Re: [BUG](-mm)pci_disable_device function clear bars_enabled element

[Grant Grundler]

On Thu, Jun 01, 2006 at 11:36:26AM -0700, Rajesh Shah wrote:
> On Thu, Jun 01, 2006 at 11:15:59AM -0600, Grant Grundler wrote:
> > +   The device driver needs to call pci_request_region() to make sure
> > +no other device is already using the same resource. The driver is expected
> > +to determine MMIO and IO Port resource availability _before_ calling
> > +pci_enable_device().  Conversely, drivers should call pci_release_region()
> > +_after_ calling pci_disable_device(). The idea is to prevent two devices
> > +colliding on the same address range.
> > +
> A quick look in the drivers directory shows that _lots_ of drivers
> violate this rule. In fact, I suspect Kaneshige-san made the original
> incorrect assumption since there were so many drivers out there
> which do it in the wrong order.

That's ok. It's not a big deal normally.
It's just when BIOS is broken and assigns overlapping ranges to
multiple devices or mis-routes MMIO address space that we need
this to work right.  The vast majority of BIOS's do get it right
and that's why I'm not terribly worried.

If there is consensus the drivers are wrong, then it's pretty
easy to fix and we don't have to panic.

Do you agree with the change in the text?

thanks,
grant
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/