checksum: Introduce AVX2 implementation, unify helpers

Provide an AVX2-based function using compiler intrinsics for
TCP/IP-style checksums. The load/unpack/add idea and implementation
is largely based on code from BESS (the Berkeley Extensible Software
Switch) licensed as 3-Clause BSD, with a number of modifications to
further decrease pipeline stalls and to minimise cache pollution.

This speeds up considerably data paths from sockets to tap
interfaces, decreasing overhead for checksum computation, with
16-64KiB packet buffers, from approximately 11% to 7%. The rest is
just syscalls at this point.

While at it, provide convenience targets in the Makefile for avx2,
avx2_debug, and debug targets -- these simply add target-specific
CFLAGS to the build.

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

1 files changed, 0 insertions, 4 deletions

diff --git a/util.h b/util.h
index 7fbce1f..1c11474 100644
--- a/util.h
+++ b/util.h
@@ -117,10 +117,6 @@ void debug(const char *format, ...);
 
 struct ctx;
 
-uint32_t sum_16b(void *buf, size_t len);
-uint16_t csum_fold(uint32_t sum);
-uint16_t csum_ip4(void *buf, size_t len);
-void csum_tcp4(struct iphdr *iph);
 char *ipv6_l4hdr(struct ipv6hdr *ip6h, uint8_t *proto);
 int sock_l4(struct ctx *c, int af, uint8_t proto, uint16_t port, int lo,
 	    uint32_t data);