Re: Re: nvme-tcp: fix a possible UAF when failing to send request【请注意，邮件由sagigrim@gmail.com代发】

["zhang.guanghui@cestc.cn" <zhang.guanghui@cestc.cn>]

Hi,



  in fact, the nvme_tcp_try_send() failure, the target may send C2HTermReq immediately. while the host receives the C2HTermReq and still starting error recovery. 



so when queue->rd_enabled is false,  can avoid starting error recovery agagin. Thanks





zhang.guanghui@cestc.cn



 



发件人： Hannes Reinecke



发送时间： 2025-03-13 15:51



收件人： zhang.guanghui@cestc.cn; Maurizio Lombardi; sagi; mgurtovoy; kbusch; sashal; chunguang.xu



抄送： linux-kernel; linux-nvme; linux-block



主题： Re: nvme-tcp: fix a possible UAF when failing to send request【请注意，邮件由sagigrim@gmail.com代发】



On 3/13/25 02:48, zhang.guanghui@cestc.cn wrote:



> Yes, the problem here is that,  despite the nvme_tcp_try_send() failure, the target sends a response capsule for the command, leading to a UAF in the host.



>



> Is it more reasonable to disable queue->rd_enabled to prevent receiving. Thanks



>  



> diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c



> index be04c5f3856d..17407eb12ad9 100644



> --- a/drivers/nvme/host/tcp.c



> +++ b/drivers/nvme/host/tcp.c



> @@ -1203,8 +1203,9 @@ static int nvme_tcp_try_send(struct nvme_tcp_queue *queue)



>          } else if (ret < 0) {



>                  dev_err(queue->ctrl->ctrl.device,



>                          "failed to send request %d\n", ret);



> -               nvme_tcp_fail_request(queue->request);



>                  nvme_tcp_done_send_req(queue);



> +              queue->rd_enabled = false;



> +              nvme_tcp_error_recovery(&queue->ctrl->ctrl);



>          }



>   out:



>          memalloc_noreclaim_restore(noreclaim_flag);



>



>



>



Hmm. In principle, yes. Problem is that network is a bi-directional



communication, and a failure on one side doesn't necessarily imply



a failure on the other.



In particular when the send side fails we should _continue_ to read



as we should be flushing the read side buffer before closing.



 



So I agree with starting error recovery, but not with disabling the



reading side (as we haven't encountered a read error).



 



Cheers,



 



Hannes



--



Dr. Hannes Reinecke                  Kernel Storage Architect



hare@suse.de                                +49 911 74053 688



SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg



HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich