sha512/sha256: remove assembler macros from AMD64 implementations

* configure.ac (gcry_cv_gcc_platform_as_ok_for_intel_syntax): Remove
assembler macro check from Intel syntax assembly support check.
* cipher/sha256-avx-amd64.S: Replace assembler macros with C
preprocessor counterparts.
* cipher/sha256-avx2-bmi2-amd64.S: Ditto.
* cipher/sha256-ssse3-amd64.S: Ditto.
* cipher/sha512-avx-amd64.S: Ditto.
* cipher/sha512-avx2-bmi2-amd64.S: Ditto.
* cipher/sha512-ssse3-amd64.S: Ditto.
--

Removing GNU assembler macros allows building these implementations with
clang.

GnuPG-bug-id: 5255
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>

1 files changed, 252 insertions, 277 deletions

diff --git a/cipher/sha256-ssse3-amd64.S b/cipher/sha256-ssse3-amd64.S
index 0fb94c1b..098b0eb6 100644
--- a/cipher/sha256-ssse3-amd64.S
+++ b/cipher/sha256-ssse3-amd64.S
@@ -70,58 +70,56 @@
 
 /* addm [mem], reg
  * Add reg to mem using reg-mem add and store */
-.macro addm p1 p2
-	add	\p2, \p1
-	mov	\p1, \p2
-.endm
+#define addm(p1, p2) \
+	add	p2, p1; \
+	mov	p1, p2;
 
 /*;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;*/
 
 /* COPY_XMM_AND_BSWAP xmm, [mem], byte_flip_mask
  * Load xmm with mem and byte swap each dword */
-.macro COPY_XMM_AND_BSWAP p1 p2 p3
-	MOVDQ \p1, \p2
-	pshufb \p1, \p3
-.endm
+#define COPY_XMM_AND_BSWAP(p1, p2, p3) \
+	MOVDQ p1, p2; \
+	pshufb p1, p3;
 
 /*;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;*/
 
-X0 = xmm4
-X1 = xmm5
-X2 = xmm6
-X3 = xmm7
+#define X0 xmm4
+#define X1 xmm5
+#define X2 xmm6
+#define X3 xmm7
 
-XTMP0 = xmm0
-XTMP1 = xmm1
-XTMP2 = xmm2
-XTMP3 = xmm3
-XTMP4 = xmm8
-XFER  = xmm9
+#define XTMP0 xmm0
+#define XTMP1 xmm1
+#define XTMP2 xmm2
+#define XTMP3 xmm3
+#define XTMP4 xmm8
+#define XFER xmm9
 
-SHUF_00BA = xmm10 /* shuffle xBxA -> 00BA */
-SHUF_DC00 = xmm11 /* shuffle xDxC -> DC00 */
-BYTE_FLIP_MASK = xmm12
+#define SHUF_00BA xmm10 /* shuffle xBxA -> 00BA */
+#define SHUF_DC00 xmm11 /* shuffle xDxC -> DC00 */
+#define BYTE_FLIP_MASK xmm12
 
-NUM_BLKS = rdx	/* 3rd arg */
-CTX = rsi	/* 2nd arg */
-INP = rdi	/* 1st arg */
+#define NUM_BLKS rdx	/* 3rd arg */
+#define CTX rsi	/* 2nd arg */
+#define INP rdi	/* 1st arg */
 
-SRND = rdi	/* clobbers INP */
-c = ecx
-d = r8d
-e = edx
+#define SRND rdi	/* clobbers INP */
+#define c ecx
+#define d r8d
+#define e edx
 
-TBL = rbp
-a = eax
-b = ebx
+#define TBL rbp
+#define a eax
+#define b ebx
 
-f = r9d
-g = r10d
-h = r11d
+#define f r9d
+#define g r10d
+#define h r11d
 
-y0 = r13d
-y1 = r14d
-y2 = r15d
+#define y0 r13d
+#define y1 r14d
+#define y2 r15d
 
 
 
@@ -138,230 +136,207 @@ y2 = r15d
 #define _XMM_SAVE	(_XFER     + _XFER_SIZE + _ALIGN_SIZE)
 #define STACK_SIZE	(_XMM_SAVE + _XMM_SAVE_SIZE)
 
-/* rotate_Xs
- * Rotate values of symbols X0...X3 */
-.macro rotate_Xs
-X_ = X0
-X0 = X1
-X1 = X2
-X2 = X3
-X3 = X_
-.endm
-
-/* ROTATE_ARGS
- * Rotate values of symbols a...h */
-.macro ROTATE_ARGS
-TMP_ = h
-h = g
-g = f
-f = e
-e = d
-d = c
-c = b
-b = a
-a = TMP_
-.endm
-
-.macro FOUR_ROUNDS_AND_SCHED
-		/* compute s0 four at a time and s1 two at a time
-		 * compute W[-16] + W[-7] 4 at a time */
-		movdqa	XTMP0, X3
-	mov	y0, e		/* y0 = e */
-	ror	y0, (25-11)	/* y0 = e >> (25-11) */
-	mov	y1, a		/* y1 = a */
-		palignr	XTMP0, X2, 4	/* XTMP0 = W[-7] */
-	ror	y1, (22-13)	/* y1 = a >> (22-13) */
-	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */
-	mov	y2, f		/* y2 = f */
-	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
-		movdqa	XTMP1, X1
-	xor	y1, a		/* y1 = a ^ (a >> (22-13) */
-	xor	y2, g		/* y2 = f^g */
-		paddd	XTMP0, X0	/* XTMP0 = W[-7] + W[-16] */
-	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
-	and	y2, e		/* y2 = (f^g)&e */
-	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
-		/* compute s0 */
-		palignr	XTMP1, X0, 4	/* XTMP1 = W[-15] */
-	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
-	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
-	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */
-		movdqa	XTMP2, XTMP1	/* XTMP2 = W[-15] */
-	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
-	add	y2, y0		/* y2 = S1 + CH */
-	add	y2, [rsp + _XFER + 0*4]	/* y2 = k + w + S1 + CH */
-		movdqa	XTMP3, XTMP1	/* XTMP3 = W[-15] */
-	mov	y0, a		/* y0 = a */
-	add	h, y2		/* h = h + S1 + CH + k + w */
-	mov	y2, a		/* y2 = a */
-		pslld	XTMP1, (32-7)
-	or	y0, c		/* y0 = a|c */
-	add	d, h		/* d = d + h + S1 + CH + k + w */
-	and	y2, c		/* y2 = a&c */
-		psrld	XTMP2, 7
-	and	y0, b		/* y0 = (a|c)&b */
-	add	h, y1		/* h = h + S1 + CH + k + w + S0 */
-		por	XTMP1, XTMP2	/* XTMP1 = W[-15] ror 7 */
-	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */
+
+#define FOUR_ROUNDS_AND_SCHED_0(X0, X1, X2, X3, a, b, c, d, e, f, g, h) \
+		/* compute s0 four at a time and s1 two at a time */; \
+		/* compute W[-16] + W[-7] 4 at a time */; \
+		movdqa	XTMP0, X3; \
+	mov	y0, e		/* y0 = e */; \
+	ror	y0, (25-11)	/* y0 = e >> (25-11) */; \
+	mov	y1, a		/* y1 = a */; \
+		palignr	XTMP0, X2, 4	/* XTMP0 = W[-7] */; \
+	ror	y1, (22-13)	/* y1 = a >> (22-13) */; \
+	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */; \
+	mov	y2, f		/* y2 = f */; \
+	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */; \
+		movdqa	XTMP1, X1; \
+	xor	y1, a		/* y1 = a ^ (a >> (22-13) */; \
+	xor	y2, g		/* y2 = f^g */; \
+		paddd	XTMP0, X0	/* XTMP0 = W[-7] + W[-16] */; \
+	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */; \
+	and	y2, e		/* y2 = (f^g)&e */; \
+	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */; \
+		/* compute s0 */; \
+		palignr	XTMP1, X0, 4	/* XTMP1 = W[-15] */; \
+	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */; \
+	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */; \
+	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */; \
+		movdqa	XTMP2, XTMP1	/* XTMP2 = W[-15] */; \
+	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */; \
+	add	y2, y0		/* y2 = S1 + CH */; \
+	add	y2, [rsp + _XFER + 0*4]	/* y2 = k + w + S1 + CH */; \
+		movdqa	XTMP3, XTMP1	/* XTMP3 = W[-15] */; \
+	mov	y0, a		/* y0 = a */; \
+	add	h, y2		/* h = h + S1 + CH + k + w */; \
+	mov	y2, a		/* y2 = a */; \
+		pslld	XTMP1, (32-7); \
+	or	y0, c		/* y0 = a|c */; \
+	add	d, h		/* d = d + h + S1 + CH + k + w */; \
+	and	y2, c		/* y2 = a&c */; \
+		psrld	XTMP2, 7; \
+	and	y0, b		/* y0 = (a|c)&b */; \
+	add	h, y1		/* h = h + S1 + CH + k + w + S0 */; \
+		por	XTMP1, XTMP2	/* XTMP1 = W[-15] ror 7 */; \
+	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */; \
 	lea	h, [h + y0]	/* h = h + S1 + CH + k + w + S0 + MAJ */
 
-ROTATE_ARGS
-		movdqa	XTMP2, XTMP3	/* XTMP2 = W[-15] */
-	mov	y0, e		/* y0 = e */
-	mov	y1, a		/* y1 = a */
-		movdqa	XTMP4, XTMP3	/* XTMP4 = W[-15] */
-	ror	y0, (25-11)	/* y0 = e >> (25-11) */
-	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */
-	mov	y2, f		/* y2 = f */
-	ror	y1, (22-13)	/* y1 = a >> (22-13) */
-		pslld	XTMP3, (32-18)
-	xor	y1, a		/* y1 = a ^ (a >> (22-13) */
-	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
-	xor	y2, g		/* y2 = f^g */
-		psrld	XTMP2, 18
-	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
-	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
-	and	y2, e		/* y2 = (f^g)&e */
-	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
-		pxor	XTMP1, XTMP3
-	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
-	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */
-		psrld	XTMP4, 3	/* XTMP4 = W[-15] >> 3 */
-	add	y2, y0		/* y2 = S1 + CH */
-	add	y2, [rsp + _XFER + 1*4]	/* y2 = k + w + S1 + CH */
-	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
-		pxor	XTMP1, XTMP2	/* XTMP1 = W[-15] ror 7 ^ W[-15] ror 18 */
-	mov	y0, a		/* y0 = a */
-	add	h, y2		/* h = h + S1 + CH + k + w */
-	mov	y2, a		/* y2 = a */
-		pxor	XTMP1, XTMP4	/* XTMP1 = s0 */
-	or	y0, c		/* y0 = a|c */
-	add	d, h		/* d = d + h + S1 + CH + k + w */
-	and	y2, c		/* y2 = a&c */
-		/* compute low s1 */
-		pshufd	XTMP2, X3, 0b11111010	/* XTMP2 = W[-2] {BBAA} */
-	and	y0, b		/* y0 = (a|c)&b */
-	add	h, y1		/* h = h + S1 + CH + k + w + S0 */
-		paddd	XTMP0, XTMP1	/* XTMP0 = W[-16] + W[-7] + s0 */
-	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */
+#define FOUR_ROUNDS_AND_SCHED_1(X0, X1, X2, X3, a, b, c, d, e, f, g, h) \
+		movdqa	XTMP2, XTMP3	/* XTMP2 = W[-15] */; \
+	mov	y0, e		/* y0 = e */; \
+	mov	y1, a		/* y1 = a */; \
+		movdqa	XTMP4, XTMP3	/* XTMP4 = W[-15] */; \
+	ror	y0, (25-11)	/* y0 = e >> (25-11) */; \
+	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */; \
+	mov	y2, f		/* y2 = f */; \
+	ror	y1, (22-13)	/* y1 = a >> (22-13) */; \
+		pslld	XTMP3, (32-18); \
+	xor	y1, a		/* y1 = a ^ (a >> (22-13) */; \
+	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */; \
+	xor	y2, g		/* y2 = f^g */; \
+		psrld	XTMP2, 18; \
+	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */; \
+	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */; \
+	and	y2, e		/* y2 = (f^g)&e */; \
+	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */; \
+		pxor	XTMP1, XTMP3; \
+	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */; \
+	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */; \
+		psrld	XTMP4, 3	/* XTMP4 = W[-15] >> 3 */; \
+	add	y2, y0		/* y2 = S1 + CH */; \
+	add	y2, [rsp + _XFER + 1*4]	/* y2 = k + w + S1 + CH */; \
+	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */; \
+		pxor	XTMP1, XTMP2	/* XTMP1 = W[-15] ror 7 ^ W[-15] ror 18 */; \
+	mov	y0, a		/* y0 = a */; \
+	add	h, y2		/* h = h + S1 + CH + k + w */; \
+	mov	y2, a		/* y2 = a */; \
+		pxor	XTMP1, XTMP4	/* XTMP1 = s0 */; \
+	or	y0, c		/* y0 = a|c */; \
+	add	d, h		/* d = d + h + S1 + CH + k + w */; \
+	and	y2, c		/* y2 = a&c */; \
+		/* compute low s1 */; \
+		pshufd	XTMP2, X3, 0b11111010	/* XTMP2 = W[-2] {BBAA} */; \
+	and	y0, b		/* y0 = (a|c)&b */; \
+	add	h, y1		/* h = h + S1 + CH + k + w + S0 */; \
+		paddd	XTMP0, XTMP1	/* XTMP0 = W[-16] + W[-7] + s0 */; \
+	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */; \
 	lea	h, [h + y0]	/* h = h + S1 + CH + k + w + S0 + MAJ */
 
-ROTATE_ARGS
-		movdqa	XTMP3, XTMP2	/* XTMP3 = W[-2] {BBAA} */
-	mov	y0, e		/* y0 = e */
-	mov	y1, a		/* y1 = a */
-	ror	y0, (25-11)	/* y0 = e >> (25-11) */
-		movdqa	XTMP4, XTMP2	/* XTMP4 = W[-2] {BBAA} */
-	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */
-	ror	y1, (22-13)	/* y1 = a >> (22-13) */
-	mov	y2, f		/* y2 = f */
-	xor	y1, a		/* y1 = a ^ (a >> (22-13) */
-	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
-		psrlq	XTMP2, 17	/* XTMP2 = W[-2] ror 17 {xBxA} */
-	xor	y2, g		/* y2 = f^g */
-		psrlq	XTMP3, 19	/* XTMP3 = W[-2] ror 19 {xBxA} */
-	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
-	and	y2, e		/* y2 = (f^g)&e */
-		psrld	XTMP4, 10	/* XTMP4 = W[-2] >> 10 {BBAA} */
-	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
-	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
-	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */
-	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
-		pxor	XTMP2, XTMP3
-	add	y2, y0		/* y2 = S1 + CH */
-	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
-	add	y2, [rsp + _XFER + 2*4]	/* y2 = k + w + S1 + CH */
-		pxor	XTMP4, XTMP2	/* XTMP4 = s1 {xBxA} */
-	mov	y0, a		/* y0 = a */
-	add	h, y2		/* h = h + S1 + CH + k + w */
-	mov	y2, a		/* y2 = a */
-		pshufb	XTMP4, SHUF_00BA	/* XTMP4 = s1 {00BA} */
-	or	y0, c		/* y0 = a|c */
-	add	d, h		/* d = d + h + S1 + CH + k + w */
-	and	y2, c		/* y2 = a&c */
-		paddd	XTMP0, XTMP4	/* XTMP0 = {..., ..., W[1], W[0]} */
-	and	y0, b		/* y0 = (a|c)&b */
-	add	h, y1		/* h = h + S1 + CH + k + w + S0 */
-		/* compute high s1 */
-		pshufd	XTMP2, XTMP0, 0b01010000 /* XTMP2 = W[-2] {DDCC} */
-	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */
+#define FOUR_ROUNDS_AND_SCHED_2(X0, X1, X2, X3, a, b, c, d, e, f, g, h) \
+		movdqa	XTMP3, XTMP2	/* XTMP3 = W[-2] {BBAA} */; \
+	mov	y0, e		/* y0 = e */; \
+	mov	y1, a		/* y1 = a */; \
+	ror	y0, (25-11)	/* y0 = e >> (25-11) */; \
+		movdqa	XTMP4, XTMP2	/* XTMP4 = W[-2] {BBAA} */; \
+	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */; \
+	ror	y1, (22-13)	/* y1 = a >> (22-13) */; \
+	mov	y2, f		/* y2 = f */; \
+	xor	y1, a		/* y1 = a ^ (a >> (22-13) */; \
+	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */; \
+		psrlq	XTMP2, 17	/* XTMP2 = W[-2] ror 17 {xBxA} */; \
+	xor	y2, g		/* y2 = f^g */; \
+		psrlq	XTMP3, 19	/* XTMP3 = W[-2] ror 19 {xBxA} */; \
+	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */; \
+	and	y2, e		/* y2 = (f^g)&e */; \
+		psrld	XTMP4, 10	/* XTMP4 = W[-2] >> 10 {BBAA} */; \
+	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */; \
+	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */; \
+	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */; \
+	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */; \
+		pxor	XTMP2, XTMP3; \
+	add	y2, y0		/* y2 = S1 + CH */; \
+	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */; \
+	add	y2, [rsp + _XFER + 2*4]	/* y2 = k + w + S1 + CH */; \
+		pxor	XTMP4, XTMP2	/* XTMP4 = s1 {xBxA} */; \
+	mov	y0, a		/* y0 = a */; \
+	add	h, y2		/* h = h + S1 + CH + k + w */; \
+	mov	y2, a		/* y2 = a */; \
+		pshufb	XTMP4, SHUF_00BA	/* XTMP4 = s1 {00BA} */; \
+	or	y0, c		/* y0 = a|c */; \
+	add	d, h		/* d = d + h + S1 + CH + k + w */; \
+	and	y2, c		/* y2 = a&c */; \
+		paddd	XTMP0, XTMP4	/* XTMP0 = {..., ..., W[1], W[0]} */; \
+	and	y0, b		/* y0 = (a|c)&b */; \
+	add	h, y1		/* h = h + S1 + CH + k + w + S0 */; \
+		/* compute high s1 */; \
+		pshufd	XTMP2, XTMP0, 0b01010000 /* XTMP2 = W[-2] {DDCC} */; \
+	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */; \
 	lea	h, [h + y0]	/* h = h + S1 + CH + k + w + S0 + MAJ */
 
-ROTATE_ARGS
-		movdqa	XTMP3, XTMP2	/* XTMP3 = W[-2] {DDCC} */
-	mov	y0, e		/* y0 = e */
-	ror	y0, (25-11)	/* y0 = e >> (25-11) */
-	mov	y1, a		/* y1 = a */
-		movdqa	X0,    XTMP2	/* X0    = W[-2] {DDCC} */
-	ror	y1, (22-13)	/* y1 = a >> (22-13) */
-	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */
-	mov	y2, f		/* y2 = f */
-	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
-		psrlq	XTMP2, 17	/* XTMP2 = W[-2] ror 17 {xDxC} */
-	xor	y1, a		/* y1 = a ^ (a >> (22-13) */
-	xor	y2, g		/* y2 = f^g */
-		psrlq	XTMP3, 19	/* XTMP3 = W[-2] ror 19 {xDxC} */
-	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
-	and	y2, e		/* y2 = (f^g)&e */
-	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
-		psrld	X0,    10	/* X0 = W[-2] >> 10 {DDCC} */
-	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
-	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
-	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */
-		pxor	XTMP2, XTMP3
-	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
-	add	y2, y0		/* y2 = S1 + CH */
-	add	y2, [rsp + _XFER + 3*4]	/* y2 = k + w + S1 + CH */
-		pxor	X0, XTMP2	/* X0 = s1 {xDxC} */
-	mov	y0, a		/* y0 = a */
-	add	h, y2		/* h = h + S1 + CH + k + w */
-	mov	y2, a		/* y2 = a */
-		pshufb	X0, SHUF_DC00	/* X0 = s1 {DC00} */
-	or	y0, c		/* y0 = a|c */
-	add	d, h		/* d = d + h + S1 + CH + k + w */
-	and	y2, c		/* y2 = a&c */
-		paddd	X0, XTMP0	/* X0 = {W[3], W[2], W[1], W[0]} */
-	and	y0, b		/* y0 = (a|c)&b */
-	add	h, y1		/* h = h + S1 + CH + k + w + S0 */
-	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */
+#define FOUR_ROUNDS_AND_SCHED_3(X0, X1, X2, X3, a, b, c, d, e, f, g, h) \
+		movdqa	XTMP3, XTMP2	/* XTMP3 = W[-2] {DDCC} */; \
+	mov	y0, e		/* y0 = e */; \
+	ror	y0, (25-11)	/* y0 = e >> (25-11) */; \
+	mov	y1, a		/* y1 = a */; \
+		movdqa	X0,    XTMP2	/* X0    = W[-2] {DDCC} */; \
+	ror	y1, (22-13)	/* y1 = a >> (22-13) */; \
+	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */; \
+	mov	y2, f		/* y2 = f */; \
+	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */; \
+		psrlq	XTMP2, 17	/* XTMP2 = W[-2] ror 17 {xDxC} */; \
+	xor	y1, a		/* y1 = a ^ (a >> (22-13) */; \
+	xor	y2, g		/* y2 = f^g */; \
+		psrlq	XTMP3, 19	/* XTMP3 = W[-2] ror 19 {xDxC} */; \
+	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */; \
+	and	y2, e		/* y2 = (f^g)&e */; \
+	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */; \
+		psrld	X0,    10	/* X0 = W[-2] >> 10 {DDCC} */; \
+	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */; \
+	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */; \
+	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */; \
+		pxor	XTMP2, XTMP3; \
+	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */; \
+	add	y2, y0		/* y2 = S1 + CH */; \
+	add	y2, [rsp + _XFER + 3*4]	/* y2 = k + w + S1 + CH */; \
+		pxor	X0, XTMP2	/* X0 = s1 {xDxC} */; \
+	mov	y0, a		/* y0 = a */; \
+	add	h, y2		/* h = h + S1 + CH + k + w */; \
+	mov	y2, a		/* y2 = a */; \
+		pshufb	X0, SHUF_DC00	/* X0 = s1 {DC00} */; \
+	or	y0, c		/* y0 = a|c */; \
+	add	d, h		/* d = d + h + S1 + CH + k + w */; \
+	and	y2, c		/* y2 = a&c */; \
+		paddd	X0, XTMP0	/* X0 = {W[3], W[2], W[1], W[0]} */; \
+	and	y0, b		/* y0 = (a|c)&b */; \
+	add	h, y1		/* h = h + S1 + CH + k + w + S0 */; \
+	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */; \
 	lea	h, [h + y0]	/* h = h + S1 + CH + k + w + S0 + MAJ */
 
-ROTATE_ARGS
-rotate_Xs
-.endm
+#define FOUR_ROUNDS_AND_SCHED(X0, X1, X2, X3, a, b, c, d, e, f, g, h) \
+	FOUR_ROUNDS_AND_SCHED_0(X0, X1, X2, X3, a, b, c, d, e, f, g, h); \
+	FOUR_ROUNDS_AND_SCHED_1(X0, X1, X2, X3, h, a, b, c, d, e, f, g); \
+	FOUR_ROUNDS_AND_SCHED_2(X0, X1, X2, X3, g, h, a, b, c, d, e, f); \
+	FOUR_ROUNDS_AND_SCHED_3(X0, X1, X2, X3, f, g, h, a, b, c, d, e);
 
 /* input is [rsp + _XFER + %1 * 4] */
-.macro DO_ROUND i1
-	mov	y0, e		/* y0 = e */
-	ror	y0, (25-11)	/* y0 = e >> (25-11) */
-	mov	y1, a		/* y1 = a */
-	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */
-	ror	y1, (22-13)	/* y1 = a >> (22-13) */
-	mov	y2, f		/* y2 = f */
-	xor	y1, a		/* y1 = a ^ (a >> (22-13) */
-	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */
-	xor	y2, g		/* y2 = f^g */
-	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */
-	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */
-	and	y2, e		/* y2 = (f^g)&e */
-	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */
-	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */
-	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */
-	add	y2, y0		/* y2 = S1 + CH */
-	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */
-	add	y2, [rsp + _XFER + \i1 * 4]	/* y2 = k + w + S1 + CH */
-	mov	y0, a		/* y0 = a */
-	add	h, y2		/* h = h + S1 + CH + k + w */
-	mov	y2, a		/* y2 = a */
-	or	y0, c		/* y0 = a|c */
-	add	d, h		/* d = d + h + S1 + CH + k + w */
-	and	y2, c		/* y2 = a&c */
-	and	y0, b		/* y0 = (a|c)&b */
-	add	h, y1		/* h = h + S1 + CH + k + w + S0 */
-	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */
+#define DO_ROUND(i1, a, b, c, d, e, f, g, h) \
+	mov	y0, e		/* y0 = e */; \
+	ror	y0, (25-11)	/* y0 = e >> (25-11) */; \
+	mov	y1, a		/* y1 = a */; \
+	xor	y0, e		/* y0 = e ^ (e >> (25-11)) */; \
+	ror	y1, (22-13)	/* y1 = a >> (22-13) */; \
+	mov	y2, f		/* y2 = f */; \
+	xor	y1, a		/* y1 = a ^ (a >> (22-13) */; \
+	ror	y0, (11-6)	/* y0 = (e >> (11-6)) ^ (e >> (25-6)) */; \
+	xor	y2, g		/* y2 = f^g */; \
+	xor	y0, e		/* y0 = e ^ (e >> (11-6)) ^ (e >> (25-6)) */; \
+	ror	y1, (13-2)	/* y1 = (a >> (13-2)) ^ (a >> (22-2)) */; \
+	and	y2, e		/* y2 = (f^g)&e */; \
+	xor	y1, a		/* y1 = a ^ (a >> (13-2)) ^ (a >> (22-2)) */; \
+	ror	y0, 6		/* y0 = S1 = (e>>6) & (e>>11) ^ (e>>25) */; \
+	xor	y2, g		/* y2 = CH = ((f^g)&e)^g */; \
+	add	y2, y0		/* y2 = S1 + CH */; \
+	ror	y1, 2		/* y1 = S0 = (a>>2) ^ (a>>13) ^ (a>>22) */; \
+	add	y2, [rsp + _XFER + i1 * 4]	/* y2 = k + w + S1 + CH */; \
+	mov	y0, a		/* y0 = a */; \
+	add	h, y2		/* h = h + S1 + CH + k + w */; \
+	mov	y2, a		/* y2 = a */; \
+	or	y0, c		/* y0 = a|c */; \
+	add	d, h		/* d = d + h + S1 + CH + k + w */; \
+	and	y2, c		/* y2 = a&c */; \
+	and	y0, b		/* y0 = (a|c)&b */; \
+	add	h, y1		/* h = h + S1 + CH + k + w + S0 */; \
+	or	y0, y2		/* y0 = MAJ = (a|c)&b)|(a&c) */; \
 	lea	h, [h + y0]	/* h = h + S1 + CH + k + w + S0 + MAJ */
-	ROTATE_ARGS
-.endm
 
 /*
 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
@@ -414,10 +389,10 @@ _gcry_sha256_transform_amd64_ssse3:
 	lea	TBL, [.LK256 ADD_RIP]
 
 	/* byte swap first 16 dwords */
-	COPY_XMM_AND_BSWAP	X0, [INP + 0*16], BYTE_FLIP_MASK
-	COPY_XMM_AND_BSWAP	X1, [INP + 1*16], BYTE_FLIP_MASK
-	COPY_XMM_AND_BSWAP	X2, [INP + 2*16], BYTE_FLIP_MASK
-	COPY_XMM_AND_BSWAP	X3, [INP + 3*16], BYTE_FLIP_MASK
+	COPY_XMM_AND_BSWAP(X0, [INP + 0*16], BYTE_FLIP_MASK)
+	COPY_XMM_AND_BSWAP(X1, [INP + 1*16], BYTE_FLIP_MASK)
+	COPY_XMM_AND_BSWAP(X2, [INP + 2*16], BYTE_FLIP_MASK)
+	COPY_XMM_AND_BSWAP(X3, [INP + 3*16], BYTE_FLIP_MASK)
 
 	mov	[rsp + _INP], INP
 
@@ -428,23 +403,23 @@ _gcry_sha256_transform_amd64_ssse3:
 	movdqa	XFER, [TBL + 0*16]
 	paddd	XFER, X0
 	movdqa	[rsp + _XFER], XFER
-	FOUR_ROUNDS_AND_SCHED
+	FOUR_ROUNDS_AND_SCHED(X0, X1, X2, X3, a, b, c, d, e, f, g, h)
 
 	movdqa	XFER, [TBL + 1*16]
-	paddd	XFER, X0
+	paddd	XFER, X1
 	movdqa	[rsp + _XFER], XFER
-	FOUR_ROUNDS_AND_SCHED
+	FOUR_ROUNDS_AND_SCHED(X1, X2, X3, X0, e, f, g, h, a, b, c, d)
 
 	movdqa	XFER, [TBL + 2*16]
-	paddd	XFER, X0
+	paddd	XFER, X2
 	movdqa	[rsp + _XFER], XFER
-	FOUR_ROUNDS_AND_SCHED
+	FOUR_ROUNDS_AND_SCHED(X2, X3, X0, X1, a, b, c, d, e, f, g, h)
 
 	movdqa	XFER, [TBL + 3*16]
-	paddd	XFER, X0
+	paddd	XFER, X3
 	movdqa	[rsp + _XFER], XFER
 	add	TBL, 4*16
-	FOUR_ROUNDS_AND_SCHED
+	FOUR_ROUNDS_AND_SCHED(X3, X0, X1, X2, e, f, g, h, a, b, c, d)
 
 	sub	SRND, 1
 	jne	.Loop1
@@ -453,17 +428,17 @@ _gcry_sha256_transform_amd64_ssse3:
 .Loop2:
 	paddd	X0, [TBL + 0*16]
 	movdqa	[rsp + _XFER], X0
-	DO_ROUND	0
-	DO_ROUND	1
-	DO_ROUND	2
-	DO_ROUND	3
+	DO_ROUND(0, a, b, c, d, e, f, g, h)
+	DO_ROUND(1, h, a, b, c, d, e, f, g)
+	DO_ROUND(2, g, h, a, b, c, d, e, f)
+	DO_ROUND(3, f, g, h, a, b, c, d, e)
 	paddd	X1, [TBL + 1*16]
 	movdqa	[rsp + _XFER], X1
 	add	TBL, 2*16
-	DO_ROUND	0
-	DO_ROUND	1
-	DO_ROUND	2
-	DO_ROUND	3
+	DO_ROUND(0, e, f, g, h, a, b, c, d)
+	DO_ROUND(1, d, e, f, g, h, a, b, c)
+	DO_ROUND(2, c, d, e, f, g, h, a, b)
+	DO_ROUND(3, b, c, d, e, f, g, h, a)
 
 	movdqa	X0, X2
 	movdqa	X1, X3
@@ -471,14 +446,14 @@ _gcry_sha256_transform_amd64_ssse3:
 	sub	SRND, 1
 	jne	.Loop2
 
-	addm	[4*0 + CTX],a
-	addm	[4*1 + CTX],b
-	addm	[4*2 + CTX],c
-	addm	[4*3 + CTX],d
-	addm	[4*4 + CTX],e
-	addm	[4*5 + CTX],f
-	addm	[4*6 + CTX],g
-	addm	[4*7 + CTX],h
+	addm([4*0 + CTX],a)
+	addm([4*1 + CTX],b)
+	addm([4*2 + CTX],c)
+	addm([4*3 + CTX],d)
+	addm([4*4 + CTX],e)
+	addm([4*5 + CTX],f)
+	addm([4*6 + CTX],g)
+	addm([4*7 + CTX],h)
 
 	mov	INP, [rsp + _INP]
 	add	INP, 64