Still RFC. Not all people for each subsystems are not included yet as
this touch quite lot of things. I would like to get feed back is this
resonable seperation and how we will land this. I have tried my best so
that it can be applied in multiple staged if needed. I have not receive
any feedback on this series and that is little bit worrying.

Introduce new ThisModule trait and THIS_MODULE impl.

So currently we have problem that we are not always filling .owner field
example for file_operations. I think we can enable const_refs_to_static
already as that is in 1.78 and is stable in 1.83. So that fits perfecly
for us.  This also seems to be quite request feature but I did not found
that no one has ever suggested that we just enable this.

So basic idea is that we will have ThisModule trait which is used kernel
side. Module side we will always use THIS_MODULE. That is completly
private for modules and kernel crate cannot use it. Currently we have
THIS_MODULE, LocalModule and

  module: &' static ThisModule

on init functions. As we anyway need THIS_MODULE just use that for all
of these things.

Patches 1-2 introduce THIS_MODULE and ThisModule trait.
Patches 3-12 can be applied any order after 1-2.
Patch 13 depends on patches 11-12.
Patches 14-15 are clean up patches and depends 1-13.

    Argillander

To: Miguel Ojeda <ojeda@kernel.org>
To: Boqun Feng <boqun.feng@gmail.com>
To: Gary Guo <gary@garyguo.net>
To: Björn Roy Baron <bjorn3_gh@protonmail.com>
To: Benno Lossin <lossin@kernel.org>
To: Andreas Hindborg <a.hindborg@kernel.org>
To: Alice Ryhl <aliceryhl@google.com>
To: Trevor Gross <tmgross@umich.edu>
To: Danilo Krummrich <dakr@kernel.org>

To: Alexandre Courbot <acourbot@nvidia.com>

Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Petr Pavlu <petr.pavlu@suse.com>
Cc: Daniel Gomez <da.gomez@kernel.org>
Cc: Sami Tolvanen <samitolvanen@google.com>
Cc: Aaron Tomlin <atomlin@atomlin.com>

Signed-off-by: Kari Argillander <kari.argillander@gmail.com>
---
Changes in v3:
- Edit cover.
- Seperate module name changes to seperate patches.
- Rebase top of next 20260109
- Seperate configfs changes to own commit.
- Fix one place still used M over TM for ThisModule.
- Link to v2: https://lore.kernel.org/r/20260106-this_module_fix-v2-0-842ac026f00b@gmail.com

Changes in v2:
- Patches are now sepereted properly.
- Removed debugfs changes as that is not so clear to me.
- Remove module parameter and just used THIS_MODULE everywhere.
- Made macro to make THIS_MODULE.
- Doc tests also have THIS_MODULE.
- Link to v1: https://lore.kernel.org/r/20260101-this_module_fix-v1-0-46ae3e5605a0@gmail.com

---
Kari Argillander (15):
      rust: enable const_refs_to_static feature
      rust: add new ThisModule trait and THIS_MODULE impl
      rust: miscdevice: fix use after free because missing .owner
      rust: block: fix missing owner field in block_device_operations
      rust: drm: fix missing owner in file_operations
      rust: configfs: use new THIS_MODULE
      rust: binder: use new THIS_MODULE
      rust: firmware: use THIS_MODULE over LocalModule for name
      gpu: nova-core: use THIS_MODULE over LocalModule for name
      samples: rust: auxiliary: use THIS_MODULE over LocalModule for name
      rust: driver: make RegistrationOps::register() to use new ThisModule
      rust: phy: make Registration::register() use new ThisModule
      rust: remove module argument from InPlaceModule::init()
      rust: remove kernel::ModuleMetadata
      rust: remove old version of ThisModule

 drivers/android/binder/rust_binder_main.rs |   5 +-
 drivers/block/rnull/configfs.rs            |   2 +-
 drivers/block/rnull/rnull.rs               |   3 +-
 drivers/gpu/drm/nova/driver.rs             |   2 +
 drivers/gpu/drm/tyr/driver.rs              |   2 +
 drivers/gpu/nova-core/nova_core.rs         |   2 +-
 lib/find_bit_benchmark_rust.rs             |   3 +-
 rust/kernel/auxiliary.rs                   |  16 +--
 rust/kernel/block/mq.rs                    |   1 +
 rust/kernel/block/mq/gen_disk.rs           |  30 +-----
 rust/kernel/block/mq/operations.rs         |  30 ++++++
 rust/kernel/configfs.rs                    |  49 ++++-----
 rust/kernel/driver.rs                      |  31 +++---
 rust/kernel/drm/device.rs                  |   2 +-
 rust/kernel/drm/driver.rs                  |   4 +
 rust/kernel/drm/gem/mod.rs                 |   5 +-
 rust/kernel/firmware.rs                    |   4 +-
 rust/kernel/i2c.rs                         |  11 +-
 rust/kernel/lib.rs                         | 161 ++++++++++++++++++++++++-----
 rust/kernel/miscdevice.rs                  |   5 +
 rust/kernel/net/phy.rs                     |  29 ++++--
 rust/kernel/pci.rs                         |  15 +--
 rust/kernel/platform.rs                    |  12 +--
 rust/kernel/prelude.rs                     |   2 +-
 rust/kernel/sync/lock/global.rs            |   4 +-
 rust/kernel/usb.rs                         |  13 +--
 rust/macros/lib.rs                         |   4 +-
 rust/macros/module.rs                      |  24 +----
 samples/rust/rust_configfs.rs              |   2 +-
 samples/rust/rust_debugfs_scoped.rs        |   2 +-
 samples/rust/rust_driver_auxiliary.rs      |   8 +-
 samples/rust/rust_driver_faux.rs           |   2 +-
 samples/rust/rust_minimal.rs               |   2 +-
 samples/rust/rust_misc_device.rs           |   3 +-
 samples/rust/rust_print_main.rs            |   2 +-
 scripts/rustdoc_test_gen.rs                |   2 +
 36 files changed, 298 insertions(+), 196 deletions(-)
---
base-commit: f417b7ffcbef7d76b0d8860518f50dae0e7e5eda
change-id: 20251230-this_module_fix-a390bff24897

Best regards,
-- 
Kari Argillander <kari.argillander@gmail.com>

On Sat, Jan 10, 2026 at 4:09 PM Kari Argillander
<kari.argillander@gmail.com> wrote:
>
> So currently we have problem that we are not always filling .owner field
> example for file_operations. I think we can enable const_refs_to_static
> already as that is in 1.78 and is stable in 1.83. So that fits perfecly
> for us.  This also seems to be quite request feature but I did not found
> that no one has ever suggested that we just enable this.

For context, we have been tracking that feature (as well as
`const_mut_refs`) precisely for `THIS_MODULE`:

    https://github.com/Rust-for-Linux/linux/issues/2

There is also a test that Ralf added upstream back then:

    https://github.com/rust-lang/rust/blob/main/tests/ui/consts/const-ref-to-static-linux-vtable.rs

(Not sure what you mean by "quite request" though)

Cheers,
Miguel

Le 10/01/2026 à 16:07, Kari Argillander a écrit :
> Still RFC. Not all people for each subsystems are not included yet as
> this touch quite lot of things. I would like to get feed back is this
> resonable seperation and how we will land this. I have tried my best so
> that it can be applied in multiple staged if needed. I have not receive
> any feedback on this series and that is little bit worrying.
> 
> Introduce new ThisModule trait and THIS_MODULE impl.

Shouldn't it be called this_module instead of ThisModule ?

I don't know much about rust but this naming looks wrong, it is the kind 
of crap we see in C++.
Linux Kernel codying style says:  ... mixed-case names are frowned upon 
... [Documentation/process/coding-style.rst, §4) Naming]

Is THIS_MODULE a constant or an enum ? 
[Documentation/process/coding-style.rst, §12) Macros, Enums and RTL]

Christophe

On Mon, Jan 12, 2026 at 6:56 PM Christophe Leroy (CS GROUP)
<chleroy@kernel.org> wrote:
>
> Shouldn't it be called this_module instead of ThisModule ?

The common Rust style, which we use in the kernel as well, uses
several kinds of casing to make it easy to distinguish between types
of entities at a glance.

Traits use CamelCase, like structs and enums. So if `ThisModule` is a
trait, that would be its case.

But, for instance, if it were a constant, then it would be `THIS_MODULE`.

Details at https://rust-lang.github.io/api-guidelines/naming.html --
sometimes we may diverge from that, but in general we follow the usual
style.

Cheers,
Miguel

On Sat Jan 10, 2026 at 3:07 PM GMT, Kari Argillander wrote:
> Still RFC. Not all people for each subsystems are not included yet as
> this touch quite lot of things. I would like to get feed back is this
> resonable seperation and how we will land this. I have tried my best so
> that it can be applied in multiple staged if needed. I have not receive
> any feedback on this series and that is little bit worrying.
>
> Introduce new ThisModule trait and THIS_MODULE impl.
>
> So currently we have problem that we are not always filling .owner field
> example for file_operations. I think we can enable const_refs_to_static
> already as that is in 1.78 and is stable in 1.83. So that fits perfecly
> for us.  This also seems to be quite request feature but I did not found
> that no one has ever suggested that we just enable this.
>
> So basic idea is that we will have ThisModule trait which is used kernel
> side. Module side we will always use THIS_MODULE. That is completly
> private for modules and kernel crate cannot use it. Currently we have
> THIS_MODULE, LocalModule and

Why is trait needed? Couldn't this just be a `const THIS_MODULE` required for
vtable?

Best,
Gary

>
>   module: &' static ThisModule
>
> on init functions. As we anyway need THIS_MODULE just use that for all
> of these things.
>
> Patches 1-2 introduce THIS_MODULE and ThisModule trait.
> Patches 3-12 can be applied any order after 1-2.
> Patch 13 depends on patches 11-12.
> Patches 14-15 are clean up patches and depends 1-13.
>
>     Argillander
>
> To: Miguel Ojeda <ojeda@kernel.org>
> To: Boqun Feng <boqun.feng@gmail.com>
> To: Gary Guo <gary@garyguo.net>
> To: Björn Roy Baron <bjorn3_gh@protonmail.com>
> To: Benno Lossin <lossin@kernel.org>
> To: Andreas Hindborg <a.hindborg@kernel.org>
> To: Alice Ryhl <aliceryhl@google.com>
> To: Trevor Gross <tmgross@umich.edu>
> To: Danilo Krummrich <dakr@kernel.org>
>
> To: Alexandre Courbot <acourbot@nvidia.com>
>
> Cc: Luis Chamberlain <mcgrof@kernel.org>
> Cc: Petr Pavlu <petr.pavlu@suse.com>
> Cc: Daniel Gomez <da.gomez@kernel.org>
> Cc: Sami Tolvanen <samitolvanen@google.com>
> Cc: Aaron Tomlin <atomlin@atomlin.com>
>
> Signed-off-by: Kari Argillander <kari.argillander@gmail.com>
> ---
> Changes in v3:
> - Edit cover.
> - Seperate module name changes to seperate patches.
> - Rebase top of next 20260109
> - Seperate configfs changes to own commit.
> - Fix one place still used M over TM for ThisModule.
> - Link to v2: https://lore.kernel.org/r/20260106-this_module_fix-v2-0-842ac026f00b@gmail.com
>
> Changes in v2:
> - Patches are now sepereted properly.
> - Removed debugfs changes as that is not so clear to me.
> - Remove module parameter and just used THIS_MODULE everywhere.
> - Made macro to make THIS_MODULE.
> - Doc tests also have THIS_MODULE.
> - Link to v1: https://lore.kernel.org/r/20260101-this_module_fix-v1-0-46ae3e5605a0@gmail.com
>
> ---
> Kari Argillander (15):
>       rust: enable const_refs_to_static feature
>       rust: add new ThisModule trait and THIS_MODULE impl
>       rust: miscdevice: fix use after free because missing .owner
>       rust: block: fix missing owner field in block_device_operations
>       rust: drm: fix missing owner in file_operations
>       rust: configfs: use new THIS_MODULE
>       rust: binder: use new THIS_MODULE
>       rust: firmware: use THIS_MODULE over LocalModule for name
>       gpu: nova-core: use THIS_MODULE over LocalModule for name
>       samples: rust: auxiliary: use THIS_MODULE over LocalModule for name
>       rust: driver: make RegistrationOps::register() to use new ThisModule
>       rust: phy: make Registration::register() use new ThisModule
>       rust: remove module argument from InPlaceModule::init()
>       rust: remove kernel::ModuleMetadata
>       rust: remove old version of ThisModule
>
>  drivers/android/binder/rust_binder_main.rs |   5 +-
>  drivers/block/rnull/configfs.rs            |   2 +-
>  drivers/block/rnull/rnull.rs               |   3 +-
>  drivers/gpu/drm/nova/driver.rs             |   2 +
>  drivers/gpu/drm/tyr/driver.rs              |   2 +
>  drivers/gpu/nova-core/nova_core.rs         |   2 +-
>  lib/find_bit_benchmark_rust.rs             |   3 +-
>  rust/kernel/auxiliary.rs                   |  16 +--
>  rust/kernel/block/mq.rs                    |   1 +
>  rust/kernel/block/mq/gen_disk.rs           |  30 +-----
>  rust/kernel/block/mq/operations.rs         |  30 ++++++
>  rust/kernel/configfs.rs                    |  49 ++++-----
>  rust/kernel/driver.rs                      |  31 +++---
>  rust/kernel/drm/device.rs                  |   2 +-
>  rust/kernel/drm/driver.rs                  |   4 +
>  rust/kernel/drm/gem/mod.rs                 |   5 +-
>  rust/kernel/firmware.rs                    |   4 +-
>  rust/kernel/i2c.rs                         |  11 +-
>  rust/kernel/lib.rs                         | 161 ++++++++++++++++++++++++-----
>  rust/kernel/miscdevice.rs                  |   5 +
>  rust/kernel/net/phy.rs                     |  29 ++++--
>  rust/kernel/pci.rs                         |  15 +--
>  rust/kernel/platform.rs                    |  12 +--
>  rust/kernel/prelude.rs                     |   2 +-
>  rust/kernel/sync/lock/global.rs            |   4 +-
>  rust/kernel/usb.rs                         |  13 +--
>  rust/macros/lib.rs                         |   4 +-
>  rust/macros/module.rs                      |  24 +----
>  samples/rust/rust_configfs.rs              |   2 +-
>  samples/rust/rust_debugfs_scoped.rs        |   2 +-
>  samples/rust/rust_driver_auxiliary.rs      |   8 +-
>  samples/rust/rust_driver_faux.rs           |   2 +-
>  samples/rust/rust_minimal.rs               |   2 +-
>  samples/rust/rust_misc_device.rs           |   3 +-
>  samples/rust/rust_print_main.rs            |   2 +-
>  scripts/rustdoc_test_gen.rs                |   2 +
>  36 files changed, 298 insertions(+), 196 deletions(-)
> ---
> base-commit: f417b7ffcbef7d76b0d8860518f50dae0e7e5eda
> change-id: 20251230-this_module_fix-a390bff24897
>
> Best regards,