From: "Michal Nazarewicz" <mina86@mina86.com>
To: "Felipe Balbi" <balbi@ti.com>,
Yongsul@vger.kernel.org, Oh@vger.kernel.org
Cc: "Greg Kroah-Hartman" <gregkh@linuxfoundation.org>,
linux-usb@vger.kernel.org, linux-kernel@vger.kernel.org,
"yongsul96.oh@samsung.com" <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
Date: Thu, 15 Mar 2012 09:19:01 +0100 [thread overview]
Message-ID: <op.wa7ddzix3l0zgt@mpn-glaptop> (raw)
In-Reply-To: <1331774635-21217-1-git-send-email-yongsul96.oh@samsung.com>
On Thu, 15 Mar 2012 02:23:55 +0100, <Yongsul@vger.kernel.org> 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 <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 +----<email/xmpp: mpn@google.com>--------------ooO--(_)--Ooo--
next prev parent reply other threads:[~2012-03-15 8:19 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-03-15 1:23 Yongsul, Oh
2012-03-15 8:19 ` Michal Nazarewicz [this message]
2012-03-15 13:57 ` Sergei Shtylyov
2012-03-16 6:27 Yongsul Oh
2012-03-16 11:14 ` Michal Nazarewicz
2012-03-19 8:34 ` Yongsul Oh
2012-03-20 1:38 Yongsul Oh
2012-03-20 13:41 ` Michal Nazarewicz
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=op.wa7ddzix3l0zgt@mpn-glaptop \
--to=mina86@mina86.com \
--cc=Oh@vger.kernel.org \
--cc=Yongsul@vger.kernel.org \
--cc=balbi@ti.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=yongsul96.oh@samsung.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®