Re: [PATCH 3/3] spi: bcm2835: Enable shared interrupt support

[Lukas Wunner <lukas@xxxxxxxxx>]

On Thu, Jun 04, 2020 at 01:24:54PM -0700, Florian Fainelli wrote:
> So we do need to know for the first time we install the interrupt
> handler whether we will be in a shared situation or not, I cannot think
> of any solution other than counting the number of available DT nodes
> with the "brcm,bcm2835-spi" compatible string.

In principle it would be possible to iterate over the entire DT using
for_each_of_allnodes() and call of_irq_parse_one() on each device_node
if it's enabled and not the one we're probing.  Then check if any of that
device_node's IRQs is identical to that of the device_node we're probing.
That would give you a generic method to test for sharedness of an
interrupt.

However the solution you've found is simpler and cheaper than such a
brute-force search, hence seems perfectly valid to me.


> I appreciate that
> Lukas has spent some tremendous amount of time working on this
> controller driver and he has a sensitivity for performance.

Thanks!  Indeed I think spi-bcm2835.c is by now among the best performing
and most featureful SPI drivers in the kernel.  I've recently had a
discussion on netdev with someone testing an SPI-attached Ethernet
chip on iMX6Q and on STM32MP1 and couldn't get it to work.
With a BCM2837 no problem at all:

https://lore.kernel.org/netdev/ac0f7227-a4ae-b6cd-36ec-3bcb02b1adbe@xxxxxxx/

Lukas