[llvm-testresults] llvm-arm-linux buildbot regression

[Jim Grosbach]

On Dec 1, 2010, at 1:29 PM, Eric Christopher wrote:

> 
> On Dec 1, 2010, at 1:22 PM, Dan Gohman wrote:
> 
>> Hello,
>> 
>> The llvm-arm-linux buildbot saw two tests recently start failing
>> these two tests:
>> 
>> 	• test/ExecutionEngine/2005-12-02-TailCallBug.ll
>> 	• test/ExecutionEngine/test-call.ll
>> 
>> These weren't reported because the buildbot is already red because
>> apm.ll has been failing for a while, for unrelated reasons.
>> 
>> The first failed build where these two tests failed is here:
>> 
>> http://google1.osuosl.org:8011/builders/llvm-arm-linux/builds/3
>> 
>> I'm guessing it's related to one of the several ARM codegen
>> commits included in that build.
>> 
>> Can one of you look into this?
> 
> Going to guess Jim's change in 120438 since it deals with BX_CALL...


Very likely something of mine, yes. That's as likely a candidate as any.

These are very tricky to fix, as those are Linux ARM JIT failures, so it's often not reproducible here. Hopefully I can get something similar to fail here. If not, we may need to xfail them until someone with access to a linux target board can look at it.