On Tue, Jun 28, 2016 at 02:38:12PM +0200, Andrew Jones wrote:
> On Tue, Jun 28, 2016 at 01:56:53PM +0200, Alexander Gordeev wrote:
> > On Tue, Jun 28, 2016 at 01:32:47PM +0200, Andrew Jones wrote:
> > > > > Hmm.. the only way I can test the early patches without pci-testdev
> > > > > is to dump the data structures, but that doesn't really test anything
> > > > > other than the DT parsing. I'd rather we wait.
> > > >
> > > > It also involves what would be PCI configuration space cycles: PCI bus
> > > > scanning and PCI devices inquiring and parsing. Not only DT.
> > >
> > > But how do I know any of that stuff is working?
> >
> > Patch 10/12:
> >
> > int main(void)
> > {
> > int ret = pci_probe();
> >
> > report("PCI bus probing", ret);
> >
> > if (ret)
> > pci_print();
> >
> > return report_summary();
> > }
>
> Well, if you plug some random PCI devices into kvm-unit-tests,
> and do the pci_print, and then compare with an lspci in a guest
> with the same devices, and get the equivalent output, then I'd
> be pretty convinced, but, even then, what's the point? Without
> the pci-testdev (the only pci device we're currently targeting)
> then none of this series is really necessary.
My point is even without pci-testdev the above is of some value,
because it only misses MMIO/PIO test and (a custom) pci-testdev
protocol - which is not a PCI test per se.
So even without pci-testdev PCI configuration space is checked,
reading/writing to devices' BARs etc. All of those is part of
QEMU PCI framwork, not pci-testdev implementation.
But surely, I am not sure how valuable this test alone is.
> Thanks,
> drew
>
> >
> > > >
> > > > > Thanks,
> > > > > drew
> > --
> > 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
--
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
