On Dec 23, 9:04am, [email protected] (Jason Thorpe) wrote: -- Subject: Re: Current best practice for testing a new kernel API?
| > Some ideas: | > - kernel module to load the test | > - use rump | | It's almost noon where you are, Christos, but I'm not sure you've had your = | coffee yet -- that's essentially what I said in my email. :-) Yes, that's true :-) | Anyway, one drawback of the module approach is that you have to be down-sec= | urelevel to load one. So, any system that wants to run this test would hav= | e to pre-load the module at boot time. Perhaps that's ok if anita can do it. | I might go ahead and implement the module approach first simply because I a= | lready have the jig written, and the ATF script can just test for the prese= | nce of the jig and skip the test if it's not present, rather than trying to= | load the module. | | For completeness, I'll explore the rump option, but I have to figure out ho= | w that stuff works, first. Once you have the module done, it should not be hard to write a rump based harness. christos
