linux-kernel - Re: [PATCH] serial: atmel: Fix Spectre v1 vulnerability reported by smatch

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] [day] [month] [year] [list]

Message-ID: <2023082220-matchless-stagnate-7dab@gregkh>
Date:   Tue, 22 Aug 2023 13:30:25 +0200
From:   Greg KH <gregkh@...uxfoundation.org>
To:     Hari Prasath Gujulan Elango <Hari.PrasathGE@...rochip.com>
Cc:     richard.genoud@...il.com, jirislaby@...nel.org,
        nicolas.ferre@...rochip.com, alexandre.belloni@...tlin.com,
        claudiu.beznea@...on.dev, linux-kernel@...r.kernel.org,
        linux-serial@...r.kernel.org, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH] serial: atmel: Fix Spectre v1 vulnerability reported by
 smatch

On Tue, Aug 22, 2023 at 04:43:21PM +0530, Hari Prasath Gujulan Elango wrote:
> smatch reports the below spectre variant 1 vulnerability.
> 
> drivers/tty/serial/atmel_serial.c:2675 atmel_console_setup() warn: potential spectre issue 'atmel_ports' [r] (local cap)
> 
> Fix the same by using the array_index_nospec() to mitigate this
> potential vulnerability especially because the console index is
> controlled by user-space.
> 
> Signed-off-by: Hari Prasath Gujulan Elango <Hari.PrasathGE@...rochip.com>
> ---
>  drivers/tty/serial/atmel_serial.c | 15 +++++++++++++--
>  1 file changed, 13 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/tty/serial/atmel_serial.c b/drivers/tty/serial/atmel_serial.c
> index 3467a875641a..25f004dd9efd 100644
> --- a/drivers/tty/serial/atmel_serial.c
> +++ b/drivers/tty/serial/atmel_serial.c
> @@ -33,6 +33,7 @@
>  #include <linux/suspend.h>
>  #include <linux/mm.h>
>  #include <linux/io.h>
> +#include <linux/nospec.h>
>  
>  #include <asm/div64.h>
>  #include <asm/ioctls.h>
> @@ -2662,13 +2663,23 @@ static void __init atmel_console_get_options(struct uart_port *port, int *baud,
>  
>  static int __init atmel_console_setup(struct console *co, char *options)
>  {
> -	struct uart_port *port = &atmel_ports[co->index].uart;
> -	struct atmel_uart_port *atmel_port = to_atmel_uart_port(port);
> +	struct uart_port *port;
> +	struct atmel_uart_port *atmel_port;
>  	int baud = 115200;
>  	int bits = 8;
>  	int parity = 'n';
>  	int flow = 'n';
>  
> +	if (unlikely(co->index < 0 || co->index >= ATMEL_MAX_UART))

Only ever use likely/unlikely if you can measure the difference with and
without the marking.  Otherwise do not use it as the compiler and cpu do
a better job than we do in figuring this out.


> +		return -ENODEV;
> +
> +	co->index = array_index_nospec(co->index, ATMEL_MAX_UART);

How exactl is index controlled by userspace such that a spectre gadget
can be used here?  You have to be able to call this multiple times in a
row, unsuccessfully and successfully, how does that happen through the
console api?

thanks,

greg k-h