Re: [PATCH v36 01/31] xhci: sideband: add initial api to register a secondary interrupter entity

[Puma Hsu]

On Thu, Mar 27, 2025 at 6:48 PM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Thu, Mar 27, 2025 at 06:14:00PM +0800, Puma Hsu wrote:
> > On Thu, Mar 27, 2025 at 3:02 PM Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> > >
> > > On Thu, Mar 27, 2025 at 02:27:00PM +0800, Puma Hsu wrote:
> > > > Hi,
> > > >
> > > > We have implemented and verified the USB audio offloading feature with
> > > > the xhci sideband driver on our Google Pixel products. We would
> > > > appreciate it if this solution can be accepted. Thank you all for the
> > > > work!
> > > >
> > >
> > > Great, can you properly send a "Tested-by:" line for this against the
> > > 00/XX email so that it will be properly saved?
> > >
> >
> > We(Google Pixel) only use the xhci sideband related changes and two
> > changes in the sound card driver. For the details, what we actually
> > tested are patch [01], [02], [03], [04], [05], [06], [08], and [12].
> > Do I still send the "Tested-by:" line to 00/31 email? Or should I just
> > send the "Tested-by:" line to the 8 changes above? (I added
> > "Tested-by" line for this [01/31] first.)
>
> Send it to the commits that you used if you only used portions.
>
> But that feels odd, why are you only using portions here?  Why not the
> whole thing?  Are you going to have to revert portions of this series in
> order for your device to work properly?
>

Google Pixels use Tensor instead of QCOM chip, so we won't enable
QCOM specific drivers from this series(i.g. patch [14~31]). What we
need from this series are common xhci stuff and portions of alsa. We
won't build QCOM driver even if this series is all merged, so we don't
have to revert anything.


> > > Also, I think a new version of the series is coming, can you test that
> > > to verify it works properly?  We have to wait until after -rc1 is out
> > > anyway.
> > >
> >
> > I think this v36 is the last version of the series as I discussed with
> > QCOM Wesley. And for sure I will test it if they do have a new
> > version.
>
> See:
>         https://lore.kernel.org/r/ee95520b-cdcc-4e10-a70e-683993cafe36@xxxxxxxxxxx
> for where the need for a new version is discussed.
>

Thank you for the information. It looks like QCOM specific, but I will
track and test if there was any change for common xhci in the new
version.

Thanks
Puma