Re: [PATCH] Add drg_why_sleeping.py

[Alexandra Petlanova Hajkova <ahajkova@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1738 bytes --]

On Tue, Sep 1, 2026 at 6:54 PM Tom Tromey <tom@tromey.com> wrote:

> >>>>> "Alexandra" == Alexandra Hájková <ahajkova@redhat.com> writes:
>
> Hi.  Thanks for the patch.
>
> I'm not sure this is really something gdb ought to ship.
>
> Alexandra>  sleep_test.c                                  | 76
> +++++++++++++++++++
>
> This file shouldn't be here.
>
> Alexandra> +           (gdb) source
> ~/binutils-gdb/gdb/python/lib/gdb/command/drgn_why_sleeping.py
>
> Normally commands are auto-installed.
>
Yeah, sorry, I originally didn't specify --data-directory so it ended up not
sourced for me.

>
> Alexandra> +           (gdb) drgn_why_sleeping
>
> This isn't a very gdb-ish name.  At the very least gdb usually uses "-"
> as a separator; I always thought maybe it inherited a Lispy style from
> the RMS days.  But maybe a prefix command would be better anyhow.
>

I'm happy to change the name, it was just the first one that came to mind.

>
>
> Extra parens?
>
> Alexandra> +                    "Can't read /proc/kcore, %s. Try running
> GDB as root. " % str(e)
>
> We can't ever recommend running gdb as root.
>
So,  you recommend just saying we can't read kcore without saying
anything about running GDB as root making things work. Sure, let's
make this bit secret.


> I don't normally like to make drastic pronouncements like that, but gdb
> is nowhere near hardened enough for this kind of thing.  In fact it's
> more accurate to say that gdb doesn't even try, it's ridiculously
> exposed.
>

So,  you recommend just saying we can't read kcore without saying
anything about running GDB as root making things work. Sure, let's
make this bit secret.




>
> Tom
>
>

[-- Attachment #2: Type: text/html, Size: 3048 bytes --]