Dmitry Adamushko wrote:
> On 19/08/06, Jan Kiszka <[EMAIL PROTECTED]> wrote:
>>
>> just realised that the output of /proc/xenomai/irq is suboptimal (no
>> names) under !CONFIG_XENO_OPT_SHIRQ_LEVEL &&
>> !CONFIG_XENO_OPT_SHIRQ_EDGE...
>
>
>
> It should be ok but just in case, Jan, pls give it a t
On 22/08/06, Jan Kiszka <[EMAIL PROTECTED]> wrote:
Dmitry Adamushko wrote:.>> diff -urp xenomai-SVN/include/nucleus/intr.h xenomai-a/include/nucleus/intr.h> --- xenomai-SVN/include/nucleus/intr.h2006-07-20 11:09:01.0 +0200> +++ xenomai-a/include/nucleus/intr.h 2006-08-22 09:32:
errr... xnintr_irq_proc() for shared-irq case is also broken
indeed. xnintr_shirq_lock(shirq) should not be used there as a
preemption may take place and another thread may happilly call
xnintr_detach() -> xnintr_shirq_detach() -> possible [SMP:
deadlock, UP: crash].
Can't believe the person who