Re: Kernel testing

David S. Miller (davem@jenolan.rutgers.edu)
Sat, 12 Apr 1997 23:07:37 -0400


Date: Sat, 12 Apr 1997 22:57:36 -0400 (EDT)
From: "Seth M. Landsman" <seth@job.cs.brandeis.edu>

I do think that the kernel lacks basic tests which quite
possibly should be there. Just because it compiles doesn't mean
that it is okay to run and use. There should be a `make test` step
in kernel compilation that ensures some basic truisms about the
kernel, such as return values of functions and the ability to boot.

What is the concensus to this?

Just a quick note. Most UNIX vendors have a directory full of stress
test toys, each test is geared specifically towards a certain
subsystem of the kernel. Essentially when you change stuff you go
into there and say "If you dicked with the buffer cache, run one of
test X Y or Z, if you changed the locking code in the networking run A
or B"

---------------------------------------------////
Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><