> In the hope of preserving state, I'll replay a suggestion which has come
> up the last time we've talked about the jiffies overflow problem:
>
> Modify the kernel to set the jiffies counter to "(unsigned) -300" at
> bootup, and then adjust the time routines appropriately. This will
> cause the jiffies overflow condition to happen very shortly after boot,
> which may help identifying any problem cases we might have.
>
> In fact, it might be a good thing to do by default in the 2.1 kernel
> series.....
If you're going to do that, then I would recommend giving it more than a
-300 start value (in this case, more meaning a more negative value). 3
seconds isn't very long after the time counters are initialized for a person
to get on the system and try to do something to break it. After all, it may
turn out that 3 seconds after these values are set up, we happen to be in
the middle of a SCSI bus reset delay or some such where the system is
idling. I would give more along the lines of a few minutes ( -12000 maybe )
in order for the system to come up, someone to get in, and for the person
testing things to have some activity running at the time of the wrap (maybe
a bonnie to test block driver problems, or a ping -f to test some network
code, etc.). Without the ability to get in and stress some particular part
of the kernel at the exact time of the wrap, chance and coincedince could
result in many a problem getting missed.
-- ***************************************************************************** * Doug Ledford * Unix, Novell, Dos, Windows 3.x, * * dledford@dialnet.net 873-DIAL * WfW, Windows 95 & NT Technician * * PPP access $14.95/month ***************************************** * Springfield, MO and surrounding * Usenet news, e-mail and shell account.* * communities. Sign-up online at * Web page creation and hosting, other * * 873-9000 V.34 * services available, call for info. * *****************************************************************************