From: Peng H. <pe...@cn...> - 2011-03-23 05:46:27
|
Murlin Wenzel said the following on 2011-3-23 4:06: >>>> diff --git a/testcases/kernel/containers/pidns/pidns05.c >>>> b/testcases/kernel/containers/pidns/pidns05.c >>>> index 8a08dce..9ffc946 100644 >>>> --- a/testcases/kernel/containers/pidns/pidns05.c >>>> +++ b/testcases/kernel/containers/pidns/pidns05.c >>>> @@ -186,13 +186,8 @@ void kill_nested_containers() >>>> >>>> /* Loops through the containers created to exit from sleep() */ >>>> for (i = 0; i < MAX_DEPTH; i++) { >>>> - if (waitpid(pids[i], &status, 0) == -1) >>>> - tst_resm(TFAIL|TERRNO, "waitpid(%d, ...) failed", >>>> - pids[i]); >>>> - else { >>>> - kill(pids[i], SIGKILL); >>>> - waitpid(pids[i], &status, 0); >>>> - } >>>> + kill(pids[i], SIGKILL); >>>> + waitpid(pids[i], &status, 0); >>>> } >>>> } >>>> >>>> @@ -222,7 +217,10 @@ int main(int argc, char *argv[]) >>>> if (waitpid(pid, &status, 0) == -1) { >>>> perror("wait failed"); >>>> } >>>> - exit(status); >>>> + if (WIFEXITED(status)) >>>> + exit(WEXITSTATUS(status)); >>>> + else >>>> + exit(status); >>>> } >>>> >>>> /* To make all the containers share the same PGID as its parent */ >>>> @@ -255,4 +253,4 @@ int main(int argc, char *argv[]) >>>> cleanup(); >>>> >>>> tst_exit(); >>>> -} >>>> \ No newline at end of file >>>> +} >>> >>> >>> I applied this patch to 20110228 release and the 'containers' test now runs >> without hanging. There are several 'TBROK' and 'TFAIL' errors, but at least >> the test runs to completion now. >>> >> >> When executing the case after applied this patch, the log is as following: >> >> pidns05 0 TINFO : 5 Nested Containers are created >> pidns05 1 TPASS : The number of containers killed are 2 >> >> Would you share your test log? >> > Attached is latest test log of containers test. I think the only actual failure now is a client test of some sort which is entirely possible since I doubt I have everything configured for network testing. > In your test log, the case pidns05 does not output 'TBROK' or 'TFAIL' errors。 This looks like the patch can work correctly:) -- Best Regards, Peng Haitao |