Hi Wenzhuo,
> Subject: [dpdk-dev] [PATCH 02/19] net/ice: support device initialization
>
> Signed-off-by: Wenzhuo Lu
> Signed-off-by: Qiming Yang
> Signed-off-by: Xiaoyun Li
> Signed-off-by: Jingjing Wu
> ---
> +static int
> +ice_dev_init(struct rte_eth_dev *dev)
> +{
> + struct rte_
https://bugs.dpdk.org/show_bug.cgi?id=110
Bug ID: 110
Summary: system hang when device deleted for hotplug
function(igb uio)
Product: DPDK
Version: 18.11
Hardware: x86
OS: Linux
Status: CONFIRMED
Hi Wenzhuo
>
> Signed-off-by: Wenzhuo Lu
> ---
> MAINTAINERS | 1 +
> doc/guides/nics/features/ice.ini | 39
> doc/guides/nics/ice.rst | 78
>
> 3 files changed, 118 insertions(+)
> create mode 10064
Signed-off-by: Wenzhuo Lu
---
MAINTAINERS | 1 +
doc/guides/nics/features/ice.ini | 39
doc/guides/nics/ice.rst | 78
3 files changed, 118 insertions(+)
create mode 100644 doc/guides/nics/features/ice.in
Add RX functions, scatter and bulk.
Add TX function, simple.
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_lan_rxtx.c | 660 -
1 file changed, 658 insertions(+), 2 delet
Add below ops,
rx_descriptor_done
rx_descriptor_status
tx_descriptor_status
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 3 ++
drivers/net/ice/ice_lan_rxtx.c | 84 +++
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 14 +
drivers/net/ice/ice_lan_rxtx.c | 568 -
drivers/net/ice/ice_rxtx.h | 8 +
3 files changed, 588 ins
Add below ops,
stats_get
stats_reset
xstats_get
xstats_get_names
xstats_reset
Signed-off-by: Wenzhuo Lu
Signed-off-by: Jia Guo
---
drivers/net/ice/ice_ethdev.c | 574 +++
1 file changed, 574 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/drive
Add below ops,
get_eeprom_length
get_eeprom
Signed-off-by: Wei Zhao
Signed-off-by: Wenzhuo Lu
---
drivers/net/ice/ice_ethdev.c | 45
1 file changed, 45 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c
index 2d5e
Add below ops,
rxq_info_get
txq_info_get
rx_queue_count
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 3 ++
drivers/net/ice/ice_lan_rxtx.c | 66 ++
drivers
Add ops fw_version_get.
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 21 +
1 file changed, 21 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethd
Add below ops,
rx_queue_intr_enable
rx_queue_intr_disable
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 236 +++
1 file changed, 236 insertions(+)
diff --git
Add below ops,
reta_update
reta_query
rss_hash_update
rss_hash_conf_get
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 248 +++
1 file changed, 248 insertions(
Add below ops,
ice_vlan_filter_set
ice_vlan_offload_set
ice_vlan_tpid_set
ice_vlan_pvid_set
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 602 +++
1 file chan
Add below ops,
mac_addr_set
mac_addr_add
mac_addr_remove
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 242 +++
1 file changed, 242 insertions(+)
diff --git
Add ops link_update.
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 37 +
1 file changed, 37 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
config/common_base | 9 +
drivers/net/Makefile| 1 +
drivers/net/ice/Makefile| 75
drivers/net/ice/ice_ethdev.c
Add ops dev_infos_get.
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 90
1 file changed, 90 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/dr
Add ops dev_supported_ptypes_get.
Signed-off-by: Wei Zhao
Signed-off-by: Wenzhuo Lu
---
drivers/net/ice/ice_ethdev.c | 2 +
drivers/net/ice/ice_lan_rxtx.c | 601 +
drivers/net/ice/ice_rxtx.h | 2 +
3 files changed, 605 insertions(+)
diff --git
Add ops link_update.
Signed-off-by: Wenzhuo Lu
Signed-off-by: Qiming Yang
Signed-off-by: Xiaoyun Li
Signed-off-by: Jingjing Wu
---
drivers/net/ice/ice_ethdev.c | 335 +++
1 file changed, 335 insertions(+)
diff --git a/drivers/net/ice/ice_ethdev.c b/dri
Normally when starting/stopping the device the queue
should be started and stopped. Support them both in
this patch.
Below ops are added,
dev_configure
dev_start
dev_stop
dev_close
dev_reset
rx_queue_start
rx_queue_stop
tx_queue_start
tx_queue_stop
rx_queue_setup
rx_queue_release
tx_queue_setup
tx
This patch set adds the support of a new net PMD,
Intel® Ethernet Network Adapters E810, also
called ice.
Besides enabling this new NIC, also some other features
supported on this NIC.
Like below,
Basic features:
1, Basic device operations: probe, initialization, start/stop, configure, info
get.
Hi Thomas,
With DPDK 18.11-rc4 I still see issues at times, Hence more test is required
before I can send the change in request.
Thanks
Vipin Varghese
> -Original Message-
> From: Thomas Monjalon
> Sent: Friday, November 23, 2018 4:25 AM
> To: Varghese, Vipin
> Cc: dev@dpdk.org; Yigit
Thanks John for ACK for documentation.
> -Original Message-
> From: Mcnamara, John
> Sent: Thursday, November 22, 2018 11:39 PM
> To: Varghese, Vipin ; dev@dpdk.org; Pattan,
> Reshma ; tho...@monjalon.net;
> step...@networkplumber.org
> Cc: Patel, Amol
> Subject: RE: [PATCH v5 9/9] doc/pr
Any update?
Ferruh Yigit 于2018年11月21日周三 下午10:48写道：
> On 11/21/2018 1:11 PM, wanlebing wrote:
> > eth rxtx callback is dynamically allocated using rte_zmalloc()
> > but not released. Fix it by calling rte_free() to free callback
> > when removing rxtx callback.
>
> I guess they are not freed inte
22/11/2018 19:05, Mcnamara, John:
> From: Shreyansh Jain
> >
> > 18.11-rc4 verification done on NXP SoCs with integrated NICs.
>
> Acked-by: John McNamara
Applied, thanks
20/11/2018 16:28, Kovacevic, Marko:
> >
> > Add tested Intel platforms with Intel NICs to the release note.
> >
> > Signed-off-by: Lijuan Tu
> > ---
> > doc/guides/rel_notes/release_18_11.rst | 70
> > ++
> > 1 file changed, 70 insertions(+)
> >
>
> Acked-by: M
22/11/2018 18:57, Mcnamara, John:
> From: Guo, Jia
> >
> > When device has been bound to igb_uio driver and application is running,
> > hot-unplugging the device may cause kernel crash.
>
>
> I agree that it should be fixed, if possible, but in the meantime having
> it documented is a good idea.
> > From: Marvin Liu
>
> Acked-by: John McNamara
Applied, thanks
22/11/2018 19:00, Mcnamara, John:
> From: Yigit, Ferruh
> >
> > Document that AVX512F has been disabled for GCC builds [1] and document
> > its potential implications on release notes, known issue section.
> >
>
> Acked-by: John McNamara
Applied, thanks
22/11/2018 19:04, Mcnamara, John:
> From: Thomas Monjalon [mailto:tho...@monjalon.net]
> >
> > The references to the figures and tables in the index are not maintained.
> >
> > It is probably better to have no list than an incomplete list.
>
> +100
>
> Acked-by: John McNamara
Applied, thanks
When starting the device, the RSS table is set. For 8 queues, the RSS hash
table would be like | 3,2,1,0,7,6,5,4 | 3,2,1,0,7,6,5,4 |... This patch
adjusts this table to set entries sequentially. Then for 8 queues, the
RSS table would be like | 0,1,2,3,4,5,6,7 | 0,1,2,3,4,5,6,7 |...
Signed-off-by:
21/11/2018 13:48, Jerin Jacob:
> > From: Nikhil Rao
> >
> > In the eth Rx adapter SW service function,
> > move the return to after the spinlock unlock.
> >
> > Coverity issue: 302857
> > Fixes: a66a83744667 ("eventdev: fix Rx SW adapter stop")
> > CC: sta...@dpdk.org
> > Signed-off-by: Nikhil R
22/11/2018 11:07, Akhil Goyal:
> Following APIs are not currently implemented by any of the
> drivers, so marking them as rte_experimental again.
>
> rte_security_get_userdata;
> rte_security_session_stats_get;
> rte_security_session_update;
>
> Signed-off-by: Akhil Goyal
Fixes: 1a81dce780a0 ("
22/11/2018 12:04, Jerin Jacob:
> From: Joyce Kong
> >
> > When running dpdk applications on cores whose ids are bigger
> > than original max_core setting, eal error as below:
> > EAL: Detected 104 lcore(s)
> > EAL: Detected 2 NUMA nodes
> > EAL: invalid core list
> > -l CORELIST core indexes betw
The pci_resource_by_index called strlen() on uninitialized
memory which would lead to the wrong size of memory allocated
for the path portion of the resource map. This would either cause
excessively large allocation, or worse memory corruption.
Coverity Issue: 300868
Fixes: ea9d56226e72 ("pci: int
> Thomas Monjalon (5):
> eal: fix build with -O1
> kni: fix possible uninitialized variable
> net/mlx4: fix possible uninitialized variable
> eventdev: fix possible uninitialized variable
> app/eventdev: fix possible uninitialized variable
Applied
On 11/6/2018 9:48 PM, Stephen Hemminger wrote:
> The pci_resource_by_index called strlen() on uninitialized
> memory which would lead to the wrong size of memory allocated
> for the path portion of the resource map. This would either cause
> excessively large allocation, or worse memory corruption.
14/11/2018 02:52, Thomas Monjalon:
> 09/11/2018 23:03, Ferruh Yigit:
> > On 10/12/2018 12:35 PM, Vipin Varghese wrote:
> > > With the latest release, support for vdev ethernet devices like pcap,
> > > tun and tap are been added. Hence collecting statistics using proc-info
> > > is possible now.
> >
22/11/2018 18:04, Thomas Monjalon:
> 21/11/2018 13:05, Anatoly Burakov:
> > Rename rte_bsf64 to rte_bsf64_safe (this is a "safe" version in
> > that it prevents undefined behavior by checking if incoming
> > parameter is zero) and move it to common header.
> >
> > Signed-off-by: Anatoly Burakov
>
22/11/2018 19:56, Ananyev, Konstantin:
> > Rename rte_bsf64 to rte_bsf64_safe (this is a "safe" version in
> > that it prevents undefined behavior by checking if incoming
> > parameter is zero) and move it to common header.
>
> Probably a stupid one: why to rename?
> Why just not fix rte_bsf64 to
On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
> The packed ring defines were declared only if kernel
> header does not declare them.
> The problem is that they are not applied in upstream kernel,
> and some changes in the names have been required.
>
> This patch declares the defines unconditionall
20/11/2018 17:18, Anatoly Burakov:
> Previous fix for rte_panic has moved setting of alarm before
> sending the message. This means that whether we send a message,
> the alarm would still trigger. The comment noted that cleanup
> would happen in the alarm handler, but that's not what actually
> hap
On 11/22/2018 7:50 PM, Maxime Coquelin wrote:
>
>
> On 11/22/18 8:20 PM, Ferruh Yigit wrote:
>> On 11/22/2018 6:51 PM, Maxime Coquelin wrote:
>>>
>>>
>>> On 11/22/18 7:23 PM, Ferruh Yigit wrote:
On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
> The packed ring defines were declared only if
On 11/22/18 8:20 PM, Ferruh Yigit wrote:
On 11/22/2018 6:51 PM, Maxime Coquelin wrote:
On 11/22/18 7:23 PM, Ferruh Yigit wrote:
On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
The packed ring defines were declared only if kernel
header does not declare them.
The problem is that they are not
On 11/22/2018 6:51 PM, Maxime Coquelin wrote:
>
>
> On 11/22/18 7:23 PM, Ferruh Yigit wrote:
>> On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
>>> The packed ring defines were declared only if kernel
>>> header does not declare them.
>>> The problem is that they are not applied in upstream kernel,
>
> Rename rte_bsf64 to rte_bsf64_safe (this is a "safe" version in
> that it prevents undefined behavior by checking if incoming
> parameter is zero) and move it to common header.
Probably a stupid one: why to rename?
Why just not fix rte_bsf64 to make it work with zero value,
and keep the sa
Changes to make ipsec-secgw to utilize librte_ipsec library.
That patch provides:
- changes in the related data structures.
- changes in the initialization code.
- new command-line parameters to enable librte_ipsec codepath
and related features.
Note that right now by default current (non-li
Changes to make ipsec-secgw data-path code to utilize librte_ipsec library.
Note that right now by default current (non-librte_ipsec) code-path will
be used. User has to run application with new command-line option ('-l')
to enable new codepath.
Signed-off-by: Mohammad Abdul Awal
Signed-off-by: K
On 11/22/18 7:23 PM, Ferruh Yigit wrote:
On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
The packed ring defines were declared only if kernel
header does not declare them.
The problem is that they are not applied in upstream kernel,
and some changes in the names have been required.
This patch
Looking at process_pkts_outbound_nosp() there seems few issues:
- accessing mbuf after it was freed
- invoking ipsec_outbound() for ipv4 packets only
- copying number of packets, but not the mbuf pointers itself
that patch provides fixes for that issues.
Fixes: 906257e965b7 ("examples/ipsec-secgw
The purpose of these scripts is to automate ipsec-secgw functional testing.
The scripts require two machines (SUT and DUT) connected through
at least 2 NICs and running linux (so far tested only on Ubuntu 18.04).
Introduced test-cases for the following scenarios:
- Transport/Tunnel modes
- AES-CBC
In some cases crypto-ops could never be dequeued from the crypto-device.
The easiest way to reproduce:
start ipsec-secgw with crypto-dev and send to it less then 32 packets.
none packets will be forwarded.
Reason for that is that the application does dequeue() from crypto-queues
only when new packe
ipsec-secgw always enables TX offloads
(DEV_TX_OFFLOAD_MULTI_SEGS, DEV_TX_OFFLOAD_SECURITY),
even when they are not requested by the config.
That causes many PMD to choose full-featured TX function,
which in many cases is much slower then one without offloads.
That patch adds checks to enabled extr
This patch series targets 19.02 release.
This patch series depends on the patch series:
http://patches.dpdk.org/patch/48044/
http://patches.dpdk.org/patch/48045/
http://patches.dpdk.org/patch/48143/
http://patches.dpdk.org/patch/48144/
http://patches.dpdk.org/patch/48145/
http://patches.dpdk.org/
In some cases it is useful to allow user to specify destination
ether address for outgoing packets.
This patch adds such ability by introducing new 'neigh' config
file option.
Signed-off-by: Konstantin Ananyev
---
examples/ipsec-secgw/ipsec-secgw.c | 21 +++--
examples/ipsec-secgw/ipsec.h
On 11/22/2018 12:09 PM, Ian Stokes wrote:
> Maximum and minimum MTU values vary between hardware devices. In
> hardware agnostic DPDK applications access to such information would
> allow a more accurate way of validating and setting supported MTU values on
> a per device basis rather than using a
On 11/22/2018 5:09 PM, Maxime Coquelin wrote:
> The packed ring defines were declared only if kernel
> header does not declare them.
> The problem is that they are not applied in upstream kernel,
> and some changes in the names have been required.
>
> This patch declares the defines unconditionall
Minutes 22 November 2018
Agenda:
* Dates
* RC4 status
* Subtrees
* OvS
* Roadmap
Release Dates
-
* RC1 was out on Monday, 29 October.
* RC2 was out on Tuesday, 6 November.
* RC3 was out on Tuesday, 13 November.
* RC4 was out on Sunday, 18 November.
* For res
> -Original Message-
> From: Thomas Monjalon [mailto:tho...@monjalon.net]
> Sent: Thursday, November 22, 2018 2:19 PM
> To: dev@dpdk.org
> Cc: Mcnamara, John ; Kovacevic, Marko
>
> Subject: [PATCH 2/2] doc: remove PCI-specific details from EAL guide
>
> The PCI bus is an independent dr
> -Original Message-
> From: Varghese, Vipin
> Sent: Thursday, November 22, 2018 2:16 PM
> To: dev@dpdk.org; Pattan, Reshma ;
> tho...@monjalon.net; step...@networkplumber.org; Mcnamara, John
>
> Cc: Patel, Amol ; Varghese, Vipin
>
> Subject: [PATCH v5 9/9] doc/procinfo: add informatio
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Marvin Liu
> Sent: Wednesday, October 31, 2018 2:49 PM
> To: tho...@monjalon.net
> Cc: dev@dpdk.org; Liu, Yong
> Subject: [dpdk-dev] [PATCH] Doc: add known issue for restricted vdev
> ethdev ops in secondary pro
> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Shreyansh Jain
> Sent: Wednesday, November 21, 2018 9:29 AM
> To: dev@dpdk.org
> Cc: Hemant Agrawal ; Shreyansh Jain
>
> Subject: [dpdk-dev] [PATCH] doc: add tested NXP NICs information
>
> 18.11-rc4 verificati
> -Original Message-
> From: Thomas Monjalon [mailto:tho...@monjalon.net]
> Sent: Thursday, November 22, 2018 2:19 PM
> To: dev@dpdk.org
> Cc: Mcnamara, John ; Kovacevic, Marko
>
> Subject: [PATCH 1/2] doc: remove lists of figure and table references
>
> The references to the figures a
> -Original Message-
> From: Yigit, Ferruh
> Sent: Tuesday, November 20, 2018 12:51 PM
> To: Mcnamara, John ; Kovacevic, Marko
>
> Cc: dev@dpdk.org; Yigit, Ferruh ; Richardson,
> Bruce ; Ananyev, Konstantin
>
> Subject: [PATCH] doc: announce disabling GCC AVX512F as known issue
>
> Do
> -Original Message-
> From: Wang, Yipeng1
> Sent: Wednesday, November 21, 2018 5:43 PM
> To: Richardson, Bruce ; Mcnamara, John
>
> Cc: dev@dpdk.org; Wang, Yipeng1 ;
> honnappa.nagaraha...@arm.com; Gobriel, Sameh ;
> Tai, Charlie
> Subject: [PATCH] doc: hash library programmer guide im
> -Original Message-
> From: Guo, Jia
> Sent: Tuesday, November 20, 2018 3:10 PM
> To: Kovacevic, Marko ; Mcnamara, John
> ; Zhang, Qi Z ; Yigit,
> Ferruh ; tho...@monjalon.net
> Cc: dev@dpdk.org; Guo, Jia ; Zhang, Helin
> ; Ananyev, Konstantin
> ; He, Shaopeng ;
> step...@networkplumber
> -Original Message-
> From: Bie, Tiwei
> Sent: Thursday, November 22, 2018 2:52 AM
> To: dev@dpdk.org
> Cc: Pattan, Reshma ; Mcnamara, John
> ; Kovacevic, Marko ;
> sta...@dpdk.org
> Subject: [PATCH] doc: update commands for pdump
>
> The --server-socket-path and --client-socket-path o
> -Original Message-
> From: Luca Boccassi [mailto:bl...@debian.org]
> Sent: Tuesday, November 20, 2018 1:08 PM
> To: Zhang, Qi Z ; dev@dpdk.org
> Cc: Lu, Wenzhuo ; Ananyev, Konstantin
> ; 3ch...@gmail.com; sta...@dpdk.org
> Subject: Re: [dpdk-stable] [PATCH v4] net/ixgbe: add LHA ID to x
On 11/20/2018 3:09 PM, Jeff Guo wrote:
> When device has been bound to igb_uio driver and application is running,
> hot-unplugging the device may cause kernel crash.
>
> Signed-off-by: Jeff Guo
Reviewed-by: Ferruh Yigit
The patch optimizes the mac swap operation by taking advantage
of SSE instructions, it only impacts x86 platform.
Signed-off-by: Qi Zhang
---
app/test-pmd/macswap.c | 4
app/test-pmd/macswap_sse.h | 43 +++
2 files changed, 47 insertions(+)
crea
Move macswap workload to dedicate function, so we can further enable
platform specific optimized version.
Signed-off-by: Qi Zhang
---
app/test-pmd/macswap.c| 32 ++--
app/test-pmd/macswap.h| 40
app/test-pmd/mac
Do four packets macswap in same loop iterate to squeeze more
CPU cycles.
Signed-off-by: Qi Zhang
---
app/test-pmd/macswap_sse.h | 65 ++
1 file changed, 54 insertions(+), 11 deletions(-)
diff --git a/app/test-pmd/macswap_sse.h b/app/test-pmd/macswap_s
**The pathset is target for 19.02**
Improved testpmd macswap performance for x86 by take advantage of SSE
instructions.
On a broadwell 1.6GHz sever with a i40e 25G NIC.
We abserve 17.7% performance improvement for testpmd's macswap test.
v2:
- remove uncessary files
Qi Zhang (3):
app/testpmd:
19/11/2018 14:10, Anatoly Burakov:
> --- /dev/null
> +++ b/doc/guides/freebsd_gsg/freebsd_eal_parameters.rst
> @@ -0,0 +1,20 @@
> +.. SPDX-License-Identifier: BSD-3-Clause
> +Copyright(c) 2018 Intel Corporation.
> +
> +EAL parameters
> +==
> +
> +This document contains a list of al
Do four packets macswap in same loop iterate to squeeze more
CPU cycles.
Signed-off-by: Qi Zhang
---
app/test-pmd/macswap_sse.h | 65 ++
1 file changed, 54 insertions(+), 11 deletions(-)
diff --git a/app/test-pmd/macswap_sse.h b/app/test-pmd/macswap_s
The patch optimizes the mac swap operation by taking advantage
of SSE instructions, it only impacts x86 platform.
Signed-off-by: Qi Zhang
---
app/test-pmd/macswap.c | 4
app/test-pmd/macswap_sse.h | 43 +++
2 files changed, 47 insertions(+)
crea
**The pathset is target for 19.02**
Improved testpmd macswap performance for x86 by take advantage
of SSE instructions.
On a broadwell 1.6GHz sever with a i40e 25G NIC.
We abserve 17.7% performance improvement for testpmd's macswap
test.
Qi Zhang (3):
app/testpmd: code refactory for macswap
a
Move macswap workload to dedicate function, so we can further enable
platform specific optimized version.
Signed-off-by: Qi Zhang
---
app/test-pmd/l2fwd.h | 40
app/test-pmd/l2fwd_common.h | 36
app/test-pmd
22/11/2018 13:11, Ferruh Yigit:
> On 11/22/2018 12:09 PM, Ian Stokes wrote:
> > Maximum and minimum MTU values vary between hardware devices. In
> > hardware agnostic DPDK applications access to such information would
> > allow a more accurate way of validating and setting supported MTU values on
>
19/11/2018 11:33, Anatoly Burakov:
> If there aren't any devices of a particular category on user's
> system, we still display them, which is bad for usability. Fix
> devbind to not print out a category unless there are devices in
> it.
>
> Signed-off-by: Anatoly Burakov
> Reviewed-by: David Hunt
The packed ring defines were declared only if kernel
header does not declare them.
The problem is that they are not applied in upstream kernel,
and some changes in the names have been required.
This patch declares the defines unconditionally, which
fixes potential build issues.
Fixes: 297b1e7350f
On 11/22/2018 1:22 PM, Olivier Matz wrote:
> On Thu, Nov 22, 2018 at 11:32:23AM +0100, Thomas Monjalon wrote:
>> After processing a kvlist in rte_kvargs_process(),
>> it may be needed to loop again over kvlist in order to know
>> whether the key is matched or not.
>> In order to simplify implementa
21/11/2018 13:05, Anatoly Burakov:
> Rename rte_bsf64 to rte_bsf64_safe (this is a "safe" version in
> that it prevents undefined behavior by checking if incoming
> parameter is zero) and move it to common header.
>
> Signed-off-by: Anatoly Burakov
> Acked-by: Cristian Dumitrescu
> Acked-by: Jas
Increase the number of addressable cores from 64 to 256. Also remove the
warning that incresing this number beyond 64 will cause problems (because
of the previous use of uint64_t masks). Now this number can be increased
significantly without causing problems.
Signed-off-by: David Hunt
---
exampl
since we're moving to allowing greater than 64 cores, the mask functions
that use uint64_t to perform functions on a masked set of cores are no
longer feasable, so removing them.
Signed-off-by: David Hunt
---
examples/vm_power_manager/channel_monitor.c | 24
examples/vm_power_manager/power
Extending the functionality to allow vms to power manage cores beyond 63.
Signed-off-by: David Hunt
---
examples/vm_power_manager/channel_manager.c | 59 -
examples/vm_power_manager/channel_manager.h | 30 ++-
examples/vm_power_manager/channel_monitor.c | 56 +++--
vm_power_manager has up to now been limited to 64-cores because of the
extensive use of uint64 type for bitmasks. This patch set converts those
uint64s to character arrays, and increases the limit of cores from 64 to
256, and allowing users to increase this by increasing the #define.
First of all
vm_power_manager currently makes use of uint64_t masks to keep track of
cores in use, limiting use of the app to only being able to manage the
first 64 cores in a multi-core system. Many modern systems have core
counts greater than 64, so this limitation needs to be removed.
This patch converts th
On 14/11/2018 4:52 PM, Singh, Jasvinder wrote:
-Original Message-
From: Burakov, Anatoly
Sent: Wednesday, November 14, 2018 4:47 PM
To: dev@dpdk.org
Cc: Neil Horman ; Mcnamara, John
; Kovacevic, Marko
; Dumitrescu, Cristian
; tho...@monjalon.net; Richardson, Bruce
; Yigit, Ferruh ; Si
On 11/22/2018 4:18 PM, Dekel Peled wrote:
> Hi,
>
> The current implementation is already validated, and since this is the last
> minute I prefer my patch to be applied as-is.
> Please ack.
Hi Dekel,
I think logic is other-way around, a patch has been acked clearly, without
question can justify
On 11/22/2018 2:51 AM, Tiwei Bie wrote:
> The --server-socket-path and --client-socket-path options
> have already been removed. So also remove them from the doc
> to avoid confusion.
>
> Fixes: 09f4aa2b955f ("app/pdump: remove unused socket path options")
> Cc: sta...@dpdk.org
>
> Signed-off-by:
Hi,
The current implementation is already validated, and since this is the last
minute I prefer my patch to be applied as-is.
Please ack.
Regards,
Dekel
> -Original Message-
> From: Adrien Mazarguil
> Sent: Thursday, November 22, 2018 12:14 PM
> To: Dekel Peled
> Cc: wenzhuo...@intel.
Hi, Hideyuki,
>Rami, thanks for your advice.
>If I understand you correctly, then
>there already exist APIs to collect statistic
>information inside dpdk including CPU usage.
Yup. I want to also note that the librte jobstats is very veteran, it
exists in DPDK versions prior to 16.04 (like in 2.2.
On 4/9/2018 7:00 AM, Kumar, Ravi1 wrote:
>>
>> -Original Message-
>> From: Ferruh Yigit [mailto:ferruh.yi...@intel.com]
>> Sent: Friday, April 06, 2018 9:26 PM
>> To: Kumar, Ravi1 ; dev@dpdk.org
>> Subject: Re: [PATCH v5 01/18] net/axgbe: add minimal dev init and uninit
>> support
>>
>> O
Thursday, November 22, 2018 3:49 PM¸ Viacheslav Ovsiienko:
> Subject: [dpdk-dev] [PATCH] doc: add mlx5 E-Switch VXLAN tunnels
> limitations
>
> This patch adds limitation notice for MLX5 PMD regarding VXLAN tunnels
> support on E-Switch Flows.
>
> Signed-off-by: Viacheslav Ovsiienko
Applied to
> > + /* iterate each object */
> > + int ret = rte_mempool_obj_iter(ptr,
> > + mempool_itr_obj, NULL);
>
> rte_mempool_obj_iter returns uint32_t, so good to use "ret" type to be
> uint32_t. So it will be in synch with below printf to
>
> Minor observations which are things that checkpatch etc won't see but make
> the code easier to read/maintain.
>
> > +/* border variable to hold for show */ char bdr_str[MAX_STRING_LEN];
Done for v5
>
> Does this have to be global, could it just be static?
>
> > + memset(&link,
Document update for debug options and information for PMD instances
like port, traffic manager, crypto, mempool and ring instances.
Signed-off-by: Vipin Varghese
---
V3:
- update document from dbg to show - Vipin Varghese
V2:
- update word style for content - Vipin Varghese
---
doc/guides/to
> > + snprintf(bdr_str, 100, " Port (%u)", i);
>
> %s/100/MAX_STRING_LEN ?
Done for v5
>
> > + ret = rte_eth_dev_rss_hash_conf_get(i, &rss_conf);
> > + if ((ret) || (rss_conf.rss_key == NULL))
> > + continue;
> > +
>
1 - 100 of 144 matches
Mail list logo
