bug in drivers/pci/hotplug/ibmphp_pci.c

bug in drivers/pci/hotplug/ibmphp_pci.c

[Jordan, William P]

I noticed what appears to be a cut/paste error in
drivers/pci/hotplug/ibmphp_pci.c:

***************
*** 969,975 ****
  			debug ("io 32\n");
  			need_io_upper = TRUE;
  		}
! 		if ((io_base & PCI_PREF_RANGE_TYPE_MASK) ==
PCI_PREF_RANGE_TYPE_64) {
  			debug ("pfmem 64\n");
  			need_pfmem_upper = TRUE;
  		}
--- 969,975 ----
  			debug ("io 32\n");
  			need_io_upper = TRUE;
  		}
! 		if ((pfmem_base & PCI_PREF_RANGE_TYPE_MASK) ==
PCI_PREF_RANGE_TYPE_64) {
  			debug ("pfmem 64\n");
  			need_pfmem_upper = TRUE;
  		}

I've verified that the bug still exists in 2.6.15-rc1

Bill Jordan
Unisys Corporation

Re: bug in drivers/pci/hotplug/ibmphp_pci.c

[Greg KH]

On Fri, Nov 18, 2005 at 09:58:50AM -0500, Jordan, William P wrote:
> 
> I noticed what appears to be a cut/paste error in
> drivers/pci/hotplug/ibmphp_pci.c:

Yes it does look like that.  Does changing this solve a problem that you
have been seeing?

If so, care to resend this, in the format specified in
Documentation/SubmittingPatches so that I can get the change into the
kernel tree?

thanks,

greg k-h

RE: bug in drivers/pci/hotplug/ibmphp_pci.c

[Jordan, William P]

> From: Greg KH [mailto:greg@kroah.com]
> Sent: Friday, November 18, 2005 8:14 PM
> 
> Yes it does look like that.  Does changing this solve a problem that
you
> have been seeing?

Actually, I am not running this code at all. I am just trying to
understand PCI. While investigating how the IO base and IO limit
registers are used, I noticed this instance where it was used
inconsistently. I can resubmit it if you like, but I cannot vouch for
the patch.

Bill Jordan

Re: bug in drivers/pci/hotplug/ibmphp_pci.c

[Greg KH]

On Mon, Nov 21, 2005 at 10:09:36AM -0500, Jordan, William P wrote:
> 
> 
> > From: Greg KH [mailto:greg@kroah.com]
> > Sent: Friday, November 18, 2005 8:14 PM
> > 
> > Yes it does look like that.  Does changing this solve a problem that
> you
> > have been seeing?
> 
> Actually, I am not running this code at all. I am just trying to
> understand PCI. While investigating how the IO base and IO limit
> registers are used, I noticed this instance where it was used
> inconsistently. I can resubmit it if you like, but I cannot vouch for
> the patch.

Yes, please resubmit it in the proper format as per
Documentation/SubmittingPatches.

thanks,

greg k-h