diff options
| author | Beniamino Galvani <bgalvani@redhat.com> | 2024-01-27 15:49:55 +0100 |
|---|---|---|
| committer | Beniamino Galvani <bgalvani@redhat.com> | 2024-02-05 20:54:24 +0100 |
| commit | d86037a4802488c133b14f97e5188cba67d63494 (patch) | |
| tree | f90cea434cdce3ca81139533202f2124cbb783b5 | |
| parent | ebe83f27bb11b70f6f08c2f5a1dd434551c0d29c (diff) | |
contrib/rpm: enable eBPF on Fedora 40bg/n-acd-ebpf
Enable eBPF for ACD in Fedora 40; this will make the daemon more
efficient because the process doesn't need to be awakened for every
ARP packet. In my limited testing I didn't encounter the issues that
were reported in the past.
| -rw-r--r-- | contrib/fedora/rpm/NetworkManager.spec | 9 |
1 files changed, 2 insertions, 7 deletions
diff --git a/contrib/fedora/rpm/NetworkManager.spec b/contrib/fedora/rpm/NetworkManager.spec index 583e962ed9..06c67d01b4 100644 --- a/contrib/fedora/rpm/NetworkManager.spec +++ b/contrib/fedora/rpm/NetworkManager.spec @@ -176,13 +176,8 @@ %global ifcfg_migrate 0 %endif -%if 0%{?fedora} -# Although eBPF would be available on Fedora's kernel, it seems -# we often get SELinux denials (rh#1651654). But even aside them, -# bpf(BPF_MAP_CREATE, ...) randomly fails with EPERM. That might -# be related to `ulimit -l`. Anyway, this is not usable at the -# moment. -%global ebpf_enabled "no" +%if 0%{?fedora} >= 40 +%global ebpf_enabled "yes" %else %global ebpf_enabled "no" %endif |