On Mon, Oct 21, 2013 at 03:58:23PM -0500, Eric Sandeen wrote: > On 10/21/13 1:32 PM, Geyslan G. Bem wrote: > > This patch puts a 'break' in the true branch, avoiding the 'icptr->ic_next' > > dereferencing. > > Reviewed-by: Eric Sandeen <sandeen@xxxxxxxxxx> Actually, NACK. > Hm, yeah - cmn_err(CE_PANIC, " " ); used to BUG_ON, but the newer > xfs_emerg() doesn't. > > Dave, was that intentional? Of course it was. ;) xfs_emerg() is only called from the debug code in xlog_verify_iclog(), xlog_verify_tail_lsn and assfail(). In the case of assfail(), it has it's own BUG() call, so it does everything just fine. In the case of xlog_verify_iclog() when icptr is NULL, it will panic immediately after the message is printed, just like the old code. i.e. this patch isn't fixing anything we need fixed. > I wonder if there are more spots after xfs_emerg()'s which aren't > defensive, because the code used to just panic there. As for the rest of the calls in xlog_verify_iclog, they are checking things that aren't immediately fatal, but indication that iclog corruption has already occurred. It's debug code, so we could add "panic immediately" code, but personally I'd prefer to see the error message being printed and then have it continue like a production system would so that we can see the types of crashes normal kernels will see as a result of iclog memory corruption.... As for xlog_verify_tail_lsn(), that's an important informational message indicating we might be leaking log space. It's not immediately fatal, but if we see it and then have a log space hang... So, really, none of the callers really need xfs_emerg to panic like CE_PANIC used to. The one case where it might be useful (i.e this patch) we panic immediately anyway.... Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
