Re: [PATCH] Fix pr57637

2013-07-30 Thread Zhenqiang Chen
On 22 July 2013 21:16, Eric Botcazou wrote: >> And if df_live is non-zero, do we need update df_lr's IN and OUT? I think >> we need another patch to make all these consistency. > > Possibly, but this would belong to another patch. I nevertheless think that > we > should set the bit in the GEN se

Re: [PATCH] Fix pr57637

2013-07-22 Thread Eric Botcazou
> And if df_live is non-zero, do we need update df_lr's IN and OUT? I think > we need another patch to make all these consistency. Possibly, but this would belong to another patch. I nevertheless think that we should set the bit in the GEN set because we'll be testing the GEN set now. The patch

Re: [PATCH] Fix pr57637

2013-07-22 Thread Zhenqiang Chen
On 22 July 2013 17:56, Eric Botcazou wrote: >> The patch is updated based the comments: it will check GEN set of LIVE >> if df_live exists. Otherwise, just give up. > > The patch is missing a ChangeLog. Otherwise it looks good, modulo: > > + /* DF_LR_BB_INFO (bb)->def does not kill the DF

Re: [PATCH] Fix pr57637

2013-07-22 Thread Eric Botcazou
> The patch is updated based the comments: it will check GEN set of LIVE > if df_live exists. Otherwise, just give up. The patch is missing a ChangeLog. Otherwise it looks good, modulo: + /* DF_LR_BB_INFO (bb)->def does not kill the DF_REF_PARTIAL and +DF_REF_CONDITIONAL defs

Re: [PATCH] Fix pr57637

2013-07-16 Thread Zhenqiang Chen
On 11 July 2013 18:31, Eric Botcazou wrote: >> Shrink-wrap optimization sinks some instructions for more >> opportunities. It uses DF_LR_BB_INFO (bb)->def to check whether BB >> clobbers SRC. But for ARM, gcc might generate cond_exec insns before >> shrink-wrapping. And DF_LR_BB_INFO (bb)->def doe

Re: [PATCH] Fix pr57637

2013-07-11 Thread Eric Botcazou
> Shrink-wrap optimization sinks some instructions for more > opportunities. It uses DF_LR_BB_INFO (bb)->def to check whether BB > clobbers SRC. But for ARM, gcc might generate cond_exec insns before > shrink-wrapping. And DF_LR_BB_INFO (bb)->def does not include def info > from cond_exec insns. So

Re: [PATCH] Fix pr57637

2013-07-05 Thread Zhenqiang Chen
On 5 July 2013 16:18, Jakub Jelinek wrote: > On Fri, Jul 05, 2013 at 04:11:00PM +0800, Zhenqiang Chen wrote: >> + FOR_BB_INSNS(bb, x) > > Just style nits: > Missing space between (. Also, please don't use uppercase names > for labels. > > Jakub Thanks for the comments. Update it as

Re: [PATCH] Fix pr57637

2013-07-05 Thread Zhenqiang Chen
Hi, The patch is updated. If there is no df_live, we still can not guarantee the correctness. So the new patch just checks the DF_INSN_DEFS. Bootstrap and no make check regression on x86-64. Bootstrap on ARM chrome book. Is it OK? Thanks! -Zhenqiang Changelog: 2013-07-05 Zhenqiang Chen

Re: [PATCH] Fix pr57637

2013-06-28 Thread Zhenqiang Chen
On 27 June 2013 21:10, Richard Earnshaw wrote: > On 27/06/13 10:02, Zhenqiang Chen wrote: >> >> Hi, >> >> Shrink-wrap optimization sinks some instructions for more >> opportunities. It uses DF_LR_BB_INFO (bb)->def to check whether BB >> clobbers SRC. But for ARM, gcc might generate cond_exec insns

Re: [PATCH] Fix pr57637

2013-06-27 Thread Richard Earnshaw
On 27/06/13 10:02, Zhenqiang Chen wrote: Hi, Shrink-wrap optimization sinks some instructions for more opportunities. It uses DF_LR_BB_INFO (bb)->def to check whether BB clobbers SRC. But for ARM, gcc might generate cond_exec insns before shrink-wrapping. And DF_LR_BB_INFO (bb)->def does not inc