[ofa-general] bitops take an unsigned long *

John Gregor john.gregor at qlogic.com
Tue May 13 16:00:49 PDT 2008


From: Roland Dreier <rdreier at cisco.com>
> Out of curiousity, was there any reason for choosing 0, 1, 2, 3 and
> then skipping to 62?

Not really.  Just that RUNNING and SHUTDOWN are conceptually different
from ABORTING, DISARMED, and DISABLED and so it seemed to make sense
at the time to cluster the bits at opposite ends of the qword.  It made
the printk() output easier to scan quickly during debugging.

-John Gregor



More information about the general mailing list