Revision e6d2e0176e1f3c1e1534851b66c0b972f03ff069 authored by Jubin John on 12 April 2016, 17:47:00 UTC, committed by Doug Ledford on 28 April 2016, 16:00:39 UTC
call_send is used to determine whether to send immediately or schedule a send for later. The current logic in rdmavt is inverted and has a negative impact on the latency of the hfi1 and qib drivers. Fix this regression by correctly calling send immediately when call_send is set. Reviewed-by: Dennis Dalessandro <dennis.dalessandro@intel.com> Reviewed-by: Mike Marciniszyn <mike.marciniszyn@intel.com> Signed-off-by: Jubin John <jubin.john@intel.com> Signed-off-by: Doug Ledford <dledford@redhat.com>
1 parent 849e3e9
| File | Mode | Size |
|---|---|---|
| Kconfig | -rw-r--r-- | 2.2 KB |
| Makefile | -rw-r--r-- | 446 bytes |
| actions.c | -rw-r--r-- | 29.2 KB |
| conntrack.c | -rw-r--r-- | 35.7 KB |
| conntrack.h | -rw-r--r-- | 2.6 KB |
| datapath.c | -rw-r--r-- | 58.1 KB |
| datapath.h | -rw-r--r-- | 6.4 KB |
| dp_notify.c | -rw-r--r-- | 2.6 KB |
| flow.c | -rw-r--r-- | 19.1 KB |
| flow.h | -rw-r--r-- | 6.7 KB |
| flow_netlink.c | -rw-r--r-- | 68.1 KB |
| flow_netlink.h | -rw-r--r-- | 2.7 KB |
| flow_table.c | -rw-r--r-- | 18.9 KB |
| flow_table.h | -rw-r--r-- | 2.8 KB |
| vport-geneve.c | -rw-r--r-- | 3.0 KB |
| vport-gre.c | -rw-r--r-- | 2.6 KB |
| vport-internal_dev.c | -rw-r--r-- | 7.8 KB |
| vport-internal_dev.h | -rw-r--r-- | 1.0 KB |
| vport-netdev.c | -rw-r--r-- | 5.3 KB |
| vport-netdev.h | -rw-r--r-- | 1.1 KB |
| vport-vxlan.c | -rw-r--r-- | 4.1 KB |
| vport.c | -rw-r--r-- | 12.6 KB |
| vport.h | -rw-r--r-- | 6.2 KB |
Computing file changes ...
