[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] [GIT PULL] pv/pcifront-2.6.32 + pv/pciback-2.6.32 = [INTx/MSI/MSI-X work now]



Hey Jermey,

I finally got an MSI-X capable machine and gotten the onboard Intel 1G
card to do MSI-X, so I was able to test the Xen PCI front and Xen PCI
backend using that.

There were two tiny bugfixes in the Xen PCI frontend driver:
pv/pcifront-2.6.32:

Konrad Rzeszutek Wilk (2):
      pci-xen: Iterate over all MSI-X interrupts, not just one.
      xen-pcifront: MSI/MSI-X vector values were copied to wrong
location.

And one change in the back-end to make it easier to spot what is
happening:

Konrad Rzeszutek Wilk (1):
      xen-pciback: Add debug statements for the MSI/MSI-X configuration
module.

After merging them in your xen/next I now have normal interrupts, MSI,
and MSI-X interrupts delievered to my guests (which are based on
xen/next + pv/pcifront-2.6.32 latest).


Bugs:
- On my own branches (pv/merge.2.6.<blah>) I can only get the normal
interrupts to work. Note, the pv/merge.2.6.<blah> are branches with
the Xen-SWIOTLB and Xen PCI Frontend and cherrypicks from various trees
to make it work. Passing in MSI and MSI-X devices makes the guest hang
so I think I am missing some patch from your trees. Haven't tracked this
down yet.

- There were reports that the PCI front in older kernels (say, SLES10)
isn't working anymore. Need to track that down too.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.