[cgl_discussion] RE: timers collaboration

Randy.Dunlap rddunlap at osdl.org
Fri Jan 10 12:31:11 PST 2003


On Wed, 8 Jan 2003, Fleischer, Julie N wrote:

| Moving this thread to cgl_discussion....
|
| Thanks for this information, Randy.  It also points out some holes in my
| knowledge of what was available for POSIX Timers.  I'd like to put some more
| time into investigating these items.
|
| Randy Dunlap wrote:
| > On Wed, 8 Jan 2003, Fleischer, Julie N wrote:
| > |
| > | Randy -
| > | I'm not sure of the details, but I'm wondering if there are some
| > | opportunities for collaboration on what you are doing for
| > evaluating timers
| > | and what we are doing with the POSIX Test Suite.
| > Specifically, I've been
| > | testing the latest timers implementations (George
| > Anzinger's and I have
| > | plans for Jim Houston's work also -- Jim is also using the
| > POSIX Test suite
| > | to test his code.).  I am wondering if some of that input
| > could feed into
| > | your decisions?
| >
| > Hi Julie,
| >
| > I certainly plan to use code from posixtest.sf.net plus George's test
| > suite.  Have you looked at any test code from the original
| > Uni. of Kansas
| > high-res-timers project, or is some of your test code from there?
| > (http://hegel.ittc.ukans.edu/projects/posix/index.html)
| >
| > I also plan to evaluate both George's and Jim's POSIX interface code.
|
| Thanks for the URL.  I had only seen the U. of Kansas test cases that were
| part of the high-res-timers package, but don't think I had seen the source
| before.  Originally, it seemed that these tests were duplicates of the tests
| we had in the POSIX Test Suite, but I'll have to look at the sources to know
| for sure.

Looks to me like the UKansas timers tests are in George's test suite,
but the signals tests might be helpful.

Do you have plans to automate the Posix timers test suite, or do you
expect to leave it as is, requiring manual intervention?
Hint:  I'm not into repeated manual intervention.  :)

| > | At the least, it might be useful for you to know that we
| > now actually have a
| > | first release of POSIX Timers conformance tests from our
| > website (URL on
| > | your website - http://posixtest.sf.net).  In addition, we
| > publish the
| > | results of the failures we find to the HRT mailing list and
| > LKML.  (You may
| > | already know all this, though....)
| >
| > Yes, I do.
| >
| > | In the future, we may be able to share more data somehow.
| > (For example,
| > | after I saw your mail about the spinlocks you were seeing,
| > I ran through our
| > | suite and noticed similar intermittent behavior.)
| > |
| > | I'm not sure exactly how we could share information in the
| > future, but
| > | thought it would be good to at least let you know what we
| > are doing and ask
| > | if there was any way you thought we could help.
| >
| > Sounds like a good idea.  I'll keep in touch and see what we can do.
| >
| **These views are not necessarily those of my employer.**

-- 
~Randy




More information about the cgl_discussion mailing list