diff options
| author | Mauri de Souza Meneguzzo <mauri870@gmail.com> | 2023-11-07 18:34:49 +0000 |
|---|---|---|
| committer | Russ Cox <rsc@golang.org> | 2023-11-07 22:36:17 +0000 |
| commit | df0bc9e1bb5eaba2300400ba5294aae0a10b6a54 (patch) | |
| tree | bf266639a07af86b4e38c98709361456dbbaa7bf | |
| parent | 74c2ba9521f1b7082a53cd34a0c4a01284cbd99b (diff) | |
| download | go-x-crypto-df0bc9e1bb5eaba2300400ba5294aae0a10b6a54.tar.xz | |
chacha20poly1305: guard PSHUFB usage with GOAMD64_v2
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.
Fixes golang/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>
| -rw-r--r-- | chacha20poly1305/chacha20poly1305_amd64.s | 24 |
1 files changed, 22 insertions, 2 deletions
diff --git a/chacha20poly1305/chacha20poly1305_amd64.s b/chacha20poly1305/chacha20poly1305_amd64.s index 541d696..731d2ac 100644 --- a/chacha20poly1305/chacha20poly1305_amd64.s +++ b/chacha20poly1305/chacha20poly1305_amd64.s @@ -183,11 +183,31 @@ GLOBL ·andMask<>(SB), (NOPTR+RODATA), $240 #define shiftD1Right BYTE $0x66; BYTE $0x45; BYTE $0x0f; BYTE $0x3a; BYTE $0x0f; BYTE $0xd2; BYTE $0x04 // PALIGNR $4, X10, X10 #define shiftD2Right BYTE $0x66; BYTE $0x45; BYTE $0x0f; BYTE $0x3a; BYTE $0x0f; BYTE $0xdb; BYTE $0x04 // PALIGNR $4, X11, X11 #define shiftD3Right BYTE $0x66; BYTE $0x45; BYTE $0x0f; BYTE $0x3a; BYTE $0x0f; BYTE $0xff; BYTE $0x04 // PALIGNR $4, X15, X15 + // Some macros + +// ROL rotates the uint32s in register R left by N bits, using temporary T. +#define ROL(N, R, T) \ + MOVO R, T; PSLLL $(N), T; PSRLL $(32-(N)), R; PXOR T, R + +// ROL16 rotates the uint32s in register R left by 16, using temporary T if needed. +#ifdef GOAMD64_v2 +#define ROL16(R, T) PSHUFB ·rol16<>(SB), R +#else +#define ROL16(R, T) ROL(16, R, T) +#endif + +// ROL8 rotates the uint32s in register R left by 8, using temporary T if needed. +#ifdef GOAMD64_v2 +#define ROL8(R, T) PSHUFB ·rol8<>(SB), R +#else +#define ROL8(R, T) ROL(8, R, T) +#endif + #define chachaQR(A, B, C, D, T) \ - PADDD B, A; PXOR A, D; PSHUFB ·rol16<>(SB), D \ + PADDD B, A; PXOR A, D; ROL16(D, T) \ PADDD D, C; PXOR C, B; MOVO B, T; PSLLL $12, T; PSRLL $20, B; PXOR T, B \ - PADDD B, A; PXOR A, D; PSHUFB ·rol8<>(SB), D \ + PADDD B, A; PXOR A, D; ROL8(D, T) \ PADDD D, C; PXOR C, B; MOVO B, T; PSLLL $7, T; PSRLL $25, B; PXOR T, B #define chachaQR_AVX2(A, B, C, D, T) \ |