[llvm-commits] [PATCH] [ARM JIT] Pseudo Instruction ARM::LEApcrel defined but not implemented

Nick Lewycky nlewycky at google.com
Wed May 26 20:55:37 PDT 2010


+   // For VFP load, the immediate offset is multiplied by 4.
+   unsigned Reloc =  ((TID.TSFlags & ARMII::FormMask) == ARMII::VFPLdStFrm)
+    ? ARM::reloc_arm_vfp_cp_entry : ARM::reloc_arm_cp_entry;

There's two spaces after the first = sign. Please remove one.

I'm a little worried that this is cut-and-pasted from getMachineOpValue().
Feel free to fix the extra space there too, by the way. Is there a sensible
way to factor this logic out?

+     // Materialize contant pool index address.

Typo: "contant".

This is similar enough in implementation to emitLEApcrelInstruction that I
think it's fine to go ahead and commit it as is, then make any changes Evan
or others ask for afterwards.

Nick

On 26 May 2010 03:19, Shih-wei Liao <sliao at google.com> wrote:

> See http://llvm.org/bugs/show_bug.cgi?id=7227: ARM JIT misses the
> implementation of ARM::LEApcrel.
>
> The patch passes both "make check-lit" and "make check".
>
> For "make check-lit",
>  Expected Passes    : 4101
>  Expected Failures  : 24
>  Unsupported Tests  : 1184
> make[1]: Leaving directory `/usr/local/google/upstream/llvm-obj/test'
>
> For "make check",
>                ===  Summary ===
>
> # of expected passes            3842
> # of expected failures          24
> make[1]: Leaving directory `/usr/local/google/upstream/llvm-obj/test'
>
> Basically, the patch added the case "ARM::LEApcrel" next to the "case
> ARM::LEApcrelJT". The latter was implemented, but not the former. With
> the added implementation, all our input programs result in correct
> execution under LLVM's ARM JIT.
> --
> Thanks,
>   Shih-wei
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20100526/667dbf05/attachment.html>


More information about the llvm-commits mailing list