<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Fri, Dec 16, 2016 at 1:26 PM, Francisco Jerez <span dir="ltr"><<a href="mailto:currojerez@riseup.net" target="_blank">currojerez@riseup.net</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>> writes:<br>
<br>
> On Dec 15, 2016 3:22 PM, "Francisco Jerez" <<a href="mailto:currojerez@riseup.net">currojerez@riseup.net</a>> wrote:<br>
><br>
> This fixes an apparent regression in a bunch of image store vulkan CTS<br>
> tests from commit ad38ba113491869ab0dffed937f7b3<wbr>dd50e8a735, which<br>
> started using OWORD block read messages to implement UBO loads. The<br>
> reason for the failure is that we were giving bogus buffer alignment<br>
> limits to the application (1B), so the CTS would happily come back<br>
> with descriptor sets pointing at not even word-aligned uniform buffer<br>
> addresses. No idea how these tests could possibly work before while<br>
> we were using the sampler to fetch pull constants.<br>
><br>
><br>
> Fun fact: the sampler has very few requirements when it comes to linear<br>
> surfaces (including buffers). One notable exception is that it doesn't<br>
> seem to care about alignment... at all. A bit surprising, I know, but I've<br>
> seen other cases where it's totally fine with nonaligned offsets and<br>
> strides (in the case of 2D).<br>
><br>
> That is for fixing this!<br>
><br>
<br>
</span>Cool feature! It would probably make sense to keep mesa-stable CC'ed<br>
regardless, because I don't think the dataport messages used to access<br>
storage buffers are so permissive with unaligned addresses.<br></blockquote><div><br></div><div>Agreed.<br></div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">
> Reviewed-by: Jason Ekstrand <<a href="mailto:jason@jlekstrand.net">jason@jlekstrand.net</a>><br>
><br>
><br>
> Cc: <<a href="mailto:mesa-stable@lists.freedesktop.org">mesa-stable@lists.<wbr>freedesktop.org</a>><br>
> Bugzilla: <a href="https://bugs.freedesktop.org/show_bug.cgi?id=99097" rel="noreferrer" target="_blank">https://bugs.freedesktop.org/<wbr>show_bug.cgi?id=99097</a><br>
> Reported-by: Mark Janes <<a href="mailto:mark.a.janes@intel.com">mark.a.janes@intel.com</a>><br>
> ---<br>
> src/intel/vulkan/anv_device.c | 4 ++--<br>
> 1 file changed, 2 insertions(+), 2 deletions(-)<br>
><br>
> diff --git a/src/intel/vulkan/anv_device.<wbr>c b/src/intel/vulkan/anv_device.<wbr>c<br>
> index e3d278d..9245e5c 100644<br>
> --- a/src/intel/vulkan/anv_device.<wbr>c<br>
> +++ b/src/intel/vulkan/anv_device.<wbr>c<br>
> @@ -582,8 +582,8 @@ void anv_<wbr>GetPhysicalDeviceProperties(<br>
> .viewportSubPixelBits = 13, /* We take a float?<br>
> */<br>
> .minMemoryMapAlignment = 4096, /* A page */<br>
> .minTexelBufferOffsetAlignment = 1,<br>
> - .<wbr>minUniformBufferOffsetAlignmen<wbr>t = 1,<br>
> - .<wbr>minStorageBufferOffsetAlignmen<wbr>t = 1,<br>
> + .<wbr>minUniformBufferOffsetAlignmen<wbr>t = 16,<br>
> + .<wbr>minStorageBufferOffsetAlignmen<wbr>t = 4,<br>
> .minTexelOffset = -8,<br>
> .maxTexelOffset = 7,<br>
> .minTexelGatherOffset = -32,<br>
> --<br>
> 2.10.2<br>
><br>
> ______________________________<wbr>_________________<br>
> mesa-dev mailing list<br>
> <a href="mailto:mesa-dev@lists.freedesktop.org">mesa-dev@lists.freedesktop.org</a><br>
> <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/<wbr>mailman/listinfo/mesa-dev</a><br>
</div></div></blockquote></div><br></div></div>