Re: [PATCH bpf-next v2 3/4] bpf, sockmap: Fix more uncharged while msg has more_data

wangyufen <wangyufen@xxxxxxxxxx> · Fri, 4 Mar 2022 15:08:35 +0800

在 2022/3/3 14:14, Cong Wang 写道:
On Wed, Mar 02, 2022 at 10:27:54AM +0800, Wang Yufen wrote:
In tcp_bpf_send_verdict(), if msg has more data after
tcp_bpf_sendmsg_redir():

tcp_bpf_send_verdict()
  tosend = msg->sg.size  //msg->sg.size = 22220
  case __SK_REDIRECT:
   sk_msg_return()  //uncharged msg->sg.size(22220) sk->sk_forward_alloc
   tcp_bpf_sendmsg_redir() //after tcp_bpf_sendmsg_redir, msg->sg.size=11000
  goto more_data;
  tosend = msg->sg.size  //msg->sg.size = 11000
  case __SK_REDIRECT:
   sk_msg_return()  //uncharged msg->sg.size(11000) to sk->sk_forward_alloc

The msg->sg.size(11000) has been uncharged twice, to fix we can charge the
remaining msg->sg.size before goto more data.
It looks like bpf_exec_tx_verdict() has the same issue.

.

In bpf_exec_tx_verdict(), case __SK_REDIRECT,  msg_redir is used and 
msg->sg.size is deducted in advance.

Therefore, this issue (more uncharged) does not exist.

However, I think that if msg_redir processing cannot be completed , that 
is msg_redir has more data,

and there is no subsequent processing,  maybe that is another problem.

Thanks.