Fwd: [RFC, PowerPC, Darwin] Clean up generation of ha16() / lo16() markers

[Don Crandall]

On 5/15/13 8:20 PM, David Fang wrote:
> Hi all,
>
>     Response below.
>
>> On 5/15/13 6:38 PM, David Fang wrote:
>>>  Hal, Ulrich,
>>>
>>>  Full llvm tests finished, and I did find 3 new failures:
>>>> *SNIP*
>>>
>>>  My branch is currently sync'd to r181825.
>>>  Is there something newer on trunk that might address this?
>>>  If so, what's the minumum revision I should re-merge from?
>>>
>>>  Fang
>>>
>>> > >   Can you please test this on your Darwin system?
>>> > > > >   Thanks again,
>>> > >   Hal
>>> > > > >   ----- Forwarded Message -----
>>> > >   From: "Ulrich Weigand" <Ulrich.Weigand at de.ibm.com>
>>> > >   To: llvm-commits at cs.uiuc.edu
>>> > >   Cc: "Hal Finkel" <hfinkel at anl.gov>
>>> > >   Sent: Wednesday, May 15, 2013 11:01:58 AM
>>> > >   Subject: [RFC, PowerPC, Darwin] Clean up generation of ha16() 
>>> / > >   lo16()
>>> > >   markers
>>> > > > >   Does this look reasonable?
>>> > > > >   Bye,
>>> > >   Ulrich
>>> > > > >   (See attached file: diff-reloc-mcexpr)
>>> > >
>
> On Wed, 15 May 2013, Don Crandall wrote:
>> Good evening,
>>
>> On Darwin9 I get the same failures without adding Ulrich's patch. I 
>> build trunk with gcc 4.8 or 4.7.
>>
>> Ulrichs' patch will be added to my G5 tree and rebuilt as soon as I 
>> finish the current unittesting with svn r181914, currently at ~57% on 
>> the G5/Darwin machine.
>>
>> The patch has also been applied to a trunk build on a PPC linux 
>> (Ubuntu) build I began within the last hour. It is currently @ ~20%. 
>> Last test results showed the exact failures indicated by David. Will 
>> have more (or detailed) information for you later this evening.
>> Best regards
>> Don
>
>     Odd, because my lastest build right before Ulrich's patch was 
> clean (w.r.t those tests):
> http://csl.cornell.edu/~fang/sw/llvm/logs/llvm-r181825-O2-powerpc-darwin8-g++-4.0.1-debug-check.log.bz2 
>
>
> This branch snapshot came from a merge with r181825.
> Built with apple gcc-4.0.1 with -g -O2, cmake+shared.
>
> Don, are you using my branch, or some derivative patches from it?
>
> I could probably just re-sync to tip-of-trunk and try again tonight.
> I do see a number of powerpc-related patches that I'm missing.
>
> Fang
>
Good evening, David;

I have not yet set up the darwin8 branch on the G5. This is only a cmake 
trunk built with homebrewed gcc 4.7 or 4.8.

The modification made to the trunk on the G5 include the 
PPCMachObjectWriter.cpp file with its associated changes to other files 
to make it work. I need to tweak it to silence a few warnings from gcc 
regarding a few relocation macros used in the file. Ulrichs' proposed 
changes implies some additional modifications. I'll get back to you on this.

G5 unit tests for the last build was 10 unexpected failures. With the 
most recent changes my build is now down to 9. I will now apply the 
patch and do some tweaking and rebuild.

More later tonight.

Best regards
Don