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.
Parameter metadata not found in the reference assembly file has been
added, leading to a diff on the lines where those symbols are
referenced.
Commands used to verify Avo output:
GOROOT=$(go env GOROOT)
ASM_PATH="chacha20poly1305/chacha20poly1305_amd64.s"
REFERENCE="b2d3a6a4b4d36521cd7f653879cf6981e7c5c340"
go tool asm -o /dev/null -I "$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)
155,157c155,157
< MOVQ dst(FP), DI
< MOVQ key+24(FP), R8
< MOVQ src+48(FP), SI
---
> MOVQ dst_base(FP), DI
> MOVQ key_base+24(FP), R8
> MOVQ src_base+48(FP), SI
159c159
< MOVQ ad+72(FP), CX
---
> MOVQ ad_base+72(FP), CX
4684,4686c4684,4686
< MOVQ dst(FP), DI
< MOVQ key+24(FP), R8
< MOVQ src+48(FP), SI
---
> MOVQ dst_base(FP), DI
> MOVQ key_base+24(FP), R8
> MOVQ src_base+48(FP), SI
4688c4688
< MOVQ ad+72(FP), CX
---
> MOVQ ad_base+72(FP), CX
Change-Id: Ia3a8e70b7440944ee739499c41ddceb70e054ef9
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/601442
Reviewed-by: Filippo Valsorda <filippo@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
The PSHUFB instruction is part of SSE which is only v2+ but it is being
used without the GOAMD64_v2 guard.
The ROL macros were copied from CL 516859 that adds internal/chacha8rand.
Fixesgolang/go#63871
Change-Id: I3c8ba75ff284cda4fc788885643246936e617b85
GitHub-Last-Rev: e235e8eae67f16b3a58817cfdff729693faf2665
GitHub-Pull-Request: golang/crypto#275
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/538786
Reviewed-by: Keith Randall <khr@google.com>
Run-TryBot: Mauri de Souza Meneguzzo <mauri870@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Russ Cox <rsc@golang.org>
This change detects BMI2 usability as an additional condition
to examine the usability of AVX2 version algorithm, fixes
the crash on the platfrom which supports AVX2 but not support BMI2.
Change-Id: I5438d4ec84265c79a51c1439265a33b1be04878a
Reviewed-on: https://go-review.googlesource.com/34852
Reviewed-by: Adam Langley <agl@golang.org>
Reviewed-by: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Ian Lance Taylor <iant@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
The names of the constant values were previously not scoped to the
package and |andMask| could collide with the variable of the same name
in the AES-GCM assembly.
Change-Id: I9387a6cace71a585ad2fe2afcc8436d44bde0917
Reviewed-on: https://go-review.googlesource.com/31265
Run-TryBot: Adam Langley <agl@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
This change adds a package, chacha20poly1305, which implements the
ChaCha20-Poly1305 AEAD from RFC 7539. This AEAD has several attractive
features:
1. It's naturally constant time. AES-GCM needs either dedicated
hardware or extreme effort to be fast and constant-time, while
this design is easy to make constant-time.
2. It's fast on modern processors: it runs at 1GB/s on my IvyBrige
system.
3. It's seeing significant use in TLS. (A change for crypto/tls is
forthcoming.)
This change merges two CLs:
https://go-review.googlesource.com/#/c/24717https://go-review.googlesource.com/#/c/26691
I took the amd64-optimised AEAD implementation from the former because
it was significantly faster. But the structure of the change is taken
from the latter.
This version will be checked into x/crypto. This package will then be
vendored into the stdlib so that it can be used from crypto/tls.
Change-Id: I5a60587958b7afeec81ca1091e603a7e8517000b
Reviewed-on: https://go-review.googlesource.com/30728
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Run-TryBot: Brad Fitzpatrick <bradfitz@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Garrett Bodley
Mauri de Souza Meneguzzo
Dmitri Shuralyov
Tobias Klauser
Tamir Duberstein
Filippo Valsorda
Neven Sajko
Michael Munday
Martin Möhrmann
Lion Yang
Mikio Hara
Shenghou Ma
Alex Vaghin
Adam Langley
Andreas Auernhammer
Brad Fitzpatrick