Return-Path: Date: Fri, 5 Jul 2013 12:41:37 +0800 From: Adam Lee To: Yang Bai , linux-bluetooth@vger.kernel.org, Wen-chien Jesse Sung , AceLan Kao , Tedd Ho-Jeong An , Anthony Wong' , Marcel Holtmann , Gustavo Padovan , Johan Hedberg , open list Subject: Re: [PATCH] btusb: fix overflow return values Message-ID: <20130705044137.GB10744@adam-laptop> References: <1372941783-30657-1-git-send-email-adam.lee@canonical.com> <20130705025947.GA10744@adam-laptop> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 In-Reply-To: <20130705025947.GA10744@adam-laptop> List-ID: On Fri, Jul 05, 2013 at 10:59:47AM +0800, Adam Lee wrote: > On Fri, Jul 05, 2013 at 10:37:07AM +0800, Yang Bai wrote: > > The return value of?btusb_setup_intel is compared with 0. Code as: > > > > drivers/bluetooth/btusb.c: > > static int btusb_probe(struct usb_interface *intf, > > const struct usb_device_id *id) > > if (id->driver_info & BTUSB_INTEL) > > hdev->setup = btusb_setup_intel; > > > > net/bluetooth/hci_core.c: > > int hci_dev_open(__u16 dev) > > if (hdev->setup && test_bit(HCI_SETUP, &hdev->dev_flags)) > > ret = hdev->setup(hdev); > > > > if (!ret) { > > Yes, for btusb_setup_intel(), the return value is compared with number > "0", doesn't break the judgement. > > But it still overflows stack without this fix. And what if the lower 32bits of PTR_ERR() are all "0"? :D > > On Thu, Jul 4, 2013 at 8:43 PM, Adam Lee wrote: > > > > PTR_ERR() returns a long type value, but btusb_setup_intel() and > > btusb_setup_intel_patching() should return an int type value. > > > > This bug makes the judgement "if (ret < 0)" not working on x86_64 > > architecture systems, leading to failure as below, even panic. -- Regards, Adam Lee Hardware Enablement