test/DebugInfo/Generic/missing-abstract-variable.ll fails on arm

[weimingz]

Summary

Does anyone else have the same issue? Should we XFail it?

[rengolin]

Hi Weiming,

The buildbots are green for a good number of years, so they're pretty good baseline for you to see if the problem happens upstream or not, meaning XFAILing tests on ARM or AArch64 are no longer accepted as a solution.

It may be that whatever revision you have in your tree broke that test, or that your environment is slightly different, but the bots are authoritative (we have to have some baseline). They have a list of commits you can look at to make sure your tree is as up-to-date as the bots, and if not, you should either update it or look down the waterfall to see if the particular commit you have fails or not.

Regardless, if the bots are currently green, XFAILing a test will break them. If it's not, fixing the problem is the best solution.

cheers,
--renato

[weimingz]

In D14782#292113, @rengolin wrote:

Hi Weiming,

The buildbots are green for a good number of years, so they're pretty good baseline for you to see if the problem happens upstream or not, meaning XFAILing tests on ARM or AArch64 are no longer accepted as a solution.

It may be that whatever revision you have in your tree broke that test, or that your environment is slightly different, but the bots are authoritative (we have to have some baseline). They have a list of commits you can look at to make sure your tree is as up-to-date as the bots, and if not, you should either update it or look down the waterfall to see if the particular commit you have fails or not.

Regardless, if the bots are currently green, XFAILing a test will break them. If it's not, fixing the problem is the best solution.

cheers,
--renato

Hi Renato,

Thanks. I'll look down to see why it fails in my local machine.

Thanks,
weiming

test/DebugInfo/Generic/missing-abstract-variable.ll

 ; REQUIRES: object-emission
 
 ; RUN: %llc_dwarf -O0 -filetype=obj < %s | llvm-dwarfdump -debug-dump=info - | FileCheck %s
 
 ; The formal parameter 'b' for Function 'x' when inlined within 'a' is lost on
 ; mips and powerpc64 (and on x86_64 at at least -O2). Presumably this is a
 ; SelectionDAG issue (do mips/powerpc64 use FastISel?).
-; XFAIL: mips, powerpc64, s390x, sparc
+; XFAIL: mips, powerpc64, s390x, sparc, arm
 
 ; Build from the following source with clang -O2.
 
 ; The important details are that 'x's abstract definition is first built during
 ; the definition of 'b', where the parameter to 'x' is constant and so 'x's 's'
 ; variable is optimized away. No abstract definition DIE for 's' is constructed.
 ; Then, during 'a' emission, the abstract DbgVariable for 's' is created, but
 ; the abstract DIE isn't (since the abstract definition for 'b' is already