1 files changed, 100 insertions, 57 deletions

diff --git a/drivers/gpu/drm/msm/msm_gem_submit.c b/drivers/gpu/drm/msm/msm_gem_submit.c
index aa5c60a7132d..d04c349d8112 100644
--- a/drivers/gpu/drm/msm/msm_gem_submit.c
+++ b/drivers/gpu/drm/msm/msm_gem_submit.c
@@ -42,6 +42,7 @@ static struct msm_gem_submit *submit_create(struct drm_device *dev,
 	if (!submit)
 		return NULL;
 
+	kref_init(&submit->ref);
 	submit->dev = dev;
 	submit->aspace = queue->ctx->aspace;
 	submit->gpu = gpu;
@@ -60,13 +61,19 @@ static struct msm_gem_submit *submit_create(struct drm_device *dev,
 	return submit;
 }
 
-void msm_gem_submit_free(struct msm_gem_submit *submit)
+void __msm_gem_submit_destroy(struct kref *kref)
 {
+	struct msm_gem_submit *submit =
+			container_of(kref, struct msm_gem_submit, ref);
+	unsigned i;
+
 	dma_fence_put(submit->fence);
-	list_del(&submit->node);
 	put_pid(submit->pid);
 	msm_submitqueue_put(submit->queue);
 
+	for (i = 0; i < submit->nr_cmds; i++)
+		kfree(submit->cmd[i].relocs);
+
 	kfree(submit);
 }
 
@@ -150,13 +157,73 @@ out:
 	return ret;
 }
 
+static int submit_lookup_cmds(struct msm_gem_submit *submit,
+		struct drm_msm_gem_submit *args, struct drm_file *file)
+{
+	unsigned i, sz;
+	int ret = 0;
+
+	for (i = 0; i < args->nr_cmds; i++) {
+		struct drm_msm_gem_submit_cmd submit_cmd;
+		void __user *userptr =
+			u64_to_user_ptr(args->cmds + (i * sizeof(submit_cmd)));
+
+		ret = copy_from_user(&submit_cmd, userptr, sizeof(submit_cmd));
+		if (ret) {
+			ret = -EFAULT;
+			goto out;
+		}
+
+		/* validate input from userspace: */
+		switch (submit_cmd.type) {
+		case MSM_SUBMIT_CMD_BUF:
+		case MSM_SUBMIT_CMD_IB_TARGET_BUF:
+		case MSM_SUBMIT_CMD_CTX_RESTORE_BUF:
+			break;
+		default:
+			DRM_ERROR("invalid type: %08x\n", submit_cmd.type);
+			return -EINVAL;
+		}
+
+		if (submit_cmd.size % 4) {
+			DRM_ERROR("non-aligned cmdstream buffer size: %u\n",
+					submit_cmd.size);
+			ret = -EINVAL;
+			goto out;
+		}
+
+		submit->cmd[i].type = submit_cmd.type;
+		submit->cmd[i].size = submit_cmd.size / 4;
+		submit->cmd[i].offset = submit_cmd.submit_offset / 4;
+		submit->cmd[i].idx  = submit_cmd.submit_idx;
+		submit->cmd[i].nr_relocs = submit_cmd.nr_relocs;
+
+		sz = array_size(submit_cmd.nr_relocs,
+				sizeof(struct drm_msm_gem_submit_reloc));
+		/* check for overflow: */
+		if (sz == SIZE_MAX) {
+			ret = -ENOMEM;
+			goto out;
+		}
+		submit->cmd[i].relocs = kmalloc(sz, GFP_KERNEL);
+		ret = copy_from_user(submit->cmd[i].relocs, userptr, sz);
+		if (ret) {
+			ret = -EFAULT;
+			goto out;
+		}
+	}
+
+out:
+	return ret;
+}
+
 static void submit_unlock_unpin_bo(struct msm_gem_submit *submit,
 		int i, bool backoff)
 {
 	struct msm_gem_object *msm_obj = submit->bos[i].obj;
 
 	if (submit->bos[i].flags & BO_PINNED)
-		msm_gem_unpin_iova(&msm_obj->base, submit->aspace);
+		msm_gem_unpin_iova_locked(&msm_obj->base, submit->aspace);
 
 	if (submit->bos[i].flags & BO_LOCKED)
 		dma_resv_unlock(msm_obj->base.resv);
@@ -259,7 +326,7 @@ static int submit_pin_objects(struct msm_gem_submit *submit)
 		uint64_t iova;
 
 		/* if locking succeeded, pin bo: */
-		ret = msm_gem_get_and_pin_iova(&msm_obj->base,
+		ret = msm_gem_get_and_pin_iova_locked(&msm_obj->base,
 				submit->aspace, &iova);
 
 		if (ret)
@@ -301,7 +368,7 @@ static int submit_bo(struct msm_gem_submit *submit, uint32_t idx,
 
 /* process the reloc's and patch up the cmdstream as needed: */
 static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *obj,
-		uint32_t offset, uint32_t nr_relocs, uint64_t relocs)
+		uint32_t offset, uint32_t nr_relocs, struct drm_msm_gem_submit_reloc *relocs)
 {
 	uint32_t i, last_offset = 0;
 	uint32_t *ptr;
@@ -318,7 +385,7 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob
 	/* For now, just map the entire thing.  Eventually we probably
 	 * to do it page-by-page, w/ kmap() if not vmap()d..
 	 */
-	ptr = msm_gem_get_vaddr(&obj->base);
+	ptr = msm_gem_get_vaddr_locked(&obj->base);
 
 	if (IS_ERR(ptr)) {
 		ret = PTR_ERR(ptr);
@@ -327,18 +394,11 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob
 	}
 
 	for (i = 0; i < nr_relocs; i++) {
-		struct drm_msm_gem_submit_reloc submit_reloc;
-		void __user *userptr =
-			u64_to_user_ptr(relocs + (i * sizeof(submit_reloc)));
+		struct drm_msm_gem_submit_reloc submit_reloc = relocs[i];
 		uint32_t off;
 		uint64_t iova;
 		bool valid;
 
-		if (copy_from_user(&submit_reloc, userptr, sizeof(submit_reloc))) {
-			ret = -EFAULT;
-			goto out;
-		}
-
 		if (submit_reloc.submit_offset % 4) {
 			DRM_ERROR("non-aligned reloc offset: %u\n",
 					submit_reloc.submit_offset);
@@ -376,7 +436,7 @@ static int submit_reloc(struct msm_gem_submit *submit, struct msm_gem_object *ob
 	}
 
 out:
-	msm_gem_put_vaddr(&obj->base);
+	msm_gem_put_vaddr_locked(&obj->base);
 
 	return ret;
 }
@@ -692,7 +752,20 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data,
 
 	ret = submit_lookup_objects(submit, args, file);
 	if (ret)
-		goto out;
+		goto out_pre_pm;
+
+	ret = submit_lookup_cmds(submit, args, file);
+	if (ret)
+		goto out_pre_pm;
+
+	/*
+	 * Thanks to dev_pm_opp opp_table_lock interactions with mm->mmap_sem
+	 * in the resume path, we need to to rpm get before we lock objs.
+	 * Which unfortunately might involve powering up the GPU sooner than
+	 * is necessary.  But at least in the explicit fencing case, we will
+	 * have already done all the fence waiting.
+	 */
+	pm_runtime_get_sync(&gpu->pdev->dev);
 
 	/* copy_*_user while holding a ww ticket upsets lockdep */
 	ww_acquire_init(&submit->ticket, &reservation_ww_class);
@@ -710,60 +783,29 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data,
 		goto out;
 
 	for (i = 0; i < args->nr_cmds; i++) {
-		struct drm_msm_gem_submit_cmd submit_cmd;
-		void __user *userptr =
-			u64_to_user_ptr(args->cmds + (i * sizeof(submit_cmd)));
 		struct msm_gem_object *msm_obj;
 		uint64_t iova;
 
-		ret = copy_from_user(&submit_cmd, userptr, sizeof(submit_cmd));
-		if (ret) {
-			ret = -EFAULT;
-			goto out;
-		}
-
-		/* validate input from userspace: */
-		switch (submit_cmd.type) {
-		case MSM_SUBMIT_CMD_BUF:
-		case MSM_SUBMIT_CMD_IB_TARGET_BUF:
-		case MSM_SUBMIT_CMD_CTX_RESTORE_BUF:
-			break;
-		default:
-			DRM_ERROR("invalid type: %08x\n", submit_cmd.type);
-			ret = -EINVAL;
-			goto out;
-		}
-
-		ret = submit_bo(submit, submit_cmd.submit_idx,
+		ret = submit_bo(submit, submit->cmd[i].idx,
 				&msm_obj, &iova, NULL);
 		if (ret)
 			goto out;
 
-		if (submit_cmd.size % 4) {
-			DRM_ERROR("non-aligned cmdstream buffer size: %u\n",
-					submit_cmd.size);
+		if (!submit->cmd[i].size ||
+			((submit->cmd[i].size + submit->cmd[i].offset) >
+				msm_obj->base.size / 4)) {
+			DRM_ERROR("invalid cmdstream size: %u\n", submit->cmd[i].size * 4);
 			ret = -EINVAL;
 			goto out;
 		}
 
-		if (!submit_cmd.size ||
-			((submit_cmd.size + submit_cmd.submit_offset) >
-				msm_obj->base.size)) {
-			DRM_ERROR("invalid cmdstream size: %u\n", submit_cmd.size);
-			ret = -EINVAL;
-			goto out;
-		}
-
-		submit->cmd[i].type = submit_cmd.type;
-		submit->cmd[i].size = submit_cmd.size / 4;
-		submit->cmd[i].iova = iova + submit_cmd.submit_offset;
-		submit->cmd[i].idx  = submit_cmd.submit_idx;
+		submit->cmd[i].iova = iova + (submit->cmd[i].offset * 4);
 
 		if (submit->valid)
 			continue;
 
-		ret = submit_reloc(submit, msm_obj, submit_cmd.submit_offset,
-				submit_cmd.nr_relocs, submit_cmd.relocs);
+		ret = submit_reloc(submit, msm_obj, submit->cmd[i].offset * 4,
+				submit->cmd[i].nr_relocs, submit->cmd[i].relocs);
 		if (ret)
 			goto out;
 	}
@@ -800,11 +842,12 @@ int msm_ioctl_gem_submit(struct drm_device *dev, void *data,
 
 
 out:
+	pm_runtime_put(&gpu->pdev->dev);
+out_pre_pm:
 	submit_cleanup(submit);
 	if (has_ww_ticket)
 		ww_acquire_fini(&submit->ticket);
-	if (ret)
-		msm_gem_submit_free(submit);
+	msm_gem_submit_put(submit);
 out_unlock:
 	if (ret && (out_fence_fd >= 0))
 		put_unused_fd(out_fence_fd);