linux-next: build failure after merge of the pwm tree

linux-next: build failure after merge of the pwm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1607 bytes --]

Hi all,

After merging the pwm tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

error[E0277]: the trait bound `core::result::Result<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>: PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not satisfied
   --> drivers/pwm/pwm_th1520.rs:331:10
    |
331 |     ) -> Result<Pin<KBox<Self>>> {
    |          ^^^^^^^^^^^^^^^^^^^^^^^ unsatisfied trait bound
    |
    = help: the trait `PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not implemented for `Result<Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, Error>`
            but trait `PinInit<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>` is implemented for it
    = help: for that trait implementation, expected `core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>`, found `Th1520PwmPlatformDriver`
note: required by a bound in `kernel::platform::Driver::{synthetic#0}`
   --> rust/kernel/platform.rs:196:15
    |
196 |     ) -> impl PinInit<Self, Error>;
    |               ^^^^^^^^^^^^^^^^^^^^ required by this bound in `Driver::{synthetic#0}`

error: aborting due to 1 previous error

For more information about this error, try `rustc --explain E0277`.

Caused by commit

  fb3957af9ec6 ("pwm: Add Rust driver for T-HEAD TH1520 SoC")

presumably interacting with something merged earlier in my tree.
If someone could provide me with an appropriate merge resolution, I will
apply it.

I have used the pwm tree from next-20251024 for today.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pwm tree

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 2766 bytes --]

Hello Stephen,

thanks for your report (and creating next each day!)

On Mon, Oct 27, 2025 at 12:51:48PM +1100, Stephen Rothwell wrote:
> After merging the pwm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> error[E0277]: the trait bound `core::result::Result<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>: PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not satisfied
>    --> drivers/pwm/pwm_th1520.rs:331:10
>     |
> 331 |     ) -> Result<Pin<KBox<Self>>> {
>     |          ^^^^^^^^^^^^^^^^^^^^^^^ unsatisfied trait bound
>     |
>     = help: the trait `PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not implemented for `Result<Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, Error>`
>             but trait `PinInit<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>` is implemented for it
>     = help: for that trait implementation, expected `core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>`, found `Th1520PwmPlatformDriver`
> note: required by a bound in `kernel::platform::Driver::{synthetic#0}`
>    --> rust/kernel/platform.rs:196:15
>     |
> 196 |     ) -> impl PinInit<Self, Error>;
>     |               ^^^^^^^^^^^^^^^^^^^^ required by this bound in `Driver::{synthetic#0}`
> 
> error: aborting due to 1 previous error
> 
> For more information about this error, try `rustc --explain E0277`.
> 
> Caused by commit
> 
>   fb3957af9ec6 ("pwm: Add Rust driver for T-HEAD TH1520 SoC")
> 
> presumably interacting with something merged earlier in my tree.
> If someone could provide me with an appropriate merge resolution, I will
> apply it.

Having no relevant clue about Rust, I bisected that. The bisection points to
0242623384c7 ("rust: driver: let probe() return impl PinInit<Self, Error>").

Translating the changes that commit does to
drivers/gpu/drm/nova/driver.rs for drivers/pwm/pwm_th1520.rs results in:

diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
index 0ad38b78be85..dd554574adc8 100644
--- a/drivers/pwm/pwm_th1520.rs
+++ b/drivers/pwm/pwm_th1520.rs
@@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
     fn probe(
         pdev: &platform::Device<Core>,
         _id_info: Option<&Self::IdInfo>,
-    ) -> Result<Pin<KBox<Self>>> {
+    ) -> impl PinInit<Self, Error> {
         let dev = pdev.as_ref();
         let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
 
@@ -365,7 +365,7 @@ fn probe(
 
         pwm::Registration::register(dev, chip)?;
 
-        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
+        Ok(Th1520PwmPlatformDriver)
     }
 }
 
which builds again.

Best regards
Uwe

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pwm tree

[Danilo Krummrich]

On Mon Oct 27, 2025 at 9:11 AM CET, Uwe Kleine-König wrote:
> Translating the changes that commit does to
> drivers/gpu/drm/nova/driver.rs for drivers/pwm/pwm_th1520.rs results in:
>
> diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
> index 0ad38b78be85..dd554574adc8 100644
> --- a/drivers/pwm/pwm_th1520.rs
> +++ b/drivers/pwm/pwm_th1520.rs
> @@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
>      fn probe(
>          pdev: &platform::Device<Core>,
>          _id_info: Option<&Self::IdInfo>,
> -    ) -> Result<Pin<KBox<Self>>> {
> +    ) -> impl PinInit<Self, Error> {
>          let dev = pdev.as_ref();
>          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
>  
> @@ -365,7 +365,7 @@ fn probe(
>  
>          pwm::Registration::register(dev, chip)?;
>  
> -        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
> +        Ok(Th1520PwmPlatformDriver)
>      }
>  }

Yes, this looks good.

Thanks,
Danilo

Re: linux-next: build failure after merge of the pwm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1229 bytes --]

Hi all,

On Mon, 27 Oct 2025 10:38:05 +0100 "Danilo Krummrich" <dakr@kernel.org> wrote:
>
> On Mon Oct 27, 2025 at 9:11 AM CET, Uwe Kleine-König wrote:
> > Translating the changes that commit does to
> > drivers/gpu/drm/nova/driver.rs for drivers/pwm/pwm_th1520.rs results in:
> >
> > diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
> > index 0ad38b78be85..dd554574adc8 100644
> > --- a/drivers/pwm/pwm_th1520.rs
> > +++ b/drivers/pwm/pwm_th1520.rs
> > @@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
> >      fn probe(
> >          pdev: &platform::Device<Core>,
> >          _id_info: Option<&Self::IdInfo>,
> > -    ) -> Result<Pin<KBox<Self>>> {
> > +    ) -> impl PinInit<Self, Error> {
> >          let dev = pdev.as_ref();
> >          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
> >  
> > @@ -365,7 +365,7 @@ fn probe(
> >  
> >          pwm::Registration::register(dev, chip)?;
> >  
> > -        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
> > +        Ok(Th1520PwmPlatformDriver)
> >      }
> >  }  
> 
> Yes, this looks good.

OK, I have applied that to linux-next from today.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pwm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1449 bytes --]

Hi all,

On Tue, 28 Oct 2025 12:51:42 +1100 Stephen Rothwell <sfr@canb.auug.org.au> wrote:
>
> On Mon, 27 Oct 2025 10:38:05 +0100 "Danilo Krummrich" <dakr@kernel.org> wrote:
> >
> > On Mon Oct 27, 2025 at 9:11 AM CET, Uwe Kleine-König wrote:  
> > > Translating the changes that commit does to
> > > drivers/gpu/drm/nova/driver.rs for drivers/pwm/pwm_th1520.rs results in:
> > >
> > > diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
> > > index 0ad38b78be85..dd554574adc8 100644
> > > --- a/drivers/pwm/pwm_th1520.rs
> > > +++ b/drivers/pwm/pwm_th1520.rs
> > > @@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
> > >      fn probe(
> > >          pdev: &platform::Device<Core>,
> > >          _id_info: Option<&Self::IdInfo>,
> > > -    ) -> Result<Pin<KBox<Self>>> {
> > > +    ) -> impl PinInit<Self, Error> {
> > >          let dev = pdev.as_ref();
> > >          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
> > >  
> > > @@ -365,7 +365,7 @@ fn probe(
> > >  
> > >          pwm::Registration::register(dev, chip)?;
> > >  
> > > -        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
> > > +        Ok(Th1520PwmPlatformDriver)
> > >      }
> > >  }    
> > 
> > Yes, this looks good.  
> 
> OK, I have applied that to linux-next from today.

This now needs to be applied to the merge of the modules tree.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pwm tree

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 2245 bytes --]

Hello Greg,

On Mon, Oct 27, 2025 at 09:11:35AM +0100, Uwe Kleine-König wrote:
> On Mon, Oct 27, 2025 at 12:51:48PM +1100, Stephen Rothwell wrote:
> > After merging the pwm tree, today's linux-next build (x86_64 allmodconfig)
> > failed like this:
> > 
> > error[E0277]: the trait bound `core::result::Result<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>: PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not satisfied
> >    --> drivers/pwm/pwm_th1520.rs:331:10
> >     |
> > 331 |     ) -> Result<Pin<KBox<Self>>> {
> >     |          ^^^^^^^^^^^^^^^^^^^^^^^ unsatisfied trait bound
> >     |
> >     = help: the trait `PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not implemented for `Result<Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, Error>`
> >             but trait `PinInit<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>` is implemented for it
> >     = help: for that trait implementation, expected `core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>`, found `Th1520PwmPlatformDriver`
> > note: required by a bound in `kernel::platform::Driver::{synthetic#0}`
> >    --> rust/kernel/platform.rs:196:15
> >     |
> > 196 |     ) -> impl PinInit<Self, Error>;
> >     |               ^^^^^^^^^^^^^^^^^^^^ required by this bound in `Driver::{synthetic#0}`
> > 
> > error: aborting due to 1 previous error
> > 
> > For more information about this error, try `rustc --explain E0277`.
> > 
> > Caused by commit
> > 
> >   fb3957af9ec6 ("pwm: Add Rust driver for T-HEAD TH1520 SoC")
> > 
> > presumably interacting with something merged earlier in my tree.
> > If someone could provide me with an appropriate merge resolution, I will
> > apply it.
> 
> Having no relevant clue about Rust, I bisected that. The bisection points to
> 0242623384c7 ("rust: driver: let probe() return impl PinInit<Self, Error>").

To make it easy for Linus to merge my pwm tree during the next merge
window, it would be great if it could contain the conflict resolution.
Can I assume that this commit is stable, will be part of your next MR
and are you ok if I merge it into my tree with the fix for the conflict?

Best regards
Uwe

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the pwm tree

[Danilo Krummrich]

On Mon Oct 27, 2025 at 2:40 PM CET, Uwe Kleine-König wrote:
> To make it easy for Linus to merge my pwm tree during the next merge
> window, it would be great if it could contain the conflict resolution.
> Can I assume that this commit is stable, will be part of your next MR
> and are you ok if I merge it into my tree with the fix for the conflict?

I think the conflict is trivial enough to resolve, so I don't think Linus would
mind.

Anyways, the commit is stable, so feel free to go ahead. Also note, when I send
the PR (which I will likely do) I usually send a very early PR to Linus, right
before the merge window opens.

- Danilo

Re: linux-next: build failure after merge of the pwm tree'

[Alice Ryhl]

On Mon, Oct 27, 2025 at 09:11:31AM +0100, Uwe Kleine-König wrote:
> Hello Stephen,
> 
> thanks for your report (and creating next each day!)
> 
> On Mon, Oct 27, 2025 at 12:51:48PM +1100, Stephen Rothwell wrote:
> > After merging the pwm tree, today's linux-next build (x86_64 allmodconfig)
> > failed like this:
> > 
> > error[E0277]: the trait bound `core::result::Result<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>: PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not satisfied
> >    --> drivers/pwm/pwm_th1520.rs:331:10
> >     |
> > 331 |     ) -> Result<Pin<KBox<Self>>> {
> >     |          ^^^^^^^^^^^^^^^^^^^^^^^ unsatisfied trait bound
> >     |
> >     = help: the trait `PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not implemented for `Result<Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, Error>`
> >             but trait `PinInit<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>` is implemented for it
> >     = help: for that trait implementation, expected `core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>`, found `Th1520PwmPlatformDriver`
> > note: required by a bound in `kernel::platform::Driver::{synthetic#0}`
> >    --> rust/kernel/platform.rs:196:15
> >     |
> > 196 |     ) -> impl PinInit<Self, Error>;
> >     |               ^^^^^^^^^^^^^^^^^^^^ required by this bound in `Driver::{synthetic#0}`
> > 
> > error: aborting due to 1 previous error
> > 
> > For more information about this error, try `rustc --explain E0277`.
> > 
> > Caused by commit
> > 
> >   fb3957af9ec6 ("pwm: Add Rust driver for T-HEAD TH1520 SoC")
> > 
> > presumably interacting with something merged earlier in my tree.
> > If someone could provide me with an appropriate merge resolution, I will
> > apply it.
> 
> Having no relevant clue about Rust, I bisected that. The bisection points to
> 0242623384c7 ("rust: driver: let probe() return impl PinInit<Self, Error>").
> 
> Translating the changes that commit does to
> drivers/gpu/drm/nova/driver.rs for drivers/pwm/pwm_th1520.rs results in:
> 
> diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
> index 0ad38b78be85..dd554574adc8 100644
> --- a/drivers/pwm/pwm_th1520.rs
> +++ b/drivers/pwm/pwm_th1520.rs
> @@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
>      fn probe(
>          pdev: &platform::Device<Core>,
>          _id_info: Option<&Self::IdInfo>,
> -    ) -> Result<Pin<KBox<Self>>> {
> +    ) -> impl PinInit<Self, Error> {
>          let dev = pdev.as_ref();
>          let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
>  
> @@ -365,7 +365,7 @@ fn probe(
>  
>          pwm::Registration::register(dev, chip)?;
>  
> -        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
> +        Ok(Th1520PwmPlatformDriver)
>      }
>  }
>  
> which builds again.

This merge resolution looks correct.

Alice

Re: linux-next: build failure after merge of the pwm tree

[Michal Wilczynski]

On 10/27/25 02:51, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the pwm tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> error[E0277]: the trait bound `core::result::Result<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>: PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not satisfied
>    --> drivers/pwm/pwm_th1520.rs:331:10
>     |
> 331 |     ) -> Result<Pin<KBox<Self>>> {
>     |          ^^^^^^^^^^^^^^^^^^^^^^^ unsatisfied trait bound
>     |
>     = help: the trait `PinInit<Th1520PwmPlatformDriver, kernel::error::Error>` is not implemented for `Result<Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, Error>`
>             but trait `PinInit<core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>, kernel::error::Error>` is implemented for it
>     = help: for that trait implementation, expected `core::pin::Pin<Box<Th1520PwmPlatformDriver, Kmalloc>>`, found `Th1520PwmPlatformDriver`
> note: required by a bound in `kernel::platform::Driver::{synthetic#0}`
>    --> rust/kernel/platform.rs:196:15
>     |
> 196 |     ) -> impl PinInit<Self, Error>;
>     |               ^^^^^^^^^^^^^^^^^^^^ required by this bound in `Driver::{synthetic#0}`
> 
> error: aborting due to 1 previous error
> 
> For more information about this error, try `rustc --explain E0277`.
> 
> Caused by commit
> 
>   fb3957af9ec6 ("pwm: Add Rust driver for T-HEAD TH1520 SoC")
> 
> presumably interacting with something merged earlier in my tree.
> If someone could provide me with an appropriate merge resolution, I will
> apply it.
> 
> I have used the pwm tree from next-20251024 for today.
> 

Hi,
I believe this would fix the problem:
diff --git a/drivers/pwm/pwm_th1520.rs b/drivers/pwm/pwm_th1520.rs
index 0ad38b78be85..dd554574adc8 100644
--- a/drivers/pwm/pwm_th1520.rs
+++ b/drivers/pwm/pwm_th1520.rs
@@ -328,7 +328,7 @@ impl platform::Driver for Th1520PwmPlatformDriver {
     fn probe(
         pdev: &platform::Device<Core>,
         _id_info: Option<&Self::IdInfo>,
-    ) -> Result<Pin<KBox<Self>>> {
+    ) -> impl PinInit<Self, Error> {
         let dev = pdev.as_ref();
         let request = pdev.io_request_by_index(0).ok_or(ENODEV)?;
 
@@ -365,7 +365,7 @@ fn probe(
 
         pwm::Registration::register(dev, chip)?;
 
-        Ok(KBox::new(Th1520PwmPlatformDriver, GFP_KERNEL)?.into())
+        Ok(Th1520PwmPlatformDriver)
     }
 }
 

This is caused by change in fn probe, caused by this commit [1].

[1] - https://lore.kernel.org/all/20251016125544.15559-1-dakr@kernel.org/


Best regards,
-- 
Michal Wilczynski <m.wilczynski@samsung.com>