2023-06-23 11:34:56

by Abel Vesa

[permalink] [raw]
Subject: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

There is a warning on dtbs check for sdm845, amongst other platforms,
about the reg-names being unevaluated. Fix that by adding reg-names to
the clocks and reg properties check for such platforms.

Fixes: 462c5c0aa798 ("dt-bindings: ufs: qcom,ufs: convert to dtschema")
Signed-off-by: Abel Vesa <[email protected]>
---
Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 4 ++++
1 file changed, 4 insertions(+)

diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
index 0209713d1f88..894b57117314 100644
--- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
+++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
@@ -166,6 +166,10 @@ allOf:
reg:
minItems: 2
maxItems: 2
+ reg-names:
+ items:
+ - const: std
+ - const: ice

- if:
properties:
--
2.34.1



2023-06-23 12:41:37

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On 23/06/2023 13:30, Abel Vesa wrote:
> There is a warning on dtbs check for sdm845, amongst other platforms,
> about the reg-names being unevaluated. Fix that by adding reg-names to
> the clocks and reg properties check for such platforms.
>
> Fixes: 462c5c0aa798 ("dt-bindings: ufs: qcom,ufs: convert to dtschema")
> Signed-off-by: Abel Vesa <[email protected]>
> ---
> Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> index 0209713d1f88..894b57117314 100644
> --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> @@ -166,6 +166,10 @@ allOf:
> reg:
> minItems: 2
> maxItems: 2
> + reg-names:
> + items:
> + - const: std
> + - const: ice

reg-names looks like a new property, so it should be defined in
top-level and just constrained per-variant.

Also there was similar approach:
https://lore.kernel.org/all/[email protected]/

but I guess no resends and it can be superseded.

Best regards,
Krzysztof


2023-06-23 12:50:21

by Luca Weiss

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Fri Jun 23, 2023 at 2:31 PM CEST, Krzysztof Kozlowski wrote:
> On 23/06/2023 13:30, Abel Vesa wrote:
> > There is a warning on dtbs check for sdm845, amongst other platforms,
> > about the reg-names being unevaluated. Fix that by adding reg-names to
> > the clocks and reg properties check for such platforms.
> >
> > Fixes: 462c5c0aa798 ("dt-bindings: ufs: qcom,ufs: convert to dtschema")
> > Signed-off-by: Abel Vesa <[email protected]>
> > ---
> > Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 4 ++++
> > 1 file changed, 4 insertions(+)
> >
> > diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > index 0209713d1f88..894b57117314 100644
> > --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > @@ -166,6 +166,10 @@ allOf:
> > reg:
> > minItems: 2
> > maxItems: 2
> > + reg-names:
> > + items:
> > + - const: std
> > + - const: ice
>
> reg-names looks like a new property, so it should be defined in
> top-level and just constrained per-variant.
>
> Also there was similar approach:
> https://lore.kernel.org/all/[email protected]/
>
> but I guess no resends and it can be superseded.

Right, the patches got reviews but was never applied... I really need to
find a strategy to keep track of sent patches until they're applied with
my work mailbox, it's not the first time that a patch has gotten
forgotten.

With my private mailbox I just have a different folder for patches that
have been sent which I archive once they're applied, but with work GMail
I don't see how I can easily replicate this since it's also not grouping
threads properly.

Also patch 4/5 in this series has an equivalent from me:
https://lore.kernel.org/all/[email protected]/
^ this might also be preferable since I guess it doesn't break
dt_binding_check?

Regards
Luca

>
> Best regards,
> Krzysztof


2023-06-23 21:43:16

by Rob Herring (Arm)

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Fri, Jun 23, 2023 at 02:38:04PM +0200, Luca Weiss wrote:
> On Fri Jun 23, 2023 at 2:31 PM CEST, Krzysztof Kozlowski wrote:
> > On 23/06/2023 13:30, Abel Vesa wrote:
> > > There is a warning on dtbs check for sdm845, amongst other platforms,
> > > about the reg-names being unevaluated. Fix that by adding reg-names to
> > > the clocks and reg properties check for such platforms.
> > >
> > > Fixes: 462c5c0aa798 ("dt-bindings: ufs: qcom,ufs: convert to dtschema")
> > > Signed-off-by: Abel Vesa <[email protected]>
> > > ---
> > > Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 4 ++++
> > > 1 file changed, 4 insertions(+)
> > >
> > > diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > index 0209713d1f88..894b57117314 100644
> > > --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > @@ -166,6 +166,10 @@ allOf:
> > > reg:
> > > minItems: 2
> > > maxItems: 2
> > > + reg-names:
> > > + items:
> > > + - const: std
> > > + - const: ice
> >
> > reg-names looks like a new property, so it should be defined in
> > top-level and just constrained per-variant.
> >
> > Also there was similar approach:
> > https://lore.kernel.org/all/[email protected]/
> >
> > but I guess no resends and it can be superseded.
>
> Right, the patches got reviews but was never applied... I really need to
> find a strategy to keep track of sent patches until they're applied with
> my work mailbox, it's not the first time that a patch has gotten
> forgotten.

There was an error reported on the above series. Why would it be
applied?

That said, I'm not sure SCSI maintainers consistently apply DT only
patch series.

> With my private mailbox I just have a different folder for patches that
> have been sent which I archive once they're applied, but with work GMail
> I don't see how I can easily replicate this since it's also not grouping
> threads properly.

Yeah, GMail sucks for that. I use 'lei' to get all my patches and
replies to them (though its caching will miss replies). Then I delete
them from the mbox when they are applied or otherwise finished. lei
updates won't re-add them to the mbox.

Rob

2023-06-24 07:55:03

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On 23/06/2023 23:17, Rob Herring wrote:
>> With my private mailbox I just have a different folder for patches that
>> have been sent which I archive once they're applied, but with work GMail
>> I don't see how I can easily replicate this since it's also not grouping
>> threads properly.
>
> Yeah, GMail sucks for that. I use 'lei' to get all my patches and
> replies to them (though its caching will miss replies). Then I delete
> them from the mbox when they are applied or otherwise finished. lei
> updates won't re-add them to the mbox.

That's interesting approach. What's your lei search query for getting
your patches? "f:rob" would get all your threads you participated in.

Best regards,
Krzysztof


2023-06-24 13:17:35

by James Bottomley

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Fri, 2023-06-23 at 14:38 +0200, Luca Weiss wrote:
> With my private mailbox I just have a different folder for patches
> that have been sent which I archive once they're applied, but with
> work GMail I don't see how I can easily replicate this since it's
> also not grouping threads properly.

I have something similar, but instead of multiple folders, I use imap
labels to achieve the same thing (and then evolution search folders to
sort out the labels). I believe GMail has some primitive labelling
system that actually works (unlike exchange), so you might be able to
get a scheme like that to work.

For my mobile phone, which doesn't have the sophisticated search
folders evolution does, I use dovecot virtual folders to achieve the
same effect. I'm afraid I don't think GMail has any equivalent of
this.

James


2023-06-26 06:57:11

by Luca Weiss

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

Hi Rob,

On Fri Jun 23, 2023 at 11:17 PM CEST, Rob Herring wrote:
> On Fri, Jun 23, 2023 at 02:38:04PM +0200, Luca Weiss wrote:
> > On Fri Jun 23, 2023 at 2:31 PM CEST, Krzysztof Kozlowski wrote:
> > > On 23/06/2023 13:30, Abel Vesa wrote:
> > > > There is a warning on dtbs check for sdm845, amongst other platforms,
> > > > about the reg-names being unevaluated. Fix that by adding reg-names to
> > > > the clocks and reg properties check for such platforms.
> > > >
> > > > Fixes: 462c5c0aa798 ("dt-bindings: ufs: qcom,ufs: convert to dtschema")
> > > > Signed-off-by: Abel Vesa <[email protected]>
> > > > ---
> > > > Documentation/devicetree/bindings/ufs/qcom,ufs.yaml | 4 ++++
> > > > 1 file changed, 4 insertions(+)
> > > >
> > > > diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > > index 0209713d1f88..894b57117314 100644
> > > > --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > > +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
> > > > @@ -166,6 +166,10 @@ allOf:
> > > > reg:
> > > > minItems: 2
> > > > maxItems: 2
> > > > + reg-names:
> > > > + items:
> > > > + - const: std
> > > > + - const: ice
> > >
> > > reg-names looks like a new property, so it should be defined in
> > > top-level and just constrained per-variant.
> > >
> > > Also there was similar approach:
> > > https://lore.kernel.org/all/[email protected]/
> > >
> > > but I guess no resends and it can be superseded.
> >
> > Right, the patches got reviews but was never applied... I really need to
> > find a strategy to keep track of sent patches until they're applied with
> > my work mailbox, it's not the first time that a patch has gotten
> > forgotten.
>
> There was an error reported on the above series. Why would it be
> applied?

The error report at [0] complains about reg-names but I'm quite sure
that patch 2/3 resolves this error. Does your bot only apply one patch
at a time and run the check or apply all of them and then run it? It's
been a while but I'm fairly sure I ran all of the checks before sending
since I also documented some other patches in the cover letter there.

[0] https://lore.kernel.org/all/[email protected]/

>
> That said, I'm not sure SCSI maintainers consistently apply DT only
> patch series.
>
> > With my private mailbox I just have a different folder for patches that
> > have been sent which I archive once they're applied, but with work GMail
> > I don't see how I can easily replicate this since it's also not grouping
> > threads properly.
>
> Yeah, GMail sucks for that. I use 'lei' to get all my patches and
> replies to them (though its caching will miss replies). Then I delete
> them from the mbox when they are applied or otherwise finished. lei
> updates won't re-add them to the mbox.

I'll try to figure something out with GMail.. Perhaps just adding a
label "not yet applied" which I manually remove once it's applied would
be sufficient.

Regards
Luca

>
> Rob


2023-06-26 07:50:57

by Krzysztof Kozlowski

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On 26/06/2023 08:38, Luca Weiss wrote:
>>>> but I guess no resends and it can be superseded.
>>>
>>> Right, the patches got reviews but was never applied... I really need to
>>> find a strategy to keep track of sent patches until they're applied with
>>> my work mailbox, it's not the first time that a patch has gotten
>>> forgotten.
>>
>> There was an error reported on the above series. Why would it be
>> applied?
>
> The error report at [0] complains about reg-names but I'm quite sure
> that patch 2/3 resolves this error. Does your bot only apply one patch
> at a time and run the check or apply all of them and then run it? It's
> been a while but I'm fairly sure I ran all of the checks before sending
> since I also documented some other patches in the cover letter there.

You did it in cover letter, not in the patch, so there is no dependency
for bots recorded.

>
> [0] https://lore.kernel.org/all/[email protected]/

Your patch 2/3 could not beĀ applied to any tree. 3/3 applied but without
previous one caused warnings.




Best regards,
Krzysztof


2023-06-26 08:30:01

by Luca Weiss

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Mon Jun 26, 2023 at 9:41 AM CEST, Krzysztof Kozlowski wrote:
> On 26/06/2023 08:38, Luca Weiss wrote:
> >>>> but I guess no resends and it can be superseded.
> >>>
> >>> Right, the patches got reviews but was never applied... I really need to
> >>> find a strategy to keep track of sent patches until they're applied with
> >>> my work mailbox, it's not the first time that a patch has gotten
> >>> forgotten.
> >>
> >> There was an error reported on the above series. Why would it be
> >> applied?
> >
> > The error report at [0] complains about reg-names but I'm quite sure
> > that patch 2/3 resolves this error. Does your bot only apply one patch
> > at a time and run the check or apply all of them and then run it? It's
> > been a while but I'm fairly sure I ran all of the checks before sending
> > since I also documented some other patches in the cover letter there.
>
> You did it in cover letter, not in the patch, so there is no dependency
> for bots recorded.

I'm not aware how to put extra comments into a patch in a series with
b4, at least last time I checked I don't think it was possible? But I
also thought the cover letter was exactly there for giving some
background of the series and documenting any dependencies on other
patches.

>
> >
> > [0] https://lore.kernel.org/all/[email protected]/
>
> Your patch 2/3 could not beĀ applied to any tree. 3/3 applied but without
> previous one caused warnings.

Anyways, just resent the series as v4, maybe this time it can get picked
up... Should have enough reviews by now :)

Regards
Luca

>
>
>
>
> Best regards,
> Krzysztof


2023-06-27 15:34:08

by Rob Herring (Arm)

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Sat, Jun 24, 2023 at 09:49:12AM +0200, Krzysztof Kozlowski wrote:
> On 23/06/2023 23:17, Rob Herring wrote:
> >> With my private mailbox I just have a different folder for patches that
> >> have been sent which I archive once they're applied, but with work GMail
> >> I don't see how I can easily replicate this since it's also not grouping
> >> threads properly.
> >
> > Yeah, GMail sucks for that. I use 'lei' to get all my patches and
> > replies to them (though its caching will miss replies). Then I delete
> > them from the mbox when they are applied or otherwise finished. lei
> > updates won't re-add them to the mbox.
>
> That's interesting approach. What's your lei search query for getting
> your patches? "f:rob" would get all your threads you participated in.

This is what I have:

q = (dfn:drivers OR dfn:sound OR dfn:tools OR dfn:kernel OR \
dfn:arch OR dfn:Documentation OR dfn:include OR dfn:scripts) AND \
f:[email protected] AND rt:3.month.ago..

Really, I'd like a 'is a patch' flag or 'dfn:*' or 'dfn:/' here, but I
didn't convince the lei maintainer such a thing is needed. Sigh.

Also, you have to disable lei's caching with the --remote-fudge-time
option because it will miss replies to the matching query. Also reported
and not fixed...

Rob

2023-06-27 16:17:36

by Rob Herring (Arm)

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Mon, Jun 26, 2023 at 10:19:09AM +0200, Luca Weiss wrote:
> On Mon Jun 26, 2023 at 9:41 AM CEST, Krzysztof Kozlowski wrote:
> > On 26/06/2023 08:38, Luca Weiss wrote:
> > >>>> but I guess no resends and it can be superseded.
> > >>>
> > >>> Right, the patches got reviews but was never applied... I really need to
> > >>> find a strategy to keep track of sent patches until they're applied with
> > >>> my work mailbox, it's not the first time that a patch has gotten
> > >>> forgotten.
> > >>
> > >> There was an error reported on the above series. Why would it be
> > >> applied?
> > >
> > > The error report at [0] complains about reg-names but I'm quite sure
> > > that patch 2/3 resolves this error. Does your bot only apply one patch
> > > at a time and run the check or apply all of them and then run it? It's
> > > been a while but I'm fairly sure I ran all of the checks before sending
> > > since I also documented some other patches in the cover letter there.
> >
> > You did it in cover letter, not in the patch, so there is no dependency
> > for bots recorded.
>
> I'm not aware how to put extra comments into a patch in a series with
> b4, at least last time I checked I don't think it was possible? But I
> also thought the cover letter was exactly there for giving some
> background of the series and documenting any dependencies on other
> patches.

I just put a '---' line and comments after that in the commit messages.
That works fine unless your git branch is going upstream directly (i.e.
via a pull request). Even when I apply my own patches, I get them from
lore and apply so the comments are dropped.

Rob

2023-06-28 09:29:11

by Luca Weiss

[permalink] [raw]
Subject: Re: [PATCH 5/5] scsi: dt-bindings: ufs: qcom: Fix warning for sdm845 by adding reg-names

On Tue Jun 27, 2023 at 5:18 PM CEST, Rob Herring wrote:
> On Mon, Jun 26, 2023 at 10:19:09AM +0200, Luca Weiss wrote:
> > On Mon Jun 26, 2023 at 9:41 AM CEST, Krzysztof Kozlowski wrote:
> > > On 26/06/2023 08:38, Luca Weiss wrote:
> > > >>>> but I guess no resends and it can be superseded.
> > > >>>
> > > >>> Right, the patches got reviews but was never applied... I really need to
> > > >>> find a strategy to keep track of sent patches until they're applied with
> > > >>> my work mailbox, it's not the first time that a patch has gotten
> > > >>> forgotten.
> > > >>
> > > >> There was an error reported on the above series. Why would it be
> > > >> applied?
> > > >
> > > > The error report at [0] complains about reg-names but I'm quite sure
> > > > that patch 2/3 resolves this error. Does your bot only apply one patch
> > > > at a time and run the check or apply all of them and then run it? It's
> > > > been a while but I'm fairly sure I ran all of the checks before sending
> > > > since I also documented some other patches in the cover letter there.
> > >
> > > You did it in cover letter, not in the patch, so there is no dependency
> > > for bots recorded.
> >
> > I'm not aware how to put extra comments into a patch in a series with
> > b4, at least last time I checked I don't think it was possible? But I
> > also thought the cover letter was exactly there for giving some
> > background of the series and documenting any dependencies on other
> > patches.
>
> I just put a '---' line and comments after that in the commit messages.
> That works fine unless your git branch is going upstream directly (i.e.
> via a pull request). Even when I apply my own patches, I get them from
> lore and apply so the comments are dropped.

Ah, didn't know this was possible/supported. In the past with git
send-email directly I'd edit the patch file and add some text under the
"---" manually but wasn't aware you can put it directly in the commit
message. But I guess if it produces the same output either way it makes
sense.

I won't have a problem with pull requests since I'm just a normal patch
submitter ;)

Thanks for the advice!

Regards
Luca

>
> Rob