Merge tag 'drm-fixes-2022-05-21' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Few final fixes for 5.18, one amdgpu, core dp mst leak fix, dma-buf
  two fixes, and i915 has a few fixes, one for a regression on older
  GM45 chipsets,

  dma-buf:
   - ioctl userspace use fix
   - fix dma-buf sysfs name generation

  core:
   - dp/mst leak fix

  amdgpu:
   - suspend/resume regression fix

  i915:
   - fix for #5806: GPU hangs and display artifacts on Intel GM45
   - reject DMC with out-of-spec MMIO
   - correctly mark guilty contexts on GuC reset"

* tag 'drm-fixes-2022-05-21' of git://anongit.freedesktop.org/drm/drm:
  drm/i915: Use i915_gem_object_ggtt_pin_ww for reloc_iomap
  drm/amd: Don't reset dGPUs if the system is going to s2idle
  drm/dp/mst: fix a possible memory leak in fetch_monitor_name()
  dma-buf: fix use of DMA_BUF_SET_NAME_{A,B} in userspace
  i915/guc/reset: Make __guc_reset_context aware of guilty engines
  drm/i915/dmc: Add MMIO range restrictions
  dma-buf: ensure unique directory name for dmabuf stats

Author: torvalds

drivers/dma-buf/dma-buf.c

@@ -407,6 +407,7 @@ static inline int is_dma_buf_file(struct file *file)
 
 static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags)
 {
+	static atomic64_t dmabuf_inode = ATOMIC64_INIT(0);
 	struct file *file;
 	struct inode *inode = alloc_anon_inode(dma_buf_mnt->mnt_sb);
 
@@ -416,6 +417,13 @@ static struct file *dma_buf_getfile(struct dma_buf *dmabuf, int flags)
 	inode->i_size = dmabuf->size;
 	inode_set_bytes(inode, dmabuf->size);
 
+	/*
+	 * The ->i_ino acquired from get_next_ino() is not unique thus
+	 * not suitable for using it as dentry name by dmabuf stats.
+	 * Override ->i_ino with the unique and dmabuffs specific
+	 * value.
+	 */
+	inode->i_ino = atomic64_add_return(1, &dmabuf_inode);
 	file = alloc_file_pseudo(inode, dma_buf_mnt, "dmabuf",
 				 flags, &dma_buf_fops);
 	if (IS_ERR(file))

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -1342,9 +1342,11 @@ static inline int amdgpu_acpi_smart_shift_update(struct drm_device *dev,
 
 #if defined(CONFIG_ACPI) && defined(CONFIG_SUSPEND)
 bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev);
+bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev);
 bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev);
 #else
 static inline bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device *adev) { return false; }
+static inline bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev) { return false; }
 static inline bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev) { return false; }
 #endif
 

drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c

@@ -1045,6 +1045,20 @@ bool amdgpu_acpi_is_s3_active(struct amdgpu_device *adev)
 		(pm_suspend_target_state == PM_SUSPEND_MEM);
 }
 
+/**
+ * amdgpu_acpi_should_gpu_reset
+ *
+ * @adev: amdgpu_device_pointer
+ *
+ * returns true if should reset GPU, false if not
+ */
+bool amdgpu_acpi_should_gpu_reset(struct amdgpu_device *adev)
+{
+	if (adev->flags & AMD_IS_APU)
+		return false;
+	return pm_suspend_target_state != PM_SUSPEND_TO_IDLE;
+}
+
 /**
  * amdgpu_acpi_is_s0ix_active
  *

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -2336,7 +2336,7 @@ static int amdgpu_pmops_suspend_noirq(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 
-	if (!adev->in_s0ix)
+	if (amdgpu_acpi_should_gpu_reset(adev))
 		return amdgpu_asic_reset(adev);
 
 	return 0;

drivers/gpu/drm/dp/drm_dp_mst_topology.c

@@ -4852,6 +4852,7 @@ static void fetch_monitor_name(struct drm_dp_mst_topology_mgr *mgr,
 
 	mst_edid = drm_dp_mst_get_edid(port->connector, mgr, port);
 	drm_edid_get_monitor_name(mst_edid, name, namelen);
+	kfree(mst_edid);
 }
 
 /**

drivers/gpu/drm/i915/display/intel_dmc.c

@@ -367,6 +367,44 @@ static void dmc_set_fw_offset(struct intel_dmc *dmc,
 	}
 }
 
+static bool dmc_mmio_addr_sanity_check(struct intel_dmc *dmc,
+				       const u32 *mmioaddr, u32 mmio_count,
+				       int header_ver, u8 dmc_id)
+{
+	struct drm_i915_private *i915 = container_of(dmc, typeof(*i915), dmc);
+	u32 start_range, end_range;
+	int i;
+
+	if (dmc_id >= DMC_FW_MAX) {
+		drm_warn(&i915->drm, "Unsupported firmware id %u\n", dmc_id);
+		return false;
+	}
+
+	if (header_ver == 1) {
+		start_range = DMC_MMIO_START_RANGE;
+		end_range = DMC_MMIO_END_RANGE;
+	} else if (dmc_id == DMC_FW_MAIN) {
+		start_range = TGL_MAIN_MMIO_START;
+		end_range = TGL_MAIN_MMIO_END;
+	} else if (DISPLAY_VER(i915) >= 13) {
+		start_range = ADLP_PIPE_MMIO_START;
+		end_range = ADLP_PIPE_MMIO_END;
+	} else if (DISPLAY_VER(i915) >= 12) {
+		start_range = TGL_PIPE_MMIO_START(dmc_id);
+		end_range = TGL_PIPE_MMIO_END(dmc_id);
+	} else {
+		drm_warn(&i915->drm, "Unknown mmio range for sanity check");
+		return false;
+	}
+
+	for (i = 0; i < mmio_count; i++) {
+		if (mmioaddr[i] < start_range || mmioaddr[i] > end_range)
+			return false;
+	}
+
+	return true;
+}
+
 static u32 parse_dmc_fw_header(struct intel_dmc *dmc,
 			       const struct intel_dmc_header_base *dmc_header,
 			       size_t rem_size, u8 dmc_id)
@@ -436,6 +474,12 @@ static u32 parse_dmc_fw_header(struct intel_dmc *dmc,
 		return 0;
 	}
 
+	if (!dmc_mmio_addr_sanity_check(dmc, mmioaddr, mmio_count,
+					dmc_header->header_ver, dmc_id)) {
+		drm_err(&i915->drm, "DMC firmware has Wrong MMIO Addresses\n");
+		return 0;
+	}
+
 	for (i = 0; i < mmio_count; i++) {
 		dmc_info->mmioaddr[i] = _MMIO(mmioaddr[i]);
 		dmc_info->mmiodata[i] = mmiodata[i];

drivers/gpu/drm/i915/gem/i915_gem_execbuffer.c

@@ -1252,14 +1252,12 @@ static void *reloc_iomap(struct i915_vma *batch,
 		 * Only attempt to pin the batch buffer to ggtt if the current batch
 		 * is not inside ggtt, or the batch buffer is not misplaced.
 		 */
-		if (!i915_is_ggtt(batch->vm)) {
+		if (!i915_is_ggtt(batch->vm) ||
+		    !i915_vma_misplaced(batch, 0, 0, PIN_MAPPABLE)) {
 			vma = i915_gem_object_ggtt_pin_ww(obj, &eb->ww, NULL, 0, 0,
 							  PIN_MAPPABLE |
 							  PIN_NONBLOCK /* NOWARN */ |
 							  PIN_NOEVICT);
-		} else if (i915_vma_is_map_and_fenceable(batch)) {
-			__i915_vma_pin(batch);
-			vma = batch;
 		}
 
 		if (vma == ERR_PTR(-EDEADLK))

drivers/gpu/drm/i915/gt/intel_reset.c

@@ -806,7 +806,7 @@ static int gt_reset(struct intel_gt *gt, intel_engine_mask_t stalled_mask)
 		__intel_engine_reset(engine, stalled_mask & engine->mask);
 	local_bh_enable();
 
-	intel_uc_reset(&gt->uc, true);
+	intel_uc_reset(&gt->uc, ALL_ENGINES);
 
 	intel_ggtt_restore_fences(gt->ggtt);
 

drivers/gpu/drm/i915/gt/uc/intel_guc.h

@@ -438,7 +438,7 @@ int intel_guc_global_policies_update(struct intel_guc *guc);
 void intel_guc_context_ban(struct intel_context *ce, struct i915_request *rq);
 
 void intel_guc_submission_reset_prepare(struct intel_guc *guc);
-void intel_guc_submission_reset(struct intel_guc *guc, bool stalled);
+void intel_guc_submission_reset(struct intel_guc *guc, intel_engine_mask_t stalled);
 void intel_guc_submission_reset_finish(struct intel_guc *guc);
 void intel_guc_submission_cancel_requests(struct intel_guc *guc);
 

drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c

@@ -1590,9 +1590,9 @@ __unwind_incomplete_requests(struct intel_context *ce)
 	spin_unlock_irqrestore(&sched_engine->lock, flags);
 }
 
-static void __guc_reset_context(struct intel_context *ce, bool stalled)
+static void __guc_reset_context(struct intel_context *ce, intel_engine_mask_t stalled)
 {
-	bool local_stalled;
+	bool guilty;
 	struct i915_request *rq;
 	unsigned long flags;
 	u32 head;
@@ -1620,22 +1620,22 @@ static void __guc_reset_context(struct intel_context *ce, bool stalled)
 		if (!intel_context_is_pinned(ce))
 			goto next_context;
 
-		local_stalled = false;
+		guilty = false;
 		rq = intel_context_find_active_request(ce);
 		if (!rq) {
 			head = ce->ring->tail;
 			goto out_replay;
 		}
 
 		if (i915_request_started(rq))
-			local_stalled = true;
+			guilty = stalled & ce->engine->mask;
 
 		GEM_BUG_ON(i915_active_is_idle(&ce->active));
 		head = intel_ring_wrap(ce->ring, rq->head);
 
-		__i915_request_reset(rq, local_stalled && stalled);
+		__i915_request_reset(rq, guilty);
 out_replay:
-		guc_reset_state(ce, head, local_stalled && stalled);
+		guc_reset_state(ce, head, guilty);
 next_context:
 		if (i != number_children)
 			ce = list_next_entry(ce, parallel.child_link);
@@ -1645,7 +1645,7 @@ static void __guc_reset_context(struct intel_context *ce, bool stalled)
 	intel_context_put(parent);
 }
 
-void intel_guc_submission_reset(struct intel_guc *guc, bool stalled)
+void intel_guc_submission_reset(struct intel_guc *guc, intel_engine_mask_t stalled)
 {
 	struct intel_context *ce;
 	unsigned long index;
@@ -4013,7 +4013,7 @@ static void guc_context_replay(struct intel_context *ce)
 {
 	struct i915_sched_engine *sched_engine = ce->engine->sched_engine;
 
-	__guc_reset_context(ce, true);
+	__guc_reset_context(ce, ce->engine->mask);
 	tasklet_hi_schedule(&sched_engine->tasklet);
 }