linux-kernel - Re: [PATCH 1/5] drm/modes: parse_cmdline: Handle empty mode name part

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 8 Jul 2022 22:06:42 +0200
From:   Geert Uytterhoeven <geert@...ux-m68k.org>
To:     Hans de Goede <hdegoede@...hat.com>
Cc:     Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Maxime Ripard <mripard@...nel.org>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        David Airlie <airlied@...ux.ie>,
        Daniel Vetter <daniel@...ll.ch>,
        DRI Development <dri-devel@...ts.freedesktop.org>,
        Linux Fbdev development list <linux-fbdev@...r.kernel.org>,
        "Linux/m68k" <linux-m68k@...r.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/5] drm/modes: parse_cmdline: Handle empty mode name part

Hi Hans.

On Fri, Jul 8, 2022 at 9:28 PM Hans de Goede <hdegoede@...hat.com> wrote:
> On 7/8/22 20:21, Geert Uytterhoeven wrote:
> > If no mode name part was specified, mode_end is zero, and the "ret ==
> > mode_end" check does the wrong thing.
> >
> > Fix this by checking for a non-zero return value instead.
>
> Which is wrong to do, since now if you have e.g. a mode list
> with:
>
> "dblntsc",
> "dblntsc-ff"
>
> in there and the cmdline contains "dblntsc-ff" then you
> will already stop with a (wrong!) match at "dblntsc".

It indeed behaves that way, and did so before, as str_has_prefix()
checks for a matching prefix, and thus may never get to the full
match.  However, can we change that to an exact match, without
introducing regressions?
This can be avoided by reverse-sorting the modelist (or iterating
backwards through a sorted modelist), though.

> > While at it, skip all named mode handling when mode_end is zero, as it
> > is futile.
>
> AFAICT, this is actually what needs to be done to fix this, while keeping
> the ret == mode_end check.

"ret == mode_end" or "ret" doesn't matter (except for the special
case of mode_end is zero), as str_has_prefix() returns either zero or
the length of the prefix.  Hence it never returns a non-zero value
smaller than the length of the prefix.

> > Fixes: 7b1cce760afe38b4 ("drm/modes: parse_cmdline: Allow specifying stand-alone options")
> > Signed-off-by: Geert Uytterhoeven <geert@...ux-m68k.org>
> > ---
> >  drivers/gpu/drm/drm_modes.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c
> > index 14b746f7ba975954..30a7be97707bfb16 100644
> > --- a/drivers/gpu/drm/drm_modes.c
> > +++ b/drivers/gpu/drm/drm_modes.c
> > @@ -1823,9 +1823,9 @@ bool drm_mode_parse_command_line_for_connector(const char *mode_option,
> >       }
> >
> >       /* First check for a named mode */
> > -     for (i = 0; i < ARRAY_SIZE(drm_named_modes_whitelist); i++) {
> > +     for (i = 0; mode_end && i < ARRAY_SIZE(drm_named_modes_whitelist); i++) {
> >               ret = str_has_prefix(name, drm_named_modes_whitelist[i]);
> > -             if (ret == mode_end) {
> > +             if (ret) {
> >                       if (refresh_ptr)
> >                               return false; /* named + refresh is invalid */

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@...ux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds