Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

next: Changes queued for the 2017.04 release #1120

Merged
merged 2,278 commits into from
Apr 10, 2017
Merged

next: Changes queued for the 2017.04 release #1120

merged 2,278 commits into from
Apr 10, 2017

Conversation

lukego
Copy link
Member

@lukego lukego commented Apr 4, 2017

Currently queued changes for 2017.04. Let's see what Hydra says.

Note: Many commits both due to importing LuaJIT history and also merging the Snabbwall branch.

aperezdc and others added 30 commits April 25, 2016 19:23
The assertion is done in make_cdata_hash_function(), so it is not needed to do
it before calling it. This removes a couple of local variables as well.
This makes the hash functions generated by make_cdata_hash_function()
use the intermediate value calculated for the previous 32-bit chunks
of the input combined with the output for the current chunk. Without
this fix, Changes at the beginning of the input would not make the
final hash result vary, which was incorrect behavior.
Contrary to the version in lib.ctable, the hash32() function here does not
need to clear the last bit of the result, because a "guard" value is unneeded
for flow keys.
Contributed by Peter Cawley.
Prerequisite for immovable IR.
Contributed by Peter Cawley.
Contributed by Peter Cawley.
This allows embedding pointers to IR constants in the machine code.
Contributed by Peter Cawley.
This is now possible due to the immovable IR.
Contributed by Peter Cawley.
Prerequisite for embedding 64 bit constants directly in the IR.
Contributed by Peter Cawley.
Contributed by Peter Cawley.
lukego and others added 21 commits March 9, 2017 09:58
…pgrade

This commit affects only the git history and not the contents of the code.

The unsquashed LuaJIT commit history is merged into our lib/luajit
subtree. This should make it easier for us to pull in changes to
LuaJIT from multiple branches, because git subtree will know which
commits we have and which ones we don't.

The code is left unchanged. There were some conflicts where we had
changed something compared with LuaJIT upstream but I resolved all of
those in favor of keeping our code the same. (If we actually don't
want changes then we can push reverts later.)
Usually Snabbwall has tracked release versions of Snabb.
I'm merging in the `next` branch instead now in order to
make it easier to upstream Snabbwall without merge
conflicts in the pull request.

Conflicts:
	lib/pflua/src/pf/anf.lua
	lib/pflua/src/pf/match.lua
One of the main motivations of ljndpi is to be used as part of
SnabbWall, a suite of Layer 7 firewall components for Snabb, which
itself is Apache-licensed. Using the same license as Snabb makes it
simpler to incorporate as part of it.
Added a note about the Slack chat & refactored the text a little.
@eugeneia
Copy link
Member

eugeneia commented Apr 7, 2017

Exciting release! :-)

@lukego
Copy link
Member Author

lukego commented Apr 9, 2017

@eugeneia I reckon this release is good to go although there are some interesting tidbits in the test results to acknowledge:

  • Hydra shows performance is the same as before (good.)
  • Hydra shows snabbnfv breakage with QEMU 2.7.1 and 2.8.0. This is a genuine problem that appears now because we did not have test coverage before (not due to changes in this Snabb release as we see from same errors with both master and next). I will create a separate issue to track this.
  • SnabbBot is failing the build due to the Snabbwall selftest.sh not passing. This seems to be a software dependency issue and the solution I see is that we need to come up with a convention for adding nix expressions that specify the required dependencies for test cases. (Could also consider running those tests under Hydra instead if that makes life easier. Currently our Hydra jobsets are a bit spread out but we could unify them.)

So overall it seems like this branch is 👍 with new features and no genuine regressions. However, it seems like we should take some kind of short-term action to make peace between SnabbBot and the Snabbwall tests, otherwise SnabbBot will be failing every PR that comes in the next release cycle. Guessing it is okay to ship this branch and then correct that in SnabbBot either by providing the necessary dependencies or disabling that specific test case. What do you think @eugeneia?

@eugeneia
Copy link
Member

Ideally I would add the missing dependency to snabb_bot.nix before doing the release, but I don’t immediately see whats missing, so I might postpone the fix until after the release to forgo further stalling from my side?

@lukego
Copy link
Member Author

lukego commented Apr 10, 2017

@eugeneia That sounds good to me. If we can make that fix without needing to hack the master branch, i.e. by editing out-of-tree code in snabb_box.nix, then it seems safe to release now to me.

@eugeneia
Copy link
Member

Heh, actually I can’t push the release onto master because SnabbBot did not sign off:

% git push snabbco master
Counting objects: 1, done.
Writing objects: 100% (1/1), 883 bytes | 0 bytes/s, done.
Total 1 (delta 0), reused 0 (delta 0)
remote: error: GH006: Protected branch update failed for refs/heads/master.        
remote: error: Required status check "davos-eugeneia/snabb-nfv-test-vanilla" is failing.        
To [email protected]:snabbco/snabb.git
 ! [remote rejected] master -> master (protected branch hook declined)
error: failed to push some refs to '[email protected]:snabbco/snabb.git'

@lukego
Copy link
Member Author

lukego commented Apr 10, 2017

@eugeneia The machines have finally risen up and asserted their dominance :-).

I temporarily disabled that requirement on the master branch. You should be able to push now.

@eugeneia eugeneia merged commit 71d14cd into master Apr 10, 2017
eugeneia added a commit that referenced this pull request Apr 10, 2017
dpino added a commit to dpino/snabb that referenced this pull request Jun 25, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants