From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760246Ab2COITZ (ORCPT ); Thu, 15 Mar 2012 04:19:25 -0400 Received: from mail-ey0-f174.google.com ([209.85.215.174]:33370 "EHLO mail-ey0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756256Ab2COITF convert rfc822-to-8bit (ORCPT ); Thu, 15 Mar 2012 04:19:05 -0400 Content-Type: text/plain; charset=utf-8; format=flowed; delsp=yes To: "Felipe Balbi" , Yongsul@vger.kernel.org, Oh@vger.kernel.org Cc: "Greg Kroah-Hartman" , linux-usb@vger.kernel.org, linux-kernel@vger.kernel.org, "yongsul96.oh@samsung.com" Subject: Re: [PATCH] USB: Gadget: Composite: Added error handling codes to prevent a memory leak case when the configuration's bind function failed References: <1331774635-21217-1-git-send-email-yongsul96.oh@samsung.com> Date: Thu, 15 Mar 2012 09:19:01 +0100 MIME-Version: 1.0 Content-Transfer-Encoding: 8BIT From: "Michal Nazarewicz" Message-ID: In-Reply-To: <1331774635-21217-1-git-send-email-yongsul96.oh@samsung.com> User-Agent: Opera Mail/11.61 (Linux) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 15 Mar 2012 02:23:55 +0100, wrote: > In some usb gadget driver, for example usb gadget serial driver(serial.c), > multifunction composite driver(multi,c), nokia composite gadget driver(nokia.c), > HID composite driver(hid.c), CDC composite driver(cdc2.c).., the configuration's > bind function by called the 'usb_add_config()' has multiple bind config functions > for each functionality, for example cdc2 configuration bind function -'cdc_do_config()' > has two functionality bind config functions -'ecm_bind_config()' & 'acm_bind_config()' > in CDC composite driver. > In each functionality bind config function, new instance for each functionality is > allocated & initialized by 'kzalloc()' ,and finally the new instance is added by > 'usb_add_function()'. After 'usb_add_function' state, already created the instance > is only handled by its configuration & freed from functionality unbind function. > So, If an error occurred during the second functionality bind config state, for example > an error occurred at 'acm_bind_config()' after succeeding 'ecm_bind_function()', > The created instance by 'acm_bind_config()' cannot be freed. > And it makes memory leak situation. It may be my poor English skills, but this is very hard to read. Also, should this be handled in usb_gadget_probe_driver() instead? Ie. on error recovery path, should usb_gadget_probe_driver() call usb_gadget_remove_driver()? In particular, if I'm not mistaken, for (some?) gadgets with two configurations (like g_multi), if the second configuration fails, the first one will not get removed, so the memory will still leak, no? > Signed-off-by: yongsul96.oh@samsung.com > --- > drivers/usb/gadget/composite.c | 13 +++++++++++++ > 1 files changed, 13 insertions(+), 0 deletions(-) > > diff --git a/drivers/usb/gadget/composite.c b/drivers/usb/gadget/composite.c > index baaebf2..4cb1801 100644 > --- a/drivers/usb/gadget/composite.c > +++ b/drivers/usb/gadget/composite.c > @@ -737,6 +737,19 @@ int usb_add_config(struct usb_composite_dev *cdev, > status = bind(config); > if (status < 0) { > + while (!list_empty(&config->functions)) { > + struct usb_function *f; > + > + f = list_first_entry(&config->functions, > + struct usb_function, list); > + list_del(&f->list); > + if (f->unbind) { > + DBG(cdev, "unbind function '%s'/%p\n", > + f->name, f); > + f->unbind(config, f); > + /* may free memory for "f" */ > + } > + } > list_del(&config->list); > config->cdev = NULL; > } else { -- Best regards, _ _ .o. | Liege of Serenely Enlightened Majesty of o' \,=./ `o ..o | Computer Science, Michał “mina86” Nazarewicz (o o) ooo +------------------ooO--(_)--Ooo--