct xe_guc_pc *pc,
-			     enum slpc_global_state state)
+			     enum slpc_global_state state,
+			     int timeout_ms)
 {
-	int timeout_us = 5000; /* rought 5ms, but no need for precision */
+	int timeout_us = 1000 * timeout_ms;
 	int slept, wait = 10;
 
 	xe_device_assert_mem_access(pc_to_xe(pc));
@@ -164,7 +170,8 @@ static int pc_action_query_task_state(struct xe_guc_pc *pc)
 	};
 	int ret;
 
-	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING))
+	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING,
+			      SLPC_RESET_TIMEOUT_MS))
 		return -EAGAIN;
 
 	/* Blocking here to ensure the results are ready before reading them */
@@ -187,7 +194,8 @@ static int pc_action_set_param(struct xe_guc_pc *pc, u8 id, u32 value)
 	};
 	int ret;
 
-	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING))
+	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING,
+			      SLPC_RESET_TIMEOUT_MS))
 		return -EAGAIN;
 
 	ret = xe_guc_ct_send(ct, action, ARRAY_SIZE(action), 0, 0);
@@ -208,7 +216,8 @@ static int pc_action_unset_param(struct xe_guc_pc *pc, u8 id)
 	struct xe_guc_ct *ct = &pc_to_guc(pc)->ct;
 	int ret;
 
-	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING))
+	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING,
+			      SLPC_RESET_TIMEOUT_MS))
 		return -EAGAIN;
 
 	ret = xe_guc_ct_send(ct, action, ARRAY_SIZE(action), 0, 0);
@@ -440,6 +449,15 @@ u32 xe_guc_pc_get_act_freq(struct xe_guc_pc *pc)
 	return freq;
 }
 
+static u32 get_cur_freq(struct xe_gt *gt)
+{
+	u32 freq;
+
+	freq = xe_mmio_read32(&gt->mmio, RPNSWREQ);
+	freq = REG_FIELD_GET(REQ_RATIO_MASK, freq);
+	return decode_freq(freq);
+}
+
 /**
  * xe_guc_pc_get_cur_freq - Get Current requested frequency
  * @pc: The GuC PC
@@ -463,10 +481,7 @@ int xe_guc_pc_get_cur_freq(struct xe_guc_pc *pc, u32 *freq)
 		return -ETIMEDOUT;
 	}
 
-	*freq = xe_mmio_read32(&gt->mmio, RPNSWREQ);
-
-	*freq = REG_FIELD_GET(REQ_RATIO_MASK, *freq);
-	*freq = decode_freq(*freq);
+	*freq = get_cur_freq(gt);
 
 	xe_force_wake_put(gt_to_fw(gt), fw_ref);
 	return 0;
@@ -1002,6 +1017,7 @@ int xe_guc_pc_start(struct xe_guc_pc *pc)
 	struct xe_gt *gt = pc_to_gt(pc);
 	u32 size = PAGE_ALIGN(sizeof(struct slpc_shared_data));
 	unsigned int fw_ref;
+	ktime_t earlier;
 	int ret;
 
 	xe_gt_assert(gt, xe_device_uc_enabled(xe));
@@ -1026,14 +1042,25 @@ int xe_guc_pc_start(struct xe_guc_pc *pc)
 	memset(pc->bo->vmap.vaddr, 0, size);
 	slpc_shared_data_write(pc, header.size, size);
 
+	earlier = ktime_get();
 	ret = pc_action_reset(pc);
 	if (ret)
 		goto out;
 
-	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING)) {
-		xe_gt_err(gt, "GuC PC Start failed\n");
-		ret = -EIO;
-		goto out;
+	if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING,
+			      SLPC_RESET_TIMEOUT_MS)) {
+		xe_gt_warn(gt, "GuC PC start taking longer than normal [freq = %dMHz (req = %dMHz), perf_limit_reasons = 0x%08X]\n",
+			   xe_guc_pc_get_act_freq(pc), get_cur_freq(gt),
+			   xe_gt_throttle_get_limit_reasons(gt));
+
+		if (wait_for_pc_state(pc, SLPC_GLOBAL_STATE_RUNNING,
+				      SLPC_RESET_EXTENDED_TIMEOUT_MS)) {
+			xe_gt_err(gt, "GuC PC Start failed: Dynamic GT frequency control and GT sleep states are now disabled.\n");
+			goto out;
+		}
+
+		xe_gt_warn(gt, "GuC PC excessive start time: %lldms",
+			   ktime_ms_delta(ktime_get(), earlier));
 	}
 
 	ret = pc_init_freqs(pc);

drivers/gpu/drm/xe/xe_guc_submit.c

@@ -1246,11 +1246,11 @@ static void __guc_exec_queue_fini_async(struct work_struct *w)
 	xe_pm_runtime_get(guc_to_xe(guc));
 	trace_xe_exec_queue_destroy(q);
 
+	release_guc_id(guc, q);
 	if (xe_exec_queue_is_lr(q))
 		cancel_work_sync(&ge->lr_tdr);
 	/* Confirm no work left behind accessing device structures */
 	cancel_delayed_work_sync(&ge->sched.base.work_tdr);
-	release_guc_id(guc, q);
 	xe_sched_entity_fini(&ge->entity);
 	xe_sched_fini(&ge->sched);
 

drivers/gpu/drm/xe/xe_hmm.c

@@ -138,13 +138,17 @@ static int xe_build_sg(struct xe_device *xe, struct hmm_range *range,
 		i += size;
 
 		if (unlikely(j == st->nents - 1)) {
+			xe_assert(xe, i >= npages);
 			if (i > npages)
 				size -= (i - npages);
+
 			sg_mark_end(sgl);
+		} else {
+			xe_assert(xe, i < npages);
 		}
+
 		sg_set_page(sgl, page, size << PAGE_SHIFT, 0);
 	}
-	xe_assert(xe, i == npages);
 
 	return dma_map_sgtable(dev, st, write ? DMA_BIDIRECTIONAL : DMA_TO_DEVICE,
 			       DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_NO_KERNEL_MAPPING);

drivers/gpu/drm/xe/xe_pm.c

@@ -267,6 +267,15 @@ int xe_pm_init_early(struct xe_device *xe)
 }
 ALLOW_ERROR_INJECTION(xe_pm_init_early, ERRNO); /* See xe_pci_probe() */
 
+static u32 vram_threshold_value(struct xe_device *xe)
+{
+	/* FIXME: D3Cold temporarily disabled by default on BMG */
+	if (xe->info.platform == XE_BATTLEMAGE)
+		return 0;
+
+	return DEFAULT_VRAM_THRESHOLD;
+}
+
 /**
  * xe_pm_init - Initialize Xe Power Management
  * @xe: xe device instance
@@ -277,6 +286,7 @@ ALLOW_ERROR_INJECTION(xe_pm_init_early, ERRNO); /* See xe_pci_probe() */
  */
 int xe_pm_init(struct xe_device *xe)
 {
+	u32 vram_threshold;
 	int err;
 
 	/* For now suspend/resume is only allowed with GuC */
@@ -290,7 +300,8 @@ int xe_pm_init(struct xe_device *xe)
 		if (err)
 			return err;
 
-		err = xe_pm_set_vram_threshold(xe, DEFAULT_VRAM_THRESHOLD);
+		vram_threshold = vram_threshold_value(xe);
+		err = xe_pm_set_vram_threshold(xe, vram_threshold);
 		if (err)
 			return err;
 	}

drivers/gpu/drm/xe/xe_vm.c

@@ -1809,9 +1809,6 @@ int xe_vm_create_ioctl(struct drm_device *dev, void *data,
 			 args->flags & DRM_XE_VM_CREATE_FLAG_FAULT_MODE))
 		return -EINVAL;
 
-	if (XE_IOCTL_DBG(xe, args->extensions))
-		return -EINVAL;
-
 	if (args->flags & DRM_XE_VM_CREATE_FLAG_SCRATCH_PAGE)
 		flags |= XE_VM_FLAG_SCRATCH_PAGE;
 	if (args->flags & DRM_XE_VM_CREATE_FLAG_LR_MODE)