Fixes compilation with musl libc.
---
src/mesa/main/querymatrix.c | 16 +---
1 file changed, 5 insertions(+), 11 deletions(-)
diff --git a/src/mesa/main/querymatrix.c b/src/mesa/main/querymatrix.c
index eb36c76..57c536f 100644
--- a/src/mesa/main/querymatrix.c
+++ b/src/mesa/main/quer
Yes, reading depth textures with sampler2DMS is allowed. There is
GL_MAX_DEPTH_TEXTURE_SAMPLES added by GL_ARB_texture_multisample I think.
Both solutions (the new CAP and using is_format_supported) are okay with me.
Marek
On Sun, Jan 18, 2015 at 8:23 PM, Ilia Mirkin wrote:
> I'm confused by wh
Looks like for ZS buffers the rule is wrong.
There's only 3 formats that can be sampled (other formats are for shadow
mapping),
there are DF16, DF24 and INTZ.
Apparently the rule described (r= depth, g=b=0, a=1) is true for DF16
and DF24,
but INTZ should be (r=g=b=depth, a=1). Tested on Win.
In d3d9, mixing render targets and depth buffer of different
multisampling is forbidden.
There's a d3d9 hack, called RESZ, which allows to copy a multisampled
depth buffer to
a single sampled depth buffer. This cap was about knowing if that
feature could be supported.
But likely the better w
I'm confused by what this flag is allowing. Are we talking about a
blit from MS -> non-MS? texelFetch(sampler2DMS) for a depth buffer?
texelFetch(sampler2D) for a MS surface? (is that even allowed in GL?)
Or are we talking about mixing MS and non-MS buffers in the fb? These
are all separate things.
gl_Position.z isn't bound by any range. It doesn't have to be clipped
(it can be clamped instead), but if it's clipped, the value outside of
the range still affects the whole primitive that can be at least
partially visible.
Marek
On Sat, Jan 17, 2015 at 5:07 PM, Thomas Helland
wrote:
> I see wh
This looks good.
Reviewed-by: Marek Olšák
Marek
On Sat, Jan 17, 2015 at 2:44 PM, Axel Davy wrote:
> Resolving a multisampled depth texture into
> a single sampled texture is supported on >= SM4.1
> hw. It is possible some previous hw support it.
>
> The ability was tested on radeonsi and nvc0.
https://bugs.freedesktop.org/show_bug.cgi?id=88535
dexcelstra...@gmail.com changed:
What|Removed |Added
Component|Drivers/DRI/nouveau |Mesa core
Assignee|no
Potentially this proposed cap should be replaced
by a is_format_supported call with PIPE_BIND_SAMPLER_VIEW
for a multisampled depth format.
Looking at the driver sources, I get the impression only
nouveau and radeon support MSAA. But they may need
some correction, because It looks like for r600 a
