Hi Fabrizio,
Thanks for working on that. I think your solution fixes the issue we had
on atmel boards [1] in a clean way (still have to find some time to
test it).
On Mon, 5 Nov 2018 13:56:48 +0000
Fabrizio Castro <fabrizio.castro@xxxxxxxxxxxxxx> wrote:
> @@ -441,6 +558,9 @@ static int sii902x_remove(struct i2c_client *client)
> {
> struct sii902x *sii902x = i2c_get_clientdata(client);
>
> + if (sii902x->i2cmux)
> + i2c_mux_del_adapters(sii902x->i2cmux);
> +
Just one tiny detail: I think you can safely call i2c_mux_del_adapters()
unconditionally since ->remove() will only be called if ->probe()
succeeded, which guarantees that sii902x->i2cmux != NULL.
> drm_bridge_remove(&sii902x->bridge);
>
> return 0;
[1]https://patchwork.kernel.org/patch/9532005/
