* [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space
@ 2024-02-01 11:28 Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 1/2] of: fdt: fix memory leak " Stefan Kerkmann
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Stefan Kerkmann @ 2024-02-01 11:28 UTC (permalink / raw)
To: Sascha Hauer, BAREBOX; +Cc: Stefan Kerkmann, Ahmad Fatoum
I have encountered the oob write while attempting to modify a large FIT
image with of_property. While hunting for the root cause I noticed that
there is a potential memory leak in fdt_ensure_space as well. Both is
fixed in this series.
Signed-off-by: Stefan Kerkmann <s.kerkmann@pengutronix.de>
---
Changes in v2:
- Review feedback from Ahmad and Sascha
- Link to v1: https://lore.kernel.org/r/20240131-fix-fdt-memory-safety-v1-0-3d3a2c797eec@pengutronix.de
---
Stefan Kerkmann (2):
of: fdt: fix memory leak in fdt_ensure_space
of: fdt: fix oob writes with large fdt properties
drivers/of/fdt.c | 32 +++++++++++++++++++++++++-------
1 file changed, 25 insertions(+), 7 deletions(-)
---
base-commit: cecd3fbea3550532d2175bc13aa479a45e605da0
change-id: 20240131-fix-fdt-memory-safety-b06f9164d953
Best regards,
--
Stefan Kerkmann <s.kerkmann@pengutronix.de>
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH v2 1/2] of: fdt: fix memory leak in fdt_ensure_space
2024-02-01 11:28 [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Stefan Kerkmann
@ 2024-02-01 11:28 ` Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 2/2] of: fdt: fix oob writes with large fdt properties Stefan Kerkmann
2024-02-01 15:10 ` [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Sascha Hauer
2 siblings, 0 replies; 4+ messages in thread
From: Stefan Kerkmann @ 2024-02-01 11:28 UTC (permalink / raw)
To: Sascha Hauer, BAREBOX; +Cc: Stefan Kerkmann, Ahmad Fatoum
If the reallocation failed the old memory remains allocated and is never
freed, this is fixed by freeing the old memory on error.
Signed-off-by: Stefan Kerkmann <s.kerkmann@pengutronix.de>
Reviewed-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
---
drivers/of/fdt.c | 28 +++++++++++++++++++++-------
1 file changed, 21 insertions(+), 7 deletions(-)
diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c
index 5c21bab5de..544294a9ac 100644
--- a/drivers/of/fdt.c
+++ b/drivers/of/fdt.c
@@ -375,24 +375,38 @@ static void *memalign_realloc(void *orig, size_t oldsize, size_t newsize)
static int fdt_ensure_space(struct fdt *fdt, int dtsize)
{
+ size_t new_size;
+ void *previous;
+
/*
* We assume strings and names have a maximum length of 1024
* whereas properties can be longer. We allocate new memory
* if we have less than 1024 bytes (+ the property size left.
*/
if (fdt->str_size - fdt->str_nextofs < 1024) {
- fdt->strings = realloc(fdt->strings, fdt->str_size * 2);
- if (!fdt->strings)
+ previous = fdt->strings;
+ new_size = fdt->str_size * 2;
+
+ fdt->strings = realloc(previous, new_size);
+ if (!fdt->strings) {
+ free(previous);
return -ENOMEM;
- fdt->str_size *= 2;
+ }
+
+ fdt->str_size = new_size;
}
if (fdt->dt_size - fdt->dt_nextofs < 1024 + dtsize) {
- fdt->dt = memalign_realloc(fdt->dt, fdt->dt_size,
- fdt->dt_size * 2);
- if (!fdt->dt)
+ previous = fdt->dt;
+ new_size = fdt->dt_size * 2;
+
+ fdt->dt = memalign_realloc(previous, fdt->dt_size, new_size);
+ if (!fdt->dt) {
+ free(previous);
return -ENOMEM;
- fdt->dt_size *= 2;
+ }
+
+ fdt->dt_size = new_size;
}
return 0;
--
2.39.2
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH v2 2/2] of: fdt: fix oob writes with large fdt properties
2024-02-01 11:28 [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 1/2] of: fdt: fix memory leak " Stefan Kerkmann
@ 2024-02-01 11:28 ` Stefan Kerkmann
2024-02-01 15:10 ` [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Sascha Hauer
2 siblings, 0 replies; 4+ messages in thread
From: Stefan Kerkmann @ 2024-02-01 11:28 UTC (permalink / raw)
To: Sascha Hauer, BAREBOX; +Cc: Stefan Kerkmann
OOB writes can be triggered when fdt->dt_size * 2 is still smaller than
the property for which memory should be allocated. This can happen under
rare circumstances when editing a fdt with the of_property command and a
property is larger than 128k in size.
This happend when editing a FIT image (which is a ftd) with the
of_property command and the Kernel image was around 8M in size.
The simplified call chain is the following:
of_property -> of_flatten_dtb -> create new fdt with 64k in size (this is
fixed) -> __of_flatten_dtb -> attempt to copy kernel image (8M) ->
fdt_ensure_space -> allocate only 128k for fdt->dt -> memcopy 8M into fdt->dt
buffer -> crash
The fix is to grow fdt->dt to hold at least the new property. The power
of 2 increment is untouched to keep the same behaviour otherwise.
Signed-off-by: Stefan Kerkmann <s.kerkmann@pengutronix.de>
---
drivers/of/fdt.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c
index 544294a9ac..cf08fa1cfd 100644
--- a/drivers/of/fdt.c
+++ b/drivers/of/fdt.c
@@ -14,6 +14,7 @@
#include <memory.h>
#include <linux/sizes.h>
#include <linux/ctype.h>
+#include <linux/log2.h>
#include <linux/overflow.h>
#include <linux/string_helpers.h>
#include <linux/err.h>
@@ -400,6 +401,9 @@ static int fdt_ensure_space(struct fdt *fdt, int dtsize)
previous = fdt->dt;
new_size = fdt->dt_size * 2;
+ if (new_size <= dtsize)
+ new_size = roundup_pow_of_two(fdt->dt_size + dtsize);
+
fdt->dt = memalign_realloc(previous, fdt->dt_size, new_size);
if (!fdt->dt) {
free(previous);
--
2.39.2
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space
2024-02-01 11:28 [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 1/2] of: fdt: fix memory leak " Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 2/2] of: fdt: fix oob writes with large fdt properties Stefan Kerkmann
@ 2024-02-01 15:10 ` Sascha Hauer
2 siblings, 0 replies; 4+ messages in thread
From: Sascha Hauer @ 2024-02-01 15:10 UTC (permalink / raw)
To: BAREBOX, Stefan Kerkmann; +Cc: Ahmad Fatoum
On Thu, 01 Feb 2024 12:28:48 +0100, Stefan Kerkmann wrote:
> I have encountered the oob write while attempting to modify a large FIT
> image with of_property. While hunting for the root cause I noticed that
> there is a potential memory leak in fdt_ensure_space as well. Both is
> fixed in this series.
>
>
Applied, thanks!
[1/2] of: fdt: fix memory leak in fdt_ensure_space
https://git.pengutronix.de/cgit/barebox/commit/?id=dd36494ae201 (link may not be stable)
[2/2] of: fdt: fix oob writes with large fdt properties
https://git.pengutronix.de/cgit/barebox/commit/?id=96628249f26e (link may not be stable)
Best regards,
--
Sascha Hauer <s.hauer@pengutronix.de>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2024-02-01 15:10 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-02-01 11:28 [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 1/2] of: fdt: fix memory leak " Stefan Kerkmann
2024-02-01 11:28 ` [PATCH v2 2/2] of: fdt: fix oob writes with large fdt properties Stefan Kerkmann
2024-02-01 15:10 ` [PATCH v2 0/2] of: fdt: fix memory leak and oob writes in fdt_ensure_space Sascha Hauer
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox