Re: [PATCH] drm/ttm: Fix race condition in ttm_bo_delayed_delete

[Thomas H. <th...@sh...>]

Luca Barbieri wrote:
> When designing this, we should also keep in mind that some drivers
> (e.g. nouveau) have multiple FIFO channels, and thus we would like a
> buffer to be referenced for reading by multiple channels at once (and
> be destroyed only when all fences are expired, obviously).
> Also, hardware may support on-GPU inter-channel synchronization, and
> then multiple references may be for writing too.
>   

In the context of the current code, I've been thinking of having a list 
of fences on each bo to support multiple readers, and also to try to 
deal with the problem of simultaneous GPU- and CPU readers.

But if the hardware supports on-GPU inter-channel synchronization, I 
figure the code should be smart enough to only wait on the "last" write 
fence?

/Thomas