Hi Bastian, On Wed, Sep 02, 2026 at 11:25:12PM +0200, Bastian Blank wrote: > Control: tags -1 moreinfo > > On Wed, Sep 02, 2026 at 09:06:03PM +0200, Helmut Grohne wrote: > > ------------[ cut here ]------------ > > memcpy: detected field-spanning write (size 28) of single field > > "_Generic(info, const struct ip_tunnel_info * : ((const void *)((info) + > > 1)), struct ip_tunnel_info * : ((void *)((info) + 1)) )" at > > include/net/ip_tunnels.h:700 (size 0) > > This should be fixed with bb5e62f2d547c4de6d1b144cbce2373a76c33f18 > (in 6.15). Could you please verify that?
I cannot easily verify that and I didn't manage to reproduce it outside our production environment yet. In any case, that commit looks very plausible. It also tells me that the kernel output likely is a red herring. The warning should not be causing a kernel hang - especially when the kernel manages to persist that output to the system journal on disk. The nftables autopkgtest likely does more things and might be crashing the kernel via other means. Given that we also run the tests on amd64, this likely requires an arm64 kvm environment for reproducing. Unless there is more information at a later point, I suggest that you close this report whenever the mentioned commit is included. Helmut
