[llvm] r226588 - Add an assertion and prefer a crash over an infinite loop.

[David Blaikie]

On Tue, Jan 20, 2015 at 11:52 AM, Adrian Prantl <aprantl at apple.com> wrote:

>
> On Jan 20, 2015, at 11:46 AM, Mehdi Amini <mehdi.amini at apple.com> wrote:
>
>
> On Jan 20, 2015, at 11:10 AM, David Blaikie <dblaikie at gmail.com> wrote:
>
>
>
> On Tue, Jan 20, 2015 at 10:58 AM, Adrian Prantl <aprantl at apple.com> wrote:
>
>>
>>
>> > Begin forwarded message:
>> >
>> > Subject: Re: [llvm] r226588 - Add an assertion and prefer a crash over
>> an infinite loop.
>> > From: "Duncan P. N. Exon Smith" <dexonsmith at apple.com>
>> > Date: January 20, 2015 at 10:53:38 AM PST
>> > Cc: llvm-commits at cs.uiuc.edu
>> > To: Adrian Prantl <aprantl at apple.com>
>> >
>> >
>> >> On 2015 Jan 20, at 10:03, Adrian Prantl <aprantl at apple.com> wrote:
>> >>
>> >> Author: adrian
>> >> Date: Tue Jan 20 12:03:37 2015
>> >> New Revision: 226588
>> >>
>> >> URL: http://llvm.org/viewvc/llvm-project?rev=226588&view=rev
>> >> Log:
>> >> Add an assertion and prefer a crash over an infinite loop.
>> >>
>> >> Modified:
>> >>   llvm/trunk/lib/IR/DebugInfo.cpp
>> >>
>> >> Modified: llvm/trunk/lib/IR/DebugInfo.cpp
>> >> URL:
>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/IR/DebugInfo.cpp?rev=226588&r1=226587&r2=226588&view=diff
>> >>
>> ==============================================================================
>> >> --- llvm/trunk/lib/IR/DebugInfo.cpp (original)
>> >> +++ llvm/trunk/lib/IR/DebugInfo.cpp Tue Jan 20 12:03:37 2015
>> >> @@ -527,12 +527,15 @@ bool DISubprogram::Verify() const {
>> >>        while ((IA = DL.getInlinedAt()))
>> >>          DL = DebugLoc::getFromDILocation(IA);
>> >>        DL.getScopeAndInlinedAt(Scope, IA);
>> >> +        assert(Scope && "debug location has no scope");
>> >>        assert(!IA);
>> >>        while (!DIDescriptor(Scope).isSubprogram()) {
>> >>          DILexicalBlockFile D(Scope);
>> >>          Scope = D.isLexicalBlockFile()
>> >>                      ? D.getScope()
>> >>                      :
>> DebugLoc::getFromDILexicalBlock(Scope).getScope();
>> >> +          if (!Scope)
>> >> +            llvm_unreachable("lexical block file has no scope");
>> >
>> > It's strange that this isn't simply an assertion:
>> >
>> >    assert(Scope && "lexical block file has no scope");
>> >
>> > But I guess that's so that it crashes even when NDEBUG?  (Is that
>> > reliable?)
>>
>> Yes, that’s the intention (and llvm_unreachable is meant to be used that
>> way).
>>
>
> I don't believe that's the intention of the LLVM style guide - generally
> we try to remove conditionals like this and replace the conditional with
> the assertion so the dead code isn't exercised in release builds.
>
>
> Isn’t the point of “unreachable” to let the compiler do whatever it wants
> in this case?
> I thought LLVM was free to just drop the test entirely in release mode?
>
>
> After going back to ErrorHandling.h it looks as if report_fatal_error()
> would be more appropriate.
>

Yeah, I've never been clear on the rules of report_fatal_error (though it
seems problematic for a library like LLVM to just bring down the process,
as I understand it) - if returning false and testing works as Duncan
suggested, it seems like the best option to me.


> llvm_unreachable() either becomes llvm_unreachable_internal() which
> invokes abort() and prints the message or __builtin_unreachable() which
> indeed drops the message.
>
> -- adrian
>
> Mehdi
>
>
>
>
>> >
>> > If that's the reason, you should probably add a comment to that
>> > effect since otherwise I'm likely to switch it to an assert next
>> > time I look at it ;).
>> >
>> > (Also, do you know when this situation arises?)
>>
>> Happened to me when I compiled an old .ll file without the MDLocation
>> upgrade.
>>
>> -- adrian
>>
>> >
>> >>        }
>> >>        if (!DISubprogram(Scope).describes(F))
>> >>          return false;
>> >>
>> >>
>> >> _______________________________________________
>> >> llvm-commits mailing list
>> >> llvm-commits at cs.uiuc.edu
>> >> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>> >
>>
>>
>> _______________________________________________
>> llvm-commits mailing list
>> llvm-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150120/24dc44c3/attachment.html>