Date: Sun, 2 Feb 2014 14:00:00 GMT From: Eugene Grosbein <eugen@grosbein.net> To: freebsd-bugs@FreeBSD.org Subject: Re: kern/186362: [panic] _mtx_lock_sleep() misses check for NULL Message-ID: <201402021400.s12E00db013129@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR kern/186362; it has been noted by GNATS. From: Eugene Grosbein <eugen@grosbein.net> To: Andriy Gapon <avg@FreeBSD.org> Cc: bug-followup@FreeBSD.org Subject: Re: kern/186362: [panic] _mtx_lock_sleep() misses check for NULL Date: Sun, 02 Feb 2014 20:54:24 +0700 On 02.02.2014 19:55, Andriy Gapon wrote: > > NULL check would be redundant there as it is already established that the lock > is owned and thus must have an owner. > What seems to be happening that is that the thread is trying to acquire a lock > that has been corrupted somehow. E.g. never initialized or its memory location > overwritten. > > Better diagnostics for that case when INVARIANTS are enabled is warranted though. > > Returning to the main issue, I wonder if the following message are related to > the problem: > > module_register: module mac_portacl already exists! > Module mac_portacl failed to register: 17 > > I seem to recall that in some cases our kernel module loading infrastructure may > refuse to load a duplicate module, bu nevertheless re-resolve a symbol name to > point to a symbol in the duplicate module rather than in the original. Which is > a bug, of course. I've disabled loading of mac_portacl in loader.conf keeping options MAC_PORTACL in my kernel and panic is no more. Do you still need additional INVARIANTS-enabled debug info? Eugene Grosbein
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201402021400.s12E00db013129>