Aurelien
/
xf86-video-intel
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

sna: Invert the function wrapping for sse64xN/sse64 

We should be consistent in making the code simpler for the compiler and
so not rely on it eliminating the dead-code for a single loop of
sse64xN!

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This commit is contained in:
Chris Wilson 2016-04-08 07:56:07 +01:00
parent 59d371a9b2
commit 65c72d9871
1 changed files with 34 additions and 36 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

70
src/sna/blt.c
View File

@ -353,35 +353,34 @@ memcpy_from_tiled_x__swizzle_0(const void *src, void *dst, int bpp,
#if defined(sse2) && defined(__x86_64__)
sse2 static force_inline void
to_sse64(uint8_t *dst, const uint8_t *src)
{
__m128i xmm1, xmm2, xmm3, xmm4;
xmm1 = xmm_load_128u((const __m128i*)src + 0);
xmm2 = xmm_load_128u((const __m128i*)src + 1);
xmm3 = xmm_load_128u((const __m128i*)src + 2);
xmm4 = xmm_load_128u((const __m128i*)src + 3);
xmm_save_128((__m128i*)dst + 0, xmm1);
xmm_save_128((__m128i*)dst + 1, xmm2);
xmm_save_128((__m128i*)dst + 2, xmm3);
xmm_save_128((__m128i*)dst + 3, xmm4);
}
sse2 static force_inline void
to_sse64xN(uint8_t *dst, const uint8_t *src, int bytes)
{
int i;
for (i = 0; i < bytes / 64; i++) {
__m128i xmm1, xmm2, xmm3, xmm4;
xmm1 = xmm_load_128u((const __m128i*)src + 0);
xmm2 = xmm_load_128u((const __m128i*)src + 1);
xmm3 = xmm_load_128u((const __m128i*)src + 2);
xmm4 = xmm_load_128u((const __m128i*)src + 3);
xmm_save_128((__m128i*)dst + 0, xmm1);
xmm_save_128((__m128i*)dst + 1, xmm2);
xmm_save_128((__m128i*)dst + 2, xmm3);
xmm_save_128((__m128i*)dst + 3, xmm4);
to_sse64(dst, src);
dst += 64;
src += 64;
}
}
sse2 static force_inline void
to_sse64(uint8_t *dst, const uint8_t *src)
{
to_sse64xN(dst, src, 64);
}
sse2 static force_inline void
to_sse32(uint8_t *dst, const uint8_t *src)
{
@ -475,35 +474,34 @@ memcpy_to_tiled_x__swizzle_0__sse2(const void *src, void *dst, int bpp,
}
}
sse2 static force_inline void
from_sse64(uint8_t *dst, const uint8_t *src)
{
__m128i xmm1, xmm2, xmm3, xmm4;
xmm1 = xmm_load_128((const __m128i*)src + 0);
xmm2 = xmm_load_128((const __m128i*)src + 1);
xmm3 = xmm_load_128((const __m128i*)src + 2);
xmm4 = xmm_load_128((const __m128i*)src + 3);
xmm_save_128u((__m128i*)dst + 0, xmm1);
xmm_save_128u((__m128i*)dst + 1, xmm2);
xmm_save_128u((__m128i*)dst + 2, xmm3);
xmm_save_128u((__m128i*)dst + 3, xmm4);
}
sse2 static force_inline void
from_sse64xN(uint8_t *dst, const uint8_t *src, int bytes)
{
int i;
for (i = 0; i < bytes / 64; i++) {
__m128i xmm1, xmm2, xmm3, xmm4;
xmm1 = xmm_load_128((const __m128i*)src + 0);
xmm2 = xmm_load_128((const __m128i*)src + 1);
xmm3 = xmm_load_128((const __m128i*)src + 2);
xmm4 = xmm_load_128((const __m128i*)src + 3);
xmm_save_128u((__m128i*)dst + 0, xmm1);
xmm_save_128u((__m128i*)dst + 1, xmm2);
xmm_save_128u((__m128i*)dst + 2, xmm3);
xmm_save_128u((__m128i*)dst + 3, xmm4);
from_sse64(dst, src);
dst += 64;
src += 64;
}
}
sse2 static force_inline void
from_sse64(uint8_t *dst, const uint8_t *src)
{
from_sse64xN(dst, src, 64);
}
sse2 static force_inline void
from_sse32(uint8_t *dst, const uint8_t *src)
{