Re: [PATCH v2 2/4] drm/vc4: Use DMA Resv to implement VC4 wait BO IOCTL

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 12/12, Maíra Canal wrote:
> As the BOs used by VC4 have DMA Reservation Objects attached to it,
> there is no need to use seqnos wait for the BOs availability. Instead,
> we can use `dma_gem_dma_resv_wait()`.

typo: drm_gem_dma_resc_wait()
> 
> Signed-off-by: Maíra Canal <mcanal@xxxxxxxxxx>
> ---
>  drivers/gpu/drm/vc4/vc4_drv.h | 10 ++++++++++
>  drivers/gpu/drm/vc4/vc4_gem.c | 31 +++++++++++++++++++------------
>  2 files changed, 29 insertions(+), 12 deletions(-)
> 
> diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h
> index 4a078ffd9f82..03ed40ab9a93 100644
> --- a/drivers/gpu/drm/vc4/vc4_drv.h
> +++ b/drivers/gpu/drm/vc4/vc4_drv.h
> @@ -924,6 +924,16 @@ struct vc4_validated_shader_info {
>  						   (Wmax))
>  #define wait_for(COND, MS)		_wait_for((COND), (MS) * 1000, 10, 1000)
>  
> +static inline unsigned long nsecs_to_jiffies_timeout(const u64 n)
> +{
> +	/* nsecs_to_jiffies64() does not guard against overflow */
> +	if ((NSEC_PER_SEC % HZ) != 0 &&
> +	    div_u64(n, NSEC_PER_SEC) >= MAX_JIFFY_OFFSET / HZ)
> +		return MAX_JIFFY_OFFSET;
> +
> +	return min_t(u64, MAX_JIFFY_OFFSET, nsecs_to_jiffies64(n) + 1);
> +}
> +

Should nsecs_to_jiffies_timeout become common code?

>  /* vc4_bo.c */
>  struct drm_gem_object *vc4_create_object(struct drm_device *dev, size_t size);
>  struct vc4_bo *vc4_bo_create(struct drm_device *dev, size_t size,
> diff --git a/drivers/gpu/drm/vc4/vc4_gem.c b/drivers/gpu/drm/vc4/vc4_gem.c
> index 1021f45cb53c..4037c65eb269 100644
> --- a/drivers/gpu/drm/vc4/vc4_gem.c
> +++ b/drivers/gpu/drm/vc4/vc4_gem.c
> @@ -1017,11 +1017,13 @@ int
>  vc4_wait_bo_ioctl(struct drm_device *dev, void *data,
>  		  struct drm_file *file_priv)
>  {
> -	struct vc4_dev *vc4 = to_vc4_dev(dev);
>  	int ret;
> +	struct vc4_dev *vc4 = to_vc4_dev(dev);
>  	struct drm_vc4_wait_bo *args = data;
> -	struct drm_gem_object *gem_obj;
> -	struct vc4_bo *bo;
> +	unsigned long timeout_jiffies =
> +		nsecs_to_jiffies_timeout(args->timeout_ns);
> +	ktime_t start = ktime_get();
> +	u64 delta_ns;
>  
>  	if (WARN_ON_ONCE(vc4->gen > VC4_GEN_4))
>  		return -ENODEV;
> @@ -1029,17 +1031,22 @@ vc4_wait_bo_ioctl(struct drm_device *dev, void *data,
>  	if (args->pad != 0)
>  		return -EINVAL;
>  
> -	gem_obj = drm_gem_object_lookup(file_priv, args->handle);
> -	if (!gem_obj) {
> -		DRM_DEBUG("Failed to look up GEM BO %d\n", args->handle);
> -		return -EINVAL;
> -	}
> -	bo = to_vc4_bo(gem_obj);
> +	ret = drm_gem_dma_resv_wait(file_priv, args->handle,
> +				    true, timeout_jiffies);
>  
> -	ret = vc4_wait_for_seqno_ioctl_helper(dev, bo->seqno,
> -					      &args->timeout_ns);
> +	/* Decrement the user's timeout, in case we got interrupted
> +	 * such that the ioctl will be restarted.
> +	 */
> +	delta_ns = ktime_to_ns(ktime_sub(ktime_get(), start));
> +	if (delta_ns < args->timeout_ns)
> +		args->timeout_ns -= delta_ns;
> +	else
> +		args->timeout_ns = 0;
> +
> +	/* Asked to wait beyond the jiffy/scheduler precision? */
> +	if (ret == -ETIME && args->timeout_ns)
> +		ret = -EAGAIN;
>  
> -	drm_gem_object_put(gem_obj);

Just left some comments above.

This change sounds sensible and the patch LGTM.
Thanks!

Reviewed-by: Melissa Wen <mwen@xxxxxxxxxx>


>  	return ret;
>  }
>  
> -- 
> 2.47.1
> 



[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux