[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <2023092552-tulip-yo-yo-cbb3@gregkh>
Date: Mon, 25 Sep 2023 11:38:38 +0200
From: Greg KH <gregkh@...uxfoundation.org>
To: Jia-Ju Bai <baijiaju@...a.edu.cn>
Cc: hminas@...opsys.com, linux-usb@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] usb: dwc2: fix possible NULL pointer dereference caused
by driver concurrency
On Mon, Sep 25, 2023 at 05:17:41PM +0800, Jia-Ju Bai wrote:
> In _dwc2_hcd_urb_enqueue(), "urb->hcpriv = NULL" is executed without
> holding the lock "hsotg->lock". In _dwc2_hcd_urb_dequeue():
>
> spin_lock_irqsave(&hsotg->lock, flags);
> ...
> if (!urb->hcpriv) {
> dev_dbg(hsotg->dev, "## urb->hcpriv is NULL ##\n");
> goto out;
> }
> rc = dwc2_hcd_urb_dequeue(hsotg, urb->hcpriv); // Use urb->hcpriv
> ...
> out:
> spin_unlock_irqrestore(&hsotg->lock, flags);
>
> When _dwc2_hcd_urb_enqueue() and _dwc2_hcd_urb_dequeue() are
> concurrently executed, the NULL check of "urb->hcpriv" can be executed
> before "urb->hcpriv = NULL". After urb->hcpriv is NULL, it can be used
> in the function call to dwc2_hcd_urb_dequeue(), which can cause a NULL
> pointer dereference.
Odd trailing spaces in your changelog text, is that intentional?
>
> To fix this possible bug, "urb->hcpriv = NULL" should be executed with
> holding the lock "hsotg->lock".
>
> Signed-off-by: Jia-Ju Bai <baijiaju@...a.edu.cn>
> ---
> drivers/usb/dwc2/hcd.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/usb/dwc2/hcd.c b/drivers/usb/dwc2/hcd.c
> index 657f1f659ffa..35c7a4df8e71 100644
> --- a/drivers/usb/dwc2/hcd.c
> +++ b/drivers/usb/dwc2/hcd.c
> @@ -4769,8 +4769,8 @@ static int _dwc2_hcd_urb_enqueue(struct usb_hcd *hcd, struct urb *urb,
> if (qh_allocated && qh->channel && qh->channel->qh == qh)
> qh->channel->qh = NULL;
> fail2:
> - spin_unlock_irqrestore(&hsotg->lock, flags);
> urb->hcpriv = NULL;
> + spin_unlock_irqrestore(&hsotg->lock, flags);
> kfree(qtd);
> fail1:
> if (qh_allocated) {
> --
> 2.34.1
>
What commit id does this fix?
And how did you test this to verify it works properly?
And how was it found?
thanks,
greg k-h
Powered by blists - more mailing lists