On Wed, 14 May 2014 10:15:38 +0100, "Jon Medhurst (Tixy)" tixy@linaro.org wrote:

On Sun, 2014-03-02 at 13:40 +0800, Grant Likely wrote:

...

On Fri, 28 Feb 2014 14:42:50 +0100, Marek Szyprowski m.szyprowski@samsung.com wrote:

...

This patch adds code for automated assignment of reserved memory regions to struct device. reserved_mem->ops->device_init()/device_cleanup() callbacks are called to perform reserved memory driver specific initialization and cleanup

Based on previous code provided by Josh Cartwright joshc@codeaurora.org

Signed-off-by: Marek Szyprowski m.szyprowski@samsung.com

Hi Marek,

I've not applied this one yet, only because there is still the open issue of whether or not these functions should be called from drivers or from core code. I don't actually have any problems with the content of this patch. Once the user is sorted out I'll merge it.

Has anything more come of these patches? I see some of the series is now in Linux 3.15, but the actual patches to let people use the feature aren't there yet, namely patches 5 though 8.

My personal immediate interest in these is as a mechanism on arm64 to limit CMA to a region of memory that is actually DMA-able devices (e.g. below 4GB for 32-bit devices without an iommu).

For reference, the mail archives for this series is at http://lkml.org/lkml/2014/2/28/237

IIRC, the issue I have with patch 5-8 is that I don't like the driver core going off and doing automagical things to attach regions to devices. I've not seen any more discussion on this topic since I merged the patches I was okay with, but I may have missed something.

g.