On 31/03/20(Tue) 23:16, Vitaliy Makkoveev wrote: > On Tue, Mar 31, 2020 at 06:15:46PM +0200, Martin Pieuchot wrote: > > [...] > > Well better fix npppd(8), no? Not crashing the kernel is priority 1. > I made patch for npppd(8) too. I include it to this email below, without > starting new thread, ok? If ioctl(PIPEXASESSION) failed and error != > ENXIO it means that pipex is enabled and session creation failed so down > this connection.
Thanks, I committed the kernel part. I'm waiting to see if other devs want to comment on the daemon part. > > Then if the daemon has a bug, should the kernel work around it? > In most common cases should :( That's an opinion. There's no true or false answer. Working around it has obvious advantages but it doesn't make us fix existing bug and instead force us to maintain the work around. It is argued that the "failing hard" model, as it is practised in OpenBSD software development, has the advantage of resulting in simpler code because every layer is responsible for handling errors and doesn't pile workaround. This bug is a nice example. Thanks for the report! If you could submit your refactoring in a new thread, I'd love to look at it.