[PATCH 1/2] rust: io: fix Bar reference in Io struct's comment

[PATCH 1/2] rust: io: fix Bar reference in Io struct's comment

[Marko Turk]

Bar was moved to a separate pci::io module, update the reference to it.

Signed-off-by: Marko Turk <mt@markoturk.info>
---
 rust/kernel/io.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/rust/kernel/io.rs b/rust/kernel/io.rs
index 98e8b84e68d1..08853f32dae6 100644
--- a/rust/kernel/io.rs
+++ b/rust/kernel/io.rs
@@ -87,7 +87,7 @@ pub fn maxsize(&self) -> usize {
 /// };
 /// use core::ops::Deref;
 ///
-/// // See also [`pci::Bar`] for a real example.
+/// // See also [`pci::io::Bar`] for a real example.
 /// struct IoMem<const SIZE: usize>(IoRaw<SIZE>);
 ///
 /// impl<const SIZE: usize> IoMem<SIZE> {
-- 
2.51.0

[PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Marko Turk]

inststance -> instance

Signed-off-by: Marko Turk <mt@markoturk.info>
---
 rust/kernel/pci/io.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/rust/kernel/pci/io.rs b/rust/kernel/pci/io.rs
index 0d55c3139b6f..fba746c4dc5d 100644
--- a/rust/kernel/pci/io.rs
+++ b/rust/kernel/pci/io.rs
@@ -20,7 +20,7 @@
 ///
 /// # Invariants
 ///
-/// `Bar` always holds an `IoRaw` inststance that holds a valid pointer to the start of the I/O
+/// `Bar` always holds an `IoRaw` instance that holds a valid pointer to the start of the I/O
 /// memory mapped PCI BAR and its size.
 pub struct Bar<const SIZE: usize = 0> {
     pdev: ARef<Device>,
-- 
2.51.0

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Danilo Krummrich]

On Sat Jan 3, 2026 at 3:31 PM CET, Marko Turk wrote:
> inststance -> instance

It's trivial in this case, but we usually write at least something along the
lines of "Fix a typo in the doc-comment of the Bar structure: 'inststance ->
instance'."

Please also add a corresponding Fixes: tag.

> Signed-off-by: Marko Turk <mt@markoturk.info>
> ---
>  rust/kernel/pci/io.rs | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/rust/kernel/pci/io.rs b/rust/kernel/pci/io.rs
> index 0d55c3139b6f..fba746c4dc5d 100644
> --- a/rust/kernel/pci/io.rs
> +++ b/rust/kernel/pci/io.rs
> @@ -20,7 +20,7 @@
>  ///
>  /// # Invariants
>  ///
> -/// `Bar` always holds an `IoRaw` inststance that holds a valid pointer to the start of the I/O
> +/// `Bar` always holds an `IoRaw` instance that holds a valid pointer to the start of the I/O
>  /// memory mapped PCI BAR and its size.
>  pub struct Bar<const SIZE: usize = 0> {
>      pdev: ARef<Device>,
> -- 
> 2.51.0

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Dirk Behme]

On 03.01.26 16:24, Danilo Krummrich wrote:
> On Sat Jan 3, 2026 at 3:31 PM CET, Marko Turk wrote:
>> inststance -> instance
> 
> It's trivial in this case, but we usually write at least something along the
> lines of "Fix a typo in the doc-comment of the Bar structure: 'inststance ->
> instance'."
> 
> Please also add a corresponding Fixes: tag.

While looking at this some days ago as well I came up with

Fixes: 3c2e31d717ac ("rust: pci: move I/O infrastructure to separate
file")

But that just moves the pre-existing typo from rust/kernel/pci.rs to
rust/kernel/pci/io.rs. So I'm unsure if that move-only commit should
be mentioned in Fixes:? Or if we should go back more to search for the
commit introducing this typo?

Best regards

Dirk

Btw: While we are at this file, do we want to add an 'is' in line 57
as well?

// `pdev` valid by the invariants of `Device`. => ... is valid ...


>> Signed-off-by: Marko Turk <mt@markoturk.info>
>> ---
>>  rust/kernel/pci/io.rs | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/rust/kernel/pci/io.rs b/rust/kernel/pci/io.rs
>> index 0d55c3139b6f..fba746c4dc5d 100644
>> --- a/rust/kernel/pci/io.rs
>> +++ b/rust/kernel/pci/io.rs
>> @@ -20,7 +20,7 @@
>>  ///
>>  /// # Invariants
>>  ///
>> -/// `Bar` always holds an `IoRaw` inststance that holds a valid pointer to the start of the I/O
>> +/// `Bar` always holds an `IoRaw` instance that holds a valid pointer to the start of the I/O
>>  /// memory mapped PCI BAR and its size.
>>  pub struct Bar<const SIZE: usize = 0> {
>>      pdev: ARef<Device>,
>> -- 
>> 2.51.0
> 
> 

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Marko Turk]

On Sat, Jan 03, 2026 at 04:38:36PM +0100, Dirk Behme wrote:
> On 03.01.26 16:24, Danilo Krummrich wrote:
> > On Sat Jan 3, 2026 at 3:31 PM CET, Marko Turk wrote:
> >> inststance -> instance
> > 
> > It's trivial in this case, but we usually write at least something along the
> > lines of "Fix a typo in the doc-comment of the Bar structure: 'inststance ->
> > instance'."
> > 
> > Please also add a corresponding Fixes: tag.
> 
> While looking at this some days ago as well I came up with
> 
> Fixes: 3c2e31d717ac ("rust: pci: move I/O infrastructure to separate
> file")
> 
> But that just moves the pre-existing typo from rust/kernel/pci.rs to
> rust/kernel/pci/io.rs. So I'm unsure if that move-only commit should
> be mentioned in Fixes:? Or if we should go back more to search for the
> commit introducing this typo?

The typo was introduced in the original commit where pci::Bar was added:
Fixes: bf9651f84b4e ("rust: pci: implement I/O mappable `pci::Bar`")

Should I use that for the Fixes: tag?

> 
> Best regards
> 
> Dirk
> 
> Btw: While we are at this file, do we want to add an 'is' in line 57
> as well?
> 
> // `pdev` valid by the invariants of `Device`. => ... is valid ...
> 

Should I do that in the same commit?

Marko

> 
> >> Signed-off-by: Marko Turk <mt@markoturk.info>
> >> ---
> >>  rust/kernel/pci/io.rs | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/rust/kernel/pci/io.rs b/rust/kernel/pci/io.rs
> >> index 0d55c3139b6f..fba746c4dc5d 100644
> >> --- a/rust/kernel/pci/io.rs
> >> +++ b/rust/kernel/pci/io.rs
> >> @@ -20,7 +20,7 @@
> >>  ///
> >>  /// # Invariants
> >>  ///
> >> -/// `Bar` always holds an `IoRaw` inststance that holds a valid pointer to the start of the I/O
> >> +/// `Bar` always holds an `IoRaw` instance that holds a valid pointer to the start of the I/O
> >>  /// memory mapped PCI BAR and its size.
> >>  pub struct Bar<const SIZE: usize = 0> {
> >>      pdev: ARef<Device>,
> >> -- 
> >> 2.51.0
> > 
> > 
> 

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Miguel Ojeda]

On Sat, Jan 3, 2026 at 10:16 PM Marko Turk <mt@markoturk.info> wrote:
>
> The typo was introduced in the original commit where pci::Bar was added:
> Fixes: bf9651f84b4e ("rust: pci: implement I/O mappable `pci::Bar`")
>
> Should I use that for the Fixes: tag?

I would add both, since it was added in both and thus different set of
stable releases may need to fix it differently (i.e. before and after
the move).

In this case, from a quick look, one is for the current release, so it
doesn't need backport, and the other would need a custom one (since
this commit wouldn't apply) if someone wants to do Option 3.

> Should I do that in the same commit?

If the patch targets the same set of stable releases with the given
`Fixes:` and is trivial (in this case, both are true from a quick
look), then it doesn't matter much either way and may simplify
handling. But, when those are not true, i.e. in the general case,
splitting where possible usually is better and simpler.

I hope that helps!

Cheers,
Miguel

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Danilo Krummrich]

(Cc: Greg, Sasha)

On Sun Jan 4, 2026 at 1:45 PM CET, Miguel Ojeda wrote:
> On Sat, Jan 3, 2026 at 10:16 PM Marko Turk <mt@markoturk.info> wrote:
>>
>> The typo was introduced in the original commit where pci::Bar was added:
>> Fixes: bf9651f84b4e ("rust: pci: implement I/O mappable `pci::Bar`")
>>
>> Should I use that for the Fixes: tag?
>
> I would add both, since it was added in both and thus different set of
> stable releases may need to fix it differently (i.e. before and after
> the move).

In general I prefer to only add a Fixes: tag for the commit that introduced the
issue.

> In this case, from a quick look, one is for the current release, so it
> doesn't need backport, and the other would need a custom one (since
> this commit wouldn't apply) if someone wants to do Option 3.

I could be wrong, but I think in trivial cases (such as code moves) the stable
team does derive custom commits themselves.

@Greg, Sasha: Is this something you prefer to do or is it something you just do
because it's easier / quicker than to get back and ask for a custom commit?

Again, I could also remember this wrongly, but I think I just recently reviewed
such a commit from Sasha. :)

>> Should I do that in the same commit?

That seems reasonable in this case, please do so.

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Miguel Ojeda]

On Sun, Jan 4, 2026 at 3:08 PM Danilo Krummrich <dakr@kernel.org> wrote:
>
> In general I prefer to only add a Fixes: tag for the commit that introduced the
> issue.

If their scripts track moves well, then it is great to avoid it, but I
am not sure how well that works or not or in which cases, i.e. it
could look like two different commits introduced the issue and thus
one backport could be missed. Not sure.

> Again, I could also remember this wrongly, but I think I just recently reviewed
> such a commit from Sasha. :)

Hmm... I also had a few cases where Sasha autoapplied, but in most
cases, I had to provide custom patches when they didn't apply cleanly,
even trivial ones.

Cheers,
Miguel

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Greg KH]

On Sun, Jan 04, 2026 at 07:30:22PM +0100, Miguel Ojeda wrote:
> On Sun, Jan 4, 2026 at 3:08 PM Danilo Krummrich <dakr@kernel.org> wrote:
> >
> > In general I prefer to only add a Fixes: tag for the commit that introduced the
> > issue.
> 
> If their scripts track moves well, then it is great to avoid it, but I
> am not sure how well that works or not or in which cases, i.e. it
> could look like two different commits introduced the issue and thus
> one backport could be missed. Not sure.
> 
> > Again, I could also remember this wrongly, but I think I just recently reviewed
> > such a commit from Sasha. :)
> 
> Hmm... I also had a few cases where Sasha autoapplied, but in most
> cases, I had to provide custom patches when they didn't apply cleanly,
> even trivial ones.

It all depends, sometimes we can handle file moves easily, sometimes we
can not.

But really, why is a comment typo being needed in stable kernels?

thanks,

greg k-h

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Miguel Ojeda]

On Mon, Jan 5, 2026 at 7:25 AM Greg KH <gregkh@linuxfoundation.org> wrote:
>
> It all depends, sometimes we can handle file moves easily, sometimes we
> can not.
>
> But really, why is a comment typo being needed in stable kernels?

It isn't (well, it is not just a comment since it does end up in the
rendered docs, so it is a bit more "visible" than in a comment, and I
imagine some projects reasonably treat them as a fix, but still, it is
just a typo).

We discussed this years ago when I noticed a typo being picked up by
stable since I wondered why. On my side, I am happy either way -- what
I currently do is explicitly tag the ones that appear in docs. That
way you can decide on your side.

For the others (the ones in comments), I think it is not really worth
it to even figure out a Fixes: tag etc.

Of course, this is for trivial typos -- for something that e.g.
completely changes the requirements of a `# Safety` precondition the
story is different.

Cheers,
Miguel

Re: [PATCH 2/2] rust: pci: fix typo in Bar struct's comment

[Danilo Krummrich]

On Mon Jan 5, 2026 at 11:39 AM CET, Miguel Ojeda wrote:
> On my side, I am happy either way -- what I currently do is explicitly tag the
> ones that appear in docs. That way you can decide on your side.

This is how I handle it as well. For doc-comments I request a Fixes: tag to be
added and leave it to the stable team to decide.

Unless explicitly requested by the stable team (which obviously did not happen
so far) I do not send separate backport patches for typos when the upstream
commit does not apply though.

Re: [PATCH 1/2] rust: io: fix Bar reference in Io struct's comment

[Danilo Krummrich]

On Sat Jan 3, 2026 at 3:31 PM CET, Marko Turk wrote:
> Bar was moved to a separate pci::io module, update the reference to it.
>
> Signed-off-by: Marko Turk <mt@markoturk.info>
> ---
>  rust/kernel/io.rs | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/rust/kernel/io.rs b/rust/kernel/io.rs
> index 98e8b84e68d1..08853f32dae6 100644
> --- a/rust/kernel/io.rs
> +++ b/rust/kernel/io.rs
> @@ -87,7 +87,7 @@ pub fn maxsize(&self) -> usize {
>  /// };
>  /// use core::ops::Deref;
>  ///
> -/// // See also [`pci::Bar`] for a real example.
> +/// // See also [`pci::io::Bar`] for a real example.

pci::io::Bar is re-exported as pci::Bar [1], so this should be fine. Given that
it is not a doc-comment the reference also doesn't do anything. We could remove
the square brackets accordingly.

[1] https://rust.docs.kernel.org/src/kernel/pci.rs.html#43