qemu-stable
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-stable] stable patches


From: Michael Roth
Subject: Re: [Qemu-stable] stable patches
Date: Tue, 30 Jul 2013 08:59:32 -0500
User-agent: alot/0.3.4

Quoting Michael S. Tsirkin (2013-07-28 01:49:45)
> On Thu, Jul 25, 2013 at 08:08:38PM -0500, Michael Roth wrote:
> > Quoting Michael S. Tsirkin (2013-07-25 17:22:37)
> > > What's the way to get patches into stable branch
> > > if not Cc'd there originally?
> > 
> > Replying to the patch submission with "CC'ing stable"
> > or something along that line makes them easier to track, but
> > if you want to collect a number of them into a digest to avoid
> > flooding qemu-devel, or just don't want to bump an old thread,
> > an email like this is fine
> > 
> > > 
> > > A couple of commits that are IMO needed there:
> > > 
> > >         e78e9ae4a933504d383703870d491615b0261801
> > 
> > Thanks
> > 
> > >         644c98587d4ccc09e7592e1688e4e7fa363c5a75
> > 
> > Wouldn't this one break migration compatibility from 1.5.3 -> 1.5.2
> > unless you explicitly add an option to disable the feature for
> > a 1.5.3 invocation?
> > 
> > The same command-line invocation should maintain migration
> > compatibility across stable releases to support things like
> > rolling back from a 'live update' or migration to updated
> > hosts.
> 
> This chunk makes it off by default, so it should be fine:

I believe that only applies to 1.4 and earlier. For 1.5 and newer it gets
switched on by default via:

+        DEFINE_PROP_BIT("ctrl_guest_offloads", _state, _field, 
VIRTIO_NET_F_CTRL_GUEST_OFFLOADS, true), \

If we applied to stable we'd want to default to false to maintain machine
compatibility, but if that's an option it makes this seem more like a feature
than a bug fix.

If it does fix a bug and really should be applied for stable we'd probably need
a patch specifically for stable that could somehow maintain migration
compatibility, but even then you're changing the guest hardware fingerprint so
I'd be hesitant to pull it in unless it's fairly serious.

> 
> diff --git a/include/hw/i386/pc.h b/include/hw/i386/pc.h
> index 2bd7090..93d8357 100644
> --- a/include/hw/i386/pc.h
> +++ b/include/hw/i386/pc.h
> @@ -217,6 +217,10 @@ int e820_add_entry(uint64_t, uint64_t, uint32_t);
>              .property = "vectors",\
>              /* DEV_NVECTORS_UNSPECIFIED as a uint32_t string */\
>              .value    = stringify(0xFFFFFFFF),\
> +        },{ \
> +            .driver   = "virtio-net-pci", \
> +            .property = "ctrl_guest_offloads", \
> +            .value    = "off", \
>          },{\
>              .driver   = "e1000",\
>              .property = "romfile",\
> 
> 
> 
> > > 
> > > -- 
> > > MST



reply via email to

[Prev in Thread] Current Thread [Next in Thread]