linux-kernel - Re: [PATCH v2] staging: gpib: Fix memory leak in ni_usb

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <2025122925-navigator-settle-9873@gregkh>
Date: Mon, 29 Dec 2025 16:28:21 +0100
From: Greg KH <gregkh@...uxfoundation.org>
To: Zilin Guan <zilin@....edu.cn>
Cc: dpenkler@...il.com, matchstick@...erthere.org, mingo@...nel.org,
	tglx@...utronix.de, linux-kernel@...r.kernel.org,
	Jianhao Xu <jianhao.xu@....edu.cn>
Subject: Re: [PATCH v2] staging: gpib: Fix memory leak in ni_usb_init()

On Mon, Dec 29, 2025 at 03:22:03PM +0000, Zilin Guan wrote:
> In ni_usb_init(), if ni_usb_setup_init() fails, the function returns
> immediately without freeing the allocated memory for writes, leading
> to a memory leak.
> 
> Fix this by freeing writes before returning the error code.
> 
> Fixes: 4e127de14fa7 ("staging: gpib: Add National Instruments USB GPIB driver")
> Suggested-by: Dave Penkler <dpenkler@...il.com>
> Co-developed-by: Jianhao Xu <jianhao.xu@....edu.cn>
> Signed-off-by: Jianhao Xu <jianhao.xu@....edu.cn>
> Signed-off-by: Zilin Guan <zilin@....edu.cn>
> ---
> Changes in v2:
> - Use early return to simplify error handling logic.
> 
>  drivers/gpib/ni_usb/ni_usb_gpib.c | 8 +++++---

As the code is no longer in staging, that prefix does not need to be
ther.

>  1 file changed, 5 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpib/ni_usb/ni_usb_gpib.c b/drivers/gpib/ni_usb/ni_usb_gpib.c
> index 1f8412de9fa3..2352c6817440 100644
> --- a/drivers/gpib/ni_usb/ni_usb_gpib.c
> +++ b/drivers/gpib/ni_usb/ni_usb_gpib.c
> @@ -1799,10 +1799,12 @@ static int ni_usb_init(struct gpib_board *board)
>  		return -ENOMEM;
>  
>  	writes_len = ni_usb_setup_init(board, writes);
> -	if (writes_len)
> -		retval = ni_usb_write_registers(ni_priv, writes, writes_len, &ibsta);
> -	else
> +	if (!writes_len) {
> +		kfree(writes);
>  		return -EFAULT;

This is not the correct error value, it should only happen if copy
to/from user fails.  I know it's not the issue here, just noticed this.

And why doesn't ni_usb_setup_init() return an error value that should be
propagated upward?

thanks,

greg k-h