Hi, On Thu, Nov 14, 2013 at 09:24:54AM +0100, Jakub Jelinek wrote: > On Wed, Nov 06, 2013 at 05:26:08PM +0100, Martin Jambor wrote: > > So, Vlad, Steven, do you think that this time I have re-computed all > > that is necessary? Do you think the patch is OK? > > > > Thanks a lot and sorry for the breakage, > > I'm afraid there are still issues left. > Last night I was bootstrapping (first r204752, then when it failed > also r204751) with: > http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01268.html > http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01437.html > http://gcc.gnu.org/ml/gcc-patches/2013-11/txturet7Hr1Ws.txt > http://gcc.gnu.org/ml/gcc-patches/2013-11/msg01553.html > patches applied on i686-linux (see > http://gcc.gnu.org/ml/gcc-patches/2013-10/msg02625.html > on how I configure/build such a gcc on x86_64-linux host), > and while it bootstrapped just fine, there were hundreds of regressions > in the vectorization tests, which worked fine if compiled with > stage1 cc1/cc1plus, but with stage2/3 cc1/cc1plus behaved strangely. > When I've reverted the ira.c part of r204698 and bootstrapped/regtested > with all those patches again, all the problems went away. > Valgrind hasn't revealed any undefined uses caused by those patches. > > As it doesn't affect bootstrap/regtest of unpatched gcc, it is not an > immediate blocker (well, would be if/once the patches make it in), > but silent wrong code (sure, just suspected) is always very important. > Can you please have a look?
I'm already in the process of debugging PR 59099 and at the moment I believe I am on the right track. Let's hope it is the same problem but if not, I will certainly have a look. Thanks for the report, Martin > > > 2013-11-04 Martin Jambor <mjam...@suse.cz> > > > > PR rtl-optimization/10474 > > * ira.c (interesting_dest_for_shprep): New function. > > (split_live_ranges_for_shrink_wrap): Likewise. > > (find_moveable_pseudos): Move calculation of dominance info, > > df_analysios and the final anlyses to... > > (ira): ...here, call split_live_ranges_for_shrink_wrap. > > > > testsuite/ > > * gcc.dg/pr10474.c: New testcase. > > * gcc.dg/ira-shrinkwrap-prep-1.c: Likewise. > > * gcc.dg/ira-shrinkwrap-prep-2.c: Likewise. > > Jakub
