Re: [PATCH net] selftests: bonding: feed link monitor to deflake prio_test

[Hangbin Liu]

On Tue, Nov 10, 2026 at 11:53:52AM +0000, Bjoern Doebel wrote:
> Hi Hangbin,
> 
> On Thu, Nov 05, 2026 at 10:19:08AM +0800, Hangbin Liu wrote:
> > If the kill fails for some reason, will we wait here forever and hang the
> > entire test?
> 
> No. bash keeps a terminated child's exit status until it is reaped, so if
> the ping already exited the kill fails but the wait returns immediately.
> Only a genuinely unkillable ping would block, independent of kill. The
> idiomatic form is kill_process() from net/lib.sh, which short-circuits
> the wait when the kill fails.

Yes, that's what I mean. We need to avoid that the kill failed and the
process still running.

Thanks
Hangbin