[llvm-commits] [PATCH 1/5] Fix arithmetic promotion rule in ELF.h

Adhemerval Zanella azanella at linux.vnet.ibm.com
Thu Sep 6 07:22:21 PDT 2012


On 09/05/2012 02:23 PM, Matt Beaumont-Gay wrote:
> On Wed, Sep 5, 2012 at 10:13 AM, Will Schmidt <will_schmidt at vnet.ibm.com> wrote:
>> On Wed, 2012-09-05 at 13:00 -0300, Adhemerval Zanella wrote:
>>> Fix for some arithmetic additions being handled as signed where it should
>>> not. Also add the PPC64 triple definitions.
>> The 0 to 0UL changes look reasonable to me.   Did you identify an
>> existing testcase that fails/works with the signed change?   I didn't
>> notice a difference out of 'gmake check-all' with/without this patch
>> applied.
>>
>> May wish to break out the PPC64 triple change, I think it's mostly
>> unrelated to the 0UL bits.
> +1, please keep unrelated changes separate.
>
I split the patch in two now:

01-elf-arithmetic-fixes.patch - The fixes for 0/0UL. I observed it wen working in the
MCJIT for PPC64, specially when MCJIT tried to load some object in a memory position
with address higher than 32-bits. I suppose the 'gmake check-all' does not trigger it.

01-01-ppc64-elf-defines.patch - The PPC64 defines itself.

Are they ok now?


-- 
Adhemerval Zanella Netto
  Software Engineer
  Linux Technology Center Brazil
  Toolchain / GLIBC on Power Architecture
  azanella at linux.vnet.ibm.com / azanella at br.ibm.com
  +55 61 8642-9890

-------------- next part --------------
A non-text attachment was scrubbed...
Name: 01-elf-arithmetic-fixes.patch
Type: text/x-patch
Size: 788 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20120906/5f216e0e/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 01-ppc64-elf-defines.patch
Type: text/x-patch
Size: 646 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20120906/5f216e0e/attachment-0001.bin>


More information about the llvm-commits mailing list