[llvm] r222727 - Fix overly aggressive type merging.

Duncan P. N. Exon Smith dexonsmith at apple.com
Thu Nov 27 09:01:53 PST 2014


> On 2014 Nov 27, at 08:57, Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote:
> 
>> 
>> On 2014 Nov 27, at 08:46, Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote:
>> 
>> Our public bots are failing to bootstrap:
>> 
>> Last passing @ r222698:
>> http://lab.llvm.org:8080/green/job/clang-Rlto_master_build/532/
>> 
>> First failing @ r222843:
>> http://lab.llvm.org:8080/green/job/clang-Rlto_master_build/533/
>> 
>> Our internal bootstraps are failing too, and they somehow came up
>> with a much tighter range: r222725 is passing, and r222731 is
>> failing.
>> 
>> This (r222727) looks like the only possible commit in the narrow
>> range pointed at by our internal builders that could have caused
>> a problem.
>> 
>> Here's the failure -- LTO crashes while handling libclang.dylib:
>> 
>> http://lab.llvm.org:8080/green/job/clang-Rlto_master_build/533/consoleFull#-158682280549ba4694-19c4-4d7e-bec5-911270d8a58c
>> 
>> Do you mind if I speculatively revert?
>> 
> 
> Now that I've looked at the error more closely, I'm pretty confident
> this is the problem, so I'm going to revert right away to get our
> builds green.  Your commit makes type merging less aggressive, and
> the error from the GEP could be fixed by making it more aggressive.
> 
> GEP is not of right type for indices!
>  %InfoObj.i.i = getelementptr inbounds %"class.llvm::OnDiskIterableChainedHashTable"* %.lcssa, i64 0, i32 0, i32 4, !dbg !123627
> %"class.clang::serialization::reader::ASTIdentifierLookupTrait" = type { %"class.clang::ASTReader.31859"*, %"class.clang::serialization::ModuleFile.31870"*, %"class.clang::IdentifierInfo"* }LLVM ERROR: Broken function found, compilation aborted!
> 
> Let me know if you need help reproducing -- although we're all on
> holiday down here, so turnaround might be poor :(.
> 
Reverted in r222895.



More information about the llvm-commits mailing list