<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Mon, Aug 8, 2016 at 11:40 AM, Nanley Chery <<a href="mailto:nanleychery@gmail.com" target="_blank">nanleychery@gmail.com</a>> wrote: <blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">On Wed, Aug 03, 2016 at 01:06:10PM -0700, Jason Ekstrand wrote: > We can't actually clear these images normally because we can't render to > them. Instead, we have to manually unpack the rgb9e5 color value on the > CPU and clear it as R32_UINT. We still have a bit of work to do to clear > non-power-of-two images, but this should get all of the power-of-two clears > working on at least Haswell. Could you mention which CTS tests this fixes? That would help me to test the correctness of this change. </blockquote><div> </div><div>The vulkan-cts-1.0-dev branch of the CTS has a bunch of tests to test clearing. This fixes 3 of the dEQP-VK.api.image_clearing.clear_color_image.* tests </div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> > > Cc: "12.0" <<a href="mailto:mesa-stable@lists.freedesktop.org">mesa-stable@lists.freedesktop.org</a>> > --- > src/intel/vulkan/anv_meta_clear.c | 16 ++++++++++++++-- > 1 file changed, 14 insertions(+), 2 deletions(-) > > diff --git a/src/intel/vulkan/anv_meta_clear.c b/src/intel/vulkan/anv_meta_clear.c > index fa07ec1..d8b5ce0 100644 > --- a/src/intel/vulkan/anv_meta_clear.c > +++ b/src/intel/vulkan/anv_meta_clear.c > @@ -25,6 +25,8 @@ > #include "anv_private.h" > #include "nir/nir_builder.h" > > +#include "gallium/auxiliary/util/u_format_rgb9e5.h" > + > /** Vertex attributes for color clears. */ > struct color_clear_vattrs { > struct anv_vue_header vue_header; > @@ -760,6 +762,16 @@ anv_cmd_clear_image(struct anv_cmd_buffer *cmd_buffer, > { > VkDevice device_h = anv_device_to_handle(cmd_buffer->device); > > + VkFormat vk_format = image->vk_format; > + if (vk_format == VK_FORMAT_E5B9G9R9_UFLOAT_PACK32) { > + /* We can't actually render to this format so we have to work around it > + * by manualy unpacking and using R32_UINT. s/manualy/manually/ <div><div class="h5"> > + */ > + clear_value.color.uint32[0] = > + float3_to_rgb9e5(clear_value.color.float32); > + vk_format = VK_FORMAT_R32_UINT; > + } > + > for (uint32_t r = 0; r < range_count; r++) { > const VkImageSubresourceRange *range = &ranges[r]; > for (uint32_t l = 0; l < anv_get_levelCount(image, range); ++l) { > @@ -773,7 +785,7 @@ anv_cmd_clear_image(struct anv_cmd_buffer *cmd_buffer, > .sType = VK_STRUCTURE_TYPE_IMAGE_VIEW_CREATE_INFO, > .image = anv_image_to_handle(image), > .viewType = anv_meta_get_view_type(image), > - .format = image->vk_format, > + .format = vk_format, > .subresourceRange = { > .aspectMask = range->aspectMask, > .baseMipLevel = range->baseMipLevel + l, > @@ -800,7 +812,7 @@ anv_cmd_clear_image(struct anv_cmd_buffer *cmd_buffer, > &fb); > > VkAttachmentDescription att_desc = { > - .format = iview.vk_format, > + .format = vk_format, > .loadOp = VK_ATTACHMENT_LOAD_OP_LOAD, > .storeOp = VK_ATTACHMENT_STORE_OP_STORE, > .stencilLoadOp = VK_ATTACHMENT_LOAD_OP_LOAD, > -- > 2.5.0.400.gff86faf > </div></div>> _______________________________________________ > mesa-dev mailing list > <a href="mailto:mesa-dev@lists.freedesktop.org">mesa-dev@lists.freedesktop.org</a> > <a href="https://lists.freedesktop.org/mailman/listinfo/mesa-dev" rel="noreferrer" target="_blank">https://lists.freedesktop.org/mailman/listinfo/mesa-dev</a> </blockquote></div> </div></div>