On 5/29/19 3:04 PM, Maxime Coquelin wrote:
Some OVS-DPDK PVP benchmarks show a performance drop
when switching from DPDK v17.11 to v18.11.
With the addition of packed ring layout support,
rte_vhost_enqueue_burst and rte_vhost_dequeue_burst
became very large, and only a part of the instructions
are executed (either packed or split ring used).
This series aims at improving the I-cache pressure,
first by un-inlining split and packed rings, but
also by moving parts considered as cold in dedicated
functions (dirty page logging, fragmented descriptors
buffer management added for CVE-2018-1059).
With the series applied, size of the enqueue and
dequeue split paths is reduced significantly:
+---------+--------------------+---------------------+
| Version | Enqueue split path | Dequeue split path |
+---------+--------------------+---------------------+
| v19.05 | 16461B | 25521B |
| +series | 7286B | 11285B |
+---------+--------------------+---------------------+
Using perf tool to monitor iTLB-load-misses event
while doing PVP benchmark with testpmd as vswitch,
we can see the number of iTLB misses being reduced:
- v19.05:
# perf stat --repeat 10 -C 2,3 -e iTLB-load-miss -- sleep 10
Performance counter stats for 'CPU(s) 2,3' (10 runs):
2,438 iTLB-load-miss
( +- 13.43% )
10.00058928 +- 0.00000336 seconds time elapsed ( +- 0.00% )
- +series:
# perf stat --repeat 10 -C 2,3 -e iTLB-load-miss -- sleep 10
Performance counter stats for 'CPU(s) 2,3' (10 runs):
55 iTLB-load-miss
( +- 10.08% )
10.00059466 +- 0.00000283 seconds time elapsed ( +- 0.00% )
The series also force the inlining of some rte_memcpy
helpers, as by adding packed ring support, some of them
were not more inlined but embedded as functions in
the virtio_net object file, which was not expected.
Finally, the series simplifies the descriptors buffers
prefetching, by doing it in the recently introduced
descriptor buffer mapping function.
v3:
===
- Prefix alloc_copy_ind_table with vhost_ (Mattias)
- Remove double new line (Tiwei)
- Fix grammar error in patch 3's commit message (Jens)
- Force noinline for hear copy functions (Mattias)
- Fix dst assignement in copy_hdr_from_desc (Tiwei)
v2:
===
- Fix checkpatch issue
- Reset author for patch 5 (David)
- Force non-inlining in patch 2 (David)
- Fix typo in path 3 commit message (David)
Maxime Coquelin (5):
vhost: un-inline dirty pages logging functions
vhost: do not inline packed and split functions
vhost: do not inline unlikely fragmented buffers code
vhost: simplify descriptor's buffer prefetching
eal/x86: force inlining of all memcpy and mov helpers
.../common/include/arch/x86/rte_memcpy.h | 18 +-
lib/librte_vhost/vdpa.c | 2 +-
lib/librte_vhost/vhost.c | 164 +++++++++++++++++
lib/librte_vhost/vhost.h | 165 ++----------------
lib/librte_vhost/virtio_net.c | 140 +++++++--------
5 files changed, 251 insertions(+), 238 deletions(-)
Applied patches 1 to 4 to dpdk-next-virtio/master.
Bruce, I'm assigning patch 5 to you in Patchwork, as this is not
vhost/virtio specific.
Thanks,
Maxime
