]> git.ipfire.org Git - thirdparty/systemd.git/commit - src/core/socket.c
bpf: beef up bpf detection, check if BPF_F_ALLOW_MULTI is supported
authorLennart Poettering <lennart@poettering.net>
Fri, 16 Feb 2018 10:55:33 +0000 (11:55 +0100)
committerLennart Poettering <lennart@poettering.net>
Wed, 21 Feb 2018 15:43:36 +0000 (16:43 +0100)
commit2ae7ee58fa03340f767821298475d8845bda0b2c
tree20ffe6635615b6e2ee6c339cf85bfd3b0ef1f2c6
parent8b15fca85b06383534d0428e2be1d64d0ca60caa
bpf: beef up bpf detection, check if BPF_F_ALLOW_MULTI is supported

This improves the BPF/cgroup detection logic, and looks whether
BPF_ALLOW_MULTI is supported. This flag allows execution of multiple
BPF filters in a recursive fashion for a whole cgroup tree. It enables
us to properly report IP accounting for slice units, as well as
delegation of BPF support to units without breaking our own IP
accounting.
src/core/bpf-firewall.c
src/core/bpf-firewall.h
src/core/dbus-cgroup.c
src/core/ip-address-access.c
src/core/socket.c
src/test/test-bpf.c