On Mon, 2008-08-04 at 14:43 -0600, Bjorn Helgaas wrote:
quoted text
> On Sunday 03 August 2008 12:02:12 pm James Bottomley wrote:
> > +static void pci_note_irq_problem(struct pci_dev *pdev, const char *reason)
> > +{
> > +	struct pci_dev *parent = to_pci_dev(pdev->dev.parent);
> > +
> > +	dev_printk(KERN_ERR, &pdev->dev,
> > +		   "Potentially misrouted IRQ (Bridge %s %04x:%04x)\n",
> > +		   parent->dev.bus_id, parent->vendor, parent->device);
> 
> Do you prefer "dev_printk(KERN_ERR, ...)" over "dev_err(...)"?  Easier
> to grep for the former, maybe?  If so, should we deprecate "dev_err()"
> and friends?  When I converted most of the PCI core to use dev_printk(),
> (80ccba1186d48f ...) I used dev_err(), but I don't really care one way
> or the other.
> 
> Maybe use pci_name(parent)?
> 
> I tried to standardize the PCI core on "[%04x/%04x]" for vendor/device ID.

To be honest I'm not really interested too much in the various API
preferences ... they can be fixed up later by the people who care.

What I am interested in is whether we can get this interface to give
sufficient information to blacklist the offending motherboard if we can
identify it as the source of the problem.  Since the usual culprit is
the bridge, that's why I'm doing parent vendor/device pairs.  However,
better suggestions for the information that should be displayed would be
gratefully received.

James


--
unsubscribe notice
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/