nvmet-tcp: Fix NULL dereference when a connect data comes in h2cdata pdu
authorZiye Yang <ziye.yang@intel.com>
Fri, 21 Aug 2020 16:48:10 +0000 (00:48 +0800)
committerSagi Grimberg <sagi@grimberg.me>
Fri, 28 Aug 2020 23:43:56 +0000 (16:43 -0700)
When handling commands without in-capsule data, we assign the ttag
assuming we already have the queue commands array allocated (based
on the queue size information in the connect data payload). However
if the connect itself did not send the connect data in-capsule we
have yet to allocate the queue commands,and we will assign a bogus
ttag and suffer a NULL dereference when we receive the corresponding
h2cdata pdu.

Fix this by checking if we already allocated commands before
dereferencing it when handling h2cdata, if we didn't, its for sure a
connect and we should use the preallocated connect command.

Signed-off-by: Ziye Yang <ziye.yang@intel.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
drivers/nvme/target/tcp.c

index 9eda911..8e0d766 100644 (file)
@@ -160,6 +160,11 @@ static void nvmet_tcp_finish_cmd(struct nvmet_tcp_cmd *cmd);
 static inline u16 nvmet_tcp_cmd_tag(struct nvmet_tcp_queue *queue,
                struct nvmet_tcp_cmd *cmd)
 {
+       if (unlikely(!queue->nr_cmds)) {
+               /* We didn't allocate cmds yet, send 0xffff */
+               return USHRT_MAX;
+       }
+
        return cmd - queue->cmds;
 }
 
@@ -866,7 +871,10 @@ static int nvmet_tcp_handle_h2c_data_pdu(struct nvmet_tcp_queue *queue)
        struct nvme_tcp_data_pdu *data = &queue->pdu.data;
        struct nvmet_tcp_cmd *cmd;
 
-       cmd = &queue->cmds[data->ttag];
+       if (likely(queue->nr_cmds))
+               cmd = &queue->cmds[data->ttag];
+       else
+               cmd = &queue->connect;
 
        if (le32_to_cpu(data->data_offset) != cmd->rbytes_done) {
                pr_err("ttag %u unexpected data offset %u (expected %u)\n",