On Mon, 24 Feb 1997, Bill Sommerfeld wrote:

> Anyone else seeing severe VM system hangs with the today's bits?

You bet!  My 386 was pretty unstable towards the end of the week (I
got lots of "unreferenced object still paging" panics), and then I
built a new kernel after the last VM mods were checked in on Sunday,
and this kernel is unrunnable on my system.

It should be added that this system is a good indicator of problems
with the VM subsystem; once it's booted and I'm logged in, it's got
about five times as much swap allocated as it's got physical RAM.

> symptom is: lockup which starts when starting memory hogs (larger X
> clients, emacs, ..).  A bunch of processes appear to be hung in
> thread_sleep_msg inside vm_fault(); more specifically, vm_fault+0x82e;
> it appears to be in the middle of the second of the two occurrances
> of:
> 
> 	UNLOCK_AND_DEALLOCATE;
> 	VM_WAIT;
> 	goto RetryFault;
> 
> .. which appears to be "wait for pageout daemon to clean pages" logic.

Ditto -- except that I've got only 8MB of RAM, so my system locks up
before it manages to go multi-user.  If I comment out absolutely all
my additions to /etc/rc.local, it'll just barely get through that, but
hang when it tries to spawn getty processes.

I get lots of "thrd_sleep", and the pagedaemon in "vospgw", according
to the kernel debugger when I break into it with ctrl-alt-shift-esc.

> My kernel sources have some local mods so they aren't pure -current,
> but nothing I've changed touches the VM system...

Same here.  Slight tweaks, nothing that should be significant.

-tih
-- 
Popularity is the hallmark of mediocrity.  --Niles Crane, "Frasier"