Skip to contentSkip to conversation

[2.4] Memleak in drivers/usb/hub.c::usb_reset_device

PATCH Applied Oleg Drokin 3 participants 5 messages Mar 12, 2003 – Mar 19, 2003 USB

lore
Share & tools
lore
#0 lore raw

Hello!

There seems to be a memleak in drivers/usb/hub.c::usb_reset_device() on error exit path. See the patch.
Found with help of smatch + enhanced unfree script.

Bye,
Oleg
===== drivers/usb/hub.c 1.19 vs edited =====

drivers/usb/hub.c +3 -1
@@ -1057,8 +1057,10 @@
 	}
 	ret = usb_get_descriptor(dev, USB_DT_DEVICE, 0, descriptor,
 			sizeof(*descriptor));
-	if (ret < 0)
+	if (ret < 0) {
+		kfree(descriptor);
 		return ret;
+	}
 
 	le16_to_cpus(&descriptor->bcdUSB);
 	le16_to_cpus(&descriptor->idVendor);
#1 lore raw
Quoted 5 lines · Oleg Drokin

Hello!

There seems to be a memleak in drivers/usb/hub.c::usb_reset_device() on error exit path. See the patch.
Found with help of smatch + enhanced unfree script.

Hmm ... and 2.5 allocates it on the stack, which is actually illegal (DMA-to-stack is nonportable). This looks like a case where it'd be good to make 2.5 look more like 2.4 (+ this patch).

- Dave

Quoted 18 lines

Bye,
Oleg
===== drivers/usb/hub.c 1.19 vs edited =====

drivers/usb/hub.c +3 -1
@@ -1057,8 +1057,10 @@
 	}
 	ret = usb_get_descriptor(dev, USB_DT_DEVICE, 0, descriptor,
 			sizeof(*descriptor));
-	if (ret < 0)
+	if (ret < 0) {
+		kfree(descriptor);
 		return ret;
+	}
 
 	le16_to_cpus(&descriptor->bcdUSB);
 	le16_to_cpus(&descriptor->idVendor);
Applied
#2 lore raw
In reply to Oleg Drokin
Quoted 5 lines · Oleg Drokin

Hello!

There seems to be a memleak in drivers/usb/hub.c::usb_reset_device() on error exit path. See the patch.
Found with help of smatch + enhanced unfree script.

Applied to my tree, thanks.

And yes, as David said, there is another kind of error in this area for 2.5. Patches to clean that up would be appreciated.

thanks,

greg k-h

#3 lore raw

Hello!

Quoted 5 lines · Greg KH
Quoted 3 lines

There seems to be a memleak in drivers/usb/hub.c::usb_reset_device() on error exit path. See the patch.
Found with help of smatch + enhanced unfree script.

And yes, as David said, there is another kind of error in this area for 2.5. Patches to clean that up would be appreciated.

Ok, I guess something like that should work:

Bye,
Oleg
===== drivers/usb/core/hub.c 1.57 vs edited =====

drivers/usb/core/hub.c +18 -9
@@ -1175,7 +1175,7 @@
 int usb_reset_device(struct usb_device *dev)
 {
 	struct usb_device *parent = dev->parent;
-	struct usb_device_descriptor descriptor;
+	struct usb_device_descriptor *descriptor;
 	int i, ret, port = -1;
 
 	if (!parent) {
@@ -1224,17 +1224,24 @@
 	 * If nothing changed, we reprogram the configuration and then
 	 * the alternate settings.
 	 */
-	ret = usb_get_descriptor(dev, USB_DT_DEVICE, 0, &descriptor,
-			sizeof(descriptor));
-	if (ret < 0)
+	descriptor = kmalloc(sizeof *descriptor, GFP_NOIO);
+	if (!descriptor) {
+		return -ENOMEM;
+	}
+	ret = usb_get_descriptor(dev, USB_DT_DEVICE, 0, descriptor,
+			sizeof(*descriptor));
+	if (ret < 0) {
+		kfree(descriptor);
 		return ret;
+	}
 
-	le16_to_cpus(&descriptor.bcdUSB);
-	le16_to_cpus(&descriptor.idVendor);
-	le16_to_cpus(&descriptor.idProduct);
-	le16_to_cpus(&descriptor.bcdDevice);
+	le16_to_cpus(&descriptor->bcdUSB);
+	le16_to_cpus(&descriptor->idVendor);
+	le16_to_cpus(&descriptor->idProduct);
+	le16_to_cpus(&descriptor->bcdDevice);
 
-	if (memcmp(&dev->descriptor, &descriptor, sizeof(descriptor))) {
+	if (memcmp(&dev->descriptor, descriptor, sizeof(*descriptor))) {
+		kfree(descriptor);
 		usb_destroy_configuration(dev);
 
 		ret = usb_get_device_descriptor(dev);
@@ -1267,6 +1274,8 @@
 
 		return 1;
 	}
+
+	kfree(descriptor);
 
 	ret = usb_set_configuration(dev, dev->actconfig->desc.bConfigurationValue);
 	if (ret < 0) {
Applied
#4 lore raw
Quoted 10 lines · Oleg Drokin

Hello!

Quoted 5 lines · Greg KH
Quoted 3 lines

There seems to be a memleak in drivers/usb/hub.c::usb_reset_device() on error exit path. See the patch.
Found with help of smatch + enhanced unfree script.

And yes, as David said, there is another kind of error in this area for 2.5. Patches to clean that up would be appreciated.

Ok, I guess something like that should work:

I've applied this to my trees, thanks.

greg k-h