[Mesa-dev] [PATCH] radv: Do not hardcode fast clear formats.
Dave Airlie
airlied at gmail.com
Mon Jun 4 23:54:58 UTC 2018
On 5 June 2018 at 08:53, Bas Nieuwenhuizen <bas at basnieuwenhuizen.nl> wrote:
> except for the odd one out.
>
> This should support many more formats.
I suppose we should have done this back at the start, but I think I was lazy.
Looks like a good idea, assuming it doesn't break anything in CTS,
Reviewed-by: Dave Airlie <airlied at redhat.com>
> ---
> src/amd/vulkan/radv_formats.c | 253 ++++++++++------------------------
> 1 file changed, 73 insertions(+), 180 deletions(-)
>
> diff --git a/src/amd/vulkan/radv_formats.c b/src/amd/vulkan/radv_formats.c
> index aafe331f27d..50ec904d514 100644
> --- a/src/amd/vulkan/radv_formats.c
> +++ b/src/amd/vulkan/radv_formats.c
> @@ -880,194 +880,87 @@ bool radv_format_pack_clear_color(VkFormat format,
> uint32_t clear_vals[2],
> VkClearColorValue *value)
> {
> - uint8_t r = 0, g = 0, b = 0, a = 0;
> const struct vk_format_description *desc = vk_format_description(format);
>
> - if (vk_format_get_component_bits(format, VK_FORMAT_COLORSPACE_RGB, 0) <= 8) {
> - if (desc->colorspace == VK_FORMAT_COLORSPACE_RGB) {
> - r = float_to_ubyte(value->float32[0]);
> - g = float_to_ubyte(value->float32[1]);
> - b = float_to_ubyte(value->float32[2]);
> - a = float_to_ubyte(value->float32[3]);
> - } else if (desc->colorspace == VK_FORMAT_COLORSPACE_SRGB) {
> - r = util_format_linear_float_to_srgb_8unorm(value->float32[0]);
> - g = util_format_linear_float_to_srgb_8unorm(value->float32[1]);
> - b = util_format_linear_float_to_srgb_8unorm(value->float32[2]);
> - a = float_to_ubyte(value->float32[3]);
> - }
> - }
> - switch (format) {
> - case VK_FORMAT_R8_UNORM:
> - case VK_FORMAT_R8_SRGB:
> - clear_vals[0] = r;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8_UNORM:
> - case VK_FORMAT_R8G8_SRGB:
> - clear_vals[0] = r | g << 8;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8B8A8_SRGB:
> - case VK_FORMAT_R8G8B8A8_UNORM:
> - clear_vals[0] = r | g << 8 | b << 16 | a << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_B8G8R8A8_SRGB:
> - case VK_FORMAT_B8G8R8A8_UNORM:
> - clear_vals[0] = b | g << 8 | r << 16 | a << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_A8B8G8R8_UNORM_PACK32:
> - case VK_FORMAT_A8B8G8R8_SRGB_PACK32:
> - clear_vals[0] = r | g << 8 | b << 16 | a << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8_UINT:
> - clear_vals[0] = value->uint32[0] & 0xff;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8_SINT:
> - clear_vals[0] = value->int32[0] & 0xff;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16_UINT:
> - clear_vals[0] = value->uint32[0] & 0xffff;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8_UINT:
> - clear_vals[0] = value->uint32[0] & 0xff;
> - clear_vals[0] |= (value->uint32[1] & 0xff) << 8;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8_SINT:
> - clear_vals[0] = value->int32[0] & 0xff;
> - clear_vals[0] |= (value->int32[1] & 0xff) << 8;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8B8A8_UINT:
> - clear_vals[0] = value->uint32[0] & 0xff;
> - clear_vals[0] |= (value->uint32[1] & 0xff) << 8;
> - clear_vals[0] |= (value->uint32[2] & 0xff) << 16;
> - clear_vals[0] |= (value->uint32[3] & 0xff) << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R8G8B8A8_SINT:
> - clear_vals[0] = value->int32[0] & 0xff;
> - clear_vals[0] |= (value->int32[1] & 0xff) << 8;
> - clear_vals[0] |= (value->int32[2] & 0xff) << 16;
> - clear_vals[0] |= (value->int32[3] & 0xff) << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_A8B8G8R8_UINT_PACK32:
> - clear_vals[0] = value->uint32[0] & 0xff;
> - clear_vals[0] |= (value->uint32[1] & 0xff) << 8;
> - clear_vals[0] |= (value->uint32[2] & 0xff) << 16;
> - clear_vals[0] |= (value->uint32[3] & 0xff) << 24;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16_UINT:
> - clear_vals[0] = value->uint32[0] & 0xffff;
> - clear_vals[0] |= (value->uint32[1] & 0xffff) << 16;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16B16A16_UINT:
> - clear_vals[0] = value->uint32[0] & 0xffff;
> - clear_vals[0] |= (value->uint32[1] & 0xffff) << 16;
> - clear_vals[1] = value->uint32[2] & 0xffff;
> - clear_vals[1] |= (value->uint32[3] & 0xffff) << 16;
> - break;
> - case VK_FORMAT_R32_UINT:
> - clear_vals[0] = value->uint32[0];
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R32G32_UINT:
> - clear_vals[0] = value->uint32[0];
> - clear_vals[1] = value->uint32[1];
> - break;
> - case VK_FORMAT_R32_SINT:
> - clear_vals[0] = value->int32[0];
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16_SFLOAT:
> - clear_vals[0] = util_float_to_half(value->float32[0]);
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16_SFLOAT:
> - clear_vals[0] = util_float_to_half(value->float32[0]);
> - clear_vals[0] |= (uint32_t)util_float_to_half(value->float32[1]) << 16;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16B16A16_SFLOAT:
> - clear_vals[0] = util_float_to_half(value->float32[0]);
> - clear_vals[0] |= (uint32_t)util_float_to_half(value->float32[1]) << 16;
> - clear_vals[1] = util_float_to_half(value->float32[2]);
> - clear_vals[1] |= (uint32_t)util_float_to_half(value->float32[3]) << 16;
> - break;
> - case VK_FORMAT_R16_UNORM:
> - clear_vals[0] = ((uint16_t)util_iround(CLAMP(value->float32[0], 0.0f, 1.0f) * 0xffff)) & 0xffff;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16_UNORM:
> - clear_vals[0] = ((uint16_t)util_iround(CLAMP(value->float32[0], 0.0f, 1.0f) * 0xffff)) & 0xffff;
> - clear_vals[0] |= ((uint16_t)util_iround(CLAMP(value->float32[1], 0.0f, 1.0f) * 0xffff)) << 16;
> - clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R16G16B16A16_UNORM:
> - clear_vals[0] = ((uint16_t)util_iround(CLAMP(value->float32[0], 0.0f, 1.0f) * 0xffff)) & 0xffff;
> - clear_vals[0] |= ((uint16_t)util_iround(CLAMP(value->float32[1], 0.0f, 1.0f) * 0xffff)) << 16;
> - clear_vals[1] = ((uint16_t)util_iround(CLAMP(value->float32[2], 0.0f, 1.0f) * 0xffff)) & 0xffff;
> - clear_vals[1] |= ((uint16_t)util_iround(CLAMP(value->float32[3], 0.0f, 1.0f) * 0xffff)) << 16;
> - break;
> - case VK_FORMAT_R16G16B16A16_SNORM:
> - clear_vals[0] = ((uint16_t)util_iround(CLAMP(value->float32[0], -1.0f, 1.0f) * 0x7fff)) & 0xffff;
> - clear_vals[0] |= ((uint16_t)util_iround(CLAMP(value->float32[1], -1.0f, 1.0f) * 0x7fff)) << 16;
> - clear_vals[1] = ((uint16_t)util_iround(CLAMP(value->float32[2], -1.0f, 1.0f) * 0x7fff)) & 0xffff;
> - clear_vals[1] |= ((uint16_t)util_iround(CLAMP(value->float32[3], -1.0f, 1.0f) * 0x7fff)) << 16;
> - break;
> - case VK_FORMAT_A2B10G10R10_UNORM_PACK32:
> - clear_vals[0] = ((uint16_t)util_iround(CLAMP(value->float32[0], 0.0f, 1.0f) * 0x3ff)) & 0x3ff;
> - clear_vals[0] |= (((uint16_t)util_iround(CLAMP(value->float32[1], 0.0f, 1.0f) * 0x3ff)) & 0x3ff) << 10;
> - clear_vals[0] |= (((uint16_t)util_iround(CLAMP(value->float32[2], 0.0f, 1.0f) * 0x3ff)) & 0x3ff) << 20;
> - clear_vals[0] |= (((uint16_t)util_iround(CLAMP(value->float32[3], 0.0f, 1.0f) * 0x3)) & 0x3) << 30;
> - clear_vals[1] = 0;
> - return true;
> - case VK_FORMAT_R32G32_SFLOAT:
> - clear_vals[0] = fui(value->float32[0]);
> - clear_vals[1] = fui(value->float32[1]);
> - break;
> - case VK_FORMAT_R32_SFLOAT:
> - clear_vals[1] = 0;
> - clear_vals[0] = fui(value->float32[0]);
> - break;
> - case VK_FORMAT_B10G11R11_UFLOAT_PACK32:
> + if (format == VK_FORMAT_B10G11R11_UFLOAT_PACK32) {
> clear_vals[0] = float3_to_r11g11b10f(value->float32);
> clear_vals[1] = 0;
> - break;
> - case VK_FORMAT_R32G32B32A32_SFLOAT:
> - if (value->float32[0] != value->float32[1] ||
> - value->float32[0] != value->float32[2])
> - return false;
> - clear_vals[0] = fui(value->float32[0]);
> - clear_vals[1] = fui(value->float32[3]);
> - break;
> - case VK_FORMAT_R32G32B32A32_UINT:
> - if (value->uint32[0] != value->uint32[1] ||
> - value->uint32[0] != value->uint32[2])
> - return false;
> + return true;
> + }
> +
> + if (desc->layout != VK_FORMAT_LAYOUT_PLAIN) {
> + fprintf(stderr, "failed to fast clear for non-plain format %d\n", format);
> + return false;
> + }
> +
> + if (!util_is_power_of_two_or_zero(desc->block.bits)) {
> + fprintf(stderr, "failed to fast clear for NPOT format %d\n", format);
> + return false;
> + }
> +
> + if (desc->block.bits > 64) {
> + /*
> + * We have a 128 bits format, check if the first 3 components are the same.
> + * Every elements has to be 32 bits since we don't support 64-bit formats,
> + * and we can skip swizzling checks as alpha always comes last for these and
> + * we do not care about the rest as they have to be the same.
> + */
> + if (desc->channel[0].type == VK_FORMAT_TYPE_FLOAT) {
> + if (value->float32[0] != value->float32[1] ||
> + value->float32[0] != value->float32[2])
> + return false;
> + } else {
> + if (value->uint32[0] != value->uint32[1] ||
> + value->uint32[0] != value->uint32[2])
> + return false;
> + }
> clear_vals[0] = value->uint32[0];
> clear_vals[1] = value->uint32[3];
> - break;
> - case VK_FORMAT_R32G32B32A32_SINT:
> - if (value->int32[0] != value->int32[1] ||
> - value->int32[0] != value->int32[2])
> + return true;
> + }
> + uint64_t clear_val = 0;
> +
> + for (unsigned c = 0; c < 4; ++c) {
> + if (desc->swizzle[c] < 0 || desc->swizzle[c] >= 4)
> + continue;
> +
> + const struct vk_format_channel_description *channel = &desc->channel[desc->swizzle[c]];
> + assert(channel->size);
> +
> + uint64_t v = 0;
> + if (channel->pure_integer) {
> + v = value->uint32[c] & ((1ULL << channel->size) - 1);
> + } else if (channel->normalized) {
> + if (channel->type == VK_FORMAT_TYPE_UNSIGNED &&
> + desc->swizzle[c] < 3 &&
> + desc->colorspace == VK_FORMAT_COLORSPACE_SRGB) {
> + assert(channel->size == 8);
> +
> + v = util_format_linear_float_to_srgb_8unorm(value->float32[c]);
> + } else if (channel->type == VK_FORMAT_TYPE_UNSIGNED) {
> + v = MAX2(MIN2(value->float32[c], 1.0f), 0.0f) * ((1ULL << channel->size) - 1);
> + } else {
> + v = MAX2(MIN2(value->float32[c], 1.0f), -1.0f) * ((1ULL << (channel->size - 1)) - 1);
> + }
> + } else if (channel->type == VK_FORMAT_TYPE_FLOAT) {
> + if (channel->size == 32) {
> + memcpy(&v, &value->float32[c], 4);
> + } else if(channel->size == 16) {
> + v = util_float_to_half(value->float32[c]);
> + } else {
> + fprintf(stderr, "failed to fast clear for unhandled float size in format %d\n", format);
> + return false;
> + }
> + } else {
> + fprintf(stderr, "failed to fast clear for unhandled component type in format %d\n", format);
> return false;
> - clear_vals[0] = value->int32[0];
> - clear_vals[1] = value->int32[3];
> - break;
> - default:
> - fprintf(stderr, "failed to fast clear %d\n", format);
> - return false;
> + }
> + clear_val |= (v & ((1ULL << channel->size) - 1)) << channel->shift;
> }
> +
> + clear_vals[0] = clear_val;
> + clear_vals[1] = clear_val >> 32;
> +
> return true;
> }
>
> --
> 2.17.0
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/mesa-dev
More information about the mesa-dev
mailing list