Hi,
On 05/18/2018 04:02 AM, Yao, Lei A wrote:
Hi, Maxime
Any idea for this performance drop? Will we improve it in this release
or it will be long term work? Thanks.
No, it will not be improved for this release. I'll create a Bz to track
this so that the release note can mention it.
Regards,
Maxime
BRs
Lei
-----Original Message-----
From: Yao, Lei A
Sent: Wednesday, May 2, 2018 8:10 PM
To: Maxime Coquelin <maxime.coque...@redhat.com>; dev@dpdk.org
Cc: Bie, Tiwei <tiwei....@intel.com>
Subject: RE: [dpdk-dev] [PATCH 00/12] Vhost: CVE-2018-1059 fixes
-----Original Message-----
From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
Sent: Wednesday, May 2, 2018 5:20 PM
To: Yao, Lei A <lei.a....@intel.com>; dev@dpdk.org
Cc: Bie, Tiwei <tiwei....@intel.com>
Subject: Re: [dpdk-dev] [PATCH 00/12] Vhost: CVE-2018-1059 fixes
Hi Lei,
Thanks for the perf report.
On 05/02/2018 07:08 AM, Yao, Lei A wrote:
Hi, Maxime
During the 18.05-rc1 performance testing, I find this patch set will bring
slightly performance drop on mergeable and normal path, and big
performance
drop on vector path. Could you have a check on this? I know this patch is
important for security. Not sure if there is any way to improve the
performance.
Could you please share info about the use cases you are benchmarking?
I run vhost/virtio loopback test .
There may be ways to improve the performance, for this we would need to
profile the code to understand where the bottlenecks are.
Mergebale
packet size
64 0.80%
128 -2.75%
260 -2.93%
520 -2.72%
1024 -1.18%
1500 -0.65%
Normal
packet size
64 -1.47%
128 -7.43%
260 -3.66%
520 -2.52%
1024 -1.19%
1500 -0.78%
Vector
packet size
64 -8.60%
128 -3.54%
260 -2.63%
520 -6.12%
1024 -1.05%
1500 -1.20%
Are you sure this is only this series that induces such a big
performance drop in vector test? I.e. have you run the benchmark
just before and right after the series is applied?
Yes. The performance drop I list here is just compared before and after your
patch set. The key patch bring performance drop is this commit
" Commit hash: 41333fba5b98945b8051e7b48f8fe47432cdd356"
vhost: introduce safe API for GPA translation.
Between 18.02 and 18.05-rc1, there are some other performance drop, but
not
so large. I need more git bisect work to identify.
Thanks,
Maxime
CPU info: Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz
OS: Ubuntu 16.04
BRs
Lei
-----Original Message-----
From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Maxime
Coquelin
Sent: Monday, April 23, 2018 11:58 PM
To: dev@dpdk.org
Cc: Maxime Coquelin <maxime.coque...@redhat.com>
Subject: [dpdk-dev] [PATCH 00/12] Vhost: CVE-2018-1059 fixes
This series fixes the security vulnerability referenced
as CVE-2018-1059.
Patches are already applied to the branch, but reviews
are encouraged. Any issues spotted would be fixed on top.
Maxime Coquelin (12):
vhost: fix indirect descriptors table translation size
vhost: check all range is mapped when translating GPAs
vhost: introduce safe API for GPA translation
vhost: ensure all range is mapped when translating QVAs
vhost: add support for non-contiguous indirect descs tables
vhost: handle virtually non-contiguous buffers in Tx
vhost: handle virtually non-contiguous buffers in Rx
vhost: handle virtually non-contiguous buffers in Rx-mrg
examples/vhost: move to safe GPA translation API
examples/vhost_scsi: move to safe GPA translation API
vhost/crypto: move to safe GPA translation API
vhost: deprecate unsafe GPA translation API
examples/vhost/virtio_net.c | 94 +++++++-
examples/vhost_scsi/vhost_scsi.c | 56 ++++-
lib/librte_vhost/rte_vhost.h | 46 ++++
lib/librte_vhost/rte_vhost_version.map | 4 +-
lib/librte_vhost/vhost.c | 39 ++--
lib/librte_vhost/vhost.h | 8 +-
lib/librte_vhost/vhost_crypto.c | 65 ++++--
lib/librte_vhost/vhost_user.c | 58 +++--
lib/librte_vhost/virtio_net.c | 411
++++++++++++++++++++++++++++-
----
9 files changed, 650 insertions(+), 131 deletions(-)
--
2.14.3
