Re-enable util/mmap_anon_no_fork

The test above was disabled with #29534 (moved) in 0.4.0. We should re-enable it and make it pass everywhere. I think I know what the problem is: I think it's a matter of some flags existing in the headers but not in the kernel.

changed milestone to %Tor: 0.4.1.x-final

Trac:
Parent Ticket: #26288 (moved)

added 041-proposed-on-roadmap asn-merge component::core tor/tor milestone::Tor: 0.4.1.x-final network-team-roadmap-2019-Q1Q2 owner::nickm parent::26288 priority::medium prop289 prop289-assigned-sponsor-v resolution::fixed reviewer::mikeperry severity::normal status::closed type::defect labels

I have a patch at #29541 (moved) to fix the test and re-enable it. It works for me on Fedora 29 and on OSX 10.14.3. I've made a PR at https://github.com/torproject/tor/pull/723 -- let's see if the CI passes.

Trac:
Status: assigned to needs_review
Cc: N/A to teor

Trac:
Reviewer: N/A to mikeperry

I tested commit:

commit 5614960e94 (HEAD, tor-github/pr/723/merge)
Merge: 69238ca2da 065e7da8e6

util/mmap_anon_no_fork passes on my machine.

But #29585 (moved) and #29586 (moved) occurred once in about 10 test runs. They happened in the same test run. Could this change have affected the reliability of these other tests?

I really don't see how this test bug could have affected those tests... but the world has been known to do things that I do not expect. If they happen one time in ten, maybe checking to see whether this fix fixes them too might be reasonable.

I have no clue how this relates to prop289 (#26288 (moved)). Feel free to re-parent but explains why before?

Trac:
Keywords: prop289, prop289-assigned-sponsor-v deleted, N/A added
Parent: #26288 (moved) to N/A

This is prop289 because it's a bug in the no-fork mmap tests; I added the no-fork mmap to make the fast PRNG more fork-safe.

Restoring prop289

Trac:
Keywords: N/A deleted, prop289, prop289-assigned-sponsor-v added
Parent: N/A to #26288 (moved)

Replying to nickm:

I really don't see how this test bug could have affected those tests... but the world has been known to do things that I do not expect. If they happen one time in ten, maybe checking to see whether this fix fixes them too might be reasonable.

To be clear, #29585 (moved) and #29586 (moved) happened once. And then they haven't happened again. I think they may have happened due to a clock change during the test run. If they happen again, we can reopen those tickets.

This is a little mind-bendy with the magic behavior constants and bytes being sent back and forth, but it's well commented and looks ok.

Trac:
Status: needs_review to merge_ready

Trac:
Keywords: N/A deleted, asn-merge added

Merged. There was no changes file but that's because the bug was not released yet.

Trac:
Resolution: N/A to fixed
Status: merge_ready to closed

closed

mentioned in issue #29585 (moved)

mentioned in issue #29586 (moved)

moved to tpo/core/tor#29541 (closed)

mentioned in issue tpo/core/tor#29585 (closed)

mentioned in issue tpo/core/tor#29586 (closed)