[ovs-git] Open vSwitch: worker: Don't have worker abort when parent dies. (branch-1.9)

dev at openvswitch.org dev at openvswitch.org
Sat Jan 26 03:01:25 UTC 2013


This is an automated email from the git hooks/post-receive script. It was
generated because a ref change was pushed to the repository containing
the project "Open vSwitch".

The branch, branch-1.9 has been updated
       via  7c944cb900ec3a1f309148adb79fc339961cc6f4 (commit)
       via  fea1494c0ebaa949da0c962980dc369e82fb4b5e (commit)
      from  0d88f017c41e3bf9152f93d60d0fa179b42c2bdc (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 7c944cb900ec3a1f309148adb79fc339961cc6f4
Diffs: http://openvswitch.org/cgi-bin/gitweb.cgi?p=openvswitch;a=commitdiff;h=7c944cb900ec3a1f309148adb79fc339961cc6f4
Author: Justin Pettit <jpettit at nicira.com>
		
worker: Don't have worker abort when parent dies.
		
Depending on how forcefully the parent process is killed, the worker
could abort when trying to read or write on their shared socket.  This
changes those errors from VLOG_ABORT to VLOG_FATAL so that a core isn't
generated.

Bug #14821

Reported-by: Amey Bhide <abhide at nicira.com>
Signed-off-by: Justin Pettit <jpettit at nicira.com>


commit fea1494c0ebaa949da0c962980dc369e82fb4b5e
Diffs: http://openvswitch.org/cgi-bin/gitweb.cgi?p=openvswitch;a=commitdiff;h=fea1494c0ebaa949da0c962980dc369e82fb4b5e
Author: Ben Pfaff <blp at nicira.com>
		
ofproto: Properly refresh rule modified time when nothing else changes.
		
In Open vSwitch, a "modify" or "modify_strict" flow_mod is supposed to
refresh the flow's last-modified time even if nothing else changes, because
this interpretation makes the "learn" action more useful.  As commit
308881afb (ofproto: Reinterpret meaning of OpenFlow hard timeouts with
OFPFC_MODIFY.) notes:

    I finally found a good use for hard timeouts in OpenFlow, but they
    require a slight reinterpretation of the meaning of hard timeouts.
    Until now, a hard timeout meant that a flow would be removed the
    specified number of seconds after a flow was created.  Intervening
    modifications with OFPFC_MODIFY(_STRICT) had no effect on the hard
    timeout; the flow would still be deleted the specified number of
    seconds after its original creation.

    This commit changes the effect of OFPFC_MODIFY(_STRICT).  Now,
    modifying a flow resets its hard timeout counter.  A flow will time out
    the specified number of seconds after creation or after the last time
    it is modified, whichever comes later.

However, commit 080437614b (ofproto: Represent flow cookie changes as
operations too.) broke this behavior because it incorrectly optimized out
"modify" operations that didn't change the flow's actions or flow cookie.
This commit fixes the problem, and adds a test to prevent future
regression.

Thanks to Amar Padmanabhan <amar at nicira.com> for helping to track this
down.

Bug #14841.
Reported-by: Hiroshi Tanaka <htanaka at vmware.com>
CC: Amar Padmanabhan <amar at nicira.com>
Signed-off-by: Ben Pfaff <blp at nicira.com>


-----------------------------------------------------------------------

Summary of changes:
 lib/worker.c      |    4 +-
 ofproto/ofproto.c |   18 +++++++++---
 tests/learn.at    |   75 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 90 insertions(+), 7 deletions(-)


hooks/post-receive
-- 
Open vSwitch



More information about the git mailing list