powerpc/net: Implement powerpc specific csum_shift() to remove branch

Today's implementation of csum_shift() leads to branching based on parity of 'offset' 000002f8 <csum_block_add>: 2f8: 70 a5 00 01 andi. r5,r5,1 2fc: 41 a2 00 08 beq 304 <csum_block_add+0xc> 300: 54 84 c0 3e rotlwi r4,r4,24 304: 7c 63 20 14 addc r3,r3,r4 308: 7c 63 01 94 addze r3,r3 30c: 4e 80 00 20 blr Use first bit of 'offset' directly as input of the rotation instead of branching. 000002f8 <csum_block_add>: 2f8: 54 a5 1f 38 rlwinm r5,r5,3,28,28 2fc: 20 a5 00 20 subfic r5,r5,32 300: 5c 84 28 3e rotlw r4,r4,r5 304: 7c 63 20 14 addc r3,r3,r4 308: 7c 63 01 94 addze r3,r3 30c: 4e 80 00 20 blr And change to left shift instead of right shift to skip one more instruction. This has no impact on the final sum. 000002f8 <csum_block_add>: 2f8: 54 a5 1f 38 rlwinm r5,r5,3,28,28 2fc: 5c 84 28 3e rotlw r4,r4,r5 300: 7c 63 20 14 addc r3,r3,r4 304: 7c 63 01 94 addze r3,r3 308: 4e 80 00 20 blr Seems like only powerpc benefits from a branchless implementation. Other main architectures like ARM or X86 get better code with the generic implementation and its branch. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: David S. Miller <davem@davemloft.net>
mariux64 · Mar 11, 2022 · 3af722c · 3af722c
1 parent 8ef1dc4
commit 3af722c
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 0 deletions.
diff --git a/arch/powerpc/include/asm/checksum.h b/arch/powerpc/include/asm/checksum.h
@@ -112,6 +112,13 @@ static __always_inline __wsum csum_add(__wsum csum, __wsum addend)
 #endif
 }
 
+#define HAVE_ARCH_CSUM_SHIFT
+static __always_inline __wsum csum_shift(__wsum sum, int offset)
+{
+	/* rotate sum to align it with a 16b boundary */
+	return (__force __wsum)rol32((__force u32)sum, (offset & 1) << 3);
+}
+
 /*
  * This is a version of ip_compute_csum() optimized for IP headers,
  * which always checksum on 4 octet boundaries.  ihl is the number

diff --git a/include/net/checksum.h b/include/net/checksum.h
@@ -80,13 +80,15 @@ static __always_inline __sum16 csum16_sub(__sum16 csum, __be16 addend)
 	return csum16_add(csum, ~addend);
 }
 
+#ifndef HAVE_ARCH_CSUM_SHIFT
 static __always_inline __wsum csum_shift(__wsum sum, int offset)
 {
 	/* rotate sum to align it with a 16b boundary */
 	if (offset & 1)
 		return (__force __wsum)ror32((__force u32)sum, 8);
 	return sum;
 }
+#endif
 
 static __always_inline __wsum
 csum_block_add(__wsum csum, __wsum csum2, int offset)