wg-socks implant crash

[pooki3bear]

Describe the bug While attempting to use a (probably misconfigured) wg-socks session, the WG implant became entirely unresponsive, and eventually session lost.

During the period before the session was lost, executing wg-socks from sliver-client made the client unresponsive.

Expected behavior Expect wg-socks output to have a reasonable timeout, and provide some error instead of very slow timeout.

[moloch--]

Let us know if you find a reliable way to reproduce the issue. I suspect this is due to a blocking read() call or something.

[rkervella]

Yeah pretty sure it's the same bug that happens on the other cases, the server reading more bytes than expected and thus blocking.

[moloch--]

I think* this should be fixed in the stage branch.

[rkervella]

Hm doesn't seem so. shell still doesn't work, and the behavior seems similar.