2017-05-16 21:16+0200, David Hildenbrand: > On 16.05.2017 15:35, Radim Krčmář wrote: > > 2017-05-12 12:58+0200, David Hildenbrand: > >> Test if the general infrastructure is working. The test will fail until > >> we have proper sclp console output. > >> > >> Signed-off-by: David Hildenbrand <david@xxxxxxxxxx> > >> --- > >> diff --git a/s390x/selftest.c b/s390x/selftest.c > >> @@ -0,0 +1,31 @@ > >> +int main(int argc, char**argv) > >> +{ > >> + report_prefix_push("selftest"); > >> + > >> + if (argc != 3) > >> + report_abort("Wrong number of arguments"); > >> + > >> + if (strcmp(argv[0], "s390x/selftest.elf") != 0) > >> + report_abort("wrong program name"); > > > > This is going to fail when executed as a standalone test (argv[0] would > > be a temp file name). No point in checking, IMO. > > That name is determined during compile time, so it shouldn't change, or > am i wrong? But if you prefer, I can drop it. You are right, only x86 uses the actual kernel file name, sorry. >>> + if (strcmp(argv[1], "test") != 0) >>> + report_abort("wrong parameter value"); >>> + if (strcmp(argv[2], "123") != 0) >>> + report_abort("wrong parameter value"); >>> + >>> + report("test true", true, 0); >> ^ >> It seems you'll be doing v2 -- please remove this zero in it. > > Thanks, if that would throw an error during compile time like printf, it > would be perfect. > > Yes, I'll resend! Thinking more about it, the test would look better if we used report() instead of if()+report_abort() and a final report, something like: report("argv[0] == PROGNAME", !strcmp(argv[0], "s390x/selftest.elf")); report("argv[1] == test", !strcmp(argv[1], "test")); report("argv[2] == 123", !strcmp(argv[2], "123")); This is also a nitpick and I'll gladly accept the original version. :)