Linux Virtualization - Sep 2017 - [PATCH] KVM: s390: Disable CONFIG_S390_GUEST_OLD_TRANSPORT by default

On 26.09.2017 12:47, Heiko Carstens wrote:
> On Tue, Sep 26, 2017 at 12:41:41PM +0200, Christian Borntraeger wrote:
>>
>>
>> On 09/26/2017 12:40 PM, Heiko Carstens wrote:
>>> On Mon, Sep 25, 2017 at 08:37:36PM +0200, Christian Borntraeger
wrote:
>>>>
>>>> On 09/25/2017 07:54 PM, Halil Pasic wrote:
>>>>>
>>>>>
>>>>> On 09/25/2017 04:45 PM, Thomas Huth wrote:
>>>>>> There is no recent user space application available
anymore which still
>>>>>> supports this old virtio transport, so let's
disable this by default.
>>>>>>
>>>>>> Signed-off-by: Thomas Huth <thuth at redhat.com>
>>>>>
>>>>> I don't have any objections, but there may be something
I'm not aware of.
>>>>> Let's see what Connie says. From my side it's ack.
>>>>>
>>>>> Via whom is this supposed to go in? Looking at the
MAINTAINERS, I would
>>>>> say Martin or Heiko but I don't see them among the
recipients.
>>>>
>>>> FWIW as the original author of that transport
>>>> Acked-by: Christian Borntraeger <borntraeger at
de.ibm.com>
>>>>
>>>> I can pick this up for Martins/Heikos tree if you want.
>>>
>>> When will this code be removed?
>>>
>>> When the config option was initially added Conny said it should
survive
>>> "probably two kernel releases or so, depending on
feedback".
>>> It was merged for v4.8. Now we are five kernel releases later...
>>
>> Lets wait for one release and then get rid of it?
> 
> So it's going to be removed with the next merge window.
> Where is the patch? ;)
Hmm, so far the code was always enabled by default, so in the unlikely
case that somebody is still using it, they might not have noticed that
this is marked as deprecated. So maybe it's better to keep the code for
two more released, but disable it by default, so that people have time
to realize that it is going away? ... just my 0.02 ? ... but if you
prefer, I can also write a patch that removes it immediately instead.

 Thomas

On Tue, Sep 26, 2017 at 12:57:26PM +0200, Thomas Huth
wrote:
> On 26.09.2017 12:47, Heiko Carstens wrote:
> > So it's going to be removed with the next merge window.
> > Where is the patch? ;)
> 
> Hmm, so far the code was always enabled by default, so in the unlikely
> case that somebody is still using it, they might not have noticed that
> this is marked as deprecated. So maybe it's better to keep the code for
> two more released, but disable it by default, so that people have time
> to realize that it is going away? ... just my 0.02 ? ... but if you
> prefer, I can also write a patch that removes it immediately instead.
Switching the default from yes to no won't prevent anybody from selecting
it again. So I don't see any value in waiting even longer.

Besides the original commit already printed a message to the console that
the transport is deprecated.
See commit 3b2fbb3f06ef ("virtio/s390: deprecate old transport").

On Tue, 26 Sep 2017 13:04:03 +0200
Heiko Carstens <heiko.carstens at de.ibm.com> wrote:
> On Tue, Sep 26, 2017 at 12:57:26PM +0200, Thomas Huth wrote:
> > On 26.09.2017 12:47, Heiko Carstens wrote:  
> > > So it's going to be removed with the next merge window.
> > > Where is the patch? ;)  
> > 
> > Hmm, so far the code was always enabled by default, so in the unlikely
> > case that somebody is still using it, they might not have noticed that
> > this is marked as deprecated. So maybe it's better to keep the
code for
> > two more released, but disable it by default, so that people have time
> > to realize that it is going away? ... just my 0.02 ? ... but if you
> > prefer, I can also write a patch that removes it immediately instead.
> 
> Switching the default from yes to no won't prevent anybody from
selecting
> it again. So I don't see any value in waiting even longer.
> 
> Besides the original commit already printed a message to the console that
> the transport is deprecated.
> See commit 3b2fbb3f06ef ("virtio/s390: deprecate old transport").
> 
Would ack a removal as well.