Re: OMAP1 realtime patch

[Dirk Behme <dirk.behme@googlemail.com>]

Kevin Hilman wrote:
> On Mon, 2007-02-05 at 20:29 +0100, Dirk Behme wrote:
> 
>>For everybody interested please find in attachment my OMAP1
>>realtime patch.
>  
> Was there a reason for the raw_spinlock conversions? 

Not really ;)

> Did you run into
> specific issues for converting those?  

There was this ethernet/irq/gpio issue we fixed with 
workaround in kernel/irq/manage.c. Until we had this the 
root cause was unclear so a lot of raw_spinlock to be on the 
safe side for debugging. Then, I posted my patch to have a 
starting point people can try and improve, as you did :)

> Here is a patch which leaves out the raw_spinlock conversions.  It goes
> on the OMAP git tree on top of -rt5 (ignoring conflicts.)  I've left out
> all the raw_spinlock conversions.  This has been tested on an
> OMAP1623/H2 platform.

Many thanks for the update!

> Index: dev/kernel/printk.c
> ===================================================================
> --- dev.orig/kernel/printk.c
> +++ dev/kernel/printk.c
> @@ -464,6 +464,7 @@ static void zap_locks(void)
>  	spin_lock_init(&logbuf_lock);
>  	/* And make sure that we print immediately */
>  	init_MUTEX(&console_sem);
> +	zap_rt_locks();
>  }

What's this?

Btw: You removed the HRTIMER_MODE_REL and enum 
hrtimer_restart changes in 
drivers/input/touchscreen/ads7846.c as well?

Cheers

Dirk