netdev - [PATCH] Cleanup hso rfkill error handling [was: 2.6.28-rc2 / hso driver oops]

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: <20081030102746.GI3162@earth.li>
Date:	Thu, 30 Oct 2008 10:27:46 +0000
From:	Jonathan McDowell <noodles@...th.li>
To:	linux-kernel@...r.kernel.org, netdev@...r.kernel.org
Cc:	Greg Kroah-Hartman <gregkh@...e.de>,
	Andrew Bird <ajb@...eresystems.co.uk>
Subject: [PATCH] Cleanup hso rfkill error handling [was: 2.6.28-rc2 / hso driver oops]

On Wed, Oct 29, 2008 at 07:40:11PM +0000, Ben Hutchings wrote:
> On Wed, 2008-10-29 at 16:40 +0000, Jonathan McDowell wrote:
> > Hi.
> > 
> > Tried out 2.6.28-rc2 today on my EEE 901 and my Option Icon 225 and got
> > the following oops:
> > 
> > hso: drivers/net/usb/hso.c: 1.2 Option Wireless
> > usbcore: registered new interface driver hso
> > usb 2-2: new full speed USB device using uhci_hcd and address 3
> > usb 2-2: configuration #1 chosen from 1 choice
> > hso0: Disabled Privacy Extensions
> > BUG: unable to handle kernel NULL pointer dereference at 000000d0
> > IP: [<c03589b9>] dev_driver_string+0x1/0x2a
> 
> Something passed a null device pointer to dev_printk().
> 
> [...]
> > [<f81b8265>] ? hso_create_net_device+0x305/0x32d [hso]
> [...]
> 
> I think that hso_create_rfkill() is the culprit here (and has been
> inlined into hso_create_net_device()).  It's using hso_dev->dev as the
> first argument to dev_err() and it doesn't look like that field is
> initialised except by kzalloc.  At a guess, it should be using
> &hso_dev->usb->dev.

Yup, this appears to be the problem, thanks. I think &hso_net->net->dev
is more intuitive for the error message, so I've used that. I've also
added missing line endings on the error messages and set our local
rfkill structure element to NULL on failure so we don't try to call
rfkill_unregister on driver removal if we failed to register at all.

The patch below Works For Me (TM); the device is detected fine, can be
removed without problems and connects ok. I'll have a prod at why the
rfkill stuff isn't working next, but I believe this cleanup of the error
handling is appropriate no matter what the issue with registration is.

Signed-Off-By: Jonathan McDowell <noodles@...th.li>

-----
diff --git a/drivers/net/usb/hso.c b/drivers/net/usb/hso.c
index 1164c52..9d9622b 100644
--- a/drivers/net/usb/hso.c
+++ b/drivers/net/usb/hso.c
@@ -2184,19 +2184,20 @@ static void hso_create_rfkill(struct hso_device *hso_dev,
 			     struct usb_interface *interface)
 {
 	struct hso_net *hso_net = dev2net(hso_dev);
-	struct device *dev = hso_dev->dev;
+	struct device *dev = &hso_net->net->dev;
 	char *rfkn;
 
 	hso_net->rfkill = rfkill_allocate(&interface_to_usbdev(interface)->dev,
 				 RFKILL_TYPE_WLAN);
 	if (!hso_net->rfkill) {
-		dev_err(dev, "%s - Out of memory", __func__);
+		dev_err(dev, "%s - Out of memory\n", __func__);
 		return;
 	}
 	rfkn = kzalloc(20, GFP_KERNEL);
 	if (!rfkn) {
 		rfkill_free(hso_net->rfkill);
-		dev_err(dev, "%s - Out of memory", __func__);
+		hso_net->rfkill = NULL;
+		dev_err(dev, "%s - Out of memory\n", __func__);
 		return;
 	}
 	snprintf(rfkn, 20, "hso-%d",
@@ -2209,7 +2210,8 @@ static void hso_create_rfkill(struct hso_device *hso_dev,
 		kfree(rfkn);
 		hso_net->rfkill->name = NULL;
 		rfkill_free(hso_net->rfkill);
-		dev_err(dev, "%s - Failed to register rfkill", __func__);
+		hso_net->rfkill = NULL;
+		dev_err(dev, "%s - Failed to register rfkill\n", __func__);
 		return;
 	}
 }
-----

J.

-- 
/-\                             | 101 things you can't have too much
|@/  Debian GNU/Linux Developer |         of : 32 - Answers.
\-                              |
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html