Subject: modprobe usbcore from acpid problem 2.4.21
From: Micha Feigin <michf@post.tau.ac.il>
To: Kernel list <linux-kernel@vger.kernel.org>
Content-Type: text/plain
Message-Id: <1060889818.1511.1.camel@litshi.luna.local>
Mime-Version: 1.0
Date: 14 Aug 2003 22:39:24 +0300
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1272
Lines: 34

There is a bug in the usbcore module, such that if it is loaded using
modprobe from an acpid script in response to an event it later fails to
unload again.

I isolated the problem to calling /etc/init.d/modutils (debian) from a
script that is called in responce to the sleep button event. This in
turn loads every module in /etc/modules using modprobe.
The modules loaded this way are (in this order):
hid
mousedev
uhci
When I later try to unload the usb modules, the usbcore module locks at:
drivers/usb/hub.c:void usb_hub_cleanup(void)
at the call to
wait_for_completion(&khubd_exited);

This doesn't happen if I call /etc/init.d/modutils manually from the
command line.

This is under kernel 2.4.21 (debian version) with acpi+swsusp patches.
Any ideas what would be different between this two methods that can
cause this bug ? I have ruled out the fact that acpid is run as a
daemon and might thus cause a problem with the reparent_to_init call
(ran it as a regular program, no help).

-- 
Micha Feigin
michf@math.tau.ac.il

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