Re: [PATCH 2/3] rtc: ds1685: use correct device struct to get platform device struct

[Thomas Bogendoerfer]

On Sat, 13 Apr 2019 01:17:19 -0400
Joshua Kinard <kumba@xxxxxxxxxx> wrote:

> On 4/12/2019 07:44, Thomas Bogendoerfer wrote:
> > On Fri, 12 Apr 2019 12:11:06 +0200
> > Alexandre Belloni <alexandre.belloni@xxxxxxxxxxx> wrote:
> > 
> >> Every patch need a commit message. Maybe you could indicate that this
> >> never gave any issue because parent is the first member of struct
> >> device.
> > 
> > I'll update the commit message, I get a nice stacktrace because of that
> > bug, so the path from work_queue calling ds1685_rtc_poweroff never worked.
> > 
> > Thomas.
> 
> I'll wager that's why the thing stopped powering off my Octane.  It *used*
> to work when I wrote the driver, but stopped after some unidentified point,
> and I never found the time to try and track it down.

calling ds1685_rtc_poweroff with the correct platform device works, so the bug is
not in the poweroff function but in the work queue.

> Which machine are you testing on, out of curiosity?

SGI Octane but I'm not setting prepare_poweroff.

Thomas.

-- 
SUSE Linux GmbH
GF: Felix Imendörffer, Mary Higgins, Sri Rasiah
HRB 21284 (AG Nürnberg)