Chia-Che Tsai
e01769337c
Detecting invalid user memory given as system call inputs (#210)
|
6 years ago | |
|---|---|---|
| .. | ||
| .gitignore | 7 years ago | |
| BLOCKED | 6 years ago | |
| FLAKY | 6 years ago | |
| Makefile | 6 years ago | |
| Makefile.testcases | 7 years ago | |
| PASSED | 6 years ago | |
| README | 7 years ago | |
| TIMEOUTS | 6 years ago | |
| block_tests.awk | 7 years ago | |
| edit_sys_tests.awk | 7 years ago | |
| fetch.py | 6 years ago | |
| manifest.template | 6 years ago | |
| passed_tests_only.awk | 7 years ago | |
| run_in_graphene.awk | 7 years ago | |
| runltp.patch | 7 years ago | |
| syscalls.sh | 7 years ago | |
README
LTP FOR GRAPHENE
----------------
Running LTP test cases in Graphene :
1) run 'make' in LibOS/shim/test/apps/ltp
2) cd Graphene/LibOS/shim/test/regression
3) make regression
The above command will start running LTP system call test cases in Graphene.
It works in 2 steps:
1) Creates a testfile (Graphene/LibOS/shim/test/apps/ltp/syscalls.graphene) with all the LTP system call test cases to run, excluding the ones mentioned in Graphene/LibOS/shim/test/apps/ltp/BLOCKED. “BLOCKED” contains all the test cases which are hanging in the Graphene and thus are excluded from regression tests.
2) Calls Graphene/LibOS/shim/test/apps/ltp/fetch.py It runs the LTP test cases and display output on the terminal.
Some LTP testcases hangs in Graphene. This python script sets a timeout for each testcase referring the file 'TIMEOUTS'. Default timeout is set 20 seconds in the script.
It also shows, ‘new passed testcases’ (passed testcases which are not in Graphene/LibOS/shim/test/apps/ltp/PASSED) and ‘new failed testcases’ (failed testcases which are in ‘PASSED’).
It also creates files ‘run_output’, ‘Failed.csv’, ‘Passed.csv’, ‘Blocked.csv’ under Graphene/LibOS/shim/test/apps/ltp for reference.
Results are of 4 types :
[Pass] -> Testcase passed
[Failed] -> Testcase failed
[Hanged] -> Testcase hanged and thereafter killed
[Broken] -> Either system call is not implemented in Graphene or test preparation failed
To run one particular testcase, execute following :
~/Runtime/pal /opt/ltp/testcases/bin/
In this way, one can debug one particular syscall testcase.
As you fix bugs, keep updating PASSED BLOCKED lists.
----------------
Running LTP test cases in Graphene :
1) run 'make' in LibOS/shim/test/apps/ltp
2) cd Graphene/LibOS/shim/test/regression
3) make regression
The above command will start running LTP system call test cases in Graphene.
It works in 2 steps:
1) Creates a testfile (Graphene/LibOS/shim/test/apps/ltp/syscalls.graphene) with all the LTP system call test cases to run, excluding the ones mentioned in Graphene/LibOS/shim/test/apps/ltp/BLOCKED. “BLOCKED” contains all the test cases which are hanging in the Graphene and thus are excluded from regression tests.
2) Calls Graphene/LibOS/shim/test/apps/ltp/fetch.py It runs the LTP test cases and display output on the terminal.
Some LTP testcases hangs in Graphene. This python script sets a timeout for each testcase referring the file 'TIMEOUTS'. Default timeout is set 20 seconds in the script.
It also shows, ‘new passed testcases’ (passed testcases which are not in Graphene/LibOS/shim/test/apps/ltp/PASSED) and ‘new failed testcases’ (failed testcases which are in ‘PASSED’).
It also creates files ‘run_output’, ‘Failed.csv’, ‘Passed.csv’, ‘Blocked.csv’ under Graphene/LibOS/shim/test/apps/ltp for reference.
Results are of 4 types :
[Pass] -> Testcase passed
[Failed] -> Testcase failed
[Hanged] -> Testcase hanged and thereafter killed
[Broken] -> Either system call is not implemented in Graphene or test preparation failed
To run one particular testcase, execute following :
~/Runtime/pal /opt/ltp/testcases/bin/
In this way, one can debug one particular syscall testcase.
As you fix bugs, keep updating PASSED BLOCKED lists.