libsa: Fix IP recv timeout

[kevans]

Summary

readip() doesn't, at the moment, properly indicate to callers that it's timed out. One can tell that it's timed out if errno == EAGAIN when it returns, but this is not ideal. Restructure it a little bit to explicitly set errno to ETIMEDOUT if we've exhausted tleft.

I found two places that care about where it timed out or not: sendrecv in net.c and sendrecv_tftp. Both are structured to pass smaller timeout values to readip while tracking a larger timeout. Neither of them were able to do this properly with readip not indicating ETIMEDOUT, so fix it.

While here, straighten out the time (t/t1) usage in sendrecv_tftp.

[kevans]

Found an unrelated bug in the tftp section; if a send fails, we're supposed to wait before retrying. However, 't1' doesn't get reset, so if it times out a second time (and still before we reached MAXTMO) there won't be any further waiting before resending. because getsecs() - t1 by definition already passed after the first timeout period.

[sbruno]

Ugh, phab seems to have not take the diffs from the root of the src tree. I'll munge a bit, but can you regenerate?

[kevans]

Whoops, my bad. =)

[sbruno]

BINGO

[kevans]

For what it's worth, I would almost like to do something more like this: https://people.freebsd.org/~kevans/stand-ip.diff

It's a lot more invasive, but it gets rid sendrecv_tftp and consolidates MINTMO/MAXTMO timeout logic into one spot that everything honors.

[tsoome]

In D14350#301042, @kevans wrote:

For what it's worth, I would almost like to do something more like this: https://people.freebsd.org/~kevans/stand-ip.diff

It's a lot more invasive, but it gets rid sendrecv_tftp and consolidates MINTMO/MAXTMO timeout logic into one spot that everything honors.

take it as second round.