On Fri, 2003-03-14 at 06:57, Alan Cox wrote:
> On Thu, 2003-03-13 at 23:25, Doug Ledford wrote:
> > Mark Haverkamp wrote:
> > > Then it sounds like the aacraid driver could set cmd_per_lun to a small
> > > number like one since the real queue depth will be set later in
> > > aac_slave_configure.
> >
> > Yes. And since the driver doesn't support anything other than drives to
> > my knowledge, 1 would be appropriate.
>
> It supports both disks and non disk devices. Disks are mapped onto bus 0
> and are driven via firmware smarts as logical devices, non disks are bus 1
> and there it behaves mostly like a scsi controller
Is aac_slave_configure only called for disk devices? If its called for
all scsi devices, then the queue depth will always be set to something a
lot less than 512. I did some searching through the scsi code and I see
only two places that cmd_per_lun is used. It is used to set the queue
depth in scsi_track_queue_full and scsi_alloc_sdev. So it seems that, if
aac_slave_configure gets called for all scsi devices, that setting
cmd_per_lun in the aacraid scsi host template to 1 would be OK. Does
that make sense or did I miss something?
Thanks,
Mark.
-- Mark Haverkamp <markh@osdl.org>- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Mar 15 2003 - 22:00:40 EST