Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Tue, Apr 05, 2016 at 11:48:29AM +0200, Richard Biener wrote: > Like this? Yeah, thanks. > 2016-04-05 Richard Biener > > * dse.c (struct store_info): Remove alias_set member. > (struct read_info_type): Likewise. > (clear_alias_group,

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Tue, 5 Apr 2016, Jakub Jelinek wrote: > On Tue, Apr 05, 2016 at 11:05:44AM +0200, Richard Biener wrote: > > True. By simple constant propagation I can remove a lot of code. > > > > I'm going to bootstrap / test the following - is this ok for trunk > > now (I'm going to write a better

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Tue, Apr 05, 2016 at 11:05:44AM +0200, Richard Biener wrote: > True. By simple constant propagation I can remove a lot of code. > > I'm going to bootstrap / test the following - is this ok for trunk > now (I'm going to write a better changelog). LGTM with better Changelog, though I have

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Mon, 4 Apr 2016, Jakub Jelinek wrote: > On Mon, Apr 04, 2016 at 11:24:41AM +0200, Richard Biener wrote: > > On Fri, 1 Apr 2016, Bernd Schmidt wrote: > > > > > On 04/01/2016 11:08 AM, Richard Biener wrote: > > > > { > > > > ! if (canon_true_dependence (s_info->mem, > > > > !

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Mon, Apr 04, 2016 at 11:24:41AM +0200, Richard Biener wrote: > On Fri, 1 Apr 2016, Bernd Schmidt wrote: > > > On 04/01/2016 11:08 AM, Richard Biener wrote: > > > { > > > ! if (canon_true_dependence (s_info->mem, > > > ! GET_MODE

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Mon, 4 Apr 2016, Richard Biener wrote: > On Fri, 1 Apr 2016, Bernd Schmidt wrote: > > > On 04/01/2016 11:08 AM, Richard Biener wrote: > > > { > > > ! if (canon_true_dependence (s_info->mem, > > > ! GET_MODE (s_info->mem), > > > !

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Fri, 1 Apr 2016, Bernd Schmidt wrote: > On 04/01/2016 11:08 AM, Richard Biener wrote: > > { > > ! if (canon_true_dependence (s_info->mem, > > !GET_MODE (s_info->mem), > > !s_info->mem_addr, > > !

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On April 1, 2016 5:26:21 PM GMT+02:00, Bernd Schmidt wrote: >On 04/01/2016 11:08 AM, Richard Biener wrote: >> { >> ! if (canon_true_dependence (s_info->mem, >> ! GET_MODE (s_info->mem), >> !

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On 04/01/2016 03:44 AM, Richard Biener wrote: On Fri, 1 Apr 2016, Jakub Jelinek wrote: On Fri, Apr 01, 2016 at 11:08:09AM +0200, Richard Biener wrote: RTL DSE uses true_dependence to see whether a store may be killed by anothe store - that's obviously broken. The following patch makes it

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On 04/01/2016 11:08 AM, Richard Biener wrote: { ! if (canon_true_dependence (s_info->mem, !GET_MODE (s_info->mem), !s_info->mem_addr, !mem, mem_addr)) {

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On 04/01/2016 05:05 PM, Jakub Jelinek wrote: with my usual pair of rtl,yes checking bootstraps/regtests (x86_64-linux and i686-linux, former one with ada, latter without), both without your patch and with the patch. Without the patch got 66555 successful replace_reads, with your patch only

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Fri, Apr 01, 2016 at 11:44:16AM +0200, Richard Biener wrote: > On Fri, 1 Apr 2016, Jakub Jelinek wrote: > > > On Fri, Apr 01, 2016 at 11:08:09AM +0200, Richard Biener wrote: > > > > > > RTL DSE uses true_dependence to see whether a store may be killed by > > > anothe store - that's obviously

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Fri, 1 Apr 2016, Jakub Jelinek wrote: > On Fri, Apr 01, 2016 at 11:08:09AM +0200, Richard Biener wrote: > > > > RTL DSE uses true_dependence to see whether a store may be killed by > > anothe store - that's obviously broken. The following patch makes > > it use output_dependence instead

Re: [PATCH] Fix PR70484, RTL DSE using wrong dependence check

On Fri, Apr 01, 2016 at 11:08:09AM +0200, Richard Biener wrote: > > RTL DSE uses true_dependence to see whether a store may be killed by > anothe store - that's obviously broken. The following patch makes > it use output_dependence instead (introducing a canon_ variant of that). I think it

[PATCH] Fix PR70484, RTL DSE using wrong dependence check

RTL DSE uses true_dependence to see whether a store may be killed by anothe store - that's obviously broken. The following patch makes it use output_dependence instead (introducing a canon_ variant of that). Bootstrap & regtest running on x86_64-unknown-linux-gnu. Ok? Thanks, Richard.