Cui, Dexuan
2009-Nov-02 10:26 UTC
[Xen-devel] [PATCH] xend: passthrough: revert some parts of changeset 20348: 9ead82c46efd
Changeset 20348 breaks the device assignment (at least) in the non-stubdomain case, e.g., 1) we can't assign more than 1 device to hvm guest statically: 'xm create' complains "Error: Device 0 (pci) is already connected"; 2) after we assign 1 device to hvm guest statically and we 'xm pci-detach' the device from guest later, we can find 'xm pci-list' still shows the device is assigned to the guest, and we can't 'xm pci-attach' the device to the guest any longer. The patch reverts some parts of changeset 20348 to fix the above issues. Hope this doesn't offend the passthrough in the stub-domain case. Stefano, please have a look. Thanks, -- Dexuan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2009-Nov-02 10:49 UTC
[Xen-devel] Re: [PATCH] xend: passthrough: revert some parts of changeset 20348: 9ead82c46efd
On Mon, 2 Nov 2009, Cui, Dexuan wrote:> Changeset 20348 breaks the device assignment (at least) in the > non-stubdomain case, e.g., > 1) we can''t assign more than 1 device to hvm guest statically: ''xm create'' > complains "Error: Device 0 (pci) is already connected"; > 2) after we assign 1 device to hvm guest statically and we ''xm pci-detach'' > the device from guest later, we can find ''xm pci-list'' still shows the device > is assigned to the guest, and we can''t ''xm pci-attach'' the device to the guest > any longer. > > The patch reverts some parts of changeset 20348 to fix the above issues. > > Hope this doesn''t offend the passthrough in the stub-domain case. > Stefano, please have a look.Keir, please wait applying this patch because I am aware of the issues above and I am trying to come up with a fix that would work for stubdoms too. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel