Go to file
Stephen Hemminger 1f1f3a0078 sched: fix grinder bug
The rte_scheduler will get stuck and not deliver any more packets
if there are two active subports and then one of them stops enqueing
more packets. This is because of a bug in how the grinder state machines
are managed.

If a non-zero grinder is assigned (but not yet active), then the dequeue
would miss it and always return zero packets. The cure is to always
do a first pass over all grinders.

Signed-off-by: Stephen Hemminger <shemming@brocade.com>
Acked-by: Thomas Monjalon <thomas.monjalon@6wind.com>
2014-05-22 16:14:36 +02:00
app sched: use malloc instead of memzone for allocation 2014-05-22 16:13:30 +02:00
config config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
doc config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
examples config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
lib sched: fix grinder bug 2014-05-22 16:14:36 +02:00
mk config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
pkg config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
scripts config: rename "default" configurations as "native" 2014-05-21 16:25:06 +02:00
tools tools: rename pci_unbind script 2014-03-20 17:50:51 +01:00
.gitignore init DPDK repository 2013-03-07 10:57:42 +01:00
GNUmakefile add FreeBSD support 2014-02-25 21:29:18 +01:00
LICENSE.GPL doc: GPL/LGPL licenses 2013-07-25 14:43:06 +02:00
LICENSE.LGPL doc: fix file format (dos to unix) 2013-09-06 11:43:07 +02:00
Makefile add FreeBSD support 2014-02-25 21:29:18 +01:00