netlink: Return unsigned value for nla_len()
Subsystems (2)
PATCH v2 1/1 netlink: Return unsigned value for nla_len()
The return value from nla_len() is never expected to be negative, and can never be more than struct nlattr::nla_len (a u16). Adjust the prototype on the function. This will let GCC's value range optimization passes know that the return can never be negative, and can never be larger than u16. As recently discussed[1], this silences the following warning in GCC 12+:
net/wireless/nl80211.c: In function 'nl80211_set_cqm_rssi.isra':net/wireless/nl80211.c:12892:17: warning: 'memcpy' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overflow=]
12892 | memcpy(cqm_config->rssi_thresholds, thresholds,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12893 | flex_array_size(cqm_config, rssi_thresholds,
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
12894 | n_thresholds));
| ~~~~~~~~~~~~~~A future change would be to clamp the subtraction to make sure it never wraps around if nla_len is somehow less than NLA_HDRLEN, which would have the additional benefit of being defensive in the face of nlattr corruption or logic errors.
---
v2:
- do not clamp return value (kuba)
- adjust NLA_HDRLEN to be u16 also
v1: https://lore.kernel.org/all/[email protected]/
--- include/net/netlink.h | 2 +- include/uapi/linux/netlink.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
include/net/netlink.h +1 -1
@@ -1200,7 +1200,7 @@ static inline void *nla_data(const struct nlattr *nla)* nla_len - length of payload * @nla: netlink attribute */ -static inline int nla_len (const struct nlattr *nla) +static inline u16 nla_len (const struct nlattr *nla) {return nla->nla_len -NLA_HDRLEN ; }
include/uapi/linux/netlink.h +1 -1
@@ -247,7 +247,7 @@ struct nlattr {#define NLA_ALIGNTO 4#define NLA_ALIGN (len) (((len) + NLA_ALIGNTO - 1) & ~(NLA_ALIGNTO - 1)) -#define NLA_HDRLEN ((int) NLA_ALIGN(sizeof(struct nlattr))) +#define NLA_HDRLEN ((__u16) NLA_ALIGN(sizeof(struct nlattr)))/* Generic 32 bitflags attribute content sent to the kernel. *
Signature · 1 line
2.34.1
Le 02/12/2023 à 21:25, Kees Cook a écrit :
Quoted 62 lines
The return value from
nla_len()is never expected to be negative, and can never be more than structnlattr::nla_len(a u16). Adjust the prototype on the function. This will let GCC's value range optimization passes know that the return can never be negative, and can never be larger than u16. As recently discussed[1], this silences the following warning in GCC 12+:
net/wireless/nl80211.c: In function 'nl80211_set_cqm_rssi.isra':net/wireless/nl80211.c:12892:17: warning: 'memcpy' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overflow=]12892 | memcpy(cqm_config->rssi_thresholds, thresholds, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 12893 | flex_array_size(cqm_config, rssi_thresholds, | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 12894 | n_thresholds)); | ~~~~~~~~~~~~~~A future change would be to clamp the subtraction to make sure it never wraps around if
nla_lenis somehow less thanNLA_HDRLEN, which would have the additional benefit of being defensive in the face of nlattr corruption or logic errors.Reported-by kernel test robot <[email protected]>Cc Jakub Kicinski <[email protected]>Cc "David S. Miller" <[email protected]>Cc Eric Dumazet <[email protected]>Cc Paolo Abeni <[email protected]>Cc Johannes Berg <[email protected]>Cc Jeff Johnson <[email protected]>Cc Michael Walle <[email protected]>Cc Max Schulze <[email protected]>Signed-off-by Kees Cook <[email protected]>---
v2:
- do not clamp return value (kuba)
- adjustNLA_HDRLENto be u16 also
v1: https://lore.kernel.org/all/[email protected]/--- include/net/netlink.h | 2 +- include/uapi/linux/netlink.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)include/net/netlink.h +1
-1@@ -1200,7 +1200,7 @@ static inline void *nla_data(const struct nlattr *nla)* nla_len - length of payload * @nla: netlink attribute */ -static inline int nla_len (const struct nlattr *nla) +static inline u16 nla_len (const struct nlattr *nla) {return nla->nla_len -NLA_HDRLEN ; }include/uapi/linux/netlink.h +0
-0
include/uapi/linux/netlink.h struct nlattr { · line 250
Quoted 5 lines
include/uapi/linux/netlink.h +1
-1@@ -247,7 +247,7 @@ struct nlattr {#define NLA_ALIGNTO 4#define NLA_ALIGN (len) (((len) + NLA_ALIGNTO - 1) & ~(NLA_ALIGNTO - 1)) -#define NLA_HDRLEN ((int) NLA_ALIGN(sizeof(struct nlattr))) +#define NLA_HDRLEN ((__u16) NLA_ALIGN(sizeof(struct nlattr)))
I wonder if this may break the compilation of some userspace tools with errors like comparing signed and unsigned values.
Regards,
Nicolas
Quoted 64 lines · Nicolas Dichtel
Le 02/12/2023 à 21:25, Kees Cook a écrit :
Quoted 63 lines
The return value from
nla_len()is never expected to be negative, and can never be more than structnlattr::nla_len(a u16). Adjust the prototype on the function. This will let GCC's value range optimization passes know that the return can never be negative, and can never be larger than u16. As recently discussed[1], this silences the following warning in GCC 12+:
net/wireless/nl80211.c: In function 'nl80211_set_cqm_rssi.isra':net/wireless/nl80211.c:12892:17: warning: 'memcpy' specified bound 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Wstringop-overflow=]12892 | memcpy(cqm_config->rssi_thresholds, thresholds, | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 12893 | flex_array_size(cqm_config, rssi_thresholds, | ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 12894 | n_thresholds)); | ~~~~~~~~~~~~~~A future change would be to clamp the subtraction to make sure it never wraps around if
nla_lenis somehow less thanNLA_HDRLEN, which would have the additional benefit of being defensive in the face of nlattr corruption or logic errors.Reported-by kernel test robot <[email protected]>Cc Jakub Kicinski <[email protected]>Cc "David S. Miller" <[email protected]>Cc Eric Dumazet <[email protected]>Cc Paolo Abeni <[email protected]>Cc Johannes Berg <[email protected]>Cc Jeff Johnson <[email protected]>Cc Michael Walle <[email protected]>Cc Max Schulze <[email protected]>Signed-off-by Kees Cook <[email protected]>---
v2:
- do not clamp return value (kuba)
- adjustNLA_HDRLENto be u16 also
v1: https://lore.kernel.org/all/[email protected]/--- include/net/netlink.h | 2 +- include/uapi/linux/netlink.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)include/net/netlink.h +1
-1@@ -1200,7 +1200,7 @@ static inline void *nla_data(const struct nlattr *nla)* nla_len - length of payload * @nla: netlink attribute */ -static inline int nla_len (const struct nlattr *nla) +static inline u16 nla_len (const struct nlattr *nla) {return nla->nla_len -NLA_HDRLEN ; }include/uapi/linux/netlink.h +0
-0@@ -247,7 +247,7 @@ struct nlattr {
Quoted 6 lines
include/uapi/linux/netlink.h struct nlattr { · line 250
Quoted 4 lines
include/uapi/linux/netlink.h +1
-1#define NLA_ALIGNTO 4#define NLA_ALIGN (len) (((len) + NLA_ALIGNTO - 1) & ~(NLA_ALIGNTO - 1)) -#define NLA_HDRLEN ((int) NLA_ALIGN(sizeof(struct nlattr))) +#define NLA_HDRLEN ((__u16) NLA_ALIGN(sizeof(struct nlattr)))I wonder if this may break the compilation of some userspace tools with errors like comparing signed and unsigned values.
Should I drop this part, then?
Signature · 1 line
Kees Cook
Le 04/12/2023 à 23:21, Kees Cook a écrit :
[snip]
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines
include/uapi/linux/netlink.h +0
-0@@ -247,7 +247,7 @@ struct nlattr {
Quoted 9 lines
Quoted 6 lines
include/uapi/linux/netlink.h struct nlattr { · line 250
Quoted 4 lines
include/uapi/linux/netlink.h +1
-1#define NLA_ALIGNTO 4#define NLA_ALIGN (len) (((len) + NLA_ALIGNTO - 1) & ~(NLA_ALIGNTO - 1)) -#define NLA_HDRLEN ((int) NLA_ALIGN(sizeof(struct nlattr))) +#define NLA_HDRLEN ((__u16) NLA_ALIGN(sizeof(struct nlattr)))I wonder if this may break the compilation of some userspace tools with errors like comparing signed and unsigned values.
Should I drop this part, then?
Yes please.
Thank you,
Nicolas