Re: [Qemu-devel] [PATCH v3 3/4] virtio-blk-test.c: add hotplug subtest

[Amos Kong]

On Fri, Jun 06, 2014 at 08:05:15AM +0000, Gonglei (Arei) wrote:
> > -----Original Message-----
> > From: Amos Kong [mailto:address@hidden
> > Sent: Friday, June 06, 2014 1:29 PM
> > To: Gonglei (Arei)
> > Cc: address@hidden; address@hidden; address@hidden
> > Subject: Re: [PATCH v3 3/4] virtio-blk-test.c: add hotplug subtest
> > 
> > On Fri, Jun 06, 2014 at 01:28:59AM +0000, Gonglei (Arei) wrote:
> > > Hi, Amos
> > >
> > > > -----Original Message-----
> > > > From: Amos Kong [mailto:address@hidden
> > > > Sent: Friday, June 06, 2014 9:12 AM
> > > > To: address@hidden
> > > > Cc: address@hidden; address@hidden; Gonglei (Arei)
> > > > Subject: [PATCH v3 3/4] virtio-blk-test.c: add hotplug subtest
> > > >
> > > > This patch adds a new subtest, it hotplugs 29 * 8 = 232 virtio-blk
> > > > devices to guest, and try to hot-unplug them.
> > > >
> > > > Note: the hot-unplug can't work without cooperation of guest OS.
> > 
> > 
> > > > +static void test_blk_hotplug(void)
> > > > +{
> > > > +    int i, j;
> > > > +
> > > > +    /* start with no network/block device, slots 3~0x1f are free */
> > > > +    qtest_start("-net none");
> > > > +
> > > > +    for (i = 3; i <= 0x1f; i++) {
> > > > +        for (j = 7; j >= 0; j--) {
> > > > +            qmp_exec_hmp_cmd("OK\r\n",
> > > > +                             "drive_add 0
> > > > if=none,file=/dev/null,id=drv-%x.%x",
> > > > +                             i, j);
> > > > +            qmp_exec_hmp_cmd("",
> > > > +                       "device_add
> > > > virtio-blk-pci,id=dev-%x.%x,drive=drv-%x.%x,"
> > > > +                       "addr=0x%x.%x,multifunction=on", i, j, i, j, i,
> > j);
> > > > +        }
> > > > +    }
> > >
> > > Do the pci multifunction devices support hotplug/unplug yet?
> > 
> > Yes.
> > 
> > You can hotplug single functions one by one, guest will identify
> > all the functions when function 0 is added.
> > Hotunplug function 0, all functions in same slot will be removed.
> > 
> Thanks.
> 
> IMHO, if we hot-plug function 0 first, the other functions will be not 
> support to
> hot-plug.

Actually other functions can also be identified by guest, I had a
commit to add this support.

| commit f382a086f3129edc152b8044b69ccc6682e637bb
| Author: Amos Kong <address@hidden>
| Date:   Fri Nov 25 15:03:07 2011 +0800
| 
|     PCI: Can continually add funcs after adding func0


> Similarly, it will not be allowed if we only hot-unplug function 1~7,
> but function 0, right?

Yes. function 0 is always necessary.
 
> Best regards,
> -Gonglei

-- 
                        Amos.