On Tue, 2019-07-09 at 19:16 +0200, Michael Biebl wrote: > Am 09.07.19 um 18:44 schrieb Marc Haber: > > https://bugzilla.kernel.org/show_bug.cgi?id=204099 > > https://bugzilla.kernel.org/show_bug.cgi?id=203769 > > > > identify the kernel commit that broke things, > > See also > https://lore.kernel.org/netdev/58ac6ec1-9255-0e51-981a-195c2b1ac...@mellanox.com/T/ > > but people think this is > > actually a systemd issue: > > > > https://github.com/systemd/systemd/issues/12784 > > I'm not an expert on this matter, so I can't say if networkd is buggy > here and if the changed kernel behaviour is ok. Would appreciate Ben's > input on this.
I don't really understand the detail of what's changed, but the kernel has a strict policy of not breaking user-space (with rare exceptions being made for security fixes). So even if this is a bug in systemd-networkd, the kernel should end up with a workaround for it. Ben. > > (with patch) > > Can you test those patches to verify that they fix your issue? > > > Will I have to live with a patched systemd again during the buster > > cycle, or would that one warrant a systemd update in buster? > > Depends on whether those patches can be sanely backported to v241 or if > they are too invasive. In the end they have to be acked by the SRM. > > Michael > > -- Ben Hutchings The Peter principle: In a hierarchy, every employee tends to rise to their level of incompetence.
signature.asc
Description: This is a digitally signed message part
