linux-kernel - Re: [PATCH] usb: typec: tcpm: testing array offset 'port->logbuffer

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: <86b338ea-2c6c-a06f-c935-30e2ebdd37ea@roeck-us.net>
Date:   Thu, 24 Mar 2022 21:45:54 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Haowen Bai <baihaowen@...zu.com>, heikki.krogerus@...ux.intel.com,
        gregkh@...uxfoundation.org
Cc:     linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] usb: typec: tcpm: testing array offset
 'port->logbuffer_head' before use

On 3/24/22 21:34, Haowen Bai wrote:
> Fix possible indexing array of bound for
> port->logbuffer[port->logbuffer_head], where port->logbuffer_head boundary
>   check happens later. so we do it before.
> 
> Signed-off-by: Haowen Bai <baihaowen@...zu.com>
> ---
>   drivers/usb/typec/tcpm/tcpm.c | 15 ++++++++-------
>   1 file changed, 8 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/usb/typec/tcpm/tcpm.c b/drivers/usb/typec/tcpm/tcpm.c
> index 5fce795..541e9e4 100644
> --- a/drivers/usb/typec/tcpm/tcpm.c
> +++ b/drivers/usb/typec/tcpm/tcpm.c
> @@ -591,6 +591,14 @@ static void _tcpm_log(struct tcpm_port *port, const char *fmt, va_list args)
>   	unsigned long rem_nsec;
>   
>   	mutex_lock(&port->logbuffer_lock);
> +
> +	if (port->logbuffer_head < 0 ||
> +	    port->logbuffer_head >= LOG_BUFFER_ENTRIES) {
> +		dev_warn(port->dev,
> +			 "Bad log buffer index %d\n", port->logbuffer_head);
> +		goto abort;
> +	}
> +
>   	if (!port->logbuffer[port->logbuffer_head]) {
>   		port->logbuffer[port->logbuffer_head] =
>   				kzalloc(LOG_BUFFER_ENTRY_SIZE, GFP_KERNEL);
> @@ -607,13 +615,6 @@ static void _tcpm_log(struct tcpm_port *port, const char *fmt, va_list args)
>   		strcpy(tmpbuffer, "overflow");
>   	}
>   
> -	if (port->logbuffer_head < 0 ||
> -	    port->logbuffer_head >= LOG_BUFFER_ENTRIES) {
> -		dev_warn(port->dev,
> -			 "Bad log buffer index %d\n", port->logbuffer_head);
> -		goto abort;
> -	}
> -
>   	if (!port->logbuffer[port->logbuffer_head]) {
>   		dev_warn(port->dev,
>   			 "Log buffer index %d is NULL\n", port->logbuffer_head);

One could argue that the check is unnecessary and can be removed as it can be proven
that it logbuffer_head always in the range of [0, LOG_BUFFER_ENTRIES - 1]. Moving
the check, however, does not add any value unless you can _prove_ that it needs to
be moved, ie that logbuffer_head can be observed to be < 0 or >= LOG_BUFFER_ENTRIES.
I think that is quite unlikely.

Guenter