net: fix checksum on big endian CPUs
authorHongzhi Guo <guohongzhi1@huawei.com>
Fri, 10 Jul 2020 11:43:13 +0000 (19:43 +0800)
committerThomas Monjalon <thomas@monjalon.net>
Fri, 10 Jul 2020 22:45:20 +0000 (00:45 +0200)
commit982bb68cab2b93724f9625d53c309e73fa194e6e
treec7324be11b0d938452418b69e1a8b714188209e3
parentd5df2ae0428a147b80bbb65d623f88f75d28b226
net: fix checksum on big endian CPUs

With current code, the checksum of odd-length buffers is wrong on
big endian CPUs: the last byte is not properly summed to the
accumulator.

Fix this by left-shifting the remaining byte by 8. For instance,
if the last byte is 0x42, we should add 0x4200 to the accumulator
on big endian CPUs.

This change is similar to what is suggested in Errata 3133 of
RFC 1071.

Fixes: 6006818cfb26("net: new checksum functions")
Cc: stable@dpdk.org
Signed-off-by: Hongzhi Guo <guohongzhi1@huawei.com>
Reviewed-by: Morten Brørup <mb@smartsharesystems.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
lib/librte_net/rte_ip.h