I am looking at the patch that solve the long wait time problem.
-/* The hardware supports interrupt driven mode too, but I haven't
- implemented that. */
-#define POLLED_MODE 1
-#define POLL_TIMEOUT (HZ)
+#define POLL_TIMEOUT (HZ/5)
Isn't the POLL_TIMEOUT too short after I applied the patch? It is now
1/5 of the original value.
Punky Tse wrote:
Hi Beat,
I encounter the same problem right now - temperature reading all
return zero values under high load. And there are loads of i2c
timeout messages in /var/log/kern.log. Sounds like the kernel is busy
to response i2c commands?
- Punky
Beat Meier wrote:
It seems that the problem of the timeouts occures when the cpu is
heavy loaded.
I do a scp to 3 clients in the background in parallel endless and the
problem is there.
At least the timeouts but not the bad value of the temp.
Will investigate more ...
Beat
On Wed, 2006-01-18 at 06:12 +0100, Michael Renzmann wrote:
There have been some patches to the scx200_acb driver in the
lm-sensors
mailing list recently, which are said to fix some bugs in the driver.
I'll give them a try (after backporting them to 2.4.x) and let you
know
if that helped us.
I'm getting old. I already applied these patches, so they didn't
help us
with the problem.
Could you please check what value POLL_TIMEOUT is set to in your
version
of scx200_acb?
Bye, Mike
_______________________________________________
Voyage-linux mailing list
[email protected]
http://list.voyage.hk/mailman/listinfo/voyage-linux
--
Regards,
Punky
P U N K N ! X . c o m
Technology + Lifestyle
(http://www.punknix.com)
Voyage Linux
(http://www.voyage.hk/software/voyage.html)
_______________________________________________
Voyage-linux mailing list
[email protected]
http://list.voyage.hk/mailman/listinfo/voyage-linux
