Re: [PATCH 2/2] I-pipe: x86 port

[Philippe Gerum <rpm@xenomai.org>]

Daniel Walker wrote:
> On Sat, 2005-06-18 at 01:21 +0200, Philippe Gerum wrote:
> 
> 
>>+#define local_save_flags(x)	((x) = __ipipe_test_root())
>>+#define local_irq_save(x)	((x) = __ipipe_test_and_stall_root())
>>+#define local_irq_restore(x)	__ipipe_restore_root(x)
>>+#define local_irq_disable()	__ipipe_stall_root()
>>+#define local_irq_enable()	__ipipe_unstall_root()
>>+
>>+#define irqs_disabled()		__ipipe_test_root()
> 
> 
> 
> If you want to integrate with newer RT patches , (I'm not sure which one
> this is on) ..

This patch only inherits the basic conditionals for integrating with 
PREEMPT_RT from a recent Adeos combo experimenting the issue w/ 
0.7.44-03, but there is indeed more to do for completing the integration 
with recent PREEMPT_RT patches.

  You could modify the above so that it uses the soft irq's
> off flags  . So if linux is in a soft irq disable than it's equal to a
> stall on the ipipe ..
> 

Ack. I will provide a newer combo patch using PREEMPT_RT's virtualized 
local_irq_disable for the root domain. Thanks for the hint.

-- 

Philippe.