[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YNoCDmHglfmN15lf@smile.fi.intel.com>
Date: Mon, 28 Jun 2021 20:08:30 +0300
From: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To: Eric Auger <eric.auger@...hat.com>
Cc: Greg KH <gregkh@...uxfoundation.org>, eric.auger.pro@...il.com,
linux-kernel@...r.kernel.org, mihai.carabas@...cle.com,
arnd@...db.de, pizhenwei@...edance.com, pbonzini@...hat.com,
joe@...ches.com
Subject: Re: [PATCH] misc/pvpanic-pci: Allow automatic loading
On Mon, Jun 28, 2021 at 05:44:13PM +0200, Eric Auger wrote:
> On 6/28/21 4:54 PM, Greg KH wrote:
> > On Mon, Jun 28, 2021 at 04:44:22PM +0200, Eric Auger wrote:
> >> The pvpanic-pci driver does not auto-load and requires manual
> >> modprobe. Let's include a device database using the
> >> MODULE_DEVICE_TABLE macro.
...
> > Is this something that you need? Or is it created by a tool?
>
> the virtual machine monitor (QEMU) exposes the pvpanic-pci device to the
> guest. On guest side the module exists but currently isn't loaded
> automatically. So the driver fails to be probed and does not its job of
> handling guest panic events. We need a SW actor that loads the module
> and I thought this should be handled that way. If not, please could you
> advise?
At least something like about should have been in the commit message.
Otherwise the selling point is unclear.
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists