[PATCH] arm64: perf: fix syscalltbl path base

[Leo Yan]

On Mon, Dec 22, 2025 at 11:42:49AM +0100, Arnd Bergmann wrote:
> On Mon, Dec 22, 2025, at 11:05, James Clark wrote:
> > On 21/12/2025 9:48 pm, Joel May wrote:
> >> On Mon, Dec 15, 2025, at 03:28, Leo Yan wrote:
> >> 
> >> This is my first time contributing to the Linux kernel, so please bare with
> >> my ignorance.  I don't know whether it's acceptable to consider patches
> >> that conceptually improve something without an actual problem they're
> >> solving.
> >> 
> >
> > No worries, thanks for sending the fix. Yes I think it could be ok to 
> > apply your patch anyway. I just wanted to make sure we were all on the 
> > same page with both fixes fixing the same thing.
> 
> I still don't think that Leo's series is correct.

As discussed, I will respin the series to use asm-generic/unistd.h for
arm64, this is consistent with other arches.

> > You might want to drop the fixes tag though to avoid confusion if Leo's 
> > one gets accepted, because then technically it doesn't fix anything any 
> > more. Keeping Leo's fixes: tag is a bit better because it fixes 
> > additional things and targets an older commit.
> 
> Before that gets applied, someone should explain why arm64 needs
> any special treatment at all. A lot of work has gone into making
> the syscall table handling consistent across architectures, so
> whether we ship a copy of the generated header for perf, we
> generate a new copy or we rely on 'make headers_install', should
> should really be done the same way across every architecture.

I agree arm64 is not a special case.

For the long term, I think "make headers_install" is worth trying.  The
main concern is that it introduces a dependency between tool builds
and the kernel.  However, if tools/selftests need the latest kernel
features, it's natural to depend on the latest headers.

Thanks,
Leo