On 2016/09/26 20:30, Alexander Bluhm wrote: > On Mon, Sep 26, 2016 at 01:24:54PM -0400, David Hill wrote: > > On Mon, Sep 26, 2016 at 01:55:01PM +0100, Stuart Henderson wrote: > > > Seen about a dozen times on a box doing ospfd + ospf6d (for internal > > > routes), bgpd (for default route) + isakmpd, shortly after startup > > > after updating: > > > > > > splassert: sorwakeup: want 5 have 4 > > > Starting stack trace... > > > splassert_check() at splassert_check+0x78 > > > sorwakeup() at sorwakeup+0x27 > > > timeout_run() at timeout_run+0x48 > > > softclock() at softclock+0x14c > > > softintr_dispatch() at softintr_dispatch+0x8b > > > Xsoftclock() at Xsoftclock+0x1f > > > --- interrupt --- > > > end of kernel > > > end trace frame: 0x51, count: 251 > > > 0x8: > > > End of stack trace. > > > > > > > Perhaps this? > > rt_senddesync() is a very good candidate to cause this stack trace. > I would prefer to put the splsoftnet() also around sbappendaddr(). > The goal is to protect all socket buffer functions with a lock. > > ok?
Yes, thanks.