Re: Kernel testing

Paul H. Hargrove (hargrove@sccm.Stanford.EDU)
Mon, 14 Apr 1997 14:03:37 -0700 (PDT)


David S. Miller writes:
[snip]
> Better yet, don't throw away that tape drive or that old klunker disk
> when it starts resetting the bus or is otherwise non-functional. It
> isn't useless trash, it is a useful debugging tool...
[snip]

Actually on media like floppies and tapes you can introduce crc errors
with a refrigerator magnet or more serious errors with a sewing needle
(scratch not stab and be sure to remove the resulting loose
particles). Of course the sewing needle works on those stupid AOL
CDROMs too (I knew they had some use other than as coasters). Errors
can be introduced on really old Segate IDE hard drives by pulling the
plug while the Segate format utility runs. Other hard drives are
harder to damage, but I imagine the paint shaker at the hardware store
could induce a head crash. Human intervention can eject floppies (on
Intel platforms anyway) during read or writes.

Network errors can be introduced by software on another machine (only
do this on a private network!)

Hardware errors can be simulated by software as well. Imagine a
kernel thread (kerrord?) which would mess with pending I/O requests to
simulate hardware errors?

-- 
Paul H. Hargrove                   All material not otherwise attributed
hargrove@sccm.stanford.edu         is the opinion of the author or a typo.