Re: [PATCH 01/18] dma-buf: cleanup dma_fence_describe v3

[Philipp Stanner <phasta@mailbox.org>]

On Thu, 2025-11-13 at 15:51 +0100, Christian König wrote:
> The driver and timeline name are meaningless for signaled fences.
> 
> Drop them and also print the context number.
> 
> v2: avoid the calls when the BO is already signaled.
> v3: use same format as trace points for context and seqno.
> 
> Signed-off-by: Christian König <christian.koenig@amd.com>
> ---

I still think that more fleshed out commit messages are great :)


>  drivers/dma-buf/dma-fence.c | 20 +++++++++++---------
>  1 file changed, 11 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c
> index c5d1478b28dd..b4f5c8635276 100644
> --- a/drivers/dma-buf/dma-fence.c
> +++ b/drivers/dma-buf/dma-fence.c
> @@ -997,19 +997,21 @@ EXPORT_SYMBOL(dma_fence_set_deadline);
>   */
>  void dma_fence_describe(struct dma_fence *fence, struct seq_file *seq)
>  {
> -	const char __rcu *timeline;
> -	const char __rcu *driver;
> +	const char __rcu *timeline = "";
> +	const char __rcu *driver = "";
> +	const char *signaled = "";
>  
>  	rcu_read_lock();
>  
> -	timeline = dma_fence_timeline_name(fence);
> -	driver = dma_fence_driver_name(fence);
> +	if (!dma_fence_is_signaled(fence)) {
> +		timeline = dma_fence_timeline_name(fence);
> +		driver = dma_fence_driver_name(fence);
> +		signaled = "un";
> +	}
>  
> -	seq_printf(seq, "%s %s seq %llu %ssignalled\n",
> -		   rcu_dereference(driver),
> -		   rcu_dereference(timeline),

Why does usage  of rcu_dereference become unnecessary? 

> -		   fence->seqno,
> -		   dma_fence_is_signaled(fence) ? "" : "un");
> +	seq_printf(seq, "%llu:%llu %s %s %ssignalled\n",
> +		   fence->context, fence->seqno, timeline, driver,
> +		   signaled);
>  
>  	rcu_read_unlock();
>  }