Yu Watanabe

@YHNdnzj Thank you for the review. Upgrading the green label.

I have not reviewed this in detail, but the basic concept looks good to me.

@YHNdnzj Thank you for the review. - document is extended, - fixed typo.

@YHNdnzj Do you want this in v256?

Note, the pidfd is actually leaked. ```diff diff --git a/src/libsystemd/sd-bus/test-bus-peersockaddr.c b/src/libsystemd/sd-bus/test-bus-peersockaddr.c index a7bba17a98..255ae68244 100644 --- a/src/libsystemd/sd-bus/test-bus-peersockaddr.c +++ b/src/libsystemd/sd-bus/test-bus-peersockaddr.c @@ -6,6 +6,7 @@ #include "sd-bus.h" #include "bus-dump.h" +#include "bus-internal.h" #include "bus-util.h"...

Hm. eth0 is not renamed... ``` May 18 14:46:47 H systemd-udevd[725]: eth0: Device ready for processing (SEQNUM=2659, ACTION=add) May 18 14:46:47 H (udev-worker)[786]: eth0: Processing device (SEQNUM=2659, ACTION=add) May 18...

Or kernel bug?

From https://github.com/systemd/systemd/actions/runs/9150396388/job/25155113839?pr=32928 for PR #32928, this is a regression in kernel 6.9. Fedora 40 (kernel 6.8) : PASS Fedora rawhide (kernel 6.9) : FAIL

Ah, caused by https://github.com/torvalds/linux/commit/8debcf5832c3e8a6baaea27c75ad8a6ba5077beb. systemd-udevd checks if a network interface is a stacked netdev or not with `iflink` sysattr. If it is `0`, then assume it is a stacked device....

Yeah, it should return `ifindex` rather than `0` if there is no peer.