linux-kernel - Re: [PATCH v2 1/2] printk: Check valid console index for preferred console

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: <20231011161713.GE34982@atomide.com>
Date:   Wed, 11 Oct 2023 19:17:13 +0300
From:   Tony Lindgren <tony@...mide.com>
To:     Petr Mladek <pmladek@...e.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Steven Rostedt <rostedt@...dmis.org>,
        John Ogness <john.ogness@...utronix.de>,
        Sergey Senozhatsky <senozhatsky@...omium.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/2] printk: Check valid console index for preferred
 console

* Petr Mladek <pmladek@...e.com> [231011 15:26]:
> On Wed 2023-10-11 12:18:03, Tony Lindgren wrote:
> > * Greg Kroah-Hartman <gregkh@...uxfoundation.org> [231011 07:53]:
> > > On Wed, Oct 11, 2023 at 10:43:25AM +0300, Tony Lindgren wrote:
> > > > Let's check for valid console index values to avoid bogus console index
> > > > numbers from kernel command line. While struct console uses short for
> > > > index, and negative index values are used by some device drivers, we do
> > > > not want to allow negative values for preferred console.
> > > 
> > > What drivers use a negative index for the console?
> > 
> > This is based on grepping with $ git grep "co->index =" drivers/tty/
> > 
> > Not sure what all might be stopping making struct console index unsigned.
> 
> The value -1 is used for initializing struct console, see:
> 
> $> git grep -A10 "struct console.*=" | \
>    grep -e "struct console" -e index | \
>    grep -B1 index
> [...]
> drivers/tty/serial/8250/8250_core.c:static struct console univ8250_console = {
> drivers/tty/serial/8250/8250_core.c-    .index          = -1,
> [...]
> drivers/tty/serial/imx.c:static struct console imx_uart_console = {
> drivers/tty/serial/imx.c-       .index          = -1,
> drivers/tty/serial/ip22zilog.c:static struct console ip22zilog_console = {
> drivers/tty/serial/ip22zilog.c- .index  =       -1,
> drivers/tty/serial/kgdb_nmi.c:static struct console kgdb_nmi_console = {
> drivers/tty/serial/kgdb_nmi.c-  .index  = -1,
> drivers/tty/serial/lantiq.c:static struct console lqasc_console = {
> drivers/tty/serial/lantiq.c-    .index =        -1,
> drivers/tty/serial/liteuart.c:static struct console liteuart_console = {
> drivers/tty/serial/liteuart.c-  .index = -1,
> drivers/tty/serial/lpc32xx_hs.c:static struct console lpc32xx_hsuart_console = {
> drivers/tty/serial/lpc32xx_hs.c-        .index          = -1,
> drivers/tty/serial/ma35d1_serial.c:static struct console ma35d1serial_console = {
> drivers/tty/serial/ma35d1_serial.c-     .index   = -1,
> drivers/tty/serial/mcf.c:static struct console mcf_console = {
> drivers/tty/serial/mcf.c-       .index          = -1,
> [...]
> 
> It means that the index still has be get assigned. For example, it is
> used here:
> 
> static int try_enable_preferred_console(struct console *newcon,
> 					bool user_specified)
> {
> [...]
> 			if (newcon->index < 0)
> 				newcon->index = c->index;
> [...]

OK

> Resume:
> 
> 1. We must either keep signed short in struct console or
>    use another check for non-yet assigned index.

OK thanks for clarifying the usage for it.

> 2. We should fix the commit message and the comment. We should
>    explain that negative value is used in struct console
>    to distinguish a non-yet-registered/assigned index/port.

I'll send v3 patches tomorrow with updated commit message and comments.

Thanks,

Tony