[llvm] r243960 - Update test suite to make "ninja check" succeed without native backend builtin

Mehdi Amini via llvm-commits llvm-commits at lists.llvm.org
Sun Aug 23 15:12:57 PDT 2015


> On Aug 23, 2015, at 2:38 PM, Robinson, Paul <Paul_Robinson at playstation.sony.com> wrote:
> 
> 
> 
>> -----Original Message-----
>> From: Mehdi Amini [mailto:mehdi.amini at apple.com]
>> Sent: Friday, August 21, 2015 5:04 PM
>> To: Robinson, Paul
>> Cc: llvm-commits
>> Subject: Re: [llvm] r243960 - Update test suite to make "ninja check"
>> succeed without native backend builtin
>> 
>> Hi Paul,
>> 
>>> On Aug 21, 2015, at 4:47 PM, Robinson, Paul
>> <Paul_Robinson at playstation.sony.com> wrote:
>>> 
>>> 
>>>> -----Original Message-----
>>>> From: llvm-commits-bounces at cs.uiuc.edu [mailto:llvm-commits-
>>>> bounces at cs.uiuc.edu] On Behalf Of Mehdi Amini
>>>> Sent: Monday, August 03, 2015 11:33 PM
>>>> To: llvm-commits at cs.uiuc.edu
>>>> Subject: [llvm] r243960 - Update test suite to make "ninja check"
>> succeed
>>>> without native backend builtin
>>>> 
>>>> Author: mehdi_amini
>>>> Date: Tue Aug  4 01:32:54 2015
>>>> New Revision: 243960
>>>> 
>>>> URL: https://urldefense.proofpoint.com/v2/url?u=http-
>> 3A__llvm.org_viewvc_llvm-2Dproject-3Frev-3D243960-26view-
>> 3Drev&d=BQIFAg&c=eEvniauFctOgLOKGJOplqw&r=v-
>> ruWq0KCv2O3thJZiK6naxuXK8mQHZUmGq5FBtAmZ4&m=jPiVoCJ0z7ZAiMhVjRCYQnQLtDHGH6
>> lSekDm2CE-bcY&s=uS_bxpYJ1Txsi1Tifw0hFQXb_0YnHQ4DKWzUekHnk3U&e=
>>>> Log:
>>>> Update test suite to make "ninja check" succeed without native backend
>>>> builtin
>>>> 
>>>> Requires "native" feature in most places that were failing.
>>>> 
>>>> From: Mehdi Amini <mehdi.amini at apple.com>
>>>> 
>>>> Added:
>>>>  llvm/trunk/test/CodeGen/Generic/lit.local.cfg
>>>>  llvm/trunk/test/DebugInfo/lit.local.cfg
>>> 
>>> Sorry, why did you need to disable all DebugInfo tests here?
>> 
>> As the commit message said, I’d like to be able to build LLVM on X86,
>> without the X86 backend enabled/configured/built-in, and be able to run
>> make check.
>> So the intention was to enable that, and I had failures with DebugInfo.
>> But it is also possible that I messed up somewhere and went wrong with
>> this folder, sorry if it is the case.
>> 
>> I’ll try to revert on DebugInfo and see how it goes.
>> 
>>> We run with different host and target triples, and these tests
>>> run fine for us.  (Host and target arch are the same, if that
>>> matters.)
>> 
>> Not sure I understand exactly what you mean here, are you running “ninja
>> check” on arch XYZ with the backend XYZ builtin?
> 
> Yes, we build and test on X86 Windows with a default triple of x86_64-scei-ps4.
> So the X86 backend is present, but the host and target triples do not match.
> 
> llc is probably using the host triple as the default target triple, so if
> you don't have a backend matching the host arch, that is likely to be a
> problem.  But, having the "native" arch backend should be sufficient, you
> should not have to have default target triple match the host triple.

Agree, I don’t know why “native” requires to have the default target triple match the host triple.
I only added the restriction that the native backend needs to be built in.
I can try to lift the triple restriction and see how it goes.

— 
Mehdi

> 
>> 
>>> It took a little while to notice this because the tests went from
>>> PASS to UNSUPPORTED and we weren't looking for that.
>> 
>> Yes, it is hard to detect, I had the same problem with a directory
>> (internal) that has “REQUIRES: Assert” and I was trying to validate my
>> changes on a release build. `ninja check` is successful but it does not
>> mean anything…
>> 
>>>> Mehdi
>> 
>> 
>>> 
>>> Thanks,
>>> --paulr
>>> 
>>>> Modified:
>>>>  llvm/trunk/test/Feature/optnone-llc.ll
>>>>  llvm/trunk/test/Feature/optnone.ll
>>>>  llvm/trunk/test/Linker/subprogram-linkonce-weak-odr.ll
>>>>  llvm/trunk/test/Linker/subprogram-linkonce-weak.ll
>>>>  llvm/trunk/test/Linker/type-unique-odr-a.ll
>>>>  llvm/trunk/test/Linker/type-unique-simple-a.ll
>>>>  llvm/trunk/test/Linker/type-unique-simple2-a.ll
>>>>  llvm/trunk/test/Linker/type-unique-simple2.ll
>>>>  llvm/trunk/test/Linker/type-unique-type-array-a.ll
>>>>  llvm/trunk/test/Other/2010-05-06-Printer.ll
>>>>  llvm/trunk/test/Transforms/BranchFolding/2007-10-19-
>>>> InlineAsmDirectives.ll
>>>>  llvm/trunk/test/tools/llvm-mc/fatal_warnings.test
>>>>  llvm/trunk/test/tools/llvm-mc/line_end_with_space.test
>>>>  llvm/trunk/test/tools/llvm-mc/no_warnings.test
>>>>  llvm/trunk/test/tools/lto/opt-level.ll
>>>> 
>>>> Added: llvm/trunk/test/CodeGen/Generic/lit.local.cfg
>>>> URL: https://urldefense.proofpoint.com/v2/url?u=http-
>> 3A__llvm.org_viewvc_llvm-2D&d=BQIFAg&c=eEvniauFctOgLOKGJOplqw&r=v-
>> ruWq0KCv2O3thJZiK6naxuXK8mQHZUmGq5FBtAmZ4&m=jPiVoCJ0z7ZAiMhVjRCYQnQLtDHGH6
>> lSekDm2CE-bcY&s=s9ppoRUJJgAgXYfSO-QA1zZDcys4P28PO11EdADYJrw&e=
>>>> 
>> project/llvm/trunk/test/CodeGen/Generic/lit.local.cfg?rev=243960&view=auto
>>>> 
>> ==========================================================================
>>>> ====
>>>> --- llvm/trunk/test/CodeGen/Generic/lit.local.cfg (added)
>>>> +++ llvm/trunk/test/CodeGen/Generic/lit.local.cfg Tue Aug  4 01:32:54
>> 2015
>>>> @@ -0,0 +1,3 @@
>>>> +if 'native' not in config.available_features:
>>>> +    config.unsupported = True
>>>> +
>>>> 
>>>> Added: llvm/trunk/test/DebugInfo/lit.local.cfg
>>>> URL: https://urldefense.proofpoint.com/v2/url?u=http-
>> 3A__llvm.org_viewvc_llvm-2D&d=BQIFAg&c=eEvniauFctOgLOKGJOplqw&r=v-
>> ruWq0KCv2O3thJZiK6naxuXK8mQHZUmGq5FBtAmZ4&m=jPiVoCJ0z7ZAiMhVjRCYQnQLtDHGH6
>> lSekDm2CE-bcY&s=s9ppoRUJJgAgXYfSO-QA1zZDcys4P28PO11EdADYJrw&e=
>>>> project/llvm/trunk/test/DebugInfo/lit.local.cfg?rev=243960&view=auto
>>>> 
>> ==========================================================================
>>>> ====
>>>> --- llvm/trunk/test/DebugInfo/lit.local.cfg (added)
>>>> +++ llvm/trunk/test/DebugInfo/lit.local.cfg Tue Aug  4 01:32:54 2015
>>>> @@ -0,0 +1,3 @@
>>>> +if 'native' not in config.available_features:
>>>> +    config.unsupported = True
>>>> +



More information about the llvm-commits mailing list