blake2s: remove NOSPLIT from asm functions

Now that the frame sizes of these functions are correct, the compiler
is detecting that they overflow the NOSPLIT stack. They don't need to
be NOSPLIT, so remove that flag.

This fixes the 1.7 build. The build is not broken at tip because the
stack guard is larger.

Change-Id: Ie32779c819f63b27ec0a77a44b7d20ee2d550843
Reviewed-on: https://go-review.googlesource.com/31667
TryBot-Result: Gobot Gobot <gobot@golang.org>
Run-TryBot: Austin Clements <austin@google.com>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

blake2s/blake2s_386.s

@@ -290,7 +290,7 @@ GLOBL counter<>(SB), (NOPTR+RODATA), $16
 	MOVL t, 8*4+off+576(dst)
 
 // func hashBlocksSSE2(h *[8]uint32, c *[2]uint32, flag uint32, blocks []byte)
-TEXT ·hashBlocksSSE2(SB), 4, $672-24 // frame = 656 + 16 byte alignment
+TEXT ·hashBlocksSSE2(SB), 0, $672-24 // frame = 656 + 16 byte alignment
 	MOVL h+0(FP), AX
 	MOVL c+4(FP), BX
 	MOVL flag+8(FP), CX
@@ -359,7 +359,7 @@ loop:
 	RET
 
 // func hashBlocksSSSE3(h *[8]uint32, c *[2]uint32, flag uint32, blocks []byte)
-TEXT ·hashBlocksSSSE3(SB), 4, $704-24 // frame = 688 + 16 byte alignment
+TEXT ·hashBlocksSSSE3(SB), 0, $704-24 // frame = 688 + 16 byte alignment
 	MOVL h+0(FP), AX
 	MOVL c+4(FP), BX
 	MOVL flag+8(FP), CX

blake2s/blake2s_amd64.s

@@ -423,17 +423,17 @@ GLOBL counter<>(SB), (NOPTR+RODATA), $16
 	MOVQ  BP, SP
 
 // func hashBlocksSSE2(h *[8]uint32, c *[2]uint32, flag uint32, blocks []byte)
-TEXT ·hashBlocksSSE2(SB), 4, $672-48 // frame = 656 + 16 byte alignment
+TEXT ·hashBlocksSSE2(SB), 0, $672-48 // frame = 656 + 16 byte alignment
 	HASH_BLOCKS(h+0(FP), c+8(FP), flag+16(FP), blocks_base+24(FP), blocks_len+32(FP), BLAKE2s_SSE2)
 	RET
 
 // func hashBlocksSSSE3(h *[8]uint32, c *[2]uint32, flag uint32, blocks []byte)
-TEXT ·hashBlocksSSSE3(SB), 4, $672-48 // frame = 656 + 16 byte alignment
+TEXT ·hashBlocksSSSE3(SB), 0, $672-48 // frame = 656 + 16 byte alignment
 	HASH_BLOCKS(h+0(FP), c+8(FP), flag+16(FP), blocks_base+24(FP), blocks_len+32(FP), BLAKE2s_SSSE3)
 	RET
 
 // func hashBlocksSSE4(h *[8]uint32, c *[2]uint32, flag uint32, blocks []byte)
-TEXT ·hashBlocksSSE4(SB), 4, $16-48 // frame = 0 + 16 byte alignment
+TEXT ·hashBlocksSSE4(SB), 0, $16-48 // frame = 0 + 16 byte alignment
 	HASH_BLOCKS(h+0(FP), c+8(FP), flag+16(FP), blocks_base+24(FP), blocks_len+32(FP), BLAKE2s_SSE4)
 	RET