linux-kernel - Re: [PATCH] leds: uleds: fix unchecked copy_to_user() in uleds

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: <20250509084652.GA2492385@google.com>
Date: Fri, 9 May 2025 09:46:52 +0100
From: Lee Jones <lee@...nel.org>
To: David Lechner <david@...hnology.com>
Cc: Ivan Stepchenko <sid@....spb.ru>, Pavel Machek <pavel@...nel.org>,
	Jacek Anaszewski <j.anaszewski@...sung.com>,
	linux-leds@...r.kernel.org, linux-kernel@...r.kernel.org,
	lvc-project@...uxtesting.org
Subject: Re: [PATCH] leds: uleds: fix unchecked copy_to_user() in uleds_read()

On Thu, 08 May 2025, David Lechner wrote:

> On 5/8/25 9:34 AM, Lee Jones wrote:
> > On Mon, 05 May 2025, Ivan Stepchenko wrote:
> > 
> >> The copy_to_user() is annotated with __must_check, indicating that
> >> its return value must be checked by the caller. Currently, uleds_read()
> >> ignores it. If the userspace buffer is invalid and copy_to_user() fails,
> >> the userspace application may assume it has received fresh data, while
> >> in fact copying has failed. This can leave applications out of sync
> >> with the actual device state.
> >>
> >> Found by Linux Verification Center (linuxtesting.org) with SVACE.
> >>
> >> Fixes: e381322b0190 ("leds: Introduce userspace LED class driver")
> >> Signed-off-by: Ivan Stepchenko <sid@....spb.ru>
> >> ---
> >>  drivers/leds/uleds.c | 11 +++++++----
> >>  1 file changed, 7 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/leds/uleds.c b/drivers/leds/uleds.c
> >> index 374a841f18c3..41bfce43136c 100644
> >> --- a/drivers/leds/uleds.c
> >> +++ b/drivers/leds/uleds.c
> >> @@ -147,10 +147,13 @@ static ssize_t uleds_read(struct file *file, char __user *buffer, size_t count,
> >>  		} else if (!udev->new_data && (file->f_flags & O_NONBLOCK)) {
> >>  			retval = -EAGAIN;
> >>  		} else if (udev->new_data) {
> >> -			retval = copy_to_user(buffer, &udev->brightness,
> >> -					      sizeof(udev->brightness));
> >> -			udev->new_data = false;
> >> -			retval = sizeof(udev->brightness);
> >> +			if (copy_to_user(buffer, &udev->brightness,
> >> +					 sizeof(udev->brightness))) {
> > 
> > This is not good.
> > 
> > Please store the result into a variable and return that instead.
> 
> Every other caller of copy_to_user() in the kernel I've seen ignores the actual
> return value and returns -EFAULT, so I thought this looked correct and it was
> just a quirk of copy_to_user().

Yes, I think you're right.  Interesting.

So my counterproposal is as follows:

--- a/drivers/leds/uleds.c
+++ b/drivers/leds/uleds.c
@@ -147,10 +147,11 @@ static ssize_t uleds_read(struct file *file, char __user *buffer, size_t count,
                } else if (!udev->new_data && (file->f_flags & O_NONBLOCK)) {
                        retval = -EAGAIN;
                } else if (udev->new_data) {
-                       retval = copy_to_user(buffer, &udev->brightness,
-                                             sizeof(udev->brightness));
-                       udev->new_data = false;
-                       retval = sizeof(udev->brightness);
+                       ssize_t size = retval = sizeof(udev->brightness);
+                       if (copy_to_user(buffer, &udev->brightness, size))
+                               retval = -EFAULT;
+                       else
+                               udev->new_data = false;
                }

                mutex_unlock(&udev->mutex);

> >> +				retval = -EFAULT;
> >> +			} else {
> >> +				udev->new_data = false;
> >> +				retval = sizeof(udev->brightness);
> >> +			}
> >>  		}
> >>  
> >>  		mutex_unlock(&udev->mutex);
> >> -- 
> >> 2.39.5
> >>
> > 
> 

-- 
Lee Jones [李琼斯]