On Tue, 8 Jul 2014, Mark Brown wrote:
> On Tue, Jul 08, 2014 at 10:15:20AM +0200, Julia Lawall wrote: > > On Tue, 8 Jul 2014, Mark Brown wrote: > > > > It should be fairly clear given what they do I'd have thought - the > > > devm_ functions tie the deallocation of a resource to the unbinding of > > > a driver from a device so they can only be used to replace things that > > > get cleaned up in a device model unbind path. There's not usually a > > > great deal of indirection going on in those. > > > It is completely clear what they do. What is not clear is what device > > libraries are set up to call the freeing functions at what point. For > > example, I know that that platform drivers are set up for this, but once I > > tried to find the lines of code that would justify that, but I could not. > > Perhaps I was not patient enough or missed something. > > All devices do this - it's done as part of the driver model core code so > there is no need for individual buses to do anything. How should one realize that this does not apply to the original file under discussion, sound/soc/codecs/sgtl5000.c? The associated structure is snd_soc_codec_driver. What code could one look for at the call sites of the probe and remove functions to know that managed memory can be used? thanks, julia -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/