ac100 team mailing list archive

[Gilles Grandou <gilles@xxxxxxxxxxx>]

Le 08/08/2011 21:57, Marc Dietrich a écrit :
> mmh, I think I added some suspend code to the ec driver:
>
> static int tegra_nvec_suspend(struct platform_device *pdev, pm_message_t state)
> {
>          struct nvec_chip *nvec = platform_get_drvdata(pdev);
>
>          dev_dbg(nvec->dev, "suspending\n");
>          nvec_write_async(nvec, EC_DISABLE_EVENT_REPORTING, 3);
>          nvec_write_async(nvec, "\x04\x02", 2);
>          nvec_disable_i2c_slave(nvec);
>
>          return 0;
> }
>
> static int tegra_nvec_resume(struct platform_device *pdev)
> {
>          struct nvec_chip *nvec = platform_get_drvdata(pdev);
>
>          dev_dbg(nvec->dev, "resuming\n");
>          tegra_init_i2c_slave(nvec);
>          nvec_write_async(nvec, EC_ENABLE_EVENT_REPORTING, 3);
>
>          return 0;
> }
>
> not sure if it's enough.

from what I remind, we're both around the same code ;)

> > 2- I don't know about voltage regulators status, but it's likely that
> > the cpu is still powered up.
>
> looking at harmony/seaboard, they make heavy use of the regulators (e.g. to
> switch off), but for ac100 everything seems to be magicly controlled by the ec.

that's not completly clear from what I read.
I'm afraid some black magic is done in userspace by nvrm_daemon...

> > 3- linked to 1, there is no wakeup event to resume the CPU.
>
> I think this needs a deeper review (especially check against ec_odm.c and
> paz00_diag.c

There are already known commands to programm the EC to wakeup, it's just 
that today we don't seem to be able to put the EC in suspend mode...

> > I've some very minor fixes in a gitorious branch, but 1-3 still needs to
> > be understood and done...
>
> cool! I will happily merge everything which improves the situtation.

nothing exciting, mainly the details I've seen while reading the toshiba 
.32 code. it's in something like dev-suspend branch in my personal 
gitorious tree.

Gilles.