From
d8b81d5393df36085009bf9f69d41fa85e2ae58a Mon Sep 17 00:00:00 2001
From: Nitin A Kamble <nitin.a.kamble@intel.com>
Date: Sat, 10 Dec 2011 09:09:06 +0100
Make assembly syntax compatible to the X32 toolchain and fix the
following kind of compilations errors with X32 gcc.
| pulsecore/svolume_mmx.c: Assembler messages:
| pulsecore/svolume_mmx.c:107: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:135: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:161: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:162: Error: `8(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:180: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:210: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:244: Error: `(%esi,%rdi,4)' is not a valid base/index expression
| pulsecore/svolume_mmx.c:245: Error: `8(%esi,%rdi,4)' is not a valid base/index expression
| make[3]: *** [libpulsecore_1.1_la-svolume_mmx.lo] Error 1
Originally these assembly lines were written for x86_64 ABI, now they
are also compatible with X32 ABI [3][4].
The patch was submitted to the OpenEmbedded-Core list [1][2].
[1] http://lists.linuxtogo.org/pipermail/openembedded-core/2011-December/014189.html
[2] http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=nitin/x32&id=
2d8eec54f755c51f2eff600390f5a4b3cc2a7662
[3] https://wiki.yoctoproject.org/wiki/X32_abi
[4] http://en.wikipedia.org/wiki/X32_ABI
" je 2f \n\t"
"1: \n\t"
- " movups (%2, %0, 2), %%xmm0 \n\t" /* read 8 floats */
- " movups 16(%2, %0, 2), %%xmm2 \n\t"
+ " movups (%q2, %0, 2), %%xmm0 \n\t" /* read 8 floats */
+ " movups 16(%q2, %0, 2), %%xmm2 \n\t"
" minps %%xmm5, %%xmm0 \n\t" /* clamp to 1.0 */
" minps %%xmm5, %%xmm2 \n\t"
" maxps %%xmm6, %%xmm0 \n\t" /* clamp to -1.0 */
" packssdw %%mm1, %%mm0 \n\t" /* pack parts */
" packssdw %%mm3, %%mm2 \n\t"
- " movq %%mm0, (%3, %0) \n\t"
- " movq %%mm2, 8(%3, %0) \n\t"
+ " movq %%mm0, (%q3, %0) \n\t"
+ " movq %%mm2, 8(%q3, %0) \n\t"
" add $16, %0 \n\t"
" dec %1 \n\t"
" je 4f \n\t"
"3: \n\t"
- " movss (%2, %0, 2), %%xmm0 \n\t"
+ " movss (%q2, %0, 2), %%xmm0 \n\t"
" minss %%xmm5, %%xmm0 \n\t"
" maxss %%xmm6, %%xmm0 \n\t"
" mulss %%xmm7, %%xmm0 \n\t"
" cvtss2si %%xmm0, %4 \n\t"
- " movw %w4, (%3, %0) \n\t"
+ " movw %w4, (%q3, %0) \n\t"
" add $2, %0 \n\t"
" dec %1 \n\t"
" jne 3b \n\t"
" je 2f \n\t"
"1: \n\t"
- " movups (%2, %0, 2), %%xmm0 \n\t" /* read 8 floats */
- " movups 16(%2, %0, 2), %%xmm2 \n\t"
+ " movups (%q2, %0, 2), %%xmm0 \n\t" /* read 8 floats */
+ " movups 16(%q2, %0, 2), %%xmm2 \n\t"
" minps %%xmm5, %%xmm0 \n\t" /* clamp to 1.0 */
" minps %%xmm5, %%xmm2 \n\t"
" maxps %%xmm6, %%xmm0 \n\t" /* clamp to -1.0 */
" cvtps2dq %%xmm2, %%xmm2 \n\t"
" packssdw %%xmm2, %%xmm0 \n\t"
- " movdqu %%xmm0, (%3, %0) \n\t"
+ " movdqu %%xmm0, (%q3, %0) \n\t"
" add $16, %0 \n\t"
" dec %1 \n\t"
" je 4f \n\t"
"3: \n\t"
- " movss (%2, %0, 2), %%xmm0 \n\t"
+ " movss (%q2, %0, 2), %%xmm0 \n\t"
" minss %%xmm5, %%xmm0 \n\t"
" maxss %%xmm6, %%xmm0 \n\t"
" mulss %%xmm7, %%xmm0 \n\t"
" cvtss2si %%xmm0, %4 \n\t"
- " movw %w4, (%3, %0) \n\t"
+ " movw %w4, (%q3, %0) \n\t"
" add $2, %0 \n\t"
" dec %1 \n\t"
" jne 3b \n\t"
" test $1, %2 \n\t" /* check for odd samples */
" je 2f \n\t"
- " movd (%1, %3, 4), %%mm0 \n\t" /* | v0h | v0l | */
+ " movd (%q1, %3, 4), %%mm0 \n\t" /* | v0h | v0l | */
" movw (%0), %w4 \n\t" /* .. | p0 | */
" movd %4, %%mm1 \n\t"
VOLUME_32x16 (%%mm1, %%mm0)
" je 4f \n\t"
"3: \n\t" /* do samples in groups of 2 */
- " movq (%1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq (%q1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
" movd (%0), %%mm1 \n\t" /* .. | p1 | p0 | */
VOLUME_32x16 (%%mm1, %%mm0)
" movd %%mm0, (%0) \n\t" /* .. | p1*v1 | p0*v0 | */
" je 6f \n\t"
"5: \n\t" /* do samples in groups of 4 */
- " movq (%1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
- " movq 8(%1, %3, 4), %%mm2 \n\t" /* | v3h | v3l | v2h | v2l | */
+ " movq (%q1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq 8(%q1, %3, 4), %%mm2 \n\t" /* | v3h | v3l | v2h | v2l | */
" movd (%0), %%mm1 \n\t" /* .. | p1 | p0 | */
" movd 4(%0), %%mm3 \n\t" /* .. | p3 | p2 | */
VOLUME_32x16 (%%mm1, %%mm0)
" test $1, %2 \n\t" /* check for odd samples */
" je 2f \n\t"
- " movd (%1, %3, 4), %%mm0 \n\t" /* | v0h | v0l | */
+ " movd (%q1, %3, 4), %%mm0 \n\t" /* | v0h | v0l | */
" movw (%0), %w4 \n\t" /* .. | p0 | */
" rorw $8, %w4 \n\t"
" movd %4, %%mm1 \n\t"
" je 4f \n\t"
"3: \n\t" /* do samples in groups of 2 */
- " movq (%1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq (%q1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
" movd (%0), %%mm1 \n\t" /* .. | p1 | p0 | */
SWAP_16 (%%mm1)
VOLUME_32x16 (%%mm1, %%mm0)
" je 6f \n\t"
"5: \n\t" /* do samples in groups of 4 */
- " movq (%1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
- " movq 8(%1, %3, 4), %%mm2 \n\t" /* | v3h | v3l | v2h | v2l | */
+ " movq (%q1, %3, 4), %%mm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq 8(%q1, %3, 4), %%mm2 \n\t" /* | v3h | v3l | v2h | v2l | */
" movd (%0), %%mm1 \n\t" /* .. | p1 | p0 | */
" movd 4(%0), %%mm3 \n\t" /* .. | p3 | p2 | */
SWAP_16_2 (%%mm1, %%mm3)
" test $1, %2 \n\t" /* check for odd samples */
" je 2f \n\t"
- " movd (%1, %3, 4), %%xmm0 \n\t" /* | v0h | v0l | */
+ " movd (%q1, %3, 4), %%xmm0 \n\t" /* | v0h | v0l | */
" movw (%0), %w4 \n\t" /* .. | p0 | */
" movd %4, %%xmm1 \n\t"
VOLUME_32x16 (%%xmm1, %%xmm0)
" je 4f \n\t"
"3: \n\t" /* do samples in groups of 2 */
- " movq (%1, %3, 4), %%xmm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq (%q1, %3, 4), %%xmm0 \n\t" /* | v1h | v1l | v0h | v0l | */
" movd (%0), %%xmm1 \n\t" /* .. | p1 | p0 | */
VOLUME_32x16 (%%xmm1, %%xmm0)
" movd %%xmm0, (%0) \n\t" /* .. | p1*v1 | p0*v0 | */
* that the array is 16 bytes aligned, we probably have to do the odd values
* after this then. */
"5: \n\t" /* do samples in groups of 4 */
- " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
+ " movdqu (%q1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
" movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */
VOLUME_32x16 (%%xmm1, %%xmm0)
" movq %%xmm0, (%0) \n\t" /* .. | p3*v3 .. p0*v0 | */
" je 8f \n\t"
"7: \n\t" /* do samples in groups of 8 */
- " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
- " movdqu 16(%1, %3, 4), %%xmm2 \n\t" /* | v7h | v7l .. v4h | v4l | */
+ " movdqu (%q1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
+ " movdqu 16(%q1, %3, 4), %%xmm2 \n\t" /* | v7h | v7l .. v4h | v4l | */
" movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */
" movq 8(%0), %%xmm3 \n\t" /* .. | p7 .. p4 | */
VOLUME_32x16 (%%xmm1, %%xmm0)
" test $1, %2 \n\t" /* check for odd samples */
" je 2f \n\t"
- " movd (%1, %3, 4), %%xmm0 \n\t" /* | v0h | v0l | */
+ " movd (%q1, %3, 4), %%xmm0 \n\t" /* | v0h | v0l | */
" movw (%0), %w4 \n\t" /* .. | p0 | */
" rorw $8, %w4 \n\t"
" movd %4, %%xmm1 \n\t"
" je 4f \n\t"
"3: \n\t" /* do samples in groups of 2 */
- " movq (%1, %3, 4), %%xmm0 \n\t" /* | v1h | v1l | v0h | v0l | */
+ " movq (%q1, %3, 4), %%xmm0 \n\t" /* | v1h | v1l | v0h | v0l | */
" movd (%0), %%xmm1 \n\t" /* .. | p1 | p0 | */
SWAP_16 (%%xmm1)
VOLUME_32x16 (%%xmm1, %%xmm0)
* that the array is 16 bytes aligned, we probably have to do the odd values
* after this then. */
"5: \n\t" /* do samples in groups of 4 */
- " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
+ " movdqu (%q1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
" movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */
SWAP_16 (%%xmm1)
VOLUME_32x16 (%%xmm1, %%xmm0)
" je 8f \n\t"
"7: \n\t" /* do samples in groups of 8 */
- " movdqu (%1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
- " movdqu 16(%1, %3, 4), %%xmm2 \n\t" /* | v7h | v7l .. v4h | v4l | */
+ " movdqu (%q1, %3, 4), %%xmm0 \n\t" /* | v3h | v3l .. v0h | v0l | */
+ " movdqu 16(%q1, %3, 4), %%xmm2 \n\t" /* | v7h | v7l .. v4h | v4l | */
" movq (%0), %%xmm1 \n\t" /* .. | p3 .. p0 | */
" movq 8(%0), %%xmm3 \n\t" /* .. | p7 .. p4 | */
SWAP_16_2 (%%xmm1, %%xmm3)