Merge tag 'drm-intel-next-2017-09-29' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

2nd batch of v4.15 features:

- lib/scatterlist updates, use for userptr allocations (Tvrtko)
- Fixed point wrapper cleanup (Mahesh)
- Gen9+ transition watermarks, watermark optimization and fixes (Mahesh)
- Display IPC (Isochronous Priority Control) support (Mahesh)
- GEM workaround fixes (Oscar)
- GVT: PCI config sanitize series (Changbin)
- GVT: Workload submission error handling series (Fred)
- PSR fixes and refactoring (Rodrigo)
- HWSP based optimizations (Chris)
- Private PAT management (Zhi)
- IRQ handling fixes and refactoring (Ville)
- Module parameter refactoring and variable name clash fix (Michal)
- Execlist refactoring, incomplete request unwinding on reset (Chris)
- GuC scheduling improvements (Michal)
- OA updates (Lionel)
- Coffeelake out of alpha support (Rodrigo)
- seqno fixes (Chris)
- Execlist refactoring (Mika)
- DP and DP MST cleanups (Dhinakaran)
- Cannonlake slice/sublice config (Ben)
- Numerous fixes all around (Everyone)

* tag 'drm-intel-next-2017-09-29' of git://anongit.freedesktop.org/drm/drm-intel: (168 commits)
  drm/i915: Update DRIVER_DATE to 20170929
  drm/i915: Use memset64() to prefill the GTT page
  drm/i915: Also discard second CRC on gen8+ platforms.
  drm/i915/psr: Set frames before SU entry for psr2
  drm/dp: Add defines for latency in sink
  drm/i915: Allow optimized platform checks
  drm/i915: Avoid using dev_priv->info.gen directly.
  i915: Use %pS printk format for direct addresses
  drm/i915/execlists: Notify context-out for lost requests
  drm/i915/cnl: Add support slice/subslice/eu configs
  drm/i915: Compact device info access by a small re-ordering
  drm/i915: Add IS_PLATFORM macro
  drm/i915/selftests: Try to recover from a wedged GPU during reset tests
  drm/i915/huc: Reorganize HuC authentication
  drm/i915: Fix default values of some modparams
  drm/i915: Extend I915_PARAMS_FOR_EACH with default member value
  drm/i915: Make I915_PARAMS_FOR_EACH macro more flexible
  drm/i915: Enable scanline read based on frame timestamps
  drm/i915/execlists: Microoptimise execlists_cancel_port_request()
  drm/i915: Don't rmw PIPESTAT enable bits
  ...

drivers/gpu/drm/i915/Kconfig

@@ -12,6 +12,7 @@ config DRM_I915
 	select DRM_PANEL
 	select DRM_MIPI_DSI
 	select RELAY
+	select IRQ_WORK
 	# i915 depends on ACPI_VIDEO when ACPI is enabled
 	# but for select to work, need to select ACPI_VIDEO's dependencies, ick
 	select BACKLIGHT_LCD_SUPPORT if ACPI

drivers/gpu/drm/i915/Makefile

@@ -139,7 +139,8 @@ i915-y += i915_perf.o \
 	  i915_oa_bxt.o \
 	  i915_oa_kblgt2.o \
 	  i915_oa_kblgt3.o \
-	  i915_oa_glk.o
+	  i915_oa_glk.o \
+	  i915_oa_cflgt2.o
 
 ifeq ($(CONFIG_DRM_I915_GVT),y)
 i915-y += intel_gvt.o

drivers/gpu/drm/i915/gvt/cfg_space.c

@@ -101,7 +101,7 @@ int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset,
 	if (WARN_ON(bytes > 4))
 		return -EINVAL;
 
-	if (WARN_ON(offset + bytes > INTEL_GVT_MAX_CFG_SPACE_SZ))
+	if (WARN_ON(offset + bytes > vgpu->gvt->device_info.cfg_space_size))
 		return -EINVAL;
 
 	memcpy(p_data, vgpu_cfg_space(vgpu) + offset, bytes);
@@ -110,13 +110,25 @@ int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset,
 
 static int map_aperture(struct intel_vgpu *vgpu, bool map)
 {
-	u64 first_gfn, first_mfn;
+	phys_addr_t aperture_pa = vgpu_aperture_pa_base(vgpu);
+	unsigned long aperture_sz = vgpu_aperture_sz(vgpu);
+	u64 first_gfn;
 	u64 val;
 	int ret;
 
 	if (map == vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_APERTURE].tracked)
 		return 0;
 
+	if (map) {
+		vgpu->gm.aperture_va = memremap(aperture_pa, aperture_sz,
+						MEMREMAP_WC);
+		if (!vgpu->gm.aperture_va)
+			return -ENOMEM;
+	} else {
+		memunmap(vgpu->gm.aperture_va);
+		vgpu->gm.aperture_va = NULL;
+	}
+
 	val = vgpu_cfg_space(vgpu)[PCI_BASE_ADDRESS_2];
 	if (val & PCI_BASE_ADDRESS_MEM_TYPE_64)
 		val = *(u64 *)(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_2);
@@ -124,14 +136,16 @@ static int map_aperture(struct intel_vgpu *vgpu, bool map)
 		val = *(u32 *)(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_2);
 
 	first_gfn = (val + vgpu_aperture_offset(vgpu)) >> PAGE_SHIFT;
-	first_mfn = vgpu_aperture_pa_base(vgpu) >> PAGE_SHIFT;
 
 	ret = intel_gvt_hypervisor_map_gfn_to_mfn(vgpu, first_gfn,
-						  first_mfn,
-						  vgpu_aperture_sz(vgpu) >>
-						  PAGE_SHIFT, map);
-	if (ret)
+						  aperture_pa >> PAGE_SHIFT,
+						  aperture_sz >> PAGE_SHIFT,
+						  map);
+	if (ret) {
+		memunmap(vgpu->gm.aperture_va);
+		vgpu->gm.aperture_va = NULL;
 		return ret;
+	}
 
 	vgpu->cfg_space.bar[INTEL_GVT_PCI_BAR_APERTURE].tracked = map;
 	return 0;
@@ -275,7 +289,7 @@ int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset,
 	if (WARN_ON(bytes > 4))
 		return -EINVAL;
 
-	if (WARN_ON(offset + bytes > INTEL_GVT_MAX_CFG_SPACE_SZ))
+	if (WARN_ON(offset + bytes > vgpu->gvt->device_info.cfg_space_size))
 		return -EINVAL;
 
 	/* First check if it's PCI_COMMAND */

drivers/gpu/drm/i915/gvt/cmd_parser.c

@@ -1576,11 +1576,11 @@ static int batch_buffer_needs_scan(struct parser_exec_state *s)
 	return 1;
 }
 
-static uint32_t find_bb_size(struct parser_exec_state *s)
+static int find_bb_size(struct parser_exec_state *s)
 {
 	unsigned long gma = 0;
 	struct cmd_info *info;
-	uint32_t bb_size = 0;
+	int bb_size = 0;
 	uint32_t cmd_len = 0;
 	bool met_bb_end = false;
 	struct intel_vgpu *vgpu = s->vgpu;
@@ -1637,6 +1637,8 @@ static int perform_bb_shadow(struct parser_exec_state *s)
 
 	/* get the size of the batch buffer */
 	bb_size = find_bb_size(s);
+	if (bb_size < 0)
+		return -EINVAL;
 
 	/* allocate shadow batch buffer */
 	entry_obj = kmalloc(sizeof(*entry_obj), GFP_KERNEL);
@@ -2603,7 +2605,8 @@ static int shadow_workload_ring_buffer(struct intel_vgpu_workload *workload)
 {
 	struct intel_vgpu *vgpu = workload->vgpu;
 	unsigned long gma_head, gma_tail, gma_top, guest_rb_size;
-	u32 *cs;
+	void *shadow_ring_buffer_va;
+	int ring_id = workload->ring_id;
 	int ret;
 
 	guest_rb_size = _RING_CTL_BUF_SIZE(workload->rb_ctl);
@@ -2616,34 +2619,42 @@ static int shadow_workload_ring_buffer(struct intel_vgpu_workload *workload)
 	gma_tail = workload->rb_start + workload->rb_tail;
 	gma_top = workload->rb_start + guest_rb_size;
 
-	/* allocate shadow ring buffer */
-	cs = intel_ring_begin(workload->req, workload->rb_len / sizeof(u32));
-	if (IS_ERR(cs))
-		return PTR_ERR(cs);
+	if (workload->rb_len > vgpu->reserve_ring_buffer_size[ring_id]) {
+		void *va = vgpu->reserve_ring_buffer_va[ring_id];
+		/* realloc the new ring buffer if needed */
+		vgpu->reserve_ring_buffer_va[ring_id] =
+			krealloc(va, workload->rb_len, GFP_KERNEL);
+		if (!vgpu->reserve_ring_buffer_va[ring_id]) {
+			gvt_vgpu_err("fail to alloc reserve ring buffer\n");
+			return -ENOMEM;
+		}
+		vgpu->reserve_ring_buffer_size[ring_id] = workload->rb_len;
+	}
+
+	shadow_ring_buffer_va = vgpu->reserve_ring_buffer_va[ring_id];
 
 	/* get shadow ring buffer va */
-	workload->shadow_ring_buffer_va = cs;
+	workload->shadow_ring_buffer_va = shadow_ring_buffer_va;
 
 	/* head > tail --> copy head <-> top */
 	if (gma_head > gma_tail) {
 		ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm,
-				      gma_head, gma_top, cs);
+				      gma_head, gma_top, shadow_ring_buffer_va);
 		if (ret < 0) {
 			gvt_vgpu_err("fail to copy guest ring buffer\n");
 			return ret;
 		}
-		cs += ret / sizeof(u32);
+		shadow_ring_buffer_va += ret;
 		gma_head = workload->rb_start;
 	}
 
 	/* copy head or start <-> tail */
-	ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm, gma_head, gma_tail, cs);
+	ret = copy_gma_to_hva(vgpu, vgpu->gtt.ggtt_mm, gma_head, gma_tail,
+				shadow_ring_buffer_va);
 	if (ret < 0) {
 		gvt_vgpu_err("fail to copy guest ring buffer\n");
 		return ret;
 	}
-	cs += ret / sizeof(u32);
-	intel_ring_advance(workload->req, cs);
 	return 0;
 }
 

drivers/gpu/drm/i915/gvt/execlist.c

@@ -368,7 +368,7 @@ static void free_workload(struct intel_vgpu_workload *workload)
 #define get_desc_from_elsp_dwords(ed, i) \
 	((struct execlist_ctx_descriptor_format *)&((ed)->data[i * 2]))
 
-static void prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
+static int prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 {
 	const int gmadr_bytes = workload->vgpu->gvt->device_info.gmadr_bytes_in_cmd;
 	struct intel_shadow_bb_entry *entry_obj;
@@ -379,7 +379,7 @@ static void prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 
 		vma = i915_gem_object_ggtt_pin(entry_obj->obj, NULL, 0, 4, 0);
 		if (IS_ERR(vma)) {
-			return;
+			return PTR_ERR(vma);
 		}
 
 		/* FIXME: we are not tracking our pinned VMA leaving it
@@ -392,6 +392,7 @@ static void prepare_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 		if (gmadr_bytes == 8)
 			entry_obj->bb_start_cmd_va[2] = 0;
 	}
+	return 0;
 }
 
 static int update_wa_ctx_2_shadow_ctx(struct intel_shadow_wa_ctx *wa_ctx)
@@ -420,7 +421,7 @@ static int update_wa_ctx_2_shadow_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 	return 0;
 }
 
-static void prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+static int prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 {
 	struct i915_vma *vma;
 	unsigned char *per_ctx_va =
@@ -428,12 +429,12 @@ static void prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 		wa_ctx->indirect_ctx.size;
 
 	if (wa_ctx->indirect_ctx.size == 0)
-		return;
+		return 0;
 
 	vma = i915_gem_object_ggtt_pin(wa_ctx->indirect_ctx.obj, NULL,
 				       0, CACHELINE_BYTES, 0);
 	if (IS_ERR(vma)) {
-		return;
+		return PTR_ERR(vma);
 	}
 
 	/* FIXME: we are not tracking our pinned VMA leaving it
@@ -447,26 +448,7 @@ static void prepare_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
 	memset(per_ctx_va, 0, CACHELINE_BYTES);
 
 	update_wa_ctx_2_shadow_ctx(wa_ctx);
-}
-
-static int prepare_execlist_workload(struct intel_vgpu_workload *workload)
-{
-	struct intel_vgpu *vgpu = workload->vgpu;
-	struct execlist_ctx_descriptor_format ctx[2];
-	int ring_id = workload->ring_id;
-
-	intel_vgpu_pin_mm(workload->shadow_mm);
-	intel_vgpu_sync_oos_pages(workload->vgpu);
-	intel_vgpu_flush_post_shadow(workload->vgpu);
-	prepare_shadow_batch_buffer(workload);
-	prepare_shadow_wa_ctx(&workload->wa_ctx);
-	if (!workload->emulate_schedule_in)
-		return 0;
-
-	ctx[0] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 1);
-	ctx[1] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 0);
-
-	return emulate_execlist_schedule_in(&vgpu->execlist[ring_id], ctx);
+	return 0;
 }
 
 static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload)
@@ -489,13 +471,62 @@ static void release_shadow_batch_buffer(struct intel_vgpu_workload *workload)
 	}
 }
 
-static void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+static int prepare_execlist_workload(struct intel_vgpu_workload *workload)
 {
-	if (!wa_ctx->indirect_ctx.obj)
-		return;
+	struct intel_vgpu *vgpu = workload->vgpu;
+	struct execlist_ctx_descriptor_format ctx[2];
+	int ring_id = workload->ring_id;
+	int ret;
+
+	ret = intel_vgpu_pin_mm(workload->shadow_mm);
+	if (ret) {
+		gvt_vgpu_err("fail to vgpu pin mm\n");
+		goto out;
+	}
+
+	ret = intel_vgpu_sync_oos_pages(workload->vgpu);
+	if (ret) {
+		gvt_vgpu_err("fail to vgpu sync oos pages\n");
+		goto err_unpin_mm;
+	}
 
-	i915_gem_object_unpin_map(wa_ctx->indirect_ctx.obj);
-	i915_gem_object_put(wa_ctx->indirect_ctx.obj);
+	ret = intel_vgpu_flush_post_shadow(workload->vgpu);
+	if (ret) {
+		gvt_vgpu_err("fail to flush post shadow\n");
+		goto err_unpin_mm;
+	}
+
+	ret = prepare_shadow_batch_buffer(workload);
+	if (ret) {
+		gvt_vgpu_err("fail to prepare_shadow_batch_buffer\n");
+		goto err_unpin_mm;
+	}
+
+	ret = prepare_shadow_wa_ctx(&workload->wa_ctx);
+	if (ret) {
+		gvt_vgpu_err("fail to prepare_shadow_wa_ctx\n");
+		goto err_shadow_batch;
+	}
+
+	if (!workload->emulate_schedule_in)
+		return 0;
+
+	ctx[0] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 1);
+	ctx[1] = *get_desc_from_elsp_dwords(&workload->elsp_dwords, 0);
+
+	ret = emulate_execlist_schedule_in(&vgpu->execlist[ring_id], ctx);
+	if (!ret)
+		goto out;
+	else
+		gvt_vgpu_err("fail to emulate execlist schedule in\n");
+
+	release_shadow_wa_ctx(&workload->wa_ctx);
+err_shadow_batch:
+	release_shadow_batch_buffer(workload);
+err_unpin_mm:
+	intel_vgpu_unpin_mm(workload->shadow_mm);
+out:
+	return ret;
 }
 
 static int complete_execlist_workload(struct intel_vgpu_workload *workload)
@@ -511,8 +542,10 @@ static int complete_execlist_workload(struct intel_vgpu_workload *workload)
 	gvt_dbg_el("complete workload %p status %d\n", workload,
 			workload->status);
 
-	release_shadow_batch_buffer(workload);
-	release_shadow_wa_ctx(&workload->wa_ctx);
+	if (!workload->status) {
+		release_shadow_batch_buffer(workload);
+		release_shadow_wa_ctx(&workload->wa_ctx);
+	}
 
 	if (workload->status || (vgpu->resetting_eng & ENGINE_MASK(ring_id))) {
 		/* if workload->status is not successful means HW GPU
@@ -820,10 +853,21 @@ static void clean_workloads(struct intel_vgpu *vgpu, unsigned long engine_mask)
 
 void intel_vgpu_clean_execlist(struct intel_vgpu *vgpu)
 {
+	enum intel_engine_id i;
+	struct intel_engine_cs *engine;
+
 	clean_workloads(vgpu, ALL_ENGINES);
 	kmem_cache_destroy(vgpu->workloads);
+
+	for_each_engine(engine, vgpu->gvt->dev_priv, i) {
+		kfree(vgpu->reserve_ring_buffer_va[i]);
+		vgpu->reserve_ring_buffer_va[i] = NULL;
+		vgpu->reserve_ring_buffer_size[i] = 0;
+	}
+
 }
 
+#define RESERVE_RING_BUFFER_SIZE		((1 * PAGE_SIZE)/8)
 int intel_vgpu_init_execlist(struct intel_vgpu *vgpu)
 {
 	enum intel_engine_id i;
@@ -843,7 +887,26 @@ int intel_vgpu_init_execlist(struct intel_vgpu *vgpu)
 	if (!vgpu->workloads)
 		return -ENOMEM;
 
+	/* each ring has a shadow ring buffer until vgpu destroyed */
+	for_each_engine(engine, vgpu->gvt->dev_priv, i) {
+		vgpu->reserve_ring_buffer_va[i] =
+			kmalloc(RESERVE_RING_BUFFER_SIZE, GFP_KERNEL);
+		if (!vgpu->reserve_ring_buffer_va[i]) {
+			gvt_vgpu_err("fail to alloc reserve ring buffer\n");
+			goto out;
+		}
+		vgpu->reserve_ring_buffer_size[i] = RESERVE_RING_BUFFER_SIZE;
+	}
 	return 0;
+out:
+	for_each_engine(engine, vgpu->gvt->dev_priv, i) {
+		if (vgpu->reserve_ring_buffer_size[i]) {
+			kfree(vgpu->reserve_ring_buffer_va[i]);
+			vgpu->reserve_ring_buffer_va[i] = NULL;
+			vgpu->reserve_ring_buffer_size[i] = 0;
+		}
+	}
+	return -ENOMEM;
 }
 
 void intel_vgpu_reset_execlist(struct intel_vgpu *vgpu,

drivers/gpu/drm/i915/gvt/gtt.c

@@ -1647,14 +1647,13 @@ int intel_vgpu_pin_mm(struct intel_vgpu_mm *mm)
 	if (WARN_ON(mm->type != INTEL_GVT_MM_PPGTT))
 		return 0;
 
-	atomic_inc(&mm->pincount);
-
 	if (!mm->shadowed) {
 		ret = shadow_mm(mm);
 		if (ret)
 			return ret;
 	}
 
+	atomic_inc(&mm->pincount);
 	list_del_init(&mm->lru_list);
 	list_add_tail(&mm->lru_list, &mm->vgpu->gvt->gtt.mm_lru_list_head);
 	return 0;
@@ -1972,7 +1971,7 @@ static int alloc_scratch_pages(struct intel_vgpu *vgpu,
 		 */
 		se.val64 |= _PAGE_PRESENT | _PAGE_RW;
 		if (type == GTT_TYPE_PPGTT_PDE_PT)
-			se.val64 |= PPAT_CACHED_INDEX;
+			se.val64 |= PPAT_CACHED;
 
 		for (i = 0; i < page_entry_num; i++)
 			ops->set_entry(scratch_pt, &se, i, false, 0, vgpu);

drivers/gpu/drm/i915/gvt/gvt.c

@@ -111,7 +111,7 @@ static void init_device_info(struct intel_gvt *gvt)
 	if (IS_BROADWELL(gvt->dev_priv) || IS_SKYLAKE(gvt->dev_priv)
 		|| IS_KABYLAKE(gvt->dev_priv)) {
 		info->max_support_vgpus = 8;
-		info->cfg_space_size = 256;
+		info->cfg_space_size = PCI_CFG_SPACE_EXP_SIZE;
 		info->mmio_size = 2 * 1024 * 1024;
 		info->mmio_bar = 0;
 		info->gtt_start_offset = 8 * 1024 * 1024;

drivers/gpu/drm/i915/gvt/gvt.h

@@ -80,6 +80,7 @@ struct intel_gvt_device_info {
 struct intel_vgpu_gm {
 	u64 aperture_sz;
 	u64 hidden_sz;
+	void *aperture_va;
 	struct drm_mm_node low_gm_node;
 	struct drm_mm_node high_gm_node;
 };
@@ -99,7 +100,6 @@ struct intel_vgpu_mmio {
 	bool disable_warn_untrack;
 };
 
-#define INTEL_GVT_MAX_CFG_SPACE_SZ 256
 #define INTEL_GVT_MAX_BAR_NUM 4
 
 struct intel_vgpu_pci_bar {
@@ -108,7 +108,7 @@ struct intel_vgpu_pci_bar {
 };
 
 struct intel_vgpu_cfg_space {
-	unsigned char virtual_cfg_space[INTEL_GVT_MAX_CFG_SPACE_SZ];
+	unsigned char virtual_cfg_space[PCI_CFG_SPACE_EXP_SIZE];
 	struct intel_vgpu_pci_bar bar[INTEL_GVT_MAX_BAR_NUM];
 };
 
@@ -165,6 +165,9 @@ struct intel_vgpu {
 	struct list_head workload_q_head[I915_NUM_ENGINES];
 	struct kmem_cache *workloads;
 	atomic_t running_workload_num;
+	/* 1/2K for each reserve ring buffer */
+	void *reserve_ring_buffer_va[I915_NUM_ENGINES];
+	int reserve_ring_buffer_size[I915_NUM_ENGINES];
 	DECLARE_BITMAP(tlb_handle_pending, I915_NUM_ENGINES);
 	struct i915_gem_context *shadow_ctx;
 	DECLARE_BITMAP(shadow_ctx_desc_updated, I915_NUM_ENGINES);
@@ -474,6 +477,13 @@ int intel_vgpu_emulate_cfg_read(struct intel_vgpu *vgpu, unsigned int offset,
 int intel_vgpu_emulate_cfg_write(struct intel_vgpu *vgpu, unsigned int offset,
 		void *p_data, unsigned int bytes);
 
+static inline u64 intel_vgpu_get_bar_gpa(struct intel_vgpu *vgpu, int bar)
+{
+	/* We are 64bit bar. */
+	return (*(u64 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
+			PCI_BASE_ADDRESS_MEM_MASK;
+}
+
 void intel_gvt_clean_opregion(struct intel_gvt *gvt);
 int intel_gvt_init_opregion(struct intel_gvt *gvt);
 

drivers/gpu/drm/i915/gvt/kvmgt.c

@@ -609,21 +609,20 @@ static void intel_vgpu_release_work(struct work_struct *work)
 	__intel_vgpu_release(vgpu);
 }
 
-static uint64_t intel_vgpu_get_bar0_addr(struct intel_vgpu *vgpu)
+static uint64_t intel_vgpu_get_bar_addr(struct intel_vgpu *vgpu, int bar)
 {
 	u32 start_lo, start_hi;
 	u32 mem_type;
-	int pos = PCI_BASE_ADDRESS_0;
 
-	start_lo = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + pos)) &
+	start_lo = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
 			PCI_BASE_ADDRESS_MEM_MASK;
-	mem_type = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + pos)) &
+	mem_type = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space + bar)) &
 			PCI_BASE_ADDRESS_MEM_TYPE_MASK;
 
 	switch (mem_type) {
 	case PCI_BASE_ADDRESS_MEM_TYPE_64:
 		start_hi = (*(u32 *)(vgpu->cfg_space.virtual_cfg_space
-						+ pos + 4));
+						+ bar + 4));
 		break;
 	case PCI_BASE_ADDRESS_MEM_TYPE_32:
 	case PCI_BASE_ADDRESS_MEM_TYPE_1M:
@@ -637,6 +636,21 @@ static uint64_t intel_vgpu_get_bar0_addr(struct intel_vgpu *vgpu)
 	return ((u64)start_hi << 32) | start_lo;
 }
 
+static int intel_vgpu_bar_rw(struct intel_vgpu *vgpu, int bar, uint64_t off,
+			     void *buf, unsigned int count, bool is_write)
+{
+	uint64_t bar_start = intel_vgpu_get_bar_addr(vgpu, bar);
+	int ret;
+
+	if (is_write)
+		ret = intel_gvt_ops->emulate_mmio_write(vgpu,
+					bar_start + off, buf, count);
+	else
+		ret = intel_gvt_ops->emulate_mmio_read(vgpu,
+					bar_start + off, buf, count);
+	return ret;
+}
+
 static ssize_t intel_vgpu_rw(struct mdev_device *mdev, char *buf,
 			size_t count, loff_t *ppos, bool is_write)
 {
@@ -661,20 +675,14 @@ static ssize_t intel_vgpu_rw(struct mdev_device *mdev, char *buf,
 						buf, count);
 		break;
 	case VFIO_PCI_BAR0_REGION_INDEX:
-	case VFIO_PCI_BAR1_REGION_INDEX:
-		if (is_write) {
-			uint64_t bar0_start = intel_vgpu_get_bar0_addr(vgpu);
-
-			ret = intel_gvt_ops->emulate_mmio_write(vgpu,
-						bar0_start + pos, buf, count);
-		} else {
-			uint64_t bar0_start = intel_vgpu_get_bar0_addr(vgpu);
-
-			ret = intel_gvt_ops->emulate_mmio_read(vgpu,
-						bar0_start + pos, buf, count);
-		}
+		ret = intel_vgpu_bar_rw(vgpu, PCI_BASE_ADDRESS_0, pos,
+					buf, count, is_write);
 		break;
 	case VFIO_PCI_BAR2_REGION_INDEX:
+		ret = intel_vgpu_bar_rw(vgpu, PCI_BASE_ADDRESS_2, pos,
+					buf, count, is_write);
+		break;
+	case VFIO_PCI_BAR1_REGION_INDEX:
 	case VFIO_PCI_BAR3_REGION_INDEX:
 	case VFIO_PCI_BAR4_REGION_INDEX:
 	case VFIO_PCI_BAR5_REGION_INDEX:
@@ -970,7 +978,7 @@ static long intel_vgpu_ioctl(struct mdev_device *mdev, unsigned int cmd,
 		switch (info.index) {
 		case VFIO_PCI_CONFIG_REGION_INDEX:
 			info.offset = VFIO_PCI_INDEX_TO_OFFSET(info.index);
-			info.size = INTEL_GVT_MAX_CFG_SPACE_SZ;
+			info.size = vgpu->gvt->device_info.cfg_space_size;
 			info.flags = VFIO_REGION_INFO_FLAG_READ |
 				     VFIO_REGION_INFO_FLAG_WRITE;
 			break;

drivers/gpu/drm/i915/gvt/mmio.c

@@ -45,8 +45,7 @@
  */
 int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa)
 {
-	u64 gttmmio_gpa = *(u64 *)(vgpu_cfg_space(vgpu) + PCI_BASE_ADDRESS_0) &
-			  ~GENMASK(3, 0);
+	u64 gttmmio_gpa = intel_vgpu_get_bar_gpa(vgpu, PCI_BASE_ADDRESS_0);
 	return gpa - gttmmio_gpa;
 }
 
@@ -57,6 +56,38 @@ int intel_vgpu_gpa_to_mmio_offset(struct intel_vgpu *vgpu, u64 gpa)
 	(reg >= gvt->device_info.gtt_start_offset \
 	 && reg < gvt->device_info.gtt_start_offset + gvt_ggtt_sz(gvt))
 
+static bool vgpu_gpa_is_aperture(struct intel_vgpu *vgpu, uint64_t gpa)
+{
+	u64 aperture_gpa = intel_vgpu_get_bar_gpa(vgpu, PCI_BASE_ADDRESS_2);
+	u64 aperture_sz = vgpu_aperture_sz(vgpu);
+
+	return gpa >= aperture_gpa && gpa < aperture_gpa + aperture_sz;
+}
+
+static int vgpu_aperture_rw(struct intel_vgpu *vgpu, uint64_t gpa,
+			    void *pdata, unsigned int size, bool is_read)
+{
+	u64 aperture_gpa = intel_vgpu_get_bar_gpa(vgpu, PCI_BASE_ADDRESS_2);
+	u64 offset = gpa - aperture_gpa;
+
+	if (!vgpu_gpa_is_aperture(vgpu, gpa + size - 1)) {
+		gvt_vgpu_err("Aperture rw out of range, offset %llx, size %d\n",
+			     offset, size);
+		return -EINVAL;
+	}
+
+	if (!vgpu->gm.aperture_va) {
+		gvt_vgpu_err("BAR is not enabled\n");
+		return -ENXIO;
+	}
+
+	if (is_read)
+		memcpy(pdata, vgpu->gm.aperture_va + offset, size);
+	else
+		memcpy(vgpu->gm.aperture_va + offset, pdata, size);
+	return 0;
+}
+
 static void failsafe_emulate_mmio_rw(struct intel_vgpu *vgpu, uint64_t pa,
 		void *p_data, unsigned int bytes, bool read)
 {
@@ -133,6 +164,12 @@ int intel_vgpu_emulate_mmio_read(struct intel_vgpu *vgpu, uint64_t pa,
 	}
 	mutex_lock(&gvt->lock);
 
+	if (vgpu_gpa_is_aperture(vgpu, pa)) {
+		ret = vgpu_aperture_rw(vgpu, pa, p_data, bytes, true);
+		mutex_unlock(&gvt->lock);
+		return ret;
+	}
+
 	if (atomic_read(&vgpu->gtt.n_write_protected_guest_page)) {
 		struct intel_vgpu_guest_page *gp;
 
@@ -224,6 +261,12 @@ int intel_vgpu_emulate_mmio_write(struct intel_vgpu *vgpu, uint64_t pa,
 
 	mutex_lock(&gvt->lock);
 
+	if (vgpu_gpa_is_aperture(vgpu, pa)) {
+		ret = vgpu_aperture_rw(vgpu, pa, p_data, bytes, false);
+		mutex_unlock(&gvt->lock);
+		return ret;
+	}
+
 	if (atomic_read(&vgpu->gtt.n_write_protected_guest_page)) {
 		struct intel_vgpu_guest_page *gp;
 

drivers/gpu/drm/i915/gvt/render.c

@@ -293,7 +293,7 @@ static void switch_mmio_to_vgpu(struct intel_vgpu *vgpu, int ring_id)
 		 */
 		if (mmio->in_context &&
 				((ctx_ctrl & inhibit_mask) != inhibit_mask) &&
-				i915.enable_execlists)
+				i915_modparams.enable_execlists)
 			continue;
 
 		if (mmio->mask)

drivers/gpu/drm/i915/gvt/scheduler.c

@@ -87,7 +87,7 @@ static int populate_shadow_context(struct intel_vgpu_workload *workload)
 			return -EINVAL;
 		}
 
-		page = i915_gem_object_get_page(ctx_obj, LRC_PPHWSP_PN + i);
+		page = i915_gem_object_get_page(ctx_obj, LRC_HEADER_PAGES + i);
 		dst = kmap(page);
 		intel_gvt_hypervisor_read_gpa(vgpu, context_gpa, dst,
 				GTT_PAGE_SIZE);
@@ -201,6 +201,43 @@ static void shadow_context_descriptor_update(struct i915_gem_context *ctx,
 	ce->lrc_desc = desc;
 }
 
+static int copy_workload_to_ring_buffer(struct intel_vgpu_workload *workload)
+{
+	struct intel_vgpu *vgpu = workload->vgpu;
+	void *shadow_ring_buffer_va;
+	u32 *cs;
+
+	/* allocate shadow ring buffer */
+	cs = intel_ring_begin(workload->req, workload->rb_len / sizeof(u32));
+	if (IS_ERR(cs)) {
+		gvt_vgpu_err("fail to alloc size =%ld shadow  ring buffer\n",
+			workload->rb_len);
+		return PTR_ERR(cs);
+	}
+
+	shadow_ring_buffer_va = workload->shadow_ring_buffer_va;
+
+	/* get shadow ring buffer va */
+	workload->shadow_ring_buffer_va = cs;
+
+	memcpy(cs, shadow_ring_buffer_va,
+			workload->rb_len);
+
+	cs += workload->rb_len / sizeof(u32);
+	intel_ring_advance(workload->req, cs);
+
+	return 0;
+}
+
+void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx)
+{
+	if (!wa_ctx->indirect_ctx.obj)
+		return;
+
+	i915_gem_object_unpin_map(wa_ctx->indirect_ctx.obj);
+	i915_gem_object_put(wa_ctx->indirect_ctx.obj);
+}
+
 /**
  * intel_gvt_scan_and_shadow_workload - audit the workload by scanning and
  * shadow it as well, include ringbuffer,wa_ctx and ctx.
@@ -214,8 +251,10 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 	int ring_id = workload->ring_id;
 	struct i915_gem_context *shadow_ctx = workload->vgpu->shadow_ctx;
 	struct drm_i915_private *dev_priv = workload->vgpu->gvt->dev_priv;
+	struct intel_engine_cs *engine = dev_priv->engine[ring_id];
 	struct drm_i915_gem_request *rq;
 	struct intel_vgpu *vgpu = workload->vgpu;
+	struct intel_ring *ring;
 	int ret;
 
 	lockdep_assert_held(&dev_priv->drm.struct_mutex);
@@ -231,35 +270,56 @@ int intel_gvt_scan_and_shadow_workload(struct intel_vgpu_workload *workload)
 		shadow_context_descriptor_update(shadow_ctx,
 					dev_priv->engine[ring_id]);
 
-	rq = i915_gem_request_alloc(dev_priv->engine[ring_id], shadow_ctx);
-	if (IS_ERR(rq)) {
-		gvt_vgpu_err("fail to allocate gem request\n");
-		ret = PTR_ERR(rq);
-		goto out;
-	}
-
-	gvt_dbg_sched("ring id %d get i915 gem request %p\n", ring_id, rq);
-
-	workload->req = i915_gem_request_get(rq);
-
 	ret = intel_gvt_scan_and_shadow_ringbuffer(workload);
 	if (ret)
-		goto out;
+		goto err_scan;
 
 	if ((workload->ring_id == RCS) &&
 	    (workload->wa_ctx.indirect_ctx.size != 0)) {
 		ret = intel_gvt_scan_and_shadow_wa_ctx(&workload->wa_ctx);
 		if (ret)
-			goto out;
+			goto err_scan;
+	}
+
+	/* pin shadow context by gvt even the shadow context will be pinned
+	 * when i915 alloc request. That is because gvt will update the guest
+	 * context from shadow context when workload is completed, and at that
+	 * moment, i915 may already unpined the shadow context to make the
+	 * shadow_ctx pages invalid. So gvt need to pin itself. After update
+	 * the guest context, gvt can unpin the shadow_ctx safely.
+	 */
+	ring = engine->context_pin(engine, shadow_ctx);
+	if (IS_ERR(ring)) {
+		ret = PTR_ERR(ring);
+		gvt_vgpu_err("fail to pin shadow context\n");
+		goto err_shadow;
 	}
 
 	ret = populate_shadow_context(workload);
 	if (ret)
-		goto out;
+		goto err_unpin;
 
+	rq = i915_gem_request_alloc(dev_priv->engine[ring_id], shadow_ctx);
+	if (IS_ERR(rq)) {
+		gvt_vgpu_err("fail to allocate gem request\n");
+		ret = PTR_ERR(rq);
+		goto err_unpin;
+	}
+
+	gvt_dbg_sched("ring id %d get i915 gem request %p\n", ring_id, rq);
+
+	workload->req = i915_gem_request_get(rq);
+	ret = copy_workload_to_ring_buffer(workload);
+	if (ret)
+		goto err_unpin;
 	workload->shadowed = true;
+	return 0;
 
-out:
+err_unpin:
+	engine->context_unpin(engine, shadow_ctx);
+err_shadow:
+	release_shadow_wa_ctx(&workload->wa_ctx);
+err_scan:
 	return ret;
 }
 
@@ -269,8 +329,6 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
 	struct i915_gem_context *shadow_ctx = workload->vgpu->shadow_ctx;
 	struct drm_i915_private *dev_priv = workload->vgpu->gvt->dev_priv;
 	struct intel_engine_cs *engine = dev_priv->engine[ring_id];
-	struct intel_vgpu *vgpu = workload->vgpu;
-	struct intel_ring *ring;
 	int ret = 0;
 
 	gvt_dbg_sched("ring id %d prepare to dispatch workload %p\n",
@@ -284,22 +342,10 @@ static int dispatch_workload(struct intel_vgpu_workload *workload)
 
 	if (workload->prepare) {
 		ret = workload->prepare(workload);
-		if (ret)
+		if (ret) {
+			engine->context_unpin(engine, shadow_ctx);
 			goto out;
-	}
-
-	/* pin shadow context by gvt even the shadow context will be pinned
-	 * when i915 alloc request. That is because gvt will update the guest
-	 * context from shadow context when workload is completed, and at that
-	 * moment, i915 may already unpined the shadow context to make the
-	 * shadow_ctx pages invalid. So gvt need to pin itself. After update
-	 * the guest context, gvt can unpin the shadow_ctx safely.
-	 */
-	ring = engine->context_pin(engine, shadow_ctx);
-	if (IS_ERR(ring)) {
-		ret = PTR_ERR(ring);
-		gvt_vgpu_err("fail to pin shadow context\n");
-		goto out;
+		}
 	}
 
 out:
@@ -408,7 +454,7 @@ static void update_guest_context(struct intel_vgpu_workload *workload)
 			return;
 		}
 
-		page = i915_gem_object_get_page(ctx_obj, LRC_PPHWSP_PN + i);
+		page = i915_gem_object_get_page(ctx_obj, LRC_HEADER_PAGES + i);
 		src = kmap(page);
 		intel_gvt_hypervisor_write_gpa(vgpu, context_gpa, src,
 				GTT_PAGE_SIZE);

drivers/gpu/drm/i915/gvt/scheduler.h

@@ -140,4 +140,5 @@ int intel_vgpu_init_gvt_context(struct intel_vgpu *vgpu);
 
 void intel_vgpu_clean_gvt_context(struct intel_vgpu *vgpu);
 
+void release_shadow_wa_ctx(struct intel_shadow_wa_ctx *wa_ctx);
 #endif

drivers/gpu/drm/i915/i915_debugfs.c

@@ -67,7 +67,7 @@ static int i915_capabilities(struct seq_file *m, void *data)
 #undef PRINT_FLAG
 
 	kernel_param_lock(THIS_MODULE);
-#define PRINT_PARAM(T, x) seq_print_param(m, #x, #T, &i915.x);
+#define PRINT_PARAM(T, x, ...) seq_print_param(m, #x, #T, &i915_modparams.x);
 	I915_PARAMS_FOR_EACH(PRINT_PARAM);
 #undef PRINT_PARAM
 	kernel_param_unlock(THIS_MODULE);
@@ -1267,7 +1267,7 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused)
 	if (waitqueue_active(&dev_priv->gpu_error.reset_queue))
 		seq_puts(m, "struct_mutex blocked for reset\n");
 
-	if (!i915.enable_hangcheck) {
+	if (!i915_modparams.enable_hangcheck) {
 		seq_puts(m, "Hangcheck disabled\n");
 		return 0;
 	}
@@ -1422,6 +1422,9 @@ static int i915_forcewake_domains(struct seq_file *m, void *data)
 	struct intel_uncore_forcewake_domain *fw_domain;
 	unsigned int tmp;
 
+	seq_printf(m, "user.bypass_count = %u\n",
+		   i915->uncore.user_forcewake.count);
+
 	for_each_fw_domain(fw_domain, i915, tmp)
 		seq_printf(m, "%s.wake_count = %u\n",
 			   intel_uncore_forcewake_domain_to_str(fw_domain->id),
@@ -1699,7 +1702,7 @@ static int i915_ips_status(struct seq_file *m, void *unused)
 	intel_runtime_pm_get(dev_priv);
 
 	seq_printf(m, "Enabled by kernel parameter: %s\n",
-		   yesno(i915.enable_ips));
+		   yesno(i915_modparams.enable_ips));
 
 	if (INTEL_GEN(dev_priv) >= 8) {
 		seq_puts(m, "Currently: unknown\n");
@@ -2014,7 +2017,7 @@ static int i915_dump_lrc(struct seq_file *m, void *unused)
 	enum intel_engine_id id;
 	int ret;
 
-	if (!i915.enable_execlists) {
+	if (!i915_modparams.enable_execlists) {
 		seq_printf(m, "Logical Ring Contexts are disabled\n");
 		return 0;
 	}
@@ -2443,12 +2446,8 @@ static void i915_guc_client_info(struct seq_file *m,
 
 	seq_printf(m, "\tPriority %d, GuC stage index: %u, PD offset 0x%x\n",
 		client->priority, client->stage_id, client->proc_desc_offset);
-	seq_printf(m, "\tDoorbell id %d, offset: 0x%lx, cookie 0x%x\n",
-		client->doorbell_id, client->doorbell_offset, client->doorbell_cookie);
-	seq_printf(m, "\tWQ size %d, offset: 0x%x, tail %d\n",
-		client->wq_size, client->wq_offset, client->wq_tail);
-
-	seq_printf(m, "\tWork queue full: %u\n", client->no_wq_space);
+	seq_printf(m, "\tDoorbell id %d, offset: 0x%lx\n",
+		client->doorbell_id, client->doorbell_offset);
 
 	for_each_engine(engine, dev_priv, id) {
 		u64 submissions = client->submissions[id];
@@ -2594,7 +2593,7 @@ static int i915_guc_log_control_get(void *data, u64 *val)
 	if (!dev_priv->guc.log.vma)
 		return -EINVAL;
 
-	*val = i915.guc_log_level;
+	*val = i915_modparams.guc_log_level;
 
 	return 0;
 }
@@ -3312,7 +3311,9 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 		seq_printf(m, "\tBBADDR: 0x%08x_%08x\n",
 			   upper_32_bits(addr), lower_32_bits(addr));
 
-		if (i915.enable_execlists) {
+		if (i915_modparams.enable_execlists) {
+			const u32 *hws = &engine->status_page.page_addr[I915_HWS_CSB_BUF0_INDEX];
+			struct intel_engine_execlists * const execlists = &engine->execlists;
 			u32 ptr, read, write;
 			unsigned int idx;
 
@@ -3323,8 +3324,10 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 			ptr = I915_READ(RING_CONTEXT_STATUS_PTR(engine));
 			read = GEN8_CSB_READ_PTR(ptr);
 			write = GEN8_CSB_WRITE_PTR(ptr);
-			seq_printf(m, "\tExeclist CSB read %d, write %d, interrupt posted? %s\n",
-				   read, write,
+			seq_printf(m, "\tExeclist CSB read %d [%d cached], write %d [%d from hws], interrupt posted? %s\n",
+				   read, execlists->csb_head,
+				   write,
+				   intel_read_status_page(engine, intel_hws_csb_write_index(engine->i915)),
 				   yesno(test_bit(ENGINE_IRQ_EXECLIST,
 						  &engine->irq_posted)));
 			if (read >= GEN8_CSB_ENTRIES)
@@ -3335,18 +3338,19 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 				write += GEN8_CSB_ENTRIES;
 			while (read < write) {
 				idx = ++read % GEN8_CSB_ENTRIES;
-				seq_printf(m, "\tExeclist CSB[%d]: 0x%08x, context: %d\n",
+				seq_printf(m, "\tExeclist CSB[%d]: 0x%08x [0x%08x in hwsp], context: %d [%d in hwsp]\n",
 					   idx,
 					   I915_READ(RING_CONTEXT_STATUS_BUF_LO(engine, idx)),
-					   I915_READ(RING_CONTEXT_STATUS_BUF_HI(engine, idx)));
+					   hws[idx * 2],
+					   I915_READ(RING_CONTEXT_STATUS_BUF_HI(engine, idx)),
+					   hws[idx * 2 + 1]);
 			}
 
 			rcu_read_lock();
-			for (idx = 0; idx < ARRAY_SIZE(engine->execlist_port); idx++) {
+			for (idx = 0; idx < execlists_num_ports(execlists); idx++) {
 				unsigned int count;
 
-				rq = port_unpack(&engine->execlist_port[idx],
-						 &count);
+				rq = port_unpack(&execlists->port[idx], &count);
 				if (rq) {
 					seq_printf(m, "\t\tELSP[%d] count=%d, ",
 						   idx, count);
@@ -3359,7 +3363,7 @@ static int i915_engine_info(struct seq_file *m, void *unused)
 			rcu_read_unlock();
 
 			spin_lock_irq(&engine->timeline->lock);
-			for (rb = engine->execlist_first; rb; rb = rb_next(rb)){
+			for (rb = execlists->first; rb; rb = rb_next(rb)) {
 				struct i915_priolist *p =
 					rb_entry(rb, typeof(*p), node);
 
@@ -3403,7 +3407,7 @@ static int i915_semaphore_status(struct seq_file *m, void *unused)
 	enum intel_engine_id id;
 	int j, ret;
 
-	if (!i915.semaphores) {
+	if (!i915_modparams.semaphores) {
 		seq_puts(m, "Semaphores are disabled\n");
 		return 0;
 	}
@@ -3523,6 +3527,57 @@ static int i915_wa_registers(struct seq_file *m, void *unused)
 	return 0;
 }
 
+static int i915_ipc_status_show(struct seq_file *m, void *data)
+{
+	struct drm_i915_private *dev_priv = m->private;
+
+	seq_printf(m, "Isochronous Priority Control: %s\n",
+			yesno(dev_priv->ipc_enabled));
+	return 0;
+}
+
+static int i915_ipc_status_open(struct inode *inode, struct file *file)
+{
+	struct drm_i915_private *dev_priv = inode->i_private;
+
+	if (!HAS_IPC(dev_priv))
+		return -ENODEV;
+
+	return single_open(file, i915_ipc_status_show, dev_priv);
+}
+
+static ssize_t i915_ipc_status_write(struct file *file, const char __user *ubuf,
+				     size_t len, loff_t *offp)
+{
+	struct seq_file *m = file->private_data;
+	struct drm_i915_private *dev_priv = m->private;
+	int ret;
+	bool enable;
+
+	ret = kstrtobool_from_user(ubuf, len, &enable);
+	if (ret < 0)
+		return ret;
+
+	intel_runtime_pm_get(dev_priv);
+	if (!dev_priv->ipc_enabled && enable)
+		DRM_INFO("Enabling IPC: WM will be proper only after next commit\n");
+	dev_priv->wm.distrust_bios_wm = true;
+	dev_priv->ipc_enabled = enable;
+	intel_enable_ipc(dev_priv);
+	intel_runtime_pm_put(dev_priv);
+
+	return len;
+}
+
+static const struct file_operations i915_ipc_status_fops = {
+	.owner = THIS_MODULE,
+	.open = i915_ipc_status_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+	.write = i915_ipc_status_write
+};
+
 static int i915_ddb_info(struct seq_file *m, void *unused)
 {
 	struct drm_i915_private *dev_priv = node_to_i915(m->private);
@@ -4674,26 +4729,26 @@ static int i915_sseu_status(struct seq_file *m, void *unused)
 
 static int i915_forcewake_open(struct inode *inode, struct file *file)
 {
-	struct drm_i915_private *dev_priv = inode->i_private;
+	struct drm_i915_private *i915 = inode->i_private;
 
-	if (INTEL_GEN(dev_priv) < 6)
+	if (INTEL_GEN(i915) < 6)
 		return 0;
 
-	intel_runtime_pm_get(dev_priv);
-	intel_uncore_forcewake_get(dev_priv, FORCEWAKE_ALL);
+	intel_runtime_pm_get(i915);
+	intel_uncore_forcewake_user_get(i915);
 
 	return 0;
 }
 
 static int i915_forcewake_release(struct inode *inode, struct file *file)
 {
-	struct drm_i915_private *dev_priv = inode->i_private;
+	struct drm_i915_private *i915 = inode->i_private;
 
-	if (INTEL_GEN(dev_priv) < 6)
+	if (INTEL_GEN(i915) < 6)
 		return 0;
 
-	intel_uncore_forcewake_put(dev_priv, FORCEWAKE_ALL);
-	intel_runtime_pm_put(dev_priv);
+	intel_uncore_forcewake_user_put(i915);
+	intel_runtime_pm_put(i915);
 
 	return 0;
 }
@@ -4859,7 +4914,8 @@ static const struct i915_debugfs_files {
 	{"i915_dp_test_type", &i915_displayport_test_type_fops},
 	{"i915_dp_test_active", &i915_displayport_test_active_fops},
 	{"i915_guc_log_control", &i915_guc_log_control_fops},
-	{"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops}
+	{"i915_hpd_storm_ctl", &i915_hpd_storm_ctl_fops},
+	{"i915_ipc_status", &i915_ipc_status_fops}
 };
 
 int i915_debugfs_register(struct drm_i915_private *dev_priv)

drivers/gpu/drm/i915/i915_drv.c

@@ -58,12 +58,12 @@ static unsigned int i915_load_fail_count;
 
 bool __i915_inject_load_failure(const char *func, int line)
 {
-	if (i915_load_fail_count >= i915.inject_load_failure)
+	if (i915_load_fail_count >= i915_modparams.inject_load_failure)
 		return false;
 
-	if (++i915_load_fail_count == i915.inject_load_failure) {
+	if (++i915_load_fail_count == i915_modparams.inject_load_failure) {
 		DRM_INFO("Injecting failure at checkpoint %u [%s:%d]\n",
-			 i915.inject_load_failure, func, line);
+			 i915_modparams.inject_load_failure, func, line);
 		return true;
 	}
 
@@ -106,8 +106,8 @@ __i915_printk(struct drm_i915_private *dev_priv, const char *level,
 
 static bool i915_error_injected(struct drm_i915_private *dev_priv)
 {
-	return i915.inject_load_failure &&
-	       i915_load_fail_count == i915.inject_load_failure;
+	return i915_modparams.inject_load_failure &&
+	       i915_load_fail_count == i915_modparams.inject_load_failure;
 }
 
 #define i915_load_error(dev_priv, fmt, ...)				     \
@@ -321,7 +321,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = USES_PPGTT(dev_priv);
 		break;
 	case I915_PARAM_HAS_SEMAPHORES:
-		value = i915.semaphores;
+		value = i915_modparams.semaphores;
 		break;
 	case I915_PARAM_HAS_SECURE_BATCHES:
 		value = capable(CAP_SYS_ADMIN);
@@ -340,7 +340,8 @@ static int i915_getparam(struct drm_device *dev, void *data,
 			return -ENODEV;
 		break;
 	case I915_PARAM_HAS_GPU_RESET:
-		value = i915.enable_hangcheck && intel_has_gpu_reset(dev_priv);
+		value = i915_modparams.enable_hangcheck &&
+			intel_has_gpu_reset(dev_priv);
 		if (value && intel_has_reset_engine(dev_priv))
 			value = 2;
 		break;
@@ -869,6 +870,10 @@ static int i915_driver_init_early(struct drm_i915_private *dev_priv,
 	memcpy(device_info, match_info, sizeof(*device_info));
 	device_info->device_id = dev_priv->drm.pdev->device;
 
+	BUILD_BUG_ON(INTEL_MAX_PLATFORMS >
+		     sizeof(device_info->platform_mask) * BITS_PER_BYTE);
+	device_info->platform_mask = BIT(device_info->platform);
+
 	BUG_ON(device_info->gen > sizeof(device_info->gen_mask) * BITS_PER_BYTE);
 	device_info->gen_mask = BIT(device_info->gen - 1);
 
@@ -1031,9 +1036,9 @@ static void i915_driver_cleanup_mmio(struct drm_i915_private *dev_priv)
 
 static void intel_sanitize_options(struct drm_i915_private *dev_priv)
 {
-	i915.enable_execlists =
+	i915_modparams.enable_execlists =
 		intel_sanitize_enable_execlists(dev_priv,
-						i915.enable_execlists);
+						i915_modparams.enable_execlists);
 
 	/*
 	 * i915.enable_ppgtt is read-only, so do an early pass to validate the
@@ -1041,12 +1046,15 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv)
 	 * do this now so that we can print out any log messages once rather
 	 * than every time we check intel_enable_ppgtt().
 	 */
-	i915.enable_ppgtt =
-		intel_sanitize_enable_ppgtt(dev_priv, i915.enable_ppgtt);
-	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt);
+	i915_modparams.enable_ppgtt =
+		intel_sanitize_enable_ppgtt(dev_priv,
+					    i915_modparams.enable_ppgtt);
+	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915_modparams.enable_ppgtt);
 
-	i915.semaphores = intel_sanitize_semaphores(dev_priv, i915.semaphores);
-	DRM_DEBUG_DRIVER("use GPU semaphores? %s\n", yesno(i915.semaphores));
+	i915_modparams.semaphores =
+		intel_sanitize_semaphores(dev_priv, i915_modparams.semaphores);
+	DRM_DEBUG_DRIVER("use GPU semaphores? %s\n",
+			 yesno(i915_modparams.semaphores));
 
 	intel_uc_sanitize_options(dev_priv);
 
@@ -1277,7 +1285,7 @@ int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
 	int ret;
 
 	/* Enable nuclear pageflip on ILK+ */
-	if (!i915.nuclear_pageflip && match_info->gen < 5)
+	if (!i915_modparams.nuclear_pageflip && match_info->gen < 5)
 		driver.driver_features &= ~DRIVER_ATOMIC;
 
 	ret = -ENOMEM;
@@ -1341,7 +1349,7 @@ int i915_driver_load(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	intel_runtime_pm_enable(dev_priv);
 
-	dev_priv->ipc_enabled = false;
+	intel_init_ipc(dev_priv);
 
 	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG))
 		DRM_INFO("DRM_I915_DEBUG enabled\n");
@@ -2609,6 +2617,8 @@ static int intel_runtime_resume(struct device *kdev)
 	if (!IS_VALLEYVIEW(dev_priv) && !IS_CHERRYVIEW(dev_priv))
 		intel_hpd_init(dev_priv);
 
+	intel_enable_ipc(dev_priv);
+
 	enable_rpm_wakeref_asserts(dev_priv);
 
 	if (ret)

drivers/gpu/drm/i915/i915_gem.c

@@ -179,7 +179,7 @@ i915_gem_object_get_pages_phys(struct drm_i915_gem_object *obj)
 	 * the alignment of the buddy allocation will naturally match.
 	 */
 	phys = drm_pci_alloc(obj->base.dev,
-			     obj->base.size,
+			     roundup_pow_of_two(obj->base.size),
 			     roundup_pow_of_two(obj->base.size));
 	if (!phys)
 		return ERR_PTR(-ENOMEM);
@@ -694,10 +694,10 @@ flush_write_domain(struct drm_i915_gem_object *obj, unsigned int flush_domains)
 
 	switch (obj->base.write_domain) {
 	case I915_GEM_DOMAIN_GTT:
-		if (INTEL_GEN(dev_priv) >= 6 && !HAS_LLC(dev_priv)) {
+		if (!HAS_LLC(dev_priv)) {
 			intel_runtime_pm_get(dev_priv);
 			spin_lock_irq(&dev_priv->uncore.lock);
-			POSTING_READ_FW(RING_ACTHD(dev_priv->engine[RCS]->mmio_base));
+			POSTING_READ_FW(RING_HEAD(dev_priv->engine[RCS]->mmio_base));
 			spin_unlock_irq(&dev_priv->uncore.lock);
 			intel_runtime_pm_put(dev_priv);
 		}
@@ -2303,7 +2303,7 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	struct sgt_iter sgt_iter;
 	struct page *page;
 	unsigned long last_pfn = 0;	/* suppress gcc warning */
-	unsigned int max_segment;
+	unsigned int max_segment = i915_sg_segment_size();
 	gfp_t noreclaim;
 	int ret;
 
@@ -2314,10 +2314,6 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj)
 	GEM_BUG_ON(obj->base.read_domains & I915_GEM_GPU_DOMAINS);
 	GEM_BUG_ON(obj->base.write_domain & I915_GEM_GPU_DOMAINS);
 
-	max_segment = swiotlb_max_segment();
-	if (!max_segment)
-		max_segment = rounddown(UINT_MAX, PAGE_SIZE);
-
 	st = kmalloc(sizeof(*st), GFP_KERNEL);
 	if (st == NULL)
 		return ERR_PTR(-ENOMEM);
@@ -2819,8 +2815,8 @@ i915_gem_reset_prepare_engine(struct intel_engine_cs *engine)
 	 * Turning off the engine->irq_tasklet until the reset is over
 	 * prevents the race.
 	 */
-	tasklet_kill(&engine->irq_tasklet);
-	tasklet_disable(&engine->irq_tasklet);
+	tasklet_kill(&engine->execlists.irq_tasklet);
+	tasklet_disable(&engine->execlists.irq_tasklet);
 
 	if (engine->irq_seqno_barrier)
 		engine->irq_seqno_barrier(engine);
@@ -2999,7 +2995,7 @@ void i915_gem_reset(struct drm_i915_private *dev_priv)
 
 void i915_gem_reset_finish_engine(struct intel_engine_cs *engine)
 {
-	tasklet_enable(&engine->irq_tasklet);
+	tasklet_enable(&engine->execlists.irq_tasklet);
 	kthread_unpark(engine->breadcrumbs.signaler);
 }
 
@@ -3026,9 +3022,6 @@ static void nop_submit_request(struct drm_i915_gem_request *request)
 
 static void engine_set_wedged(struct intel_engine_cs *engine)
 {
-	struct drm_i915_gem_request *request;
-	unsigned long flags;
-
 	/* We need to be sure that no thread is running the old callback as
 	 * we install the nop handler (otherwise we would submit a request
 	 * to hardware that will never complete). In order to prevent this
@@ -3038,40 +3031,7 @@ static void engine_set_wedged(struct intel_engine_cs *engine)
 	engine->submit_request = nop_submit_request;
 
 	/* Mark all executing requests as skipped */
-	spin_lock_irqsave(&engine->timeline->lock, flags);
-	list_for_each_entry(request, &engine->timeline->requests, link)
-		if (!i915_gem_request_completed(request))
-			dma_fence_set_error(&request->fence, -EIO);
-	spin_unlock_irqrestore(&engine->timeline->lock, flags);
-
-	/*
-	 * Clear the execlists queue up before freeing the requests, as those
-	 * are the ones that keep the context and ringbuffer backing objects
-	 * pinned in place.
-	 */
-
-	if (i915.enable_execlists) {
-		struct execlist_port *port = engine->execlist_port;
-		unsigned long flags;
-		unsigned int n;
-
-		spin_lock_irqsave(&engine->timeline->lock, flags);
-
-		for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++)
-			i915_gem_request_put(port_request(&port[n]));
-		memset(engine->execlist_port, 0, sizeof(engine->execlist_port));
-		engine->execlist_queue = RB_ROOT;
-		engine->execlist_first = NULL;
-
-		spin_unlock_irqrestore(&engine->timeline->lock, flags);
-
-		/* The port is checked prior to scheduling a tasklet, but
-		 * just in case we have suspended the tasklet to do the
-		 * wedging make sure that when it wakes, it decides there
-		 * is no work to do by clearing the irq_posted bit.
-		 */
-		clear_bit(ENGINE_IRQ_EXECLIST, &engine->irq_posted);
-	}
+	engine->cancel_requests(engine);
 
 	/* Mark all pending requests as complete so that any concurrent
 	 * (lockless) lookup doesn't try and wait upon the request as we
@@ -4778,7 +4738,7 @@ bool intel_sanitize_semaphores(struct drm_i915_private *dev_priv, int value)
 		return false;
 
 	/* TODO: make semaphores and Execlists play nicely together */
-	if (i915.enable_execlists)
+	if (i915_modparams.enable_execlists)
 		return false;
 
 	if (value >= 0)
@@ -4799,7 +4759,7 @@ int i915_gem_init(struct drm_i915_private *dev_priv)
 
 	dev_priv->mm.unordered_timeline = dma_fence_context_alloc(1);
 
-	if (!i915.enable_execlists) {
+	if (!i915_modparams.enable_execlists) {
 		dev_priv->gt.resume = intel_legacy_submission_resume;
 		dev_priv->gt.cleanup_engine = intel_engine_cleanup;
 	} else {

drivers/gpu/drm/i915/i915_gem_context.c

@@ -314,7 +314,7 @@ __create_hw_context(struct drm_i915_private *dev_priv,
 	 * present or not in use we still need a small bias as ring wraparound
 	 * at offset 0 sometimes hangs. No idea why.
 	 */
-	if (HAS_GUC(dev_priv) && i915.enable_guc_loading)
+	if (HAS_GUC(dev_priv) && i915_modparams.enable_guc_loading)
 		ctx->ggtt_offset_bias = GUC_WOPCM_TOP;
 	else
 		ctx->ggtt_offset_bias = I915_GTT_PAGE_SIZE;
@@ -407,7 +407,7 @@ i915_gem_context_create_gvt(struct drm_device *dev)
 	i915_gem_context_set_closed(ctx); /* not user accessible */
 	i915_gem_context_clear_bannable(ctx);
 	i915_gem_context_set_force_single_submission(ctx);
-	if (!i915.enable_guc_submission)
+	if (!i915_modparams.enable_guc_submission)
 		ctx->ring_size = 512 * PAGE_SIZE; /* Max ring buffer size */
 
 	GEM_BUG_ON(i915_gem_context_is_kernel(ctx));
@@ -431,7 +431,7 @@ int i915_gem_contexts_init(struct drm_i915_private *dev_priv)
 
 	if (intel_vgpu_active(dev_priv) &&
 	    HAS_LOGICAL_RING_CONTEXTS(dev_priv)) {
-		if (!i915.enable_execlists) {
+		if (!i915_modparams.enable_execlists) {
 			DRM_INFO("Only EXECLIST mode is supported in vgpu.\n");
 			return -EINVAL;
 		}
@@ -483,7 +483,7 @@ void i915_gem_contexts_lost(struct drm_i915_private *dev_priv)
 	}
 
 	/* Force the GPU state to be restored on enabling */
-	if (!i915.enable_execlists) {
+	if (!i915_modparams.enable_execlists) {
 		struct i915_gem_context *ctx;
 
 		list_for_each_entry(ctx, &dev_priv->contexts.list, link) {
@@ -568,7 +568,7 @@ mi_set_context(struct drm_i915_gem_request *req, u32 flags)
 	enum intel_engine_id id;
 	const int num_rings =
 		/* Use an extended w/a on gen7 if signalling from other rings */
-		(i915.semaphores && INTEL_GEN(dev_priv) == 7) ?
+		(i915_modparams.semaphores && INTEL_GEN(dev_priv) == 7) ?
 		INTEL_INFO(dev_priv)->num_rings - 1 :
 		0;
 	int len;
@@ -837,7 +837,7 @@ int i915_switch_context(struct drm_i915_gem_request *req)
 	struct intel_engine_cs *engine = req->engine;
 
 	lockdep_assert_held(&req->i915->drm.struct_mutex);
-	if (i915.enable_execlists)
+	if (i915_modparams.enable_execlists)
 		return 0;
 
 	if (!req->ctx->engine[engine->id].state) {

drivers/gpu/drm/i915/i915_gem_execbuffer.c

@@ -58,6 +58,7 @@ enum {
 
 #define __EXEC_HAS_RELOC	BIT(31)
 #define __EXEC_VALIDATED	BIT(30)
+#define __EXEC_INTERNAL_FLAGS	(~0u << 30)
 #define UPDATE			PIN_OFFSET_FIXED
 
 #define BATCH_OFFSET_BIAS (256*1024)
@@ -679,7 +680,7 @@ static int eb_select_context(struct i915_execbuffer *eb)
 static int eb_lookup_vmas(struct i915_execbuffer *eb)
 {
 	struct radix_tree_root *handles_vma = &eb->ctx->handles_vma;
-	struct drm_i915_gem_object *uninitialized_var(obj);
+	struct drm_i915_gem_object *obj;
 	unsigned int i;
 	int err;
 
@@ -725,19 +726,17 @@ static int eb_lookup_vmas(struct i915_execbuffer *eb)
 			goto err_obj;
 		}
 
+		/* transfer ref to ctx */
 		vma->open_count++;
 		list_add(&lut->obj_link, &obj->lut_list);
 		list_add(&lut->ctx_link, &eb->ctx->handles_list);
 		lut->ctx = eb->ctx;
 		lut->handle = handle;
 
-		/* transfer ref to ctx */
-		obj = NULL;
-
 add_vma:
 		err = eb_add_vma(eb, i, vma);
 		if (unlikely(err))
-			goto err_obj;
+			goto err_vma;
 
 		GEM_BUG_ON(vma != eb->vma[i]);
 		GEM_BUG_ON(vma->exec_flags != &eb->flags[i]);
@@ -766,8 +765,7 @@ static int eb_lookup_vmas(struct i915_execbuffer *eb)
 	return eb_reserve(eb);
 
 err_obj:
-	if (obj)
-		i915_gem_object_put(obj);
+	i915_gem_object_put(obj);
 err_vma:
 	eb->vma[i] = NULL;
 	return err;
@@ -1587,7 +1585,7 @@ static int eb_prefault_relocations(const struct i915_execbuffer *eb)
 	const unsigned int count = eb->buffer_count;
 	unsigned int i;
 
-	if (unlikely(i915.prefault_disable))
+	if (unlikely(i915_modparams.prefault_disable))
 		return 0;
 
 	for (i = 0; i < count; i++) {
@@ -2188,6 +2186,7 @@ i915_gem_do_execbuffer(struct drm_device *dev,
 	int out_fence_fd = -1;
 	int err;
 
+	BUILD_BUG_ON(__EXEC_INTERNAL_FLAGS & ~__I915_EXEC_ILLEGAL_FLAGS);
 	BUILD_BUG_ON(__EXEC_OBJECT_INTERNAL_FLAGS &
 		     ~__EXEC_OBJECT_UNKNOWN_FLAGS);
 

drivers/gpu/drm/i915/i915_gem_gtt.h

@@ -126,13 +126,13 @@ typedef u64 gen8_ppgtt_pml4e_t;
  * tables */
 #define GEN8_PDPE_MASK			0x1ff
 
-#define PPAT_UNCACHED_INDEX		(_PAGE_PWT | _PAGE_PCD)
-#define PPAT_CACHED_PDE_INDEX		0 /* WB LLC */
-#define PPAT_CACHED_INDEX		_PAGE_PAT /* WB LLCeLLC */
-#define PPAT_DISPLAY_ELLC_INDEX		_PAGE_PCD /* WT eLLC */
+#define PPAT_UNCACHED			(_PAGE_PWT | _PAGE_PCD)
+#define PPAT_CACHED_PDE			0 /* WB LLC */
+#define PPAT_CACHED			_PAGE_PAT /* WB LLCeLLC */
+#define PPAT_DISPLAY_ELLC		_PAGE_PCD /* WT eLLC */
 
 #define CHV_PPAT_SNOOP			(1<<6)
-#define GEN8_PPAT_AGE(x)		(x<<4)
+#define GEN8_PPAT_AGE(x)		((x)<<4)
 #define GEN8_PPAT_LLCeLLC		(3<<2)
 #define GEN8_PPAT_LLCELLC		(2<<2)
 #define GEN8_PPAT_LLC			(1<<2)
@@ -143,6 +143,11 @@ typedef u64 gen8_ppgtt_pml4e_t;
 #define GEN8_PPAT_ELLC_OVERRIDE		(0<<2)
 #define GEN8_PPAT(i, x)			((u64)(x) << ((i) * 8))
 
+#define GEN8_PPAT_GET_CA(x) ((x) & 3)
+#define GEN8_PPAT_GET_TC(x) ((x) & (3 << 2))
+#define GEN8_PPAT_GET_AGE(x) ((x) & (3 << 4))
+#define CHV_PPAT_GET_SNOOP(x) ((x) & (1 << 6))
+
 struct sg_table;
 
 struct intel_rotation_info {
@@ -536,6 +541,37 @@ i915_vm_to_ggtt(struct i915_address_space *vm)
 	return container_of(vm, struct i915_ggtt, base);
 }
 
+#define INTEL_MAX_PPAT_ENTRIES 8
+#define INTEL_PPAT_PERFECT_MATCH (~0U)
+
+struct intel_ppat;
+
+struct intel_ppat_entry {
+	struct intel_ppat *ppat;
+	struct kref ref;
+	u8 value;
+};
+
+struct intel_ppat {
+	struct intel_ppat_entry entries[INTEL_MAX_PPAT_ENTRIES];
+	DECLARE_BITMAP(used, INTEL_MAX_PPAT_ENTRIES);
+	DECLARE_BITMAP(dirty, INTEL_MAX_PPAT_ENTRIES);
+	unsigned int max_entries;
+	u8 clear_value;
+	/*
+	 * Return a score to show how two PPAT values match,
+	 * a INTEL_PPAT_PERFECT_MATCH indicates a perfect match
+	 */
+	unsigned int (*match)(u8 src, u8 dst);
+	void (*update_hw)(struct drm_i915_private *i915);
+
+	struct drm_i915_private *i915;
+};
+
+const struct intel_ppat_entry *
+intel_ppat_get(struct drm_i915_private *i915, u8 value);
+void intel_ppat_put(const struct intel_ppat_entry *entry);
+
 int i915_gem_init_aliasing_ppgtt(struct drm_i915_private *i915);
 void i915_gem_fini_aliasing_ppgtt(struct drm_i915_private *i915);
 

drivers/gpu/drm/i915/i915_gem_request.c

@@ -1021,12 +1021,28 @@ static bool busywait_stop(unsigned long timeout, unsigned int cpu)
 	return this_cpu != cpu;
 }
 
-bool __i915_spin_request(const struct drm_i915_gem_request *req,
-			 u32 seqno, int state, unsigned long timeout_us)
+static bool __i915_spin_request(const struct drm_i915_gem_request *req,
+				u32 seqno, int state, unsigned long timeout_us)
 {
 	struct intel_engine_cs *engine = req->engine;
 	unsigned int irq, cpu;
 
+	GEM_BUG_ON(!seqno);
+
+	/*
+	 * Only wait for the request if we know it is likely to complete.
+	 *
+	 * We don't track the timestamps around requests, nor the average
+	 * request length, so we do not have a good indicator that this
+	 * request will complete within the timeout. What we do know is the
+	 * order in which requests are executed by the engine and so we can
+	 * tell if the request has started. If the request hasn't started yet,
+	 * it is a fair assumption that it will not complete within our
+	 * relatively short timeout.
+	 */
+	if (!i915_seqno_passed(intel_engine_get_seqno(engine), seqno - 1))
+		return false;
+
 	/* When waiting for high frequency requests, e.g. during synchronous
 	 * rendering split between the CPU and GPU, the finite amount of time
 	 * required to set up the irq and wait upon it limits the response
@@ -1040,12 +1056,8 @@ bool __i915_spin_request(const struct drm_i915_gem_request *req,
 	irq = atomic_read(&engine->irq_count);
 	timeout_us += local_clock_us(&cpu);
 	do {
-		if (seqno != i915_gem_request_global_seqno(req))
-			break;
-
-		if (i915_seqno_passed(intel_engine_get_seqno(req->engine),
-				      seqno))
-			return true;
+		if (i915_seqno_passed(intel_engine_get_seqno(engine), seqno))
+			return seqno == i915_gem_request_global_seqno(req);
 
 		/* Seqno are meant to be ordered *before* the interrupt. If
 		 * we see an interrupt without a corresponding seqno advance,
@@ -1156,7 +1168,7 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 	GEM_BUG_ON(!i915_sw_fence_signaled(&req->submit));
 
 	/* Optimistic short spin before touching IRQs */
-	if (i915_spin_request(req, state, 5))
+	if (__i915_spin_request(req, wait.seqno, state, 5))
 		goto complete;
 
 	set_current_state(state);
@@ -1213,7 +1225,7 @@ long i915_wait_request(struct drm_i915_gem_request *req,
 			continue;
 
 		/* Only spin if we know the GPU is processing this request */
-		if (i915_spin_request(req, state, 2))
+		if (__i915_spin_request(req, wait.seqno, state, 2))
 			break;
 
 		if (!intel_wait_check_request(&wait, req)) {

drivers/gpu/drm/i915/i915_gem_request.h

@@ -312,26 +312,6 @@ static inline bool i915_seqno_passed(u32 seq1, u32 seq2)
 	return (s32)(seq1 - seq2) >= 0;
 }
 
-static inline bool
-__i915_gem_request_started(const struct drm_i915_gem_request *req, u32 seqno)
-{
-	GEM_BUG_ON(!seqno);
-	return i915_seqno_passed(intel_engine_get_seqno(req->engine),
-				 seqno - 1);
-}
-
-static inline bool
-i915_gem_request_started(const struct drm_i915_gem_request *req)
-{
-	u32 seqno;
-
-	seqno = i915_gem_request_global_seqno(req);
-	if (!seqno)
-		return false;
-
-	return __i915_gem_request_started(req, seqno);
-}
-
 static inline bool
 __i915_gem_request_completed(const struct drm_i915_gem_request *req, u32 seqno)
 {
@@ -352,21 +332,6 @@ i915_gem_request_completed(const struct drm_i915_gem_request *req)
 	return __i915_gem_request_completed(req, seqno);
 }
 
-bool __i915_spin_request(const struct drm_i915_gem_request *request,
-			 u32 seqno, int state, unsigned long timeout_us);
-static inline bool i915_spin_request(const struct drm_i915_gem_request *request,
-				     int state, unsigned long timeout_us)
-{
-	u32 seqno;
-
-	seqno = i915_gem_request_global_seqno(request);
-	if (!seqno)
-		return 0;
-
-	return (__i915_gem_request_started(request, seqno) &&
-		__i915_spin_request(request, seqno, state, timeout_us));
-}
-
 /* We treat requests as fences. This is not be to confused with our
  * "fence registers" but pipeline synchronisation objects ala GL_ARB_sync.
  * We use the fences to synchronize access from the CPU with activity on the

drivers/gpu/drm/i915/i915_gem_userptr.c

@@ -399,64 +399,42 @@ struct get_pages_work {
 	struct task_struct *task;
 };
 
-#if IS_ENABLED(CONFIG_SWIOTLB)
-#define swiotlb_active() swiotlb_nr_tbl()
-#else
-#define swiotlb_active() 0
-#endif
-
-static int
-st_set_pages(struct sg_table **st, struct page **pvec, int num_pages)
-{
-	struct scatterlist *sg;
-	int ret, n;
-
-	*st = kmalloc(sizeof(**st), GFP_KERNEL);
-	if (*st == NULL)
-		return -ENOMEM;
-
-	if (swiotlb_active()) {
-		ret = sg_alloc_table(*st, num_pages, GFP_KERNEL);
-		if (ret)
-			goto err;
-
-		for_each_sg((*st)->sgl, sg, num_pages, n)
-			sg_set_page(sg, pvec[n], PAGE_SIZE, 0);
-	} else {
-		ret = sg_alloc_table_from_pages(*st, pvec, num_pages,
-						0, num_pages << PAGE_SHIFT,
-						GFP_KERNEL);
-		if (ret)
-			goto err;
-	}
-
-	return 0;
-
-err:
-	kfree(*st);
-	*st = NULL;
-	return ret;
-}
-
 static struct sg_table *
-__i915_gem_userptr_set_pages(struct drm_i915_gem_object *obj,
-			     struct page **pvec, int num_pages)
+__i915_gem_userptr_alloc_pages(struct drm_i915_gem_object *obj,
+			       struct page **pvec, int num_pages)
 {
-	struct sg_table *pages;
+	unsigned int max_segment = i915_sg_segment_size();
+	struct sg_table *st;
 	int ret;
 
-	ret = st_set_pages(&pages, pvec, num_pages);
-	if (ret)
+	st = kmalloc(sizeof(*st), GFP_KERNEL);
+	if (!st)
+		return ERR_PTR(-ENOMEM);
+
+alloc_table:
+	ret = __sg_alloc_table_from_pages(st, pvec, num_pages,
+					  0, num_pages << PAGE_SHIFT,
+					  max_segment,
+					  GFP_KERNEL);
+	if (ret) {
+		kfree(st);
 		return ERR_PTR(ret);
+	}
 
-	ret = i915_gem_gtt_prepare_pages(obj, pages);
+	ret = i915_gem_gtt_prepare_pages(obj, st);
 	if (ret) {
-		sg_free_table(pages);
-		kfree(pages);
+		sg_free_table(st);
+
+		if (max_segment > PAGE_SIZE) {
+			max_segment = PAGE_SIZE;
+			goto alloc_table;
+		}
+
+		kfree(st);
 		return ERR_PTR(ret);
 	}
 
-	return pages;
+	return st;
 }
 
 static int
@@ -540,7 +518,8 @@ __i915_gem_userptr_get_pages_worker(struct work_struct *_work)
 		struct sg_table *pages = ERR_PTR(ret);
 
 		if (pinned == npages) {
-			pages = __i915_gem_userptr_set_pages(obj, pvec, npages);
+			pages = __i915_gem_userptr_alloc_pages(obj, pvec,
+							       npages);
 			if (!IS_ERR(pages)) {
 				__i915_gem_object_set_pages(obj, pages);
 				pinned = 0;
@@ -661,7 +640,7 @@ i915_gem_userptr_get_pages(struct drm_i915_gem_object *obj)
 		pages = __i915_gem_userptr_get_pages_schedule(obj);
 		active = pages == ERR_PTR(-EAGAIN);
 	} else {
-		pages = __i915_gem_userptr_set_pages(obj, pvec, num_pages);
+		pages = __i915_gem_userptr_alloc_pages(obj, pvec, num_pages);
 		active = !IS_ERR(pages);
 	}
 	if (active)
@@ -834,7 +813,9 @@ int i915_gem_init_userptr(struct drm_i915_private *dev_priv)
 	hash_init(dev_priv->mm_structs);
 
 	dev_priv->mm.userptr_wq =
-		alloc_workqueue("i915-userptr-acquire", WQ_HIGHPRI, 0);
+		alloc_workqueue("i915-userptr-acquire",
+				WQ_HIGHPRI | WQ_MEM_RECLAIM,
+				0);
 	if (!dev_priv->mm.userptr_wq)
 		return -ENOMEM;
 

drivers/gpu/drm/i915/i915_gpu_error.c

@@ -396,6 +396,8 @@ static void error_print_context(struct drm_i915_error_state_buf *m,
 static void error_print_engine(struct drm_i915_error_state_buf *m,
 			       const struct drm_i915_error_engine *ee)
 {
+	int n;
+
 	err_printf(m, "%s command stream:\n", engine_str(ee->engine_id));
 	err_printf(m, "  START: 0x%08x\n", ee->start);
 	err_printf(m, "  HEAD:  0x%08x [0x%08x]\n", ee->head, ee->rq_head);
@@ -465,8 +467,11 @@ static void error_print_engine(struct drm_i915_error_state_buf *m,
 		   jiffies_to_msecs(jiffies - ee->hangcheck_timestamp));
 	err_printf(m, "  engine reset count: %u\n", ee->reset_count);
 
-	error_print_request(m, "  ELSP[0]: ", &ee->execlist[0]);
-	error_print_request(m, "  ELSP[1]: ", &ee->execlist[1]);
+	for (n = 0; n < ee->num_ports; n++) {
+		err_printf(m, "  ELSP[%d]:", n);
+		error_print_request(m, " ", &ee->execlist[n]);
+	}
+
 	error_print_context(m, "  Active context: ", &ee->context);
 }
 
@@ -567,7 +572,7 @@ static __always_inline void err_print_param(struct drm_i915_error_state_buf *m,
 static void err_print_params(struct drm_i915_error_state_buf *m,
 			     const struct i915_params *p)
 {
-#define PRINT(T, x) err_print_param(m, #x, #T, &p->x);
+#define PRINT(T, x, ...) err_print_param(m, #x, #T, &p->x);
 	I915_PARAMS_FOR_EACH(PRINT);
 #undef PRINT
 }
@@ -861,7 +866,7 @@ void __i915_gpu_state_free(struct kref *error_ref)
 	kfree(error->overlay);
 	kfree(error->display);
 
-#define FREE(T, x) free_param(#T, &error->params.x);
+#define FREE(T, x, ...) free_param(#T, &error->params.x);
 	I915_PARAMS_FOR_EACH(FREE);
 #undef FREE
 
@@ -1327,17 +1332,19 @@ static void engine_record_requests(struct intel_engine_cs *engine,
 static void error_record_engine_execlists(struct intel_engine_cs *engine,
 					  struct drm_i915_error_engine *ee)
 {
-	const struct execlist_port *port = engine->execlist_port;
+	const struct intel_engine_execlists * const execlists = &engine->execlists;
 	unsigned int n;
 
-	for (n = 0; n < ARRAY_SIZE(engine->execlist_port); n++) {
-		struct drm_i915_gem_request *rq = port_request(&port[n]);
+	for (n = 0; n < execlists_num_ports(execlists); n++) {
+		struct drm_i915_gem_request *rq = port_request(&execlists->port[n]);
 
 		if (!rq)
 			break;
 
 		record_request(rq, &ee->execlist[n]);
 	}
+
+	ee->num_ports = n;
 }
 
 static void record_context(struct drm_i915_error_context *e,
@@ -1554,7 +1561,7 @@ static void i915_gem_capture_guc_log_buffer(struct drm_i915_private *dev_priv,
 					    struct i915_gpu_state *error)
 {
 	/* Capturing log buf contents won't be useful if logging was disabled */
-	if (!dev_priv->guc.log.vma || (i915.guc_log_level < 0))
+	if (!dev_priv->guc.log.vma || (i915_modparams.guc_log_level < 0))
 		return;
 
 	error->guc_log = i915_error_object_create(dev_priv,
@@ -1696,8 +1703,8 @@ static int capture(void *data)
 		ktime_to_timeval(ktime_sub(ktime_get(),
 					   error->i915->gt.last_init_time));
 
-	error->params = i915;
-#define DUP(T, x) dup_param(#T, &error->params.x);
+	error->params = i915_modparams;
+#define DUP(T, x, ...) dup_param(#T, &error->params.x);
 	I915_PARAMS_FOR_EACH(DUP);
 #undef DUP
 
@@ -1751,7 +1758,7 @@ void i915_capture_error_state(struct drm_i915_private *dev_priv,
 	struct i915_gpu_state *error;
 	unsigned long flags;
 
-	if (!i915.error_capture)
+	if (!i915_modparams.error_capture)
 		return;
 
 	if (READ_ONCE(dev_priv->gpu_error.first_error))

drivers/gpu/drm/i915/i915_oa_cflgt2.c

+/*
+ * Autogenerated file by GPU Top : https://github.com/rib/gputop
+ * DO NOT EDIT manually!
+ *
+ *
+ * Copyright (c) 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/sysfs.h>
+
+#include "i915_drv.h"
+#include "i915_oa_cflgt2.h"
+
+static const struct i915_oa_reg b_counter_config_test_oa[] = {
+	{ _MMIO(0x2740), 0x00000000 },
+	{ _MMIO(0x2744), 0x00800000 },
+	{ _MMIO(0x2714), 0xf0800000 },
+	{ _MMIO(0x2710), 0x00000000 },
+	{ _MMIO(0x2724), 0xf0800000 },
+	{ _MMIO(0x2720), 0x00000000 },
+	{ _MMIO(0x2770), 0x00000004 },
+	{ _MMIO(0x2774), 0x00000000 },
+	{ _MMIO(0x2778), 0x00000003 },
+	{ _MMIO(0x277c), 0x00000000 },
+	{ _MMIO(0x2780), 0x00000007 },
+	{ _MMIO(0x2784), 0x00000000 },
+	{ _MMIO(0x2788), 0x00100002 },
+	{ _MMIO(0x278c), 0x0000fff7 },
+	{ _MMIO(0x2790), 0x00100002 },
+	{ _MMIO(0x2794), 0x0000ffcf },
+	{ _MMIO(0x2798), 0x00100082 },
+	{ _MMIO(0x279c), 0x0000ffef },
+	{ _MMIO(0x27a0), 0x001000c2 },
+	{ _MMIO(0x27a4), 0x0000ffe7 },
+	{ _MMIO(0x27a8), 0x00100001 },
+	{ _MMIO(0x27ac), 0x0000ffe7 },
+};
+
+static const struct i915_oa_reg flex_eu_config_test_oa[] = {
+};
+
+static const struct i915_oa_reg mux_config_test_oa[] = {
+	{ _MMIO(0x9840), 0x00000080 },
+	{ _MMIO(0x9888), 0x11810000 },
+	{ _MMIO(0x9888), 0x07810013 },
+	{ _MMIO(0x9888), 0x1f810000 },
+	{ _MMIO(0x9888), 0x1d810000 },
+	{ _MMIO(0x9888), 0x1b930040 },
+	{ _MMIO(0x9888), 0x07e54000 },
+	{ _MMIO(0x9888), 0x1f908000 },
+	{ _MMIO(0x9888), 0x11900000 },
+	{ _MMIO(0x9888), 0x37900000 },
+	{ _MMIO(0x9888), 0x53900000 },
+	{ _MMIO(0x9888), 0x45900000 },
+	{ _MMIO(0x9888), 0x33900000 },
+};
+
+static ssize_t
+show_test_oa_id(struct device *kdev, struct device_attribute *attr, char *buf)
+{
+	return sprintf(buf, "1\n");
+}
+
+void
+i915_perf_load_test_config_cflgt2(struct drm_i915_private *dev_priv)
+{
+	strncpy(dev_priv->perf.oa.test_config.uuid,
+		"74fb4902-d3d3-4237-9e90-cbdc68d0a446",
+		UUID_STRING_LEN);
+	dev_priv->perf.oa.test_config.id = 1;
+
+	dev_priv->perf.oa.test_config.mux_regs = mux_config_test_oa;
+	dev_priv->perf.oa.test_config.mux_regs_len = ARRAY_SIZE(mux_config_test_oa);
+
+	dev_priv->perf.oa.test_config.b_counter_regs = b_counter_config_test_oa;
+	dev_priv->perf.oa.test_config.b_counter_regs_len = ARRAY_SIZE(b_counter_config_test_oa);
+
+	dev_priv->perf.oa.test_config.flex_regs = flex_eu_config_test_oa;
+	dev_priv->perf.oa.test_config.flex_regs_len = ARRAY_SIZE(flex_eu_config_test_oa);
+
+	dev_priv->perf.oa.test_config.sysfs_metric.name = "74fb4902-d3d3-4237-9e90-cbdc68d0a446";
+	dev_priv->perf.oa.test_config.sysfs_metric.attrs = dev_priv->perf.oa.test_config.attrs;
+
+	dev_priv->perf.oa.test_config.attrs[0] = &dev_priv->perf.oa.test_config.sysfs_metric_id.attr;
+
+	dev_priv->perf.oa.test_config.sysfs_metric_id.attr.name = "id";
+	dev_priv->perf.oa.test_config.sysfs_metric_id.attr.mode = 0444;
+	dev_priv->perf.oa.test_config.sysfs_metric_id.show = show_test_oa_id;
+}

drivers/gpu/drm/i915/i915_oa_cflgt2.h

+/*
+ * Autogenerated file by GPU Top : https://github.com/rib/gputop
+ * DO NOT EDIT manually!
+ *
+ *
+ * Copyright (c) 2015 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __I915_OA_CFLGT2_H__
+#define __I915_OA_CFLGT2_H__
+
+extern void i915_perf_load_test_config_cflgt2(struct drm_i915_private *dev_priv);
+
+#endif