linux-kernel - Re: [PATCH] net: hso: fix null-ptr-deref during tty device unregistration

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: <YGxqxddOyyDM9ueu@kroah.com>
Date:   Tue, 6 Apr 2021 16:05:57 +0200
From:   Greg KH <gregkh@...uxfoundation.org>
To:     Anirudh Rayabharam <mail@...rudhrb.com>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Oliver Neukum <oneukum@...e.com>,
        Emil Renner Berthing <kernel@...il.dk>,
        Geert Uytterhoeven <geert@...ux-m68k.org>,
        Zheng Yongjun <zhengyongjun3@...wei.com>,
        Rustam Kovhaev <rkovhaev@...il.com>,
        syzbot+c49fe6089f295a05e6f8@...kaller.appspotmail.com,
        linux-usb@...r.kernel.org, netdev@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net: hso: fix null-ptr-deref during tty device
 unregistration

On Tue, Apr 06, 2021 at 06:13:59PM +0530, Anirudh Rayabharam wrote:
> Multiple ttys try to claim the same the minor number causing a double
> unregistration of the same device. The first unregistration succeeds
> but the next one results in a null-ptr-deref.
> 
> The get_free_serial_index() function returns an available minor number
> but doesn't assign it immediately. The assignment is done by the caller
> later. But before this assignment, calls to get_free_serial_index()
> would return the same minor number.
> 
> Fix this by modifying get_free_serial_index to assign the minor number
> immediately after one is found to be and rename it to obtain_minor()
> to better reflect what it does. Similary, rename set_serial_by_index()
> to release_minor() and modify it to free up the minor number of the
> given hso_serial. Every obtain_minor() should have corresponding
> release_minor() call.
> 
> Reported-by: syzbot+c49fe6089f295a05e6f8@...kaller.appspotmail.com
> Tested-by: syzbot+c49fe6089f295a05e6f8@...kaller.appspotmail.com
> 
> Signed-off-by: Anirudh Rayabharam <mail@...rudhrb.com>
> ---
>  drivers/net/usb/hso.c | 32 ++++++++++++--------------------
>  1 file changed, 12 insertions(+), 20 deletions(-)
> 
> diff --git a/drivers/net/usb/hso.c b/drivers/net/usb/hso.c
> index 31d51346786a..295ca330e70c 100644
> --- a/drivers/net/usb/hso.c
> +++ b/drivers/net/usb/hso.c
> @@ -611,7 +611,7 @@ static struct hso_serial *get_serial_by_index(unsigned index)
>  	return serial;
>  }
>  
> -static int get_free_serial_index(void)
> +static int obtain_minor(struct hso_serial *serial)
>  {
>  	int index;
>  	unsigned long flags;
> @@ -619,8 +619,10 @@ static int get_free_serial_index(void)
>  	spin_lock_irqsave(&serial_table_lock, flags);
>  	for (index = 0; index < HSO_SERIAL_TTY_MINORS; index++) {
>  		if (serial_table[index] == NULL) {
> +			serial_table[index] = serial->parent;
> +			serial->minor = index;
>  			spin_unlock_irqrestore(&serial_table_lock, flags);
> -			return index;
> +			return 0;

Minor note, you might want to convert this to use an idr structure in
the future, this "loop and find a free minor" isn't really needed now
that we have a data structure that does this all for us :)

But that's not going to fix this issue, that's for future changes.

thanks,

greg k-h