From d3f139c46280035509600900a6ead3d7e680218f Mon Sep 17 00:00:00 2001 From: Ofir Bitton Date: Wed, 18 Nov 2020 15:46:57 +0200 Subject: [PATCH] habanalabs: add driver support for internal cb scheduling In order to support scnenarios in which driver needs access to HW components but it cannot access them directly, we add support for scheduling command buffers internally. These command buffers will be transmitted upon next user command submission context. Signed-off-by: Ofir Bitton Reviewed-by: Oded Gabbay Signed-off-by: Oded Gabbay --- .../habanalabs/common/command_submission.c | 139 ++++++++++++++++++ drivers/misc/habanalabs/common/context.c | 8 + drivers/misc/habanalabs/common/habanalabs.h | 26 ++++ 3 files changed, 173 insertions(+) diff --git a/drivers/misc/habanalabs/common/command_submission.c b/drivers/misc/habanalabs/common/command_submission.c index 3affb350070c..a5e9bb0a4855 100644 --- a/drivers/misc/habanalabs/common/command_submission.c +++ b/drivers/misc/habanalabs/common/command_submission.c @@ -585,6 +585,18 @@ void hl_cs_rollback_all(struct hl_device *hdev) } } +void hl_pending_cb_list_flush(struct hl_ctx *ctx) +{ + struct hl_pending_cb *pending_cb, *tmp; + + list_for_each_entry_safe(pending_cb, tmp, + &ctx->pending_cb_list, cb_node) { + list_del(&pending_cb->cb_node); + hl_cb_put(pending_cb->cb); + kfree(pending_cb); + } +} + static void job_wq_completion(struct work_struct *work) { struct hl_cs_job *job = container_of(work, struct hl_cs_job, @@ -954,6 +966,129 @@ out: return rc; } +static int pending_cb_create_job(struct hl_device *hdev, struct hl_ctx *ctx, + struct hl_cs *cs, struct hl_cb *cb, u32 size, u32 hw_queue_id) +{ + struct hw_queue_properties *hw_queue_prop; + struct hl_cs_counters_atomic *cntr; + struct hl_cs_job *job; + + hw_queue_prop = &hdev->asic_prop.hw_queues_props[hw_queue_id]; + cntr = &hdev->aggregated_cs_counters; + + job = hl_cs_allocate_job(hdev, hw_queue_prop->type, true); + if (!job) { + atomic64_inc(&ctx->cs_counters.out_of_mem_drop_cnt); + atomic64_inc(&cntr->out_of_mem_drop_cnt); + dev_err(hdev->dev, "Failed to allocate a new job\n"); + return -ENOMEM; + } + + job->id = 0; + job->cs = cs; + job->user_cb = cb; + atomic_inc(&job->user_cb->cs_cnt); + job->user_cb_size = size; + job->hw_queue_id = hw_queue_id; + job->patched_cb = job->user_cb; + job->job_cb_size = job->user_cb_size; + + /* increment refcount as for external queues we get completion */ + cs_get(cs); + + cs->jobs_in_queue_cnt[job->hw_queue_id]++; + + list_add_tail(&job->cs_node, &cs->job_list); + + hl_debugfs_add_job(hdev, job); + + return 0; +} + +static int hl_submit_pending_cb(struct hl_fpriv *hpriv) +{ + struct hl_device *hdev = hpriv->hdev; + struct hl_ctx *ctx = hpriv->ctx; + struct hl_pending_cb *pending_cb, *tmp; + struct list_head local_cb_list; + struct hl_cs *cs; + struct hl_cb *cb; + u32 hw_queue_id; + u32 cb_size; + int process_list, rc = 0; + + if (list_empty(&ctx->pending_cb_list)) + return 0; + + process_list = atomic_cmpxchg(&ctx->thread_pending_cb_token, 1, 0); + + /* Only a single thread is allowed to process the list */ + if (!process_list) + return 0; + + if (list_empty(&ctx->pending_cb_list)) + goto free_pending_cb_token; + + /* move all list elements to a local list */ + INIT_LIST_HEAD(&local_cb_list); + spin_lock(&ctx->pending_cb_lock); + list_for_each_entry_safe(pending_cb, tmp, &ctx->pending_cb_list, + cb_node) + list_move_tail(&pending_cb->cb_node, &local_cb_list); + spin_unlock(&ctx->pending_cb_lock); + + rc = allocate_cs(hdev, ctx, CS_TYPE_DEFAULT, &cs); + if (rc) + goto add_list_elements; + + hl_debugfs_add_cs(cs); + + /* Iterate through pending cb list, create jobs and add to CS */ + list_for_each_entry(pending_cb, &local_cb_list, cb_node) { + cb = pending_cb->cb; + cb_size = pending_cb->cb_size; + hw_queue_id = pending_cb->hw_queue_id; + + rc = pending_cb_create_job(hdev, ctx, cs, cb, cb_size, + hw_queue_id); + if (rc) + goto free_cs_object; + } + + rc = hl_hw_queue_schedule_cs(cs); + if (rc) { + if (rc != -EAGAIN) + dev_err(hdev->dev, + "Failed to submit CS %d.%llu (%d)\n", + ctx->asid, cs->sequence, rc); + goto free_cs_object; + } + + /* pending cb was scheduled successfully */ + list_for_each_entry_safe(pending_cb, tmp, &local_cb_list, cb_node) { + list_del(&pending_cb->cb_node); + kfree(pending_cb); + } + + cs_put(cs); + + goto free_pending_cb_token; + +free_cs_object: + cs_rollback(hdev, cs); + cs_put(cs); +add_list_elements: + spin_lock(&ctx->pending_cb_lock); + list_for_each_entry_safe_reverse(pending_cb, tmp, &local_cb_list, + cb_node) + list_move(&pending_cb->cb_node, &ctx->pending_cb_list); + spin_unlock(&ctx->pending_cb_lock); +free_pending_cb_token: + atomic_set(&ctx->thread_pending_cb_token, 1); + + return rc; +} + static int hl_cs_ctx_switch(struct hl_fpriv *hpriv, union hl_cs_args *args, u64 *cs_seq) { @@ -1353,6 +1488,10 @@ int hl_cs_ioctl(struct hl_fpriv *hpriv, void *data) if (rc) goto out; + rc = hl_submit_pending_cb(hpriv); + if (rc) + goto out; + cs_type = hl_cs_get_cs_type(args->in.cs_flags & ~HL_CS_FLAGS_FORCE_RESTORE); chunks = (void __user *) (uintptr_t) args->in.chunks_execute; diff --git a/drivers/misc/habanalabs/common/context.c b/drivers/misc/habanalabs/common/context.c index 3d86b83f4ca6..829fe98eed61 100644 --- a/drivers/misc/habanalabs/common/context.c +++ b/drivers/misc/habanalabs/common/context.c @@ -15,6 +15,11 @@ static void hl_ctx_fini(struct hl_ctx *ctx) u64 idle_mask = 0; int i; + /* Release all allocated pending cb's, those cb's were never + * scheduled so it is safe to release them here + */ + hl_pending_cb_list_flush(ctx); + /* * If we arrived here, there are no jobs waiting for this context * on its queues so we can safely remove it. @@ -142,8 +147,11 @@ int hl_ctx_init(struct hl_device *hdev, struct hl_ctx *ctx, bool is_kernel_ctx) kref_init(&ctx->refcount); ctx->cs_sequence = 1; + INIT_LIST_HEAD(&ctx->pending_cb_list); + spin_lock_init(&ctx->pending_cb_lock); spin_lock_init(&ctx->cs_lock); atomic_set(&ctx->thread_ctx_switch_token, 1); + atomic_set(&ctx->thread_pending_cb_token, 1); ctx->thread_ctx_switch_wait_token = 0; ctx->cs_pending = kcalloc(hdev->asic_prop.max_pending_cs, sizeof(struct hl_fence *), diff --git a/drivers/misc/habanalabs/common/habanalabs.h b/drivers/misc/habanalabs/common/habanalabs.h index eb43fb3065ce..8e0553bf3e0e 100644 --- a/drivers/misc/habanalabs/common/habanalabs.h +++ b/drivers/misc/habanalabs/common/habanalabs.h @@ -1016,6 +1016,20 @@ struct hl_cs_counters_atomic { atomic64_t validation_drop_cnt; }; +/** + * struct hl_pending_cb - pending command buffer structure + * @cb_node: cb node in pending cb list + * @cb: command buffer to send in next submission + * @cb_size: command buffer size + * @hw_queue_id: destination queue id + */ +struct hl_pending_cb { + struct list_head cb_node; + struct hl_cb *cb; + u32 cb_size; + u32 hw_queue_id; +}; + /** * struct hl_ctx - user/kernel context. * @mem_hash: holds mapping from virtual address to virtual memory area @@ -1031,6 +1045,8 @@ struct hl_cs_counters_atomic { * @mmu_lock: protects the MMU page tables. Any change to the PGT, modifying the * MMU hash or walking the PGT requires talking this lock. * @debugfs_list: node in debugfs list of contexts. + * pending_cb_list: list of pending command buffers waiting to be sent upon + * next user command submission context. * @cs_counters: context command submission counters. * @cb_va_pool: device VA pool for command buffers which are mapped to the * device's MMU. @@ -1039,11 +1055,17 @@ struct hl_cs_counters_atomic { * index to cs_pending array. * @dram_default_hops: array that holds all hops addresses needed for default * DRAM mapping. + * @pending_cb_lock: spinlock to protect pending cb list * @cs_lock: spinlock to protect cs_sequence. * @dram_phys_mem: amount of used physical DRAM memory by this context. * @thread_ctx_switch_token: token to prevent multiple threads of the same * context from running the context switch phase. * Only a single thread should run it. + * @thread_pending_cb_token: token to prevent multiple threads from processing + * the pending CB list. Only a single thread should + * process the list since it is protected by a + * spinlock and we don't want to halt the entire + * command submission sequence. * @thread_ctx_switch_wait_token: token to prevent the threads that didn't run * the context switch phase from moving to their * execution phase before the context switch phase @@ -1062,13 +1084,16 @@ struct hl_ctx { struct mutex mem_hash_lock; struct mutex mmu_lock; struct list_head debugfs_list; + struct list_head pending_cb_list; struct hl_cs_counters_atomic cs_counters; struct gen_pool *cb_va_pool; u64 cs_sequence; u64 *dram_default_hops; + spinlock_t pending_cb_lock; spinlock_t cs_lock; atomic64_t dram_phys_mem; atomic_t thread_ctx_switch_token; + atomic_t thread_pending_cb_token; u32 thread_ctx_switch_wait_token; u32 asid; u32 handle; @@ -2143,6 +2168,7 @@ int hl_cb_va_pool_init(struct hl_ctx *ctx); void hl_cb_va_pool_fini(struct hl_ctx *ctx); void hl_cs_rollback_all(struct hl_device *hdev); +void hl_pending_cb_list_flush(struct hl_ctx *ctx); struct hl_cs_job *hl_cs_allocate_job(struct hl_device *hdev, enum hl_queue_type queue_type, bool is_kernel_allocated_cb); void hl_sob_reset_error(struct kref *ref); -- 2.39.5