In your message on you write:
> Andrew,
>
> I cannot help on the maintainer, but someone last week pointed me to the
> sistina site that had a work around you can try for this problem.
> http://www.sistina.com/gfs/howtos/fibrechannel_howto/node7.html
>
Mike,
We did some code reading and compared how the vendor's version of the
qlogic driver handles determining if the SCSI command queue is full
versus how it is done in the kernel version. We independantly came to
the conclusion that increasing the queue size would probably solve our
problem (The vendor driver uses a queue size of 128 while the vendor
driver uses 1024) as pointed out in the Sistina site. We also think
that the kernel driver is not accounting for scatter/gather data
segments when determing if the the queue is full. If the queue is big
enough, you may never run into this problem.
Thanks for the pointer.
> I personally run the driver from the qlogic site which seems to have
> worked better for me YMMV.
>
We have heard that the vendor driver doesn't handle some error
conditions as well. For example, when a retry is required after LIP
occurs.
Andrew
> -Mike
> Andrew Patterson [[EMAIL PROTECTED]] wrote:
> >
> > Anybody know who is maintaining the Qlogic FC SCSI driver for Linux 2.4?
> > We get the following error when run under load:
> >
> >
> > qlogicfc0: no handle slots, this should not happen.
> > hostdata-> queued is 2b, in_prt : 07
> >
> > It only seems to happen when the VM system starts thrashing, but from
> > our code reading is (as the message says) should never happen.
> >
> > Andrew Patterson
> >
> >
> >
> >
> > -
> > To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> > the body of a message to [EMAIL PROTECTED]
>
> --
> Michael Anderson
> [EMAIL PROTECTED]
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
