On Fri, Nov 25, 2016 at 10:15:25AM +0100, Richard Biener wrote: > On Thu, Nov 24, 2016 at 5:14 PM, Segher Boessenkool > <seg...@kernel.crashing.org> wrote: > > On Thu, Nov 24, 2016 at 08:48:04AM -0700, Jeff Law wrote: > >> On 11/24/2016 07:53 AM, Segher Boessenkool wrote: > >> > > >> >That we compare different kinds of costs (which really has no meaning at > >> >all, it's a heuristic at best) in various places is a known problem, not > >> >a regression. > >> But the problems with the costing system exhibit themselves as a code > >> quality regression. In the end that's what the end-users see -- a > >> regression in the quality of the code GCC generates. > > > > Yes, exactly -- and I fear this all-encompassing change will cause just > > such a regression for many users. Tests are running, will know more > > later today (or tomorrow). > > > > The PR is about a very specific problem; the patch is not. The patch > > is not a bug fix. If we allow anything that "makes things better" in > > stage 3, what make it different from stage 1 then? > > That's a good question ;) The stage 3 definition has a loophole via > "go file a bug about feature X, then it's a bugfix!". > > I'm all open for a more sensible definition, like constraining the kind > of non-regression fixes that we want to allow, but I fear the most > sensible option would be to simply ditch the notion of different > "stages" and make it "general development" and "regression fixing". > (though if you try hard enough and go back in time you'll find that > almost all non-enhancement bugs are regressions in some sense)
The scale goes: early stage 1, anything goes; ...; until stage 4, only very narrow regression fixes are allowed. Let's try to keep that spirit, and not behave like politicians following the "rules" (or not). > And yes, current stage3 still feels too much like stage1 ;) Yes, very much so. Well, at least trunk bootstraps on more targets now. -- So IMNSHO this rtx costing change belongs in early stage 1, and should be reverted. If ifcvt should use full rtx cost instead of rtx_src_cost, fix *that*, that is a much more local change. And even then, test on more targets please. Segher
