linux-kernel - Re: [PATVH] media, dvb, IX2505V: Remember to free allocated memory in failure path (ix2505v

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: <alpine.LNX.2.00.1101022012520.11481@swampdragon.chaosbits.net>
Date:	Sun, 2 Jan 2011 20:14:03 +0100 (CET)
From:	Jesper Juhl <jj@...osbits.net>
To:	Malcolm Priestley <tvboxspy@...il.com>
cc:	linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
	Mauro Carvalho Chehab <mchehab@...radead.org>
Subject: Re: [PATVH] media, dvb, IX2505V: Remember to free allocated memory
 in failure path (ix2505v_attach()).

On Sun, 2 Jan 2011, Jesper Juhl wrote:

> On Fri, 31 Dec 2010, Malcolm Priestley wrote:
> 
> > On Fri, 2010-12-31 at 15:51 +0100, Jesper Juhl wrote:
> > > On Fri, 31 Dec 2010, Malcolm Priestley wrote:
> > > 
> > > > On Fri, 2010-12-31 at 00:11 +0100, Jesper Juhl wrote:
> > > > > Hi,
> > > > > 
> > > > > We may leak the storage allocated to 'state' in 
> > > > > drivers/media/dvb/frontends/ix2505v.c::ix2505v_attach() on error.
> > > > > This patch makes sure we free the allocated memory in the failure case.
> > > > > 
> > > > > 
> > > > > Signed-off-by: Jesper Juhl <jj@...osbits.net>
> > > > > ---
> > > > >  ix2505v.c |    1 +
> > > > >  1 file changed, 1 insertion(+)
> > > > > 
> > > > >   Compile tested only.
> > > > > 
> > > > > diff --git a/drivers/media/dvb/frontends/ix2505v.c b/drivers/media/dvb/frontends/ix2505v.c
> > > > > index 55f2eba..fcb173d 100644
> > > > > --- a/drivers/media/dvb/frontends/ix2505v.c
> > > > > +++ b/drivers/media/dvb/frontends/ix2505v.c
> > > > > @@ -293,6 +293,7 @@ struct dvb_frontend *ix2505v_attach(struct dvb_frontend *fe,
> > > > >  		ret = ix2505v_read_status_reg(state);
> > > > >  
> > > > >  		if (ret & 0x80) {
> > > > > +			kfree(state);
> > > > >  			deb_i2c("%s: No IX2505V found\n", __func__);
> > > > >  			goto error;
> > > > >  		}
> > > > > 
> > > > Memory is freed in... 
> > > > 
> > > > error:
> > > > 	ix2505v_release(fe);
> > > > 	return NULL;
> > > > 
> > > > via...
> > > > 
> > > > static int ix2505v_release(struct dvb_frontend *fe)
> > > > {
> > > > 	struct ix2505v_state *state = fe->tuner_priv;
> > > > 
> > > > 	fe->tuner_priv = NULL;
> > > > 	kfree(state);
> > > > 
> > > > 	return 0;
> > > > }
> > > > 
> > > 
> > > Except that 'state' has not been assigned to fe->tuner_priv at this 
> > > point, so ix2505v_release() cannot free the memory that was just 
> > > allocated with kzalloc().
> > > 
> > > 
> > >   state is a local variable:
> > >   		struct ix2505v_state *state = NULL;
> > > 		...
> > > 
> > >   we allocate memory and assign it to 'state' here:
> > >   		state = kzalloc(sizeof(struct ix2505v_state), GFP_KERNEL);
> > >   		if (NULL == state)
> > >   			return NULL;
> > >   	
> > >   		state->config = config;
> > >   		state->i2c = i2c;
> > >   	
> > >   here 'state' is used, but not in a way that saves it anywhere:
> > >   		if (state->config->tuner_write_only) {
> > >   			if (fe->ops.i2c_gate_ctrl)
> > >   				fe->ops.i2c_gate_ctrl(fe, 1);
> > >   	
> > >   this function call involves 'state' but it does not save it anywhere
> > >   either:
> > >   			ret = ix2505v_read_status_reg(state);
> > >   	
> > >   			if (ret & 0x80) {
> > >   				deb_i2c("%s: No IX2505V found\n", __func__);
> > >   so when we jump to error here 'state' still exists only as the local
> > >   variable, it has not been assigned to anything else.
> > >   				goto error;
> > >   			}
> > >   		...
> > >   	error:
> > >   there is no way this function call can free 'state' on this path since
> > >   it has not been assigned to fe->tuner_priv. 
> > >   		ix2505v_release(fe);
> > >   The local variable state goes out of scope here and leaks the memory it
> > >   points to:
> > >   		return NULL;
> > >   	}
> > > 
> > > Am I missing something?
> > 
> > Oh, Sorry, I see it now.
> > 
> > Now there is two options.
> > 
> > Either;
> > 
> > 1) Move fe->tuner_priv = state to below line 287, so it can be released
> > by ix2505v_release and fe->tuner_priv returned to NULL;
> > 
> > 2) or not calling ix2505v_release changing line 314 to kfree(state).
> > fe->tuner_priv will remain NULL through out.
> > 
> [...]
> 
> How about this?
> 
Or actually, I think this is better:


Signed-off-by: Jesper Juhl <jj@...osbits.net>
---
 ix2505v.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/media/dvb/frontends/ix2505v.c b/drivers/media/dvb/frontends/ix2505v.c
index 55f2eba..846ffd3 100644
--- a/drivers/media/dvb/frontends/ix2505v.c
+++ b/drivers/media/dvb/frontends/ix2505v.c
@@ -311,7 +311,7 @@ struct dvb_frontend *ix2505v_attach(struct dvb_frontend *fe,
 	return fe;
 
 error:
-	ix2505v_release(fe);
+	kfree(state);
 	return NULL;
 }
 EXPORT_SYMBOL(ix2505v_attach);



-- 
Jesper Juhl <jj@...osbits.net>            http://www.chaosbits.net/
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/