On Fri, Mar 8, 2024 at 12:39 AM Stephen Hemminger
wrote:
>
> From: Isaac Boukris
>
> If request to set promiscious mode failed at startup,
> then it is not necessary to disable it when shutting down.
>
> This should only be issue if with a buggy driver because if
> driver does not support setting
Acked-by: Chengwen Feng
On 2024/3/8 13:37, Wenwu Ma wrote:
> The structure rte_dma_dev needs cacheline alignment, but the return
> value of malloc may not be aligned to the cacheline. Therefore,
> extra memory is applied for realignment.
>
> Fixes: b36970f2e13e ("dmadev: introduce DMA device lib
The structure rte_dma_dev needs cacheline alignment, but the return
value of malloc may not be aligned to the cacheline. Therefore,
extra memory is applied for realignment.
Fixes: b36970f2e13e ("dmadev: introduce DMA device library")
Cc: sta...@dpdk.org
Signed-off-by: Wenwu Ma
---
lib/dmadev/rt
When creating the drop action Devx object, the global counter set
is also used as in the regular or hairpin queue creation.
The drop action should be destroyed before the global counter set
release procedure. Or else, the counter set object is still
referenced and cannot be released successfully.
From: Zerun Fu
The previous logic did not distinguish the firmware type when
reading the value of "app_cap" during the initialization of the
PF. This operation is not supported by the flower firmware, and
this will lead to a startup failure.
Fix this by adding the logic to check the firmware type
- printf("\nTotal Bandwidth: %.3lf Gbps, Total MOps: %.3lf\n",
bandwidth_total, mops_total);
+ printf("\nAverage Cycles/op per worker: %.1f, Total Bandwidth: %.3lf Gbps,
Total MOps: %.3lf\n",
+ (avg_cycles_total * (float) 1.0) / nb_workers,
bandwidth_total, mops_to
https://bugs.dpdk.org/show_bug.cgi?id=1387
jiang,yu (yux.ji...@intel.com) changed:
What|Removed |Added
Status|CONFIRMED |RESOLVED
Resolution|---
Hi Vipin,
On 2024/3/6 23:06, Vipin Varghese wrote:
> Modify the user display data with total average latency per worker.
>
> Signed-off-by: Vipin Varghese
>
> V2 Changes:
> - keep one-digit precision is enough, suggested by Chengwen
> ---
> app/test-dma-perf/benchmark.c | 3 ++-
> 1 file chan
Acked-by: Chengwen Feng
On 2024/3/8 0:14, David Marchand wrote:
> All types exposed in a public header must be prefixed with rte_.
>
> Fixes: e3e579f5bab5 ("argparse: introduce argparse library")
>
> Signed-off-by: David Marchand
> ---
> lib/argparse/rte_argparse.h | 4 ++--
> 1 file changed,
From: Isaac Boukris
If request to set promiscious mode failed at startup,
then it is not necessary to disable it when shutting down.
This should only be issue if with a buggy driver because if
driver does not support setting promiscious it would just
ignore request to disable it as well.
Fixes:
A new DPDK release candidate is ready for testing:
https://git.dpdk.org/dpdk/tag/?id=v24.03-rc2
There are 276 new patches in this snapshot.
Release notes:
https://doc.dpdk.org/guides/rel_notes/release_24_03.html
There were a lot of updates in drivers.
The driver features should b
On Thu, Mar 07, 2024 at 08:53:49PM +, Bruce Richardson wrote:
> On Thu, Mar 07, 2024 at 10:34:42AM -0800, Tyler Retzlaff wrote:
> > * Initialize const int NS_PER_SEC with an integer literal instead of
> > double thereby avoiding implicit conversion from double to int.
> >
> > * Cast the resu
On Thu, Mar 07, 2024 at 10:34:42AM -0800, Tyler Retzlaff wrote:
> * Initialize const int NS_PER_SEC with an integer literal instead of
> double thereby avoiding implicit conversion from double to int.
>
> * Cast the result of the expression assigned to timespec.tv_nsec to long.
>
> Signed-off-b
Test compresses and decompresses pointers with various combinations
of memory regions and alignments and verifies the pointers are
recovered correctly.
Signed-off-by: Paul Szczepanek
Reviewed-by: Honnappa Nagarahalli
Reviewed-by: Nathan Brown
---
app/test/meson.build | 1 +
app/test/
Documentation added in the prog guide for the new
utility functions for pointer compression
showing example code and potential usecases.
Signed-off-by: Paul Szczepanek
Reviewed-by: Honnappa Nagarahalli
Reviewed-by: Nathan Brown
---
doc/guides/prog_guide/ptr_compress_lib.rst | 144 +
Add a test that runs a zero copy burst enqueue and dequeue on a ring
of raw pointers and compressed pointers at different burst sizes to
showcase performance benefits of newly added pointer compression APIs.
Refactored threading code to pass more parameters to threads to
reuse existing code. Added
Add a new utility header for compressing pointers. The provided
functions can store pointers in 32-bit or 16-bit offsets.
The compression takes advantage of the fact that pointers are
usually located in a limited memory region (like a mempool).
We can compress them by converting them to offsets fr
This patchset is proposing adding a new EAL header with utility functions
that allow compression of arrays of pointers.
When passing caches full of pointers between threads, memory containing
the pointers is copied multiple times which is especially costly between
cores. A compression method will
On Thu, 7 Mar 2024 20:50:26 +0100
Mattias Rönnblom wrote:
> On 2024-03-05 22:02, Tyler Retzlaff wrote:
> > On Tue, Mar 05, 2024 at 09:18:20PM +0100, Mattias Rönnblom wrote:
> >> Shouldn't we have a DPDK-native mutex API, rather than using direct
> >> POSIX mutex lock calls?
> >
> > David rai
On Thu, 7 Mar 2024 18:59:11 +0100
David Marchand wrote:
> On Thu, Mar 7, 2024 at 6:37 PM Tyler Retzlaff
> wrote:
> > On Wed, Mar 06, 2024 at 05:36:42PM -0800, Stephen Hemminger wrote:
> > > Make the GFNI stub functions always built. This solves the conditional
> > > linking problem. If GFNI is
Make the GFNI stub functions always built. This solves the conditional
linking problem. If GFNI is available, they will never get used.
Fixes: 07d836e5929d ("hash: uninline GFNI stubs")
Signed-off-by: Stephen Hemminger
---
v4 - fix build on aarch64 and __rte_internal
lib/hash/rte_thash_gfni.c |
On 2024-03-05 22:02, Tyler Retzlaff wrote:
On Tue, Mar 05, 2024 at 09:18:20PM +0100, Mattias Rönnblom wrote:
Shouldn't we have a DPDK-native mutex API, rather than using direct
POSIX mutex lock calls?
David raised this a while back and the consensus is yes. I admit it's
been on my radar for a
> From: Tyler Retzlaff [mailto:roret...@linux.microsoft.com]
> Sent: Thursday, 7 March 2024 19.37
>
> On Thu, Mar 07, 2024 at 07:06:26PM +0100, David Marchand wrote:
> > On Thu, Mar 7, 2024 at 3:16 PM David Marchand
> wrote:
> > >
> > > Busy looping on RTE_MAX_LCORES threads is too heavy in some
On Thu, Mar 07, 2024 at 06:58:20PM +0100, David Marchand wrote:
> On Thu, Mar 7, 2024 at 6:43 PM Tyler Retzlaff
> wrote:
> >
> > On Thu, Mar 07, 2024 at 01:07:40PM +, Chengwen Feng wrote:
> > > The enum rte_argparse_flag's value is u64, but an enum in C is
> > > represented as an int. This com
On Thu, Mar 07, 2024 at 07:06:26PM +0100, David Marchand wrote:
> On Thu, Mar 7, 2024 at 3:16 PM David Marchand
> wrote:
> >
> > Busy looping on RTE_MAX_LCORES threads is too heavy in some CI or build
> > systems running the fast-test testsuite.
> > Ask for a reschedule at the threads synchronisa
* Initialize const int NS_PER_SEC with an integer literal instead of
double thereby avoiding implicit conversion from double to int.
* Cast the result of the expression assigned to timespec.tv_nsec to long.
Signed-off-by: Tyler Retzlaff
Acked-by: Dmitry Kozlyuk
---
v2:
* update commit mess
On Thu, Mar 7, 2024 at 3:16 PM David Marchand wrote:
>
> Busy looping on RTE_MAX_LCORES threads is too heavy in some CI or build
> systems running the fast-test testsuite.
> Ask for a reschedule at the threads synchronisation points.
>
> Signed-off-by: David Marchand
> Acked-by: Luca Boccassi
I
On Thu, Mar 7, 2024 at 6:37 PM Tyler Retzlaff
wrote:
> On Wed, Mar 06, 2024 at 05:36:42PM -0800, Stephen Hemminger wrote:
> > Make the GFNI stub functions always built. This solves the conditional
> > linking problem. If GFNI is available, they will never get used.
> >
> > Fixes: 07d836e5929d ("ha
On Thu, Mar 7, 2024 at 6:43 PM Tyler Retzlaff
wrote:
>
> On Thu, Mar 07, 2024 at 01:07:40PM +, Chengwen Feng wrote:
> > The enum rte_argparse_flag's value is u64, but an enum in C is
> > represented as an int. This commit replace these enum values with
> > macro.
> >
> > Fixes: e3e579f5bab5 ("
On Thu, Mar 07, 2024 at 05:14:02PM +0100, David Marchand wrote:
> All types exposed in a public header must be prefixed with rte_.
>
> Fixes: e3e579f5bab5 ("argparse: introduce argparse library")
>
> Signed-off-by: David Marchand
> ---
Acked-by: Tyler Retzlaff
On Thu, Mar 07, 2024 at 01:07:40PM +, Chengwen Feng wrote:
> The enum rte_argparse_flag's value is u64, but an enum in C is
> represented as an int. This commit replace these enum values with
> macro.
>
> Fixes: e3e579f5bab5 ("argparse: introduce argparse library")
> Fixes: 5357c248c960 ("argp
On Wed, Mar 06, 2024 at 05:36:42PM -0800, Stephen Hemminger wrote:
> Make the GFNI stub functions always built. This solves the conditional
> linking problem. If GFNI is available, they will never get used.
>
> Fixes: 07d836e5929d ("hash: uninline GFNI stubs")
> Signed-off-by: Stephen Hemminger
>
On Thu, 7 Mar 2024 15:16:06 +0100
David Marchand wrote:
> +#ifndef _POSIX_PRIORITY_SCHEDULING
> +/* sched_yield(2):
> + * POSIX systems on which sched_yield() is available define _POSIX_PRIOR‐
> + * ITY_SCHEDULING in .
> + */
> +#define sched_yield()
> +#endif
Could you fix the awkward line br
On Mon, Mar 4, 2024 at 10:22 AM Aaron Conole wrote:
>
> zhoumin writes:
>
> > On Wed, Feb 21, 2024 at 2:24AM, Patrick Robb wrote:
> >
> > On Tue, Feb 20, 2024 at 1:12 PM Aaron Conole wrote:
> >
> > Why not something like:
> >
> > Recheck-request: [attribute-list],[test-list]...
> >
> > For e
On Thu, 7 Mar 2024 15:16:06 +0100
David Marchand wrote:
> Busy looping on RTE_MAX_LCORES threads is too heavy in some CI or build
> systems running the fast-test testsuite.
> Ask for a reschedule at the threads synchronisation points.
>
> Signed-off-by: David Marchand
> Acked-by: Luca Boccassi
On Thu, 7 Mar 2024 12:33:24 +0100
David Marchand wrote:
> Busy looping on RTE_MAX_LCORES threads is too heavy in some CI
> environments running the fast-test testsuite.
> Ask for a reschedule at the threads synchronisation points.
>
> Signed-off-by: David Marchand
You could use rte_pause here
On Thu, 7 Mar 2024 10:25:48 +
Ferruh Yigit wrote:
> > I got 4 queues setup, but looks like they are trash in secondary.
> > Probably best to revert this and fix it by bumping RTE_MP_MAX_FD_NUM.
> > This is better, but does take some ABI issue handling.
> >
>
> We can increase RTE_MP_MAX_FD
On Thu, 7 Mar 2024 11:32:36 +0100
David Marchand wrote:
> On Tue, Mar 5, 2024 at 6:53 PM Tyler Retzlaff
> wrote:
> >
> > On Tue, Mar 05, 2024 at 11:14:45AM +0100, David Marchand wrote:
> > > On Mon, Mar 4, 2024 at 7:45 PM Stephen Hemminger
> > > wrote:
> > > >
> > > > This reverts commit 07
>
> If the correct version is being installed and picked up, maybe there is
> something missing for that function definition in arm-ipsec-mb repo.
> Wathsala, can you check that please?
We are working on reproducing this issue. Will update asap.
All types exposed in a public header must be prefixed with rte_.
Fixes: e3e579f5bab5 ("argparse: introduce argparse library")
Signed-off-by: David Marchand
---
lib/argparse/rte_argparse.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/argparse/rte_argparse.h b/lib/a
- printf("Error: Source or destination numa exceeds the acture numa
nodes.\n");
+ printf("Error: %s numa exceeds the available numa nodes.\n",
+ (cfg->src_numa_node >= nr_sockets) ? "Source" :
"Destination");
Thank you for comments, pleas
- The size of the memory footprint.
+ The size of the memory footprint in megabytes (MB) for source and
destination.
I prefer not add "for source and destination", it makes sense without this, and
maybe future test like fill don't have source buffer.
The current dma-perf application, mak
07/03/2024 15:03, Honnappa Nagarahalli:
>
> > On Mar 7, 2024, at 12:02 AM, Anoob Joseph wrote:
> >
> > Move 'soc_cn9k' above 'soc_cn10k' to have the list alphabetically
> > sorted.
> >
> > Signed-off-by: Anoob Joseph
> Thank you
>
> Reviewed-by: Honnappa Nagarahalli
Applied, thanks.
07/03/2024 10:40, sk...@marvell.com:
> From: Sunil Kumar Kori
>
> Use the dpdk-cmdline-gen script to autogenerate all the boilerplate
> structs and defines for the commandline part of the application.
>
> Signed-off-by: Sunil Kumar Kori
> Acked-by: Bruce Richardson
> Acked-by: Nithin Dabilpura
07/03/2024 11:14, Konstantin Ananyev:
>
> > When configuring RX queues, the default port configuration was used,
> > even though it was modified before. This results in the
> > 'relax-rx-offload' not being respected for RX queues.
> > This commit uses 'rte_eth_dev_conf_get()' to obtain the device
10/01/2024 15:42, jspew...@iol.unh.edu:
> dts/framework/config/conf_yaml_schema.json| 3 +-
> dts/framework/exception.py| 7 +
> dts/framework/remote_session/testpmd_shell.py | 149 +-
> dts/framework/test_suite.py | 15 +-
> dts/frame
> Juraj Linkeš (7):
> dts: convert dts.py methods to class
> dts: move test suite execution logic to DTSRunner
> dts: filter test suites in executions
> dts: reorganize test result
> dts: block all test cases when earlier setup fails
> dts: refactor logging configuration
> dts: improv
On 07/03/2024 10:36, David Marchand wrote:
> As the vhost library may receive a request for an unsupported request,
> it is necessary to check msg_handler before checking if locking queue
> pairs is requested.
>
> Coverity issue: 415049
> Fixes: 5e8fcc60b59d ("vhost: enhance virtqueue access lock
Busy looping on RTE_MAX_LCORES threads is too heavy in some CI or build
systems running the fast-test testsuite.
Ask for a reschedule at the threads synchronisation points.
Signed-off-by: David Marchand
Acked-by: Luca Boccassi
---
Changes since v1:
- fix build with mingw,
---
app/test/test_lco
On 16/02/2024 12:02, Kamil Vojanec wrote:
> When configuring RX queues, the default port configuration was used,
> even though it was modified before. This results in the
> 'relax-rx-offload' not being respected for RX queues.
> This commit uses 'rte_eth_dev_conf_get()' to obtain the device
> confi
> On Mar 7, 2024, at 12:02 AM, Anoob Joseph wrote:
>
> Move 'soc_cn9k' above 'soc_cn10k' to have the list alphabetically
> sorted.
>
> Signed-off-by: Anoob Joseph
Thank you
Reviewed-by: Honnappa Nagarahalli
> ---
> config/arm/meson.build | 20 ++--
> 1 file changed, 10 inser
Hi Fengchengwen,
> > Waiting for a confirmation that this series is good to go.
>
> In the discuss of thread [1], I hope this patchset continue take a step
> forward
> (means new version) to support bi-direction test just by modify config.ini
> file.
>
This patch set already exposes all confi
Hi Thomas,
> 29/02/2024 14:48, Gowrishankar Muthukrishnan:
> > Improve dma-perf application to support PCI dev and SG copy, along
> > with additional supports as below:
> > - validate copied memory
> > - skip tests if not opted.
> >
> > v11:
> > - Review suggestions.
> >
> > Gowrishankar Muthukr
Hi Thomas and Gowrishankar,
On 2024/3/7 3:50, Thomas Monjalon wrote:
> 29/02/2024 14:48, Gowrishankar Muthukrishnan:
>> Improve dma-perf application to support PCI dev and SG copy,
>> along with additional supports as below:
>> - validate copied memory
>> - skip tests if not opted.
>>
>> v11:
>>
Hi Amit,
On 2024/3/1 18:59, Amit Prakash Shukla wrote:
> Hi Chengwen,
>
> Please find my reply in-line.
>
> Thanks,
> Amit Shukla
>
>> Hi Amit,
>>
>> On 2024/3/1 16:31, Amit Prakash Shukla wrote:
>>> Hi Chengwen,
>>>
>>> If I'm not wrong, your concern was about config file additions and not
>>>
Hi Vipin,
On 2023/12/12 18:45, Vipin Varghese wrote:
> For configuration parameters `mem_size` and `buf_size` are represented
> as megabytes and bytes respectively in application. Update the
> documentation to represent the same.
>
> Signed-off-by: Vipin Varghese
> ---
> doc/guides/tools/dmaper
Hi Vipin,
On 2024/3/6 23:02, Vipin Varghese wrote:
> Current commit decalres either `source or destination numa is greater
> than acture numa` as cause of error. Rephrase as `Source or Destination`
> is incorrect numa by checking which is greater than available numa.
>
> Signed-off-by: Vipin Varg
> From: Thomas Monjalon [mailto:tho...@monjalon.net]
> Sent: Thursday, 7 March 2024 14.13
>
> 07/03/2024 11:44, Ferruh Yigit:
> > On 3/6/2024 8:48 PM, Thomas Monjalon wrote:
> > > 26/01/2024 09:57, Chengwen Feng:
> > >> Add support scatter gather copy.
> > >>
> > >> Signed-off-by: Chengwen Feng
>
Hi David,
On 2024/3/7 1:34, David Marchand wrote:
> On Tue, Feb 20, 2024 at 2:16 PM Chengwen Feng wrote:
>>
>> The struct rte_argparse_arg's flags was 64bit type, uint64_t should be
>> used instead of uint32_t where the operation happened.
>
> Something is strange.
> An enum in C is represented
07/03/2024 11:44, Ferruh Yigit:
> On 3/6/2024 8:48 PM, Thomas Monjalon wrote:
> > 26/01/2024 09:57, Chengwen Feng:
> >> Add support scatter gather copy.
> >>
> >> Signed-off-by: Chengwen Feng
> >> ---
> >> --- a/drivers/dma/skeleton/skeleton_dmadev.c
> >> +++ b/drivers/dma/skeleton/skeleton_dmadev
This patch refines the error message.
Signed-off-by: Chengwen Feng
---
lib/argparse/rte_argparse.c | 31 +--
1 file changed, 17 insertions(+), 14 deletions(-)
diff --git a/lib/argparse/rte_argparse.c b/lib/argparse/rte_argparse.c
index 2d953f1694..48738cd07b 100644
-
The enum rte_argparse_flag's value is u64, but an enum in C is
represented as an int. This commit replace these enum values with
macro.
Fixes: e3e579f5bab5 ("argparse: introduce argparse library")
Fixes: 5357c248c960 ("argparse: parse unsigned integers")
Signed-off-by: Chengwen Feng
---
lib/arg
Refine testcases, including:
1. add testcase comment.
2. argv[0] should set obj->prog_name.
3. set val_set as NULL in test_argparse_invalid_arg_flags, let it
test to the specified code logic.
4. enable index verification in opt_callback_parse_int_of_no_val.
Fixes: 6c5c6571601c ("argparse: verify a
The struct rte_argparse_arg's flags was 64bit type, uint64_t should be
used instead of uint32_t where the operation happened.
Also, the flags' bit16 was also unused, so don't test bit16 in testcase
test_argparse_invalid_arg_flags.
Fixes: 6c5c6571601c ("argparse: verify argument config")
Fixes: 31
I found a couple of issues when I revisited the argparse_autotest
output, so got this patchset.
Chengwen Feng (5):
argparse: refine error message
argparse: remove dead code
argparse: replace flag enum with marco
argparse: fix argument flags operate as uint32 type
test/argparse: refine te
The judgement "obj->callback == NULL" is dead code which can't be
reached, because verify_arg_saver() already make sure obj->callback
must not be NULL when arg->val_saver is NULL.
Signed-off-by: Chengwen Feng
---
app/test/test_argparse.c| 7 ---
lib/argparse/rte_argparse.c | 6 --
2
> -Original Message-
> From: Bing Zhao
> Sent: Thursday, March 7, 2024 8:14 AM
> To: Slava Ovsiienko ; dev@dpdk.org; Raslan
> Darawsheh
> Cc: Ori Kam ; Dariusz Sosnowski
> ; Suanming Mou ; Matan
> Azrad ; Eli Britstein ; Ophir Munk
> ; sta...@dpdk.org
> Subject: [PATCH] net/mlx5: prevent
On Thu, 7 Mar 2024 at 11:33, David Marchand wrote:
>
> Busy looping on RTE_MAX_LCORES threads is too heavy in some CI
> environments running the fast-test testsuite.
> Ask for a reschedule at the threads synchronisation points.
>
> Signed-off-by: David Marchand
> ---
> Note: this is a quick patch
On Thu, Mar 07, 2024 at 11:46:51AM +0100, David Marchand wrote:
> On Wed, Mar 6, 2024 at 11:17 PM Paul Szczepanek
> wrote:
> >
> > Allow header only libraries.
> >
> > Signed-off-by: Paul Szczepanek
> > Reviewed-by: Dhruv Tripathi
>
> Don't forget to Cc: Bruce, he is the build system maintaine
Busy looping on RTE_MAX_LCORES threads is too heavy in some CI
environments running the fast-test testsuite.
Ask for a reschedule at the threads synchronisation points.
Signed-off-by: David Marchand
---
Note: this is a quick patch with no validation, except that it runs fine
on my laptop.
Luca, c
On Fri, Mar 1, 2024 at 11:21 AM Paul Szczepanek wrote:
>
> Test compresses and decompresses pointers with various combinations
> of memory regions and alignments and verifies the pointers are
> recovered correctly.
>
> Signed-off-by: Paul Szczepanek
> ---
> app/test/meson.build | 1
On Fri, Mar 1, 2024 at 11:21 AM Paul Szczepanek wrote:
>
> Add a test that runs a zero copy burst enqueue and dequeue on a ring
> of raw pointers and compressed pointers at different burst sizes to
> showcase performance benefits of newly added pointer compression APIs.
>
> Refactored threading co
Hello Paul,
Before sending the v8, please fix small issues I found.
On Fri, Mar 1, 2024 at 11:21 AM Paul Szczepanek wrote:
>
> Add a new utility header for compressing pointers. The provided
> functions can store pointers in 32-bit offsets.
>
> The compression takes advantage of the fact that p
Hello,
On Thu, Mar 7, 2024 at 2:37 AM Stephen Hemminger
wrote:
>
> Make the GFNI stub functions always built. This solves the conditional
> linking problem. If GFNI is available, they will never get used.
>
> Fixes: 07d836e5929d ("hash: uninline GFNI stubs")
> Signed-off-by: Stephen Hemminger
>
01/03/2024 18:43, Jeremy Spewock:
> On Fri, Mar 1, 2024 at 12:41 PM Jeremy Spewock wrote:
> >
> > For the whole series:
> >
> > Reviewed-by: Jeremy Spewock
> >
> Apologies, I forgot to remove the content of the patch when I added my tag.
The best is to keep the commit message,
and put your tag b
21/02/2024 04:40, Patrick Robb:
> Tested-by: Patrick Robb
>
> The approach for mounting in ssh keys and managing the test engine
> environment has been agreed upon by the CI group, and this patch has been
> tested by a few people at the Community Lab. I believe it is ready to be
> merged.
Applie
On Wed, Mar 6, 2024 at 11:17 PM Paul Szczepanek wrote:
>
> Allow header only libraries.
>
> Signed-off-by: Paul Szczepanek
> Reviewed-by: Dhruv Tripathi
Don't forget to Cc: Bruce, he is the build system maintainer.
>
> ---
> lib/meson.build | 176 +
On 3/6/2024 8:48 PM, Thomas Monjalon wrote:
> 26/01/2024 09:57, Chengwen Feng:
>> Add support scatter gather copy.
>>
>> Signed-off-by: Chengwen Feng
>> ---
>> --- a/drivers/dma/skeleton/skeleton_dmadev.c
>> +++ b/drivers/dma/skeleton/skeleton_dmadev.c
>> @@ -1,5 +1,5 @@
>> /* SPDX-License-Identi
As the vhost library may receive a request for an unsupported request,
it is necessary to check msg_handler before checking if locking queue
pairs is requested.
Coverity issue: 415049
Fixes: 5e8fcc60b59d ("vhost: enhance virtqueue access lock asserts")
Signed-off-by: David Marchand
---
lib/vhos
For vDPA devices, vq are not locked once the device has been configured
at runtime.
On the other hand, we need to hold the vq lock to evaluate vq->access_ok,
invalidate vring addresses and translate them.
Move vring address update earlier and, when vDPA is configured, skip parts
which expect lock
On Tue, Mar 5, 2024 at 6:53 PM Tyler Retzlaff
wrote:
>
> On Tue, Mar 05, 2024 at 11:14:45AM +0100, David Marchand wrote:
> > On Mon, Mar 4, 2024 at 7:45 PM Stephen Hemminger
> > wrote:
> > >
> > > This reverts commit 07d836e5929d18ad6640ebae90dd2f81a2cafb71.
> > >
> > > Tyler found build issues w
Testpmd calls the same function to create legacy indirect action and
indirect list action.
The function did not identify required action correctly.
The patch adds the `indirect_list` boolean function parameter that is
derived from the action type.
Fixes: ee752a734d14 ("app/testpmd: add INDIRECT_L
On 3/6/2024 8:21 PM, Stephen Hemminger wrote:
> On Wed, 6 Mar 2024 16:14:51 +
> Ferruh Yigit wrote:
>
>> On 2/29/2024 5:56 PM, Stephen Hemminger wrote:
>>> The tap device needs to exchange file descriptors for tx and rx.
>>> But the EAL MP layer has limit of 8 file descriptors per message.
>>
MLX5 PMD driver supports 2 types of indirect actions:
legacy INDIRECT and INDIRECT_LIST.
PMD has different handlers for each of indirection actions types.
Therefore PMD marks async `job::indirect_type` with relevant value.
PMD set the type only during indirect action creation.
Legacy INDIRECT quer
Synchronous calls for meter ASO try to pull pending completions
from CQ, submit WR and return to caller. That avoids delays between
WR post and HW response.
If the template API was activated, PMD will use control queue for
sync operations.
PMD has different formats for the `user_data` context in
Implement sync meter ASO interface for async calls
in the template API environment.
Gregory Etelson (3):
net/mlx5: fix HWS meter actions availability
net/mlx5: fix sync meter processing in HWS setup
net/mlx5: fix indirect action async job initialization
drivers/net/mlx5/mlx5.h|
Allow compilation of HWS meter code only on platforms
that support HWS.
Fixes: 24865366e495 ("net/mlx5: support flow meter action for HWS")
Cc: sta...@dpdk.org
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5_flow_meter.c | 28
> When configuring RX queues, the default port configuration was used,
> even though it was modified before. This results in the
> 'relax-rx-offload' not being respected for RX queues.
> This commit uses 'rte_eth_dev_conf_get()' to obtain the device
> configuration structure instead.
>
> Fixes:
MLX5 PMD driver supports 2 types of indirect actions:
legacy INDIRECT and INDIRECT_LIST.
PMD has different handlers for each of indirection actions types.
Therefore PMD marks async `job::indirect_type` with relevant value.
PMD set the type only during indirect action creation.
Legacy INDIRECT quer
Synchronous calls for meter ASO try to pull pending completions
from CQ, submit WR and return to caller. That avoids delays between
WR post and HW response.
If the template API was activated, PMD will use control queue for
sync operations.
PMD has different formats for the `user_data` context in
Implement sync meter ASO interface for async calls
in the template API environment.
Gregory Etelson (3):
net/mlx5: fix HWS meter actions availability
net/mlx5: fix sync meter processing in HWS setup
net/mlx5: fix indirect action async job initialization
drivers/net/mlx5/mlx5.h|
Allow compilation of HWS meter code only on platforms
that support HWS.
Fixes: 24865366e495 ("net/mlx5: support flow meter action for HWS")
Cc: sta...@dpdk.org
Signed-off-by: Gregory Etelson
Acked-by: Dariusz Sosnowski
---
drivers/net/mlx5/mlx5_flow_meter.c | 28
Hi Thomas,
Following patch is also rebased on ToT. Please do the needful.
Regards
Sunil Kumar Kori
> -Original Message-
> From: sk...@marvell.com
> Sent: Thursday, March 7, 2024 3:36 PM
> To: Sunil Kumar Kori ; Rakesh Kudurumalla
>
> Cc: dev@dpdk.org
> Subject: [PATCH v2 1/1] doc: updat
From: Sunil Kumar Kori
Set of CLI commands are classified into following types;
- Commands which must be used in script only.
- Commands which must be used via telnet session only.
- Commands which can be used either in script or via telnet session.
Rename "Dynamic" column to "Scope" to provide
Release status meeting minutes 2024-03-07
=
Agenda:
* Release Dates
* Subtrees
* Roadmaps
* LTS
* Defects
* Opens
Participants:
* AMD
* ARM
* Debian/Microsoft
* Intel
* Marvell
* Nvidia
* Red Hat
Release Dates
-
The following are the current/u
> Subject: [PATCH] doc: update link to DevX integration on Windows
>
> The older link no longer works.
>
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ali Alnubani
> ---
> doc/guides/platform/mlx5.rst | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/doc/guides/platform/mlx5
> From: David Marchand [mailto:david.march...@redhat.com]
> Sent: Thursday, 7 March 2024 10.22
restriction
>
> On Thu, Mar 7, 2024 at 10:16 AM Morten Brørup
> wrote:
> >
> > > From: David Marchand [mailto:david.march...@redhat.com]
> > > Sent: Thursday, 7 March 2024 09.34
> > >
> > > On Mon, Dec
23/02/2024 09:30, Juraj Linkeš:
> Add hyphen to the regex, which is needed for drivers such as vfio-pci.
>
> Fixes: 88489c0501af ("dts: add smoke tests")
> Cc: jspew...@iol.unh.edu
> Signed-off-by: Juraj Linkeš
Reviewed-by: Jeremy Spewock
Tested-by: Nicholas Pratte
Applied, thanks.
From: Sunil Kumar Kori
Use the dpdk-cmdline-gen script to autogenerate all the boilerplate
structs and defines for the commandline part of the application.
Signed-off-by: Sunil Kumar Kori
Acked-by: Bruce Richardson
Acked-by: Nithin Dabilpuram
---
v2..v3:
- Rebase on top of the tree.
v1..v2:
Hi Thomas,
v3 is sent. Please do the needful.
Regards
Sunil Kumar Kori
> -Original Message-
> From: Sunil Kumar Kori
> Sent: Thursday, March 7, 2024 12:36 PM
> To: Thomas Monjalon
> Cc: Rakesh Kudurumalla ; dev@dpdk.org;
> Bruce Richardson
> Subject: RE: [EXTERNAL] Re: [PATCH v2 1/1] a
1 - 100 of 113 matches
Mail list logo
