Re: [PATCH net] kcm: do not attach PF_KCM sockets to avoid deadlock

2017-08-30 Thread David Miller
From: Eric Dumazet Date: Wed, 30 Aug 2017 09:29:31 -0700 > From: Eric Dumazet > > syzkaller had no problem to trigger a deadlock, attaching a KCM socket > to another one (or itself). (original syzkaller report was a very > confusing lockdep splat during a sendmsg()) > > It seems KCM claims to

Re: [PATCH net] kcm: do not attach PF_KCM sockets to avoid deadlock

2017-08-30 Thread Tom Herbert
On Wed, Aug 30, 2017 at 9:29 AM, Eric Dumazet wrote: > From: Eric Dumazet > > syzkaller had no problem to trigger a deadlock, attaching a KCM socket > to another one (or itself). (original syzkaller report was a very > confusing lockdep splat during a sendmsg()) > > It seems KCM claims to only su

[PATCH net] kcm: do not attach PF_KCM sockets to avoid deadlock

2017-08-30 Thread Eric Dumazet
From: Eric Dumazet syzkaller had no problem to trigger a deadlock, attaching a KCM socket to another one (or itself). (original syzkaller report was a very confusing lockdep splat during a sendmsg()) It seems KCM claims to only support TCP, but no enforcement is done, so we might need to add add