Re: [PATCH] [PATCH v6] USB: serial: pl2303: Add new PID to support PL2303HXN (TYPE_HXN)

[Charles Yeh <charlesyeh522@gmail.com>]

>
>         pl2303_update_reg(serial, PL2303_HXN_FLOWCTRL,
>                         PL2303_HXN_FLOWCTRL_MASK,
>                         PL2303_HXN_CTRL_RTS_CTS);
>

Indent continuation lines at least two tabs<<--Done.

> I'm asking why you write the value 0 instead of 3 (or say, 0xfc)? Your
> documentation said bit 0 and 1 are used to reset the up and downstream
> pipes.
>
> To be more specific; what happens if I
>
>         1. set bit 0
>         2. clear bit 0?
>
> and leave the other bits alone (write back the same value, e.g. 0xfe).
>

You are right..
set "1" is reset.
set "0" is nothing.

I have used pl2303_update_reg instead pl2303_vendor_write which to reset
the upstream and downstream pipe data

Charles.


Johan Hovold <johan@kernel.org> 於 2019年7月1日 週一 下午11:29寫道：
>
> On Mon, Jul 01, 2019 at 11:11:02PM +0800, Charles Yeh wrote:
> > > > +             if (spriv->quirks & PL2303_QUIRK_LEGACY) {
> > > > +                     pl2303_update_reg(serial, 0, PL2303_FLOWCTRL_MASK,
> > > > +                             0x40);
> > >
> > > No need to break this line even if you end up with slightly more than 80
> > > chars.
> > >
> > OK. I will ignore "WARNING: line over 80 characters".
>
> Yeah, it's ok to go slightly above 80 chars when it improves readability.
>
> > > > +             } else if (spriv->type == &pl2303_type_data[TYPE_HXN]) {
> > > > +                     pl2303_update_reg(serial, PL2303_HXN_FLOWCTRL,
> > > > +                             PL2303_HXN_FLOWCTRL_MASK,
> > > > +                             PL2303_HXN_CTRL_RTS_CTS);
> > >
> > > Again, continuation lines should be indented at least two tabs further
> > > (you only use one tab now).
> > >
> >
> > I have done it the way you do today...but after checking
> > ./scripts/checkpatch.pl.. I got another warning message...
> > So I am a little confused now...
> > Previously before submitting.. must first pass ./scripts/checkpatch.pl
> > check. No ERROR, or WARRING message...
> >
> > I will return to the office tomorrow ... I will post another warning
> > message (according to the way you mentioned)
>
> checkpatch isn't always right. Just remember to indent continuation
> lines at least two tabs further, such as
>
>         pl2303_update_reg(serial, PL2303_HXN_FLOWCTRL,
>                         PL2303_HXN_FLOWCTRL_MASK,
>                         PL2303_HXN_CTRL_RTS_CTS);
>
> > >         } else if (spriv->type == &pl2303_type_data[TYPE_HXN]) {
> > >
> > > > +                     pl2303_vendor_write(serial, PL2303_HXN_RESET_CONTROL,
> > > > +                             0);
> > >
> > > You again completely ignored my question about why you're wring 0
> > > instead of 3 here.
> > >
> > > I'll ignore your patch until you explain.
> >
> > 3. In pl2303_open: Because TYPE_HXN is different from the instruction of reset
> >    down/up stream used by TYPE_HX.
> >    Therefore, we will also execute different instructions here.
> >    The default of chip Reset Control is 0xFF(TYPE_HXN), therefore we will
> >    write 0x00 to reset down/up stream(TYPE_HXN).
>
> I'm asking why you write the value 0 instead of 3 (or say, 0xfc)? Your
> documentation said bit 0 and 1 are used to reset the up and downstream
> pipes.
>
> To be more specific; what happens if I
>
>         1. set bit 0
>         2. clear bit 0?
>
> and leave the other bits alone (write back the same value, e.g. 0xfe).
>
> Johan