>> I don't know if this patch is necessary. I think Alan told about this

>> some days ago. In my case I setting my irq_type with a device_tree and I
>> know other boards, which are not using a device_tree, do the same in the
>> boardspecific code file in arch/arm/**/*_board.c.
> 
> The at86rf230 supports generating multiple types of IRQ (eg: rising and
> falling). His patch makes it configurable, in case it's being connected
> to a board or SoC that is not as configurable.


To be honest, at the moment I'm not sure what would be the "right"
solution here. The patches don't change the way the IRQ are generated by
the rf230, but change the way the driver handle them. The rf230 always
sets the IRQ line to high when a interrupt cause occurs and resets the
line when we read the interrupt status register.

Patch 2/3 configures the interrupt type of the board, which (my guess)
most people like to see in device_tree / board files. Patch 3/3 changes
the way the interrupts are handled by the driver. For boards configured
to level type interrupts we need to disable interrupts until the irqwork
function is done (else the interrupt would fire multiple times until we
are done with the work). For boards with edge type interrupts, the
interrupt shouldn't be disabled because we may miss some interrupts
while handling the first one.

For patch 3/3 we need to tell the driver how the IRQ is configured.
Removing the irq_set_irq_type() call from patch 2/3 would force the user
to configure the irq type twice. Alas there seem to be no
irq_get_irq_type() function available in the kernel :(

What do you think about it? Maybe I should post it as rfc patch to netdev?

Thanks,
Sascha

> 
> In the device tree or board file (in your example), you're setting the
> IRQ type that the board or SoC should expect. This patch sets the IRQ
> type that the at86rf230 will generate.
> 
>> I would only add this, if this is really necessary to setup irq_type at
>> probing time.
> 
> If you want something other than the default from the at86rf230, then it
> is necessary.
> 
> Alan.
> 
> 
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_mar
> _______________________________________________
> Linux-zigbee-devel mailing list
> Linux-zigbee-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/linux-zigbee-devel



-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature to help me
spread!

------------------------------------------------------------------------------
Everyone hates slow websites. So do we.
Make your web apps faster with AppDynamics
Download AppDynamics Lite for free today:
http://p.sf.net/sfu/appdyn_d2d_mar
_______________________________________________
Linux-zigbee-devel mailing list
Linux-zigbee-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-zigbee-devel

Reply via email to