Re: [x86_64] strange delays since 2.6.15 (was: Re: ohci1394: abortingtransmission)

From: Christian Kujau
Date: Wed Jul 12 2006 - 17:02:10 EST


Hello all,

I am still having these strange boot-delays which I reported in [0] when I was under the assumption that ohci1394 was to blame, simply because "ohci1394: aborting transmission" was the last message before the 3minute delay during bootup happened. I've found out that the actual problem started way earlier (2.6.14->2.6.15) and I have bisected my way down to this patchset:

a7fd67062efc5b0fc9a61368c607fa92d1d57f9e is first bad commit
diff-tree a7fd67062efc5b0fc9a61368c607fa92d1d57f9e (from d8539d81aeee4dbdc0624a798321e822fb2df7ae)
Author: Kay Sievers <kay.sievers@xxxxxxx>
Date: Sat Oct 1 14:49:43 2005 +0200

[PATCH] add sysfs attr to re-emit device hotplug event

A "coldplug + udevstart" can be simple like this:
for i in /sys/block/*/*/uevent; do echo 1 > $i; done
for i in /sys/class/*/*/uevent; do echo 1 > $i; done
for i in /sys/bus/*/devices/*/uevent; do echo 1 > $i; done

Signed-off-by: Kay Sievers <kay.sievers@xxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxx>

:040000 040000 7edb92ad6f55113bac2e1919e6b2364e17dfaa0b 43b3108675db828a2ce7df87f08a26aa7d7ac0fc M drivers
:040000 040000 382f692398a9fe4bea4ab5333dea67f8151b0b82 58bb09d0051236695e6987f12e389ce75387ddf7 M fs
:040000 040000 72bcc7fddd5087cadf70b9d2582c7faea7ebb364 2ec812818920645cb58f1694818019516ea6e3b6 M include

Wtihout this patch, booting is fine, no delays. With this patchset, I notice a 3 minute delay.

However, as the patch-description tells me, it only introduces something in /sys/..../*/uevent and the delay is in fact triggered by userspace!
My distribution of choice (Ubuntu/6.06 LTS) does "something" in /sys during execution of /etc/init.d/udev (from: udev-079-0ubuntu34). Please see this url for details, logs, .config and more:

http://nerdbynature.de/bits/2.6.15/

Many thanks for hints on what to do here. This issue is 100% reproducible and is kinda annoying now that I happen to be on-site when the machine boots up. As mentioned earlier I had no access to the box for a few months and did not pay attention to the boot-process and did not track -current too, that's why I'm whining about 2.6.15-problems when we're actually at 2.6.18-* already....

Thanks,
Christian.

[0] http://www.ussg.iu.edu/hypermail/linux/kernel/0607.1/1708.html
--
BOFH excuse #156:

Zombie processes haunting the computer
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/