On Wed, Aug 01, 2012 at 07:00:54AM +0200, Marek Vasut wrote:

Dear Shawn Guo,

quoted

On Tue, Jul 31, 2012 at 08:35:59PM -0700, Guenter Roeck wrote:

quoted

I think the call to spi_master_put() triggers the call to
spi_master_release(). If so, kfree() would not be needed at all, and the
documentation is wrong.

Also those drivers calling kfree in probe.

Looks like that to me ...

Doesn't seem to be far spread, fortunately. Only spi-davinci.c, spi-imx.c, and
spi-omap2-mcspi.c as far as I can see, plus the misleading comment in spi.c.

Anyone up for writing some patches ? If not, I'll do it.

Thanks,
Guenter