When running TIPC in large clusters we experience behavior that
may potentially become problematic in the future. This series
picks some low-hanging fruit in this regard, and also fixes a
couple of other minor issues.
v2: #1: Found even more remnants to remove.
#2: Clarified problem description, and ensured that even the
first discovery broadcast is sent as it should be.
#3: Introduce a "half" disabling of the bearer to be reset,
and moved the fix to l2 layer code only, where the problem
has manifested itself. We cannot do a full bearer disable
here, since we cannot know when to enable it again unless
we have kept the bearer instance. Hopefully resolved the
atomicity problem (which I frankly don't see as a real
problem) as described in feedback from Ying.
v3: #3: Just clearing/setting the 'tipc_ptr' field in struct netdev
when an interface goes up or down, instead of calling
l2_enable/l2_disable(). Otherwise, l2_disable() will be
called twice when we call bearer_disable(), and we end up
with a negative reference counter in the device.
v4: #3: Moved call to tipc_node_delete_links() in bearer_disable()
back to the original location, after dev->tipc_ptr but before
bearer_list[bearer_id] have been set to NULL. Thanks Ying.
Jon Maloy (4):
tipc: remove remnants of old broadcast code
tipc: eliminate buffer leak in bearer layer
tipc: stricter filtering of packets in bearer layer
tipc: reduce transmission rate of RESET messages when link is down
net/tipc/bearer.c | 98 +++++++++++++++++++++++++++++++----------------------
net/tipc/bearer.h | 15 --------
net/tipc/discover.c | 7 ++--
net/tipc/discover.h | 2 +-
net/tipc/link.c | 5 ++-
net/tipc/msg.h | 5 +++
6 files changed, 69 insertions(+), 63 deletions(-)
--
1.9.1
------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
_______________________________________________
tipc-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/tipc-discussion
