Hi Heiner,
On Wed, Sep 20, 2023 at 09:29:28AM +0200, Heiner Kallweit wrote:
> When entering the suspend callback, at first we should ensure that
> transfers are finished and I2C core can't start further transfers.
> Use i2c_mark_adapter_suspended() for this purpose, and complement it
> with a call to i2c_mark_adapter_resumed() in the resume path.
>
> Signed-off-by: Heiner Kallweit <hkallweit1@xxxxxxxxx>
> ---
> Rebased version of a previously discussed patch, now w/o touching
> the remove and shutdown path.
> ---
> drivers/i2c/busses/i2c-i801.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c
> index 6d02a8b88..26f132277 100644
> --- a/drivers/i2c/busses/i2c-i801.c
> +++ b/drivers/i2c/busses/i2c-i801.c
> @@ -1818,6 +1818,7 @@ static int i801_suspend(struct device *dev)
> {
> struct i801_priv *priv = dev_get_drvdata(dev);
>
> + i2c_mark_adapter_suspended(&priv->adapter);
> i801_restore_regs(priv);
>
> return 0;
> @@ -1829,6 +1830,7 @@ static int i801_resume(struct device *dev)
>
> i801_setup_hstcfg(priv);
> i801_enable_host_notify(&priv->adapter);
> + i2c_mark_adapter_resumed(&priv->adapter);
I think I already reviewed this patch previously and I had same
concerns as Jean, anyway,
Reviewed-by: Andi Shyti <andi.shyti@xxxxxxxxxx>
Andi
