[uclibc-ng-devel] [ork1] issue with Binutils 2.32 and gcc 9.x

Stafford Horne shorne at gmail.com
Mon Aug 26 23:02:32 CEST 2019


On Sat, Aug 24, 2019 at 02:45:59PM +0200, Romain Naour wrote:
> Hi Stafford,
> 
> Le 23/08/2019 à 23:01, Stafford Horne a écrit :
> > On Fri, Aug 23, 2019 at 05:23:12PM +0200, Romain Naour wrote:
> >> Hi Stafford,
> >>
> >> Le 23/08/2019 à 15:47, Stafford Horne a écrit :
> >>> On Fri, Aug 16, 2019 at 07:35:33PM +0200, Romain Naour wrote:
> >>>> Hi Waldemar,
> >>>>
> >>>> I discovered an issue with uClibc and binutils 2.32 and gcc 9.1 or 9.2.
> >>>
> >>> Hi Romain,
> >>>
> >>> Thank you for reporting and doing the initial investigation.  Sorry, I haven't
> >>> built uclibc-ng for a while.
> >>
> >> You're welcome, your help is very appreciated.
> >>>
> >>> Have you been able to make much progress?
> >>
> >> Yes and no...
> > 
> > I was able to reproduce the issue and did a temporary fix by just deleting the
> > __syscall_error from or1k_clone.S (2 places).  This fixed the build but probably
> > not something we really want to do.  But it does point out that the only bad
> > symbol is the one in or1k_clone.S.
> 
> I don't get it, there is only SYSCALL_ERROR_NAME from or1k_clone.S.
> Maybe post the patch diff here.

Sorry I meant SYSCALL_ERROR_NAME which defined __syscall_error.  It actually
defines it incorrectly as it looks for a definition of PIC, but in uclibc-ng
__PIC__ is defined.

> > 
> > There is a glibc port I am working on getting working right now:
> > 
> >   https://github.com/stffrdhrn/or1k-glibc/blob/upstream-rebase/sysdeps/unix/sysv/linux/or1k/or1k_clone.S
> > 
> > It seems to have the same code.  But it does not have this problem.  The macro
> > for SYSCALL_ERROR_NAME is probably getting evaluated differently in glibc vs
> > uclibc-ng.
> 
> Maybe something doesn't work with the way we define SYSCALL_ERROR_NAME
> https://cgit.uclibc-ng.org/cgi/cgit/uclibc-ng.git/tree/libc/sysdeps/linux/or1k/sysdep.h#n67
> > 
> > This fix should not be too hard.  SYSCALL_ERROR_NAME is only being used on
> > or1k_clone.S.  Let's try to see what is the right thing to replace it with for
> > uclibc-ng.
> 
> I haven't looked in depth in the code, I'm not sure how to fix this :-/
>

No problems,

I just posted a patch which I think fixes this can you give it a try?

 https://mailman.uclibc-ng.org/pipermail/devel/2019-August/001892.html

-Stafford


More information about the devel mailing list