Re: [PATCH AUTOSEL 6.6 52/75] PCI: Mark LSI FW643 to avoid bus reset

[Sasha Levin]

On Fri, Mar 29, 2024 at 10:17:02AM -0500, Bjorn Helgaas wrote:
> On Fri, Mar 29, 2024 at 08:42:33AM -0400, Sasha Levin wrote:
> > From: Edmund Raile <edmund.raile@xxxxxxxxx>
> >
> > [ Upstream commit 29a43dc130ce65d365a8ea9e1cc4bc51005a353e ]
> >
> > Apparently the LSI / Agere FW643 can't recover after a Secondary Bus Reset
> > and requires a power-off or suspend/resume and rescan.
> >
> > VFIO resets a device before assigning it to a VM, and the FW643 doesn't
> > support any other reset methods, so this problem prevented assignment of
> > FW643 to VMs.
> >
> > Prevent use of Secondary Bus Reset for this device.
> >
> > With this change, the FW643 can be assigned to VMs with VFIO.  Note that it
> > will not be reset, resulting in leaking state between VMs and host.
> >
> > Link: https://lore.kernel.org/r/20240227131401.17913-1-edmund.raile@xxxxxxxxx
> > Signed-off-by: Edmund Raile <edmund.raile@xxxxxxxxx>
> > [bhelgaas: commit log, comment]
> > Signed-off-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx>
> > Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
>
> We're about to revert this upstream, so I wouldn't backport this to
> any stable trees:
>
> https://lore.kernel.org/r/20240328212302.1582483-1-helgaas@xxxxxxxxxx

I'll drop it, thanks!

--
Thanks,
Sasha