Subject: Re: Processes getting stuck in disk wait on current currents
To: Wolfgang Rupprecht <wolfgang@wsrcc.com>
From: Chuck Silvers <chuq@chuq.com>
List: current-users
Date: 02/18/2001 19:50:51
this kinda looks like the IDE driver is losing the i/o...
even if the hardware isn't responding, I would still think the
driver should time out the request eventually.

manuel, got any ideas?

-Chuck


On Sun, Feb 18, 2001 at 07:46:43PM -0800, Wolfgang Rupprecht wrote:
> 
> Chuck Silvers writes:
> > I just copied 500 MB from NFS to a local filesystem with "tar | (cd && tar)"
> > on an x86 and it worked fine.
> 
> It doesn't seem to happen on my other 3 machines either, and they all
> run the same fs code and mostly the same kernel.
> 
> > can you get a crashdump while some processes are hung in biowait?
> 
> I tried to initiate a "sync" from DDB and got a panic.  (transcribing
> by hand) the panic was:
> 
> 	syncing disks... 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 ... giving up
> 	dumping to dev 0,1 offset 153,331
> 	dump panic: wddump: polled command has been queued.
> 	stopped at cpu_debugger+0x4 leave
> 
> Hmm.  Seems like when this happens the disk is really OTL.
> 
> -wolfgang
> -- 
> Wolfgang Rupprecht    <wolfgang@wsrcc.com>     http://www.wsrcc.com/wolfgang/
> Coming soon: GPS mapping tools for Open Systems. http://www.gnomad-mapping.com/