linux-kernel - Re: Question about dynamic minor number of misc device

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]

Message-ID: <CAK8P3a3DwaZnRff7CCrJoSxP_MeVUn1S6nRd+hb5rHnv9dBgLQ@mail.gmail.com>
Date:   Mon, 20 Jan 2020 11:03:29 +0100
From:   Arnd Bergmann <arnd@...db.de>
To:     Zhenzhong Duan <zhenzhong.duan@...il.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: Question about dynamic minor number of misc device

On Mon, Jan 20, 2020 at 9:33 AM Zhenzhong Duan <zhenzhong.duan@...il.com> wrote:
>
> Hi Maintainers,
>
> I see there are 64 free slots(0-63) used for misc devices with dynamic
> minor number. But PSMOUSE_MINOR(1) overlaps with that dynamic range.
>
> So if the dynamic minor number exhaust, psaux driver will fail with
> "could not register psaux device, error: -16", is this expected?
> Should we preserve a slot for psaux and serio_raw which use static
> minor number PSMOUSE_MINOR?

Is this a theoretical question, or are you actually running out of dynamic
minor numbers? I would guess that if we change the limit to only allow
dynamic minors 2 through 63, that would technically be a correct
change, but the result would be that in the same situation another random
driver fails, which is not much of an improvement, unless the dynamic
minor numbers also continue into the range above 255.

On a  related note, I checked for drivers that call misc_register()
with a minor number that is not defined in include/linux/misc.h
and found a bunch, including some that have conflicting numbers,
conflicting names  or numbers from the dynamic range:

drivers/staging/speakup/devsynth.c:#define SYNTH_MINOR 25
drivers/staging/speakup/speakup_soft.c:#define SOFTSYNTH_MINOR 26 /*
might as well give it one more than /dev/synth */
drivers/staging/speakup/speakup_soft.c:#define SOFTSYNTHU_MINOR 27 /*
might as well give it one more than /dev/synth */
drivers/macintosh/via-pmu.c:#define PMU_MINOR           154
drivers/macintosh/ans-lcd.h:#define ANSLCD_MINOR                156
drivers/auxdisplay/charlcd.c:#define LCD_MINOR          156
drivers/char/applicom.c:#define AC_MINOR 157
drivers/char/nwbutton.h:#define BUTTON_MINOR 158
arch/arm/include/asm/nwflash.h:#define FLASH_MINOR               160
drivers/sbus/char/envctrl.c:#define ENVCTRL_MINOR       162
drivers/sbus/char/flash.c:#define FLASH_MINOR   152
drivers/sbus/char/uctrl.c:#define UCTRL_MINOR   174
drivers/char/toshiba.c:#define TOSH_MINOR_DEV 181
arch/um/drivers/random.c:#define RNG_MISCDEV_MINOR              183 /*
official */
drivers/auxdisplay/panel.c:#define KEYPAD_MINOR         185
drivers/video/fbdev/pxa3xx-gcu.c:#define MISCDEV_MINOR  197
kernel/power/user.c:#define SNAPSHOT_MINOR      231
drivers/parisc/eisa_eeprom.c:#define    EISA_EEPROM_MINOR 241

If you would like to help clean that up, you are definitely welcome
to send patches.

      Arnd