On Thu, Sep 18, 2014 at 08:34:41PM -0700, Guenter Roeck wrote:

On 09/18/2014 07:54 PM, Guenter Roeck wrote:

quoted

On 09/18/2014 03:32 PM, Josh Cartwright wrote:

quoted

By converting to the restart_notifier mechanism for restart, we allow
for other mechanisms, like the watchdog, to be used for restart in the
case where PS_HOLD has failed to reset the chip.

Choose priority 128, as according to documentation, this mechanism "is
sufficient to restart the entire system".

Cc: Pramod Gurav <redacted>
Cc: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Josh Cartwright <redacted>
---

[..]

quoted

quoted

 int msm_pinctrl_probe(struct platform_device *pdev,
               const struct msm_pinctrl_soc_data *soc_data)

@@ -943,6 +948,15 @@ int msm_pinctrl_remove(struct platform_device *pdev)

     pinctrl_unregister(pctrl->pctrl);

+    if (pctrl->restart_nb.notifier_call) {

One more comment: The conditional is really unnecessary. Just let
unregister_restart_handler deal with it ...

quoted

quoted

+        ret = unregister_restart_handler(&pctrl->restart_nb);

and just ignore the error return. The function will only return an error
if the entry was not found, and then it is a don't care.

Awesome, thanks.  I like simplifications :).  I was hijacking
notifier_call to indicate whether or not the restart notifier was
registered at all (because it's conditional on the particular chipset
having a "ps_hold" function).  But, nice to know
unregister_restart_handler() does the right thing if the handler wasn't
registered in the first place.

  Josh

-- 
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
hosted by The Linux Foundation