Re: [PATCH] Fold x86 bzhi with zero last operand (PR rtl-optimization/87817)

2018-11-14 Thread Richard Biener
On Wed, 14 Nov 2018, Jakub Jelinek wrote: > On Wed, Nov 14, 2018 at 09:29:15AM +0100, Richard Biener wrote: > > So where do the zero_extracts come from? Can we somehow avoid > > zero-sized bit-extracts at expansion time by folding them to zero? > > The following patch implements the missing fold

Re: [PATCH] Fold x86 bzhi with zero last operand (PR rtl-optimization/87817)

2018-11-14 Thread Uros Bizjak
On Wed, Nov 14, 2018 at 1:21 PM Jakub Jelinek wrote: > > On Wed, Nov 14, 2018 at 09:29:15AM +0100, Richard Biener wrote: > > So where do the zero_extracts come from? Can we somehow avoid > > zero-sized bit-extracts at expansion time by folding them to zero? > > The following patch implements the

[PATCH] Fold x86 bzhi with zero last operand (PR rtl-optimization/87817)

2018-11-14 Thread Jakub Jelinek
On Wed, Nov 14, 2018 at 09:29:15AM +0100, Richard Biener wrote: > So where do the zero_extracts come from? Can we somehow avoid > zero-sized bit-extracts at expansion time by folding them to zero? The following patch implements the missing folding. Note, as I've mentioned, it is just an optimizat