From: Brian Starkey <rodrigosiqueiramelo@xxxxxxxxx> To use writeback buffers as a CRC source, we need to be able to hash them. Implement a simple FVA-1a hashing routine for this purpose. Doing a bytewise hash on the framebuffer directly can be very slow if the memory is noncached. By making a copy of each line in the FB first (which can take advantage of word-access speedup), we can do the hash on a cached copy, which is much faster (10x speedup on my platform). Changes since V7 (Maxime Ripard): * Maxime Ripard: - Replace `gem_mmap__gtt` by `igt_fb_map_buffer` on `igt_fb_get_crc()` * Rodrigo Siqueira: - Rename igt_fb_get_crc to igt_fb_get_fnv1a_crc Changes since V6 (Simon Ser): * Use plain uint32_t for FNV1a_OFFSET_BIAS and FNV1a_PRIME * Return -EINVAL in case fb->num_planes != 1 * Just copy fb->width * cpp instead of copy the whole stride v5: use igt_memcpy_from_wc() instead of plain memcpy, as suggested by Chris Wilson Signed-off-by: Brian Starkey <brian.starkey@xxxxxxx> [rebased and updated to the most recent API] Signed-off-by: Liviu Dudau <liviu.dudau@xxxxxxx> [rebased and updated] Signed-off-by: Rodrigo Siqueira <rodrigosiqueiramelo@xxxxxxxxx> Reviewed-by: Simon Ser <simon.ser@xxxxxxxxx> Reviewed-by: Liviu Dudau <liviu.dudau@xxxxxxx> --- lib/igt_fb.c | 63 ++++++++++++++++++++++++++++++++++++++++++++++++++++ lib/igt_fb.h | 2 ++ 2 files changed, 65 insertions(+) diff --git a/lib/igt_fb.c b/lib/igt_fb.c index 3864b7a1..5e3706e9 100644 --- a/lib/igt_fb.c +++ b/lib/igt_fb.c @@ -3783,6 +3783,69 @@ bool igt_fb_supported_format(uint32_t drm_format) return true; } +/* + * This implements the FNV-1a hashing algorithm instead of CRC, for + * simplicity + * http://www.isthe.com/chongo/tech/comp/fnv/index.html + * + * hash = offset_basis + * for each octet_of_data to be hashed + * hash = hash xor octet_of_data + * hash = hash * FNV_prime + * return hash + * + * 32 bit offset_basis = 2166136261 + * 32 bit FNV_prime = 224 + 28 + 0x93 = 16777619 + */ +int igt_fb_get_fnv1a_crc(struct igt_fb *fb, igt_crc_t *crc) +{ + const uint32_t FNV1a_OFFSET_BIAS = 2166136261; + const uint32_t FNV1a_PRIME = 16777619; + uint32_t hash; + void *map; + char *ptr, *line = NULL; + int x, y, cpp = igt_drm_format_to_bpp(fb->drm_format) / 8; + uint32_t stride = calc_plane_stride(fb, 0); + + if (fb->num_planes != 1) + return -EINVAL; + + ptr = igt_fb_map_buffer(fb->fd, fb); + igt_assert(ptr); + map = ptr; + + /* + * Framebuffers are often uncached, which can make byte-wise accesses + * very slow. We copy each line of the FB into a local buffer to speed + * up the hashing. + */ + line = malloc(stride); + if (!line) { + munmap(map, fb->size); + return -ENOMEM; + } + + hash = FNV1a_OFFSET_BIAS; + + for (y = 0; y < fb->height; y++, ptr += stride) { + + igt_memcpy_from_wc(line, ptr, fb->width * cpp); + + for (x = 0; x < fb->width * cpp; x++) { + hash ^= line[x]; + hash *= FNV1a_PRIME; + } + } + + crc->n_words = 1; + crc->crc[0] = hash; + + free(line); + igt_fb_unmap_buffer(fb, map); + + return 0; +} + /** * igt_format_is_yuv: * @drm_format: drm fourcc diff --git a/lib/igt_fb.h b/lib/igt_fb.h index 5ec906b7..debfb5ad 100644 --- a/lib/igt_fb.h +++ b/lib/igt_fb.h @@ -207,5 +207,7 @@ int igt_format_plane_bpp(uint32_t drm_format, int plane); void igt_format_array_fill(uint32_t **formats_array, unsigned int *count, bool allow_yuv); +int igt_fb_get_fnv1a_crc(struct igt_fb *fb, igt_crc_t *crc); + #endif /* __IGT_FB_H__ */ -- 2.28.0 _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx