[Mesa-dev] [PATCH] radeonsi: fix discard-only fragment shaders
Marek Olšák
maraeo at gmail.com
Tue Jan 19 14:13:41 PST 2016
On Tue, Jan 19, 2016 at 9:13 PM, Nicolai Hähnle <nhaehnle at gmail.com> wrote:
> From: Nicolai Hähnle <nicolai.haehnle at amd.com>
>
> When a fragment shader is used that has no outputs but does conditional
> discard (KILL_IF), all fragments are killed without this patch.
>
> By comparing various register settings, my conclusion is that the exec mask
> is either not properly forwarded to the DB by NULL exports or ends up being
> unused, at least when there is _only_ a NULL export (the ISA documentation
> claims that NULL exports can be used to override a previously exported exec
> mask).
>
> Of the various approaches I have tried to work around the problem, this one
> seems to be the least invasive one.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93761
> ---
> I've condensed the bug reporter's test case to a Piglit test that I will
> send out soon.
>
> This should probably go to stable as well, but won't apply as is because of
> the many recent changes to fragment shader export handling. I will follow
> this mail up with a version that applies to 11.1 and that I've confirmed
> fixes the problem there as well.
>
> src/gallium/drivers/radeonsi/si_state_shaders.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/gallium/drivers/radeonsi/si_state_shaders.c b/src/gallium/drivers/radeonsi/si_state_shaders.c
> index 8ff70b4..0f0cbb2 100644
> --- a/src/gallium/drivers/radeonsi/si_state_shaders.c
> +++ b/src/gallium/drivers/radeonsi/si_state_shaders.c
> @@ -456,6 +456,7 @@ static void si_shader_ps(struct si_shader *shader)
> info->writes_samplemask ? V_028710_SPI_SHADER_32_ABGR :
> info->writes_stencil ? V_028710_SPI_SHADER_32_GR :
> info->writes_z ? V_028710_SPI_SHADER_32_R :
> + (info->uses_kill && !info->num_outputs) ? V_028710_SPI_SHADER_32_R :
I guess you know this already: It should also take alpha-test and
SPI_SHADER_Z_FORMAT into account.
Marek
More information about the mesa-dev
mailing list