Hi Kevin,
> The board is running a 2.4 kernel so if some of the suggestions are specific
> for 2.6, I couldn't use them. Adding /proc nodes was the easier method.
You'll find that adding /dev or /proc nodes is almost identical.
The call-backs are the only things that change.
One trick I have found useful is to make sure that the node
understands ASCII and binary so that;
echo 0 > /dev/mydev
and a low-level write of a binary 0 have the same effect.
>> So did you end up adding a spinlock to protect the
>> resource, or queue I/O requests and let the ISR
>> update the I/O pins?
>
> No spinlocks. I just write to a /proc node and the value is saved to an
> internal variable. This variable is then used by the ISR when it updates the
> I/O pins. The 500uS to 1mS delay between the proc write and the change of
> state in the I/O pins is not a problem for my app.
>
No ... if you are using the variable from write() and from the
ISR, then you will generally need to protect it with a spinlock.
There are some rare cases where you might have used an atomic
variable, or its write-only from the user-space code and read-only
from the ISR, and you can guarantee that the read or write are
atomic. But then you go an add a debug statement somewhere
that uses the variable, and then you mess that assumption up.
Its generally good programming practice to add protection where
it is needed, and is likely to be needed. So in this case,
I you will want to add a spinlock to protect access to
your variable.
Feel free to convince me otherwise.
Cheers,
Dave
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/ts-7000/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/ts-7000/join
(Yahoo! ID required)
<*> To change settings via email:
<*> To unsubscribe from this group, send an email to:
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
|