On Wed, Mar 17, 2010 at 03:46:04PM -0600, Rick McNeal wrote:
>
> On 17-Mar-2010, at 3:20 PM, Edward Pilatowicz wrote:
>
> <sniped>
> >>
> >> I'm still thinking I must be doing something wrong in my hba driver or 
> >> with it's configuration, but I don't see how. As you say, the hba driver 
> >> works with the SCSA framework and it's that framework which is invoking 
> >> the probe function in the 'sd' driver which is panic'ing.
> >>
> >
> > it's also possible that the scsi hba device enumeration works
> > differently from other device enumeration and instance assignment
> > doesn't happen until sometime after probe.  (in which case the hvm sd
> > driver won't work in that environment, but as i mentioned before the hvm
> > sd driver was really only designed to work with ide cdrom devices.) i'm
> > not a scsi device enumeration expert.
> >
> > looking at the native sd driver, i see that it doesn't actually use
> > instance numbers in it's probe routine, so it's unlikely that you'll see
> > the same panic with the native driver.
> >
>
> Unless I'm looking at something different the native 'sd' driver DOES use 
> instance numbers in it's probe routine. That code is commented out if 
> XPV_HVM_DRIVER is defined. This is also consistent with the panic I get when 
> I removed the i86hvm/kernel/drv/sd driver.
>
> The kernel now panic's with sd:sd_probe() on the stack. The assert is still 
> in ddi_get_soft_state().
>

oops.  your right.  i was reading that ifndef backwards.  (i was
thinking we only accessed the instance number if XPV_HVM_DRIVER was
defined.)

ed
_______________________________________________
driver-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/driver-discuss

Reply via email to