On Mon, 2018-05-14 at 20:38 +0300, Andy Shevchenko wrote:

First of all, do not remove mailing lists from Cc and people if you
are not sure they do not need your stuff.

Sorry. My mistake. 

On Mon, May 14, 2018 at 11:11 AM, Radu Pirea
[off-list ref] wrote:

quoted

On Sun, 2018-05-13 at 16:33 +0300, Andy Shevchenko wrote:

quoted

On Fri, May 11, 2018 at 1:38 PM, Radu Pirea <radu.pirea@microchip
.com

quoted

wrote:
+static void at91_usart_spi_cleanup(struct spi_device *spi)
+{
+       struct at91_usart_spi_device *ausd = spi-

quoted

controller_state;

+
+       if (!ausd)
+               return;

Is it even possible?

Theoretically yes.

I would like to know real circumstances when it might happen.

That check was used in debug stage of driver. I will remove.

quoted

quoted

Anyway the code below will work fine even if it's the case.

quoted

+
+       spi->controller_state = NULL;
+       kfree(ausd);
+}

The question is, why you didn't utilize what SPI core provides
you?

I tried, but it did not work the way I expected.

So, what is not going as expected in "SPI core takes care of CSs"
case?
Did you use oscilloscope for that?

Yes, I used and CSs was not asserted. Anyway, I will will try again.