Message ID | 20170405015328.5476-3-jdarnley@obe.tv |
---|---|
State | Superseded |
Headers | show |
Hi, On Tue, Apr 4, 2017 at 9:53 PM, James Darnley <jdarnley@obe.tv> wrote: > The labels get stripped leading to (slightly) nicer disassembly from > objdump. > [..] > - jz .cycle%1end > + jz %%skip Can you preserve the leading dot? I don't mind the %%skip, but please make it .%%skip. Ronald
On 2017-04-05 13:41, Ronald S. Bultje wrote: > Hi, > > On Tue, Apr 4, 2017 at 9:53 PM, James Darnley <jdarnley@obe.tv> wrote: > >> The labels get stripped leading to (slightly) nicer disassembly from >> objdump. >> > [..] > >> - jz .cycle%1end >> + jz %%skip > > > Can you preserve the leading dot? I don't mind the %%skip, but please make > it .%%skip. That makes the patch pointless because those symbols don't get stripped. If you want the leading dot then I will drop this patch.
Hi, On Thu, Apr 6, 2017 at 10:22 AM, James Darnley <jdarnley@obe.tv> wrote: > On 2017-04-05 13:41, Ronald S. Bultje wrote: > > Hi, > > > > On Tue, Apr 4, 2017 at 9:53 PM, James Darnley <jdarnley@obe.tv> wrote: > > > >> The labels get stripped leading to (slightly) nicer disassembly from > >> objdump. > >> > > [..] > > > >> - jz .cycle%1end > >> + jz %%skip > > > > > > Can you preserve the leading dot? I don't mind the %%skip, but please > make > > it .%%skip. > > That makes the patch pointless because those symbols don't get stripped. > If you want the leading dot then I will drop this patch. Oh I see, right... So, yes, that has disadvantages as well as advantages. I don't really have an opinion on it. So if nobody else does either, and you like it, I guess you could commit it? Ronald
diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm index 878ff02..dde40e9 100644 --- a/libavcodec/x86/h264_idct.asm +++ b/libavcodec/x86/h264_idct.asm @@ -846,7 +846,7 @@ h264_add8x4_idct_sse2: %macro add16_sse2_cycle 2 movzx r0, word [r4+%2] test r0, r0 - jz .cycle%1end + jz %%skip mov r0d, dword [r1+%1*8] %if ARCH_X86_64 add r0, r5 @@ -854,7 +854,7 @@ h264_add8x4_idct_sse2: add r0, r0m %endif call h264_add8x4_idct_sse2 -.cycle%1end: +%%skip: %if %1 < 7 add r2, 64 %endif @@ -883,7 +883,7 @@ REP_RET %macro add16intra_sse2_cycle 2 movzx r0, word [r4+%2] test r0, r0 - jz .try%1dc + jz %%trydc mov r0d, dword [r1+%1*8] %if ARCH_X86_64 add r0, r7 @@ -891,11 +891,11 @@ REP_RET add r0, r0m %endif call h264_add8x4_idct_sse2 - jmp .cycle%1end -.try%1dc: + jmp %%skip +%%trydc: movsx r0, word [r2 ] or r0w, word [r2+32] - jz .cycle%1end + jz %%skip mov r0d, dword [r1+%1*8] %if ARCH_X86_64 add r0, r7 @@ -903,7 +903,7 @@ REP_RET add r0, r0m %endif call h264_idct_dc_add8_mmxext -.cycle%1end: +%%skip: %if %1 < 7 add r2, 64 %endif @@ -930,7 +930,7 @@ REP_RET %macro add8_sse2_cycle 2 movzx r0, word [r4+%2] test r0, r0 - jz .try%1dc + jz %%trydc %if ARCH_X86_64 mov r0d, dword [r1+(%1&1)*8+64*(1+(%1>>1))] add r0, [r7] @@ -940,11 +940,11 @@ REP_RET add r0, dword [r1+(%1&1)*8+64*(1+(%1>>1))] %endif call h264_add8x4_idct_sse2 - jmp .cycle%1end -.try%1dc: + jmp %%cycle_end +%%trydc: movsx r0, word [r2 ] or r0w, word [r2+32] - jz .cycle%1end + jz %%cycle_end %if ARCH_X86_64 mov r0d, dword [r1+(%1&1)*8+64*(1+(%1>>1))] add r0, [r7] @@ -954,7 +954,7 @@ REP_RET add r0, dword [r1+(%1&1)*8+64*(1+(%1>>1))] %endif call h264_idct_dc_add8_mmxext -.cycle%1end: +%%cycle_end: %if %1 == 1 add r2, 384+64 %elif %1 < 3