This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [trans-mem] make alias oracle understand TM read/writes
- From: Aldy Hernandez <aldyh at redhat dot com>
- To: Richard Henderson <rth at redhat dot com>
- Cc: Richard Guenther <rguenther at suse dot de>, Daniel Berlin <dberlin at dberlin dot org>, gcc-patches at gcc dot gnu dot org
- Date: Tue, 1 Sep 2009 10:54:48 -0400
- Subject: Re: [trans-mem] make alias oracle understand TM read/writes
- References: <alpine.LNX.2.00.0908311725450.28140@zhemvz.fhfr.qr> <20090831203402.GA615@redhat.com> <4aca3dc20908311355r6b680e26h820e5815357cc47c@mail.gmail.com> <alpine.LNX.2.00.0908312328460.28140@zhemvz.fhfr.qr> <4A9C44A2.5020403@redhat.com> <alpine.LNX.2.00.0908312348300.28140@zhemvz.fhfr.qr> <4A9C4F07.7050603@redhat.com> <alpine.LNX.2.00.0909011010540.28140@zhemvz.fhfr.qr> <20090901142001.GA10863@redhat.com> <4A9D3393.8020608@redhat.com>
On Tue, Sep 01, 2009 at 07:45:39AM -0700, Richard Henderson wrote:
> On 09/01/2009 07:20 AM, Aldy Hernandez wrote:
>> So, is the patch approved? I think it's safe to say you (Richard G) can
>> approve it and rth won't have any issues. Otherwise I can take the
>> blame.
> ...
>> - return 0;
>> + return TODO_rebuild_alias;
>
> Everything except this bit is approved.
> We don't need to recompute at all.
Except if we don't recompute, my testcase will fail because we haven't
seen any TM builtins in tree-ssa-structalias.c:find_func_aliases().
I guess if what y'all want is alias info for builtins created by hand
then I'll just rewrite my testcase with manually created builtin calls.
Is this what you want?
Aldy