Clang Stage 2: configure, RLTO, using Stage 1 RA (Check) - Build # 8590 - Failure!

[Hal Finkel via llvm-commits]

Hi everyone,

Could someone please revert r275678 (LLVM) and 275679 (Clang) for me? I can't do this myself right now. If someone with this configuration could bugpoint these regressions test failures (which I imagine should be possible given that this is a stage-2 LTO build), that would certainly be appreciated.

Interestingly, this was the only builder that complained. Is this the only self-host w/ LTO builder that we have? Also, what do I click on in Jenkins (or elsewhere) to see the full build parameters, system information, etc.?

Thanks again,
Hal

----- Original Message -----
> From: "llvm greendragon" <llvm.greendragon at gmail.com>
> To: "Tobias Grosser" <tobias at grosser.es>, "Hal Finkel" <hfinkel at anl.gov>
> Cc: "Lang Hames" <lhames at apple.com>
> Sent: Saturday, July 16, 2016 6:28:29 AM
> Subject: Clang Stage 2: configure, RLTO, using Stage 1 RA (Check) - Build # 8590 - Failure!
> 
> [ FAILURE ] clang-stage2-configure-Rlto_check [#8590]
> Build URL:
> 	http://lab.llvm.org:8080/green/job/clang-stage2-configure-Rlto_check/8590/
> Project: 	clang-stage2-configure-Rlto_check
> Date of build: 	Sat, 16 Jul 2016 04:15:29 -0700
> Build duration: 	12 min Identified problems:
> 
> 
>     * Compile Error: This build failed because of a compile error.
>     Below is a list of all errors in the build log:
>         * Indication 1
>     * Regression test failed: This build failed because a regression
>     test in the test suite FAILed. See the test report for details.
>         * Indication 1
>     * Ninja target failed: Below is a link to the first failed ninja
>     target.
>         * Indication 1
> Tests:
> 
> Name: Clang
> Failed: 0 test(s), Passed: 8074 test(s), Total: 8074 test(s)
> 
> Name: Clang-Unit
> Failed: 0 test(s), Passed: 1570 test(s), Total: 1570 test(s)
> 
> Name: LLVM
> Failed: 2 test(s), Passed: 15597 test(s), Total: 15599 test(s)
> 
> 
>     * Failed: LLVM.Analysis/DependenceAnalysis.GCD.ll
> 
> 
> 
>     * Failed: LLVM.Transforms/ObjCARC.basic.ll
> 
> 
> Name: LLVM-Unit
> Failed: 0 test(s), Passed: 1699 test(s), Total: 1699 test(s)
> 
> Changes
> 
> 
>     * Commit 275680 by grosser:
> 
> GPGPU: Abort if any dummy function is called This ensures that
> accidental calls to these functions will break loadly instead of
> corrupting the stack with invalid return values. These functions
> have been introduced earlier as replacement of pet and parts of ppcg
> which we will never use and consequently have not been imported or
> compiled into Polly.
>         * edit :
>         llvm-revision.src/polly/trunk/lib/External/ppcg/external.c
>     * Commit 275679 by hfinkel:
> 
> Revert "Revert r275029 - Update Clang tests after adding inference
> for the returned argument attribute" This reverts commit r275043
> after reapplying the underlying LLVM commit.
>         * edit :
>         llvm-revision.src/cfe/trunk/test/CodeGen/ppc64-struct-onevect.c
>         * edit :
>         llvm-revision.src/cfe/trunk/test/CodeGenCXX/wasm-args-returns.cpp
>         * edit :
>         llvm-revision.src/cfe/trunk/test/CodeGenOpenCL/as_type.cl
>     * Commit 275678 by hfinkel:
> 
> Revert "Revert r275027 - Let FuncAttrs infer the 'returned' argument
> attribute" This reverts commit r275042; the initial commit triggered
> self-hosting failures on ARM/AArch64. James Molloy identified the
> problematic backend code, which has been disabled in r275677. Trying
> again... Original commit message: Let FuncAttrs infer the 'returned'
> argument attribute A function can have one argument with the
> 'returned' attribute, indicating that the associated argument is
> always the return value of the function. Add FuncAttrs inference
> logic.
>         * edit :
>         llvm-revision.src/llvm/trunk/lib/Transforms/IPO/FunctionAttrs.cpp
>         * edit :
>         llvm-revision.src/llvm/trunk/test/Transforms/FunctionAttrs/2009-01-02-LocalStores.ll
>         * edit :
>         llvm-revision.src/llvm/trunk/test/Transforms/FunctionAttrs/nocapture.ll
>         * edit :
>         llvm-revision.src/llvm/trunk/test/Transforms/FunctionAttrs/readattrs.ll
>     * Commit 275677 by hfinkel:
> 
> Disable this-return argument forwarding on ARM/AArch64 r275042
> reverted function-attribute inference for the 'returned' attribute
> because the feature triggered self-hosting failures on ARM and
> AArch64. James Molloy determined that the this-return argument
> forwarding feature, which directly ties the returned input argument
> to the returned value, was the cause. It seems likely that this
> forwarding code contains, or triggers, a subtle bug. Disabling for
> now until we can track that down.
>         * edit :
>         llvm-revision.src/llvm/trunk/lib/Target/AArch64/AArch64ISelLowering.cpp
>         * edit :
>         llvm-revision.src/llvm/trunk/lib/Target/ARM/ARMISelLowering.cpp
>         * edit :
>         llvm-revision.src/llvm/trunk/test/CodeGen/AArch64/arm64-this-return.ll
>         * edit :
>         llvm-revision.src/llvm/trunk/test/CodeGen/ARM/returned-ext.ll
>         * edit :
>         llvm-revision.src/llvm/trunk/test/CodeGen/ARM/this-return.ll
> 

-- 
Hal Finkel
Assistant Computational Scientist
Leadership Computing Facility
Argonne National Laboratory