Intravision/reactos
1
0
Fork 0
mirror of https://github.com/reactos/reactos.git synced 2025-08-05 09:13:01 +00:00
Code Issues Projects Releases Packages Wiki Activity

[LIBM] Fix up some asm files

Browse source 
This allows to compile them with GAS after translation
This commit is contained in:
Timo Kreuzer 2022-06-25 13:00:01 +03:00
parent af6d3bb18f
commit 105426b81a
5 changed files with 32 additions and 32 deletions
Download patch file Download diff file Expand all files Collapse all files

10
sdk/lib/crt/math/libm_sse2/fmod.asm
View file

@ -90,12 +90,12 @@ fmod PROC FRAME
fnclex ; clear exception flags fnclex ; clear exception flags
; in preparation for fprem ; in preparation for fprem
@@: @again:
DB 0d9h, 0f8h ; fprem DB 0d9h, 0f8h ; fprem
DB 09bh, 0dfh, 0e0h ; fstsw ax DB 09bh, 0dfh, 0e0h ; fstsw ax
test ax, 4 SHL X87SW_C test ax, 4 SHL X87SW_C
jnz @b ; do it again in case of partial result jnz @again ; do it again in case of partial result
DB 0ddh, 01ch, 024h ; fstp QWORD PTR [rsp] DB 0ddh, 01ch, 024h ; fstp QWORD PTR [rsp]
movlpd xmm0, QWORD PTR [rsp] ; result movlpd xmm0, QWORD PTR [rsp] ; result
@ -133,14 +133,14 @@ fmod PROC FRAME
@raise: @raise:
mov eax, INVALID ; raise exception mov eax, INVALID ; raise exception
mov r8, FPIND mov r8, FPIND
jmp @f jmp @fail
@error: @error:
xor eax, eax ; no exception xor eax, eax ; no exception
movd r8, xmm0 movd r8, xmm0
jmp @f jmp @fail
@@: @fail:
lea rcx, [@fmodz] ; fname lea rcx, [@fmodz] ; fname
mov edx, FPCODEFMOD ; opcode mov edx, FPCODEFMOD ; opcode
; mov r8, INDEF ; value ; mov r8, INDEF ; value

10
sdk/lib/crt/math/libm_sse2/fmodf.asm
View file

@ -90,12 +90,12 @@ fmodf PROC FRAME
fnclex ; clear exception flags fnclex ; clear exception flags
; in preparation for fprem ; in preparation for fprem
@@: @again:
DB 0d9h, 0f8h ; fprem DB 0d9h, 0f8h ; fprem
DB 9bh, 0dfh, 0e0h ; fstsw ax DB 9bh, 0dfh, 0e0h ; fstsw ax
test ax, 00400h test ax, 00400h
jnz @b ; do it again in case of partial result jnz @again ; do it again in case of partial result
DB 0d9h, 1ch, 24h ; fstp DWORD PTR [rsp] DB 0d9h, 1ch, 24h ; fstp DWORD PTR [rsp]
movss xmm0, DWORD PTR [rsp] ; result movss xmm0, DWORD PTR [rsp] ; result
@ -133,14 +133,14 @@ fmodf PROC FRAME
@raise: @raise:
mov eax, INVALID ; raise exception mov eax, INVALID ; raise exception
mov r8d, FPIND mov r8d, FPIND
jmp @f jmp @fail
@error: @error:
xor eax, eax ; no exception xor eax, eax ; no exception
movd r8d, xmm0 movd r8d, xmm0
jmp @f jmp @fail
@@: @fail:
lea rcx, [@fmodfz] ; fname lea rcx, [@fmodfz] ; fname
mov edx, FPCODEFMOD ; opcode mov edx, FPCODEFMOD ; opcode
; mov r8d, [rsp] ; value ; mov r8d, [rsp] ; value

20
sdk/lib/crt/math/libm_sse2/log.asm
View file

@ -391,11 +391,11 @@ Llog_fma3:
xor rax,rax xor rax,rax
vpsrlq xmm3,xmm0,52 vpsrlq xmm3,xmm0,52
vmovq rax,xmm0 vmovq rax,xmm0
vpsubq xmm3,xmm3,QWORD PTR __mask_1023 vpsubq xmm3,xmm3,XMMWORD PTR __mask_1023
vcvtdq2pd xmm6,xmm3 ; xexp vcvtdq2pd xmm6,xmm3 ; xexp
; NaN or inf ; NaN or inf
vpand xmm5,xmm0,QWORD PTR __real_inf vpand xmm5,xmm0,XMMWORD PTR __real_inf
vcomisd xmm5,QWORD PTR __real_inf vcomisd xmm5,QWORD PTR __real_inf
je Llog_fma3_x_is_inf_or_nan je Llog_fma3_x_is_inf_or_nan
@ -404,7 +404,7 @@ Llog_fma3:
vcomisd xmm0,xmm5 vcomisd xmm0,xmm5
jbe Llog_fma3_x_is_zero_or_neg jbe Llog_fma3_x_is_zero_or_neg
vpand xmm2,xmm0,QWORD PTR __real_mant vpand xmm2,xmm0,XMMWORD PTR __real_mant
vsubsd xmm4,xmm0,QWORD PTR __real_one vsubsd xmm4,xmm0,QWORD PTR __real_one
vcomisd xmm6,QWORD PTR __mask_1023_f vcomisd xmm6,QWORD PTR __mask_1023_f
@ -412,21 +412,21 @@ Llog_fma3:
Llog_fma3_continue_common: Llog_fma3_continue_common:
; compute index into the log tables ; compute index into the log tables
vpand xmm1,xmm0,QWORD PTR __mask_mant_all8 vpand xmm1,xmm0,XMMWORD PTR __mask_mant_all8
vpand xmm3,xmm0,QWORD PTR __mask_mant9 vpand xmm3,xmm0,XMMWORD PTR __mask_mant9
vpsllq xmm3,xmm3,1 vpsllq xmm3,xmm3,1
vpaddq xmm1,xmm3,xmm1 vpaddq xmm1,xmm3,xmm1
vmovq rax,xmm1 vmovq rax,xmm1
; near one codepath ; near one codepath
vpand xmm4,xmm4,QWORD PTR __real_notsign vpand xmm4,xmm4,XMMWORD PTR __real_notsign
vcomisd xmm4,QWORD PTR __real_threshold vcomisd xmm4,QWORD PTR __real_threshold
jb Llog_fma3_near_one jb Llog_fma3_near_one
; F,Y ; F,Y
shr rax,44 shr rax,44
vpor xmm2,xmm2,QWORD PTR __real_half vpor xmm2,xmm2,XMMWORD PTR __real_half
vpor xmm1,xmm1,QWORD PTR __real_half vpor xmm1,xmm1,XMMWORD PTR __real_half
lea r9,QWORD PTR __log_F_inv_qword lea r9,QWORD PTR __log_F_inv_qword
; f = F - Y,r = f * inv ; f = F - Y,r = f * inv
@ -501,10 +501,10 @@ Llog_fma3_near_one:
Llog_fma3_denormal_adjust: Llog_fma3_denormal_adjust:
vpor xmm2,xmm2,QWORD PTR __real_one vpor xmm2,xmm2,XMMWORD PTR __real_one
vsubsd xmm2,xmm2,QWORD PTR __real_one vsubsd xmm2,xmm2,QWORD PTR __real_one
vpsrlq xmm5,xmm2,52 vpsrlq xmm5,xmm2,52
vpand xmm2,xmm2,QWORD PTR __real_mant vpand xmm2,xmm2,XMMWORD PTR __real_mant
vmovapd xmm0,xmm2 vmovapd xmm0,xmm2
vpsubd xmm5,xmm5,XMMWORD PTR __mask_2045 vpsubd xmm5,xmm5,XMMWORD PTR __mask_2045
vcvtdq2pd xmm6,xmm5 vcvtdq2pd xmm6,xmm5

20
sdk/lib/crt/math/libm_sse2/log10.asm
View file

@ -382,7 +382,7 @@ Llog10_fma3:
xor rax,rax xor rax,rax
vpsrlq xmm3,xmm0,52 vpsrlq xmm3,xmm0,52
vmovq rax,xmm0 vmovq rax,xmm0
vpsubq xmm3,xmm3,QWORD PTR __int_1023 vpsubq xmm3,xmm3,XMMWORD PTR __int_1023
vcvtdq2pd xmm6,xmm3 ; xmm6 <-- (double)xexp vcvtdq2pd xmm6,xmm3 ; xmm6 <-- (double)xexp
; NaN or Inf? ; NaN or Inf?
@ -404,21 +404,21 @@ Llog10_fma3:
Llog10_fma3_continue_common: Llog10_fma3_continue_common:
; compute index into the log tables ; compute index into the log tables
vpand xmm1,xmm0,DWORD PTR __mask_mant_top8 vpand xmm1,xmm0,XMMWORD PTR __mask_mant_top8
vpand xmm3,xmm0,DWORD PTR __mask_mant9 vpand xmm3,xmm0,XMMWORD PTR __mask_mant9
vpsllq xmm3,xmm3,1 vpsllq xmm3,xmm3,1
vpaddq xmm1,xmm3,xmm1 vpaddq xmm1,xmm3,xmm1
vmovq rax,xmm1 vmovq rax,xmm1
; near one codepath ; near one codepath
vpand xmm4,xmm4,DWORD PTR __real_notsign vpand xmm4,xmm4,XMMWORD PTR __real_notsign
vcomisd xmm4,QWORD PTR __real_threshold vcomisd xmm4,QWORD PTR __real_threshold
jb Llog10_fma3_near_one jb Llog10_fma3_near_one
; F,Y ; F,Y
shr rax,44 shr rax,44
vpor xmm2,xmm2,DWORD PTR __real_half vpor xmm2,xmm2,XMMWORD PTR __real_half
vpor xmm1,xmm1,DWORD PTR __real_half vpor xmm1,xmm1,XMMWORD PTR __real_half
lea r9,DWORD PTR __log_F_inv_qword lea r9,DWORD PTR __log_F_inv_qword
; f = F - Y,r = f * inv ; f = F - Y,r = f * inv
@ -486,7 +486,7 @@ Llog10_fma3_near_one:
vmulsd xmm5,xmm5,xmm2 vmulsd xmm5,xmm5,xmm2
vaddsd xmm4,xmm4,xmm5 vaddsd xmm4,xmm4,xmm5
vsubsd xmm4,xmm4,xmm6 vsubsd xmm4,xmm4,xmm6
vpand xmm3,xmm0,QWORD PTR __mask_lower vpand xmm3,xmm0,XMMWORD PTR __mask_lower
vsubsd xmm0,xmm0,xmm3 vsubsd xmm0,xmm0,xmm3
vaddsd xmm4,xmm4,xmm0 vaddsd xmm4,xmm4,xmm0
@ -505,12 +505,12 @@ Llog10_fma3_near_one:
Llog10_fma3_denormal_adjust: Llog10_fma3_denormal_adjust:
vpor xmm2,xmm2,QWORD PTR __real_one vpor xmm2,xmm2,XMMWORD PTR __real_one
vsubsd xmm2,xmm2,QWORD PTR __real_one vsubsd xmm2,xmm2,QWORD PTR __real_one
vpsrlq xmm5,xmm2,52 vpsrlq xmm5,xmm2,52
vpand xmm2,xmm2,QWORD PTR __mask_mant vpand xmm2,xmm2,XMMWORD PTR __mask_mant
vmovapd xmm0,xmm2 vmovapd xmm0,xmm2
vpsubd xmm5,xmm5,DWORD PTR __mask_2045 vpsubd xmm5,xmm5,XMMWORD PTR __mask_2045
vcvtdq2pd xmm6,xmm5 vcvtdq2pd xmm6,xmm5
jmp Llog10_fma3_continue_common jmp Llog10_fma3_continue_common

4
sdk/lib/crt/math/libm_sse2/pow.asm
View file

@ -1219,7 +1219,7 @@ Lpow_sse2_near_one:
movsd xmm3, QWORD PTR [r9+r8*8] movsd xmm3, QWORD PTR [r9+r8*8]
addsd xmm3, QWORD PTR [rdx+r8*8] addsd xmm3, QWORD PTR [rdx+r8*8]
mulsd xmm4, xmm3 mulsd xmm4, xmm3
andpd xmm4, QWORD PTR __real_fffffffff8000000 andpd xmm4, XMMWORD PTR __real_fffffffff8000000
movsd xmm5, xmm4 ; r1 movsd xmm5, xmm4 ; r1
mulsd xmm4, xmm0 mulsd xmm4, xmm0
subsd xmm1, xmm4 subsd xmm1, xmm4
@ -2028,7 +2028,7 @@ Lpow_fma3_near_one:
vmovsd xmm3, QWORD PTR [r9+r8*8] vmovsd xmm3, QWORD PTR [r9+r8*8]
vaddsd xmm3, xmm3, QWORD PTR [rdx+r8*8] vaddsd xmm3, xmm3, QWORD PTR [rdx+r8*8]
vmulsd xmm4, xmm4, xmm3 ; xmm4 <-- r = f*inv vmulsd xmm4, xmm4, xmm3 ; xmm4 <-- r = f*inv
vandpd xmm4, xmm4, QWORD PTR __real_fffffffff8000000 ; r1 vandpd xmm4, xmm4, XMMWORD PTR __real_fffffffff8000000 ; r1
vmovapd xmm5, xmm4 ; xmm5 <-- copy of r1 vmovapd xmm5, xmm4 ; xmm5 <-- copy of r1
; mulsd xmm4, xmm0 ; xmm4 <-- F*r1 ; mulsd xmm4, xmm0 ; xmm4 <-- F*r1
; subsd xmm1, xmm4 ; xmm1 <-- f - F*r1 ; subsd xmm1, xmm4 ; xmm1 <-- f - F*r1