Re: [RFC] btf: split core BTF parsing out of BPF subsystem into kernel/btf/

[Alexei Starovoitov]

On Wed, Apr 22, 2026 at 3:44 AM Alan Maguire <alan.maguire@xxxxxxxxxx> wrote:
>
> On 25/03/2026 01:18, Sasha Levin wrote:
> > Move BTF type format parsing and inspection code out of the BPF
> > subsystem into its own kernel/bf/ directory, separating core BTF
> > functionality from BPF-specific extensions.
> >
> > CONFIG_DEBUG_INFO_BTF currently depends on CONFIG_BPF_SYSCALL, which
> > prevents embedded, automotive, and safety-critical environments from
> > using BTF. These platforms often disable BPF for security and size
> > reasons but would benefit from BTF type information for crash
> > diagnostics and debugging.
> >
>
> I had a go at a refactor in this area too, and it's pretty tricky. How about
> we start with a smaller goal; making /sys/kernel/btf available to userspace
> on systems without CONFIG_BPF_SYSCALL? This would just involve a small refactor
> of the BTF module-related stuff in kernel/bpf/btf.c, moving it to btf_module.c
> or similar say. No need to split out BTF parsing APIs from those actively used in
> validating BPF etc, since a non-BPF_SYSCALL system would not need to parse BTF
> in the kernel (just make it available in sysfs.

Small refactor is ok, but in general I'm not interested
in complicating code for systems without CONFIG_BPF_SYSCALL.