It looks like my Compaq Contura 410 notebook generates an unexpected
intr each time it spins down the disk, misusing the busy flag to
signal the spindown state. This results in a printk message, and a
disk spinup to write it to the logfile.
Before 2.0.9, this also produced a cascade of unexpected intr messages
with status 0x00 (no kidding :-) while the disk was spinning up again.
This is gone since some kind soul set WAIT_READY to 5s if CONFIG_APM
is set (I could bite my backside for not trying this before :-). To
make my notebook completely happy I've suppressed the status 0x80
message:
--- ide.c.old Sat Jul 27 10:23:41 1996
+++ ide.c Sun Jul 28 13:16:16 1996
@@ -1644,7 +1644,7 @@
static unsigned long last_msgtime = 0;
if ((last_msgtime + (HZ/2)) < jiffies) {
last_msgtime = jiffies;
- (void) ide_dump_status(drive, "unexpected_intr", stat);
+ if (stat != 0x80) (void) ide_dump_status(drive, "unexpected_intr", stat);
}
}
if ((stat & DRQ_STAT))
I know that this not the way it should be done, but I'm willing to
take this kluge for now, because I'm too lazy to analyze the ide code
to the extend that I know what is happening where and why. :-) Maybe
some adept can use this information.
Chris
-- Chris Blum http://www.phil.uni-sb.de/~chris/ PGP encrypted mails welcome