[PATCH] of/device: Fix of_device_get_modalias() buffer handling

[PATCH] of/device: Fix of_device_get_modalias() buffer handling

[Bjorn Andersson]

of_device_request_module() calls of_device_get_modalias() with "len" 0,
to calculate the size of the buffer needed to store the result, but due
to integer promotion the ssize_t "len" will be compared as unsigned with
strlen(compat) and the loop will generally never break. This results in
a call to snprintf() with a negative len, which triggers below warning,
followed by a dereference of a invalid pointer:

  [    3.060067] WARNING: CPU: 0 PID: 51 at lib/vsprintf.c:2122 vsnprintf+0x348/0x6d8
  ...
  [    3.060301] [<ffffff800891ede8>] vsnprintf+0x348/0x6d8
  [    3.060308] [<ffffff800891f248>] snprintf+0x48/0x50
  [    3.060316] [<ffffff80086a7c80>] of_device_get_modalias+0x108/0x160
  [    3.060322] [<ffffff80086a7cf8>] of_device_request_module+0x20/0x88
  ...

Further more of_device_get_modalias() is supposed to return the number
of bytes needed to store the entire modalias, so the loop needs to
continue accumulate the total size even though the buffer is full.

Finally the function is not expected to ensure space for the NUL, nor
include it in the returned size, so only 1 should be added to the length
of "compat" in the loop (to account for the character 'C').

Fixes: bc575064d688 ("of/device: use of_property_for_each_string to parse compatible strings")
Cc: Rob Herring <robh@kernel.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
 drivers/of/device.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/drivers/of/device.c b/drivers/of/device.c
index 6f33a0e0d351..7cff599a9c6a 100644
--- a/drivers/of/device.c
+++ b/drivers/of/device.c
@@ -195,10 +195,11 @@ EXPORT_SYMBOL(of_device_get_match_data);
 
 static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len)
 {
-	const char *compat, *start = str;
+	const char *compat;
 	char *c;
 	struct property *p;
 	ssize_t csize;
+	ssize_t tsize;
 
 	if ((!dev) || (!dev->of_node))
 		return -ENODEV;
@@ -206,12 +207,16 @@ static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len
 	/* Name & Type */
 	csize = snprintf(str, len, "of:N%sT%s", dev->of_node->name,
 			 dev->of_node->type);
+	tsize = csize;
 	len -= csize;
-	str += csize;
+	if (str)
+		str += csize;
 
 	of_property_for_each_string(dev->of_node, "compatible", p, compat) {
-		if (strlen(compat) + 2 > len)
-			break;
+		csize = strlen(compat) + 1;
+		tsize += csize;
+		if (csize > len)
+			continue;
 
 		csize = snprintf(str, len, "C%s", compat);
 		for (c = str; c; ) {
@@ -223,7 +228,7 @@ static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len
 		str += csize;
 	}
 
-	return str - start;
+	return tsize;
 }
 
 int of_device_request_module(struct device *dev)
-- 
2.12.0

Re: [PATCH] of/device: Fix of_device_get_modalias() buffer handling

[Rob Herring]

On Wed, Aug 23, 2017 at 8:03 PM, Bjorn Andersson
<bjorn.andersson@linaro.org> wrote:
> of_device_request_module() calls of_device_get_modalias() with "len" 0,
> to calculate the size of the buffer needed to store the result, but due
> to integer promotion the ssize_t "len" will be compared as unsigned with
> strlen(compat) and the loop will generally never break. This results in
> a call to snprintf() with a negative len, which triggers below warning,
> followed by a dereference of a invalid pointer:
>
>   [    3.060067] WARNING: CPU: 0 PID: 51 at lib/vsprintf.c:2122 vsnprintf+0x348/0x6d8
>   ...
>   [    3.060301] [<ffffff800891ede8>] vsnprintf+0x348/0x6d8
>   [    3.060308] [<ffffff800891f248>] snprintf+0x48/0x50
>   [    3.060316] [<ffffff80086a7c80>] of_device_get_modalias+0x108/0x160
>   [    3.060322] [<ffffff80086a7cf8>] of_device_request_module+0x20/0x88
>   ...
>
> Further more of_device_get_modalias() is supposed to return the number
> of bytes needed to store the entire modalias, so the loop needs to
> continue accumulate the total size even though the buffer is full.
>
> Finally the function is not expected to ensure space for the NUL, nor
> include it in the returned size, so only 1 should be added to the length
> of "compat" in the loop (to account for the character 'C').
>
> Fixes: bc575064d688 ("of/device: use of_property_for_each_string to parse compatible strings")
> Cc: Rob Herring <robh@kernel.org>
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
>  drivers/of/device.c | 15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/of/device.c b/drivers/of/device.c
> index 6f33a0e0d351..7cff599a9c6a 100644
> --- a/drivers/of/device.c
> +++ b/drivers/of/device.c
> @@ -195,10 +195,11 @@ EXPORT_SYMBOL(of_device_get_match_data);
>
>  static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len)
>  {
> -       const char *compat, *start = str;
> +       const char *compat;
>         char *c;
>         struct property *p;
>         ssize_t csize;
> +       ssize_t tsize;
>
>         if ((!dev) || (!dev->of_node))
>                 return -ENODEV;
> @@ -206,12 +207,16 @@ static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len
>         /* Name & Type */
>         csize = snprintf(str, len, "of:N%sT%s", dev->of_node->name,
>                          dev->of_node->type);
> +       tsize = csize;
>         len -= csize;
> -       str += csize;
> +       if (str)
> +               str += csize;
>
>         of_property_for_each_string(dev->of_node, "compatible", p, compat) {
> -               if (strlen(compat) + 2 > len)
> -                       break;
> +               csize = strlen(compat) + 1;
> +               tsize += csize;
> +               if (csize > len)
> +                       continue;
>
>                 csize = snprintf(str, len, "C%s", compat);

We could just use the snprintf to give us the length. Something like
this following the snprintf:

tsize +=csize;
if (csize > len) {
  if (len) {
    str[0] = '\0';
    len = 0;
  }
  continue;
}

You'd need to prevent len from going negative up above too. It ends up
being more lines but you save a strlen call. So perhaps it's fine as
is, but since I've written it already throwing it out there...

Rob

Re: [PATCH] of/device: Fix of_device_get_modalias() buffer handling

[Bjorn Andersson]

On Thu 24 Aug 09:51 PDT 2017, Rob Herring wrote:

> On Wed, Aug 23, 2017 at 8:03 PM, Bjorn Andersson
> <bjorn.andersson@linaro.org> wrote:
[..]
> > @@ -206,12 +207,16 @@ static ssize_t of_device_get_modalias(struct device *dev, char *str, ssize_t len
> >         /* Name & Type */
> >         csize = snprintf(str, len, "of:N%sT%s", dev->of_node->name,
> >                          dev->of_node->type);
> > +       tsize = csize;
> >         len -= csize;
> > -       str += csize;
> > +       if (str)
> > +               str += csize;
> >
> >         of_property_for_each_string(dev->of_node, "compatible", p, compat) {
> > -               if (strlen(compat) + 2 > len)
> > -                       break;
> > +               csize = strlen(compat) + 1;
> > +               tsize += csize;
> > +               if (csize > len)
> > +                       continue;
> >
> >                 csize = snprintf(str, len, "C%s", compat);
> 
> We could just use the snprintf to give us the length. Something like
> this following the snprintf:
> 
> tsize +=csize;
> if (csize > len) {
>   if (len) {
>     str[0] = '\0';
>     len = 0;
>   }
>   continue;
> }
> 
> You'd need to prevent len from going negative up above too. It ends up
> being more lines but you save a strlen call. So perhaps it's fine as
> is, but since I've written it already throwing it out there...
> 

Indeed, I did attempt a few variants of this as well, but ended up
taking the cost of an extra strlen() to keep the complexity of the code
down.

Regards,
Bjorn