[Intel-gfx] [PATCH 31/43] drm/i915/bdw: Handle context switch events

[Daniel Vetter]

On Thu, Jul 24, 2014 at 05:04:39PM +0100, Thomas Daniel wrote:
> Handle all context status events in the context status buffer on every
> context switch interrupt. We only remove work from the execlist queue
> after a context status buffer reports that it has completed and we only
> attempt to schedule new contexts on interrupt when a previously submitted
> context completes (unless no contexts are queued, which means the GPU is
> free).
> 
> We canot call intel_runtime_pm_get() in an interrupt (or with a spinlock
> grabbed, FWIW), because it might sleep, which is not a nice thing to do.
> Instead, do the runtime_pm get/put together with the create/destroy request,
> and handle the forcewake get/put directly.
> 
> Signed-off-by: Thomas Daniel <thomas.daniel at intel.com>
> 
> v2: Unreferencing the context when we are freeing the request might free
> the backing bo, which requires the struct_mutex to be grabbed, so defer
> unreferencing and freeing to a bottom half.
> 
> v3:
> - Ack the interrupt inmediately, before trying to handle it (fix for
> missing interrupts by Bob Beckett <robert.beckett at intel.com>).
> - Update the Context Status Buffer Read Pointer, just in case (spotted
> by Damien Lespiau).
> 
> v4: New namespace and multiple rebase changes.
> 
> v5: Squash with "drm/i915/bdw: Do not call intel_runtime_pm_get() in an
> interrupt", as suggested by Daniel.
> 
> Signed-off-by: Oscar Mateo <oscar.mateo at intel.com>

One more ...

> +void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring)

Please rename this to intel_execlist_ctx_events_irq_handler or similar for
consistency with all the other irq handler functions in a follow-up patch.
That kind of consistency helps a lot when reviewing the locking of
irq-save spinlocks.
-Daniel

> +{
> +	struct drm_i915_private *dev_priv = ring->dev->dev_private;
> +	u32 status_pointer;
> +	u8 read_pointer;
> +	u8 write_pointer;
> +	u32 status;
> +	u32 status_id;
> +	u32 submit_contexts = 0;
> +
> +	status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring));
> +
> +	read_pointer = ring->next_context_status_buffer;
> +	write_pointer = status_pointer & 0x07;
> +	if (read_pointer > write_pointer)
> +		write_pointer += 6;
> +
> +	spin_lock(&ring->execlist_lock);
> +
> +	while (read_pointer < write_pointer) {
> +		read_pointer++;
> +		status = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
> +				(read_pointer % 6) * 8);
> +		status_id = I915_READ(RING_CONTEXT_STATUS_BUF(ring) +
> +				(read_pointer % 6) * 8 + 4);
> +
> +		if (status & GEN8_CTX_STATUS_COMPLETE) {
> +			if (execlists_check_remove_request(ring, status_id))
> +				submit_contexts++;
> +		}
> +	}
> +
> +	if (submit_contexts != 0)
> +		execlists_context_unqueue(ring);
> +
> +	spin_unlock(&ring->execlist_lock);
> +
> +	WARN(submit_contexts > 2, "More than two context complete events?\n");
> +	ring->next_context_status_buffer = write_pointer % 6;
> +
> +	I915_WRITE(RING_CONTEXT_STATUS_PTR(ring),
> +			((u32)ring->next_context_status_buffer & 0x07) << 8);
> +}
> +
> +static void execlists_free_request_task(struct work_struct *work)
> +{
> +	struct intel_ctx_submit_request *req =
> +			container_of(work, struct intel_ctx_submit_request, work);
> +	struct drm_device *dev = req->ring->dev;
> +	struct drm_i915_private *dev_priv = dev->dev_private;
> +
> +	intel_runtime_pm_put(dev_priv);
> +
> +	mutex_lock(&dev->struct_mutex);
> +	i915_gem_context_unreference(req->ctx);
> +	mutex_unlock(&dev->struct_mutex);
> +
> +	kfree(req);
> +}
> +
>  static int execlists_context_queue(struct intel_engine_cs *ring,
>  				   struct intel_context *to,
>  				   u32 tail)
> @@ -261,6 +375,8 @@ static int execlists_context_queue(struct intel_engine_cs *ring,
>  	i915_gem_context_reference(req->ctx);
>  	req->ring = ring;
>  	req->tail = tail;
> +	INIT_WORK(&req->work, execlists_free_request_task);
> +	intel_runtime_pm_get(dev_priv);
>  
>  	spin_lock_irqsave(&ring->execlist_lock, flags);
>  
> @@ -908,6 +1024,7 @@ static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *rin
>  
>  	INIT_LIST_HEAD(&ring->execlist_queue);
>  	spin_lock_init(&ring->execlist_lock);
> +	ring->next_context_status_buffer = 0;
>  
>  	ret = intel_lr_context_deferred_create(dctx, ring);
>  	if (ret)
> diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h
> index 14492a9..2e8929f 100644
> --- a/drivers/gpu/drm/i915/intel_lrc.h
> +++ b/drivers/gpu/drm/i915/intel_lrc.h
> @@ -66,6 +66,9 @@ struct intel_ctx_submit_request {
>  	u32 tail;
>  
>  	struct list_head execlist_link;
> +	struct work_struct work;
>  };
>  
> +void intel_execlists_handle_ctx_events(struct intel_engine_cs *ring);
> +
>  #endif /* _INTEL_LRC_H_ */
> diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
> index 6358823..905d1ba 100644
> --- a/drivers/gpu/drm/i915/intel_ringbuffer.h
> +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
> @@ -225,6 +225,7 @@ struct  intel_engine_cs {
>  	/* Execlists */
>  	spinlock_t execlist_lock;
>  	struct list_head execlist_queue;
> +	u8 next_context_status_buffer;
>  	u32             irq_keep_mask; /* bitmask for interrupts that should not be masked */
>  	int		(*emit_request)(struct intel_ringbuffer *ringbuf);
>  	int		(*emit_flush)(struct intel_ringbuffer *ringbuf,
> -- 
> 1.7.9.5
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch