There are two reasons why ide-disk crashes at system shutdown.
1. idedisk_devdrv.lock is not initialized (RW_LOCK_UNLOCKED is not all
zeroes on x86).
2. The ide-disk module "manually" sets device->driver, but
drivers/base/core.c assumes that if device->driver != NULL then
a bunch of other fields also have valid data (driver->driver_list,
for example). In this patch, I just clear drive->device.driver
before calling put_device as a first step. Eventually, I would
like to use {driver,device}_register than their generic
matching facility, but that might take a day to get right and
I would like to get this first step into the kernel tree and
get your input on this idea.
In my patch, I also reindented the fields for idedisk_devdrv.
It's just a suggestion.
I'll cc this to linux-kernel since others have reported
this kernel BUG().
By the way, with this patch I now get a crash at about
the same time in usbdevfs, but I believe that is not part of
this problem.
Adam J. Richter __ ______________ 575 Oroville Road
[email protected] \ / Milpitas, California 95035
+1 408 309-6081 | g g d r a s i l United States of America
"Free Software For The Rest Of Us."
--- linux-2.5.21/drivers/ide/ide-disk.c 2002-06-08 22:26:29.000000000 -0700
+++ linux/drivers/ide/ide-disk.c 2002-06-13 02:08:46.000000000 -0700
@@ -822,8 +822,9 @@
*/
static struct device_driver idedisk_devdrv = {
- suspend: idedisk_suspend,
- resume: idedisk_resume,
+ suspend: idedisk_suspend,
+ resume: idedisk_resume,
+ lock: RW_LOCK_UNLOCKED,
};
/*
@@ -1196,6 +1197,9 @@
if (!drive)
return 0;
+ /* Kludge: we never registered with device_attach, so prevent
+ put_device from trying to remove us from device->driver_list */
+ drive->device.driver = NULL;
put_device(&drive->device);
if ((drive->id->cfs_enable_2 & 0x3000) && drive->wcache)
if (idedisk_flushcache(drive))