From: <cz...@li...> - 2012-03-09 05:08:10
|
Hi All Not sure if anyone else has seen this yet but I just did my first final run on an mls 64 bit x86_64 architecture platform. I had a clean run except for two testcases in syscalls. They are odd in that their error and failure appear to be related to the calls directly in front of them. I have not seen these errors on previous version runs such as ks22 with a late February audit-test drop. The test are tgkill__mac_process_sigkill_fail_subj_domby_obj, and tkill__mac_process_sigstop_success_subj_eq_obj (numbers 412 and 413 on an x86_64 mls run). Each test will pass when run individually every time. 412 gets an error "unexpected test result and 413 fails with "missing syscall record". If a "run.bash 412 413" is executed so only 412 and 413 are run together 412 will pass and 413 will fail with "missing syscall record" If 411 (tgkill__mac_process_sigkill_success_subj_eq_obj) and 412 are run together 412 will usually show the error however I did get one test run with this combination where both passed. If 411 is run followed by 413 they will both pass each time. So it appears that running 411 in front of 412 will cause 412 to error and running 412 in front of 413 will cause 413 to fail. I need to look into this further but I was just curious if anyone else has seen this behavior on the current build or any previous? Other than this it was a spotless run including ltp. Thanks Jim |