Better handling of fds 0/1/2 #1233
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Discovered when mosh-server is started with a closed stderr, /dev/urandom is opened on fd 2, and the daemon code reopens fd 2 on /dev/null. In that case, mosh-server loops forever throwing a non-fatal exception.
Fixes unpleasant behavior in mosh-server when, say, stderr is closed.
|
@cgull would you mind rebasing this PR? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I ran into this while working on what is now #1232.
mosh-serverdoes not adequately handle the problem of standard file descriptors being closed. In my development, er, over two years ago, I ran into a situation wheremosh-serverhad started with fd 2 closed, it opened/dev/urandomon fd 2, and then later reopened fd 2 on/dev/null. Then, when it tried to actually read fd 2 to get entropy, it got 0 bytes instead, and got caught in a never-ending loop involving an exception being caught and the action being retried. I don't remember for sure, but I think that left the Mosh session hung inmosh.This work does 3 things:
CryptoExceptionthrown on failure to read fatal-- there's no point in continuing if we can't get a random seed./dev/nullon the closed fd,dup2()it to 0/1/2, and then close the temporary file. Oops.I can't now reproduce the hang, unfortunately. Both Mosh and my OSes have changed since hten. I tried to write a test for this issue, but between not being able to reproduce the problem and
mosh-serverbeing difficult to keep track of because it double-forks, I've given up for now. Should any of you doubt there's a problem,strace -o strace.log -f mosh-server new 2>&-will convince you fairly quickly. And while I didn't experience any actual crypto failure, even the possibility of confused crypto code is a rather bad smell.(Plus, we burned a couple of weeks chasing this problem down at $WORK once.)