RE: [PATCH v2 3/5] usb: renesas_usbhs: Update usbhs pipe configuration for RZ/G2L family

Hi Geert,

Thanks for the feedback.

> -----Original Message-----
> From: Geert Uytterhoeven <[email protected]>
> Sent: Thursday, March 14, 2024 1:04 PM
> Subject: Re: [PATCH v2 3/5] usb: renesas_usbhs: Update usbhs pipe configuration for RZ/G2L family
> 
> Hi Biju,
> 
> On Thu, Mar 14, 2024 at 1:49 PM Biju Das <[email protected]> wrote:
> > > From: Geert Uytterhoeven <[email protected]> On Wed, Mar 13, 2024
> > > at 7:16 PM Biju Das <[email protected]> wrote:
> > > > From: Huy Nguyen <[email protected]> The RZ/G2L family
> > > > SoCs has 10 PIPE buffers compared to 16 pipe buffers on RZ/A2M.
> > > > Update the pipe configuration for RZ/G2L family SoCs and use
> > > > family SoC specific compatible to handle this difference.
> > > >
> > > > Added SoC specific compatible to OF table toavoid ABI breakage
> > > > with old DTB. To optimize memory usage the SoC specific compatible
> > > > will be removed later.
> > > >
> > > > Signed-off-by: Huy Nguyen <[email protected]>
> > > > Signed-off-by: Biju Das <[email protected]>
> 
> > > > --- a/drivers/usb/renesas_usbhs/common.c
> > > > +++ b/drivers/usb/renesas_usbhs/common.c
> > > > @@ -640,8 +656,13 @@ static int usbhs_probe(struct platform_device
> > > > *pdev)
> 
> > Since [1] is better compared to [2], if SH can work with [1], we can
> > replace [2] with [1], do we have any SH platform to test this?
> 
> I don't have an sh7757lcr or ecovec24 to test. But the risk looks low.
> 
> So it looks like a good idea to have two patches:
>   1. Improve usbhsc_default_pipe[] for isochronous transfers,
>   2. Fix support for RZ/G2L using the default 10-entry pipe.
> 

Agreed. Will send v3 with these changes.

Cheers,
Biju