[Mesa-dev] [PATCH 1/1] radeonsi: Bump number of allowed global buffers to 32

Marek Olšák maraeo at gmail.com
Thu Oct 18 19:44:38 UTC 2018


This is OK.

Reviewed-by: Marek Olšák <marek.olsak at amd.com>

Marek

On Thu, Oct 18, 2018 at 3:43 PM Jan Vesely <jan.vesely at rutgers.edu> wrote:

> Fixes assertion failure/crash when running luxmark/luxball on clover.
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108272
> CC: mesa-stable at lists.freedesktop.org
> Signed-off-by: Jan Vesely <jan.vesely at rutgers.edu>
> ---
> This is just a stop gap measure. OpenCL does not limit the number of
> global buffers, but we can return CL_MEM_OBJECT_ALLOCATION_FAILURE in
> clEnqueueNDRangeKernel.
> What would be the preferred way? Add PIPE_CAP to query this? allow
> set_global_binding to fail with too many buffers?
>
> thanks,
> Jan
>
>  src/gallium/drivers/radeonsi/si_compute.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/gallium/drivers/radeonsi/si_compute.h
> b/src/gallium/drivers/radeonsi/si_compute.h
> index 99b501673c..57c0bde4ac 100644
> --- a/src/gallium/drivers/radeonsi/si_compute.h
> +++ b/src/gallium/drivers/radeonsi/si_compute.h
> @@ -29,7 +29,7 @@
>
>  #include "si_shader.h"
>
> -#define MAX_GLOBAL_BUFFERS 22
> +#define MAX_GLOBAL_BUFFERS 32
>
>  struct si_compute {
>         struct pipe_reference reference;
> --
> 2.17.2
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20181018/95e3010b/attachment.html>


More information about the mesa-dev mailing list