On Tue, Sep 29, 2015 at 3:05 AM, Tomeu Vizoso
[off-list ref] wrote:

On 26 September 2015 at 20:15, Rob Herring [off-list ref] wrote:

quoted

On 09/21/2015 09:03 AM, Tomeu Vizoso wrote:

quoted

Add a field to struct device that instructs the device-driver core to
defer the probe of this device until the late_initcall level.

This is true until the next patch with moves deferred probe processing
to device_initcall_sync. So disabling this option alone won't totally
revert to current behaviour. I guess patch 21 could be reverted if
necessary.

Actually, the goal with that commit was to prevent potential problems
due to the increased pressure on late_initcall, as suggested by
Grygorii Strashko, but I haven't found yet any evidence of it being
needed, and in my testing the series boot all boards in kernelci with
or without this commit. So I would just not commit it for now and only
consider applying it later if someone reports a problem.

I had similar concerns with assumptions about ordering WRT
late_initcall. I would keep this for now.

Rob