Date: Wed, 30 Dec 1998 22:27:37 -0600 (CST) From: Font <font@Mcs.Net> To: freebsd-questions@FreeBSD.ORG Subject: Another "wd0: interrupt timeout" message data point Message-ID: <Pine.BSF.3.95.981230220821.25777A-100000@Jupiter.mcs.net>
next in thread | raw e-mail | index | archive | help
I've been getting the popular interrupt timeout messages, a la
Dec 30 21:39:04 kibbeh /kernel: wd0: interrupt timeout:
Dec 30 21:39:04 kibbeh /kernel: wd0: status 52<rdy,seekdone,index> error 0
Dec 30 21:39:35 kibbeh /kernel: wd0: interrupt timeout:
Dec 30 21:39:35 kibbeh /kernel: wd0: status 52<rdy,seekdone,index> error
1<no_dam>
which many people have seen. Apparently it's often the case that this
happens when, after a period of inactivity, a machine with APM or other
power management has spun down the disk, and activity causes the disk to
be spun up again. But this isn't the case all the time, or so I read;
this is one of those other cases.
So, I have what might be a new data point to add. The interrupt timeouts
happen when my computer is active, and APM is completely disabled, but
only after I have been playing with SoundBlaster 16 bit audio.
The machine is an IBM Intellistation Pro M model 6898-22U, a 333 MHz
Pentium II. The on-board Crystal audio is disabled and a SoundBlaster
AWE32 is installed with (what I believe to be) the default correct PNP
settings, and the 2.2.8-RELEASE-based kernel is recompiled to reflect
this.
I have the port for amp compiled, which I use to play the occasional mp3
from the command line. During play of the mp3, frequent pauses (every
minute or three) will result, and the first two times the pauses occur,
the disk activity light is on, the rest of the machine pauses, then the
interrupt timeout error message results. After the first two times, the
interrupt timeout error may not result (but sometimes does), but the
pauses still occur. The mp3 file is stored on the local disk.
Sometimes, after these timeouts occur, I will be unable to perform a "df"
command (even as root), and the system will not shut down properly
afterwards.
If, while playing the mp3 file, I continuously generate disk activity (for
instance, a du -s from root, with all sources installed), the interrupt
timeouts and pauses seem to occur much less often (but they still do
occur).
The disk drive is a Seagate ST36530A, which according to the web site, may
have Ultra DMA problems. I don't know much about UDMA, and the IBM BIOS
doesn't really say anything directly about UDMA, only something about
read-ahead on/off and "Standard" or "High Performance" mode for the disk
(not the controller). I've tried all four combinations of these settings
and there seems to have been no difference in the pauses.
When not playing mp3's at all, the system seems very stable. I'm thinking
my next move may be to buy the OpenSound drivers and use the on-board
Crystal audio, and see if there is an improvement.
I don't think the pauses are occurring under Windows 95 (no Myth II Demo
for FreeBSD yet), but I am not certain, plus Windows doesn't provide any
diagnostic messages. I'd try it more under Windows, but the machine is
busy running more important tasks...
Anyway, that's my data point, and hopefully someone can use this
information, or, if it enables anyone to help me to bypass the problems
(so I can play mp3s again, my newfound joy at work; I was going to try to
rip my whole Buzzcocks collection), I thank you in advance.
dw
A bug in my MUA causes news.announce.newusers font
to be sent to beneficiaries and senders of UCE/SPAM. @
mcs.net
Wishes are like dishes.
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.981230220821.25777A-100000>