[LTP] get_mempolicy failing to compile

[Jan Stancek <jstancek@redhat.com>]

----- Original Message -----
> From: "$rik@nth" <srikanth007m@gmail.com>
> To: "Jan Stancek" <jstancek@redhat.com>
> Cc: ltp@lists.linux.it
> Sent: Wednesday, 27 January, 2016 9:39:26 AM
> Subject: Re: [LTP] get_mempolicy failing to compile
> 
> Now it is failing here.. on the same setup i am able to compile the
> sample numa program statically.
> 
> make[4]: Leaving directory
> '/home/linaro/Srikanth/ltp/testcases/kernel/include'
> gcc -static -g -O2 -fno-strict-aliasing -pipe -Wall -W
> -Wold-style-definition
> -I/home/linaro/Srikanth/ltp/testcases/kernel/include -D_GNU_SOURCE
> -DNUMA_VERSION1_COMPATIBILITY -I../../../include -I../../../include
> -static -pthread  -L../../../lib  numa_node_size.c   -lltp -lnuma -o
> numa_node_size
> numa_node_size.c: In function ?fmt_mem?:
> numa_node_size.c:44:10: warning: comparison between signed and
> unsigned integer expressions [-Wsign-compare]
>   if (mem == -1L)
>           ^
> /usr/lib/gcc/aarch64-linux-gnu/4.9/../../../aarch64-linux-gnu/libnuma.a(libnuma.o):(.bss+0x38):
> multiple definition of `numa_exit_on_error'

That's likely because "numa_exit_on_error" is supposed to be defined
by libnuma library:
  http://linux.die.net/man/3/numa_exit_on_error

Test defines it too, which creates conflict. We should consider removing it,
(and that it doesn't break older libnuma releases):

diff --git a/testcases/kernel/numa/numa_node_size.c b/testcases/kernel/numa/numa_node_size.c
index 6338bc7..1375193 100644
--- a/testcases/kernel/numa/numa_node_size.c
+++ b/testcases/kernel/numa/numa_node_size.c
@@ -38,7 +38,7 @@
 #include <numa.h>
 #endif
 
-int numa_exit_on_error = 0;
 char *fmt_mem(unsigned long long mem, char *buf)
 {
        if (mem == -1L)

Regards,
Jan

> /tmp/ccRABETR.o:(.bss+0x0): first defined here
> /usr/lib/gcc/aarch64-linux-gnu/4.9/../../../aarch64-linux-gnu/libnuma.a(affinity.o):
> In function `affinity_ip':
> (.text+0x6dc): warning: Using 'getaddrinfo' in statically linked
> applications requires at runtime the shared libraries from the glibc
> version used for linking
> collect2: error: ld returned 1 exit status
> <builtin>: recipe for target 'numa_node_size' failed
> make[3]: *** [numa_node_size] Error 1
> make[3]: Leaving directory '/home/linaro/Srikanth/ltp/testcases/kernel/numa'
> ../../include/mk/generic_trunk_target.inc:93: recipe for target 'all' failed
> make[2]: *** [all] Error 2
> make[2]: Leaving directory '/home/linaro/Srikanth/ltp/testcases/kernel'
> ../include/mk/generic_trunk_target.inc:93: recipe for target 'all' failed
> make[1]: *** [all] Error 2
> make[1]: Leaving directory '/home/linaro/Srikanth/ltp/testcases'
> Makefile:114: recipe for target 'testcases-all' failed
> make: *** [testcases-all] Error 2
> 
> On Fri, Jan 22, 2016 at 12:10 PM, $rik@nth <srikanth007m@gmail.com> wrote:
> > I did an softlink to arm-linux gcc after my initial mail.. So on
> > latest one you will be not seeing.
> >
> > I will try on another development board. This is my 3rd board having
> > problem :(
> >
> > On Fri, Jan 22, 2016 at 2:00 PM, Jan Stancek <jstancek@redhat.com> wrote:
> >>
> >>
> >> ----- Original Message -----
> >>> From: "$rik@nth" <srikanth007m@gmail.com>
> >>> To: "Jan Stancek" <jstancek@redhat.com>
> >>> Sent: Friday, 22 January, 2016 5:25:54 AM
> >>> Subject: Re: [LTP] get_mempolicy failing to compile
> >>>
> >>> My environment is ARM ubuntu and all my libs and packages are in AMR.
> >>
> >> config.log you gave us is from x86_64 system named "hyd-lnxbld84".
> >>
> >>> Yes able to compile the source  static as well
> >>> linaro@linaro-alip:~/Srikanth$ gcc -static numatest.c -lnuma
> >>
> >> This is a different system, isn't it?
> >>
> >> What I meant was "hyd-lnxbld84" system and gcc you use to
> >> compile LTP:
> >>   arm-linux-gnueabihf-gcc -static numatest.c -lnuma
> >
> >
> >
> > --
> > Thanks & Regards,
> > M.Srikanth Kumar.
> 
> 
> 
> --
> Thanks & Regards,
> M.Srikanth Kumar.
>