mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Daniel Sedlak <daniel@sedlak.dev>
To: Daniel Almeida <daniel.almeida@collabora.com>,
	ojeda@kernel.org, alex.gaynor@gmail.com, boqun.feng@gmail.com,
	gary@garyguo.net, bjorn3_gh@protonmail.mco,
	benno.lossin@proton.me, a.hindborg@kernel.org,
	aliceryhl@google.com, tmgross@umich.edu,
	gregkh@linuxfoundation.org, rafael@kernel.org, dakr@kernel.org,
	boris.brezillon@collabora.com, robh@kernel.org
Cc: rust-for-linux@vger.kernel.org, linux-kernel@vger.kernel.org,
	Fiona Behrens <me@kloenk.dev>
Subject: Re: [PATCH v6 1/3] rust: io: add resource abstraction
Date: Fri, 31 Jan 2025 11:02:03 +0100	[thread overview]
Message-ID: <0623e789-ff5d-4d6f-a73a-7d514e0dc6d7@sedlak.dev> (raw)
In-Reply-To: <20250130220529.665896-2-daniel.almeida@collabora.com>

Hi,

On 1/30/25 11:05 PM, Daniel Almeida wrote:

> +/// Returns a reference to the global `iomem_resource` variable.
> +pub fn iomem_resource() -> &'static Resource {
> +    // SAFETY: `bindings::iomem_resoure` has global lifetime and is of type Resource.
> +    unsafe { Resource::from_ptr(core::ptr::addr_of_mut!(bindings::iomem_resource)) }
> +}
> +
> +/// Resource Size type.
> +/// This is a type alias to `u64`
> +/// depending on the config option `CONFIG_PHYS_ADDR_T_64BIT`.

The comment seems weirdly formatted, shouldn't it be rather:

/// Resource size type. This is a type alias to `u64`
/// depending on the config option `CONFIG_PHYS_ADDR_T_64BIT`.

or

/// Resource size type.
///
/// This is a type alias to `u64` depending on the config
/// option `CONFIG_PHYS_ADDR_T_64BIT`.

> +#[cfg(CONFIG_PHYS_ADDR_T_64BIT)]
> +pub type ResourceSize = u64;
> +
> +/// Resource Size type.
> +/// This is a type alias to `u32`
> +/// depending on the config option `CONFIG_PHYS_ADDR_T_64BIT`.

Similar to the previous one.

> +#[cfg(not(CONFIG_PHYS_ADDR_T_64BIT))]
> +pub type ResourceSize = u32;
> +
> +/// A region allocated from a parent resource.
> +///
> +/// # Invariants
> +/// - `self.0` points to a valid `bindings::resource` that was obtained through
> +/// `__request_region`.

Shouldn't be there an extra newline after # Invariants, to be consistent 
with others in the patch?

> +pub struct Region(NonNull<bindings::resource>);
> +
> +impl Deref for Region {
> +    type Target = Resource;
> +
> +    fn deref(&self) -> &Self::Target {
> +        // SAFETY: Safe as per the invariant of `Region`
> +        unsafe { Resource::from_ptr(self.0.as_ptr()) }
> +    }
> +}
> +
> +impl Drop for Region {
> +    fn drop(&mut self) {
> +        // SAFETY: Safe as per the invariant of `Region`
> +        let res = unsafe { Resource::from_ptr(self.0.as_ptr()) };
> +        let flags = res.flags();
> +
> +        let release_fn = if flags.contains(flags::IORESOURCE_MEM) {
> +            bindings::release_mem_region
> +        } else {
> +            bindings::release_region
> +        };
> +
> +        // SAFETY: Safe as per the invariant of `Region`
> +        unsafe { release_fn(res.start(), res.size()) };
> +    }
> +}
> +
> +// SAFETY: `Region` only holds a pointer to a C `struct resource`, which is safe to be used from
> +// any thead.

typo thead -> thread
> +unsafe impl Send for Region {}
> +
> +// SAFETY: `Region` only holds a pointer to a C `struct resource`, references to which are
> +// safe to be used from any thead.

typo thead -> thread

> +unsafe impl Sync for Region {}
> +
> +/// A resource abstraction.
> +///
> +/// # Invariants
> +///
> +/// `Resource` is a transparent wrapper around a valid `bindings::resource`.
> +#[repr(transparent)]
> +pub struct Resource(Opaque<bindings::resource>);
> +
> +impl Resource {
> +    /// Creates a reference to a [`Resource`] from a valid pointer.
> +    ///
> +    /// # Safety
> +    ///
> +    /// The caller must ensure that for the duration of 'a, the pointer will
> +    /// point at a valid `bindings::resource`
> +    ///
> +    /// The caller must also ensure that the `Resource` is only accessed via the
> +    /// returned reference for the duration of 'a.
> +    pub(crate) const unsafe fn from_ptr<'a>(ptr: *mut bindings::resource) -> &'a Self {
> +        // SAFETY: Self is a transparent wrapper around `Opaque<bindings::resource>`.
> +        unsafe { &*ptr.cast() }
> +    }
> +
> +    /// A helper to abstract the common pattern of requesting a region.
> +    fn request_region_checked(
> +        &self,
> +        start: ResourceSize,
> +        size: ResourceSize,
> +        name: &CStr,
> +        request_fn: RequestFn,
> +    ) -> Option<Region> {
> +        // SAFETY: Safe as per the invariant of `Resource`
> +        let region = unsafe { request_fn(start, size, name.as_char_ptr()) };
> +
> +        Some(Region(NonNull::new(region)?))
> +    }
> +
> +    /// Requests a resource region.
> +    ///
> +    /// Exclusive access will be given and the region will be marked as busy.
> +    /// Further calls to `request_region` will return `None` if the region, or a
> +    /// part of it, is already in use.
> +    pub fn request_region(
> +        &self,
> +        start: ResourceSize,
> +        size: ResourceSize,
> +        name: &CStr,
> +    ) -> Option<Region> {
> +        self.request_region_checked(start, size, name, bindings::request_region)
> +    }
> +
> +    /// Requests a resource region with the IORESOURCE_MUXED flag.

formatting: IORESOURCE_MUXED -> `IORESOURCE_MUXED`

> +    ///
> +    /// Exclusive access will be given and the region will be marked as busy.
> +    /// Further calls to `request_region` will return `None` if the region, or a
> +    /// part of it, is already in use.
> +    pub fn request_muxed_region(
> +        &self,
> +        start: ResourceSize,
> +        size: ResourceSize,
> +        name: &CStr,
> +    ) -> Option<Region> {
> +        self.request_region_checked(start, size, name, bindings::request_muxed_region)
> +    }
> +
> +    /// Requests a memory resource region, i.e.: a resource of type
> +    /// IORESOURCE_MEM.

formatting: IORESOURCE_MEM -> `IORESOURCE_MEM`

> +    ///
> +    /// Exclusive access will be given and the region will be marked as busy.
> +    /// Further calls to `request_region` will return `None` if the region, or a
> +    /// part of it, is already in use.
> +    pub fn request_mem_region(
> +        &self,
> +        start: ResourceSize,
> +        size: ResourceSize,
> +        name: &CStr,
> +    ) -> Option<Region> {
> +        self.request_region_checked(start, size, name, bindings::request_mem_region)
> +    }
> +
> +    /// Returns the size of the resource.
> +    pub fn size(&self) -> ResourceSize {
> +        let inner = self.0.get();
> +        // SAFETY: safe as per the invariants of `Resource`
> +        unsafe { bindings::resource_size(inner) }
> +    }
> +
> +    /// Returns the start address of the resource.
> +    pub fn start(&self) -> u64 {

Should the address be of type `usize`?

> +        let inner = self.0.get();
> +        // SAFETY: safe as per the invariants of `Resource`
> +        unsafe { *inner }.start
> +    }
> +
> +    /// Returns the name of the resource.
> +    pub fn name(&self) -> &CStr {
> +        let inner = self.0.get();
> +        // SAFETY: safe as per the invariants of `Resource`
> +        unsafe { CStr::from_char_ptr((*inner).name) }
> +    }
> +
> +    /// Returns the flags associated with the resource.
> +    pub fn flags(&self) -> Flags {
> +        let inner = self.0.get();
> +        // SAFETY: safe as per the invariants of `Resource`
> +        let flags = unsafe { *inner }.flags;
> +
> +        Flags(flags)
> +    }
> +}
> +
> +// SAFETY: `Resource` only holds a pointer to a C `struct resource`, which is safe to be used from
> +// any thead.

typo: thead -> thread

> +unsafe impl Send for Resource {}
> +
> +// SAFETY: `Resource` only holds a pointer to a C `struct resource`, references to which are
> +// safe to be used from any thead.

typo: thead -> thread

> +unsafe impl Sync for Resource {}
> +
> +/// Resource flags as stored in the C `struct resource::flags` field.
> +///
> +/// They can be combined with the operators `|`, `&`, and `!`.
> +///
> +/// Values can be used from the [`flags`] module.
> +#[derive(Clone, Copy, PartialEq)]
> +pub struct Flags(u64);
> +
> +impl Flags {
> +    /// Check whether `flags` is contained in `self`.
> +    pub fn contains(self, flags: Flags) -> bool {
> +        (self & flags) == flags
> +    }
> +}
> +
> +impl core::ops::BitOr for Flags {
> +    type Output = Self;
> +    fn bitor(self, rhs: Self) -> Self::Output {
> +        Self(self.0 | rhs.0)
> +    }
> +}
> +
> +impl core::ops::BitAnd for Flags {
> +    type Output = Self;
> +    fn bitand(self, rhs: Self) -> Self::Output {
> +        Self(self.0 & rhs.0)
> +    }
> +}
> +
> +impl core::ops::Not for Flags {
> +    type Output = Self;
> +    fn not(self) -> Self::Output {
> +        Self(!self.0)
> +    }
> +}
> +
> +/// Resource flags as stored in the `struct resource::flags` field.
> +pub mod flags {
> +    use super::Flags;
> +
> +    /// PCI/ISA I/O ports

formatting: period at the end

> +    pub const IORESOURCE_IO: Flags = Flags(bindings::IORESOURCE_IO as u64);
> +
> +    /// Resource is software muxed.
> +    pub const IORESOURCE_MUXED: Flags = Flags(bindings::IORESOURCE_MUXED as u64);
> +
> +    /// Resource represents a memory region.
> +    pub const IORESOURCE_MEM: Flags = Flags(bindings::IORESOURCE_MEM as u64);
> +}

	Daniel

  reply	other threads:[~2025-01-31 10:02 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-01-30 22:05 [PATCH v6 0/3] rust: platform: add Io support Daniel Almeida
2025-01-30 22:05 ` [PATCH v6 1/3] rust: io: add resource abstraction Daniel Almeida
2025-01-31 10:02   ` Daniel Sedlak [this message]
2025-02-09 11:45   ` Guangbo Cui
2025-01-30 22:05 ` [PATCH v6 2/3] rust: io: mem: add a generic iomem abstraction Daniel Almeida
2025-01-31 10:09   ` Daniel Sedlak
2025-02-02 22:45   ` Asahi Lina
2025-02-03  9:26     ` Alice Ryhl
2025-02-03 14:14       ` Asahi Lina
2025-02-03  9:32   ` Alice Ryhl
2025-02-05 14:56   ` Guangbo Cui
2025-02-06 15:43     ` Alice Ryhl
2025-02-06 15:58       ` Miguel Ojeda
2025-02-06 15:58       ` Guangbo Cui
2025-02-06 16:11         ` Miguel Ojeda
     [not found]           ` <tencent_E1DC219DB45DC03A8454E2124D238DCEC705@qq.com>
2025-02-06 17:13             ` Danilo Krummrich
2025-02-07 13:25               ` Daniel Almeida
2025-02-06 15:57     ` Daniel Almeida
2025-02-06 16:05       ` Miguel Ojeda
2025-04-01 15:57       ` Joel Fernandes
2025-04-01 16:44         ` Danilo Krummrich
2025-04-01 17:07           ` Joel Fernandes
2025-01-30 22:05 ` [PATCH v6 3/3] rust: platform: allow ioremap of platform resources Daniel Almeida
2025-01-31 10:19   ` Daniel Sedlak
2025-01-31 11:36     ` Alice Ryhl

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=0623e789-ff5d-4d6f-a73a-7d514e0dc6d7@sedlak.dev \
    --to=daniel@sedlak.dev \
    --cc=a.hindborg@kernel.org \
    --cc=alex.gaynor@gmail.com \
    --cc=aliceryhl@google.com \
    --cc=benno.lossin@proton.me \
    --cc=bjorn3_gh@protonmail.mco \
    --cc=boqun.feng@gmail.com \
    --cc=boris.brezillon@collabora.com \
    --cc=dakr@kernel.org \
    --cc=daniel.almeida@collabora.com \
    --cc=gary@garyguo.net \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=me@kloenk.dev \
    --cc=ojeda@kernel.org \
    --cc=rafael@kernel.org \
    --cc=robh@kernel.org \
    --cc=rust-for-linux@vger.kernel.org \
    --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

all inboxes | Powered by JetHome®