On Tue, Jan 4, 2011 at 2:18 AM, Luck, Tony <tony.luck@xxxxxxxxx> wrote:
>> if you can get the patch merged into your test branch
>
> My test branch is pulled into each linux-next build ... and there
> are some generic changes at the end of your patch that I don't
> understand - but they look like they will affect other architectures.
Oh....then would you like to create a new branch for stuff' like lockdep?
It would be very helpful for me.
>
> What are INIT_{SOFT,HARD}_IRQFLAGS for? Why do we need to add them
> for ia64?
I found a lockdep warning in kernel boot where a kernel process (don't
remember if it is IDLE) is hit by the following check in copy_process
(kernel/fork.c)
#ifdef CONFIG_PROVE_LOCKING
DEBUG_LOCKS_WARN_ON(!p->hardirqs_enabled);
DEBUG_LOCKS_WARN_ON(!p->softirqs_enabled);
#endif
I noticed that there is a Macro in INIT_TASK that has softirqs_enable
initialized to 1. (by default it is 0 ,which is copied from
INIT_TASK). Both softirqs_enabled and hardirqs_enabled must be
explicitly initialized to 1 in relavent kenrel paths by the lockdep
and irqflag trace patch. I still need to find a right place to do
that.
Do you want me to fix this as the firs step to merge?
Thanks,
Luming
--
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
|