On Tue, Jan 24, 2017 at 8:59 AM, Tony Lindgren <tony@xxxxxxxxxxx> wrote:
> * Adam Ford <aford173@xxxxxxxxx> [170124 05:49]:
>> On Mon, Jan 23, 2017 at 3:01 PM, Tony Lindgren <tony@xxxxxxxxxxx> wrote:
>> > * Adam Ford <aford173@xxxxxxxxx> [170123 11:51]:
>> >> On Mon, Jan 23, 2017 at 12:58 PM, Tony Lindgren <tony@xxxxxxxxxxx> wrote:
>> >> > * aford173@xxxxxxxxx <aford173@xxxxxxxxx> [170120 10:10]:
>> >> >> From: Adam Ford <aford173@xxxxxxxxx>
>> >> >>
>> >> >> This fixes commit ab8dd3aed011 ("ARM: DTS: Add minimal Support for
>> >> >> Logic PD DM3730 SOM-LV") where the system may not successfully
>> >> >> start after a reboot.
>> >> >
>> >> > Hmm so what goes wrong here? And which compatible do you end up
>> >> > using that works?
>> >>
>> >> The issue was the bootloader (MLO) would load, but it would throw an
>> >> error unable to load anything (like U-boot) from the MMC. After
>> >> reading the binding, the twl4030-power-reset seemed like a good
>> >> option, so I added it.
>> >
>> > OK..
>> >
>> >> I didn't think the compatible options were mutually exclusive, because
>> >> the original device tree you did the for the Torpedo had both
>> >> "ti,twl4030-power-idle-osc-off", "ti,twl4030-power-idle" enabled, so I
>> >> added the "ti,twl4030-power-reset" option.
>> >>
>> >> Looking at binding, I thought I should have the more generic
>> >> compatible = "ti,twl4030-power" added, but it doesn't appear to change
>> >> the behavior, so I'm ok with removing that.
>> >>
>> >> Either way, it appears as if II need the "ti,twl4030-power-reset" to
>> >> address the bootloader hanging. Does that mean I have to remove the
>> >> other "ti,twl4030-power-idle-osc-off", "ti,twl4030-power-idle" flags?
>> >
>> > Hmm I think we may have a bug somewhere. The "ti,twl4030-power-reset"
>> > should be a subset of "ti,twl4030-power-idle" which is a subset of
>> > "ti,twl4030-power-idle-osc-off".
>> >
>> > Maybe try adding some printks to the driver to see if the power-reset
>> > features are configured with power-idle-osc-off?
>>
>> I did some testing with the stable 4.9.y branch. I set
>>
>>
>> Using only the power-idle-osc-off, I was able to observe the following:
>>
>> [ 2.425842] TWL4030_WAKEUP12_SCRIPT
>> [ 2.434448] TWL4030_WAKEUP3_SCRIPT
>> [ 2.453979] TWL4030_WRST_SCRIPT
>> [ 2.460754] TWL4030_SLEEP_SCRIPT
>>
>> Reboot was not successful.
>>
>> The "ti,twl4030-power-reset" flag only showed
>> [ 2.439392] TWL4030_WRST_SCRIPT
>>
>> Reboot was successful.
>>
>>
>> and with just the "ti,twl4030-power-idle" flag, I received:
>>
>> [ 2.426177] TWL4030_WAKEUP12_SCRIPT
>> [ 2.434783] TWL4030_WAKEUP3_SCRIPT
>> [ 2.454315] TWL4030_WRST_SCRIPT
>> [ 2.461090] TWL4030_SLEEP_SCRIPT
>>
>> Reboot was not successful.
>
> OK so the TWL4030_WRST_SCRIPT is configured in all cases. So most
> likely it's one of the regulators that either can't idle or can't be
> shut off during idle.
>
>> > From PM point of view, you really want to use all features with
>> > "ti,twl4030-power-idle-osc-off" if your hardware allows cutting off
>> > the oscillator during idle :)
>>
>> I haven't observed any issues with "ti,twl4030-power-idle-osc-off"
>> during normal operation or during sleep, but if the system cannot
>> reboot, we can't use it.
>>
>> Do you have any suggestions? If not, I may have to switch the
>> compatible flat to "ti,twl4030-power-reset" for compatibility reason.
>> I have a couple customers inquiring with the same reboot issue, and
>> this fixes their issue in both instances.
>
> Yes comment out one line at a time in omap3_idle_rconfig and see
> which regulator needs to stay on for you for reboot to work.
I appreciate you help on this. I commented out every line and the
issue never resolved itself. If anything commenting out
TWL_REMAP_OFF(RES_VPLL1, DEV_GRP_P1, 3, 1), made the
problem worse, because MLO didn't even load.
> Also see if adding or leaving out "ti,system-power-controller" or
> "ti,use_poweroff" makes a difference.
I tried removing that before I did anything else, and it didn't appear
to make any difference. If you have any other ideas, I'm willing to try stuff.
I'll continue to dig around in here.
> Regards,
>
> Tony
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
