Re: [Xen-devel] [PATCH] hotplug/Linux: update to new ip command syntax.

[George Dunlap <George.Dunlap@xxxxxxxxxxxxx>]

On Wed, Nov 20, 2013 at 2:53 PM, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:
> On Wed, 2013-11-20 at 14:37 +0000, Ian Jackson wrote:
>> Ian Campbell writes ("Re: [Xen-devel] [PATCH] hotplug/Linux: update to new 
>> ip command syntax."):
>> > On Mon, 2013-08-19 at 18:01 +0100, Ian Campbell wrote:
>> > > On Mon, 2013-08-19 at 17:44 +0100, Ian Jackson wrote:
>> > > > Do we know whether this new syntax is backwards-compatible ?
>> > >
>> > > Good question. Squeeze's ip(8) doesn't mention dev for "ip link show" in
>> > > the synopsis by virtue of not defining DEVICE anywhere but it does show
>> > > the use of dev later on and the tool accepts it. Likewise for "ip link
>> > > set". Given this, as Mike says, makes it impossible to name a device
>> > > "dev" it seems likely to be a documentation bug.
>> > >
>> > > "ip addr flush" is correctly documented as needing the dev (by saying
>> > > "dev STRING" and not "DEVICE" , I didn't try without but I assume it
>> > > works since our scripts do that right now.
>> > >
>> > > I wasn't easily able to come up with an older machine.
>> >
>> > Ian, was that a satisfactory explanation?
>>
>> Yes, it was.  Sorry.
>>
>> Acked-by: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>
>
> Thanks.
>
> Now, WRT 4.4...
>
> With reference to bugs.debian.org/cgi-bin/bugreport.cgi?bug=705659 the
> patch allows the creation of a vif named "dev". Which although I'm sure
> useful it is easily worked around. But on the other hand we are early in
> the freeze and this could easily be considered a bug fix.
>
> At this stage I'm leaning towards say lets take it.

Should I say 'no' once just to let people know who's in charge? :-)

I might be inclined to do so actually, just on principle; except that
the best way to get this actually tested will be the upcoming test
days, when (hopefully) it will be tested on all the major distros.  If
we put it off until 4.5, I doubt it will get much more testing than it
would right now.

One thing I'm not happy with the patch though -- it doesn't have a
proper description of the problem and what the patch is actually doing
in the changeset itself.  Links to supplemental information are fine,
I think, but we have to assume that 1) people will be browsing the
source code offline, and 2) links will disappear; so the critical
information needs to be included.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel