On Monday 02 November 2009 08:33:55 Subrata Modak wrote:
> gcc -g -O2 -g -O2 -fno-strict-aliasing -pipe -Wall
> -I/root/subrata/ltp/ltp-full-20091031/testcases/kernel/include
> -I../../../../include -I../../../../include   -L../../../../lib
> ptrace04.c   -lltp -o ptrace04
> ptrace03.c: In function 'main':
> ptrace03.c:182: warning: format '%d' expects type 'int', but argument 3
> has type 'long int'
> ptrace04.c: In function 'compare_registers':
> ptrace04.c:50: error: storage size of 'pt_regs' isn't known
> ptrace04.c:50: warning: unused variable 'pt_regs'
> make[4]: *** [ptrace04] Error 1

the ptrace stuff has always been fairly unfriendly to cross-arch work.  we'll 
probably have to add header tests for a bunch of files and unify it in a local 
ptrace.h header.

> gcc -g -O2 -g -O2 -fno-strict-aliasing -pipe -Wall
> -I../../../../include -I../../../../include   -L../../../../lib
> tomoyo_file_test.c   -o tomoyo_file_test
> /tmp/ccvzdsNd.o: In function `main':
> /root/subrata/ltp/ltp-full-20091031/testcases/kernel/security/tomoyo/newns.
> c:43: undefined reference to `clone' collect2: ld returned 1 exit status

these test cases need to be changed to the new common ltp clone code.  as part 
of that, we should probably add a macro that changes clone() to a compile 
error so that people stop using it by accident
#define clone(...) DO_NOT_USE_CLONE
-mike