Re: [PATCH 02/17] raid6: remove __KERNEL__ ifdefs

[Eric Biggers <ebiggers@kernel.org>]

On Wed, Mar 25, 2026 at 09:13:16AM -0700, H. Peter Anvin wrote:
> On 2026-03-25 08:13, H. Peter Anvin wrote:
> > On 2026-03-23 23:40, Christoph Hellwig wrote:
> >> With the test code ported to kernel space, none of this is required.
> > 
> > I really *really* don't like this.
> > 
> > The ability of running in user space is really useful when it comes to
> > developing new code for new platforms, which happens often enough for this code.
> > 
> 
> That being said -- and I do say this as the original author of this code --
> this should be reduced to the maximum extent possible to a (minimal) set of
> #ifndef __KERNEL__, which should be localized as much as possible. The actual
> user space components, even such a thing as a simple #include, should be moved
> to a separate user space header.
> 
> But pretty please do leave the ability to debug the algorithms in user space.
> This is hard code to write and debug; it is not just about regression testing.

While I do like developing code in userspace as well, I've been doing it
less over time as the kernel's tests and benchmarks have been improved.
Running the KUnit tests is pretty straightforward and fast.

The issues with providing userspace build support in the tree are that:

- It has to be maintained.
- It's fundamentally a bit of a hack that is used only by developers
  (who always have the option of doing something locally that is
  tailored to the specific function they're working on)
- It diverts effort from the kernel's actual test and benchmark.

So while the faster iteration speed that userspace development enables
is definitely nice, I do think we should be cautious with committing to
maintain it in the kernel tree.  If it's causing problems for the
ongoing refactoring, dropping it for now seems reasonable to me.

I would suggest adding a benchmark to the KUnit test similar to the
crypto and CRC ones, though.

- Eric