[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20180928080810.Horde.QIGMZmp54QWS1jzs1zoa7_a@webmail.kiener-muenchen.de>
Date: Fri, 28 Sep 2018 08:08:10 +0000
From: guido@...ner-muenchen.de
To: Colin King <colin.king@...onical.com>
Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Steve Bayless <steve_bayless@...sight.com>,
linux-usb@...r.kernel.org, kernel-janitors@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] usb: usbtmc: check size before allocating buffer and
remove duplicated allocation
Zitat von Colin King <colin.king@...onical.com>:
> From: Colin Ian King <colin.king@...onical.com>
>
> Currently the allocation of a buffer is performed before a sanity check on
> the required buffer size and if the buffer size is too large the error exit
> return leaks the allocated buffer. Fix this by checking the size before
> allocating.
>
> Also, the same buffer is allocated again inside an if statement, causing
> the first allocation to be leaked. Fix this by removing this second
> redundant allocation.
>
> Detected by CoverityScan, CID#1473697 ("Resource leak")
>
> Fixes: 658f24f4523e ("usb: usbtmc: Add ioctl for generic requests on
> control")
> Signed-off-by: Colin Ian King <colin.king@...onical.com>
> ---
> drivers/usb/class/usbtmc.c | 10 +++-------
> 1 file changed, 3 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/usb/class/usbtmc.c b/drivers/usb/class/usbtmc.c
> index 0fcb81a1399b..c01edff190d2 100644
> --- a/drivers/usb/class/usbtmc.c
> +++ b/drivers/usb/class/usbtmc.c
> @@ -1895,18 +1895,14 @@ static int usbtmc_ioctl_request(struct
> usbtmc_device_data *data,
> if (res)
> return -EFAULT;
>
> + if (request.req.wLength > USBTMC_BUFSIZE)
> + return -EMSGSIZE;
> +
> buffer = kmalloc(request.req.wLength, GFP_KERNEL);
> if (!buffer)
> return -ENOMEM;
>
> - if (request.req.wLength > USBTMC_BUFSIZE)
> - return -EMSGSIZE;
> -
> if (request.req.wLength) {
> - buffer = kmalloc(request.req.wLength, GFP_KERNEL);
> - if (!buffer)
> - return -ENOMEM;
> -
> if ((request.req.bRequestType & USB_DIR_IN) == 0) {
> /* Send control data to device */
> res = copy_from_user(buffer, request.data,
> --
> 2.17.1
Good to know that there are some tools finding this memory leak.
I have already sent a patch series here:
https://patchwork.kernel.org/cover/10612963/
This includes a similar patch to your proposal:
https://patchwork.kernel.org/patch/10612965/
Thank you!
Regards,
Guido
Powered by blists - more mailing lists