alias analysis

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

alias analysis

Ryan M. Lefever
Since Data Structure alias analysis inherits from ModulePass rather than
FunctionPass, does that mean that it performs interprocedural alias
analysis?  Further, if I compare pointers from two different functions,
will I sometimes get a NoAlias response.

Regards,
Ryan

--
Ryan M. Lefever  [http://www.ews.uiuc.edu/~lefever]

_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Reply | Threaded
Open this post in threaded view
|

Re: alias analysis

Chris Lattner
On Mon, 16 Oct 2006, Ryan M. Lefever wrote:
> Since Data Structure alias analysis inherits from ModulePass rather than
> FunctionPass, does that mean that it performs interprocedural alias
> analysis?

yes.

> Further, if I compare pointers from two different functions,
> will I sometimes get a NoAlias response.

Don't do that :).  That's probably a bug in DSA, but fixing it would be
impractical.

-Chris

--
http://nondot.org/sabre/
http://llvm.org/
_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Reply | Threaded
Open this post in threaded view
|

Re: alias analysis

Vikram S. Adve-2
Ryan's second question should be more generally phrased for the  
entire AliasAnalysis interface, not just DSA:  will it work to ask  
alias queries or mod/ref queries for a pair of pointers or a pair of  
instructions in two different functions?

--Vikram
http://www.cs.uiuc.edu/~vadve
http://llvm.cs.uiuc.edu/


On Oct 16, 2006, at 1:45 PM, Chris Lattner wrote:

> On Mon, 16 Oct 2006, Ryan M. Lefever wrote:
>> Since Data Structure alias analysis inherits from ModulePass  
>> rather than
>> FunctionPass, does that mean that it performs interprocedural alias
>> analysis?
>
> yes.
>
>> Further, if I compare pointers from two different functions,
>> will I sometimes get a NoAlias response.
>
> Don't do that :).  That's probably a bug in DSA, but fixing it  
> would be
> impractical.
>
> -Chris
>
> --
> http://nondot.org/sabre/
> http://llvm.org/
> _______________________________________________
> LLVM Developers mailing list
> [hidden email]         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
Reply | Threaded
Open this post in threaded view
|

Re: alias analysis

Chris Lattner
On Mon, 16 Oct 2006, Vikram Adve wrote:
> Ryan's second question should be more generally phrased for the
> entire AliasAnalysis interface, not just DSA:  will it work to ask
> alias queries or mod/ref queries for a pair of pointers or a pair of
> instructions in two different functions?

Yes, an implementation should strive to implement that.  I believe that
globalmodref-aa (the only 'not research' IP alias analysis) does this
right.

-Chris

> On Oct 16, 2006, at 1:45 PM, Chris Lattner wrote:
>
>> On Mon, 16 Oct 2006, Ryan M. Lefever wrote:
>>> Since Data Structure alias analysis inherits from ModulePass
>>> rather than
>>> FunctionPass, does that mean that it performs interprocedural alias
>>> analysis?
>>
>> yes.
>>
>>> Further, if I compare pointers from two different functions,
>>> will I sometimes get a NoAlias response.
>>
>> Don't do that :).  That's probably a bug in DSA, but fixing it
>> would be
>> impractical.
>>
>> -Chris
>>
>> --
>> http://nondot.org/sabre/
>> http://llvm.org/
>> _______________________________________________
>> LLVM Developers mailing list
>> [hidden email]         http://llvm.cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>
> _______________________________________________
> LLVM Developers mailing list
> [hidden email]         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev
>

-Chris

--
http://nondot.org/sabre/
http://llvm.org/
_______________________________________________
LLVM Developers mailing list
[hidden email]         http://llvm.cs.uiuc.edu
http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev