[llvm] r212203 - Don't try to construct debug LexicalScopes hierarchy for functions that do not have top level debug information.

David Blaikie dblaikie at gmail.com
Mon Jul 7 20:22:00 PDT 2014


On Thu, Jul 3, 2014 at 6:01 PM, Aditya Nandakumar
<aditya_nandakumar at apple.com> wrote:
> Thanks Duncan
> Bug point reduced it from 27MB to about 3.7. I ran it through opt and
> brought it down to 2.4MB and am unable to reduce further. Hopefully this is
> not too big.

I'll have to take a look, though I'm a bit concerned that a reduced IR
is going to be tricky to actually pinpoint the original source of the
problem as the IR may already contain the bad debug info metadata at
this point... and I won't be able to discover where that was
introduced (by which optimization, etc).

But I'll take a look, see if I can create something by hand, etc...
and see how I go.

Thanks,
- David

>
> Running:
> bin/llvm-lto -exported-symbols=“_main” bugpoint-reduced-opt.bc should
> reproduce the issue.
>
> Attached:
>
> On Jul 3, 2014, at 5:40 PM, Eric Christopher <echristo at gmail.com> wrote:
>
> On Thu, Jul 3, 2014 at 5:39 PM, Duncan P. N. Exon Smith
> <dexonsmith at apple.com> wrote:
>
>
> On 2014-Jul-03, at 15:34, Eric Christopher <echristo at gmail.com> wrote:
>
> On Thu, Jul 3, 2014 at 1:37 PM, Duncan P. N. Exon Smith
> <dexonsmith at apple.com> wrote:
>
> Hi David,
>
> Aditya is working on a reproduction.  Hopefully you'll have it soon.
>
> In the meantime, this has gummed up our internal buildbots.  Any chance
> you'd revert proactively?
>
>
> I've gone ahead and reverted this for now in r212298,
>
>
> Thanks Eric!
>
> but a
> reproducible testcase would be preferred since it's not showing up on
> the external bots.
>
>
> I'll make sure you get one.
>
> Aditya finally has something reproducible on trunk that's huge (~25MB
> bitcode).  He's just running bugpoint now.  Let's hope it shrinks a bit?
>
>
> Totally. And thanks a ton :)
>
> -eric
>
> Our lab.llvm.org LTO builder [1] would reproduce it, but the config
> change I made yesterday in our ailing public lab didn't seem to take
> effect -- I must have restarted the wrong buildmaster or something.
> I'll look into it.
>
> [1]: http://lab.llvm.org:8013/builders/clang-x86_64-darwin11-Rlto
>
>
>




More information about the llvm-commits mailing list