Re: [PATCH v2 2/4] x86/MSI-X: access MSI-X table only after having enabled MSI-X

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Thu, Apr 16, 2015 at 08:43:30AM +0100, Jan Beulich wrote:
> >>> On 15.04.15 at 19:41, <konrad.wilk@oracle.com> wrote:
> > On Mon, Apr 13, 2015 at 10:05:14AM +0100, Jan Beulich wrote:
> >> You mentioning XSA-120 and its addendum - are these requirements
> >> for the problem to be seen? I admit I may have tested a PV guest
> >> only with an SR-IOV VF (and only a HVM guest also with an "ordinary"
> >> device), but I'd like to be clear about the validity of the connection.
> > 
> > No. I just tried with v4.0-rc5 (and then also v4.0) and just 
> > using SR-IOV to make this simpler.
> 
> Good.
> 
> > With staging  + two of your patches:
> > a10cc68 TODO: drop //temp-s
> > 1b8721c x86/MSI-X: be more careful during teardown
> > 
> > When trying to enable SR-IOV I get this error:
> 
> Okay, this definitely works for me, albeit I know I had to do
> adjustments to avoid running into the (debug) warning you've
> hit. But (looking at the call stack) it surely would be a mistake to
> set up an MSI-X IRQ on the device without first enabling MSI-X
> on the it (i.e. the error returned could be considered legitimate).
> While we may want the hypervisor to cope with this (by enabling
> MSI-X on this path, which I'd have to add to that patch), is this
> hypervisor change perhaps uncovering a pv-ops kernel issue (in
> that other than what drivers/pci/msi.c does as of the commit
> mentioned in the description of that second patch some Xen-
> specific path fails to enable MSI-X before setting up any of the
> entries)?

Everything is possible :-)

What kernel are you using? Or better yet - what branch/tree could
one find it at?
> 
> Jan
>