Merge pull request #167 from harshavardhana/pr_out_merge_with_intel_isal_changes_from_github_com_minio_io_isal

Makefile

@@ -14,9 +14,6 @@ build-erasure:
 	@$(MAKE) $(MAKE_OPTIONS) -C pkg/erasure/isal lib
 	@godep go test -race -coverprofile=cover.out github.com/minio-io/minio/pkg/erasure
 
-build-signify:
-	@$(MAKE) $(MAKE_OPTIONS) -C pkg/signify
-
 build-cpu:
 	@godep go test -race -coverprofile=cover.out github.com/minio-io/minio/pkg/cpu
 
@@ -53,7 +50,7 @@ build-storage-append:
 build-storage-encoded:
 	@godep go test -race -coverprofile=cover.out github.com/minio-io/minio/pkg/storage/encodedstorage
 
-cover: build-erasure build-signify build-split build-crc32c build-cpu build-scsi build-storage build-md5 build-sha1 build-sha256 build-sha512
+cover: build-erasure build-split build-crc32c build-cpu build-scsi build-storage build-md5 build-sha1 build-sha256 build-sha512
 	@godep go test -race -coverprofile=cover.out github.com/minio-io/minio/pkg/gateway
 
 install: build-erasure

pkg/checksum/crc32c/crc32c_amd64.S

@@ -168,7 +168,7 @@ continue_block:
 
 	## branch into array
 	lea	jump_table(%rip), bufp
-	movzxw  (bufp, %rax, 2), len
+	movzwq  (bufp, %rax, 2), len
 	offset=crc_array-jump_table
 	lea     offset(bufp, len, 1), bufp
 	jmp     *bufp
@@ -194,18 +194,22 @@ full_block:
 crc_array:
 	i=128
 .rept 128-1
+#if !defined(__clang__)
 .altmacro
 LABEL crc_ %i
 .noaltmacro
+#endif
 	crc32q   -i*8(block_0), crc_init
 	crc32q   -i*8(block_1), crc1
 	crc32q   -i*8(block_2), crc2
 	i=(i-1)
 .endr
 
+#if !defined(__clang__)
 .altmacro
 LABEL crc_ %i
 .noaltmacro
+#endif
 	crc32q   -i*8(block_0), crc_init
 	crc32q   -i*8(block_1), crc1
 # SKIP  crc32  -i*8(block_2), crc2 ; Don't do this one yet

pkg/cpu/cpu_amd64.S

@@ -14,9 +14,18 @@
  * limitations under the License.
  */
 
+#ifdef __APPLE__
+#define HAS_SSE _has_sse41
+#define HAS_AVX _has_avx
+#define HAS_AVX2 _has_avx2
+#else
+#define HAS_SSE has_sse41
+#define HAS_AVX has_avx
+#define HAS_AVX2 has_avx2
+#endif
+
 	.file	"cpufeatures.c"
 	.text
-	.type	cpuid, @function
 cpuid:
 .LFB2:
 	.cfi_startproc
@@ -56,10 +65,8 @@ cpuid:
 	ret
 	.cfi_endproc
 .LFE2:
-	.size	cpuid, .-cpuid
+	.globl	HAS_SSE
-	.globl	has_sse41
+HAS_SSE:
-	.type	has_sse41, @function
-has_sse41:
 .LFB3:
 	.cfi_startproc
 	pushq	%rbp
@@ -82,10 +89,8 @@ has_sse41:
 	ret
 	.cfi_endproc
 .LFE3:
-	.size	has_sse41, .-has_sse41
+	.globl	HAS_AVX
-	.globl	has_avx
+HAS_AVX:
-	.type	has_avx, @function
-has_avx:
 .LFB4:
 	.cfi_startproc
 	pushq	%rbp
@@ -108,10 +113,8 @@ has_avx:
 	ret
 	.cfi_endproc
 .LFE4:
-	.size	has_avx, .-has_avx
+	.globl	HAS_AVX2
-	.globl	has_avx2
+HAS_AVX2:
-	.type	has_avx2, @function
-has_avx2:
 .LFB5:
 	.cfi_startproc
 	pushq	%rbp
@@ -133,7 +136,3 @@ has_avx2:
 	.cfi_def_cfa 7, 8
 	ret
 	.cfi_endproc
-.LFE5:
-	.size	has_avx2, .-has_avx2
-	.ident	"GCC: (Ubuntu 4.8.2-19ubuntu1) 4.8.2"
-	.section	.note.GNU-stack,"",@progbits

pkg/erasure/Makefile

@@ -1,12 +1,17 @@
 all: build test
 .PHONY: all
 
+SYSTEM_NAME := $(shell uname -s)
+
 test:
 	@godep go test -race -coverprofile=cover.out
 
 isal/isal-l.a:
+ifeq ($(SYSTEM_NAME), Darwin)
+	@$(MAKE) -C isal arch=osx lib
+else
 	@$(MAKE) -C isal lib
-
+endif
 build: isal/isal-l.a
 	@godep go build
 

pkg/erasure/decode.c

@@ -60,6 +60,8 @@ int32_t minio_get_source_target (int errs, int k, int m,
 
         *source = tmp_source;
         *target = tmp_target;
+
+	return 0;
 }
 
 /*

pkg/erasure/isal/.gitignore

@@ -1,3 +1,22 @@
 *.o
 *.a
 *.so
+*~
+*.dSYM
+erasure-code-base-test
+erasure-code-sse-test
+erasure-code-test
+gf-2vect-dot-prod-sse-test
+gf-3vect-dot-prod-sse-test
+gf-4vect-dot-prod-sse-test
+gf-5vect-dot-prod-sse-test
+gf-6vect-dot-prod-sse-test
+gf-inverse-test
+gf-vect-dot-prod-avx-test
+gf-vect-dot-prod-base-test
+gf-vect-dot-prod-sse-test
+gf-vect-dot-prod-test
+gf-vect-mul-avx-test
+gf-vect-mul-base-test
+gf-vect-mul-sse-test
+gf-vect-mul-test

pkg/erasure/isal/Makefile

@@ -30,7 +30,7 @@
 
 units = src
 
-default: slib
+default: lib
 
 include $(foreach unit,$(units), $(unit)/Makefile)
 

pkg/erasure/isal/include/erasure/types.h

@@ -41,7 +41,7 @@
 extern "C" {
 #endif
 
-#ifndef __unix__
+#if !defined(__unix__) && !defined(__APPLE__)
 #ifdef __MINGW32__
 # include <_mingw.h>
 #endif
@@ -59,7 +59,7 @@ typedef unsigned char      UINT8;
 #endif
 
 
-#ifdef __unix__
+#if defined(__unix__) || defined(__APPLE__)
 # define DECLARE_ALIGNED(decl, alignval) decl __attribute__((aligned(alignval)))
 # define __forceinline static inline
 #else

pkg/erasure/isal/make.inc

@@ -63,6 +63,11 @@ ASFLAGS_win64 = -f win64
 CFLAGS_icl    = -Qstd=c99
 ARFLAGS_win64 = -out:$@
 
+# arch=osx build options
+ASFLAGS_osx = -f macho64
+ARFLAGS_osx = -r $@
+STRIP_gcc = strip -d $@
+
 # arch=mingw build options
 ASFLAGS_mingw = -f win64
 ARFLAGS_mingw = cr $@
@@ -101,9 +106,13 @@ ifeq ($(arch),win64)
   lib_name := $(basename $(lib_name)).lib
 endif
 lsrcwin64 = $(lsrc)
+lsrcosx = $(lsrc)
 unit_testswin64 = $(unit_tests)
+unit_testsosx = $(unit_tests)
 exampleswin64 = $(examples)
+examplesosx = $(examples)
 perf_testswin64 = $(perf_tests)
+perf_testsosx = $(perf_tests)
 
 # Build and run unit tests, performance tests, etc.
 all_tests = $(sort $(perf_tests$(arch)) $(unit_tests$(arch)) $(examples$(arch)) $(other_tests))
@@ -199,7 +208,6 @@ perf_report:
 	@echo Summary:
 	-grep runtime $(rpt_name)
 
-
 clean:
 	@echo Cleaning up
 	@$(RM) -r $(O)/*.o *.a $(all_tests) $(lib_name) $(so_lib_name)

pkg/erasure/isal/src/ec-multibinary.asm

@@ -33,6 +33,42 @@
 %define WRT_OPT
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+%define EC_ENCODE_DATA_SSE _ec_encode_data_sse
+%define EC_ENCODE_DATA_AVX _ec_encode_data_avx
+%define EC_ENCODE_DATA_AVX2 _ec_encode_data_avx2
+%define GF_VECT_MUL_SSE _gf_vect_mul_sse
+%define GF_VECT_MUL_AVX _gf_vect_mul_avx
+%define GF_VECT_DOT_PROD_SSE _gf_vect_dot_prod_sse
+%define GF_VECT_DOT_PROD_AVX _gf_vect_dot_prod_avx
+%define GF_VECT_DOT_PROD_AVX2 _gf_vect_dot_prod_avx2
+%define GF_VECT_MUL_BASE _gf_vect_mul_base
+%define EC_ENCODE_DATA_BASE _ec_encode_data_base
+%define GF_VECT_DOT_PROD_BASE _gf_vect_dot_prod_base
+
+%define EC_ENCODE_DATA _ec_encode_data
+%define GF_VECT_MUL _gf_vect_mul
+%define GF_VECT_DOT_PROD _gf_vect_dot_prod
+
+%else
+%define EC_ENCODE_DATA_SSE ec_encode_data_sse
+%define EC_ENCODE_DATA_AVX ec_encode_data_avx
+%define EC_ENCODE_DATA_AVX2 ec_encode_data_avx2
+%define GF_VECT_MUL_SSE gf_vect_mul_sse
+%define GF_VECT_MUL_AVX gf_vect_mul_avx
+%define GF_VECT_DOT_PROD_SSE gf_vect_dot_prod_sse
+%define GF_VECT_DOT_PROD_AVX gf_vect_dot_prod_avx
+%define GF_VECT_DOT_PROD_AVX2 gf_vect_dot_prod_avx2
+%define GF_VECT_MUL_BASE gf_vect_mul_base
+%define EC_ENCODE_DATA_BASE ec_encode_data_base
+%define GF_VECT_DOT_PROD_BASE gf_vect_dot_prod_base
+
+%define EC_ENCODE_DATA ec_encode_data
+%define GF_VECT_MUL gf_vect_mul
+%define GF_VECT_DOT_PROD gf_vect_dot_prod
+
+%endif
+
 %ifidn __OUTPUT_FORMAT__, elf32
 
 [bits 32]
@@ -51,19 +87,19 @@ default rel
 %define wrd_sz  	qword
 %define arg1		rsi
 
-extern ec_encode_data_sse
+extern EC_ENCODE_DATA_SSE
-extern ec_encode_data_avx
+extern EC_ENCODE_DATA_AVX
-extern ec_encode_data_avx2
+extern EC_ENCODE_DATA_AVX2
-extern gf_vect_mul_sse
+extern GF_VECT_MUL_SSE
-extern gf_vect_mul_avx
+extern GF_VECT_MUL_AVX
-extern gf_vect_dot_prod_sse
+extern GF_VECT_DOT_PROD_SSE
-extern gf_vect_dot_prod_avx
+extern GF_VECT_DOT_PROD_AVX
-extern gf_vect_dot_prod_avx2
+extern GF_VECT_DOT_PROD_AVX2
 %endif
 
-extern gf_vect_mul_base
+extern GF_VECT_MUL_BASE
-extern ec_encode_data_base
+extern EC_ENCODE_DATA_BASE
-extern gf_vect_dot_prod_base
+extern GF_VECT_DOT_PROD_BASE
 
 section .data
 ;;; *_mbinit are initial values for *_dispatched; is updated on first call.
@@ -82,33 +118,33 @@ section .text
 ;;;;
 ; ec_encode_data multibinary function
 ;;;;
-global ec_encode_data:function
+global EC_ENCODE_DATA:function
 ec_encode_data_mbinit:
 	call	ec_encode_data_dispatch_init
 
-ec_encode_data:
+EC_ENCODE_DATA:
 	jmp	wrd_sz [ec_encode_data_dispatched]
 
 ec_encode_data_dispatch_init:
 	push    arg1
 %ifidn __OUTPUT_FORMAT__, elf32		;; 32-bit check
-	lea     arg1, [ec_encode_data_base]
+	lea     arg1, [EC_ENCODE_DATA_BASE]
 %else
 	push    rax
 	push    rbx
 	push    rcx
 	push    rdx
-	lea     arg1, [ec_encode_data_base WRT_OPT] ; Default
+	lea     arg1, [EC_ENCODE_DATA_BASE WRT_OPT] ; Default
 
 	mov     eax, 1
 	cpuid
-	lea     rbx, [ec_encode_data_sse WRT_OPT]
+	lea     rbx, [EC_ENCODE_DATA_BASE WRT_OPT]
 	test    ecx, FLAG_CPUID1_ECX_SSE4_1
 	cmovne  arg1, rbx
 
 	and	ecx, (FLAG_CPUID1_ECX_AVX | FLAG_CPUID1_ECX_OSXSAVE)
 	cmp	ecx, (FLAG_CPUID1_ECX_AVX | FLAG_CPUID1_ECX_OSXSAVE)
-	lea	rbx, [ec_encode_data_avx WRT_OPT]
+	lea	rbx, [EC_ENCODE_DATA_AVX WRT_OPT]
 
 	jne	_done_ec_encode_data_init
 	mov	rsi, rbx
@@ -118,7 +154,7 @@ ec_encode_data_dispatch_init:
 	mov	eax, 7
 	cpuid
 	test	ebx, FLAG_CPUID1_EBX_AVX2
-	lea     rbx, [ec_encode_data_avx2 WRT_OPT]
+	lea     rbx, [EC_ENCODE_DATA_AVX2 WRT_OPT]
 	cmovne	rsi, rbx
 
 	;; Does it have xmm and ymm support
@@ -127,7 +163,7 @@ ec_encode_data_dispatch_init:
 	and	eax, FLAG_XGETBV_EAX_XMM_YMM
 	cmp	eax, FLAG_XGETBV_EAX_XMM_YMM
 	je	_done_ec_encode_data_init
-	lea     rsi, [ec_encode_data_sse WRT_OPT]
+	lea     rsi, [EC_ENCODE_DATA_SSE WRT_OPT]
 
 _done_ec_encode_data_init:
 	pop     rdx
@@ -142,28 +178,28 @@ _done_ec_encode_data_init:
 ;;;;
 ; gf_vect_mul multibinary function
 ;;;;
-global gf_vect_mul:function
+global GF_VECT_MUL:function
 gf_vect_mul_mbinit:
 	call    gf_vect_mul_dispatch_init
 
-gf_vect_mul:
+GF_VECT_MUL:
 	jmp	wrd_sz [gf_vect_mul_dispatched]
 
 gf_vect_mul_dispatch_init:
 	push    arg1
 %ifidn __OUTPUT_FORMAT__, elf32		;; 32-bit check
-	lea     arg1, [gf_vect_mul_base]
+	lea     arg1, [GF_VECT_MUL_BASE]
 %else
 	push    rax
 	push    rbx
 	push    rcx
 	push    rdx
-	lea     arg1, [gf_vect_mul_base WRT_OPT] ; Default
+	lea     arg1, [GF_VECT_MUL_BASE WRT_OPT] ; Default
 
 	mov     eax, 1
 	cpuid
 	test    ecx, FLAG_CPUID1_ECX_SSE4_2
-	lea     rbx, [gf_vect_mul_sse WRT_OPT]
+	lea     rbx, [GF_VECT_MUL_SSE WRT_OPT]
 	je      _done_gf_vect_mul_dispatch_init
 	mov     arg1, rbx
 
@@ -178,7 +214,7 @@ gf_vect_mul_dispatch_init:
 	and     eax, FLAG_XGETBV_EAX_XMM_YMM
 	cmp     eax, FLAG_XGETBV_EAX_XMM_YMM
 	jne     _done_gf_vect_mul_dispatch_init
-	lea     arg1, [gf_vect_mul_avx WRT_OPT]
+	lea     arg1, [GF_VECT_MUL_AVX WRT_OPT]
 
 _done_gf_vect_mul_dispatch_init:
         pop     rdx
@@ -194,33 +230,33 @@ _done_gf_vect_mul_dispatch_init:
 ;;;;
 ; gf_vect_dot_prod multibinary function
 ;;;;
-global gf_vect_dot_prod:function
+global GF_VECT_DOT_PROD:function
 gf_vect_dot_prod_mbinit:
 	call    gf_vect_dot_prod_dispatch_init
 
-gf_vect_dot_prod:
+GF_VECT_DOT_PROD:
 	jmp     wrd_sz [gf_vect_dot_prod_dispatched]
 
 gf_vect_dot_prod_dispatch_init:
 	push    arg1
 %ifidn __OUTPUT_FORMAT__, elf32         ;; 32-bit check
-	lea     arg1, [gf_vect_dot_prod_base]
+	lea     arg1, [GF_VECT_DOT_PROD_BASE]
 %else
 	push	rax
 	push	rbx
 	push	rcx
 	push	rdx
-	lea     arg1, [gf_vect_dot_prod_base WRT_OPT] ; Default
+	lea     arg1, [GF_VECT_DOT_PROD_BASE WRT_OPT] ; Default
 
 	mov     eax, 1
 	cpuid
-	lea     rbx, [gf_vect_dot_prod_sse WRT_OPT]
+	lea     rbx, [GF_VECT_DOT_PROD_SSE WRT_OPT]
 	test    ecx, FLAG_CPUID1_ECX_SSE4_1
 	cmovne  arg1, rbx
 
 	and	ecx, (FLAG_CPUID1_ECX_AVX | FLAG_CPUID1_ECX_OSXSAVE)
 	cmp	ecx, (FLAG_CPUID1_ECX_AVX | FLAG_CPUID1_ECX_OSXSAVE)
-	lea     rbx, [gf_vect_dot_prod_avx WRT_OPT]
+	lea     rbx, [GF_VECT_DOT_PROD_AVX WRT_OPT]
 
 	jne     _done_gf_vect_dot_prod_init
 	mov	rsi, rbx
@@ -230,7 +266,7 @@ gf_vect_dot_prod_dispatch_init:
 	mov	eax, 7
 	cpuid
 	test	ebx, FLAG_CPUID1_EBX_AVX2
-	lea     rbx, [gf_vect_dot_prod_avx2 WRT_OPT]
+	lea     rbx, [GF_VECT_DOT_PROD_AVX2 WRT_OPT]
 	cmovne	rsi, rbx
 
 	;; Does it have xmm and ymm support
@@ -239,7 +275,7 @@ gf_vect_dot_prod_dispatch_init:
 	and	eax, FLAG_XGETBV_EAX_XMM_YMM
 	cmp	eax, FLAG_XGETBV_EAX_XMM_YMM
 	je      _done_gf_vect_dot_prod_init
-	lea     rsi, [gf_vect_dot_prod_sse WRT_OPT]
+	lea     rsi, [GF_VECT_DOT_PROD_SSE WRT_OPT]
 
 _done_gf_vect_dot_prod_init:
 	pop     rdx
@@ -261,6 +297,6 @@ global %1_slver
 %endmacro
 
 ;;;       func                  core, ver, snum
-slversion ec_encode_data,	00,   02,  0133
+slversion EC_ENCODE_DATA,	00,   02,  0133
-slversion gf_vect_mul,		00,   02,  0134
+slversion GF_VECT_MUL,		00,   02,  0134
-slversion gf_vect_dot_prod,	00,   01,  0138
+slversion GF_VECT_DOT_PROD,	00,   01,  0138

pkg/erasure/isal/src/gf-2vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_2VECT_DOT_PROD_AVX _gf_2vect_dot_prod_avx
+%else
+ %define GF_2VECT_DOT_PROD_AVX gf_2vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -58,6 +63,31 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r9
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -143,9 +173,8 @@ section .text
 %define xp2    xmm3
 
 align 16
-global gf_2vect_dot_prod_avx:function
+global GF_2VECT_DOT_PROD_AVX:function
-
+func(GF_2VECT_DOT_PROD_AVX)
-func(gf_2vect_dot_prod_avx)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -231,4 +260,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_2vect_dot_prod_avx, 02,  03,  0191
+slversion GF_2VECT_DOT_PROD_AVX, 02,  03,  0191

pkg/erasure/isal/src/gf-2vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_2VECT_DOT_PROD_AVX2 _gf_2vect_dot_prod_avx2
+%else
+ %define GF_2VECT_DOT_PROD_AVX2 gf_2vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -60,6 +65,33 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r9
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -149,9 +181,8 @@ section .text
 %define xp2    ymm3
 
 align 16
-global gf_2vect_dot_prod_avx2:function
+global GF_2VECT_DOT_PROD_AVX2:function
-
+func(GF_2VECT_DOT_PROD_AVX2)
-func(gf_2vect_dot_prod_avx2)
 	FUNC_SAVE
 	sub	len, 32
 	jl	.return_fail
@@ -243,4 +274,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                   core, ver, snum
-slversion gf_2vect_dot_prod_avx2, 04,  03,  0196
+slversion GF_2VECT_DOT_PROD_AVX2, 04,  03,  0196

pkg/erasure/isal/src/gf-2vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_2VECT_DOT_PROD_SSE _gf_2vect_dot_prod_sse
+%else
+ %define GF_2VECT_DOT_PROD_SSE gf_2vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -58,6 +63,31 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r9
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -144,9 +174,8 @@ section .text
 %define xp2    xmm3
 
 align 16
-global gf_2vect_dot_prod_sse:function
+global GF_2VECT_DOT_PROD_SSE:function
-
+func(GF_2VECT_DOT_PROD_SSE)
-func(gf_2vect_dot_prod_sse)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -233,4 +262,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_2vect_dot_prod_sse, 00,  02,  0062
+slversion GF_2VECT_DOT_PROD_SSE, 00,  02,  0062

pkg/erasure/isal/src/gf-3vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_3VECT_DOT_PROD_AVX _gf_3vect_dot_prod_avx
+%else
+ %define GF_3VECT_DOT_PROD_AVX gf_3vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -60,6 +65,33 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -157,8 +189,8 @@ section .text
 %define xp3    xmm4
 
 align 16
-global gf_3vect_dot_prod_avx:function
+global GF_3VECT_DOT_PROD_AVX:function
-func(gf_3vect_dot_prod_avx)
+func(GF_3VECT_DOT_PROD_AVX)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -255,4 +287,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_3vect_dot_prod_avx, 02,  03,  0192
+slversion GF_3VECT_DOT_PROD_AVX, 02,  03,  0192

pkg/erasure/isal/src/gf-3vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_3VECT_DOT_PROD_AVX2 _gf_3vect_dot_prod_avx2
+%else
+ %define GF_3VECT_DOT_PROD_AVX2 gf_3vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -62,6 +67,35 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -162,8 +196,8 @@ section .text
 %define xp3    ymm4
 
 align 16
-global gf_3vect_dot_prod_avx2:function
+global GF_3VECT_DOT_PROD_AVX2:function
-func(gf_3vect_dot_prod_avx2)
+func(GF_3VECT_DOT_PROD_AVX2)
 	FUNC_SAVE
 	sub	len, 32
 	jl	.return_fail
@@ -268,4 +302,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                   core, ver, snum
-slversion gf_3vect_dot_prod_avx2, 04,  03,  0197
+slversion GF_3VECT_DOT_PROD_AVX2, 04,  03,  0197

pkg/erasure/isal/src/gf-3vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_3VECT_DOT_PROD_SSE _gf_3vect_dot_prod_sse
+%else
+ %define GF_3VECT_DOT_PROD_SSE gf_3vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -60,6 +65,33 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -157,8 +189,8 @@ section .text
 %define xp3    xmm4
 
 align 16
-global gf_3vect_dot_prod_sse:function
+global GF_3VECT_DOT_PROD_SSE:function
-func(gf_3vect_dot_prod_sse)
+func(GF_3VECT_DOT_PROD_SSE)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -256,4 +288,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_3vect_dot_prod_sse, 00,  03,  0063
+slversion GF_3VECT_DOT_PROD_SSE, 00,  03,  0063

pkg/erasure/isal/src/gf-4vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_4VECT_DOT_PROD_AVX _gf_4vect_dot_prod_avx
+%else
+ %define GF_4VECT_DOT_PROD_AVX gf_4vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -182,8 +220,8 @@ section .text
 %define xp4    xmm5
 
 align 16
-global gf_4vect_dot_prod_avx:function
+global GF_4VECT_DOT_PROD_AVX:function
-func(gf_4vect_dot_prod_avx)
+func(GF_4VECT_DOT_PROD_AVX)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -293,4 +331,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_4vect_dot_prod_avx, 00,  02,  0064
+slversion GF_4VECT_DOT_PROD_AVX, 00,  02,  0064

pkg/erasure/isal/src/gf-4vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_4VECT_DOT_PROD_AVX2 _gf_4vect_dot_prod_avx2
+%else
+ %define GF_4VECT_DOT_PROD_AVX2 gf_4vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -68,6 +73,41 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -187,8 +227,8 @@ section .text
 %define xp4    ymm5
 
 align 16
-global gf_4vect_dot_prod_avx2:function
+global GF_4VECT_DOT_PROD_AVX2:function
-func(gf_4vect_dot_prod_avx2)
+func(GF_4VECT_DOT_PROD_AVX2)
 	FUNC_SAVE
 	sub	len, 32
 	jl	.return_fail
@@ -302,4 +342,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                   core, ver, snum
-slversion gf_4vect_dot_prod_avx2, 04,  03,  0064
+slversion GF_4VECT_DOT_PROD_AVX2, 04,  03,  0064

pkg/erasure/isal/src/gf-4vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_4VECT_DOT_PROD_SSE _gf_4vect_dot_prod_sse
+%else
+ %define GF_4VECT_DOT_PROD_SSE gf_4vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -182,8 +220,8 @@ section .text
 %define xp4    xmm5
 
 align 16
-global gf_4vect_dot_prod_sse:function
+global GF_4VECT_DOT_PROD_SSE:function
-func(gf_4vect_dot_prod_sse)
+func(GF_4VECT_DOT_PROD_SSE)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -293,4 +331,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_4vect_dot_prod_sse, 00,  03,  0064
+slversion GF_4VECT_DOT_PROD_SSE, 00,  03,  0064

pkg/erasure/isal/src/gf-5vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_5VECT_DOT_PROD_AVX _gf_5vect_dot_prod_avx
+%else
+ %define GF_5VECT_DOT_PROD_AVX gf_5vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx

pkg/erasure/isal/src/gf-5vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_5VECT_DOT_PROD_AVX2 _gf_5vect_dot_prod_avx2
+%else
+ %define GF_5VECT_DOT_PROD_AVX2 gf_5vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -68,6 +73,41 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -189,8 +229,8 @@ section .text
 %define xp5    ymm6
 
 align 16
-global gf_5vect_dot_prod_avx2:function
+global GF_5VECT_DOT_PROD_AVX2:function
-func(gf_5vect_dot_prod_avx2)
+func(GF_5VECT_DOT_PROD_AVX2)
 	FUNC_SAVE
 	sub	len, 32
 	jl	.return_fail
@@ -320,4 +360,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_5vect_dot_prod_avx2, 04,  03,  0199
+slversion GF_5VECT_DOT_PROD_AVX2, 04,  03,  0199

pkg/erasure/isal/src/gf-5vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_5VECT_DOT_PROD_SSE _gf_5vect_dot_prod_sse
+%else
+ %define GF_5VECT_DOT_PROD_SSE gf_5vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -184,8 +222,8 @@ section .text
 %define xp5    xmm6
 
 align 16
-global gf_5vect_dot_prod_sse:function
+global GF_5VECT_DOT_PROD_SSE:function
-func(gf_5vect_dot_prod_sse)
+func(GF_5VECT_DOT_PROD_SSE)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -309,4 +347,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_5vect_dot_prod_sse, 00,  03,  0065
+slversion GF_5VECT_DOT_PROD_SSE, 00,  03,  0065

pkg/erasure/isal/src/gf-6vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_6VECT_DOT_PROD_AVX _gf_6vect_dot_prod_avx
+%else
+ %define GF_6VECT_DOT_PROD_AVX gf_6vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -182,8 +220,8 @@ section .text
 %define xp6    xmm7
 
 align 16
-global gf_6vect_dot_prod_avx:function
+global GF_6VECT_DOT_PROD_AVX:function
-func(gf_6vect_dot_prod_avx)
+func(GF_6VECT_DOT_PROD_AVX)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -320,4 +358,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_6vect_dot_prod_avx, 02,  03,  0195
+slversion GF_6VECT_DOT_PROD_AVX, 02,  03,  0195

pkg/erasure/isal/src/gf-6vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_6VECT_DOT_PROD_AVX2 _gf_6vect_dot_prod_avx2
+%else
+ %define GF_6VECT_DOT_PROD_AVX2 gf_6vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -68,6 +73,41 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx

pkg/erasure/isal/src/gf-6vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_6VECT_DOT_PROD_SSE _gf_6vect_dot_prod_sse
+%else
+ %define GF_6VECT_DOT_PROD_SSE gf_6vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -66,6 +71,39 @@
  %endmacro
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r13		; must be saved and restored
+ %define tmp4  r12		; must be saved and restored
+ %define tmp5  r14		; must be saved and restored
+ %define tmp6  r15		; must be saved and restored
+ %define return rax
+ %define PS 8
+ %define LOG_PS 3
+
+ %define func(x) x:
+ %macro FUNC_SAVE 0
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ %endmacro
+ %macro FUNC_RESTORE 0
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ %endmacro
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -182,8 +220,8 @@ section .text
 %define xp6    xmm7
 
 align 16
-global gf_6vect_dot_prod_sse:function
+global GF_6VECT_DOT_PROD_SSE:function
-func(gf_6vect_dot_prod_sse)
+func(GF_6VECT_DOT_PROD_SSE)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -320,4 +358,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_6vect_dot_prod_sse, 00,  03,  0066
+slversion GF_6VECT_DOT_PROD_SSE, 00,  03,  0066

pkg/erasure/isal/src/gf-vect-dot-prod-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_VECT_DOT_PROD_AVX _gf_vect_dot_prod_avx
+%else
+ %define GF_VECT_DOT_PROD_AVX gf_vect_dot_prod_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -51,6 +56,24 @@
  %define FUNC_RESTORE
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r9
+ %define return rax
+ %define PS 8
+ %define func(x) x:
+ %define FUNC_SAVE
+ %define FUNC_RESTORE
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -121,8 +144,8 @@ section .text
 %define xp     xmm2
 
 align 16
-global gf_vect_dot_prod_avx:function
+global GF_VECT_DOT_PROD_AVX:function
-func(gf_vect_dot_prod_avx)
+func(GF_VECT_DOT_PROD_AVX)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -195,4 +218,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                 core, ver, snum
-slversion gf_vect_dot_prod_avx, 02,  03,  0061
+slversion GF_VECT_DOT_PROD_AVX, 02,  03,  0061

pkg/erasure/isal/src/gf-vect-dot-prod-avx2.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_VECT_DOT_PROD_AVX2 _gf_vect_dot_prod_avx2
+%else
+ %define GF_VECT_DOT_PROD_AVX2 gf_vect_dot_prod_avx2
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -53,6 +58,26 @@
  %define FUNC_RESTORE
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+
+ %define tmp   r11
+ %define tmp.w r11d
+ %define tmp.b r11b
+ %define tmp2  r10
+ %define tmp3  r9
+ %define return rax
+ %define PS 8
+ %define func(x) x:
+ %define FUNC_SAVE
+ %define FUNC_RESTORE
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -126,8 +151,8 @@ section .text
 %define xp     ymm2
 
 align 16
-global gf_vect_dot_prod_avx2:function
+global GF_VECT_DOT_PROD_AVX2:function
-func(gf_vect_dot_prod_avx2)
+func(GF_VECT_DOT_PROD_AVX2)
 	FUNC_SAVE
 	sub	len, 32
 	jl	.return_fail
@@ -200,4 +225,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                  core, ver, snum
-slversion gf_vect_dot_prod_avx2, 04,  03,  0190
+slversion GF_VECT_DOT_PROD_AVX2, 04,  03,  0190

pkg/erasure/isal/src/gf-vect-dot-prod-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_VECT_DOT_PROD_SSE _gf_vect_dot_prod_sse
+%else
+ %define GF_VECT_DOT_PROD_SSE gf_vect_dot_prod_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -50,6 +55,23 @@
  %define FUNC_RESTORE
 %endif
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+
+ %define tmp   r11
+ %define tmp2  r10
+ %define tmp3  r9
+ %define return rax
+ %define PS 8
+ %define func(x) x:
+ %define FUNC_SAVE
+ %define FUNC_RESTORE
+%endif
+
 %ifidn __OUTPUT_FORMAT__, win64
  %define arg0   rcx
  %define arg1   rdx
@@ -121,8 +143,8 @@ section .text
 %define xp     xmm2
 
 align 16
-global gf_vect_dot_prod_sse:function
+global GF_VECT_DOT_PROD_SSE:function
-func(gf_vect_dot_prod_sse)
+func(GF_VECT_DOT_PROD_SSE)
 	FUNC_SAVE
 	sub	len, 16
 	jl	.return_fail
@@ -192,4 +214,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func                 core, ver, snum
-slversion gf_vect_dot_prod_sse, 00,  03,  0060
+slversion GF_VECT_DOT_PROD_SSE, 00,  03,  0060

pkg/erasure/isal/src/gf-vect-mul-avx.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_VECT_MUL_AVX _gf_vect_mul_avx
+%else
+ %define GF_VECT_MUL_AVX gf_vect_mul_avx
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -46,6 +51,19 @@
  %define FUNC_SAVE
  %define FUNC_RESTORE
 
+%elifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+ %define tmp   r11
+ %define return rax
+ %define func(x) x:
+ %define FUNC_SAVE
+ %define FUNC_RESTORE
+
 %elifidn __OUTPUT_FORMAT__, win64
  %define arg0  rcx
  %define arg1  rdx
@@ -111,8 +129,8 @@ section .text
 %define xtmp2c xmm7
 
 align 16
-global gf_vect_mul_avx:function
+global GF_VECT_MUL_AVX:function
-func(gf_vect_mul_avx)
+func(GF_VECT_MUL_AVX)
 	FUNC_SAVE
 	mov	pos, 0
 	vmovdqa	xmask0f, [mask0f]	;Load mask of lower nibble in each byte
@@ -169,4 +187,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func             core, ver, snum
-slversion gf_vect_mul_avx, 01,   02,  0036
+slversion GF_VECT_MUL_AVX, 01,   02,  0036

pkg/erasure/isal/src/gf-vect-mul-sse.asm

@@ -32,6 +32,11 @@
 ;;;
 ;;; Author: Gregory Tucker
 
+%ifidn __OUTPUT_FORMAT__, macho64
+ %define GF_VECT_MUL_SSE _gf_vect_mul_sse
+%else
+ %define GF_VECT_MUL_SSE gf_vect_mul_sse
+%endif
 
 %ifidn __OUTPUT_FORMAT__, elf64
  %define arg0  rdi
@@ -46,6 +51,19 @@
  %define FUNC_SAVE
  %define FUNC_RESTORE
 
+%elifidn __OUTPUT_FORMAT__, macho64
+ %define arg0  rdi
+ %define arg1  rsi
+ %define arg2  rdx
+ %define arg3  rcx
+ %define arg4  r8
+ %define arg5  r9
+ %define tmp   r11
+ %define return rax
+ %define func(x) x:
+ %define FUNC_SAVE
+ %define FUNC_RESTORE
+
 %elifidn __OUTPUT_FORMAT__, win64
  %define arg0  rcx
  %define arg1  rdx
@@ -112,8 +130,8 @@ section .text
 
 
 align 16
-global gf_vect_mul_sse:function
+global GF_VECT_MUL_SSE:function
-func(gf_vect_mul_sse)
+func(GF_VECT_MUL_SSE)
 	FUNC_SAVE
 	mov	pos, 0
 	movdqa	xmask0f, [mask0f]	;Load mask of lower nibble in each byte
@@ -175,4 +193,4 @@ global %1_slver
 	db 0x%3, 0x%2
 %endmacro
 ;;;       func        core, ver, snum
-slversion gf_vect_mul_sse, 00,   02,  0034
+slversion GF_VECT_MUL_SSE, 00,   02,  0034

pkg/utils/execpipe_test.go

@@ -23,7 +23,7 @@ func (s *MySuite) TestPiping(c *C) {
 	// Run the command on each directory
 	for _, dir := range dirs {
 		// find $DIR -type f # Find all files
-		ls := exec.Command("ls", dir, "-l")
+		ls := exec.Command("ls", "-l", dir)
 
 		// | sort -t. -k2 # Sort by file extension
 		sort := exec.Command("sort", "-t.", "-k2")