From: Henrik Gramner <[email protected]>
Store XMM6 and XMM7 in the shadow space in functions that clobbers them.
This way we don't have to adjust the stack pointer as often,
reducing the number of instructions as well as code size.
---
libavcodec/x86/fft.asm | 4 +-
libavcodec/x86/h264_deblock.asm | 19 ++++------
libavutil/x86/x86inc.asm | 83 +++++++++++++++++++++++------------------
3 files changed, 55 insertions(+), 51 deletions(-)
diff --git a/libavcodec/x86/fft.asm b/libavcodec/x86/fft.asm
index c87752b..e4744a3 100644
--- a/libavcodec/x86/fft.asm
+++ b/libavcodec/x86/fft.asm
@@ -667,13 +667,13 @@ cglobal imdct_calc, 3,5,3
push r1
push r0
%else
- sub rsp, 8
+ sub rsp, 8+32*WIN64 ; allocate win64 shadow space
%endif
call r4
%if ARCH_X86_32
add esp, 12
%else
- add rsp, 8
+ add rsp, 8+32*WIN64
%endif
POP r1
POP r3
diff --git a/libavcodec/x86/h264_deblock.asm b/libavcodec/x86/h264_deblock.asm
index fc6c983..6e29ce7 100644
--- a/libavcodec/x86/h264_deblock.asm
+++ b/libavcodec/x86/h264_deblock.asm
@@ -331,16 +331,14 @@ cglobal deblock_v_luma_8, 5,5,10
; void deblock_h_luma( uint8_t *pix, int stride, int alpha, int beta, int8_t
*tc0 )
;-----------------------------------------------------------------------------
INIT_MMX cpuname
-cglobal deblock_h_luma_8, 5,9
+cglobal deblock_h_luma_8, 5,9,0,0x60+16*WIN64
movsxd r7, r1d
lea r8, [r7+r7*2]
lea r6, [r0-4]
lea r5, [r0-4+r8]
%if WIN64
- sub rsp, 0x98
- %define pix_tmp rsp+0x30
+ %define pix_tmp rsp+0x30 ; shadow space + r4
%else
- sub rsp, 0x68
%define pix_tmp rsp
%endif
@@ -379,11 +377,6 @@ cglobal deblock_h_luma_8, 5,9
movq m3, [pix_tmp+0x40]
TRANSPOSE8x4B_STORE PASS8ROWS(r6, r5, r7, r8)
-%if WIN64
- add rsp, 0x98
-%else
- add rsp, 0x68
-%endif
RET
%endmacro
@@ -704,13 +697,16 @@ INIT_MMX cpuname
;-----------------------------------------------------------------------------
; void deblock_h_luma_intra( uint8_t *pix, int stride, int alpha, int beta )
;-----------------------------------------------------------------------------
-cglobal deblock_h_luma_intra_8, 4,9
+cglobal deblock_h_luma_intra_8, 4,9,0,0x80
movsxd r7, r1d
lea r8, [r7*3]
lea r6, [r0-4]
lea r5, [r0-4+r8]
- sub rsp, 0x88
+%if WIN64
+ %define pix_tmp rsp+0x20 ; shadow space
+%else
%define pix_tmp rsp
+%endif
; transpose 8x16 -> tmp space
TRANSPOSE8x8_MEM PASS8ROWS(r6, r5, r7, r8), PASS8ROWS(pix_tmp,
pix_tmp+0x30, 0x10, 0x30)
@@ -730,7 +726,6 @@ cglobal deblock_h_luma_intra_8, 4,9
sub r5, r7
shr r7, 3
TRANSPOSE8x8_MEM PASS8ROWS(pix_tmp, pix_tmp+0x30, 0x10, 0x30),
PASS8ROWS(r6, r5, r7, r8)
- add rsp, 0x88
RET
%else
cglobal deblock_h_luma_intra_8, 2,4,8,0x80
diff --git a/libavutil/x86/x86inc.asm b/libavutil/x86/x86inc.asm
index 633a140..860906c 100644
--- a/libavutil/x86/x86inc.asm
+++ b/libavutil/x86/x86inc.asm
@@ -317,14 +317,18 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
%if stack_size < 0
%assign stack_size -stack_size
%endif
- %if mmsize != 8
- %assign xmm_regs_used %2
+ %assign stack_size_padded stack_size
+ %if WIN64
+ %assign stack_size_padded stack_size_padded + 32 ; reserve 32
bytes for shadow space
+ %if mmsize != 8
+ %assign xmm_regs_used %2
+ %if xmm_regs_used > 8
+ %assign stack_size_padded stack_size_padded +
(xmm_regs_used-8)*16
+ %endif
+ %endif
%endif
%if mmsize <= 16 && HAVE_ALIGNED_STACK
- %assign stack_size_padded stack_size + %%stack_alignment -
gprsize - (stack_offset & (%%stack_alignment - 1))
- %if xmm_regs_used > 6
- %assign stack_size_padded stack_size_padded +
(xmm_regs_used - 6) * 16
- %endif
+ %assign stack_size_padded stack_size_padded +
%%stack_alignment - gprsize - (stack_offset & (%%stack_alignment - 1))
SUB rsp, stack_size_padded
%else
%assign %%reg_num (regs_used - 1)
@@ -334,14 +338,6 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
; stack in a single instruction (i.e. mov rsp, rstk or mov
; rsp, [rsp+stack_size_padded])
mov rstk, rsp
- %assign stack_size_padded stack_size
- %if xmm_regs_used > 6
- %assign stack_size_padded stack_size_padded +
(xmm_regs_used - 6) * 16
- %if mmsize == 32 && xmm_regs_used & 1
- ; re-align to 32 bytes
- %assign stack_size_padded (stack_size_padded + 16)
- %endif
- %endif
%if %1 < 0 ; need to store rsp on stack
sub rsp, gprsize+stack_size_padded
and rsp, ~(%%stack_alignment-1)
@@ -353,9 +349,7 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
%xdefine rstkm rstk
%endif
%endif
- %if xmm_regs_used > 6
- WIN64_PUSH_XMM
- %endif
+ WIN64_PUSH_XMM
%endif
%endif
%endmacro
@@ -416,40 +410,55 @@ DECLARE_REG 14, R15, 120
%endmacro
%macro WIN64_PUSH_XMM 0
- %assign %%i xmm_regs_used
- %rep (xmm_regs_used-6)
- %assign %%i %%i-1
- movaps [rsp + (%%i-6)*16 + stack_size + (~stack_offset&8)], xmm %+ %%i
- %endrep
+ ; Use the shadow space to store XMM6 and XMM7, the rest needs stack space
allocated.
+ %if xmm_regs_used > 6
+ movaps [rstk + stack_offset + 8], xmm6
+ %endif
+ %if xmm_regs_used > 7
+ movaps [rstk + stack_offset + 24], xmm7
+ %endif
+ %if xmm_regs_used > 8
+ %assign %%i 8
+ %rep xmm_regs_used-8
+ movaps [rsp + (%%i-8)*16 + stack_size + 32], xmm %+ %%i
+ %assign %%i %%i+1
+ %endrep
+ %endif
%endmacro
%macro WIN64_SPILL_XMM 1
%assign xmm_regs_used %1
ASSERT xmm_regs_used <= 16
- %if xmm_regs_used > 6
- SUB rsp, (xmm_regs_used-6)*16+16
- WIN64_PUSH_XMM
+ %if xmm_regs_used > 8
+ %assign stack_size_padded (xmm_regs_used-8)*16 + (~stack_offset&8) + 32
+ SUB rsp, stack_size_padded
%endif
+ WIN64_PUSH_XMM
%endmacro
%macro WIN64_RESTORE_XMM_INTERNAL 1
- %if xmm_regs_used > 6
+ %assign %%pad_size 0
+ %if xmm_regs_used > 8
%assign %%i xmm_regs_used
- %rep (xmm_regs_used-6)
+ %rep xmm_regs_used-8
%assign %%i %%i-1
- movaps xmm %+ %%i, [%1 + (%%i-6)*16+stack_size+(~stack_offset&8)]
+ movaps xmm %+ %%i, [%1 + (%%i-8)*16 + stack_size + 32]
%endrep
- %if stack_size_padded == 0
- add %1, (xmm_regs_used-6)*16+16
- %endif
%endif
%if stack_size_padded > 0
%if stack_size > 0 && (mmsize == 32 || HAVE_ALIGNED_STACK == 0)
mov rsp, rstkm
%else
add %1, stack_size_padded
+ %assign %%pad_size stack_size_padded
%endif
%endif
+ %if xmm_regs_used > 7
+ movaps xmm7, [%1 + stack_offset - %%pad_size + 24]
+ %endif
+ %if xmm_regs_used > 6
+ movaps xmm6, [%1 + stack_offset - %%pad_size + 8]
+ %endif
%endmacro
%macro WIN64_RESTORE_XMM 1
@@ -666,12 +675,12 @@ BRANCH_INSTR jz, je, jnz, jne, jl, jle, jnl, jnle, jg,
jge, jng, jnge, ja, jae,
%endif
align function_align
%2:
- RESET_MM_PERMUTATION ; not really needed, but makes disassembly somewhat
nicer
- %xdefine rstk rsp
- %assign stack_offset 0
- %assign stack_size 0
- %assign stack_size_padded 0
- %assign xmm_regs_used 0
+ RESET_MM_PERMUTATION ; needed for x86-64, also makes disassembly
somewhat nicer
+ %xdefine rstk rsp ; copy of the original stack pointer, used
when greater alignment than the known stack alignment is required
+ %assign stack_offset 0 ; stack pointer offset relative to the return
address
+ %assign stack_size 0 ; amount of stack space that can be freely
used inside a function
+ %assign stack_size_padded 0 ; total amount of allocated stack space,
including space for callee-saved xmm registers on WIN64 and alignment padding
+ %assign xmm_regs_used 0 ; number of XMM registers requested, used for
dealing with callee-saved registers on WIN64
%ifnidn %3, ""
PROLOGUE %3
%endif
--
1.8.3.msysgit.0
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
