Re: [LTP] proc01 failure

[Mike F. <va...@ge...>]

On Thursday 22 May 2008, Cai Qian wrote:
> Hi,
>
> From: Subrata Modak <su...@li...>
> Subject: Re: [LTP] proc01 failure
> Date: Thu, 22 May 2008 19:20:58 +0530
>
> > Also wanted to forward this info to the test case author and to others
> > who have modified/patched this later.
> >
> > Stephane/Mike/Cai,
> >
> > Can you please look in to the following error in:
> >
> > ltp/testcases/kernel/fs/proc/proc01.c
> > (http://ltp.cvs.sourceforge.net/*checkout*/ltp/ltp/testcases/kernel/fs/pr
> >oc/proc01.c)
> >
> > test case.
> >
> >
> >
> > ...........................
> > ...........................
> >
> > Thanks Li for the patch. The same has been applied. Can you also look
> > into one test case which has been giving some weired failure. It was
> > regarding the:
> >
> > ltp/testcases/kernel/fs/proc/proc01.c test case.
> >
> > I am just pasting what Max Stirling <vic...@gm...> reported on
> > the mailing list some days back:
> >
> > Max´s mail.....................
> >
> > Does proc01 testcase pass for anyone?
> >
> > I ran the binary on my machine and it fails. Please see the below log.
> >
> > Am I _missing_ something?
> >
> > ubuntu@/home/ubuntu/Documents/work/ltp/ltp-full-20071130/testcases/kernel
> >/fs/proc ->sudo ./proc01
> > proc01      0  INFO  :  /proc/acpi/event: open: Device or resource busy
> > proc01      0  INFO  :  /proc/sys/fs/binfmt_misc/register: is
> > write-only.
> > proc01      0  INFO  :  /proc/sysrq-trigger: is write-only.
> > proc01      0  INFO  :  /proc/kmsg: read: Resource temporarily
> > unavailable
> > proc01      0  INFO  :  /proc/self/task/5273/mem: read: Input/output
> > error
> > proc01      0  INFO  :  /proc/self/task/5273/clear_refs: is write-only.
> > proc01      0  INFO  :  /proc/self/task/5273/attr/prev: read: Invalid
> > argument
> > proc01      0  INFO  :  /proc/self/task/5273/attr/exec: read: Invalid
> > argument
> > proc01      0  INFO  :  /proc/self/task/5273/attr/fscreate: read:
> > Invalid argument
> > proc01      0  INFO  :  /proc/self/task/5273/attr/keycreate: read:
> > Invalid argument
> > proc01      0  INFO  :  /proc/self/task/5273/attr/sockcreate: read:
> > Invalid argument
> > proc01      0  INFO  :  /proc/self/mem: read: Input/output error
> > proc01      0  INFO  :  /proc/self/clear_refs: is write-only.
> > proc01      0  INFO  :  /proc/self/attr/prev: read: Invalid argument
> > proc01      0  INFO  :  /proc/self/attr/exec: read: Invalid argument
> > proc01      0  INFO  :  /proc/self/attr/fscreate: read: Invalid argument
> > proc01      0  INFO  :  /proc/self/attr/keycreate: read: Invalid
> > argument
> > proc01      0  INFO  :  /proc/self/attr/sockcreate: read: Invalid
> > argument
> > proc01      1  FAIL  :  readproc() failed with 14 errors.
>
> Those failures are known issues and probably too subtle to be fixed in
> kernel side. However, You could try to apply the following two patches
> in attachments which let the test pass at least for RHEL.
>
> Patch (1/2) expect_fail.patch
> Patch (2/2) expect_fail_more.patch

ugh, so fugly.  not your patches' fault though.  the code should probably be 
changed to declare an array of file names and expected failures like so many 
other test cases instead of having an ugly (and ever growing) list of if 
statements.  bonus points if you do it, otherwise i'd just add a TODO comment 
to the top ;).
-mike