+48
−2
+1
−1
Loading
With previous implementation connect was blocking, time was not deterministic and depending on global networking settings. Checked VM to VM connect blocking time just to have some reference. 1) with allowed traffic; few ms (4-6) 2) not allowed traffic from 4420/tcp; as above; EHOSTUNREACH errno 3) no listener; as above, ECONNREFUSED errno 4) drop rule on target; 135s; ETIMEDOUT errno 5) link down on target; 70s; EHOSTUNREACH errno 6) link down on target + arp table flush; 3s; EHOSTUNREACH errno 7) non existing ip; as above Blocking behavior is maintained with this change however it is done on posix poll() and not connect(). Following patch will introduce timeout opt. 1-3 scenarios results are the same with the change, however 4-7 can be significantly mitigated. This is a step in the middle of transitioning to async connect. Change-Id: Ia2f347a3cfdcd427b8302e0220448b0f34d19dd4 Signed-off-by:Jacek Kalwas <jacek.kalwas@nutanix.com> Reviewed-on: https://review.spdk.io/c/spdk/spdk/+/25794 Reviewed-by:
Ben Walker <ben@nvidia.com> Reviewed-by:
Tomasz Zawadzki <tomasz@tzawadzki.com> Reviewed-by:
Jim Harris <jim.harris@nvidia.com> Community-CI: Mellanox Build Bot Tested-by:
SPDK Automated Test System <spdkbot@gmail.com>