On Fri, Nov 04, 2016 at 05:47:20PM +0100, Andrew Jones wrote:
> On Wed, Oct 26, 2016 at 03:47:10PM +0800, Peter Xu wrote:
> > Let's provide a more general way to scan PCI bars, rather than read the
> > config registers every time.
> >
> > Signed-off-by: Peter Xu <peterx@xxxxxxxxxx>
> > ---
> > lib/pci.c | 14 ++++++++++++++
> > lib/pci.h | 3 +++
> > 2 files changed, 17 insertions(+)
> >
> > diff --git a/lib/pci.c b/lib/pci.c
> > index bb41e0f..95a6581 100644
> > --- a/lib/pci.c
> > +++ b/lib/pci.c
> > @@ -35,6 +35,20 @@ int pci_find_dev(struct pci_dev *pci_dev, uint16_t vendor_id, uint16_t device_id
> > return -1;
> > }
> >
> > +void pci_scan_bars(struct pci_dev *dev)
> > +{
> > + int i = 0;
> > +
> > + for (i = 0; i < PCI_BAR_NUM; i++) {
>
> I don't think we need a PCI_BAR_NUM def, but if you do introduce it
> then please change all uses of '6' to it that should be changed.
> Same goes for PCI_DEVFN_MAX.
Then I will replace corresponding number "6" into PCI_BAR_NUM when
necessary. Same for PCI_DEVFN_MAX.
>
> > + if (!pci_bar_is_valid(dev, i))
> > + continue;
> > + dev->pci_bar[i] = pci_bar_get_addr(dev, i);
> > + printf("PCI: init dev 0x%04x BAR %d [%s] addr 0x%lx\n",
> > + dev->pci_bdf, i, pci_bar_is_memory(dev, i) ?
> > + "MEM" : "IO", dev->pci_bar[i]);
>
> v9 of Alex's series has introduced pci_bar_print. You'll want to
> use that here after modifying it to get the bars from struct.
Ok. Will call pci_bar_print() here.
Thanks,
-- peterx
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
