[Mesa-dev] [PATCH] i965: Allocate at least some URB space even when max_vertices = 0.
Ian Romanick
idr at freedesktop.org
Tue Dec 6 00:33:51 UTC 2016
On 12/05/2016 01:39 PM, Kenneth Graunke wrote:
> Allocating zero URB space is a really bad idea. The hardware has to
> give threads a handle to their URB space, and threads have to use that
> to terminate the thread. Having it be an empty region just breaks a
> lot of assumptions. Hence, why we asserted that it isn't possible.
>
> Unfortunately, it /is/ possible prior to Gen8, if max_vertices = 0.
> In theory a geometry shader could do SSBO/image access and maybe
> still accomplish something. In reality, this is tripped up by
> conformance tests.
>
> Gen8+ already avoids this problem by placing the vertex count DWord
> in the URB entry header. This fixes things on earlier generations.
>
> Signed-off-by: Kenneth Graunke <kenneth at whitecape.org>
Assuming this is the same patch you sent me a couple months ago,
Tested-by: Ian Romanick <ian.d.romanick at intel.com>
Possibly candidate for stable?
> ---
> src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp | 8 +++++++-
> 1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
> index 900d9d3..3894a63 100644
> --- a/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
> +++ b/src/mesa/drivers/dri/i965/brw_vec4_gs_visitor.cpp
> @@ -781,7 +781,13 @@ brw_compile_gs(const struct brw_compiler *compiler, void *log_data,
> if (compiler->devinfo->gen >= 8)
> output_size_bytes += 32;
>
> - assert(output_size_bytes >= 1);
> + /* Shaders can technically set max_vertices = 0, at which point we
> + * may have a URB size of 0 bytes. Nothing good can come from that,
> + * so enforce a minimum size.
> + */
> + if (output_size_bytes == 0)
> + output_size_bytes = 1;
> +
> unsigned max_output_size_bytes = GEN7_MAX_GS_URB_ENTRY_SIZE_BYTES;
> if (compiler->devinfo->gen == 6)
> max_output_size_bytes = GEN6_MAX_GS_URB_ENTRY_SIZE_BYTES;
>
More information about the mesa-dev
mailing list