Garrett Bodley
bcb0f91bbc
This implementation utilizes the same registers found in the reference
implementation, aiming to produce a minimal semantic diff between the
Avo-generated output and the original hand-written assembly.
To verify the Avo implementation, the reference and Avo-generated
assembly files are fed to `go tool asm`, capturing the debug output into
corresponding temp files. The debug output contains supplementary
metadata (line numbers, instruction offsets, and source file references)
that must be removed in order to obtain a semantic diff of the two
files. This is accomplished via a small utility script written in awk.
Commands used to verify Avo output:
GOROOT=$(go env GOROOT)
ASM_PATH="internal/poly1305/sum_amd64.s"
REFERENCE="b2d3a6a4b4d36521cd7f653879cf6981e7c5c340"
go tool asm -o /dev/null -I "$GOROOT"/src/runtime -debug \
<(git cat-file -p "$REFERENCE:$ASM_PATH") \
> /tmp/reference.s
go tool asm -o /dev/null -I "$GOROOT"/src/runtime -debug \
"$ASM_PATH" \
> /tmp/avo.s
normalize(){
awk '{
$1=$2=$3="";
print substr($0,4)
}'
}
diff <(normalize < /tmp/reference.s) <(normalize < /tmp/avo.s)
Change-Id: I80212c95d1b05335d7f6b73a3030b6f812f6105b
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/600035
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
94 lines
1.5 KiB
ArmAsm
94 lines
1.5 KiB
ArmAsm
// Code generated by command: go run sum_amd64_asm.go -out ../sum_amd64.s -pkg poly1305. DO NOT EDIT.
|
|
|
|
//go:build gc && !purego
|
|
|
|
// func update(state *macState, msg []byte)
|
|
TEXT ·update(SB), $0-32
|
|
MOVQ state+0(FP), DI
|
|
MOVQ msg_base+8(FP), SI
|
|
MOVQ msg_len+16(FP), R15
|
|
MOVQ (DI), R8
|
|
MOVQ 8(DI), R9
|
|
MOVQ 16(DI), R10
|
|
MOVQ 24(DI), R11
|
|
MOVQ 32(DI), R12
|
|
CMPQ R15, $0x10
|
|
JB bytes_between_0_and_15
|
|
|
|
loop:
|
|
ADDQ (SI), R8
|
|
ADCQ 8(SI), R9
|
|
ADCQ $0x01, R10
|
|
LEAQ 16(SI), SI
|
|
|
|
multiply:
|
|
MOVQ R11, AX
|
|
MULQ R8
|
|
MOVQ AX, BX
|
|
MOVQ DX, CX
|
|
MOVQ R11, AX
|
|
MULQ R9
|
|
ADDQ AX, CX
|
|
ADCQ $0x00, DX
|
|
MOVQ R11, R13
|
|
IMULQ R10, R13
|
|
ADDQ DX, R13
|
|
MOVQ R12, AX
|
|
MULQ R8
|
|
ADDQ AX, CX
|
|
ADCQ $0x00, DX
|
|
MOVQ DX, R8
|
|
MOVQ R12, R14
|
|
IMULQ R10, R14
|
|
MOVQ R12, AX
|
|
MULQ R9
|
|
ADDQ AX, R13
|
|
ADCQ DX, R14
|
|
ADDQ R8, R13
|
|
ADCQ $0x00, R14
|
|
MOVQ BX, R8
|
|
MOVQ CX, R9
|
|
MOVQ R13, R10
|
|
ANDQ $0x03, R10
|
|
MOVQ R13, BX
|
|
ANDQ $-4, BX
|
|
ADDQ BX, R8
|
|
ADCQ R14, R9
|
|
ADCQ $0x00, R10
|
|
SHRQ $0x02, R14, R13
|
|
SHRQ $0x02, R14
|
|
ADDQ R13, R8
|
|
ADCQ R14, R9
|
|
ADCQ $0x00, R10
|
|
SUBQ $0x10, R15
|
|
CMPQ R15, $0x10
|
|
JAE loop
|
|
|
|
bytes_between_0_and_15:
|
|
TESTQ R15, R15
|
|
JZ done
|
|
MOVQ $0x00000001, BX
|
|
XORQ CX, CX
|
|
XORQ R13, R13
|
|
ADDQ R15, SI
|
|
|
|
flush_buffer:
|
|
SHLQ $0x08, BX, CX
|
|
SHLQ $0x08, BX
|
|
MOVB -1(SI), R13
|
|
XORQ R13, BX
|
|
DECQ SI
|
|
DECQ R15
|
|
JNZ flush_buffer
|
|
ADDQ BX, R8
|
|
ADCQ CX, R9
|
|
ADCQ $0x00, R10
|
|
MOVQ $0x00000010, R15
|
|
JMP multiply
|
|
|
|
done:
|
|
MOVQ R8, (DI)
|
|
MOVQ R9, 8(DI)
|
|
MOVQ R10, 16(DI)
|
|
RET
|