On 5/29/23 10:38, Greg Kroah-Hartman wrote:
On Mon, May 29, 2023 at 01:48:05AM +0200, Marek Vasut wrote:
On 5/28/23 21:11, Greg Kroah-Hartman wrote:
From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
This reverts commit bd2573ee0f91c0e6d2bee8599110453e2909060e which is
commit 5a51e1f2b083423f75145c512ee284862ab33854 upstream.
Marc writes:
can you please revert this patch, without the corresponding driver patch
[1] it breaks probing of the device, as no one populates the sub-nodes.
[1] 9cb6d1b39a8f ("soc: imx: imx8m-blk-ctrl: Scan subnodes and bind
drivers to them")
Would it make more sense to pick the missing blk-ctrl patch instead ?
If you want that to happen, sure, but it seems like a new feature to me,
right?
From my point of view, it is neither a feature nor a fix, although it
is banking toward the 'fix' side.
The imx8mp.dtsi 'simple-bus', 'syscon' triggers scanning of DT subnodes,
so that drivers can be bound to them, but it was deemed as not the right
approach by DT maintainers and moving that scan trigger into the block
controller driver was considered better. That block controller patch
adds that scan trigger.
So, I would argue that in order to make the kernel compatible with DTs
old and new, the block controller patch should be picked. Whether we
want to keep the DT patch that is being reverted here or not, I would
say either way is fine, as with the block controller patch in place, it
is just a clean up.
