[PATCH] [RFC][1/17]TILE-Gx: recognize tilegx triple

[Jiong Wang]

于 2013/3/12 17:17, Dmitri Gribenko 写道:
> On Tue, Mar 12, 2013 at 10:00 AM, Jiong Wang <jiwang at tilera.com> wrote:
>> 于 2013/3/12 15:48, Joerg Sonnenberger 写道:
>>
>>> On Tue, Mar 12, 2013 at 09:41:13AM +0800, Jiong Wang wrote:
>>>> I know the maturity of a new target is very important to be
>>>> accepted, then what's the requirement for a target to be enabled by
>>>> default?
>>> Getting everything is obviously the first part. After that, a buildbot
>>> should show that it works and some of the existing bots should have it
>>> enabled as well.
>>>
>>> My suggestion is to keep it disabled for the initial patch so that it
>>> can go in now. This makes the change to enable it a very small patch,
>>> which can also be easily reverted in case there is a problem. It is not
>>> meant to say anything about the quality of the backend or how likely it
>>> is to create problems. I hope this clarifies it a bit?
>> Hi Joerg,
>>
>> I see, thanks for your clarification. It's OK to keep tilegx default
>> disabled, I will confirm with the company and try to setup a buildbot.
> Hi Jiong,
>
> I can enable it on my x86-64 debian buildbot (of course, if it does
> not break anything).  It builds with two build systems:
> configure+Makefile and cmake+ninja+ccache.
>
> Of course, if you can set up a fast buildbot -- please do it.  That's
> a good thing to do in any case.  Is it possible and reasonable (in
> terms of build times) to set up a buildbot on the actual target
> hardware?  This is something that community can not do.
Hi Dimit,

  thanks for your help.

   yes, I am trying to setup a tilegx 36 core build slave. I am learning 
buildbot documents, and will confirm with the company whether it's OK to 
send the result back to coummunity server.

---
Regards,
Jiong

>
> Dmitri
>