Re: [PATCH v3 2/3] kgdb: Add NOKPROBE labels on the trap handler functions

[Daniel Thompson]

On Mon, Sep 14, 2020 at 05:14:22PM -0700, Doug Anderson wrote:
> Hi,
> 
> On Mon, Sep 14, 2020 at 6:02 AM Daniel Thompson
> <daniel.thompson@xxxxxxxxxx> wrote:
> >
> > Currently kgdb honours the kprobe blocklist but doesn't place its own
> > trap handling code on the list. Add labels to discourage attempting to
> > use kgdb to debug itself.
> >
> > Not every functions that executes from the trap handler needs to be
> > marked up: relatively early in the trap handler execution (just after
> > we bring the other CPUs to a halt) all breakpoints are replaced with
> > the original opcodes. This patch marks up code in the debug_core that
> > executes between trap entry and the breakpoints being deactivated
> > and, also, code that executes between breakpoint activation and trap
> > exit.
> 
> Other functions that seem to be missing from a quick skim:
> * kgdb_io_ready()
> * kgdb_roundup_cpus()
> * kgdb_call_nmi_hook()

I've grabbed those (and a one or two extras after paying better attention
to the entry logic).

> I'm not confident in my ability to spot every code path, though, so
> I'm not sure at what point we stop looking.  I only spent a few
> minutes and, if important, I could dig more.  Did you have any chance
> to see if there was any way to have a magic linker script just add
> this to everything under "kernel/debug" or something like that where
> we just use a heavier hammer to whack a whole bunch?

I think one could play games with linker sections but it would involve
adding extra infrastructure for the kprobe blocklist. I'm not convinced
that is worth the effort whilst there are acknowledged (and bigger) gaps
elsewhere.

> In general any extra annotation here is better than no annotation, I
> suppose.  ...so if you just want to commit what you have (maybe with
> the above 3 extra functions) then I suppose it'd be fine.

This wasn't quite confident enough for me to convert into an Acked-by:
but I plan to pull v4 into -next very shortly after posting it (since
everything else is agreed).


Daniel.