future branch

[Bernhard Reutner-Fischer]

On 9 February 2012 19:24, Bernhard Reutner-Fischer
<rep.dot.nop at gmail.com> wrote:
> On 8 February 2012 16:15, Bernhard Reutner-Fischer
> <rep.dot.nop at gmail.com> wrote:
>>>>> On Thu, Feb 2, 2012 at 1:38 PM, Peter Mazinger <ps.m at gmx.net> wrote:
>>
>>>>>>>> fine, tell me about problems you encounter...
>>
>> Peter,
>> I am merging your future branch into master ATM. A couple of questions:
>>
>> db97b8bb3859fb28e2b36934e48de90a17c06222 removed the guard for
>> __NR_socketcall but there are arches that have individual "socketcall"
>> impls (like x86_64, alpha for sure) and i don't see __NR_socketcall
>> for e.g. xtensa, c6x either. reinstalled.
>
> re-added guard and pushed to the future branch.
>
>> 1ee4b98476a83d5f139e863d6f71a9356a046edd can you explain why always
>> hardcoding the interp path is a good idea?
>> You seem to lose (and have lost in the past) ", or (at your option)
>> any later version." from touching boilerplates, which is a bad thing.
>
> I could use your help with f69319d5a7d3a3ccb46b28ee2b0fd9053c6415ac from master:
> gcc -Wl,-EL -shared -Wl,--warn-common -Wl,--warn-once -Wl,-z,combreloc
> -Wl,-z,relro -Wl,-z,now -Wl,-z,defs -Wl,-s  -Wl,-fini,dl_cleanup
> -Wl,-soname=libdl.so.0 -nostdlib -o lib/libdl-0.9.34-git.so
> -Wl,--whole-archive ldso/libdl/libdl_so.a -Wl,--no-whole-archive
> ./lib/interp.os -L./lib ./lib/libc.so ./lib/ld64-uClibc.so.0
> /usr/lib/gcc/x86_64-linux-gnu/4.6/libgcc.a
> /usr/bin/ld: ldso/libdl/libdl_so.a(libdl.oS): relocation R_X86_64_PC32
> against undefined hidden symbol `__pthread_mutex_unlock' can not be
> used when making a shared object
> Thoughts?

I see this with NPTL. Trivial testcase:
#include <dlfcn.h>
int main(void) {return !!!dlopen("lib/libc.so.0",0);}