[llvm-commits] [llvm] r71077 - in /llvm/trunk: lib/Transforms/IPO/FunctionAttrs.cpp test/Transforms/FunctionAttrs/2009-05-06-Malloc.ll

[Chris Lattner]

On May 7, 2009, at 10:09 AM, Duncan Sands wrote:

> Hi Owen,
>
>> We don't do much in the way of optimization of MallocInst anyways.
>> The only thing that this helps is that it prevents a call to malloc()
>> from clobbering results of a memory dependence query.  As I've
>> advocated on IRC,  I think we should just add  LibCallAA support for
>> saying that malloc() clobbers no known pointers, just like we can use
>> it to say that libm functions only clobber errno.  This seems like  
>> the
>> "right" way to go to me.
>
> "malloc" is already marked as returning a "noalias" result IIRC.
> Thus there is nothing to be done for malloc: it has already been
> done!  MallocInst is different to malloc, and should just be
> removed from LLVM.  The only problem I see with this is how to
> autoupgrade existing llvm that uses the malloc instruction.  Some
> trick (GEP?) will be needed to get the size of the object being
> allocated.

Gep should work fine.  There is also no need to handle this with  
libcallaa, basicaa can handle "call malloc" directly.

-Chris