Hi Peter,
Peter Seiderer wrote,

> Hello Waldemar,
> 
> On Fri, 21 Jan 2022 11:54:01 +0100, Peter Seiderer <[email protected]> wrote:
> 
> > Hello Waldemar,
> >
> > On Fri, 21 Jan 2022 11:04:57 +0100, Waldemar Brodkorb <[email protected]> 
> > wrote:
> >
> > > Hi Peter,
> > >
> > > sorry about coming so late up with a regression.
> >
> > No problem...
> >
> > >
> > > I get following failure when compiling the defconfig 
> > > stm32f429_disco_xip_defconfig in buildroot with your patch:
> > >
> > > In file included from libc/inet/_res_state.c:8:
> > > libc/inet/resolv.c: In function '__res_state':
> > > libc/inet/resolv.c:3903:15: error: '__resp' undeclared (first use in this 
> > > function); did you mean '_res'?
> > >  3903 |        return __resp;
> > >       |               ^~~~~~
> > >       |               _res
> > > libc/inet/resolv.c:3903:15: note: each undeclared identifier is reported 
> > > only once for each function it appears in
> > > libc/inet/resolv.c:3904:1: warning: control reaches end of non-void 
> > > function [-Wreturn-type]
> > >  3904 | }
> > >       | ^
> > > make[1]: *** [Makerules:368: libc/inet/_res_state.o] Error 1
> > > make[1]: Leaving directory 
> > > '/home/wbx/buildroot/output/build/uclibc-1.0.39'
> > > make: *** [package/pkg-generic.mk:295: 
> > > /home/wbx/buildroot/output/build/uclibc-1.0.39/.stamp_built] Error 2
> > >
> > > In this case Linuxthreads are used, because the target is noMMU and FLAT 
> > > binary format.
> > >
> > > Any idea?
> >
> > No...not yet, but will take a look at it ;-)
> 
> Quick (compile fix, not run-time tested):
> 
> index cf170fb..92a65d0 100644
> --- a/libc/inet/resolv.c
> +++ b/libc/inet/resolv.c
> @@ -3897,6 +3897,7 @@ res_ninit(res_state statp)
> 
>  #ifdef L_res_state
>  # if !defined __UCLIBC_HAS_TLS__
> +extern struct __res_state *__resp;
>  struct __res_state *
>  __res_state (void)
>  {

Thanks, that worked. Tested on stm board and in Qemu.
Pushed it with your SoB.

best regards
 Waldemar
> 
_______________________________________________
devel mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to