Skip to contentSkip to conversation

Revert "i2c: fix stack buffer overflow vulnerability in i2c md command"

PATCH In review Marek Vasut 4 participants 7 messages Aug 26, 2022 – Aug 31, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 Revert "i2c: fix stack buffer overflow vulnerability in i2c md command"

This reverts commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409.

The commit is largely wrong and breaks most of i2c command functionality. The problem described in the aforementioned commit commit message is valid, however the commit itself does many more changes unrelated to fixing that one problem it describes. Those extra changes, namely the handling of i2c device address length as unsigned instead of signed integer, breaks the expectation that address length may be negative value. The negative value is used by DM to indicate that address length of device does not change.

The actual bug documented in commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 can be fixed by extra sanitization in separate patch.

Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>
---
 cmd/i2c.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
cmd/i2c.c +12 -12
index bd04b14024b..9050b2b8d27 100644
@@ -200,10 +200,10 @@ void i2c_init_board(void)
  *
  * Returns the address length.
  */
-static uint get_alen(char *arg, uint default_len)
+static uint get_alen(char *arg, int default_len)
 {
-	uint	j;
-	uint	alen;
+	int	j;
+	int	alen;
 
 	alen = default_len;
 	for (j = 0; j < 8; j++) {
@@ -247,7 +247,7 @@ static int do_i2c_read(struct cmd_tbl *cmdtp, int flag, int argc,
 {
 	uint	chip;
 	uint	devaddr, length;
-	uint	alen;
+	int alen;
 	u_char  *memaddr;
 	int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
@@ -301,7 +301,7 @@ static int do_i2c_write(struct cmd_tbl *cmdtp, int flag, int argc,
 {
 	uint	chip;
 	uint	devaddr, length;
-	uint	alen;
+	int alen;
 	u_char  *memaddr;
 	int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
@@ -469,8 +469,8 @@ static int do_i2c_md(struct cmd_tbl *cmdtp, int flag, int argc,
 {
 	uint	chip;
 	uint	addr, length;
-	uint	alen;
-	uint	j, nbytes, linebytes;
+	int alen;
+	int	j, nbytes, linebytes;
 	int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
 	struct udevice *dev;
@@ -589,9 +589,9 @@ static int do_i2c_mw(struct cmd_tbl *cmdtp, int flag, int argc,
 {
 	uint	chip;
 	ulong	addr;
-	uint	alen;
+	int	alen;
 	uchar	byte;
-	uint	count;
+	int	count;
 	int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
 	struct udevice *dev;
@@ -676,8 +676,8 @@ static int do_i2c_crc(struct cmd_tbl *cmdtp, int flag, int argc,
 {
 	uint	chip;
 	ulong	addr;
-	uint	alen;
-	uint	count;
+	int	alen;
+	int	count;
 	uchar	byte;
 	ulong	crc;
 	ulong	err;
@@ -985,7 +985,7 @@ static int do_i2c_loop(struct cmd_tbl *cmdtp, int flag, int argc,
 		       char *const argv[])
 {
 	uint	chip;
-	uint	alen;
+	int alen;
 	uint	addr;
 	uint	length;
 	u_char	bytes[16];
Signature · 1 line
2.35.1
#1 lore raw

PATCH 2/2 i2c: fix stack buffer overflow vulnerability in i2c md command

This reinstates fix from commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 without the changes unrelated to the actual fix. Avoid the underflow by setting only nbytes and linebytes as unsigned integers.

Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>
---
 cmd/i2c.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
cmd/i2c.c +2 -1
index 9050b2b8d27..e196a73efa6 100644
@@ -470,7 +470,8 @@ static int do_i2c_md(struct cmd_tbl *cmdtp, int flag, int argc,
 	uint	chip;
 	uint	addr, length;
 	int alen;
-	int	j, nbytes, linebytes;
+	int j;
+	uint nbytes, linebytes;
 	int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
 	struct udevice *dev;
Signature · 1 line
2.35.1
#2 lore raw

Re: [PATCH 2/2] i2c: fix stack buffer overflow vulnerability in i2c md command

Quoted 29 lines · Marek Vasut

This reinstates fix from commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 without the changes unrelated to the actual fix. Avoid the underflow by setting only nbytes and linebytes as unsigned integers.

Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>
---
 cmd/i2c.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
cmd/i2c.c +2 -2
index 9050b2b8d27..e196a73efa6 100644
@@ -470,7 +470,8 @@ static int do_i2c_md(struct cmd_tbl *cmdtp, int flag, int argc,
        uint    chip;
        uint    addr, length;
        int alen;
-       int     j, nbytes, linebytes;
+       int j;
+       uint nbytes, linebytes;
        int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
        struct udevice *dev;
--
Quoted 2 lines

2.35.1

Marek,

Did 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 get reverted then?

Best Regards,

Tim

Acked-by
#3 lore raw

Re: [PATCH 2/2] i2c: fix stack buffer overflow vulnerability in i2c md command

Quoted 36 lines · Tim Harvey
Quoted 29 lines · Marek Vasut

This reinstates fix from commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 without the changes unrelated to the actual fix. Avoid the underflow by setting only nbytes and linebytes as unsigned integers.

Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>
---
 cmd/i2c.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
cmd/i2c.c +2 -2
index 9050b2b8d27..e196a73efa6 100644
@@ -470,7 +470,8 @@ static int do_i2c_md(struct cmd_tbl *cmdtp, int flag, int argc,
        uint    chip;
        uint    addr, length;
        int alen;
-       int     j, nbytes, linebytes;
+       int j;
+       uint nbytes, linebytes;
        int ret;
 #if CONFIG_IS_ENABLED(DM_I2C)
        struct udevice *dev;
--
Quoted 2 lines

2.35.1

Marek,

Quoted 2 lines

Did 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 get reverted then?

Oh gosh... my bad for not seeing your first patch that reverts it!

Best Regards,

Tim

#4 lore raw

Re: [PATCH 2/2] i2c: fix stack buffer overflow vulnerability in i2c md command

In reply to Marek Vasut
Quoted 10 lines · Marek Vasut

This reinstates fix from commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 without the changes unrelated to the actual fix. Avoid the underflow by setting only nbytes and linebytes as unsigned integers.

Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>

Applied to u-boot/master, thanks!

Signature · 1 line
Tom
Reviewed-by on v1 1/2
#5 lore raw
In reply to Marek Vasut
Quoted 14 lines · Marek Vasut

This reverts commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409.

The commit is largely wrong and breaks most of i2c command functionality. The problem described in the aforementioned commit commit message is valid, however the commit itself does many more changes unrelated to fixing that one problem it describes. Those extra changes, namely the handling of i2c device address length as unsigned instead of signed integer, breaks the expectation that address length may be negative value. The negative value is used by DM to indicate that address length of device does not change.

The actual bug documented in commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 can be fixed by extra sanitization in separate patch.

Quoted 8 lines
Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>
---
 cmd/i2c.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)
#6 lore raw
In reply to Marek Vasut
Quoted 13 lines · Marek Vasut

This reverts commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409.

The commit is largely wrong and breaks most of i2c command functionality. The problem described in the aforementioned commit commit message is valid, however the commit itself does many more changes unrelated to fixing that one problem it describes. Those extra changes, namely the handling of i2c device address length as unsigned instead of signed integer, breaks the expectation that address length may be negative value. The negative value is used by DM to indicate that address length of device does not change.

The actual bug documented in commit 8f8c04bf1ebbd2f72f1643e7ad9617dafa6e5409 can be fixed by extra sanitization in separate patch.

Quoted 6 lines
Cc Heiko Schocher <[email protected]>
Cc Nicolas Iooss <[email protected]>
Cc Simon Glass <[email protected]>
Cc Tim Harvey <[email protected]>

Applied to u-boot/master, thanks!

Signature · 1 line
Tom