Re: [patch] (un)fix loop unswitching.

[Richard Biener <richard dot guenther at gmail dot com>]

On Wed, Oct 16, 2013 at 2:56 PM, Andrew MacLeod <amacleod@redhat.com> wrote:
> Bah, bounced due to html or something. My latest email client upgrade must
> have lost the 'plain text only' attribute..
>
>
> On 10/16/2013 08:02 AM, Andrew MacLeod wrote:
>>
>> More or less  the same patch that I put in bugzilla
>> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58697
>>
>> I've bootstrapped it and run it through regressions.  Jakubs test case
>> fails again like it did before my changes, so it no longer masks the
>> problem.
>>
>> Loop unswitching was calling estimate_loop_iterations_int, which for
>> various annoying reasons ended up in cfgloop and called
>> get_estimate_loop_iterations, so it wasn't doing the calculations. Hopefully
>> its clearer now.
>>
>> The end result now is that cfgloop.c has 4  accessor routines which just
>> retrieve info:
>> get_estimated_loop_iterations, get_estimated_loop_iterations_int,
>> get_max_loop_iterations, and get_max_loop_iterations_int.
>>
>> tree-ssa-loop-niter.c has 4 routines which create the information first,
>> if they can, then retrieve the information.
>> estimated_loop_iterations, estimated_loop_iterations_int,
>> max_loop_iterations, and max_loop_iterations_int.
>>
>> all the cfg*  routines call the 'get_' variety, and all the tree-ssa
>> routines call the non 'get_' variety to calculate things if need be.
>>
>> Bootstraps on x86_64-unknown-linux-gnu and no regressions. OK? I

Ok.

Thanks,
Richard.

presume
>> Jakub's test case isn't in the testsuite anywhere then?  I presume it would
>> go in along with a fix to the original problem
>>
>> Andrew
>
>