lkml.org 
[lkml]   [2021]   [Jul]   [14]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 29/29] arm64: dts: qcom: Harmonize DWC USB3 DT nodes name
Hello John,

On Tue, Jul 13, 2021 at 05:07:00PM -0700, John Stultz wrote:
> On Tue, Oct 20, 2020 at 5:10 AM Serge Semin
> <Sergey.Semin@baikalelectronics.ru> wrote:
> >
> > In accordance with the DWC USB3 bindings the corresponding node
> > name is suppose to comply with the Generic USB HCD DT schema, which
> > requires the USB nodes to have the name acceptable by the regexp:
> > "^usb(@.*)?" . Make sure the "snps,dwc3"-compatible nodes are correctly
> > named.
> >
> > Signed-off-by: Serge Semin <Sergey.Semin@baikalelectronics.ru>
>

> I know folks like to ignore this, but this patch breaks AOSP on db845c. :(

Sorry to hear that. Alas there is no much can be done about it.
DT-nodes name is a subject of DT-schema convention and as we've finally
unified USB-controller nodes it shouldn't be reverted back. You can
find the final USB-controller bindings in:
Documentation/devicetree/bindings/usb/usb.yaml
It strictly defines to have USB-nodes with names like "usb(@.*)".
Reverting this patch will cause the DT-bindings check procedure
failure. You can also find the naming convention defined in the
latest DT spec:
https://github.com/devicetree-org/devicetree-specification/releases/tag/v0.3

See also device-tree bindings requirements listed in the file:
Documentation/devicetree/bindings/writing-bindings.rst
It says: "DO use node names matching the class of the device. Many
standard names are defined in the DT Spec. If there isn't one,
consider adding it."

>
> In the exact same way an earlier patch broke HiKey960:
> https://lore.kernel.org/lkml/CALAqxLWGujgR7p8Vb5S_RimRVYxwm5XF-c4NkKgMH-43wEBaWg@mail.gmail.com/
>
> (which I still have to carry a revert for).
>
> I get that this change is useful so more dynamic userland can find
> devices using consistent naming with future kernels (but doesn't the
> dynamic userland have to handle the case for older kernels as well?)
> But for userland that uses static configs, its painful as updating
> userland to use the new node ids then causes older kernels to fail.
>
> I'm looking into how we might be able to probe and set the property
> dynamically, but AOSP's init system is far more aligned to static
> configs.
>

As Krzysztof said in
https://lore.kernel.org/lkml/20201221210423.GA2504@kozik-lap/
and Bjorn noted in his response to your email, the only way to solve
the problem is to fix the user-land app so one would be able to deal
with both old and new DT-nodes name. Alternatively you can just
replace the dts with older one, where the name still have
the "dwc3"-prefix.

-Sergey

> This will probably be ignored again, but it would be nice if we could
> have a release where DTS changes don't break userland for one of my
> boards. As it feels like its been awhile.
>
> thanks
> -john

\
 
 \ /
  Last update: 2021-07-14 14:56    [W:0.183 / U:0.056 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site