Re: [PATCH] Fix PR56035

2013-01-22 Thread Richard Biener
Marek Polacek wrote: >On Tue, Jan 22, 2013 at 06:44:40PM +0100, Zdenek Dvorak wrote: >> > I was however worried that we might end up with an edge >> > coming out of BB >> > from different loop heading into the header. In this case setting >up >> > loop's latch as the source of the latch edge wou

Re: [PATCH] Fix PR56035

2013-01-22 Thread Marek Polacek
On Tue, Jan 22, 2013 at 06:44:40PM +0100, Zdenek Dvorak wrote: > > I was however worried that we might end up with an edge > > coming out of BB > > from different loop heading into the header. In this case setting up > > loop's latch as the source of the latch edge would be wrong. > > Actually, t

Re: [PATCH] Fix PR56035

2013-01-22 Thread Marek Polacek
On Tue, Jan 22, 2013 at 03:30:06PM +0100, Richard Biener wrote: > Marek Polacek wrote: > > >We ICEd on attached testcase because we didn't properly detected the > >latch edge. Furthermore, I think the code for re-computing latches > >is somehow broken at the moment. In fix_loop_structure we hav

Re: [PATCH] Fix PR56035

2013-01-22 Thread Zdenek Dvorak
Hi, > > > We ICEd on attached testcase because we didn't properly detected the > > > latch edge. Furthermore, I think the code for re-computing latches > > > is somehow broken at the moment. In fix_loop_structure we have > > > /* If there was no latch, schedule the loop for removal. */ >

Re: [PATCH] Fix PR56035

2013-01-22 Thread Marek Polacek
On Tue, Jan 22, 2013 at 05:26:02PM +0100, Zdenek Dvorak wrote: > Hi, > > > We ICEd on attached testcase because we didn't properly detected the > > latch edge. Furthermore, I think the code for re-computing latches > > is somehow broken at the moment. In fix_loop_structure we have > > /* I

Re: [PATCH] Fix PR56035

2013-01-22 Thread Zdenek Dvorak
Hi, > We ICEd on attached testcase because we didn't properly detected the > latch edge. Furthermore, I think the code for re-computing latches > is somehow broken at the moment. In fix_loop_structure we have > /* If there was no latch, schedule the loop for removal. */ > if (!first

Re: [PATCH] Fix PR56035

2013-01-22 Thread Richard Biener
Marek Polacek wrote: >We ICEd on attached testcase because we didn't properly detected the >latch edge. Furthermore, I think the code for re-computing latches >is somehow broken at the moment. In fix_loop_structure we have > /* If there was no latch, schedule the loop for removal. */ >

[PATCH] Fix PR56035

2013-01-22 Thread Marek Polacek
We ICEd on attached testcase because we didn't properly detected the latch edge. Furthermore, I think the code for re-computing latches is somehow broken at the moment. In fix_loop_structure we have /* If there was no latch, schedule the loop for removal. */ if (!first_latch)