Hi Marek,
On Fri, Aug 10, 2018 at 5:09 PM Marek Szyprowski
<m.szyprowski@xxxxxxxxxxx> wrote:
>
> Regulators, which are marked as 'on-in-suspend' seems to be critical for
> board operation, thus they must not be disabled anytime. This can be
> only assured by marking them as 'always-on', because otherwise some
> actions of their clients might result in turning them off. This patch
> restores suspend/resume operation on Peach-Pit Chromebook board. It
> partially reverts 'always-on' property removal done by the commit
> mentioned in the Fixes tag.
>
> Fixes: 665c441eea3d ("ARM: dts: exynos: Remove unneded always-on for regulators on Peach boards")
> Signed-off-by: Marek Szyprowski <m.szyprowski@xxxxxxxxxxx>
> ---
> This partial revert restores suspend/resume operation on Linux v4.7
> release. To make suspend/resume working on current linux-next (tested
> on next-20180806), one has to additionally revert following commits:
>
> 925ffff2ea8b "Input: cros_ec_keyb - remove check before calling pm_wakeup_event"
> 38ba34a43dbc "Input: cros_ec_keyb - mark cros_ec_keyb driver as wake enabled device."
> 47b7de2f6c18 "mmc: dw_mmc: fix falling from idmac to PIO mode when dw_mci_reset occurs"
> 2695578b896a "net: usbnet: fix potential deadlock on 32bit hosts"
>
> I'm investigating those issues and separate reports/fixes will be
> posted for them.
Without this patch I'm seeing tpm suspend failures on Peach Pi, but
after applying it suspend/resume seems to work fine. The 4 reverts
above are needed indeed, otherwise the device fails to wake up from
suspend.
Tested-by: Tomasz Figa <tfiga@xxxxxxxxxxxx>
Thanks for figuring this out.
Best regards,
Tomasz
