RE: [PATCH v2 REGRESSION RESEND] usb: ohci-at91: use descriptor-based gpio APIs correctly

From: Wenyou.Yang
Date: Mon Dec 26 2016 - 00:19:40 EST




> -----Original Message-----
> From: Greg Kroah-Hartman [mailto:gregkh@xxxxxxxxxxxxxxxxxxx]
> Sent: 2016年12月24日 2:05
> To: Peter Rosin <peda@xxxxxxxxxx>
> Cc: linux-kernel@xxxxxxxxxxxxxxx; Alan Stern <stern@xxxxxxxxxxxxxxxxxxx>;
> linux-usb@xxxxxxxxxxxxxxx; Wenyou Yang - A41535
> <Wenyou.Yang@xxxxxxxxxxxxx>
> Subject: Re: [PATCH v2 REGRESSION RESEND] usb: ohci-at91: use descriptor-
> based gpio APIs correctly
>
> On Thu, Dec 22, 2016 at 09:38:08PM +0100, Peter Rosin wrote:
> > On 2016-12-22 18:27, Greg Kroah-Hartman wrote:
> > > On Thu, Dec 22, 2016 at 08:43:55AM +0100, Peter Rosin wrote:
> > >> The gpiod_get* function family does not want the -gpio suffix.
> > >> Use devm_gpiod_get_index_optional instead of devm_gpiod_get_optional.
> > >> The descriptor based APIs handle active high/low automatically.
> > >> The vbus-gpios are output, request enable while getting the gpio.
> > >> Don't try to get any vbus-gpios for ports outside num-ports.
> > >>
> > >> WTF? Big sigh.
> > >>
> > >> Fixes: 054d4b7b577d ("usb: ohci-at91: Use descriptor-based gpio
> > >> APIs")
> > >> Signed-off-by: Peter Rosin <peda@xxxxxxxxxx>
> > >> ---
> > >>
> > >> Hi!
> > >>
> > >> Resending this, since the only response I've got is that the merge
> > >> window is open and that this patch has been put on hold due to that.
> > >> But I think this regression (which happend between v4.9 and current
> > >> master) should be fixed before the merge window closes.
> > >
> > > I don't merge patches before -rc1 comes out, sorry, people should
> > > have tested linux-next better :)
> >
> > Neat, shift the blame for the shit patch over to the messenger :)
>
> Not at all, I blame the original developer :)

I am very very sorry. It is my ignorance. Sorry.

I tested this patch on linux-next branch on the SAMA5D3 and SAMA5D4 Xplained board.

>
> > > I'll catch up the first week of January, relax.
> >
> > As we all know, unrelated regressions are painful when trying to
> > locate other problems. It's seems silly to have a few extra for no good reason.
>
> I am supposed to be on vacation and not reading email until the 3rd of January,
> relax, we will catch up on stuff like this, and other minor things, soon enough, in
> plenty of time for 4.10-final.
>
> thanks,
>
> greg k-h


Best Regards,
Wenyou Yang