* [PATCHv1 0/2] xen-pciback: allow device reset to work more often
@ 2014-07-10 13:03 David Vrabel
0 siblings, 0 replies; 4+ messages in thread
From: David Vrabel @ 2014-07-10 13:03 UTC (permalink / raw)
To: linux-pci; +Cc: Bjorn Helgaas, xen-devel, Boris Ostrovsky, David Vrabel
When a PCI device is passed through to a different domain, it must be
reset to ensure it works correctly in the new domain. Usually the
toolstack will request a device reset via the "reset" sysfs file. But
this file is not present if a function reset is not available.
A common use case is a GPU device and its associated audio function.
In the case where these two devices are simultaneously co-assigned to
the same domain, it is safe to reset them with an SBR.
Patch #1 export pci_probe_reset_function() so pciback can know if an
alternate reset mechanism needs to be proved.
Patch #2 adds a "reset" file to devices bound to pciback that performs
the SBR if it is safe to do so.
David
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCHv1 0/2] xen-pciback: allow device reset to work more often
2014-07-10 13:03 David Vrabel
@ 2014-07-10 16:25 ` Bjorn Helgaas
2014-07-10 16:25 ` Bjorn Helgaas
1 sibling, 0 replies; 4+ messages in thread
From: Bjorn Helgaas @ 2014-07-10 16:25 UTC (permalink / raw)
To: David Vrabel
Cc: linux-pci, xen-devel, Konrad Rzeszutek Wilk, Boris Ostrovsky,
alex.williamson
[+cc Alex]
On Thu, Jul 10, 2014 at 7:03 AM, David Vrabel <david.vrabel@citrix.com> wrote:
> When a PCI device is passed through to a different domain, it must be
> reset to ensure it works correctly in the new domain. Usually the
> toolstack will request a device reset via the "reset" sysfs file. But
> this file is not present if a function reset is not available.
>
> A common use case is a GPU device and its associated audio function.
> In the case where these two devices are simultaneously co-assigned to
> the same domain, it is safe to reset them with an SBR.
>
> Patch #1 export pci_probe_reset_function() so pciback can know if an
> alternate reset mechanism needs to be proved.
>
> Patch #2 adds a "reset" file to devices bound to pciback that performs
> the SBR if it is safe to do so.
I haven't looked at your patches yet, but I added Alex because he did
a lot of similar work for vfio.
Bjorn
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCHv1 0/2] xen-pciback: allow device reset to work more often
2014-07-10 13:03 David Vrabel
2014-07-10 16:25 ` Bjorn Helgaas
@ 2014-07-10 16:25 ` Bjorn Helgaas
1 sibling, 0 replies; 4+ messages in thread
From: Bjorn Helgaas @ 2014-07-10 16:25 UTC (permalink / raw)
To: David Vrabel; +Cc: linux-pci, Boris Ostrovsky, alex.williamson, xen-devel
[+cc Alex]
On Thu, Jul 10, 2014 at 7:03 AM, David Vrabel <david.vrabel@citrix.com> wrote:
> When a PCI device is passed through to a different domain, it must be
> reset to ensure it works correctly in the new domain. Usually the
> toolstack will request a device reset via the "reset" sysfs file. But
> this file is not present if a function reset is not available.
>
> A common use case is a GPU device and its associated audio function.
> In the case where these two devices are simultaneously co-assigned to
> the same domain, it is safe to reset them with an SBR.
>
> Patch #1 export pci_probe_reset_function() so pciback can know if an
> alternate reset mechanism needs to be proved.
>
> Patch #2 adds a "reset" file to devices bound to pciback that performs
> the SBR if it is safe to do so.
I haven't looked at your patches yet, but I added Alex because he did
a lot of similar work for vfio.
Bjorn
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCHv1 0/2] xen-pciback: allow device reset to work more often
@ 2014-07-10 13:03 David Vrabel
2014-07-10 16:25 ` Bjorn Helgaas
2014-07-10 16:25 ` Bjorn Helgaas
0 siblings, 2 replies; 4+ messages in thread
From: David Vrabel @ 2014-07-10 13:03 UTC (permalink / raw)
To: linux-pci
Cc: David Vrabel, Bjorn Helgaas, xen-devel, Konrad Rzeszutek Wilk,
Boris Ostrovsky
When a PCI device is passed through to a different domain, it must be
reset to ensure it works correctly in the new domain. Usually the
toolstack will request a device reset via the "reset" sysfs file. But
this file is not present if a function reset is not available.
A common use case is a GPU device and its associated audio function.
In the case where these two devices are simultaneously co-assigned to
the same domain, it is safe to reset them with an SBR.
Patch #1 export pci_probe_reset_function() so pciback can know if an
alternate reset mechanism needs to be proved.
Patch #2 adds a "reset" file to devices bound to pciback that performs
the SBR if it is safe to do so.
David
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2014-07-10 16:25 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-07-10 13:03 [PATCHv1 0/2] xen-pciback: allow device reset to work more often David Vrabel
-- strict thread matches above, loose matches on Subject: below --
2014-07-10 13:03 David Vrabel
2014-07-10 16:25 ` Bjorn Helgaas
2014-07-10 16:25 ` Bjorn Helgaas
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.