Re: Is: Discussion about doing it in Xen 4.5 or Xen 4.6 Was:Re: [PATCH] tools: remove blktap1

[George Dunlap <george.dunlap@eu.citrix.com>]

On 11/05/2014 06:16 PM, Konrad Rzeszutek Wilk wrote:
> On Wed, Nov 05, 2014 at 09:20:44AM +0000, Ian Campbell wrote:
>> On Tue, 2014-11-04 at 15:00 -0500, Konrad Rzeszutek Wilk wrote:
>>> On Tue, Nov 04, 2014 at 06:45:12PM +0000, Andrew Cooper wrote:
>>>> On 04/11/14 18:42, Konrad Rzeszutek Wilk wrote:
>>>>> On Tue, Nov 04, 2014 at 06:07:19PM +0000, Andrew Cooper wrote:
>>>>>> On 04/11/14 17:32, Konrad Rzeszutek Wilk wrote:
>>>>>>> On Tue, Nov 04, 2014 at 11:52:47AM +0000, Ian Campbell wrote:
>>>>>>>> This was disabled by default in Xen 4.4. Since xend has now been removed from
>>>>>>>> the tree I don't believe anything is using it.
>>>>>>> What about XenServer?
>>>>>> We are most definitely not using it, and haven’t used it in a very long
>>>>>> time. We explicitly nuke BLKTAP1 and BLKTAP2 from the Xen build.
>>>>>>
>>>>>>> And isn't there some blktap3 ?
>>>>>> https://github.com/xapi-project/blktap
>>>>>>
>>>>>>>> We need to pass an explicit CONFIG_BLKTAP1=n to qemu-xen-traditional otherwise
>>>>>>>> it defaults to y and doesn't build.
>>>>>>>>
>>>>>>>> Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
>>>>>>>> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
>>>>>>>> ---
>>>>>>>> I think this has probably missed the boat for 4.5 and there isn't much harm in
>>>>>>>> waiting for 4.6. I'm open to being told otherwise though ;-)
>>>>>>> You really want to be at the top of the commit list with the most deleted
>>>>>>> code, eh?
>>>>>> /me suspects that he is already, but I for one am a fan of pruning dead
>>>>>> code.
>>>>> True, but we did talk about xend removal for quite a while before doing it.
>>>>>
>>>>> However, I believe that the folks that did Remus look to be using it
>>>>> (And they have tons of patches against it).
>>>>>
>>>>>
>>>>>     It is unclear to me whether they:
>>>>> 	- want to be the maintainers of it?
>>>>> 	- want to use blktap3 but haven't yet backported their patches.
>>>> The remus patches are against blktap2, not blktap1.  We currently have
>>>> both in-tree.
>>> <slaps his head>
>>>
>>> OK, so blktap1 - dead. That looks like it could go under the knife now.
>> IMHO yes it could.
>>
>>> blktap2 - waiting for response
>> This should stay, at least for the time being, and certainly for 4.5.
>>
>>> Is the long-term idea to put 'blktap3' in the tree?
>> I don't think so.
>>
>> George was going to make a proposal about future plans for blktap*.
> I belive it makes sense to wait for George's proposal.
>
> While I believe the outcome of it is going to be rm *blktap1* you
> never know who is going to come out of the woodshed when code is
> to be deleted.

My proposal doesn't consider blktap1.

While I'm all in favor of nuking useless functionality, I think the 
timing of this is pretty bad -- we've already had an RC; if anyone *is* 
using it, the probability of them not noticing that it's gone missing 
between now and the release is pretty high.  I'd much rather wait and 
nuke it at the beginning of the cycle, like we did with xend.

  -George

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