oprofile broken by sysfs updates

oprofile broken by sysfs updates

[Pavel Machek]

Hi!

I only tried reading the diffs, but:

arch/i386/oprofile/nmi_int.c must be suspended before
arch/i386/kernel/apic.c is.

How is that guaranteed with new code?

-static struct device device_nmi = {
-       .name   = "oprofile",
-       .bus_id = "oprofile",
-       .driver = &nmi_driver,
-       .parent = &device_lapic.dev,
+static struct sys_device device_oprofile = {
+       .id     = 0,
+       .cls    = &oprofile_sysclass,
 };


							Pavel
-- 
When do you have a heart between your knees?
[Johanka's followup: and *two* hearts?]

Re: oprofile broken by sysfs updates

[Patrick Mochel]

> I only tried reading the diffs, but:

Then maybe you should

a) Read the entire initial thread (which you participated in), esp this 
message:

http://marc.theaimsgroup.com/?l=linux-kernel&m=105518049424749&w=2

b) Read the comments in the code (from drivers/base/sys.c):

/**
 *      sysdev_shutdown - Shut down all system devices.
 *
 *      Loop over each class of system devices, and the devices in each
 *      of those classes. For each device, we call the shutdown method for
 *      each driver registered for the device - the globals, the auxillaries,
 *      and the class driver. 
 *
 *      Note: The list is iterated in reverse order, so that we shut down
 *      child devices before we shut down thier parents. The list ordering
 *      is guaranteed by virtue of the fact that child devices are registered
 *      after their parents. 
 */

c) Try using the code and stop being a troll.


Thanks,

	-pat

Re: oprofile broken by sysfs updates

[John Levon]

On Wed, Jun 11, 2003 at 11:12:20PM +0200, Pavel Machek wrote:

> arch/i386/oprofile/nmi_int.c must be suspended before
> arch/i386/kernel/apic.c is.
> 
> How is that guaranteed with new code?

http://marc.theaimsgroup.com/?l=linux-kernel&m=105517556119093&w=2

regards
john