]> git.baikalelectronics.ru Git - kernel.git/commit
Merge branch 'bpf-fix-sock-field-tests'
authorDaniel Borkmann <daniel@iogearbox.net>
Fri, 18 Mar 2022 14:46:59 +0000 (15:46 +0100)
committerDaniel Borkmann <daniel@iogearbox.net>
Fri, 18 Mar 2022 14:47:04 +0000 (15:47 +0100)
commitd6d427d399ac289f1160065c0064b4e616e0cf63
tree0c5f7fffe2b308ba022fd0263a5cb412a88c0e09
parentd1f367db623fac80bb41f0f7106ddb1c9630f220
parenteadc4a1002b4b882513324d5834b022b107bbca4
Merge branch 'bpf-fix-sock-field-tests'

Jakub Sitnicki says:

====================
I think we have reached a consensus [1] on how the test for the 4-byte load from
bpf_sock->dst_port and bpf_sk_lookup->remote_port should look, so here goes v3.

I will submit a separate set of patches for bpf_sk_lookup->remote_port tests.

This series has been tested on x86_64 and s390 on top of recent bpf-next -
1651a812dc16 ("selftests/bpf: Test subprog jit when toggle bpf_jit_harden
repeatedly").

  [1] https://lore.kernel.org/bpf/87k0cwxkzs.fsf@cloudflare.com/

v2 -> v3:
- Split what was previously patch 2 which was doing two things
- Use BPF_TCP_* constants (Martin)
- Treat the result of 4-byte load from dst_port as a 16-bit value (Martin)
- Typo fixup and some rewording in patch 4 description
v1 -> v2:
- Limit read_sk_dst_port only to client traffic (patch 2)
- Make read_sk_dst_port pass on litte- and big-endian (patch 3)

v1: https://lore.kernel.org/bpf/20220225184130.483208-1-jakub@cloudflare.com/
v2: https://lore.kernel.org/bpf/20220227202757.519015-1-jakub@cloudflare.com/
====================

Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>