Hi Dan
The patch dc3c78e43469: "HID: usbhid: Check HID report descriptor contents after device reset" from Apr 3, 2012, leads to the following static checker warning: drivers/hid/usbhid/hid-core.c:1591 hid_reset_resume() warn: bool comparison is always 'true' drivers/hid/usbhid/hid-core.c 1421 static int hid_post_reset(struct usb_interface *intf) 1422 { 1423 struct usb_device *dev = interface_to_usbdev (intf); 1424 struct hid_device *hid = usb_get_intfdata(intf); 1425 struct usbhid_device *usbhid = hid->driver_data; 1426 struct usb_host_interface *interface = intf->cur_altsetting; 1427 int status; 1428 char *rdesc; 1429 1430 /* Fetch and examine the HID report descriptor. If this 1431 * has changed, then rebind. Since usbcore's check of the 1432 * configuration descriptors passed, we already know that 1433 * the size of the HID report descriptor has not changed. 1434 */ 1435 rdesc = kmalloc(hid->dev_rsize, GFP_KERNEL); 1436 if (!rdesc) { 1437 dbg_hid("couldn't allocate rdesc memory (post_reset)\n"); 1438 return 1; We return 1 on error now in this function. I can't figure out if this is intentional.
Originally, this was intentional. The thinking was that if we're stopped from determining whether or not the report descriptor contents have changed, we should act as though they have and indicate to usb_reset_device() that a rebind should occur.
< snip >
1463 1464 return 0; 1465 } [ snip ] 1589 clear_bit(HID_SUSPENDED, &usbhid->iofl); 1590 status = hid_post_reset(intf); 1591 if (status >= 0 && hid->driver && hid->driver->reset_resume) { ^^^^^^^^^^^ This condition is always true. 1592 int ret = hid->driver->reset_resume(hid); 1593 if (ret < 0) 1594 status = ret; Maybe the ->reset_resume() can succeed even though hid_post_reset() failed? In that case, we return the "1" error code and print it in dmesg in usb_resume_interface(). Strange and uncommented.
Yes, it does look like hid_post_reset() should be returning proper error codes on failure, to allow correct behaviour here. usb_reset_device() should still interpret such error code returns as a need to attempt a rebind.
1595 } 1596 return status; 1597 } regards, dan carpenter
Many thanks Simon -- To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html