This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [tree-ssa] Removing redundant loads


In message <1054265433.21692.117.camel@frodo.toronto.redhat.com>, Diego Novillo
 writes:
 >I'm attaching the .i.gz file so that you don't have to go through the
 >whole cycle.  I've run into odd things with PPC in the past.  A few
 >weeks ago STATEMENT_CODE_P() was casting the index to 'int' instead of
 >'size_t' and that was causing random failures.  Today, it was the
 >pushing of int instead of host int.  Oh, well.  I guess this means that
 >if it works on PPC, we're doing pretty good :)
Tee hee hee.  I know what's causing this.  With work I could probably
expose it on just about any target.  Give me a few minutes to think a
solution through.

The fundamental problem is while iterative_hash_expr knows how to deal with
commutative operations, avail_expr_hash does not.  Thus we get different
hashes because the order of operands in the use operands array is different
between the two statements.  Ugh.  We're probably going to have the same 
problem with VUSE/VDEFs.

Jeff


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]