public inbox for linux-pci@vger.kernel.org
 help / color / mirror / Atom feed
From: Philipp Stanner <pstanner@redhat.com>
To: Danilo Krummrich <dakr@kernel.org>,
	gregkh@linuxfoundation.org,  rafael@kernel.org,
	bhelgaas@google.com, ojeda@kernel.org, alex.gaynor@gmail.com,
	 boqun.feng@gmail.com, gary@garyguo.net,
	bjorn3_gh@protonmail.com,  benno.lossin@proton.me,
	tmgross@umich.edu, a.hindborg@samsung.com,  aliceryhl@google.com,
	airlied@gmail.com, fujita.tomonori@gmail.com,
	 lina@asahilina.net, ajanulgu@redhat.com, lyude@redhat.com,
	robh@kernel.org,  daniel.almeida@collabora.com,
	saravanak@google.com, dirk.behme@de.bosch.com,  j@jannau.net,
	fabien.parent@linaro.org, chrisi.schrefl@gmail.com
Cc: rust-for-linux@vger.kernel.org, linux-kernel@vger.kernel.org,
	 linux-pci@vger.kernel.org, devicetree@vger.kernel.org
Subject: Re: [PATCH v4 09/13] rust: pci: implement I/O mappable `pci::Bar`
Date: Fri, 06 Dec 2024 11:44:48 +0100	[thread overview]
Message-ID: <900dcafa7ee5113145f1ee7dda02de32432388e7.camel@redhat.com> (raw)
In-Reply-To: <20241205141533.111830-10-dakr@kernel.org>

btw the PCI subsystem consistently always writes the acronym all-
uppercase (PCI) in commit messages and comments.

If you go for a v5, might make sense to adjust that.

Same applies for patch 8.


P.


On Thu, 2024-12-05 at 15:14 +0100, Danilo Krummrich wrote:
> Implement `pci::Bar`, `pci::Device::iomap_region` and
> `pci::Device::iomap_region_sized` to allow for I/O mappings of PCI
> BARs.
> 
> To ensure that a `pci::Bar`, and hence the I/O memory mapping, can't
> out-live the PCI device, the `pci::Bar` type is always embedded into
> a
> `Devres` container, such that the `pci::Bar` is revoked once the
> device
> is unbound and hence the I/O mapped memory is unmapped.
> 
> A `pci::Bar` can be requested with
> (`pci::Device::iomap_region_sized`) or
> without (`pci::Device::iomap_region`) a const generic representing
> the
> minimal requested size of the I/O mapped memory region. In case of
> the
> latter only runtime checked I/O reads / writes are possible.
> 
> Co-developed-by: Philipp Stanner <pstanner@redhat.com>
> Signed-off-by: Philipp Stanner <pstanner@redhat.com>
> Signed-off-by: Danilo Krummrich <dakr@kernel.org>
> ---
>  rust/kernel/pci.rs | 145
> +++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 145 insertions(+)
> 
> diff --git a/rust/kernel/pci.rs b/rust/kernel/pci.rs
> index 748912c1c994..0a32a5935c9c 100644
> --- a/rust/kernel/pci.rs
> +++ b/rust/kernel/pci.rs
> @@ -5,10 +5,14 @@
>  //! C header: [`include/linux/pci.h`](srctree/include/linux/pci.h)
>  
>  use crate::{
> +    alloc::flags::*,
>      bindings, container_of, device,
>      device_id::RawDeviceId,
> +    devres::Devres,
>      driver,
>      error::{to_result, Result},
> +    io::Io,
> +    io::IoRaw,
>      str::CStr,
>      types::{ARef, ForeignOwnable},
>      ThisModule,
> @@ -239,9 +243,115 @@ pub trait Driver {
>  ///
>  /// A PCI device is based on an always reference counted
> `device:Device` instance. Cloning a PCI
>  /// device, hence, also increments the base device' reference count.
> +///
> +/// # Invariants
> +///
> +/// `Device` hold a valid reference of `ARef<device::Device>` whose
> underlying `struct device` is a
> +/// member of a `struct pci_dev`.
>  #[derive(Clone)]
>  pub struct Device(ARef<device::Device>);
>  
> +/// A PCI BAR to perform I/O-Operations on.
> +///
> +/// # Invariants
> +///
> +/// `Bar` always holds an `IoRaw` inststance 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: Device,
> +    io: IoRaw<SIZE>,
> +    num: i32,
> +}
> +
> +impl<const SIZE: usize> Bar<SIZE> {
> +    fn new(pdev: Device, num: u32, name: &CStr) -> Result<Self> {
> +        let len = pdev.resource_len(num)?;
> +        if len == 0 {
> +            return Err(ENOMEM);
> +        }
> +
> +        // Convert to `i32`, since that's what all the C bindings
> use.
> +        let num = i32::try_from(num)?;
> +
> +        // SAFETY:
> +        // `pdev` is valid by the invariants of `Device`.
> +        // `num` is checked for validity by a previous call to
> `Device::resource_len`.
> +        // `name` is always valid.
> +        let ret = unsafe {
> bindings::pci_request_region(pdev.as_raw(), num, name.as_char_ptr())
> };
> +        if ret != 0 {
> +            return Err(EBUSY);
> +        }
> +
> +        // SAFETY:
> +        // `pdev` is valid by the invariants of `Device`.
> +        // `num` is checked for validity by a previous call to
> `Device::resource_len`.
> +        // `name` is always valid.
> +        let ioptr: usize = unsafe {
> bindings::pci_iomap(pdev.as_raw(), num, 0) } as usize;
> +        if ioptr == 0 {
> +            // SAFETY:
> +            // `pdev` valid by the invariants of `Device`.
> +            // `num` is checked for validity by a previous call to
> `Device::resource_len`.
> +            unsafe { bindings::pci_release_region(pdev.as_raw(),
> num) };
> +            return Err(ENOMEM);
> +        }
> +
> +        let io = match IoRaw::new(ioptr, len as usize) {
> +            Ok(io) => io,
> +            Err(err) => {
> +                // SAFETY:
> +                // `pdev` is valid by the invariants of `Device`.
> +                // `ioptr` is guaranteed to be the start of a valid
> I/O mapped memory region.
> +                // `num` is checked for validity by a previous call
> to `Device::resource_len`.
> +                unsafe { Self::do_release(&pdev, ioptr, num) };
> +                return Err(err);
> +            }
> +        };
> +
> +        Ok(Bar { pdev, io, num })
> +    }
> +
> +    /// # Safety
> +    ///
> +    /// `ioptr` must be a valid pointer to the memory mapped PCI bar
> number `num`.
> +    unsafe fn do_release(pdev: &Device, ioptr: usize, num: i32) {
> +        // SAFETY:
> +        // `pdev` is valid by the invariants of `Device`.
> +        // `ioptr` is valid by the safety requirements.
> +        // `num` is valid by the safety requirements.
> +        unsafe {
> +            bindings::pci_iounmap(pdev.as_raw(), ioptr as _);
> +            bindings::pci_release_region(pdev.as_raw(), num);
> +        }
> +    }
> +
> +    fn release(&self) {
> +        // SAFETY: The safety requirements are guaranteed by the
> type invariant of `self.pdev`.
> +        unsafe { Self::do_release(&self.pdev, self.io.addr(),
> self.num) };
> +    }
> +}
> +
> +impl Bar {
> +    fn index_is_valid(index: u32) -> bool {
> +        // A `struct pci_dev` owns an array of resources with at
> most `PCI_NUM_RESOURCES` entries.
> +        index < bindings::PCI_NUM_RESOURCES
> +    }
> +}
> +
> +impl<const SIZE: usize> Drop for Bar<SIZE> {
> +    fn drop(&mut self) {
> +        self.release();
> +    }
> +}
> +
> +impl<const SIZE: usize> Deref for Bar<SIZE> {
> +    type Target = Io<SIZE>;
> +
> +    fn deref(&self) -> &Self::Target {
> +        // SAFETY: By the type invariant of `Self`, the MMIO range
> in `self.io` is properly mapped.
> +        unsafe { Io::from_raw(&self.io) }
> +    }
> +}
> +
>  impl Device {
>      /// Create a PCI Device instance from an existing
> `device::Device`.
>      ///
> @@ -275,6 +385,41 @@ pub fn set_master(&self) {
>          // SAFETY: `self.as_raw` is guaranteed to be a pointer to a
> valid `struct pci_dev`.
>          unsafe { bindings::pci_set_master(self.as_raw()) };
>      }
> +
> +    /// Returns the size of the given PCI bar resource.
> +    pub fn resource_len(&self, bar: u32) ->
> Result<bindings::resource_size_t> {
> +        if !Bar::index_is_valid(bar) {
> +            return Err(EINVAL);
> +        }
> +
> +        // SAFETY:
> +        // - `bar` is a valid bar number, as guaranteed by the above
> call to `Bar::index_is_valid`,
> +        // - by its type invariant `self.as_raw` is always a valid
> pointer to a `struct pci_dev`.
> +        Ok(unsafe { bindings::pci_resource_len(self.as_raw(),
> bar.try_into()?) })
> +    }
> +
> +    /// Mapps an entire PCI-BAR after performing a region-request on
> it. I/O operation bound checks
> +    /// can be performed on compile time for offsets (plus the
> requested type size) < SIZE.
> +    pub fn iomap_region_sized<const SIZE: usize>(
> +        &self,
> +        bar: u32,
> +        name: &CStr,
> +    ) -> Result<Devres<Bar<SIZE>>> {
> +        let bar = Bar::<SIZE>::new(self.clone(), bar, name)?;
> +        let devres = Devres::new(self.as_ref(), bar, GFP_KERNEL)?;
> +
> +        Ok(devres)
> +    }
> +
> +    /// Mapps an entire PCI-BAR after performing a region-request on
> it.
> +    pub fn iomap_region(&self, bar: u32, name: &CStr) ->
> Result<Devres<Bar>> {
> +        self.iomap_region_sized::<0>(bar, name)
> +    }
> +
> +    /// Returns a new `ARef` of the base `device::Device`.
> +    pub fn as_dev(&self) -> ARef<device::Device> {
> +        self.0.clone()
> +    }
>  }
>  
>  impl AsRef<device::Device> for Device {


  reply	other threads:[~2024-12-06 10:44 UTC|newest]

Thread overview: 45+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-12-05 14:14 [PATCH v4 00/13] Device / Driver PCI / Platform Rust abstractions Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 01/13] rust: pass module name to `Module::init` Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 02/13] rust: implement generic driver registration Danilo Krummrich
2024-12-06 13:57   ` Alice Ryhl
2024-12-06 18:13     ` Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 03/13] rust: implement `IdArray`, `IdTable` and `RawDeviceId` Danilo Krummrich
2024-12-07  1:14   ` Fabien Parent
2024-12-09 10:45     ` Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 04/13] rust: add rcu abstraction Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 05/13] rust: add `Revocable` type Danilo Krummrich
2024-12-06 15:11   ` Alice Ryhl
2024-12-09 10:40     ` Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 06/13] rust: add `io::{Io, IoRaw}` base types Danilo Krummrich
2024-12-06 14:13   ` Alice Ryhl
2024-12-11 14:52   ` Daniel Almeida
2024-12-05 14:14 ` [PATCH v4 07/13] rust: add devres abstraction Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 08/13] rust: pci: add basic PCI device / driver abstractions Danilo Krummrich
2024-12-06 14:01   ` Alice Ryhl
2024-12-09 10:44     ` Danilo Krummrich
2024-12-10 10:55       ` Alice Ryhl
2024-12-10 22:38         ` Danilo Krummrich
2024-12-11 13:06           ` Alice Ryhl
2024-12-11 14:32             ` Danilo Krummrich
2024-12-11 14:41               ` Greg KH
2024-12-11 14:42                 ` Greg KH
2024-12-11 14:44               ` Alice Ryhl
2024-12-06 15:25   ` Alice Ryhl
2024-12-05 14:14 ` [PATCH v4 09/13] rust: pci: implement I/O mappable `pci::Bar` Danilo Krummrich
2024-12-06 10:44   ` Philipp Stanner [this message]
2024-12-05 14:14 ` [PATCH v4 10/13] samples: rust: add Rust PCI sample driver Danilo Krummrich
2024-12-05 14:14 ` [PATCH v4 11/13] rust: of: add `of::DeviceId` abstraction Danilo Krummrich
2024-12-09 21:22   ` Rob Herring
2024-12-05 14:14 ` [PATCH v4 12/13] rust: platform: add basic platform device / driver abstractions Danilo Krummrich
2024-12-09 22:37   ` Rob Herring
2024-12-09 23:13     ` Danilo Krummrich
2024-12-10  7:46     ` Greg KH
2024-12-10  9:34       ` Danilo Krummrich
2024-12-10  9:40         ` Greg KH
2024-12-05 14:14 ` [PATCH v4 13/13] samples: rust: add Rust platform sample driver Danilo Krummrich
2024-12-05 17:09   ` Dirk Behme
2024-12-05 18:03     ` Rob Herring
2024-12-06  6:39       ` Dirk Behme
2024-12-06  8:33     ` Danilo Krummrich
2024-12-06  9:29       ` Dirk Behme
2024-12-10 22:59   ` Rob Herring (Arm)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=900dcafa7ee5113145f1ee7dda02de32432388e7.camel@redhat.com \
    --to=pstanner@redhat.com \
    --cc=a.hindborg@samsung.com \
    --cc=airlied@gmail.com \
    --cc=ajanulgu@redhat.com \
    --cc=alex.gaynor@gmail.com \
    --cc=aliceryhl@google.com \
    --cc=benno.lossin@proton.me \
    --cc=bhelgaas@google.com \
    --cc=bjorn3_gh@protonmail.com \
    --cc=boqun.feng@gmail.com \
    --cc=chrisi.schrefl@gmail.com \
    --cc=dakr@kernel.org \
    --cc=daniel.almeida@collabora.com \
    --cc=devicetree@vger.kernel.org \
    --cc=dirk.behme@de.bosch.com \
    --cc=fabien.parent@linaro.org \
    --cc=fujita.tomonori@gmail.com \
    --cc=gary@garyguo.net \
    --cc=gregkh@linuxfoundation.org \
    --cc=j@jannau.net \
    --cc=lina@asahilina.net \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-pci@vger.kernel.org \
    --cc=lyude@redhat.com \
    --cc=ojeda@kernel.org \
    --cc=rafael@kernel.org \
    --cc=robh@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --cc=saravanak@google.com \
    --cc=tmgross@umich.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox