libata/ipr/powerpc: regression between 2.6.39-rc4 and 2.6.39-rc5

[Nishanth Aravamudan]

On 15.06.2011 [15:02:18 -0500], Brian King wrote:
> On 06/15/2011 02:17 PM, Nishanth Aravamudan wrote:
> > From what I can tell, the only place that explicitly clears the FROZEN
> > flag is the error-handling code via ata_eh_thaw_port().
> > 
> > So I thought either we're not invoking the error-handler at probe time
> > correctly to kick the port or perhaps the SAS code is not clearing the
> > flag?
> > 
> > I tried the following patch:
> > 
> > diff --git a/drivers/ata/libata-scsi.c b/drivers/ata/libata-scsi.c
> > index d51f979..abd0e0b 100644
> > --- a/drivers/ata/libata-scsi.c
> > +++ b/drivers/ata/libata-scsi.c
> > @@ -3797,6 +3797,7 @@ EXPORT_SYMBOL_GPL(ata_sas_port_alloc);
> >   */
> >  int ata_sas_port_start(struct ata_port *ap)
> >  {
> > +       ap->pflags &= ~ATA_PFLAG_FROZEN;
> >         return 0;
> >  }
> >  EXPORT_SYMBOL_GPL(ata_sas_port_start);
> > 
> > and the CD-ROM drive works, but I have no idea if it's the right thing
> > to do. I chose this particular change, FWIW, because we call
> > ata_sas_port_start before we probe in ata_sas_port_init and it seems
> > like we need to mark the port as not frozen before we init it? Perhaps
> > that should really be a call to a thaw function, not sure. Let's just
> > say the ATA/SAS/SCSI interactions are a bit hard to follow at first :)
> 
> That looks like the right thing to do. For ipr's usage of
> libata, we don't have the concept of a port frozen state, so this flag
> should really never get set. The alternate way to fix this would be to
> only set ATA_PFLAG_FROZEN in ata_port_alloc if ap->ops->error_handler
> is not NULL.

It seemed like ipr is as you say, but I wasn't sure if it was
appropriate to make the change above in the common libata-scis code or
not. I don't want to break some other device on accident.

Also, I tried your suggestion, but I don't think that can happen in
ata_port_alloc? ata_port_alloc is allocated ap itself, and it seems like
ap->ops typically gets set only after ata_port_alloc returns?

Thanks,
Nish

-- 
Nishanth Aravamudan <nacc at us.ibm.com>
IBM Linux Technology Center