v2: patch the correct write helper and add bounds-checking; v1 mistakenly guarded e1000_get_eeprom() (read path).
---
Mikael Wessel (1): e1000e: fix heap overflow in e1000_set_eeprom()
drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
The ETHTOOL_SETEEPROM ioctl copies user data into a kmalloc'ed buffer without validating eeprom->len and eeprom->offset. A CAP_NET_ADMIN user can overflow the heap and crash the kernel or gain code execution.
Validate length and offset before memcpy().
Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)") Reported-by: Mikael Wessel post@mikaelkw.online Signed-off-by: Mikael Wessel post@mikaelkw.online Cc: stable@vger.kernel.org --- drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c index 9364bc2b4eb1..98e541e39730 100644 --- a/drivers/net/ethernet/intel/e1000e/ethtool.c +++ b/drivers/net/ethernet/intel/e1000e/ethtool.c @@ -596,6 +596,9 @@ static int e1000_set_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) le16_to_cpus(&eeprom_buff[i]);
+ if (eeprom->len > max_len || + eeprom->offset > max_len - eeprom->len) + return -EINVAL; memcpy(ptr, bytes, eeprom->len);
for (i = 0; i < last_word - first_word + 1; i++)
-----Original Message----- From: Intel-wired-lan intel-wired-lan-bounces@osuosl.org On Behalf Of Mikael Wessel Sent: Tuesday, May 27, 2025 10:56 AM To: netdev@vger.kernel.org Cc: intel-wired-lan@lists.osuosl.org; torvalds@linuxfoundation.org; Nguyen, Anthony L anthony.l.nguyen@intel.com; Kitszel, Przemyslaw przemyslaw.kitszel@intel.com; andrew@lunn.ch; kuba@kernel.org; pabeni@redhat.com; security@kernel.org; stable@vger.kernel.org; davem@davemloft.net; edumazet@google.com; linux- kernel@vger.kernel.org; Mikael Wessel post@mikaelkw.online Subject: [Intel-wired-lan] [PATCH v2 1/1] e1000e: fix heap overflow in e1000_set_eeprom()
The ETHTOOL_SETEEPROM ioctl copies user data into a kmalloc'ed buffer without validating eeprom->len and eeprom->offset. A CAP_NET_ADMIN user can overflow the heap and crash the kernel or gain code execution.
Validate length and offset before memcpy().
Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)") Reported-by: Mikael Wessel post@mikaelkw.online Signed-off-by: Mikael Wessel post@mikaelkw.online
Reviewed-by: Aleksandr Loktionov aleksandr.loktionov@intel.com
Cc: stable@vger.kernel.org
drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c index 9364bc2b4eb1..98e541e39730 100644 --- a/drivers/net/ethernet/intel/e1000e/ethtool.c +++ b/drivers/net/ethernet/intel/e1000e/ethtool.c @@ -596,6 +596,9 @@ static int e1000_set_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) le16_to_cpus(&eeprom_buff[i]);
if (eeprom->len > max_len ||eeprom->offset > max_len - eeprom->len)return -EINVAL;memcpy(ptr, bytes, eeprom->len);
for (i = 0; i < last_word - first_word + 1; i++)
-- 2.48.1
Dear Mikael,
Thank you for your patch.
Am 27.05.25 um 10:56 schrieb Mikael Wessel:
The ETHTOOL_SETEEPROM ioctl copies user data into a kmalloc'ed buffer without validating eeprom->len and eeprom->offset. A CAP_NET_ADMIN user can overflow the heap and crash the kernel or gain code execution.
Validate length and offset before memcpy().
Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)") Reported-by: Mikael Wessel post@mikaelkw.online Signed-off-by: Mikael Wessel post@mikaelkw.online Cc: stable@vger.kernel.org
drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c index 9364bc2b4eb1..98e541e39730 100644 --- a/drivers/net/ethernet/intel/e1000e/ethtool.c +++ b/drivers/net/ethernet/intel/e1000e/ethtool.c @@ -596,6 +596,9 @@ static int e1000_set_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) le16_to_cpus(&eeprom_buff[i]);
if (eeprom->len > max_len ||eeprom->offset > max_len - eeprom->len)return -EINVAL;
I think you used spaces instead of tabs for indentation. It’d be great if you could fix this, and send v3 tomorrow. Running `scripts/checkpatch.pl` with the patch as an argument, should catch these things.
memcpy(ptr, bytes, eeprom->len); for (i = 0; i < last_word - first_word + 1; i++)
Kind regards,
Paul
[one addition]
Am 27.05.25 um 16:43 schrieb Paul Menzel:
Dear Mikael,
Thank you for your patch.
Am 27.05.25 um 10:56 schrieb Mikael Wessel:
The ETHTOOL_SETEEPROM ioctl copies user data into a kmalloc'ed buffer without validating eeprom->len and eeprom->offset. A CAP_NET_ADMIN user can overflow the heap and crash the kernel or gain code execution.
Validate length and offset before memcpy().
Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)") Reported-by: Mikael Wessel post@mikaelkw.online Signed-off-by: Mikael Wessel post@mikaelkw.online Cc: stable@vger.kernel.org
drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/ net/ethernet/intel/e1000e/ethtool.c index 9364bc2b4eb1..98e541e39730 100644 --- a/drivers/net/ethernet/intel/e1000e/ethtool.c +++ b/drivers/net/ethernet/intel/e1000e/ethtool.c @@ -596,6 +596,9 @@ static int e1000_set_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) le16_to_cpus(&eeprom_buff[i]); + if (eeprom->len > max_len || + eeprom->offset > max_len - eeprom->len) + return -EINVAL;
I think you used spaces instead of tabs for indentation. It’d be great if you could fix this, and send v3 tomorrow. Running `scripts/ checkpatch.pl` with the patch as an argument, should catch these things.
Should a warning/error be logged if the condition is true?
memcpy(ptr, bytes, eeprom->len); for (i = 0; i < last_word - first_word + 1; i++)
Kind regards,
Paul
On 5/27/2025 1:56 AM, Mikael Wessel wrote:
The ETHTOOL_SETEEPROM ioctl copies user data into a kmalloc'ed buffer without validating eeprom->len and eeprom->offset. A CAP_NET_ADMIN user can overflow the heap and crash the kernel or gain code execution.
Validate length and offset before memcpy().
Fixes: bc7f75fa9788 ("[E1000E]: New pci-express e1000 driver (currently for ICH9 devices only)") Reported-by: Mikael Wessel post@mikaelkw.online Signed-off-by: Mikael Wessel post@mikaelkw.online Cc: stable@vger.kernel.org
drivers/net/ethernet/intel/e1000e/ethtool.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/intel/e1000e/ethtool.c b/drivers/net/ethernet/intel/e1000e/ethtool.c index 9364bc2b4eb1..98e541e39730 100644 --- a/drivers/net/ethernet/intel/e1000e/ethtool.c +++ b/drivers/net/ethernet/intel/e1000e/ethtool.c @@ -596,6 +596,9 @@ static int e1000_set_eeprom(struct net_device *netdev, for (i = 0; i < last_word - first_word + 1; i++) le16_to_cpus(&eeprom_buff[i]);
if (eeprom->len > max_len ||eeprom->offset > max_len - eeprom->len)return -EINVAL;
This is going to cause 'eeprom_buff' to leak. You should use the goto out, however, seems like these checks can be moved up before the allocation is done. Also, indentation looks off.
Thanks, Tony
memcpy(ptr, bytes, eeprom->len); for (i = 0; i < last_word - first_word + 1; i++)
linux-stable-mirror@lists.linaro.org
-
Loktionov, Aleksandr -
Mikael Wessel -
Paul Menzel -
Tony Nguyen