On Wed, Dec 29, 2010 at 8:21 PM, Nicolae Mogoreanu <[email protected]> wrote:
> Just to illustrate what I try to say by 'test stability' is for example the
> fact that 'btreplay.py' references 'blktrace-git-latest.tar.gz' which is
> actually not checked in. It means that either people down run btreplay tests
> or they have to find the blktrace-git-latest themselves

That sounds like a simple bug, probably by me or someone else forgetting
to do an svn add before a commit.

> You would help me tremendously if you could tell me what tests people
> usually run and what tests are considered important.

These look reasonable (thought sleeptest doesn't do anything really)

>> compilebench, cyclictest, disktest, dma_memtest, hwclock, linus_stress,
>> pktgen, scrashme, sleeptest, stress
>> Here are some that do not work right now for various reasons, but seem
>> interesting for testing:
>> cerberus, ebizzy, ltp, pi_tests, signaltest, synctest, tiobench, btreplay,
>> bonnie, fs_mark, aiostress, ... maybe others
>> 1. Is there a set of tests that is guaranteed to be stable and is
>> recommended to run?

I'm not working on the kernel any more, so I don't have current correct
knowledge, but it boils down to what others are using.

kernbench, tbench, dbench, disktest, bonnie, aiostress

were what I recall running.
_______________________________________________
Autotest mailing list
[email protected]
http://test.kernel.org/cgi-bin/mailman/listinfo/autotest

Reply via email to