Re: [LTP] [003][RFC] [TESTING linux-2.6.30] TOMOYO Linux and pathname-basedsecurity

[Subrata Modak <subrata@linux.vnet.ibm.com>]

On Fri, 2009-06-19 at 15:37 +0900, Tetsuo Handa wrote: 
> Hello.
> 
> Subrata Modak wrote:
> > Do you feel the above information is sufficient for your to write tests
> > for LTP ? Please feel free to ask us questions regarding anything to LTP
> > tests.
> Yes. I'm reading it now. Some random topics I noticed...
> 
> 
> 
> In http://ltp.sourceforge.net/ltpman.php ...
> The "LTP test driver PAN" points to
> http://ltp.sourceforge.net/ltpmantemp.php?file=man1/pan.html
> but the content seems empty.
> Same result for other topics.
> Is ltpmantemp.php working?

I need to check out, as, i rarely give a look to the documentation stuff
on the website due to lack of time ;-)

> 
> 
> 
> In "Developing tests in ANSI-C using LTP test harness API's" at
> http://ltp.sourceforge.net/documentation/how-to/ltp.php ...
> 
> > TFAIL
> >    Print the test case result. The tst_resm/tst_res function prints related messages to the screen indicating a test failure. The test will exit with a non-zero exit value when tst_exit function is called.
> >
> >    test01    0  PASS  :  uname command exited with error code = 255
> >
> I think this is "FAIL  :  uname command exited with error code = 255".
> 
> >TWARN
> >    Print warning messages. For example, when system resources are low.
> >
> >    test01    0  PASS  :  System low on memory, test case might fail.
> I think this is "WARN  :  System low on memory, test case might fail."
> 
> 
> 
> There are several "amp;" which I think it meant '&'.
> 
> 
> 
> In "Recommendation on coding style"...
> 
> > Check for return values from system calls. Report reason appropriately using the tst_brkm() API. The recommended style is as follows:
> > Use spaces instead of tab characters. Typically four spaces may be used instead of a tab character.
> 
> Ouch! I'm writing testcases in accordance with script/checkpatch.pl in Linux

Actually, you should do according to script/checkpatch.pl. Recently we
agreed to that. Alas, we have too many code to fix for this :-(

> kernel source tree. I expect that those who update/maintain testcases for
> kernel source code likely have some experience in writing kernel patches.
> Is it preferable to change codyng style in accordance with LTP?
> 
> 
> 
> http://ltp.cvs.sourceforge.net/viewvc/ltp/ltp/testcases/kernel/selinux-testsuite/
> seems to be moved to
> http://ltp.cvs.sourceforge.net/viewvc/ltp/ltp/testcases/kernel/security/selinux-testsuite/
> 
> 
> 
> Many programs under
> http://ltp.cvs.sourceforge.net/viewvc/ltp/ltp/testcases/kernel/security/ seems
> not using harness APIs...
> Are these programs intended for executed by general LTP users?

Correct, there are many programs within LTP which do not use the
harness. The main reason being:

Porting would have been a huge exercise, involving huge man hours. And
simply people did not have time. So, looking into the usefulness of
those tests, they were added to LTP. But, we are careful about the
incremental addition of tests in LTP, in which case, it is easy to
adhere to the harness specific standards. And hence we insist on using
those harness when new tests keep coming in.

And hence, any test without harness can be added to LTP if it returns
proper exit code, thus enabling the LTP provisioning engine to actually
report back PASS/FAIL.

Regards--
Subrata



------------------------------------------------------------------------------
Are you an open source citizen? Join us for the Open Source Bridge conference!
Portland, OR, June 17-19. Two days of sessions, one day of unconference: $250.
Need another reason to go? 24-hour hacker lounge. Register today!
http://ad.doubleclick.net/clk;215844324;13503038;v?http://opensourcebridge.org
_______________________________________________
Ltp-list mailing list
Ltp-list@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ltp-list