On Fri, 2016-03-11 at 07:34 +0900, Oleg Endo wrote: > > On Mar 11, 2016, at 7:15 AM, John Paul Adrian Glaubitz < > glaub...@physik.fu-berlin.de> wrote: > > > Hi Aurelien! > > > > (CC'ing Helmut Grohne since he has been seeing this issue in > > rebootstrap > > as well as Oleg Endo and Kaz Kojima as they are most likely > > interested > > in fixing the issue in gcc being gcc/sh upstream developers) > > > > On 03/10/2016 11:02 PM, Aurelien Jarno wrote: > > > The glibc package in version 2.22 which has been uploaded to sid > > > a few > > > days ago fails to build on sh4 [1]. I have tracked it down to a > > > missing > > > target specific implementation of __builtin_trap() in GCC. I have > > > workarounded the issue by passing the -fno-delete-null-pointer > > > -checks > > > option to GCC. This will be available in the next upload. > > > > Wow, thanks a lot for tracking this down. Helmut Grohne has been > > seeing > > this issue in rebootstrap for a long time now and it was not until > > glibc > > 2.22 was uploaded when we saw this issue arise for native builds. > > > > I (and probably Helmut as well) are very glad this has been > > debugged > > now! I owe you a DebConf beer :). > > > > > It would be better instead if this can be fixed in GCC. An > > > example of > > > such a patch (for hppa) is available [3]. > > > > Great. Maybe Oleg and Kaz could have a look at that. > > > > @Oleg: Should I file a gcc bug report? > > Yes, please.
I went ahead and created a GCC PR https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70216 Let's take further discussions there. Cheers, Oleg
