mirror of
https://sourceware.org/git/glibc.git
synced 2025-01-14 21:10:19 +00:00
x86_64: Fix svml_d_atan24_core_avx2.S code formatting
This commit contains following formatting changes 1. Instructions proceeded by a tab. 2. Instruction less than 8 characters in length have a tab between it and the first operand. 3. Instruction greater than 7 characters in length have a space between it and the first operand. 4. Tabs after `#define`d names and their value. 5. 8 space at the beginning of line replaced by tab. 6. Indent comments with code. 7. Remove redundent .text section. 8. 1 space between line content and line comment. 9. Space after all commas. Reviewed-by: Noah Goldstein <goldstein.w.n@gmail.com>
This commit is contained in:
Sunil K Pandey
parent
1447e84caf
commit
4f75454074
@ -60,8 +60,7 @@
|
||||
|
||||
#include <sysdep.h>
|
||||
|
||||
.text
|
||||
.section .text.avx2,"ax",@progbits
|
||||
.section .text.avx2, "ax", @progbits
|
||||
ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
pushq %rbp
|
||||
cfi_def_cfa_offset(16)
|
||||
@ -72,7 +71,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
subq $128, %rsp
|
||||
xorl %edx, %edx
|
||||
|
||||
/*
|
||||
/*
|
||||
* #define NO_VECTOR_ZERO_ATAN2_ARGS
|
||||
* Declarations
|
||||
* Variables
|
||||
@ -85,7 +84,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
*/
|
||||
vmovupd dABS_MASK+__svml_datan2_data_internal(%rip), %ymm5
|
||||
|
||||
/* Argument signs */
|
||||
/* Argument signs */
|
||||
vmovupd dSIGN_MASK+__svml_datan2_data_internal(%rip), %ymm4
|
||||
vmovups iCHK_WORK_SUB+__svml_datan2_data_internal(%rip), %xmm13
|
||||
vmovupd %ymm0, (%rsp)
|
||||
@ -94,7 +93,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vandpd %ymm5, %ymm0, %ymm1
|
||||
vcmpnlt_uqpd %ymm2, %ymm1, %ymm15
|
||||
|
||||
/*
|
||||
/*
|
||||
* 1) If y<x then a= y, b=x, PIO2=0
|
||||
* 2) If y>x then a=-x, b=y, PIO2=Pi/2
|
||||
*/
|
||||
@ -108,7 +107,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vandpd %ymm4, %ymm8, %ymm5
|
||||
vandpd dPIO2+__svml_datan2_data_internal(%rip), %ymm15, %ymm4
|
||||
|
||||
/* Check if y and x are on main path. */
|
||||
/* Check if y and x are on main path. */
|
||||
vextractf128 $1, %ymm2, %xmm9
|
||||
vextractf128 $1, %ymm1, %xmm10
|
||||
vshufps $221, %xmm9, %xmm2, %xmm11
|
||||
@ -122,7 +121,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vpor %xmm6, %xmm15, %xmm11
|
||||
vpor %xmm3, %xmm10, %xmm12
|
||||
|
||||
/* Polynomial. */
|
||||
/* Polynomial. */
|
||||
vmulpd %ymm14, %ymm14, %ymm10
|
||||
vpor %xmm12, %xmm11, %xmm3
|
||||
vmovupd dA18+__svml_datan2_data_internal(%rip), %ymm9
|
||||
@ -130,7 +129,7 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vmovupd dA16+__svml_datan2_data_internal(%rip), %ymm15
|
||||
vmulpd %ymm10, %ymm10, %ymm11
|
||||
|
||||
/* if x<0, dPI = Pi, else dPI =0 */
|
||||
/* if x<0, dPI = Pi, else dPI =0 */
|
||||
vcmple_oqpd dZERO+__svml_datan2_data_internal(%rip), %ymm8, %ymm13
|
||||
vmovmskps %xmm3, %eax
|
||||
vmulpd %ymm11, %ymm11, %ymm0
|
||||
@ -152,13 +151,13 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vfmadd213pd dA01+__svml_datan2_data_internal(%rip), %ymm0, %ymm12
|
||||
vfmadd213pd dA03+__svml_datan2_data_internal(%rip), %ymm0, %ymm13
|
||||
|
||||
/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */
|
||||
/* A00=1.0, account for it later VQFMA(D, dP4, dP4, dR8, dA00); */
|
||||
vmulpd %ymm15, %ymm0, %ymm0
|
||||
vfmadd213pd %ymm9, %ymm10, %ymm13
|
||||
vfmadd213pd %ymm0, %ymm10, %ymm12
|
||||
vfmadd213pd %ymm12, %ymm11, %ymm13
|
||||
|
||||
/*
|
||||
/*
|
||||
* Reconstruction.
|
||||
* dP=(R+R*dP) + dPIO2
|
||||
*/
|
||||
@ -168,29 +167,29 @@ ENTRY(_ZGVdN4vv_atan2_avx2)
|
||||
vaddpd %ymm0, %ymm6, %ymm9
|
||||
vorpd %ymm7, %ymm9, %ymm0
|
||||
|
||||
/* Special branch for fast (vector) processing of zero arguments */
|
||||
/* Special branch for fast (vector) processing of zero arguments */
|
||||
testl %eax, %eax
|
||||
|
||||
/* Go to auxilary branch */
|
||||
/* Go to auxilary branch */
|
||||
jne L(AUX_BRANCH)
|
||||
# LOE rbx r12 r13 r14 r15 edx xmm3 ymm0 ymm1 ymm2 ymm4 ymm5 ymm6 ymm7 ymm8
|
||||
|
||||
/* Return from auxilary branch
|
||||
/* Return from auxilary branch
|
||||
* for out of main path inputs
|
||||
*/
|
||||
|
||||
L(AUX_BRANCH_RETURN):
|
||||
/*
|
||||
/*
|
||||
* Special branch for fast (vector) processing of zero arguments
|
||||
* The end of implementation
|
||||
*/
|
||||
testl %edx, %edx
|
||||
|
||||
/* Go to special inputs processing branch */
|
||||
/* Go to special inputs processing branch */
|
||||
jne L(SPECIAL_VALUES_BRANCH)
|
||||
# LOE rbx r12 r13 r14 r15 edx ymm0 ymm8
|
||||
|
||||
/* Restore registers
|
||||
/* Restore registers
|
||||
* and exit the function
|
||||
*/
|
||||
|
||||
@ -203,7 +202,7 @@ L(EXIT):
|
||||
cfi_def_cfa(6, 16)
|
||||
cfi_offset(6, -16)
|
||||
|
||||
/* Branch to process
|
||||
/* Branch to process
|
||||
* special inputs
|
||||
*/
|
||||
|
||||
@ -231,18 +230,18 @@ L(SPECIAL_VALUES_BRANCH):
|
||||
.cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22
|
||||
# LOE rbx r15 r12d r13d
|
||||
|
||||
/* Range mask
|
||||
/* Range mask
|
||||
* bits check
|
||||
*/
|
||||
|
||||
L(RANGEMASK_CHECK):
|
||||
btl %r12d, %r13d
|
||||
|
||||
/* Call scalar math function */
|
||||
/* Call scalar math function */
|
||||
jc L(SCALAR_MATH_CALL)
|
||||
# LOE rbx r15 r12d r13d
|
||||
|
||||
/* Special inputs
|
||||
/* Special inputs
|
||||
* processing loop
|
||||
*/
|
||||
|
||||
@ -250,7 +249,7 @@ L(SPECIAL_VALUES_LOOP):
|
||||
incl %r12d
|
||||
cmpl $4, %r12d
|
||||
|
||||
/* Check bits in range mask */
|
||||
/* Check bits in range mask */
|
||||
jl L(RANGEMASK_CHECK)
|
||||
# LOE rbx r15 r12d r13d
|
||||
|
||||
@ -262,7 +261,7 @@ L(SPECIAL_VALUES_LOOP):
|
||||
cfi_restore(14)
|
||||
vmovupd 96(%rsp), %ymm0
|
||||
|
||||
/* Go to exit */
|
||||
/* Go to exit */
|
||||
jmp L(EXIT)
|
||||
/* DW_CFA_expression: r12 (r12) (DW_OP_lit8; DW_OP_minus; DW_OP_const4s: -32; DW_OP_and; DW_OP_const4s: -112; DW_OP_plus) */
|
||||
.cfi_escape 0x10, 0x0c, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x90, 0xff, 0xff, 0xff, 0x22
|
||||
@ -272,37 +271,37 @@ L(SPECIAL_VALUES_LOOP):
|
||||
.cfi_escape 0x10, 0x0e, 0x0e, 0x38, 0x1c, 0x0d, 0xe0, 0xff, 0xff, 0xff, 0x1a, 0x0d, 0x80, 0xff, 0xff, 0xff, 0x22
|
||||
# LOE rbx r12 r13 r14 r15 ymm0
|
||||
|
||||
/* Scalar math fucntion call
|
||||
/* Scalar math fucntion call
|
||||
* to process special input
|
||||
*/
|
||||
|
||||
L(SCALAR_MATH_CALL):
|
||||
movl %r12d, %r14d
|
||||
movsd 32(%rsp,%r14,8), %xmm0
|
||||
movsd 64(%rsp,%r14,8), %xmm1
|
||||
movsd 32(%rsp, %r14, 8), %xmm0
|
||||
movsd 64(%rsp, %r14, 8), %xmm1
|
||||
call atan2@PLT
|
||||
# LOE rbx r14 r15 r12d r13d xmm0
|
||||
|
||||
movsd %xmm0, 96(%rsp,%r14,8)
|
||||
movsd %xmm0, 96(%rsp, %r14, 8)
|
||||
|
||||
/* Process special inputs in loop */
|
||||
/* Process special inputs in loop */
|
||||
jmp L(SPECIAL_VALUES_LOOP)
|
||||
cfi_restore(12)
|
||||
cfi_restore(13)
|
||||
cfi_restore(14)
|
||||
# LOE rbx r15 r12d r13d
|
||||
|
||||
/* Auxilary branch
|
||||
/* Auxilary branch
|
||||
* for out of main path inputs
|
||||
*/
|
||||
|
||||
L(AUX_BRANCH):
|
||||
vmovupd (%rsp), %ymm11
|
||||
|
||||
/* Check if at least on of Y or Y is zero: iAXAYZERO */
|
||||
/* Check if at least on of Y or Y is zero: iAXAYZERO */
|
||||
vmovupd dZERO+__svml_datan2_data_internal(%rip), %ymm10
|
||||
|
||||
/* Check if both X & Y are not NaNs: iXYnotNAN */
|
||||
/* Check if both X & Y are not NaNs: iXYnotNAN */
|
||||
vcmpordpd %ymm8, %ymm8, %ymm12
|
||||
vcmpordpd %ymm11, %ymm11, %ymm13
|
||||
vcmpeqpd %ymm10, %ymm2, %ymm2
|
||||
@ -314,26 +313,26 @@ L(AUX_BRANCH):
|
||||
vshufps $221, %xmm15, %xmm14, %xmm9
|
||||
vshufps $221, %xmm11, %xmm2, %xmm12
|
||||
|
||||
/*
|
||||
/*
|
||||
* Path for zero arguments (at least one of both)
|
||||
* Check if both args are zeros (den. is zero)
|
||||
*/
|
||||
vcmpeqpd 32(%rsp), %ymm10, %ymm2
|
||||
|
||||
/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */
|
||||
/* Check if at least on of Y or Y is zero and not NaN: iAXAYZEROnotNAN */
|
||||
vpand %xmm9, %xmm12, %xmm1
|
||||
|
||||
/* Exclude from previous callout mask zero (and not NaN) arguments */
|
||||
/* Exclude from previous callout mask zero (and not NaN) arguments */
|
||||
vpandn %xmm3, %xmm1, %xmm3
|
||||
|
||||
/* Go to callout */
|
||||
/* Go to callout */
|
||||
vmovmskps %xmm3, %edx
|
||||
|
||||
/* Set sPIO2 to zero if den. is zero */
|
||||
/* Set sPIO2 to zero if den. is zero */
|
||||
vblendvpd %ymm2, %ymm10, %ymm4, %ymm4
|
||||
vorpd %ymm5, %ymm4, %ymm5
|
||||
|
||||
/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */
|
||||
/* Res = sign(Y)*(X<0)?(PIO2+PI):PIO2 */
|
||||
vextractf128 $1, %ymm10, %xmm2
|
||||
vextractf128 $1, %ymm8, %xmm3
|
||||
vshufps $221, %xmm2, %xmm10, %xmm4
|
||||
@ -346,13 +345,13 @@ L(AUX_BRANCH):
|
||||
vaddpd %ymm6, %ymm5, %ymm2
|
||||
vorpd %ymm7, %ymm2, %ymm2
|
||||
|
||||
/* Merge results from main and spec path */
|
||||
/* Merge results from main and spec path */
|
||||
vpshufd $80, %xmm1, %xmm7
|
||||
vpshufd $250, %xmm1, %xmm1
|
||||
vinsertf128 $1, %xmm1, %ymm7, %ymm3
|
||||
vblendvpd %ymm3, %ymm2, %ymm0, %ymm0
|
||||
|
||||
/* Return to main vector processing path */
|
||||
/* Return to main vector processing path */
|
||||
jmp L(AUX_BRANCH_RETURN)
|
||||
# LOE rbx r12 r13 r14 r15 edx ymm0 ymm8
|
||||
END(_ZGVdN4vv_atan2_avx2)
|
||||
@ -393,9 +392,9 @@ typedef struct {
|
||||
} __svml_datan2_data_internal;
|
||||
#endif
|
||||
__svml_datan2_data_internal:
|
||||
.quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 //dPI
|
||||
.quad 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18, 0x400921FB54442D18 // dPI
|
||||
.align 32
|
||||
.quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 //dPIO2
|
||||
.quad 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18, 0x3FF921FB54442D18 // dPIO2
|
||||
.align 32
|
||||
.quad 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3, 0xBEF4FDB537ABC7A3 // dA19
|
||||
.align 32
|
||||
@ -437,15 +436,15 @@ __svml_datan2_data_internal:
|
||||
.align 32
|
||||
.quad 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000, 0x3FF0000000000000 // dA00
|
||||
.align 32
|
||||
.quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 //dSIGN_MASK
|
||||
.quad 0x8000000000000000, 0x8000000000000000, 0x8000000000000000, 0x8000000000000000 // dSIGN_MASK
|
||||
.align 32
|
||||
.long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 //iCHK_WORK_SUB
|
||||
.long 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000, 0x80300000 // iCHK_WORK_SUB
|
||||
.align 32
|
||||
.long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 //iCHK_WORK_CMP
|
||||
.long 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000, 0xfdd00000 // iCHK_WORK_CMP
|
||||
.align 32
|
||||
.quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff //dABS_MASK
|
||||
.quad 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff, 0x7fffffffffffffff // dABS_MASK
|
||||
.align 32
|
||||
.quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 //dZERO
|
||||
.quad 0x0000000000000000, 0x0000000000000000, 0x0000000000000000, 0x0000000000000000 // dZERO
|
||||
.align 32
|
||||
.type __svml_datan2_data_internal,@object
|
||||
.size __svml_datan2_data_internal,.-__svml_datan2_data_internal
|
||||
.type __svml_datan2_data_internal, @object
|
||||
.size __svml_datan2_data_internal, .-__svml_datan2_data_internal
|
||||
|
||||
Loading…
Reference in New Issue
Block a user