Since dt maintainer give comments at old thread
https://lore.kernel.org/linux-usb/[email protected]/
The patch v4 already merged.
https://lore.kernel.org/linux-usb/[email protected]/
So submit new patch to rename snps,host-vbus-glitches-quirk to
snps,host-vbus-glitches to align dt maintainer's comments.
Signed-off-by: Frank Li <[email protected]>
---
Frank Li (2):
dt-bindings: usb: dwc3: drop 'quirk' suffix at snps,host-vbus-glitches-quirk
usb: dwc3: drop 'quirk' suffix at snps,host-vbus-glitches-quirk
Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 5 +++--
drivers/usb/dwc3/core.c | 2 +-
2 files changed, 4 insertions(+), 3 deletions(-)
---
base-commit: 2ae0a045e6814c8c1d676d6153c605a65746aa29
change-id: 20240207-vbus-glitch-97fd340d8a66
Best regards,
--
Frank Li <[email protected]>
On Wed, Feb 07, 2024 at 05:00:17PM -0500, Frank Li wrote:
> Since dt maintainer give comments at old thread
> https://lore.kernel.org/linux-usb/[email protected]/
>
> The patch v4 already merged.
> https://lore.kernel.org/linux-usb/[email protected]/
>
> So submit new patch to rename snps,host-vbus-glitches-quirk to
> snps,host-vbus-glitches to align dt maintainer's comments.
I thought the last comment left on the v1 was Thinh agreeing that a
DT property was not needed here and we should be able to apply this
conditionally?
On Wed, Feb 07, 2024 at 10:05:23PM +0000, Conor Dooley wrote:
> On Wed, Feb 07, 2024 at 05:00:17PM -0500, Frank Li wrote:
> > Since dt maintainer give comments at old thread
> > https://lore.kernel.org/linux-usb/[email protected]/
> >
> > The patch v4 already merged.
> > https://lore.kernel.org/linux-usb/[email protected]/
> >
> > So submit new patch to rename snps,host-vbus-glitches-quirk to
> > snps,host-vbus-glitches to align dt maintainer's comments.
>
> I thought the last comment left on the v1 was Thinh agreeing that a
> DT property was not needed here and we should be able to apply this
> conditionally?
I don't think so. This is workaround. We can use this track which chip
actually need this. If some year later, such chips already end of life.
We have chance to clear up these code. Otherwise, it will keep there for
ever.
And I am not sure that the side effect for other chips. Workaround should
be applied as less as possible.
Frank
On Wed, Feb 07, 2024 at 05:14:01PM -0500, Frank Li wrote:
> On Wed, Feb 07, 2024 at 10:05:23PM +0000, Conor Dooley wrote:
> > On Wed, Feb 07, 2024 at 05:00:17PM -0500, Frank Li wrote:
> > > Since dt maintainer give comments at old thread
> > > https://lore.kernel.org/linux-usb/[email protected]/
> > >
> > > The patch v4 already merged.
> > > https://lore.kernel.org/linux-usb/[email protected]/
> > >
> > > So submit new patch to rename snps,host-vbus-glitches-quirk to
> > > snps,host-vbus-glitches to align dt maintainer's comments.
> >
> > I thought the last comment left on the v1 was Thinh agreeing that a
> > DT property was not needed here and we should be able to apply this
> > conditionally?
>
> I don't think so. This is workaround. We can use this track which chip
> actually need this. If some year later, such chips already end of life.
> We have chance to clear up these code. Otherwise, it will keep there for
> ever.
> And I am not sure that the side effect for other chips. Workaround should
> be applied as less as possible.
I'd rather do it unconditionally if we can, but if you and Thinh think
that we cannot do it unconditionally then sure, keep the property.
Acked-by: Conor Dooley <[email protected]>
Cheers,
Conor.
On Thu, Feb 08, 2024, Conor Dooley wrote:
> On Wed, Feb 07, 2024 at 05:14:01PM -0500, Frank Li wrote:
> > On Wed, Feb 07, 2024 at 10:05:23PM +0000, Conor Dooley wrote:
> > > On Wed, Feb 07, 2024 at 05:00:17PM -0500, Frank Li wrote:
> > > > Since dt maintainer give comments at old thread
> > > > https://lore.kernel.org/linux-usb/[email protected]/
> > > >
> > > > The patch v4 already merged.
> > > > https://lore.kernel.org/linux-usb/[email protected]/
> > > >
> > > > So submit new patch to rename snps,host-vbus-glitches-quirk to
> > > > snps,host-vbus-glitches to align dt maintainer's comments.
> > >
> > > I thought the last comment left on the v1 was Thinh agreeing that a
> > > DT property was not needed here and we should be able to apply this
> > > conditionally?
> >
> > I don't think so. This is workaround. We can use this track which chip
> > actually need this. If some year later, such chips already end of life.
> > We have chance to clear up these code. Otherwise, it will keep there for
> > ever.
>
> > And I am not sure that the side effect for other chips. Workaround should
> > be applied as less as possible.
>
> I'd rather do it unconditionally if we can, but if you and Thinh think
> that we cannot do it unconditionally then sure, keep the property.
>
Perhaps I wasn't clear. I meant I agree that we don't need a new quirk
property. If anything, it should be safer to keep vbus disabled before
handing over to xhci driver. We should be able to do this
unconditionally.
BR,
Thinh
On Thu, Feb 08, 2024 at 10:28:22PM +0000, Thinh Nguyen wrote:
> On Thu, Feb 08, 2024, Conor Dooley wrote:
> > On Wed, Feb 07, 2024 at 05:14:01PM -0500, Frank Li wrote:
> > > On Wed, Feb 07, 2024 at 10:05:23PM +0000, Conor Dooley wrote:
> > > > On Wed, Feb 07, 2024 at 05:00:17PM -0500, Frank Li wrote:
> > > > > Since dt maintainer give comments at old thread
> > > > > https://lore.kernel.org/linux-usb/[email protected]/
> > > > >
> > > > > The patch v4 already merged.
> > > > > https://lore.kernel.org/linux-usb/[email protected]/
> > > > >
> > > > > So submit new patch to rename snps,host-vbus-glitches-quirk to
> > > > > snps,host-vbus-glitches to align dt maintainer's comments.
> > > >
> > > > I thought the last comment left on the v1 was Thinh agreeing that a
> > > > DT property was not needed here and we should be able to apply this
> > > > conditionally?
> > >
> > > I don't think so. This is workaround. We can use this track which chip
> > > actually need this. If some year later, such chips already end of life.
> > > We have chance to clear up these code. Otherwise, it will keep there for
> > > ever.
> >
> > > And I am not sure that the side effect for other chips. Workaround should
> > > be applied as less as possible.
> >
> > I'd rather do it unconditionally if we can, but if you and Thinh think
> > that we cannot do it unconditionally then sure, keep the property.
> >
>
> Perhaps I wasn't clear. I meant I agree that we don't need a new quirk
> property. If anything, it should be safer to keep vbus disabled before
> handing over to xhci driver. We should be able to do this
> unconditionally.
Okay, if everyone think unconditional is good. I can submit new patch to
remove these.
Frank
>
> BR,
> Thinh