Re: [PATCH v5 1/3] Provide in-kernel headers to make extending kernel easier

[Olof Johansson]

On Tue, Apr 16, 2019 at 6:32 AM Karim Yaghmour
<karim.yaghmour@xxxxxxxxxxx> wrote:
>
>
> On 4/16/19 9:04 AM, Joel Fernandes wrote:
> > On Tue, Apr 16, 2019 at 02:49:39PM +0200, Greg Kroah-Hartman wrote:
> >> On Tue, Apr 16, 2019 at 08:33:06AM -0400, Steven Rostedt wrote:
> >>> On Mon, 15 Apr 2019 22:50:10 -0500
> >>> Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> >>>
> >>>> On Mon, Apr 15, 2019 at 9:41 AM Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
> >>>>> I agree with this assessment. We shouldn't use config.gz as precedence
> >>>>> for this solution. config.gz should have been in debugfs to begin with,
> >>>>> but I don't believe debugfs was around when config.gz was introduced.
> >>>>> (Don't have time to look into the history of the two).
> >>>>
> >>>> I don't agree with this: /proc/config.gz is used by a lot of tools
> >>>> that do sanity-check of running systems. This isn't _debugging_...
> >>>> it's verifying correct kernel builds. It's a fancy version of checking
> >>>> /proc/version.
> >>>>
> >>>
> >>> Then we should perhaps make a new file system call tarballs ;-)
> >>>
> >>>   /sys/kernel/tarballs/
> >>>
> >>> and place everything there. That way it removes it from /proc (which is
> >>> the worse place for that) and also makes it something other than debug.
> >>> That's what I did for tracefs.
> >>
> >> As horrible as that suggestion is, it does kind of make sense :)
> >>
> >> We can't put this in debugfs as that's only for debugging and systems
> >> should never have that mounted for normal operations (users want to
> >> build ebpf programs), and /proc really should be for processes but that
> >> horse is long left the barn.
> >>
> >> But, I'm willing to consider putting this either in a system-fs-like
> >> filesystem, or just in sysfs itself, we do have /sys/kernel/ to play
> >> around in if the main objection is that we should not be cluttering up
> >> /proc with stuff like this.
> >>
> >
> > I am ok with the suggestion of /sys/kernel for the archive. That also seems
> > to fit well with the idea that the headers are kernel related and probably
> > belong here more strictly speaking, than /proc.
>
> This makes sense. And if it alleviates concerns regarding extending
> /proc ABIs then might as well switch to this.
>
> Olof, what do you think of this?

In practice we've been more lenient with changes to /sys over time, so
I think this might be a reasonable compromise.

I still think that a filesystem view is the cleanest way to do this,
but I won't push back from this going in. It does solve a real
problem, and if we want a different format later we can revisit it
then.


-Olof