Re: Lockup with -current via SSH
Allen Briggs <[email protected]> Tue, 25 Feb 2003 00:05:05 -0500
| Newsgroups | gmane.os.netbsd.devel.smp,gmane.os.netbsd.current |
|---|---|
| Message-ID | <[email protected]> |
On Tue, Feb 25, 2003 at 11:33:52AM +0900, enami tsugutomo wrote: > > Until...I log in remotely via ssh. Enter username, enter password. Solid > > lockup. No console output, nothing in /var/log. Reset to recover. > > Repeatable every time. > Probably, this is locking against myself in unp_internalize. I think it may be something different. I have run out of time to track this right now, but a kernel with LOCKDEBUG DEBUG DIAGNOSTIC is blowing up in biodone() with locks that are either uninitialized or not equal to SIMPLELOCK_LOCKED or SIMPLELOCK_UNLOCKED. With raidframe configured, I got a "simple_lock: uninitialized lock" on boot right after the "Searching for RAID components..." message. Without raidframe configured, I get a KDASSERT on 'v == SIMPLELOCK_LOCKED || v == SIMPLELOCK_UNLOCKED' Both from the same simple_lock(&bp->b_interlock); at the beginning of biodone(). If no one else has found the source of this by my tomorrow afternoon, I'll look at it some more. -allen -- Allen Briggs [email protected] Wasabi Systems, Inc. http://www.wasabisystems.com/