Hi, Tyler
May I know why you need to change the legacy atomics? rte_atomic32_xxx series
action is performance optimized, I don't think the change is needed as no bug
or other requirement.
Qiming
> -Original Message-
> From: David Marchand
> Sent: Tuesday, June 13, 2023 5:51 PM
> To: Ty
Hi,
Thanks for your comments, replies are inline.
Thanks a again,
Cheng
> -Original Message-
> From: huangdengdui
> Sent: Tuesday, June 13, 2023 8:55 PM
> To: Jiang, Cheng1 ; tho...@monjalon.net;
> Richardson, Bruce ;
> m...@smartsharesystems.com; Xia, Chenbo ;
> amitpraka...@marvell.co
> -Original Message-
> From: Zhang, Qi Z
> Sent: Monday, June 12, 2023 8:40 PM
> To: Ye, MingjinX ; dev@dpdk.org
> Cc: Yang, Qiming ; Zhou, YidingX
> ; sta...@dpdk.org; Wu, Jingjing
> ; Xing, Beilei
> Subject: RE: [PATCH] net/iavf: fix abnormal disable HW interrupt
>
>
>
> > -Or
> -Original Message-
> From: Tang, Yaqi
> Sent: Wednesday, June 14, 2023 10:55 AM
> To: Xing, Beilei ; Wu, Jingjing
> Cc: dev@dpdk.org; Xing, Beilei
> Subject: RE: [PATCH] net/cpfl: release memzone for rxq
>
>
> > -Original Message-
> > From: beilei.x...@intel.com
> > Sent:
> -Original Message-
> From: Wu, Wenjun1
> Sent: Friday, June 9, 2023 6:54 PM
> To: dev@dpdk.org; Wu, Jingjing ; Xing, Beilei
> ; Zhang, Qi Z
> Cc: Wu, Wenjun1 ; sta...@dpdk.org
> Subject: [PATCH v2] common/idpf: fix memory leak on AVX512 TX queue
> close
>
> When releasing mbufs for
Add support for modify field in tunnel MPLS header.
For now it is supported only to copy from.
Signed-off-by: Michael Baum
Acked-by: Matan Azrad
---
doc/guides/nics/mlx5.rst| 2 ++
drivers/common/mlx5/mlx5_prm.h | 5 +
drivers/net/mlx5/mlx5_flow.h| 1 +
drivers/net/mlx5/mlx5
Add support for MPLS header modification.
RFC:
https://patchwork.dpdk.org/project/dpdk/patch/20230420094347.523784-1-michae...@nvidia.com/
Michael Baum (2):
net/mlx5: align implementation with modify API
net/mlx5: add MPLS modify field support
doc/guides/nics/mlx5.rst| 2 ++
driver
Modify field action API has been changed to use "tag_index" instead of
"level" fields to represent the tag array for TAG modification type.
Although the old API still work, this patch move all internal API usage
to use "tag_index".
Signed-off-by: Michael Baum
Acked-by: Matan Azrad
---
drivers/
From: Qin Ke
Performance is reduced because of adding redundant read configure
operations in nfp_net_parse_ptype() called by nfp_net_recv_pkts(),
which cause every received packet to spend more time.
Fix the performance issue by deleting the redundant code directly.
Fixes: 2e7c36128be ("net/nfp
> -Original Message-
> From: Ivan Malov
> Sent: Tuesday, June 13, 2023 2:38 PM
> To: Zhang, Qi Z
> Cc: tho...@monjalon.net; or...@nvidia.com; david.march...@redhat.com;
> Richardson, Bruce ; jer...@marvell.com;
> ferruh.yi...@amd.com; Mcnamara, John ;
> Zhang, Helin ; techbo...@dpdk.or
> 16/05/2023 17:24, Ciara Power:
> > From: Pablo de Lara
> >
> > When creating a session, get the session ID that
> > defines the fixed session parameters and store it in the private data.
> > When retrieving IMB_JOB's, if their internal session ID matches
> > the one in the private session data,
Good morning,
Today we have our biweekly DPDK Techboard meeting - 8am Pacfici/11am
Eastern/1500h UTC.
A read-only agenda can be found here:
https://annuel.framapad.org/p/r.0c3cc4d1e011214183872a98f6b5c7db
And our login link:
https://meet.jit.si/dpdk
See you soon
Thanks,
Nathan
Nathan C. So
During EAL cleanup stage, all bus devices are cleaned up properly.
Based on this, there is no need to do extra device close process,
which may call the dev ops '*dev->dev_ops->dev_close' twice.
If this dev ops for ntb was called twice, the interrupt handle for
EAL will be disabled twice and will l
From my experience, inline is used with “static” as a compiler hint.
In case of a raw inline, which is used by both insert_new_tcp_item()
and delete_tcp_item() now, it seems you need to provide an external
one to avoid linking error. But I haven’t used it in this way before.
If inline is the root
Replace the maintainers for OcteonTX EP and cnxk DPI DMA drivers.
Signed-off-by: Vamsi Attunuru
---
MAINTAINERS | 7 ++-
1 file changed, 2 insertions(+), 5 deletions(-)
diff --git a/MAINTAINERS b/MAINTAINERS
index 906b31f97c..827733a1aa 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -845,9 +
On Tue, Jun 13, 2023 at 7:56 AM Hu, Jiayu wrote:
> Hi Kumara,
>
> Overall, the patch looks good to me. But you need to update the doc and
> there
> are some minor changes needed to be made. Please see replies inline.
>
>> Thanks Jiayu. I see some linking errors for windows. It was a for
>> ineser
> -Original Message-
> From: beilei.x...@intel.com
> Sent: Tuesday, June 13, 2023 4:30 PM
> To: Wu, Jingjing
> Cc: dev@dpdk.org; Xing, Beilei
> Subject: [PATCH] net/cpfl: release memzone for rxq
>
> From: Beilei Xing
>
> Memzone of split Rx queue is not released, this commit fixes t
Remove redundant function ngbe_dev_setup_link_alarm_handler() and related
codes.
Fixes: b9246b8fa280 ("net/ngbe: support link update")
Cc: sta...@dpdk.org
Signed-off-by: Jiawen Wu
---
drivers/net/ngbe/ngbe_ethdev.c | 28
drivers/net/ngbe/ngbe_ethdev.h | 1 -
2 file
When using no LSC mode, device link status is sometimes get incorrectly.
Introduce hw->lsc to fix this issue.
Fixes: 3d0af7066759 ("net/ngbe: setup PHY link")
Fixes: 3518df5774c7 ("net/ngbe: support device start/stop")
Cc: sta...@dpdk.org
Signed-off-by: Jiawen Wu
---
drivers/net/ngbe/base/ngbe_
In the new firmware, MNG veto bit is set by default for new features. It
causes the PHY configuration not to be reset after LAN reset. So PHY power
will be down after stopping device. The internal PHY then cannot be
initialized successfully at device start. Add setting power on for internal
PHY to
Fix to show tx_broadcast_packets and tx_multicast_packets counters.
Fixes: 8b433d04adc9 ("net/ngbe: support device xstats")
Cc: sta...@dpdk.org
Signed-off-by: Jiawen Wu
---
drivers/net/ngbe/ngbe_ethdev.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ngbe/ngbe_ethdev.c b/driv
Fix to show tx_broadcast_packets and tx_multicast_packets counters.
Fixes: 91fe49c87d76 ("net/txgbe: support device xstats")
Cc: sta...@dpdk.org
Signed-off-by: Jiawen Wu
---
drivers/net/txgbe/txgbe_ethdev.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/txgbe/txgbe_ethdev.c b
Autoneg was always on for 1G speed, fix to set autoneg off.
Fixes: 01c3cf5c85a7 ("net/txgbe: add autoneg control read and write")
Fixes: e4c515a7bc7e ("net/txgbe: add multi-speed link setup")
Cc: sta...@dpdk.org
Signed-off-by: Jiawen Wu
---
drivers/net/txgbe/base/txgbe_hw.c | 14 +
In the new firmware, MNG veto bit is set by default for new features.
It leads to several issues, so driver makes the following changes:
1. Change the way by sending reset command to firmware to LAN reset.
So that MNG domain will not be reset.
2. Change the hardware flush register since the orig
When using no LSC mode, GPIO interrupt can still cause
txgbe_dev_interrupt_delayed_handler() to be called, with
TXGBE_FLAG_NEED_LINK_UPDATE to be set. So add a LSC condition on
intr->mask_misc.
Fixes: 2fc745e6b606 ("net/txgbe: add interrupt operation")
Fixes: d3bb4a04eac1 ("net/txgbe: add SFP hotp
Under heavy traffic, unpluging and pluging fiber optic cables may cause Tx
failure. The reason is that there is still traffic passing through at PCS
VR reset, during txgbe_set_link_to_*. So totally disable Rx and Tx before
PCS VR reset to fix it.
Fixes: 01c3cf5c85a7 ("net/txgbe: add autoneg contro
Fix some bugs for txgbe and ngbe.
Jiawen Wu (9):
net/txgbe: fix Tx failure with fiber hotplug
net/txgbe: fix interrupt enable mask
net/txgbe: fix issues caused by MNG veto bit setting
net/txgbe: fix to set autoneg for 1G speed
net/txgbe: fix device extended stats
net/ngbe: fix device e
The TX descriptor of tunnel packet filled incorrectly due to the MACLEN
is not set.
This patch fixes this issue by setting MACLEN to correctly fill the
TX descriptor.
Fixes: c59faf3fe8fe ("net/i40e: support TSO on tunneling packet")
Cc: sta...@dpdk.org
Signed-off-by: Shiyang He
---
drivers/net
Hi Ferruh,
On Wed, June 7, 2023 at 2:09AM, Ferruh Yigit wrote:
On 5/8/2023 2:21 AM, zhoumin wrote:
Kindly ping.
Hi Min,
Sorry for the delay.
Should we solve this problem by modifying PMDs?
Is it suitable to maintain the RSS configurations for NIC by testpmd in
order to set DCB?
On Thur,
> -Original Message-
> From: Jerin Jacob
> Sent: Wednesday, June 14, 2023 12:09 AM
> To: David Marchand
> Cc: Yan, Zhirun ; jer...@marvell.com; dev@dpdk.org;
> kirankum...@marvell.com; ndabilpu...@marvell.com;
> step...@networkplumber.org; pbhagavat...@marvell.com; Liang, Cunming
> ; Wa
This patch enables Rx timestamp offload on the SSE data path.
It significantly reduces the performance drop when
RTE_ETH_RX_OFFLOAD_TIMESTAMP is enabled.
---
v5: fix CI errors
---
v4: rework avx2 patch based on offload path
---
v3: logging with driver dedicated macro
---
v2: fix compile warning a
This patch enables Rx timestamp offload on the AVX2 data path.
It significantly reduces the performance drop when
RTE_ETH_RX_OFFLOAD_TIMESTAMP is enabled.
---
v5: fix CI errors
---
v4: rework avx2 patch based on offload path
---
v3: logging with driver dedicated macro
---
v2: fix compile warning
This patch enables Rx timestamp offload on the AVX512 data path.
It significantly reduces the performance drop when
RTE_ETH_RX_OFFLOAD_TIMESTAMP is enabled.
Signed-off-by: Zhichao Zeng
---
v5: fix CI errors
---
v4: rework avx2 patch based on offload path
---
v3: logging with driver dedicated ma
This patch enables Rx timestamp offload on the vector data path.
It significantly reduces the performance drop when
RTE_ETH_RX_OFFLOAD_TIMESTAMP is enabled.
---
v5: fix CI errors
---
v4: rework avx2 patch based on offload path
---
v3: logging with driver dedicated macro
---
v2: fix compile warnin
> From: Shiyang He
> Sent: Thursday, June 8, 2023 6:19 PM
> To: dev@dpdk.org
> Cc: Zhou, YidingX ; He, ShiyangX
> ; sta...@dpdk.org; Yang, Qiming
> ; Zhang, Qi Z ; Xing, Beilei
>
> Subject: [PATCH v2] net/ice:fix tunnel packet TX descriptor error
>
> The TX descriptor of tunnel packet filled inc
On Tue, 13 Jun 2023 22:00:25 +0200
Maxime Coquelin wrote:
> >>
> >> If we are going to break the ABI, better to do the right rework directly.
> >> Otherwise
> >> we'll end-up breaking it again next year.
> >
> > With the suggested change, this will not break ABI next year. Any future
> > fun
Hi,
On 6/13/23 19:16, Chautru, Nicolas wrote:
Hi Maxime,
-Original Message-
From: Maxime Coquelin
On 6/12/23 22:53, Chautru, Nicolas wrote:
Hi Maxime, David,
-Original Message-
From: Maxime Coquelin
On 6/6/23 23:01, Chautru, Nicolas wrote:
Hi David,
-Original M
> From: Bruce Richardson [mailto:bruce.richard...@intel.com]
> Sent: Tuesday, 13 June 2023 16.34
>
> As discussed by DPDK technical board [1], out contributor guide should
Typo: out -> our
> include some details as to when to use logging vs tracing vs telemetry
> to provide the end user with inf
Hi Maxime,
> -Original Message-
> From: Maxime Coquelin
>
> On 6/12/23 22:53, Chautru, Nicolas wrote:
> > Hi Maxime, David,
> >
> >> -Original Message-
> >> From: Maxime Coquelin
> >>
> >> On 6/6/23 23:01, Chautru, Nicolas wrote:
> >>> Hi David,
> >>>
> -Original Mess
When libnuma development package is missing on a system, DPDK can still
be built but will be missing much-needed support for NUMA memory
management. This may later cause issues at runtime if the resulting
binary is run on a NUMA system.
We can reduce the incidence of such runtime errors by ensurin
The patch adds tracing capability to Tx datapath.
To engage this tracing capability the following steps
should be taken:
- meson option -Denable_trace_fp=true
- meson option -Dc_args='-DALLOW_EXPERIMENTAL_API'
- EAL command line parameter --trace=pmd.net.mlx5.tx.*
The Tx datapath tracing allows t
The Python script is intended to analyze mlx5 PMD
datapath traces and report:
- tx_burst routine timings
- how packets are pushed to WQEs
- how packet sending is completed with timings
Signed-off-by: Viacheslav Ovsiienko
---
drivers/net/mlx5/tools/mlx5_trace.py | 271 ++
The "save_trace" CLI command is added to trigger
saving the trace dumps to the trace directory.
Signed-off-by: Viacheslav Ovsiienko
---
app/test-pmd/cmdline.c | 38 ++
1 file changed, 38 insertions(+)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline
There is the demand to trace the send completions of
every WQE if time scheduling is enabled.
The patch extends the size of completion queue and
requests completion on every issued WQE in the
send queue. As the result hardware provides CQE on
each completed WQE and driver is able to fetch
completi
The mlx5 provides the send scheduling on specific moment of time,
and for the related kind of applications it would be extremely useful
to have extra debug information - when and how packets were scheduled
and when the actual sending was completed by the NIC hardware (it helps
application to track
When libnuma development package is missing on a system, DPDK can still
be built but will be missing much-needed support for NUMA memory
management. This may later cause issues at runtime if the resulting
binary is run on a NUMA system.
We can reduce the incidence of such runtime errors by ensurin
There is an intention to engage DPDK tracing capabilities
for mlx5 PMDs monitoring and profiling in various modes.
The patch introduces tracepoints for the Tx datapath in
the ethernet device driver.
Signed-off-by: Viacheslav Ovsiienko
---
drivers/common/mlx5/meson.build | 1 +
drivers/common/m
On Tue, 16 Apr 2019 10:59:12 +0900
ogawa.yasuf...@lab.ntt.co.jp wrote:
> From: Yasufumi Ogawa
>
> In secondary_msl_create_walk(), it creates a file for fbarrays with its
> PID for reserving unique name among secondary processes. However, it
> does not work as expected if secondary is run as app
On 6/13/2023 2:24 PM, David Marchand wrote:
> Displaying a bitmask is terrible for users.
> Prefer offload names when refusing some offloads in
> rte_eth_dev_configure.
>
> Before:
> Ethdev port_id=0 requested Rx offloads 0x621 doesn't match Rx offloads
> capabilities 0x0 in rte_eth_dev_conf
On Thu, 10 Dec 2020 23:15:53 +
dheemanth wrote:
> In order to improve performance, the KNI is made to
> support multiple fifos, So that multiple threads pinned
> to multiple cores can process packets in parallel.
>
> Signed-off-by: dheemanth
Rejecting this old patch because there was no fo
Several commands have a -q option but the description in
the documentation was unclear. It is the upper limit on
the number of queues per lcore.
Suggested-by: Vipin Varghese
Signed-off-by: Stephen Hemminger
---
doc/guides/sample_app_ug/ip_frag.rst | 2 +-
doc/guides/sample_app_u
On Tue, Jun 13, 2023 at 8:10 PM David Marchand
wrote:
>
> On Tue, Jun 13, 2023 at 4:12 PM Zhirun Yan wrote:
> >
> > V14:
> > Rebase to main and fix build issues.(align header name in patch 01,02)
>
> And, again, compilation is broken between some patches.
Same here.
Applying: graph: introduce s
On Tue, Jun 13, 2023 at 9:29 PM Slava Ovsiienko wrote:
>
> Hi,
>
> > -Original Message-
> > From: Jerin Jacob
> > Sent: Tuesday, June 13, 2023 6:53 PM
> > To: Slava Ovsiienko
> > Cc: dev@dpdk.org
> > Subject: Re: [RFC 2/5] common/mlx5: introduce tracepoints for mlx5 drivers
> >
> > On Tu
Hi,
> -Original Message-
> From: Jerin Jacob
> Sent: Tuesday, June 13, 2023 6:53 PM
> To: Slava Ovsiienko
> Cc: dev@dpdk.org
> Subject: Re: [RFC 2/5] common/mlx5: introduce tracepoints for mlx5 drivers
>
> On Tue, Jun 13, 2023 at 9:20 PM Slava Ovsiienko
> wrote:
> >
> > Hi,
> >
> > <..
On Tue, Jun 13, 2023 at 9:20 PM Slava Ovsiienko wrote:
>
> Hi,
>
> <..snip..>
> > >
> > > mlx5_os_interrupt_handler_create; # WINDOWS_NO_EXPORT
> > > mlx5_os_interrupt_handler_destroy; # WINDOWS_NO_EXPORT
> > > +
> > > + __rte_pmd_mlx5_trace_tx_entry;
> > > + __rte_pmd_
Hi,
<..snip..>
> >
> > mlx5_os_interrupt_handler_create; # WINDOWS_NO_EXPORT
> > mlx5_os_interrupt_handler_destroy; # WINDOWS_NO_EXPORT
> > +
> > + __rte_pmd_mlx5_trace_tx_entry;
> > + __rte_pmd_mlx5_trace_tx_exit;
> > + __rte_pmd_mlx5_trace_tx_wqe;
> > + __
Currently, in the case when we search for a bit set after a particular
value, the bitmap has to be scanned from the beginning and
rte_bitmap_scan() has to be called multiple times until we hit the value.
Add a new rte_bitmap_scan_from_offset() function to initialize scan
state at the given offset
The has_libfdt flag was using 0 and 1 integer values, instead of the
more appropriate boolean type. Change to use true/false instead.
Signed-off-by: Bruce Richardson
---
config/meson.build | 4 ++--
drivers/net/ipn3ke/meson.build | 2 +-
drivers/raw/ifpga/meson.build | 2 +-
3 files
in meson.build files, some flags were being treated as integers rather
than using the more logical option of booleans. Fix this.
V3: drop the patch requiring libnuma from this set, to make this only
a simple set changing the types of the flag. The other patch will be
submitted separately for
The has_libnuma flag was using 0 and 1 integer values, instead of the
more appropriate boolean type. Change to use true/false instead.
Signed-off-by: Bruce Richardson
Reviewed-by: David Marchand
---
config/arm/meson.build| 2 +-
config/meson.build| 4 ++--
lib/eal/linux/meson.build
On Tue, Jun 13, 2023 at 3:52 PM Akhil Goyal wrote:
>
> Added support for MACsec SC/flow/session stats.
>
> Signed-off-by: Akhil Goyal
> ---
> doc/guides/rel_notes/release_23_07.rst | 5 ++
> drivers/net/cnxk/cn10k_ethdev_sec.c| 11 +++--
> drivers/net/cnxk/cnxk_ethdev_mcs.c | 64 +++
On Tue, 13 Jun 2023 15:33:55 +0100
Bruce Richardson wrote:
> +Below are some guidelines for when each should be used:
> +
> +* For reporting error conditions, or other abnormal runtime issues,
> *logging* should be used.
> + Depending on the severity of the issue, the appropriate log level, for
Extract VLAN TCI provided by the HW in the prefix and put it to mbuf.
VLAN stripping is supported for ef100 datapath only. This is device
level offload.
Signed-off-by: Artemii Morozov
Reviewed-by: Viacheslav Galaktionov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
doc/guides/nics/sfc
To enable VLAN stripping, two conditions must be met:
the corresponding flag must be set and the appropriate
RX prefix should be requested.
VLAN stripping is supported for ef100 datapath only.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
drivers/common/s
These changes are necessary in order to add support for stripping
VLAN tags in the future.
Signed-off-by: Artemii Morozov
Reviewed-by: Ivan Malov
Reviewed-by: Andy Moreton
---
drivers/common/sfc_efx/base/ef10_nic.c | 6 ++
drivers/common/sfc_efx/base/efx.h | 1 +
drivers/common/sfc_
This patch series adds VLAN stripping offload. Note that this
offload are device level offload.
v5:
* fixed problems with naming
* fixed problems with abbreviations
* fixed problems with isolated mode
* fixed problems with consistency
v4:
* fix apply patch failure warning
v3:
* fix a
Add support for HW reassembly of multi-seg pkts.
Also optimize the code for normal reassembly path.
Signed-off-by: Nithin Dabilpuram
---
v3:
- No changes, resend after rebase
v2:
- Removed change id
drivers/event/cnxk/cn10k_worker.h | 21 +-
drivers/net/cnxk/cn10k_rx.h | 714 +++
Acked-by: Gowrishankar Muthukrishnan
Thanks.
> -Original Message-
> From: David Marchand
> Sent: Tuesday, June 13, 2023 1:48 PM
> To: dev@dpdk.org
> Cc: tho...@monjalon.net; bruce.richard...@intel.com; sta...@dpdk.org;
> Brian Dooley ; Gowrishankar Muthukrishnan
>
> Subject: [EXT] [PAT
On Tue, Jun 13, 2023 at 4:12 PM Zhirun Yan wrote:
>
> V14:
> Rebase to main and fix build issues.(align header name in patch 01,02)
And, again, compilation is broken between some patches.
--
David Marchand
As discussed by DPDK technical board [1], out contributor guide should
include some details as to when to use logging vs tracing vs telemetry
to provide the end user with information about the running process and
the DPDK libraries it uses.
[1] https://mails.dpdk.org/archives/dev/2023-March/265204
While the section on dynamic logging in the contributors guide covered
the details of the logging naming scheme, it failed to cover exactly how
the component developer, i.e. the contributor, could actually use
dynamic logging in their component.
Fix this by splitting the details of the naming sche
Add details to our documentation on how to register logging for a component.
Also provide guidelines on when to use logs vs tracing vs telemetry.
Feedback welcome on the guidelines.
Bruce Richardson (2):
doc/contributing: provide coding details for dynamic logging
doc/contributing: guidelines
On Mon, Mar 6, 2023 at 2:19 PM Slava Ovsiienko wrote:
>
> Hi, Stephen
>
> We have another place of using RTE_LOG in mlx5_glue_path().
> Should we update this one as well ?
I think we should.
Stephen can you send a new revision?
--
David Marchand
Add new parameter "model" to choose mcore dispatch or rtc model.
And in dispatch model, the node will affinity to worker core successively.
RTE_GRAPH_MODEL_SELECT is set to RTE_GRAPH_MODEL_RTC by default. Must set
model the same as RTE_GRAPH_MODEL_SELECT if set it as rtc or mcore
dispatch explicit
Add functional test for mcore dispatch model including graph clone,
graph model set/get, node worker affinity, graph worker binding/unbinding.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Pavan Nikhilesh
---
app/test/test_graph.c | 130 ++
Add stats for mcore dispatch model if stats collection is
enabled.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
Acked-by: Pavan Nikhilesh
---
lib/graph/graph_debug.c| 6 ++
lib/graph/graph_stats.c
This patch enables to chose new scheduler model. Must define
RTE_GRAPH_MODEL_SELECT before including rte_graph_worker.h
to enable specific model choosing.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
Acked-by: Pavan Nikhilesh
---
doc
This patch introduces the task scheduler mechanism to enable dispatching
tasks to another worker cores. Currently, there is only a local work
queue for one graph to walk. We introduce a scheduler worker queue in
each worker core for dispatching tasks. It will perform the walk on
scheduler work queu
This patch enables to create and destroy scheduling workqueue into
common graph operations.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Pavan Nikhilesh
---
lib/graph/graph.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/lib/gra
This patch introduces key functions to allow a worker thread to
enable enqueue and move streams of objects to the next nodes over
different cores for mcore dispatch model.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Pavan Nikhilesh
---
lib/graph
Add graph_mcore_dispatch_wq_node to hold graph scheduling workqueue
node.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
---
lib/graph/graph.c | 2 ++
lib/graph/graph_populate.c | 1 +
lib/graph/graph_private.h | 12
This patch adds graph API for supporting to clone the graph object for
a specified worker core. The new graph will also clone all nodes.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
lib/graph/graph.c | 89 +
Move clone_name() into graph_private.h as a common function for both node
and graph to naming a new cloned object.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
lib/graph/graph_private.h | 41 +++
Add lcore_id for graph to hold affinity core id where graph would run on.
Add bind/unbind API to set/unset graph affinity attribute. lcore_id will
be set as MAX by default, it means not enable this attribute.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acke
Add lcore_id for node to hold affinity core id and impl
rte_graph_model_mcore_dispatch_lcore_affinity_set to set node affinity
with specific lcore.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
lib/graph/graph_private.h
Add new get/set APIs to configure graph worker model which is used to
determine which model will be chosen.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
lib/graph/meson.build | 1 +
lib/graph/rte_graph_worker.c
Node process is a single and reusable block, move the code into an inline
function.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
lib/graph/rte_graph_model_rtc.h | 20 ++---
lib/graph/rte_graph_worker_common.h | 33
To support multiple graph worker model, split graph into common
and default. Naming the current walk function as rte_graph_model_rtc
cause the default model is RTC(Run-to-completion).
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
a
Rename rte_graph_work.h to rte_graph_work_common.h for supporting
multiple graph worker model.
Signed-off-by: Haiyue Wang
Signed-off-by: Cunming Liang
Signed-off-by: Zhirun Yan
Acked-by: Jerin Jacob
---
MAINTAINERS | 3 ++-
app/test/test_graph.c
V14:
Rebase to main and fix build issues.(align header name in patch 01,02)
V13:
Add sub header into meson indirect_headers list to export.(change meson.build
in patch02, 05)
V12:
Fix compilation broken at patch 1.(keep renamed header align with patch 1,2)
V11:
Update comments and fix to add ex
Rename sub directories in this example.
This removes the only special case in examples and will ease a next
change.
Signed-off-by: David Marchand
Acked-by: Bruce Richardson
---
doc/guides/sample_app_ug/server_node_efd.rst | 14 +++---
examples/meson.build
Added sources were not referenced in the makefile used when compiling
this example externally.
Fixes: 36128a67c27e ("examples/fips_validation: add asymmetric validation")
Fixes: b455d261eb89 ("examples/fips_validation: validate ECDSA")
Cc: sta...@dpdk.org
Signed-off-by: David Marchand
Acked-by:
Enhance our CI coverage by building examples against an installed DPDK.
Signed-off-by: David Marchand
Acked-by: Bruce Richardson
---
Changes since v1:
- reworked built examples discovery,
- added comment for people who are not sed fluent,
---
.ci/linux-build.sh | 27 ++
So far, users of test-meson-builds.sh had to define their own set of
examples to build externally. This is not that great because users need
to maintain this list when examples are removed/added.
Rework the script so that the 'all' value triggers an automatic
discovery based on what was configured
As DPDK provides examples compiled with makefiles, we need some tests in
the CI. So far, a few maintainers have been testing them but a simple
issue has been missed for some time and there was no way to try to build
all examples that were built with meson.
Additionnally, this series can help in id
Signed-off-by: Ivan Malov
---
.mailmap | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/.mailmap b/.mailmap
index 8e3940a253..d97cdc670c 100644
--- a/.mailmap
+++ b/.mailmap
@@ -105,7 +105,7 @@ Andriy Berestovskyy
Andrzej Ostruszka
Andy Gospodarek
Andy Green
-Andy Mo
On Tue, Jun 13, 2023 at 03:43:34PM +0200, David Marchand wrote:
> On Mon, Jun 12, 2023 at 7:15 PM Bruce Richardson
> wrote:
> >
> > The has_libnuma flag was using 0 and 1 integer values, instead of the
> > more appropriate boolean type. Change to use true/false instead.
> >
> > Signed-off-by: Bruc
On Mon, Jun 12, 2023 at 7:59 PM Bruce Richardson
wrote:
>
> On Mon, Jun 12, 2023 at 06:14:56PM +0100, Bruce Richardson wrote:
> > When libnuma development package is missing on a system, DPDK can still
> > be built but will be missing much-needed support for NUMA memory
> > management. This may la
On Mon, Jun 12, 2023 at 7:15 PM Bruce Richardson
wrote:
>
> The has_libnuma flag was using 0 and 1 integer values, instead of the
> more appropriate boolean type. Change to use true/false instead.
>
> Signed-off-by: Bruce Richardson
This patch lgtm (with or without the suggestion on config/arm b
Displaying a bitmask is terrible for users.
Prefer offload names when refusing some offloads in
rte_eth_dev_configure.
Before:
Ethdev port_id=0 requested Rx offloads 0x621 doesn't match Rx offloads
capabilities 0x0 in rte_eth_dev_configure()
After:
Ethdev port_id=0 does not support Rx off
1 - 100 of 192 matches
Mail list logo
