Re: [PATCH] MAINTAINERS: Associate GDB scripts with their subsystems

[Florian Fainelli]

On 6/25/25 13:22, Andrew Morton wrote:
> On Wed, 25 Jun 2025 21:38:20 +0200 Jan Kiszka <jan.kiszka@xxxxxxxxxxx> wrote:
>
> > On 25.06.25 19:52, Florian Fainelli wrote:
> > > The GDB scripts under scripts/gdb/linux are very useful for inspecting
> > > kernel data structures however they depend upon the internal APIs and
> > > data structures which are updated without much consideration for those
> > > scripts. This results in a near constant catching up with fixing the
> > > scripts so they continue to work.
> > >
> > > Associate the GDB scripts with their subsystems in the hope that they
> > > get more love and attention.
> >
> > ...
> >
> > I will surely support any proposal that helps connecting the scripts
> > with subsystems they address. However, you should likely break up this
> > one here into per-subsystem patches and address each affected
> > maintainer. They should have a chance to accept or reject this potential
> > extension of their responsibility.
>
> I agree - this proposal doesn't seem very practical, really.

Yes, that's a good point it should be on a per-subsystem basis to decide 
whether they are willing to take on the maintenance, if nothing else the 
reporting.

> It might actually be harmful - if someone has an issue with a gdb script
> they'll report that to the subsystem maintainer rather than to the GDB
> script maintainers who are better equipped to address the issue.

If they run scripts/get_maintainer.pl they will get both subsystems to 
be listed as recipients to reach out to, so that should help cross 
pollinate and ease the pain of fixing.

> And I'm not sure there's really a problem to fix here.  I'm seeing 13
> commits to scripts/gdb this year and afaict only one (e0349c46cb4f
> ("scripts/gdb/linux/symbols.py: address changes to module_sect_attrs"))
> looks like it is fixing up such a problem.

There are a few more that result from breakage that could have been avoided:

https://lore.kernel.org/all/20250601055027.3661480-1-tony.ambardar@xxxxxxxxx/

https://lore.kernel.org/all/20250619225105.320729-1-florian.fainelli@xxxxxxxxxxxx/

and the recent ones for interrupts.py.
--
Florian