From: Randy MacLeod <randy.macleod@windriver.com>
To: Harish Sadineni <Harish.Sadineni@windriver.com>,
Richard Purdie <richard.purdie@linuxfoundation.org>,
bruce.ashfield@gmail.com
Cc: Sundeep.Kokkonda@windriver.com, yoann.congal@smile.fr,
elmehdi.younes@smile.fr,
openembedded-core@lists.openembedded.org
Subject: Re: [OE-core] [PATCH v2 03/15] rust: install Rust library sources for 'make rustavailable' support
Date: Wed, 7 Jan 2026 13:21:51 -0500 [thread overview]
Message-ID: <980e869b-15e4-4b6a-a2ac-3e13ecb43b00@windriver.com> (raw)
In-Reply-To: <663fe060-072a-43a4-b0f9-20e172007c47@windriver.com>
[-- Attachment #1: Type: text/plain, Size: 8404 bytes --]
On 2026-01-07 11:34 a.m., Harish Sadineni wrote:
>
> On 1/7/2026 12:29 AM, Randy MacLeod wrote:
>> On 2026-01-05 11:24 a.m., Harish Sadineni wrote:
>>>
>>> On 12/30/2025 9:28 PM, Richard Purdie wrote:
>>>> CAUTION: This email comes from a non Wind River email account!
>>>> Do not click links or open attachments unless you recognize the
>>>> sender and know the content is safe.
>>>>
>>>> On Tue, 2025-12-30 at 06:15 -0800, Sadineni, Harish via
>>>> lists.openembedded.org wrote:
>>>>> From: Harish Sadineni <Harish.Sadineni@windriver.com>
>>>>>
>>>>> The `make rustavailable` process (1) expects the Rust standard
>>>>> library source files (e.g., `lib.rs`)
>>>>> to be present in the `library/` directory under `rustlib/src/rust/`.
>>>>>
>>>>> This patch ensures the required sources are available by:
>>>>> - Copying the `library/` directory from the Rust source tree into
>>>>> `${TMPDIR}/work-shared/rust`
>>>>> during the snapshot setup.
>>>>> - Installing the `library/` directory into
>>>>> `${SDKPATHNATIVE}/usr/lib/rustlib/src/rust` for the
>>>>> `nativesdk` class, making them available in them available in sdk
>>>>>
>>>>> 1) See the kernel tree for Documentation/rust/quick-start.rst in
>>>>> the section: Requirements: Building
>>>>>
>>>>> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/rust/quick-start.rst#n145
>>>>>
>>>>>
>>>>> Signed-off-by: Harish Sadineni <Harish.Sadineni@windriver.com>
>>>>> ---
>>>>> meta/recipes-devtools/rust/rust_1.91.1.bb | 17 +++++++++++++++++
>>>>> 1 file changed, 17 insertions(+)
>>>>>
>>>>> diff --git a/meta/recipes-devtools/rust/rust_1.91.1.bb
>>>>> b/meta/recipes-devtools/rust/rust_1.91.1.bb
>>>>> index a25f65f674..7644ecf2d2 100644
>>>>> --- a/meta/recipes-devtools/rust/rust_1.91.1.bb
>>>>> +++ b/meta/recipes-devtools/rust/rust_1.91.1.bb
>>>>> @@ -63,6 +63,16 @@ do_rust_setup_snapshot () {
>>>>> done
>>>>> fi
>>>>> }
>>>>> +
>>>>> +do_rust_setup_snapshot:append:class-native () {
>>>>> + if ${@bb.utils.contains('DISTRO_FEATURES', 'rust-kernel',
>>>>> 'true', 'false', d)}; then
>>>>> + if [ ! -d "${TMPDIR}/work-shared/rust" ]; then
>>>>> + mkdir -p ${TMPDIR}/work-shared/rust
>>>>> + cp -r ${RUSTSRC}/library
>>>>> ${TMPDIR}/work-shared/rust/.
>>>>> + fi
>>>>> + fi
>>>>> +}
>>>>> +
>>>>> addtask rust_setup_snapshot after do_unpack before do_configure
>>>>> addtask do_test_compile after do_configure do_rust_gen_targets
>>>>> do_rust_setup_snapshot[dirs] += "${WORKDIR}/rust-snapshot"
>>>>> @@ -314,6 +324,13 @@ rust_do_install:class-nativesdk() {
>>>>> export
>>>>> CARGO_TARGET_${RUST_HOST_TRIPLE}_RUNNER="\$OECORE_NATIVE_SYSROOT/lib/${SDKLOADER}"
>>>>> export CC_$RUST_HOST_CC="${CCACHE}${HOST_PREFIX}gcc"
>>>>> EOF
>>>>> +
>>>>> + if ${@bb.utils.contains('DISTRO_FEATURES', 'rust-kernel',
>>>>> 'true', 'false', d)}; then
>>>>> + if [ ! -d
>>>>> ${D}${SDKPATHNATIVE}/usr/lib/rustlib/src/rust ]; then
>>>>> + mkdir -p
>>>>> ${D}${SDKPATHNATIVE}/usr/lib/rustlib/src/rust
>>>>> + cp -r --no-preserve=ownership ${S}/library
>>>>> ${D}${SDKPATHNATIVE}/usr/lib/rustlib/src/rust/
>>>>> + fi
>>>>> + fi
>>>>> }
>>>>>
>>>>> FILES:${PN} += "${base_prefix}/environment-setup.d"
>>>> The commit message should mention the size of these files.
>>> Ok sure, I will add file size in v3.
>>>> Does this make sense as a distro feature or should we just do this
>>>> all the time?
>>> This is suggestion from Bruce that we take it as distro feature.
>>> https://lists.openembedded.org/g/openembedded-core/message/225256
>>
>>
>> Richard mentioned this thread in today's tech call when I asked for
>> commentson the rust-kernel PR.
>>
>> Yes, the high level requirement is to have a DISTRO_FEATURE but
>> common, infrastructure parts
>> such as this code that just copies a hopefully small number of files
>> around,
>> and is part of the rust recipe, could and likely be done regardless
>> of the rust-kernel DISTRO_FEATURE.
>>
>
> Ok sure, We will remove the dependency on DISTRO_FEATURE for copying
> the library directory from rust recipe.
Hold on, I said "small number of files...". See below.
>>
>>
>> We don't want the rust recipe to change based on a kernel config
>> unless we *really* have to
>> since that essentially doubles the testing that should be done or
>> leaves a gap in testing of the
>> rust builds. If you do that for the kernel first, then another recipe
>> later, soon you have a maintenance mess.
>>
>> Also if the kernel needs these files, then it's likely that other
>> software will need it as well.
>> You should analyze why the kernel needs these files and why other
>> recipes do not. Perhaps any
>> kernel-like image will have the same requirement. Is there a
>> baremetal image using rust anywhere
>> that you can use to check on that? I looked but all I found was:
>> https://github.com/ahcbb6/baremetal-helloqemu-rust
>> Anyway, let's focus on the linux kernel's requirements for now.
>>
>>
>> So, how many files are needed and how much FS space do they use?
>>
>
> The file size of the library directory is around 50MB.
>
>
I've been around since the 1990s, 50 MB doesn't seem small to me but
let's see what other people think.
Also, how may files is that?
What's the content? ls -lR if the list isn't too long.
Does the kernel build need each and every file ? How did you check?
Can we automate the generation of the list of required files by scraping
the data from the kernel perhaps?
>>
>>
>> What are other build systems (gentoo for example) doing with their
>> Rust builds to satisfy the kernel's rust requirements?
>>
>
> I will check and update on this.
Thanks.
>>
>>
>>>
>>> In future when rust is default in kernel we can change this, But
>>> till then it is good to have it as a distro feature.
>>>> Do the nativesdk components get packaged separately? If they were, we
>>>> could then make that an SDK feature instead.
>>> No, We are not packaging it separately.
>>
>>
>> The questions seems to be whether we should create a separate
>> packaging rule.
>>
>
> Now by default it is getting packaged with nativesdk-rust, Do we need
> a separate packaging for libraries/files that being installed for rust
> in kernel support?
>>
>>
>>>> What happens for on target kernel module development? Shouldn't
>>>> there be a target package too?
>>> Yes, I have made the necessary changes to include Rust library for
>>> target as well and have tested Rust-based kernel module development
>>> on the target.
>>> I will send updated patches with v3.
>>
>>
>> Before you spend time on polishing v3 please explain what your
>> workflow is, step by step,
>> so we can be sure that things makes sense from a high level.
>>
> We will update the rust recipe to copy the required libs to target
> image and then the below steps to be followed :
This part we're still trying to work out...
> - Build the image by adding the required tools via
> IMAGE_INSTALL:append ( e.g kernel-devsrc, gcc, rust, cargo,
> bindgen-cli etc..).
>
We could create / extend a packagegroup or use:
meta/recipes-extended/images/core-image-kernel-dev.bb
Bruce, what approach do you use / prefer ?
> - In /usr/src/kernel, run "make rustavailable" to verify Rust support
> (This step will check all supporting tools are available for rust
> support in kernel).
> - Run "make menuconfig" and enable "CONFIG_RUST".
> - Run "make scripts" and "make prepare".
> - Write a kernel module in Rust & build the module using "make", which
> generates the "module_name.ko" file.
> - Load the module using "insmod module_name.ko".
The rest seems sensible to a non-kernel guy like me.
Thanks,
../Randy
>
> Thanks,
> Harish
>
>> ../Randy
>>
>>
>>>
>>> Thanks,
>>> Harish
>>>> Cheers,
>>>>
>>>> Richard
>>
>>
>> --
>> # Randy MacLeod
>> # Wind River Linux
--
# Randy MacLeod
# Wind River Linux
[-- Attachment #2: Type: text/html, Size: 14838 bytes --]
next prev parent reply other threads:[~2026-01-07 18:22 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-30 14:15 [PATCH v2 00/15] Enable rust support for linux kernel Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 01/15] bindgen-cli: extend BBCLASSEXTEND to include nativesdk Harish.Sadineni
2026-01-12 0:10 ` [OE-core] " Alistair Francis
2025-12-30 14:15 ` [PATCH v2 02/15] linux-yocto: conditionally add clang/rust/bindgen-cli-native to DEPENDS Harish.Sadineni
2026-01-12 0:12 ` [OE-core] " Alistair Francis
2025-12-30 14:15 ` [PATCH v2 03/15] rust: install Rust library sources for 'make rustavailable' support Harish.Sadineni
2025-12-30 15:58 ` [OE-core] " Richard Purdie
2026-01-05 16:24 ` Harish Sadineni
2026-01-06 18:59 ` Randy MacLeod
2026-01-07 16:34 ` Harish Sadineni
2026-01-07 18:21 ` Randy MacLeod [this message]
2026-01-07 19:03 ` Richard Purdie
2026-01-12 0:42 ` Alistair Francis
2026-01-13 12:14 ` Harish Sadineni
2025-12-30 14:15 ` [PATCH v2 04/15] bitbake.conf: Include "rust-kernel" in native/nativesdk feature filters Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 05/15] kernel-yocto: enable Rust kernel support via rustavailable and staged rustlib sources Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 06/15] linux-yocto: enable Rust support in kernel configuration Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 07/15] kernel-yocto: Fix for buildpaths errors when rust is enabled for kernel Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 08/15] kernel-yocto.bbclass: Disable ccache when rust-kernel is enabled Harish.Sadineni
2026-01-14 15:41 ` alban.moizan
2025-12-30 14:15 ` [PATCH v2 09/15] kernel-devsrc: copying rust-kernel source to $kerneldir/build Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 10/15] selftest/cases/runtime_test: Add test for Linux Rust sample Harish.Sadineni
2026-01-05 9:32 ` [OE-core] " Mathieu Dubois-Briand
2026-01-08 9:39 ` Yoann Congal
2025-12-30 14:15 ` [PATCH v2 11/15] kernel.bbclass: Copy include/config/auto.conf in STAGING_KERNEL_BUILDDIR Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 12/15] kernel.bbclass: Export artifacts needed for out-of-tree Rust compilation Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 13/15] module.bbclass: Prepare out-of-tree rust module compilation Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 14/15] meta-skeleton: Add rust-out-of-tree-module recipe Harish.Sadineni
2025-12-30 14:15 ` [PATCH v2 15/15] runtime_test: Add rust-out-of-tree selftest Harish.Sadineni
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=980e869b-15e4-4b6a-a2ac-3e13ecb43b00@windriver.com \
--to=randy.macleod@windriver.com \
--cc=Harish.Sadineni@windriver.com \
--cc=Sundeep.Kokkonda@windriver.com \
--cc=bruce.ashfield@gmail.com \
--cc=elmehdi.younes@smile.fr \
--cc=openembedded-core@lists.openembedded.org \
--cc=richard.purdie@linuxfoundation.org \
--cc=yoann.congal@smile.fr \
/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