net/mlx5: add Netlink message size check in rule cleanup
authorViacheslav Ovsiienko <viacheslavo@mellanox.com>
Sat, 10 Nov 2018 09:59:26 +0000 (09:59 +0000)
committerFerruh Yigit <ferruh.yigit@intel.com>
Tue, 13 Nov 2018 23:35:53 +0000 (00:35 +0100)
This patch is preparation for the following fix, we are going to send
Netlink message from buffer in one-by-one fashion. It is highly
desirable to check multimessage buffer consistency for debug purposes.

Signed-off-by: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
Acked-by: Shahaf Shuler <shahafs@mellanox.com>
Acked-by: Yongseok Koh <yskoh@mellanox.com>
drivers/net/mlx5/mlx5_flow_tcf.c

index ba0674a..bba8aed 100644 (file)
@@ -3935,6 +3935,7 @@ flow_tcf_collect_local_cb(const struct nlmsghdr *nlh, void *arg)
        struct nlattr *na_local = NULL;
        struct nlattr *na_peer = NULL;
        unsigned char family;
+       uint32_t size;
 
        if (nlh->nlmsg_type != RTM_NEWADDR) {
                rte_errno = EINVAL;
@@ -3962,11 +3963,11 @@ flow_tcf_collect_local_cb(const struct nlmsghdr *nlh, void *arg)
        if (!na_local || !na_peer)
                return 1;
        /* Local rule found with scope link, permanent and assigned peer. */
-       cmd = flow_tcf_alloc_nlcmd(ctx, MNL_ALIGN(sizeof(struct nlmsghdr)) +
-                                       MNL_ALIGN(sizeof(struct ifaddrmsg)) +
-                                       (family == AF_INET6
-                                       ? 2 * SZ_NLATTR_DATA_OF(IPV6_ADDR_LEN)
-                                       : 2 * SZ_NLATTR_TYPE_OF(uint32_t)));
+       size = MNL_ALIGN(sizeof(struct nlmsghdr)) +
+              MNL_ALIGN(sizeof(struct ifaddrmsg)) +
+              (family == AF_INET6 ? 2 * SZ_NLATTR_DATA_OF(IPV6_ADDR_LEN)
+                                  : 2 * SZ_NLATTR_TYPE_OF(uint32_t));
+       cmd = flow_tcf_alloc_nlcmd(ctx, size);
        if (!cmd) {
                rte_errno = ENOMEM;
                return -rte_errno;
@@ -3991,6 +3992,7 @@ flow_tcf_collect_local_cb(const struct nlmsghdr *nlh, void *arg)
                mnl_attr_put(cmd, IFA_ADDRESS, IPV6_ADDR_LEN,
                        mnl_attr_get_payload(na_peer));
        }
+       assert(size == cmd->nlmsg_len);
        return 1;
 }
 
@@ -4059,6 +4061,7 @@ flow_tcf_collect_neigh_cb(const struct nlmsghdr *nlh, void *arg)
        struct nlattr *na_ip = NULL;
        struct nlattr *na_mac = NULL;
        unsigned char family;
+       uint32_t size;
 
        if (nlh->nlmsg_type != RTM_NEWNEIGH) {
                rte_errno = EINVAL;
@@ -4085,12 +4088,12 @@ flow_tcf_collect_neigh_cb(const struct nlmsghdr *nlh, void *arg)
        if (!na_mac || !na_ip)
                return 1;
        /* Neigh rule with permenent attribute found. */
-       cmd = flow_tcf_alloc_nlcmd(ctx, MNL_ALIGN(sizeof(struct nlmsghdr)) +
-                                       MNL_ALIGN(sizeof(struct ndmsg)) +
-                                       SZ_NLATTR_DATA_OF(ETHER_ADDR_LEN) +
-                                       (family == AF_INET6
-                                       ? SZ_NLATTR_DATA_OF(IPV6_ADDR_LEN)
-                                       : SZ_NLATTR_TYPE_OF(uint32_t)));
+       size = MNL_ALIGN(sizeof(struct nlmsghdr)) +
+              MNL_ALIGN(sizeof(struct ndmsg)) +
+              SZ_NLATTR_DATA_OF(ETHER_ADDR_LEN) +
+              (family == AF_INET6 ? SZ_NLATTR_DATA_OF(IPV6_ADDR_LEN)
+                                  : SZ_NLATTR_TYPE_OF(uint32_t));
+       cmd = flow_tcf_alloc_nlcmd(ctx, size);
        if (!cmd) {
                rte_errno = ENOMEM;
                return -rte_errno;
@@ -4113,6 +4116,7 @@ flow_tcf_collect_neigh_cb(const struct nlmsghdr *nlh, void *arg)
        }
        mnl_attr_put(cmd, NDA_LLADDR, ETHER_ADDR_LEN,
                     mnl_attr_get_payload(na_mac));
+       assert(size == cmd->nlmsg_len);
        return 1;
 }
 
@@ -4179,6 +4183,7 @@ flow_tcf_collect_vxlan_cb(const struct nlmsghdr *nlh, void *arg)
        struct nlattr *na_vxlan = NULL;
        bool found = false;
        unsigned int vxindex;
+       uint32_t size;
 
        if (nlh->nlmsg_type != RTM_NEWLINK) {
                rte_errno = EINVAL;
@@ -4224,8 +4229,9 @@ flow_tcf_collect_vxlan_cb(const struct nlmsghdr *nlh, void *arg)
                return 1;
        /* Attached VXLAN device found, store the command to delete. */
        vxindex = ifm->ifi_index;
-       cmd = flow_tcf_alloc_nlcmd(ctx, MNL_ALIGN(sizeof(struct nlmsghdr)) +
-                                       MNL_ALIGN(sizeof(struct ifinfomsg)));
+       size = MNL_ALIGN(sizeof(struct nlmsghdr)) +
+              MNL_ALIGN(sizeof(struct ifinfomsg));
+       cmd = flow_tcf_alloc_nlcmd(ctx, size);
        if (!cmd) {
                rte_errno = ENOMEM;
                return -rte_errno;
@@ -4236,6 +4242,7 @@ flow_tcf_collect_vxlan_cb(const struct nlmsghdr *nlh, void *arg)
        ifm = mnl_nlmsg_put_extra_header(cmd, sizeof(*ifm));
        ifm->ifi_family = AF_UNSPEC;
        ifm->ifi_index = vxindex;
+       assert(size == cmd->nlmsg_len);
        return 1;
 }