[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Getting the latest and greatest Linux perf features on every Debian kernel



On Mon, Dec 26, 2022 at 11:15 AM Ben Hutchings <ben@decadent.org.uk> wrote:
>
> On Thu, 2022-12-22 at 17:43 -0800, Ian Rogers wrote:
> [...]
> > I had a few things to ask wrt this thread and getting the best
> > possible Debian experience:
> >  - in Linux 6.2 the libtraceevent will need to become an external
> > library dependency
>
> This shouldn't be a problem; it's already packaged separately in Debian
> and we can build-depend on that.
>
> >  - are there issues with debuginfod/zstd enablement? libbfd has
> > licensing issues. libcrypto is only used by the JIT support, but still
> > may be nice to enable.
>
> - debuginfod is licensed under GPLv3+, so has the same issue as libbfd
> - libzstd is licensed under GPLv2 (with most of its code also
> permissively licensed)
> - libcrypto is now licensed under Apache 2.0, which is not compatible
> with GPLv2
>
> So I think only libzstd can be enabled in official packages.
>
> [...]
> >  - are there some residual wrapper artifacts in the package, for example:
> > ```
> > $ dpkg -L linux-perf
> > /.
> > /usr
> > /usr/bin
> > /usr/bin/perf
> > package diverts others to: /usr/bin/perf.wrapper
> > ...
> > package diverts others to: /usr/share/bash-completion/completions/perf.wrapper
> > ...
> > package diverts others to: /usr/share/man/man1/perf.wrapper.1.gz
> > ...
> > ```
> [...]
>
> These diversions are present to avoid conflicts with the wrappers in
> older versions of linux-base.  We can remove them after the next stable
> release.
>
> Ben.

Thanks again Ben! With the releases, the plan is to follow the kernel
releases? We could potentially be more aggressive with the perf tool
given its backward compatibility. That would benefit people using
older kernels who want the latest perf tool bug fixes and features.

Ian

> --
> Ben Hutchings
> Klipstein's 4th Law of Prototyping and Production:
>                                A fail-safe circuit will destroy others.


Reply to: