Thank you for long discussion.
On Thu, 15 Jan 2009 14:25:14 +0800
Qing He <qing.he@xxxxxxxxx> wrote:
> On Thu, 2009-01-15 at 10:35 +0800, Shohei Fujiwara wrote:
> > I think MSI-INTx translation can be enhanced.
> >
> > In case of using MSI enable bit as mask, guest OS need to set interrupt
> > disable bit to prevent device from asserting INTx. Actually, linux
> > 2.6.25 set interrupt disable bit in
> > drivers/pci/msi.c:msi_capability_init. If interrupt disable bit is 1,
> > we can assume guest OS might use MSI enable bit as mask. If interrupt
> > disable bit is 0 and MSI enable bit is 0, we can assume guest OS uses
> > INTx.
>
> This should be better, but cannot garauntee it. Since INTx disable bit
> was added in PCI 2.3. Early guests may exist not using it. Also, some
> devices have problems, that setting INTx disable bit also completely
> disables MSI from signalling, so OS may not touch INTx disable bit of
> these devices.
I think we can guarantee it, even if OS doesn't touch INTx disable bit.
When OS set 1 to MSI/MSI-X enable bit, qemu will disable MSI-INTx
translation.
When OS set 0 to MSI/MSI-X enable bit, qemu will re-enable MSI-INTx
translation.
> Theoretically, MSI-enable-as-masking is not correct and should not be
> used in the first place. If that were true it would be much easier for
> us. However, the existing (and widely used) guest complicates this matter...
I agree with you that we need make MSI-enable-as-masking work on HVM
domain.
> > I think we can archive both of good performance and not-using INTx, if
> > we control MSI-INTx translation as follows.
> >
> > MSI/MSI-X | Interrupt | MSI-INTx | Note
> > enable bit | disable bit | translation |
> > -----------+-------------+-------------+------------------------------
> > 0 | 0 | enable | Re-enabling MSI-INTx
> > | | | translation is needed.
> > -----------+-------------+-------------+------------------------------
> > 0 | 1 | disable | Guest OS might use MSI enable
> > | | | bit as mask.
> > -----------+-------------+-------------+------------------------------
> > 1 | 0 | disable | Guest OS use MSI/MSI-X.
> > -----------+-------------+-------------+------------------------------
> > 1 | 1 | disable | Guest OS use MSI/MSI-X.
> >
> > It is great if we can guarantee INTx interrupt isn't used.
> > We don't need to consider sharing machine gsi. We don't need to consider
> > hot-pluging I/O APIC. In long term, It will be possible even to to remove
> > I/O APIC from real machine.
>
> I am not against it, it does look more symmetric, though in the cost of
> a little overhead (4 hypercalls for a mask operation) in the worst
> (but rare) case. I'll ack the patch if you add it now, or you can wait
> for me to add it at some later time.
Currently I'm trying to enabling PCI pass-through with stub domain.
I'm happy if you can add it.
Thanks,
--
Shohei Fujiwara
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home