linux-kernel - RE: Staging: unisys/verisonic: Correct double unlock

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]

Date:	Sat, 2 Apr 2016 23:20:14 +0000
From:	"Sell, Timothy C" <Timothy.Sell@...sys.com>
To:	Iban Rodriguez <iban.rodriguez@....com>,
	"Kershner, David A" <David.Kershner@...sys.com>,
	Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
	Benjamin Romer <benjamin.romer@...sys.com>,
	Neil Horman <nhorman@...hat.com>
CC:	*S-Par-Maintainer <SParMaintainer@...sys.com>,
	"devel@...verdev.osuosl.org" <devel@...verdev.osuosl.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: Staging: unisys/verisonic: Correct double unlock

> -----Original Message-----
> From: Iban Rodriguez [mailto:iban.rodriguez@....com]
> Sent: Saturday, April 02, 2016 1:47 PM
> To: Kershner, David A; Greg Kroah-Hartman; Benjamin Romer; Sell, Timothy
> C; Neil Horman
> Cc: *S-Par-Maintainer; devel@...verdev.osuosl.org; linux-
> kernel@...r.kernel.org; Iban Rodriguez
> Subject: Staging: unisys/verisonic: Correct double unlock
> 
> 'priv_lock' is unlocked twice. The first one is removed and
> the function 'visornic_serverdown_complete' is now called with
> 'priv_lock' locked because 'devdata' is modified inside.
> 
> Signed-off-by: Iban Rodriguez <iban.rodriguez@....com>
> ---
>  drivers/staging/unisys/visornic/visornic_main.c | 1 -
>  1 file changed, 1 deletion(-)
> 
> diff --git a/drivers/staging/unisys/visornic/visornic_main.c
> b/drivers/staging/unisys/visornic/visornic_main.c
> index be0d057346c3..af03f2938fe9 100644
> --- a/drivers/staging/unisys/visornic/visornic_main.c
> +++ b/drivers/staging/unisys/visornic/visornic_main.c
> @@ -368,7 +368,6 @@ visornic_serverdown(struct visornic_devdata
> *devdata,
>  		}
>  		devdata->server_change_state = true;
>  		devdata->server_down_complete_func = complete_func;
> -		spin_unlock_irqrestore(&devdata->priv_lock, flags);
>  		visornic_serverdown_complete(devdata);
>  	} else if (devdata->server_change_state) {
>  		dev_dbg(&devdata->dev->device, "%s changing state\n",

I agree there is a bug here involving priv_lock being unlocked
twice, but this patch isn't the appropriate fix.  Reason is, we can NOT
call visornic_serverdown_complete() while holding a spinlock
(which is what this patch would cause to occur) because
visornic_serverdown_complete() might block when it calls
rtnl_lock() in this code sequence (rtnl_lock() grabs a mutex):

    rtnl_lock();
    dev_close(netdev);
    rtnl_unlock();

Blocking with a spinlock held is always a bad idea.  :-(

> --
> 1.9.1