Re: [PATCH v2 2/2] samples: rust: fwctl: add sample code for fwctl

[John Hubbard <jhubbard@nvidia.com>]

On 1/22/26 1:06 PM, Danilo Krummrich wrote:
> On Thu Jan 22, 2026 at 9:58 PM CET, Jason Gunthorpe wrote:
>> On Thu, Jan 22, 2026 at 10:42:31PM +0200, Zhi Wang wrote:
>>> This patch adds a sample Rust driver demonstrating the usage of the fwctl
>>> Rust abstractions. Add sample code for creating a FwCtl device, getting
>>> device info and issuing an RPC.
>>>
>>> Cc: Danilo Krummrich <dakr@kernel.org>
>>> Cc: Jason Gunthorpe <jgg@nvidia.com>
>>> Signed-off-by: Zhi Wang <zhiw@nvidia.com>
>>> ---
>>>  samples/rust/Kconfig              |  11 +++
>>>  samples/rust/Makefile             |   1 +
>>>  samples/rust/rust_driver_fwctl.rs | 136 ++++++++++++++++++++++++++++++
>>>  3 files changed, 148 insertions(+)
>>>  create mode 100644 samples/rust/rust_driver_fwctl.rs
>>
>> Is this normal for Rust? Can we delete it when a real driver is
>> merged?
>>
>> I'm not so keen on sample drivers, I've spent far too much time lately
>> touching sample drivers that nobody ever uses :(
> 
> I think it is pretty useful for review purposes (at least as long as the first
> real user is not yet ready) to see how it turns out from a user perspective.
> 
> It may also be pretty useful for people starting out with Rust in the kernel, as
> the samples have a very limited scope compared to real drivers.
> 
> But there is no policy or anything, i.e. no need to keep it. There's also no
> need to even land it if you prefer not to do so.

Zhi, the real code that uses this for vGPU is going to be posted soon-ish,
right? So we could use this patch as a review assist, like Danilo mentions,
but not merge it, because the real code will suffice to exercise the new
fwctl functionality.

thanks,
-- 
John Hubbard