[PATCH v3] iio: backend: fix out-of-bound write

[PATCH v3] iio: backend: fix out-of-bound write

[Markus Burri]

The buffer is set to 80 character. If a caller write more characters,
count is truncated to the max available space in "simple_write_to_buffer".
But afterwards a string terminator is written to the buffer at offset count
without boundary check. The zero termination is written OUT-OF-BOUND.

Add a check that the given buffer is smaller then the buffer to prevent.

Fixes: 035b4989211d ("iio: backend: make sure to NULL terminate stack buffer")
Signed-off-by: Markus Burri <markus.burri@mt.com>
---
 drivers/iio/industrialio-backend.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/iio/industrialio-backend.c b/drivers/iio/industrialio-backend.c
index a43c8d1bb3d0..4a364e038449 100644
--- a/drivers/iio/industrialio-backend.c
+++ b/drivers/iio/industrialio-backend.c
@@ -155,11 +155,14 @@ static ssize_t iio_backend_debugfs_write_reg(struct file *file,
 	ssize_t rc;
 	int ret;
 
+	if (count >= sizeof(buf) - 1)
+		return -ENOSPC;
+
 	rc = simple_write_to_buffer(buf, sizeof(buf) - 1, ppos, userbuf, count);
 	if (rc < 0)
 		return rc;
 
-	buf[count] = '\0';
+	buf[rc] = '\0';
 
 	ret = sscanf(buf, "%i %i", &back->cached_reg_addr, &val);
 

base-commit: b4432656b36e5cc1d50a1f2dc15357543add530e
-- 
2.39.5

Re: [PATCH v3] iio: backend: fix out-of-bound write

[Nuno Sá]

On Mon, 2025-05-05 at 22:38 +0200, Markus Burri wrote:
> The buffer is set to 80 character. If a caller write more characters,
> count is truncated to the max available space in "simple_write_to_buffer".
> But afterwards a string terminator is written to the buffer at offset count
> without boundary check. The zero termination is written OUT-OF-BOUND.
> 
> Add a check that the given buffer is smaller then the buffer to prevent.
> 
> Fixes: 035b4989211d ("iio: backend: make sure to NULL terminate stack buffer")
> Signed-off-by: Markus Burri <markus.burri@mt.com>
> ---

LGTM

Reviewed-by: Nuno Sá <nuno.sa@analog.com>

>  drivers/iio/industrialio-backend.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/industrialio-backend.c b/drivers/iio/industrialio-
> backend.c
> index a43c8d1bb3d0..4a364e038449 100644
> --- a/drivers/iio/industrialio-backend.c
> +++ b/drivers/iio/industrialio-backend.c
> @@ -155,11 +155,14 @@ static ssize_t iio_backend_debugfs_write_reg(struct file
> *file,
>  	ssize_t rc;
>  	int ret;
>  
> +	if (count >= sizeof(buf) - 1)
> +		return -ENOSPC;
> +
>  	rc = simple_write_to_buffer(buf, sizeof(buf) - 1, ppos, userbuf, count);
>  	if (rc < 0)
>  		return rc;
>  
> -	buf[count] = '\0';
> +	buf[rc] = '\0';

you could have mentioned this change in the commit message...

> 
>  	ret = sscanf(buf, "%i %i", &back->cached_reg_addr, &val);
>  
> 
> base-commit: b4432656b36e5cc1d50a1f2dc15357543add530e

Re: [PATCH v3] iio: backend: fix out-of-bound write

[David Lechner]

On 5/5/25 3:38 PM, Markus Burri wrote:
> The buffer is set to 80 character. If a caller write more characters,
> count is truncated to the max available space in "simple_write_to_buffer".
> But afterwards a string terminator is written to the buffer at offset count
> without boundary check. The zero termination is written OUT-OF-BOUND.
> 
> Add a check that the given buffer is smaller then the buffer to prevent.
> 
> Fixes: 035b4989211d ("iio: backend: make sure to NULL terminate stack buffer")
> Signed-off-by: Markus Burri <markus.burri@mt.com>
> ---
>  drivers/iio/industrialio-backend.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/iio/industrialio-backend.c b/drivers/iio/industrialio-backend.c
> index a43c8d1bb3d0..4a364e038449 100644
> --- a/drivers/iio/industrialio-backend.c
> +++ b/drivers/iio/industrialio-backend.c
> @@ -155,11 +155,14 @@ static ssize_t iio_backend_debugfs_write_reg(struct file *file,
>  	ssize_t rc;
>  	int ret;
>  
> +	if (count >= sizeof(buf) - 1)

Isn't it OK if count == sizeof(buf) - 1? In other words, should be:

	if (count >= sizeof(buf))

> +		return -ENOSPC;
> +
>  	rc = simple_write_to_buffer(buf, sizeof(buf) - 1, ppos, userbuf, count);
>  	if (rc < 0)
>  		return rc;
>  
> -	buf[count] = '\0';
> +	buf[rc] = '\0';
>  
>  	ret = sscanf(buf, "%i %i", &back->cached_reg_addr, &val);
>  
> 
> base-commit: b4432656b36e5cc1d50a1f2dc15357543add530e

It looks like we have the same or similar bugs in:

drivers/accel/ivpu/ivpu_debugfs.c
drivers/gpio/gpio-virtuser.c
drivers/iio/industrialio-core.c
drivers/iio/dac/ad3552r-hs.c

Do you plan to fix these as well? 

Re: [PATCH v3] iio: backend: fix out-of-bound write

[Nuno Sá]

On Tue, 2025-05-06 at 12:00 -0500, David Lechner wrote:
> On 5/5/25 3:38 PM, Markus Burri wrote:
> > The buffer is set to 80 character. If a caller write more characters,
> > count is truncated to the max available space in "simple_write_to_buffer".
> > But afterwards a string terminator is written to the buffer at offset count
> > without boundary check. The zero termination is written OUT-OF-BOUND.
> > 
> > Add a check that the given buffer is smaller then the buffer to prevent.
> > 
> > Fixes: 035b4989211d ("iio: backend: make sure to NULL terminate stack buffer")
> > Signed-off-by: Markus Burri <markus.burri@mt.com>
> > ---
> >  drivers/iio/industrialio-backend.c | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/iio/industrialio-backend.c b/drivers/iio/industrialio-
> > backend.c
> > index a43c8d1bb3d0..4a364e038449 100644
> > --- a/drivers/iio/industrialio-backend.c
> > +++ b/drivers/iio/industrialio-backend.c
> > @@ -155,11 +155,14 @@ static ssize_t iio_backend_debugfs_write_reg(struct file
> > *file,
> >  	ssize_t rc;
> >  	int ret;
> >  
> > +	if (count >= sizeof(buf) - 1)
> 
> Isn't it OK if count == sizeof(buf) - 1? In other words, should be:
> 
> 	if (count >= sizeof(buf))

Oh, indeed you're right. Sorry Mark! I was the one asking for it but I did not
realized the comparison was '>='. So I was thinking if (count > sizeof(buf) - 1)
which is pretty much if (count >= sizeof(buf))

Arghh... Maybe Jonathan can tweak this while applying so you do not have to spin
another version because of this.

- Nuno Sá

> 
> > +		return -ENOSPC;
> > +
> >  	rc = simple_write_to_buffer(buf, sizeof(buf) - 1, ppos, userbuf, count);
> >  	if (rc < 0)
> >  		return rc;
> >  
> > -	buf[count] = '\0';
> > +	buf[rc] = '\0';
> >  
> >  	ret = sscanf(buf, "%i %i", &back->cached_reg_addr, &val);
> >  
> > 
> > base-commit: b4432656b36e5cc1d50a1f2dc15357543add530e
> 
> It looks like we have the same or similar bugs in:
> 
> drivers/accel/ivpu/ivpu_debugfs.c
> drivers/gpio/gpio-virtuser.c
> drivers/iio/industrialio-core.c
> drivers/iio/dac/ad3552r-hs.c
> 
> Do you plan to fix these as well? 

Re: EXTERNAL - [PATCH v3] iio: backend: fix out-of-bound write

[Markus Burri]

On Tue, May 06, 2025 at 12:00:19PM -0500, David Lechner wrote:
> On 5/5/25 3:38 PM, Markus Burri wrote:
> > The buffer is set to 80 character. If a caller write more characters,
> > count is truncated to the max available space in "simple_write_to_buffer".
> > But afterwards a string terminator is written to the buffer at offset count
> > without boundary check. The zero termination is written OUT-OF-BOUND.
> > 
> > Add a check that the given buffer is smaller then the buffer to prevent.
> > 
> > Fixes: 035b4989211d ("iio: backend: make sure to NULL terminate stack buffer")
> > Signed-off-by: Markus Burri <markus.burri@mt.com>
> > ---
> >  drivers/iio/industrialio-backend.c | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/iio/industrialio-backend.c b/drivers/iio/industrialio-backend.c
> > index a43c8d1bb3d0..4a364e038449 100644
> > --- a/drivers/iio/industrialio-backend.c
> > +++ b/drivers/iio/industrialio-backend.c
> > @@ -155,11 +155,14 @@ static ssize_t iio_backend_debugfs_write_reg(struct file *file,
> >  	ssize_t rc;
> >  	int ret;
> >  
> > +	if (count >= sizeof(buf) - 1)
> 
> Isn't it OK if count == sizeof(buf) - 1? In other words, should be:
> 
> 	if (count >= sizeof(buf))
>
This was my original patch and I think it is OK.
In a situation we have 79 characters and the last one ('\n'), the '\n' will be
replaces by a '\0', therefore it is OK.
Since the given text should anyway be < buffer size and it is a little more
correct to have the -1, I would keep it.

> > +		return -ENOSPC;
> > +
> >  	rc = simple_write_to_buffer(buf, sizeof(buf) - 1, ppos, userbuf, count);
> >  	if (rc < 0)
> >  		return rc;
> >  
> > -	buf[count] = '\0';
> > +	buf[rc] = '\0';
> >  
> >  	ret = sscanf(buf, "%i %i", &back->cached_reg_addr, &val);
> >  
> > 
> > base-commit: b4432656b36e5cc1d50a1f2dc15357543add530e
> 
> It looks like we have the same or similar bugs in:
> 
> drivers/accel/ivpu/ivpu_debugfs.c
> drivers/gpio/gpio-virtuser.c
> drivers/iio/industrialio-core.c
> drivers/iio/dac/ad3552r-hs.c
> 
> Do you plan to fix these as well? 
True there are some more. I will check them later.