> Subject: [EXT] [PATCH v2] cryptodev: avoid algorithm strings null
> pointers
>
> The crypto algorithm strings identifiers that are Continuous may be
> null, so there is needed to add null judgment.
> When testing with dpdk-test-crypto-perf and passing in the parameter
> --auth-algo sm3-hmac,
The purpose of the event dispatcher is to decouple different parts of
an application (e.g., processing pipeline stages), sharing the same
underlying event device.
The event dispatcher replaces the conditional logic (often, a switch
statement) that typically follows an event device dequeue operatio
The purpose of the event dispatcher is to help reduce coupling in an
Eventdev-based DPDK application.
In addition, the event dispatcher also provides a convenient and
flexible way for the application to use service cores for
application-level processing.
Signed-off-by: Mattias Rönnblom
Tested-by
Add unit tests for the event dispatcher.
--
RFC v4: Adapt to non-const events in process function prototype.
Signed-off-by: Mattias Rönnblom
---
app/test/meson.build | 1 +
app/test/test_event_dispatcher.c | 814 +++
2 files changed, 815 insertions(+)
Provide programming guide the for the event dispatcher.
Signed-off-by: Mattias Rönnblom
--
RFC v4:
o Extend event matching section of the programming guide.
o Improve grammar and spelling.
---
doc/api/doxy-api-index.md | 1 +
doc/guides/prog_guide/event_dispatcher.rst | 44
Hi David,
> -Original Message-
> From: David Marchand
> Sent: Friday, June 9, 2023 2:58 PM
> To: Zeng, ZhichaoX
> Cc: dev@dpdk.org; Zhang, Qi Z ; Gao, DaxueX
> ; Richardson, Bruce ;
> Konstantin Ananyev ; Wu, Jingjing
> ; Xing, Beilei ; Wu, Wenjun1
>
> Subject: Re: [PATCH] net/iavf: fi
> -Original Message-
> From: liu xixin
> Sent: Friday, June 9, 2023 12:31 PM
> To: Akhil Goyal ; dev@dpdk.org
> Subject: 答复: [EXT] [PATCH v2] cryptodev: avoid algorithm strings null pointers
>
> > Subject: [EXT] [PATCH v2] cryptodev: avoid algorithm strings null
> > pointers
> >
> > The
Please reply only in plain text format.
All the values are added in the array but are jumbled up.
But still it will point to correct position in the array as we are using enum
value.
Eg. RTE_CRYPTO_CIPHER_NULL = 1,
And in the array it is there as highlighted below.
Hence there is no hole present
> Hi, Akhil
>
>
> 在 2023/6/2 11:04, Zhangfei Gao 写道:
> > By default, uadk only alloc two queues for each algorithm, which
> > will impact performance.
> > Set queue pair number as required in dev_configure.
> > The default max queue pair number is 8, which can be modified
> > via para: max_nb_que
David Marchand 于2023年6月8日周四 04:33写道：
>
> On Mon, May 29, 2023 at 1:23 PM Fengnan Chang
> wrote:
> >
> > Under legacy mode, if the number of continuous memsegs greater
> > than RTE_MAX_MEMSEG_PER_LIST, eal init will failed even though
> > another memseg list is empty, because only one memseg list
On 2023/6/8 2:12, Ferruh Yigit wrote:
On 5/29/2023 3:26 AM, Jie Hai wrote:
Each stream has a read-only "disabled" field that control if this
stream should be used to forward. This field depends on states
of Rx/Tx queues, please see
commit 3c4426db54fc ("app/testpmd: do not poll stopped queues"
Use fallthrough comments to ignore compilation warnings instead of
diagnostic pragmas, which are not recognized by old GCC(below GCC7).
Fixes: 95a1b0b23140 ("net/iavf: support Rx timestamp offload on SSE")
Fixes: 4f58266da4d3 ("net/iavf: support Rx timestamp offload on AVX2")
Fixes: d65eadb9e178 (
Here's how the problem arises.
step1: Start the app.
dpdk-testpmd -a :35:00.0 -l 0-3 -- -i --rxq=10 --txq=10
step2: Perform the following steps and send traffic. As expected,
queue 7 does not send or receive packets, and other queues do.
port 0 rxq 7 stop
port 0 txq 7 stop
set
Pexpect is not a dedicated SSH connection library while Fabric is. With
Fabric, all SSH-related logic is provided and we can just focus on
what's DTS specific.
Signed-off-by: Juraj Linkeš
---
Notes:
v3: updated passwordless sudo setup on Linux
doc/guides/tools/dts.rst
On 10/6/2021 12:06 PM, Ferruh Yigit wrote:
> On 10/1/2021 10:07 AM, Andrew Rybchenko wrote:
>> All xstats names may be retrieved passing NULL ids.
>>
>> If a driver does not support getting names by IDs, the callback should
>> return -ENOTSUP on attempt to use it. If so, the request is handled
>> o
在 2023/6/9 16:31, Akhil Goyal 写道:
Hi, Akhil
在 2023/6/2 11:04, Zhangfei Gao 写道:
By default, uadk only alloc two queues for each algorithm, which
will impact performance.
Set queue pair number as required in dev_configure.
The default max queue pair number is 8, which can be modified
via para:
https://bugs.dpdk.org/show_bug.cgi?id=1246
Bug ID: 1246
Summary: [DPDK-23.07]
runtime_vf_queue_number_kernel/set_valid_vf_queue_num_
command_line: failed information when starting testpmd
Product: DPDK
Version: 23
After doing a build, to install DPDK system-wide our documentation
recommended using the "ninja install" command. However, for anyone
building as a non-root user and only installing as root, the "meson
install" command is a better alternative, as it provides for
automatically dropping or elevating
When releasing mbufs for TX queue software ring of AVX512,
the mbuf in the range [i, tx_tail] should also be freed.
The variable i refers to the index of the last mbuf released
in the software ring.
Fixes: 5bf87b45b2c8 ("net/idpf: add AVX512 data path for single queue model")
Cc: sta...@dpdk.org
On 6/9/2023 10:03 AM, Jie Hai wrote:
> Here's how the problem arises.
> step1: Start the app.
> dpdk-testpmd -a :35:00.0 -l 0-3 -- -i --rxq=10 --txq=10
>
> step2: Perform the following steps and send traffic. As expected,
> queue 7 does not send or receive packets, and other queues do.
>
Absence of the CDX bus is not an error.
Bugzilla ID: 1246
Fixes: 45ef232af515 ("bus/cdx: introduce AMD CDX bus")
Signed-off-by: David Marchand
---
drivers/bus/cdx/cdx.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/bus/cdx/cdx.c b/drivers/bus/cdx/cdx.c
index 96
On Fri, Jun 9, 2023 at 12:52 PM Bruce Richardson
wrote:
>
> After doing a build, to install DPDK system-wide our documentation
> recommended using the "ninja install" command. However, for anyone
> building as a non-root user and only installing as root, the "meson
> install" command is a better a
[Public]
Acked-by: Nipun Gupta
> -Original Message-
> From: David Marchand
> Sent: Friday, June 9, 2023 4:41 PM
> To: dev@dpdk.org
> Cc: Gupta, Nipun ; Agarwal, Nikhil
> ; Yigit, Ferruh
> Subject: [PATCH] bus/cdx: silence bus scan when CDX is unavailable
>
> Absence of the CDX bus is n
From: Satheesh Paul
Currently PORT_ID action can redirect traffic only between
VFs of a PF. This patch extends PORT_ID action to redirect
traffic from one PF port to another PF port also.
Signed-off-by: Satheesh Paul
Reviewed-by: Kiran Kumar K
---
drivers/common/cnxk/roc_npc.c | 29 +
On Fri, Jun 09, 2023 at 01:13:15PM +0200, David Marchand wrote:
> On Fri, Jun 9, 2023 at 12:52 PM Bruce Richardson
> wrote:
> >
> > After doing a build, to install DPDK system-wide our documentation
> > recommended using the "ninja install" command. However, for anyone
> > building as a non-root u
On Fri, Jun 9, 2023 at 1:19 PM Bruce Richardson
wrote:
>
> On Fri, Jun 09, 2023 at 01:13:15PM +0200, David Marchand wrote:
> > On Fri, Jun 9, 2023 at 12:52 PM Bruce Richardson
> > wrote:
> > >
> > > After doing a build, to install DPDK system-wide our documentation
> > > recommended using the "ni
Hi,
Thanks for adding the app. Few comments inline. Please check.
Thanks,
Anoob
> -Original Message-
> From: Cheng Jiang
> Sent: Thursday, June 8, 2023 2:14 PM
> To: tho...@monjalon.net; bruce.richard...@intel.com;
> m...@smartsharesystems.com; chenbo@intel.com
> Cc: dev@dpdk.org; j
On Fri, Jun 2, 2023 at 6:22 PM Tyler Retzlaff
wrote:
> On Thu, Jun 01, 2023 at 12:23:39PM +, Zhang, Qi Z wrote:
> > The patchset looks good to me.
> >
> > I have just one question regarding the patch set targets, which include PMD
> > iavf, ice, and ixgbe. However, I noticed that some other I
On Fri, Mar 17, 2023 at 11:35 PM Tyler Retzlaff
wrote:
>
> Adopt rte thread APIs in code built for Windows to decouple it from the
> pthread shim.
>
> Remove most of the pthread_xxx lifetime shim functions, only
> pthread_create remains while we wait for rte_ctrl_thread_create removal.
>
> Tyler R
On Mon, May 29, 2023 at 1:23 PM Fengnan Chang
wrote:
>
> Under legacy mode, if the number of continuous memsegs greater
> than RTE_MAX_MEMSEG_PER_LIST, eal init will failed even though
> another memseg list is empty, because only one memseg list used
> to check in remap_needed_hugepages.
> Fix thi
On Fri, Jun 9, 2023 at 1:14 PM Gupta, Nipun wrote:
> > Absence of the CDX bus is not an error.
> >
> > Bugzilla ID: 1246
> > Fixes: 45ef232af515 ("bus/cdx: introduce AMD CDX bus")
> >
> > Signed-off-by: David Marchand
> Acked-by: Nipun Gupta
Applied.
--
David Marchand
On 6/9/2023 7:00 AM, Chaoyong He wrote:
> The former logic does not consider using several cards with flower
> firmware in one host, this patch series add the support of this
> situation.
>
> Long Wu (3):
> net/nfp: fix control mempool creation failed
> net/nfp: fix representor creation failed
After doing a build, to install DPDK system-wide our documentation
recommended using the "ninja install" command. However, for anyone
building as a non-root user and only installing as root, the "meson
install" command is a better alternative, as it provides for
automatically dropping or elevating
On Thu, Jun 8, 2023 at 5:31 PM Jerin Jacob wrote:
> On Thu, Jun 8, 2023 at 8:55 PM Zhirun Yan wrote:
> >
> > V11:
> > Update comments and fix to add experimental flags for
> > rte_graph_model_is_valid() in patch 04.
> > Update added symbols in alphabetical order in version.map with patch
> > 04
On 6/9/2023 7:19 AM, Chaoyong He wrote:
> The former logic of set IPv6 source/destination address flow action
> can't get the right IPV6 address, a 32 bit big endian value is
> expected while a 8 bit value is provided.
> This caused the offloaded packets don't have the right IPv6 address
> as expec
This small series cleans-up some crypto PMD drivers.
Maxime Coquelin (6):
crypto/qat: fix sym device prototype
crypto/qat: remove security lib presence checks
crypto/dpaa2_sec: remove security lib presence checks
crypto/dpaa_sec: remove security lib presence checks
crypto/ipsec_mb: remov
qat_dev_cmd_param parameter of qat_sym_dev_create() was
wrongly marked as unused, this patch fixes it.
Fixes: fb3b9f492205 ("crypto/qat: rework burst data path")
Cc: sta...@dpdk.org
Signed-off-by: Maxime Coquelin
---
drivers/crypto/qat/qat_sym.c | 2 +-
1 file changed, 1 insertion(+), 1 deletio
Since security library is a required dependency for
building QAT crypto PMD, remove useless #ifdefs around
rte_security APIs calls.
Signed-off-by: Maxime Coquelin
---
drivers/common/qat/qat_qp.c | 2 --
drivers/crypto/qat/dev/qat_crypto_pmd_gen2.c | 3 ---
drivers/crypto/qat/d
Since security library is a required dependency for
building DPAA2 crypto PMD, remove useless #ifdefs around
rte_security APIs calls.
Signed-off-by: Maxime Coquelin
---
drivers/crypto/dpaa2_sec/dpaa2_sec_dpseci.c | 36 ++---
drivers/crypto/dpaa2_sec/dpaa2_sec_priv.h | 9 --
iSince security library is a required dependency for
building DPAA crypto PMD, remove useless #ifdefs around
rte_security APIs calls.
Signed-off-by: Maxime Coquelin
---
drivers/crypto/dpaa_sec/dpaa_sec.c| 26 ++-
drivers/crypto/dpaa_sec/dpaa_sec.h| 9 +---
Since security library is a required dependency for
building ipsec_mb crypto PMD, remove useless check before
security lib headers inclusion.
The patch also takes the opportunity to remove unused
related defines.
Signed-off-by: Maxime Coquelin
---
drivers/crypto/ipsec_mb/ipsec_mb_private.h | 3
qat_sym_debug_log_dump() being stubbed if log level
is less than debug, it is not necessary to check the log
level before calling it.
Signed-off-by: Maxime Coquelin
---
drivers/crypto/qat/dev/qat_crypto_pmd_gen3.c | 9 +--
drivers/crypto/qat/dev/qat_crypto_pmd_gen4.c | 7 +
drivers/cry
On 6/9/2023 2:54 PM, Maxime Coquelin wrote:
qat_sym_debug_log_dump() being stubbed if log level
is less than debug, it is not necessary to check the log
level before calling it.
Signed-off-by: Maxime Coquelin
---
Acked-by: Fan Zhang
> -Original Message-
> From: Cheng Jiang
> Sent: Thursday, June 8, 2023 2:14 PM
> To: tho...@monjalon.net; bruce.richard...@intel.com;
> m...@smartsharesystems.com; chenbo@intel.com
> Cc: dev@dpdk.org; jiayu...@intel.com; xuan.d...@intel.com;
> wenwux...@intel.com; yuanx.w...@intel.c
On 6/9/2023 2:54 PM, Maxime Coquelin wrote:
qat_dev_cmd_param parameter of qat_sym_dev_create() was
wrongly marked as unused, this patch fixes it.
Fixes: fb3b9f492205 ("crypto/qat: rework burst data path")
Cc: sta...@dpdk.org
Signed-off-by: Maxime Coquelin
---
drivers/crypto/qat/qat_sym.c
On 6/9/2023 7:22 AM, Chaoyong He wrote:
> The former logic can't distinguish the TCP and UDP set action, and
> only send a TCP type ctrl message to firmware, which cause the set
> action of UDP can't offload.
>
> Fixes: fc185097bbe6 ("net/nfp: support TP source flow action")
> Fixes: 87986df09d75
On 6/9/2023 7:28 AM, Chaoyong He wrote:
> From: James Hershaw
>
> Incorrect port speed shown after a link state and interrupt of type
> RTE_ETH_EVENT_INTR_LSC occurs. This is because the interrupt handler for
> this type of event in the nfp PMD does not reconfigure the netdev.
>
> Fixed by movin
On 5/15/2023 9:16 AM, Sachin Saxena (OSS) wrote:
> On 5/8/2023 4:11 PM, Tianli Lai wrote:
>> Caution: This is an external email. Please take care when clicking
>> links or opening attachments. When in doubt, report the message using
>> the 'Report this email' button
>>
>>
>> this threshold value ca
On Fri, Jun 9, 2023 at 3:39 PM David Marchand wrote:
>
> On Thu, Jun 8, 2023 at 5:31 PM Jerin Jacob wrote:
> > On Thu, Jun 8, 2023 at 8:55 PM Zhirun Yan wrote:
> > >
> > > V11:
> > > Update comments and fix to add experimental flags for
> > > rte_graph_model_is_valid() in patch 04.
> > > Update
On Fri, 9 Jun 2023 13:41:56 +0800
Zhichao Zeng wrote:
> The code uses the Wimplicit-fallthrough compile option to ignore
> falls through warnings in switch case, but this option was introduced
> by GCC 7 and old GCC do not recognize the option.
>
> Add judgment to avoid warnings about unrecogni
After doing a build, to install DPDK system-wide our documentation
recommended using the "ninja install" command. However, for anyone
building as a non-root user and only installing as root, the "meson
install" command is a better alternative, as it provides for
automatically dropping or elevating
On Tue, Jun 6, 2023 at 11:45 PM Tyler Retzlaff
wrote:
>
> Replace the use of rte_atomic.h types and functions, instead use GCC
> supplied C++11 memory model builtins.
>
> This series covers the libraries and drivers that are built on Windows.
>
> The code has be converted to use the __atomic built
Acked-by: Hemant Agrawal
Acked-by: Hemant Agrawal
On Fri, Jun 09, 2023 at 05:01:53PM +0200, David Marchand wrote:
> On Tue, Jun 6, 2023 at 11:45 PM Tyler Retzlaff
> wrote:
> >
> > Replace the use of rte_atomic.h types and functions, instead use GCC
> > supplied C++11 memory model builtins.
> >
> > This series covers the libraries and drivers that
If DPDK application is given a bogus option, the error message
would get printed twice. Once during scan for log level and
again during parsing of arguments.
Example:
# ./build/app/dpdk-testpmd --bogus
./build/app/dpdk-testpmd: unrecognized option '--bogus'
EAL: Detected CP
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
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
The "dump_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 | 6 +-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/app/test-pmd/cmdline.c b/app/test-pmd/cmdline.c
index 7b20bef4e9
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
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 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 ++
On 6/9/2023 2:43 PM, Ferruh Yigit wrote:
> On 6/9/2023 7:19 AM, Chaoyong He wrote:
>> The former logic of set IPv6 source/destination address flow action
>> can't get the right IPV6 address, a 32 bit big endian value is
>> expected while a 8 bit value is provided.
>> This caused the offloaded packe
On Fri, Jun 09, 2023 at 03:03:25PM +0100, Zhang, Fan wrote:
>
> On 6/9/2023 2:54 PM, Maxime Coquelin wrote:
> >qat_dev_cmd_param parameter of qat_sym_dev_create() was
> >wrongly marked as unused, this patch fixes it.
> >
> >Fixes: fb3b9f492205 ("crypto/qat: rework burst data path")
> >Cc: sta...@d
> -Original Message-
> From: David Marchand
> Sent: Friday, June 9, 2023 10:37 PM
> To: Jerin Jacob
> Cc: Yan, Zhirun ; Thomas Monjalon
> ; dev@dpdk.org; jer...@marvell.com;
> kirankum...@marvell.com; ndabilpu...@marvell.com;
> step...@networkplumber.org; pbhagavat...@marvell.com; Liang
This patch is no long relevant since the current DPDK code
no longer depends on VA width. It should be rejected.
This patch is old, no longer applies and is incomplete.
There are no new tests, and no useful examples for using it.
It should be rejected.
This patch is old, no longer applies and is incomplete.
There are no new tests, and no useful examples for using it.
It should be rejected.
This patch proposed an interesting extension to DPDK, but never
got any positive reviews. It creates new API's which introduces
likelihood of new bugs; and overlaps existing functionality.
Therefore recommend that it be marked rejected.
This patch is old, and was covered by later ring changes.
It should be marked as superseded or rejected in patchwork
Unifying hotplug and device event support would be good idea
but this patch has gone stale. And supporting it would require
API changes (need to wait for next LTS) and a lot of testing.
Also, hotplug is a complex topic which has a lot of platform
dependencies. If someone wants to do a more genera
Eliminate non-burst enqueue operation from Eventdev.
The effect of this change is to reduce Eventdev code complexity
somewhat and slightly improve performance.
The single-event enqueue shortcut provided a very minor performance
advantage in some situations (e.g., with a compile time-constant burs
On 2023-06-09 16:34, Stephen Hemminger wrote:
On Fri, 9 Jun 2023 09:08:24 +0200
Mattias Rönnblom wrote:
+#define RED_MAX_PORTS_PER_LCORE 4
+#define RED_MAX_HANDLERS 32
+#define RED_MAX_FINALIZERS 16
+#define RED_AVG_PRIO_INTERVAL 2000
RED in networking usually means something else.
Is there
On 6/7/2023 6:57 PM, Kaijun Zeng wrote:
> Improve error handling
>
> Bugzilla ID: 1239
> Fixes: dfaff37fc46d ("vmxnet3: import new vmxnet3 poll mode driver
> implementation")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Kaijun Zeng
>
Reviewed-by: Ferruh Yigit
Applied to dpdk-next-net/main, thank
V11:
Update comments and fix to add experimental flags for
rte_graph_model_is_valid() in patch 04.
Update added symbols in alphabetical order in version.map with patch
04,05,06,08,10.
Update commit message in patch 16.
V10:
Add rte_graph_worker_model_no_check_get() for fast path, extract
rte_gr
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
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
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
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
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 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
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 +++
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 +
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 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
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 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 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
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
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 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
91 matches
Mail list logo
