Messages in this thread | | | Date | Mon, 30 Nov 1998 00:06:25 +0100 (MET) | From | Gerard Roudier <> | Subject | Re: typo in scsi_obsolete.c |
| |
On Sun, 29 Nov 1998, Andrea Arcangeli wrote:
> On Sun, 29 Nov 1998, Gerard Roudier wrote: > > >> + /* FIXME: last_reset == 0 is allowed */ > > This mean last_reset == 0 is the same as last_reset == 1 or last_reset == > -1.
But some codes rely on zero value to keep track on the delay being not valid. If 0 and 1 are the same, just have to prevent jiffies from being 0 by jumping to 1 after ~0ul and you will get a nice 0=1 condition that will make happy all codes relying on some zero kernel time value. :-)
> >This comment does not make sense for a variable that wraps around. > > Woops, I think that you have misunderstood the comment or my comments > sounds bad... ;)
The comment points to a bug or is just useless.
> >By the way, the ncr53c8xx driver assumes that linux will be rebooted > >before the 497 days uptime that wraps jiffies around. > > If this is true
It should not be false. But the jiffies wrap-around problem had been stated not urgent by Linus. I only want to fix it in time. :)
> (as I just said some time ago), you should at least change > the: > > jiffies++ > > with > > if (!(jiffies+1)) > { > sync(); > panic("jiffies wrap"); > } > jiffies++;
Are you sure this change would make any difference? ;-)
> You should also document the thing. I am not aware of a FAQ that say that > linux has to be resetted on day 497.
I am wondering if humans will really continue using computers after the Y2K crash which will occur in less time for jiffies to wrap around on machines running 2.2 kernels. :-)
> I think we are lucky though since the reset-scsi code is the only thing I > am aware that isn' t jiffy wrap compliant (or at least not trivially > fixable) right now. I' ll spend my next spare time on it infact.
If you have much spare time, you may have some chance to succeed, IMO. ;-)
> Ideas and comments are always welcome...
Will try to have some.
Regards, Gerard.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |