[llvm-commits] RFC: patch for PR1782 (BasicAliasAnalyis)

[Wojciech Matyjewicz]

Chris Lattner wrote:

> Yep, it looks Gordon was right :).  Thanks again Wojtek, I applied the  
> patch.  Please close the PR,

Before I close the PR, I would like to discuss one more issue discovered
while working on it.

Suppose, we have a target with 32-bit pointers and the following
instructions:

%p = getelementptr i32* %x, i32 -1
%q = getelementptr i32* %x, i32 1073741823  ;(1073741823 == 2^30 - 1)

TargetData::getIndexedOffset() uses 64-bit arithmetic to perform offset
computation and return 64-bit values. Hence, it will return -4 as an
offset for %p, and 2^32 - 4 for %q. Based on these offsets, it may seem
that %p and %q point to different memory objects. However, they don't,
taking into account that pointers are 32-bit long.

I guess, such a large positive index in GEP as seen above can be
introduced by -instcombine pass.

BasicAA seems to be affected by the issue. For the attached example opt
-aa-eval says that %p and %q don't alias.

I think, the simplest way to fix it is to truncate the computed offset
to the target pointer size before returning it in
TargetData::getIndexedOffset(). If you think this is the correct
approach, I may prepare a patch.

-Wojtek
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: example.ll
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20071209/e5ef51a0/attachment.ksh>