[llvm] r214285 - [MCJIT] Fix the ARM BR24 relocation in RuntimeDyldMachO.

Lang Hames lhames at gmail.com
Wed Jul 30 09:35:47 PDT 2014


Hi Renalto,

Looking more closely at that log message, it seems like it's the LHS that's
going wrong - the decode_operand call. That is somewhat surprising, since
there is more to go wrong on the RHS, but at least makes it easy to narrow
down. If decode_operand is returning the wrong value, it must be that
either (1) we're pointing at the wrong location (unlikely), (2) we have
assembled the bits incorrectly, or (3) we have decoded the bits incorrectly.

Could you please run the llvm-mc line from the top of the test case and
send me the resulting foo.o object? That would help me rule out the second
option.

Cheers,
Lang.


On Wed, Jul 30, 2014 at 8:50 AM, Lang Hames <lhames at gmail.com> wrote:

> Hi Renalto,
>
> This test is notionally platform independent (in the sense that it should
> run the same on any target). I'll see if I can figure out what's going
> wrong here.
>
> Thanks for the heads up!
>
> - Lang.
>
>
>
> On Wed, Jul 30, 2014 at 4:06 AM, Renato Golin <renato.golin at linaro.org>
> wrote:
>
>> On 30 July 2014 04:35, Lang Hames <lhames at gmail.com> wrote:
>> > Author: lhames
>> > Date: Tue Jul 29 22:35:05 2014
>> > New Revision: 214285
>> >
>> > URL: http://llvm.org/viewvc/llvm-project?rev=214285&view=rev
>> > Log:
>> > [MCJIT] Fix the ARM BR24 relocation in RuntimeDyldMachO.
>> >
>> > We now (1) correctly decode the branch immediate, (2) modify the
>> immediate to
>> > corretly treat it as PC-rel, and (3) properly populate the stub entry.
>> > Previously we had been doing each of these wrong.
>>
>>
>> Hi Lang,
>>
>> You've probably seen it, but just in case you didn't, this broke the ARM
>> bot:
>>
>> http://lab.llvm.org:8011/builders/clang-native-arm-cortex-a15/builds/2150
>>
>> Is this a MachO specific thing that the JIT is not getting it right on
>> Linux?
>>
>> cheers,
>> --renato
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140730/07fd3757/attachment.html>


More information about the llvm-commits mailing list