Re: [Qemu-devel] [PATCH v3 2/8] pc: acpi: decribe bridge device as not hotpluggable

[Igor Mammedov]

On Mon, 19 Jan 2015 14:46:38 +0200
"Michael S. Tsirkin" <address@hidden> wrote:

> On Fri, Dec 19, 2014 at 11:46:58AM +0000, Igor Mammedov wrote:
> > when bridge hotplug is disabled, i.e. for machine
> > types less then 2.0, bridge device was created as
> > hotpluggable by mistake since commit 133a2da (2.1).
> > Fix it by just creating it as a present device as
> > it was done in 1.7.
> > 
> > Signed-off-by: Igor Mammedov <address@hidden>
> 
> I still don't get this. Under 1.7 all devices are hotpluggable
> unless disabled by device class. Bridges were hotpluggable too.
> 
> See 72c194f7e75cb64b2558111cb111adb49fbf4097, function
> acpi_get_hotplug_info.
> 
> So if the idea is to match 1.7, this does not do it.
Yep, I was wrong about 1.7.
Let's see:

1.7 - all PCI devices created as hotpluggable entries

99fd437d - creates bridge as non hotpluggable unconditionally

133a2da4 - creates bridge as
  1 - non hotpluggable if bridge hotplug supported - subtree is creeated
  2 - if bridge hotplug isn't supported, it creates bridge as hotpluggable - 
subtree is not created


so I guess, I should maintain what 133a2da4 does, drop this patch and amend 8/8

> 
> What, then, is the purpose of this patch?
> 
> 
> > ---
> >  hw/i386/acpi-build.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
> > index a4d0c0c..c151fde 100644
> > --- a/hw/i386/acpi-build.c
> > +++ b/hw/i386/acpi-build.c
> > @@ -919,7 +919,7 @@ static void build_pci_bus_end(PCIBus *bus, void 
> > *bus_state)
> >              }
> >          }
> >  
> > -        if (!dc->hotpluggable || bridge_in_acpi) {
> > +        if (!dc->hotpluggable || pc->is_bridge) {
> >              clear_bit(slot, slot_hotplug_enable);
> >          }
> >      }
> > -- 
> > 1.8.3.1
> >