netdev - Re: BUG: stack guard page was hit in unwind_next

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

Message-ID: <20200507095949.GD9365@kadam>
Date:   Thu, 7 May 2020 12:59:49 +0300
From:   Dan Carpenter <dan.carpenter@...cle.com>
To:     Hillf Danton <hdanton@...a.com>
Cc:     syzbot <syzbot+e73ceacfd8560cc8a3ca@...kaller.appspotmail.com>,
        linux-kernel@...r.kernel.org, netdev@...r.kernel.org,
        syzkaller-bugs@...glegroups.com
Subject: Re: BUG: stack guard page was hit in unwind_next_frame

On Sun, May 03, 2020 at 06:22:20PM +0800, Hillf Danton wrote:
> Bail out if it's detected to handle the event more than once.
> 
> --- a/drivers/net/bonding/bond_main.c
> +++ b/drivers/net/bonding/bond_main.c
> @@ -3273,9 +3273,19 @@ static int bond_netdev_event(struct noti
>  			return ret;
>  	}
>  
> -	if (event_dev->flags & IFF_SLAVE)
> -		return bond_slave_netdev_event(event, event_dev);
> +	if (event_dev->flags & IFF_SLAVE) {
> +		static void *tail_spin = NULL;
                             ^^^^^^^^^^^^^^^^
assigning NULL

> +		void *token = (void *) this + (void *) event_dev;

Adding a pointer to a pointer doesn't make any sense.  But the result
is non-NULL because event_dev is non-NULL.

> +
> +		if (tail_spin == token) {
                    ^^^^^^^^^^^^^^^^^^
Impossible because tail_spin is NULL and token is non-NULL.

> +			tail_spin = NULL;
                        ^^^^^^^^^^^^^^^^
re-assigning NULL.  local variable assigned right before a return is
pointless.

> +			return NOTIFY_DONE;
> +		}
> +		if (tail_spin == NULL)

Always true condition.

> +			tail_spin = token;

Pointless assign.

>  
> +		return bond_slave_netdev_event(event, event_dev);

This whole patch is a very complicated no-op.  :P  I'm not sure at all
what was intended by this patch.

> +	}
>  	return NOTIFY_DONE;
>  }

regards,
dan carpenter