On Fri, Mar 06, 2015 at 12:21:42PM +0200, Jyri Sarha wrote:
On 03/06/15 11:58, Russell King - ARM Linux wrote:
On Fri, Mar 06, 2015 at 10:33:27AM +0200, Jyri Sarha wrote:
Would it be Ok to add a check that master->ops->add_components is defined, before calling it in find_componets() (drivers/base/component.c:120) and return 0 if it is not?
No:
http://ftp.arm.linux.org.uk/cgit/linux-arm.git/commit/?h=8c4e8764a7e3
also:
http://ftp.arm.linux.org.uk/cgit/linux-arm.git/log/?h=11eda5aaf41e
is what's planned to be merged when I can get a round tuit, and people stop using the old methods.
Ok, but could it still be allowed to add a master without any components (match == NULL)?
Or do I have to handle the configurations without any components separately?
That's not a decision I want to make in my current state. Give me a couple of week or two and re-ping me.
http://archive.arm.linux.org.uk/lurker/message/20150306.102749.fcabd2bf.en.h...
(and the reason becomes self-evident when you realise that message did not go to the right list on Tuesday evening when it was meant to.)