On Thursday, 23 April 2026 00:57:06 CEST Kuninori Morimoto wrote:
> 
> Hi Janusz
> 
> > > We need an update in _close() as well, that's getting the component from
> > > disc_data.
> > 
> > Thanks for spotting this.  We should now get the component from 
> > cx20442_codec.component in _close(), as we do in _receive(), and return if 
> > (tty->disc_data != &cx20442_codec), maybe also WARN_ON(), or even BUG_ON().  
> > Sorry for missing that.
> > 
> > Kuninori, will you take care yourself, or do you prefer me to send a fixed 
> > version?
> 
> OK, I will try.
> 
> I noticed that it seems it is inconsistent ways of handling tty->disc_data ?
> 	(A) is handling as "struct cx20442_codec",
> 	(B) is handling as "struct snd_soc_component",
> 
> 	static int cx81801_open(...)
> 	{
> 		...
> (A)		tty->disc_data = &cx20442_codec;
> 		...
> 	}
> 
> 	static void cx81801_close(...)
> 	{
> (B)		struct snd_soc_component *component = tty->disc_data;

Yeah, that's exactly what Mark pointed out as missed in my patch.

Thanks,
Janusz

> 		...
> 	};
> 
> Well.., it is the same things, anyway...
> It is confusable, I will fix this too.
> 
> (A)	struct cx20442_codec {
> (B)		struct snd_soc_component *component;
> 		bool ready;
> 	};
> 
> 
> Thank you for your help !!
> 
> Best regards
> ---
> Kuninori Morimoto
>