Re: [PATCH] mmc: core: Add support for the eMMC RTC feature in mmc_ops

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Tue, Dec 07, 2021 at 03:13:08PM +0530, Nishad Kamdar wrote:
> On Tue, Dec 07, 2021 at 09:33:46AM +0000, Avri Altman wrote:
> > > On Tue, Dec 07, 2021 at 08:28:42AM +0000, Avri Altman wrote:
> > > >
> > > > > This patch adds support to set the RTC information in the eMMC
> > > > > device. This is based on the JEDEC specification.
> > > > >
> > > > > There is no way however, to read the RTC time from the device. Hence
> > > > > we rely on the response of the CMD49 to confirm the completion of
> > > > > the operation.
> > > > >
> > > > > This patch has been tested successfully with the ioctl interface.
> > > > > This patch has also been tested suceessfully with all the three
> > > > > RTC_INFO_TYPEs.
> > > > If this is triggered from user-space via ioctl anyway, Why do we need
> > > > this command to be implemented in the kernel?
> > > > Why not just add this to mmc-utils?
> > > >
> > > > Thanks,
> > > > Avri
> > > As per the spec, B51: Section 6.6.35:
> > > Providing RTC info may be useful for internal maintainance operations.
> > > And the host should send it on the following events:
> > > - power-up
> > > - wake-up
> > > - Periodically
> > > Hence IMO, the Kernel would be the right place of peforming this operation.
> > But your patch doesn't do that, is it?
> >
> Yes, That's because this operation may be device specific. In order to know when
> to call this function may require eMMC firmware info.
> This patch only adds support so that if the info is made available
> in the future, a separate patch can be added to introduce the calling mechanism.

We do not add code that is not actually used in the kernel tree.

Please submit a user of this new function, otherwise there is no need
for it at all.

thanks,

greg k-h