[LLVMdev] 3.4.1 Release Plans

[Tom Stellard]

On Tue, Apr 01, 2014 at 06:09:15PM +0800, Jiangning Liu wrote:
> Hi,
> 
> I'd like to say it is not a new feature at all. How should we define a new
> feature? What I said is "Complete AArch64 NEON feature" rather than "new
> AArch64 NEON feature".
> 
> If we look at the list for LLVM patches, most of them are bug fixes
> addressing compiler crash. Those fixes makes AArch64 NEON more robustness.
> 
> If we don't apply those patches, AArch64 NEON would often crash. It is bug
> fix or new feature? How do you judge?
>

In principle, I don't really have a problem with this kind of change going
into the stable branch since it only modifies two backends
and not common code, but I would like to know how these changes affect
the non-neon parts of the AArch64 and ARM backends?

Also, in order to be accepted, these changes will need to be approved by
the code owners for both ARM and AARCH64 and we will need to verify that
these patches don't change the ABI.

I've been using:
http://ispras.linuxbase.org/index.php/ABI_compliance_checker
to verify ABI consistency when merging patches.  Could you test it out
on your branch?  I think changes like r200768 which add or remove
intrinsics in llvm/include/IR will end up breaking the ABI.

Also, if you use git, it would be really helpful if you could push a
branch with these changes somewhere for me to look at.

Thanks,
Tom


> You may have a try by simply applying those two monolithic patches I
> provided. You won't see any failure on LLVM side.
> 
> I think the purpose of this new 3.4.1 release is to fix all of the known
> bugs and bring value to end-user. I would be sad if we don't bring AArch64
> NEON fix into this release, and end-user would be also sad, I believe.
> 
> Thanks,
> -Jiangning
> 
> 
> 
> 2014-04-01 17:30 GMT+08:00 Sylvestre Ledru <sylvestre at debian.org>:
> 
> > On 01/04/2014 11:26, Renato Golin wrote:
> > > Hi Jiangning,
> > >
> > > Thank you very much for doing all the hard work of integrating the
> > > patches and testing it beforehand, it really helps the back-porting!
> > >
> > >
> > > On 1 April 2014 09:25, Jiangning Liu <liujiangning1 at gmail.com> wrote:
> > >> I know the patch list is little bit longer, we have the following
> > reasons,
> > >> 1) Last year, when branch 3.4 was created, actually we didn't really
> > have
> > >> time to complete all AArch64 neon work. So branch 3.4 is actually at the
> > >> middle stage of aarch64 neon implementation. Now the patches I'm
> > requesting
> > >> intends to give a complete AArch64 NEON feature.
> > > I'd be less comfortable with this change than the others, since it
> > > introduces new functionality (and may introduce incompatibility with
> > > previous behaviour), but I'll let Tom decide.
> > With my packager hat, I agree with you. I don't think we should accept
> > any new features in the stable releases.
> > Only important bugs fixes should be part of the release.
> >
> > Sylvestre
> >
> >
> 
> 
> -- 
> Thanks,
> -Jiangning

> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits