On Friday 27 April 2012 06:40 PM, Russell King - ARM Linux wrote:

On Fri, Apr 27, 2012 at 06:31:13PM +0530, Laxman Dewangan wrote:

quoted

On Friday 27 April 2012 05:03 PM, Russell King - ARM Linux wrote:

quoted

On Fri, Apr 27, 2012 at 07:26:56PM +0800, Richard Zhao wrote:

quoted

Sure, I mean, can you doc in include/linux/dmaengine.h that
dmaengine_prep_xxx may sleep?

Incorrect.  They may _not_ sleep.

But I have seen that we are using the kzalloc in the dmaengine_prep_xxx
and kzalloc is sleeping call.

Not with GFP_ATOMIC it isn't.

Yes, with GFP_ATOMIC , it will not be a sleeping calls.
I will fix my dma driver accordingly.