Re: Inline across -ffast-math boundary

2016-05-05 Thread Rainer Orth
Richard Biener writes: >> >> This new testcase does not pass on bare-metal configs (using newlib), >> >> because: >> >> warning: implicit declaration of function 'isnanf' >> >> [-Wimplicit-function-declaration] >> >> warning: incompatible implicit declaration of built-in

Re: Inline across -ffast-math boundary

2016-05-03 Thread Aldy Hernandez
On 05/03/2016 10:20 AM, Richard Biener wrote: On Tue, 3 May 2016, Jan Hubicka wrote: On 21/04/16 12:45, Jan Hubicka wrote: this patch implements the long promised logic to inline across -ffast-math boundary when eitehr caller or callee has no fp operations in it. This is needed to resolve

Re: Inline across -ffast-math boundary

2016-05-03 Thread Richard Biener
On Tue, 3 May 2016, Jan Hubicka wrote: > > On 21/04/16 12:45, Jan Hubicka wrote: > > > this patch implements the long promised logic to inline across -ffast-math > > > boundary when eitehr caller or callee has no fp operations in it. This is > > > needed t

Re: Inline across -ffast-math boundary

2016-05-03 Thread Richard Biener
On Tue, 3 May 2016, Szabolcs Nagy wrote: > On 21/04/16 12:45, Jan Hubicka wrote: > > this patch implements the long promised logic to inline across -ffast-math > > boundary when eitehr caller or callee has no fp operations in it. This is > > needed to resolve code quality

Re: Inline across -ffast-math boundary

2016-05-03 Thread Jan Hubicka
> On 21/04/16 12:45, Jan Hubicka wrote: > > this patch implements the long promised logic to inline across -ffast-math > > boundary when eitehr caller or callee has no fp operations in it. This is > > needed to resolve code quality regression on Firefox with LTO where >

Re: Inline across -ffast-math boundary

2016-05-03 Thread Szabolcs Nagy
On 21/04/16 12:45, Jan Hubicka wrote: > this patch implements the long promised logic to inline across -ffast-math > boundary when eitehr caller or callee has no fp operations in it. This is > needed to resolve code quality regression on Firefox with LTO where > -O3/-O2/-Ofast flags

Re: Inline across -ffast-math boundary

2016-05-03 Thread Richard Biener
1 Apr 2016, Jan Hubicka wrote: > >> >> > >> >> > Hi, > >> >> > this patch implements the long promised logic to inline across > >> >> > -ffast-math > >> >> > boundary when eitehr caller or callee has no fp operatio

Re: Inline across -ffast-math boundary

2016-05-03 Thread Christophe Lyon
> Hi, >> >> > this patch implements the long promised logic to inline across >> >> > -ffast-math >> >> > boundary when eitehr caller or callee has no fp operations in it. This >> >> > is >> >> > needed to resolve code quali

Re: Inline across -ffast-math boundary

2016-05-03 Thread Richard Biener
On Tue, 3 May 2016, Christophe Lyon wrote: > On 2 May 2016 at 19:01, Jan Hubicka <hubi...@ucw.cz> wrote: > >> On Thu, 21 Apr 2016, Jan Hubicka wrote: > >> > >> > Hi, > >> > this patch implements the long promised logic to inline across &g

Re: Inline across -ffast-math boundary

2016-05-03 Thread Christophe Lyon
On 2 May 2016 at 19:01, Jan Hubicka <hubi...@ucw.cz> wrote: >> On Thu, 21 Apr 2016, Jan Hubicka wrote: >> >> > Hi, >> > this patch implements the long promised logic to inline across -ffast-math >> > boundary when eitehr caller or callee has no fp oper

Re: Inline across -ffast-math boundary

2016-05-02 Thread Jan Hubicka
> On Mon, May 02, 2016 at 07:01:38PM +0200, Jan Hubicka wrote: > > This patch impleemnts the suggested check for presence of FP parameters. > > We can play with special casing the moves incrementally. > > This patch has broken the build: > > /home/marek/src/gcc/gcc/ipa-inline.c: In function

Re: Inline across -ffast-math boundary

2016-05-02 Thread Marek Polacek
On Mon, May 02, 2016 at 07:01:38PM +0200, Jan Hubicka wrote: > This patch impleemnts the suggested check for presence of FP parameters. > We can play with special casing the moves incrementally. This patch has broken the build: /home/marek/src/gcc/gcc/ipa-inline.c: In function ‘bool

Re: Inline across -ffast-math boundary

2016-05-02 Thread Jan Hubicka
> On Thu, 21 Apr 2016, Jan Hubicka wrote: > > > Hi, > > this patch implements the long promised logic to inline across -ffast-math > > boundary when eitehr caller or callee has no fp operations in it. This is > > needed to resolve code quality regression on Fire

Re: Inline across -ffast-math boundary

2016-04-22 Thread Richard Biener
On Thu, 21 Apr 2016, Jan Hubicka wrote: > Hi, > this patch implements the long promised logic to inline across -ffast-math > boundary when eitehr caller or callee has no fp operations in it. This is > needed to resolve code quality regression on Firefox with LTO where > -O3/-

Inline across -ffast-math boundary

2016-04-21 Thread Jan Hubicka
Hi, this patch implements the long promised logic to inline across -ffast-math boundary when eitehr caller or callee has no fp operations in it. This is needed to resolve code quality regression on Firefox with LTO where -O3/-O2/-Ofast flags are combined and we fail to inline a lot of comdats