RE: [PATCH/RFC] mmc: core: Issue power_off_notify for eMMC Suspend-to-RAM

[Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>]

Hi Ulf,

> From: Ulf Hansson, Sent: Monday, June 8, 2020 4:57 PM
> 
> On Tue, 19 May 2020 at 13:33, Yoshihiro Shimoda
> <yoshihiro.shimoda.uh@renesas.com> wrote:
> >
> > The commit 432356793415 ("mmc: core: Enable power_off_notify for
> > eMMC shutdown sequence") enabled the power off notification
> > even if MMC_CAP2_POWEROFF_NOTIFY (MMC_CAP2_FULL_PWR_CYCLE now) is
> > not set. However, the mmc core lacks to issue the power off
> > notificaiton when Suspend-to-{RAM,Disk} happens on the system.
> 
> This isn't an entirely correct description, I think.
> 
> If the host supports MMC_CAP2_FULL_PWR_CYCLE (both vmmc and vqmmc can
> be powered on/off), we use power-off-notification during system
> suspend, in case the eMMC card also supports it. Otherwise we send the
> sleep command.

Yes.

> This behaviour was decided on purpose and it's mainly because without
> MMC_CAP2_FULL_PWR_CYCLE, we assume that vqmmc remains always-on. In
> this case, it simply seemed better to use the sleep command, rather
> than the power-off-notification, as we aren't really going to do a
> full power off anyway.

I understood it. However, on my environment (r8a77951-salvator-xs),
while the board is entering Suspend-to-RAM, the vqmmc and vcc are turned off.
Should I add a new flag for such environment?

Best regards,
Yoshihiro Shimoda