> From: Huisong Li [mailto:lihuis...@huawei.com]
> Sent: Wednesday, 19 June 2024 08.32
>
> The deeper the idle state, the lower the power consumption, but the longer
> the resume time. Some service are delay sensitive and very except the low
> resume time, like interrupt packet receiving mode.
>
Acked-by: Abhinandan Gujjar
> -Original Message-
> From: Anoob Joseph
> Sent: Wednesday, June 19, 2024 12:03 PM
> To: Gujjar, Abhinandan S ; Akhil Goyal
>
> Cc: Jerin Jacob ; dev@dpdk.org
> Subject: [PATCH] test/event: increase default queue depth for tests
>
> With crypto adapter, lar
> From: Sriram Yagnaraman [mailto:sriram.yagnara...@ericsson.com]
>
> Hi Morten,
>
> > From: Morten Brørup
> >
> > > From: Sriram Yagnaraman [mailto:sriram.yagnara...@ericsson.com]
> > >
> > > When using ring based ethdev, mbuf->port is not set on received packets.
> > >
> > > For applications t
Wait for event with timeout (WFET) puts the CPU in a low power
mode and stays there until an event is signalled (SEV), loss of
an exclusive monitor or a timeout.
WFET is enabled selectively by checking FEAT_WFxT in Linux
auxiliary vector. If FEAT_WFxT is not available power management
will fallback
Add Arm Neoverse N3 part number to build configuration.
Signed-off-by: Wathsala Vithanage
Reviewed-by: Dhruv Tripathi
Reviewed-by: Jack Bond-Preston
---
config/arm/meson.build | 22 +-
1 file changed, 21 insertions(+), 1 deletion(-)
diff --git a/config/arm/meson.build b/c
Hello,
I want to use hairpin queues to forward high priority traffic (such as
LACP).
My goal is to ensure that this traffic is not dropped in case the
software pipeline is overwhelmed.
But during testing with dpdk-testpmd I can't achieve full throughput for
hairpin queues.
The best result I
The deeper the idle state, the lower the power consumption, but the longer
the resume time. Some service are delay sensitive and very except the low
resume time, like interrupt packet receiving mode.
And the "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
interface is used to s
The deeper the idle state, the lower the power consumption, but the longer
the resume time. Some service are delay sensitive and very except the low
resume time, like interrupt packet receiving mode.
And the "/sys/devices/system/cpu/cpuX/power/pm_qos_resume_latency_us" sysfs
interface is used to s
Add PM QoS configuration to declease the delay after sleep in case of
entering deeper idle state.
Signed-off-by: Huisong Li
---
examples/l3fwd-power/main.c | 29 +
1 file changed, 29 insertions(+)
diff --git a/examples/l3fwd-power/main.c b/examples/l3fwd-power/main.c
With crypto adapter, larger queue depths are desirable since same queue
could be used from multiple cores at the same time. With devices that
are capable of doing large bursts, larger queues would help in multi
core management of same queue.
Increase default queue depth in crypto adapter tests to
16/04/2024 02:50, Jochen Behrens:
>
> I submitted a separate patch cleaning up all vmware addresses in .mailmap.
Squashed altogether, thanks.
18/04/2024 12:46, Ferruh Yigit:
> On 4/17/2024 5:19 PM, Jochen Behrens wrote:
> > VMware was acquired by Broadcom, and vmware e-mail addresses will
> > be invalid in the future.
> >
> > Add broadcom.com addresses for those still in the company.
> > Place broadcom.com addresses first, as they are n
11/06/2024 08:04, David Marchand:
> Hello,
>
> On Mon, Jun 10, 2024 at 5:37 PM David Christensen
> wrote:
> >
> > Modify email address to route email traffic through different
> > internal email servers.
>
> Please update .mailmap too.
Applied with .mailmap update.
18/06/2024 14:09, Suanming Mou:
> From: Slava Ovsiienko
> > From: Dariusz Sosnowski
> > > Adding Bing Zhao as NVIDIA mlx5 maintainer.
> > >
> > > Signed-off-by: Dariusz Sosnowski
> >
> > Welcome, Bing
> >
> > Acked-by: Viacheslav Ovsiienko
>
> Good luck, Bing.
>
> Acked-by: Suanming Mou
A
21/05/2024 00:22, Ferruh Yigit:
> On 4/25/2024 4:03 PM, Loftus, Ciara wrote:
> >> Subject: [v1 1/1] MAINTAINERS: add another AF_XDP maintainer
> >>
> >> Add Maryam Tahhan as an additional maintainer for AF_XDP
> >> PMD and it's documentation.
> >>
> >> Signed-off-by: Maryam Tahhan
> >
> > Thanks
28/03/2024 18:22, Bruce Richardson:
> On Thu, Mar 28, 2024 at 04:06:38PM +, yuying.zh...@intel.com wrote:
> > From: Yuying Zhang
> >
> > Delete self from cpfl driver.
> >
> > Signed-off-by: Yuying Zhang
> > ---
> Sorry to see you go,
>
> Acked-by: Bruce Richardson
Applied
28/03/2024 18:22, Bruce Richardson:
> On Thu, Mar 28, 2024 at 04:10:21PM +, yuying.zh...@intel.com wrote:
> > From: Yuying Zhang
> >
> > Delete self from testpmd.
> >
> > Signed-off-by: Yuying Zhang
> > ---
> Acked-by: Bruce Richardson
Applied
28/03/2024 18:23, Bruce Richardson:
> On Thu, Mar 28, 2024 at 04:13:07PM +, yuying.zh...@intel.com wrote:
> > From: Yuying Zhang
> >
> > Delete self from i40e driver.
> >
> > Signed-off-by: Yuying Zhang
> > ---
> Acked-by: Bruce Richardson
Applied
08/05/2024 18:12, Dooley, Brian:
> From: Power, Ciara
> > Replacing myself as Crypto Perf App maintainer with Brian Dooley.
> >
> > Signed-off-by: Ciara Power
>
> Sorry to see you go!
>
> Acked-by: Brian Dooley
Series applied.
Thank you Ciara for all the work.
23/04/2024 10:27, Juraj Linkeš:
> On Mon, Apr 22, 2024 at 8:40 PM Stephen Hemminger
> wrote:
> >
> > Recent patch set for rte_memcpy, discovered that the some
> > mail addresses could not be found.
> >
> > Signed-off-by: Stephen Hemminger
> > Acked-by: Bruce Richardson
>
> Acked-by: Juraj Linke
Hello,
The mbuf structure [1] has __extension__ here,
i.e. preceding the structure following the integer field,
instead of preceding the union.
[1]: https://git.dpdk.org/dpdk/tree/lib/mbuf/rte_mbuf_core.h#n520
+1 anonymous unions are standard C so __extension__ is not necessary, it
is necessa
Update the description of match with E-Switch manager
dependency on upstream Linux and rdma-core in case user
will not install MLNX_OFED.
Signed-off-by: Suanming Mou
---
doc/guides/nics/mlx5.rst | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/doc/guides/nics/mlx5.rst b/doc/
From: Long Wu
As we already use the device argument to control the switch of
CPP service, it is time we can add the abnormal exit logic now.
Signed-off-by: Long Wu
Reviewed-by: Peng Zhang
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c | 6 --
1 file changed, 4 insertions(+), 2
From: Long Wu
The CPP service is not necessary for NFP PMD,
so add an option as the switch of it.
Signed-off-by: Long Wu
Reviewed-by: Peng Zhang
Reviewed-by: Chaoyong He
---
doc/guides/nics/nfp.rst | 9 +
drivers/net/nfp/nfp_ethdev.c | 21 -
drivers/
From: Long Wu
The CPP service is enabled in 'nfp_pf_init()', so it should be disabled
in corresponding free function 'nfp_pf_uninit()'.
Fixes: 600f6d2c7704 ("net/nfp: add service module")
Cc: sta...@dpdk.org
Signed-off-by: Long Wu
Reviewed-by: Peng Zhang
Reviewed-by: Chaoyong He
---
drivers
From: Long Wu
The flower service does not require high real-time performance,
so remove the dead loop to share service core with other services.
Signed-off-by: Long Wu
Reviewed-by: Peng Zhang
Reviewed-by: Chaoyong He
---
drivers/net/nfp/flower/nfp_flower_service.c | 29 +
From: Long Wu
There is no check for NFP device argument now, so the users may
not be aware of the wrong device argument and feels confused
with the result in some case.
Add the check logic to fix this problem.
Fixes: b301fd736003 ("net/nfp: add force reload firmware option")
Cc: peng.zh...@cori
From: Long Wu
The 'rte_eal_init()' will call 'rte_service_start_with_defaults()'
to start service core, map service core and set service state.
NFP service does not have any special needs for these setps, so
remove the redundant NFP service code.
Signed-off-by: Long Wu
Reviewed-by: Peng Zhang
This patch series refactor the service module, remove some specific
logic and use the service framework logic as much as possible.
Also add a device argument to control the enable of CPP service.
Long Wu (6):
net/nfp: fix check logic for device arguments
net/nfp: remove redundant NFP service c
From: Long Wu
Vector AVX2 Rx function supports parsing packet type and set it to mbuf.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_net_common.c| 2 +-
drivers/net/nfp/nfp_rxtx.c | 2 +-
drivers/net/nfp/nfp_rxtx.h | 3 +++
drivers/net/nfp/nfp_r
From: Long Wu
Use AVX2 instructions to accelerate Rx performance. The
acceleration only works on X86 machine.
Signed-off-by: Peng Zhang
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c| 2 +-
drivers/net/nfp/nfp_ethdev_vf.c | 2 +-
drivers/net/
From: Long Wu
Use AVX2 instructions to accelerate Tx performance. The
acceleration only works on X86 machine.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/meson.build | 15 +
drivers/net/nfp/nfdk/nfp_nfdk.h | 1 +
drivers/net/nfp/nfdk/nfp_
From: Long Wu
NFP will support NFDk vector Tx function, so move some
functions to header file for use by vector Tx function.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfdk/nfp_nfdk.h| 45 ++
drivers/net/nfp/nfdk/nfp_nfdk_dp.c | 43 +---
This patch series add the support of Rx/Tx function using the
AVX2 instruction.
Long Wu (4):
net/nfp: export more interfaces of NFDk
net/nfp: support AVX2 Tx function
net/nfp: support AVX2 Rx function
net/nfp: vector Rx function supports parsing ptype
drivers/net/nfp/meson.build
From: Long Wu
Add support for getting Tx queue information on NFP CoreNIC firmware.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c | 1 +
drivers/net/nfp/nfp_rxtx.c | 21 +
drivers/net/nfp/nfp_rxtx.h | 3 +++
3 files changed, 25 inse
From: Long Wu
Add support for getting Rx queue information on NFP CoreNIC firmware.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c | 1 +
drivers/net/nfp/nfp_rxtx.c | 18 ++
drivers/net/nfp/nfp_rxtx.h | 3 +++
3 files changed, 22 inserti
From: Long Wu
Use Rx/Tx offloads flag to control IPsec offload configuration.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c | 11 +++
1 file changed, 7 insertions(+), 4 deletions(-)
diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_eth
From: Long Wu
Use Rx offloads flag to control VXLAN offload configuration.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c | 10 +++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/drivers/net/nfp/nfp_ethdev.c b/drivers/net/nfp/nfp_ethdev.
From: Long Wu
Use Rx offloads flag to control RSS offload configuration.
Signed-off-by: Long Wu
Reviewed-by: Chaoyong He
---
drivers/net/nfp/nfp_ethdev.c| 2 +-
drivers/net/nfp/nfp_ethdev_vf.c | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/nfp/nfp_ethde
This patch series add the support of queue information get operations
for both Rx and Tx.
Long Wu (5):
net/nfp: use offload flag to control RSS configuration
net/nfp: use offload flag to control VXLAN configuration
net/nfp: use offload flag to control IPsec configuration
net/nfp: support g
> -Original Message-
> From: David Marchand
> Sent: Tuesday, June 18, 2024 9:06 PM
> To: Burakov, Anatoly ; Ye, MingjinX
>
> Cc: dev@dpdk.org; Stephen Hemminger ;
> Richardson, Bruce ; Mcnamara, John
>
> Subject: Re: [PATCH] Revert "bus/vdev: fix devargs in secondary process"
>
> On Mon
On Tue, Jun 18, 2024 at 04:25:08PM -0700, Stephen Hemminger wrote:
> The function rte_memzone_free() already checks if the argument
> is NULL. Add it to the coccinelle fix script and correct
> all the existing code.
>
> Signed-off-by: Stephen Hemminger
> ---
Reviewed-by: Tyler Retzlaff
On Tue, Jun 18, 2024 at 02:51:12PM +0200, David Marchand wrote:
> Similarly to libraries, when setting a list of desired applications,
> report if some application cannot be built because of a missing
> dependency.
>
> Signed-off-by: David Marchand
> ---
Acked-by: Tyler Retzlaff
On Tue, Jun 18, 2024 at 08:42:53AM +0200, Morten Brørup wrote:
> > From: Gregory Etelson [mailto:getel...@nvidia.com]
> > Sent: Tuesday, 18 June 2024 07.18
> >
> > DPDK IPv6 header definition combined the `version`, `traffic class`
> > and `flow label` header fields into a single 32 bits structure
Currently, in switch mode, mlx5 PMD only supports match with
dedicated vport. There is a use case which user may want to
offload the rules only to match with all the packet sent by
application not from vport.
Since the port_id info of packet sent by application is
E-Switch manager, and kernel driv
The function rte_memzone_free() already checks if the argument
is NULL. Add it to the coccinelle fix script and correct
all the existing code.
Signed-off-by: Stephen Hemminger
---
app/test-mldev/test_inference_common.c | 6 ++
devtools/cocci/nullfree.cocci | 3 +++
drivers/crypto/ip
Currently, `rte_eth_dev_info.rx_desc_lim.nb_max` as well as
`rte_eth_dev_info.tx_desc_lim.nb_max` shows 65535 as the limit,
which results in a few problems:
* It is an incorrect value
* Allocating an RX queue and passing `rx_desc_lim.nb_max` results in an
integer overflow and 0 ring size:
```
r
Hi, Slava
> 1. The absolute max descriptor number supported by ConnectX hardware is 32768.
> 2. The actual max descriptor number supported by the port (and its related
> representors)
> reported in log_max_wq_sz in HCA.caps. This value should be queried and
> save in mlx5_devx_cmd_query_hca
On Tue, 18 Jun 2024 18:41:29 +0100
Daniel Gregory wrote:
> diff --git a/config/riscv/meson.build b/config/riscv/meson.build
> index 07d7d9da23..4bda4089bd 100644
> --- a/config/riscv/meson.build
> +++ b/config/riscv/meson.build
> @@ -26,6 +26,13 @@ flags_common = [
> # read from /proc/device
https://bugs.dpdk.org/show_bug.cgi?id=1464
Bug ID: 1464
Summary: VLAN filtering bug
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: UNCONFIRMED
Severity: normal
Priority: Normal
Hi Morten,
> -Original Message-
> From: Morten Brørup
> Sent: Tuesday, 18 June 2024 18:01
> To: Sriram Yagnaraman
> Cc: dev@dpdk.org; Bruce Richardson
> Subject: RE: [PATCH] net/ring: Set mbuf->port for received packets
>
> [Du får inte e-post ofta från m...@smartsharesystems.com. Läs
When the RISC-V Zbc (carryless multiplication) extension is present, an
implementation of CRC hashing using hardware instructions is available.
Use it rather than jhash.
Signed-off-by: Daniel Gregory
---
lib/ip_frag/ip_frag_internal.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
When the RISC-V Zbc (carryless multiplication) extension is present, an
implementation of CRC hashing using hardware instructions is available.
Use it rather than jhash.
Signed-off-by: Daniel Gregory
---
examples/l3fwd/l3fwd_em.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git
Using carryless multiply instructions (clmul) from RISC-V's Zbc
extension, implement CRC-32 and CRC-16 calculations on buffers.
Based on the approach described in Intel's whitepaper on "Fast CRC
Computation for Generic Polynomails Using PCLMULQDQ Instructions", we
perfom repeated folds-by-1 whilst
Using carryless multiply instructions from RISC-V's Zbc extension,
implement a Barrett reduction that calculates CRC-32C checksums.
Based on the approach described by Intel's whitepaper on "Fast CRC
Computation for Generic Polynomials Using PCLMULQDQ Instruction", which
is also described here
(htt
The RISC-V Zbc extension adds carry-less multiply instructions we can
use to implement more efficient CRC hashing algorithms.
Signed-off-by: Daniel Gregory
---
config/riscv/meson.build | 7 +++
1 file changed, 7 insertions(+)
diff --git a/config/riscv/meson.build b/config/riscv/meson.build
The RISC-V Zbc extension adds instructions for carry-less multiplication
we can use to implement CRC in hardware. This patchset contains two new
implementations:
- one in lib/hash/rte_crc_riscv64.h that uses a Barrett reduction to
implement the four rte_hash_crc_* functions
- one in lib/net/net_
On Wed, Jun 12, 2024 at 04:01:59PM +0100, Anatoly Burakov wrote:
> To avoid triggering compile errors reported on some targets, use
> explicit type casts to specify type conversion.
>
> Signed-off-by: Anatoly Burakov
> ---
> drivers/net/ice/base/ice_flow.h | 2 +-
> drivers/net/ice/base/ice_pa
On Tue, Jun 18, 2024 at 11:54 AM Juraj Linkeš
wrote:
>
> > @@ -15,18 +18,34 @@ class InteractiveShell(SingleActiveInteractiveShell):
>
> > +def _start_application(self, get_privileged_command: Callable[[str],
> > str] | None) -> None:
> > +"""Overrides :meth:`_start_application` in th
Configuration to run vlan test suite
Signed-off-by: Dean Marx
---
dts/framework/config/conf_yaml_schema.json | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/dts/framework/config/conf_yaml_schema.json
b/dts/framework/config/conf_yaml_schema.json
index 4731f4511d..eca8244f27
Tweaked logic on sending and verifying packets for
more concise code, added verbose and promisc
function calls from pmd shell module.
Signed-off-by: Dean Marx
---
dts/tests/TestSuite_vlan.py | 167
1 file changed, 167 insertions(+)
create mode 100644 dts/t
Ported over the promisc and verbose mode functions
from v2 of the queue start/stop suite to use for the
VLAN suite. Tweaked some of the verification methods to be
more concise, changed some docstrings to be more specific.
Signed-off-by: Dean Marx
---
dts/framework/remote_session/testpmd_shell.
On Wed, Jun 12, 2024 at 04:01:38PM +0100, Anatoly Burakov wrote:
> From: Ian Stokes
>
> Support configurations with multiple ETH56G PHYs per a MAC complex. These PHYs
> have their own addresses and are supposed to have their ports numbered
> contiguously. This commit mainly provides port number t
> From: Sriram Yagnaraman [mailto:sriram.yagnara...@ericsson.com]
>
> When using ring based ethdev, mbuf->port is not set on received packets.
>
> For applications that use the mbuf->port to identify the incoming port,
> especially when eventdev RX adapter pulls the packet on a different core
> a
Hi David,
> > This patchset adds SVE support for the signature comparison in the cuckoo
> > hash lookup and improves the existing NEON implementation. These
> > optimizations required changes to the data format and signature of the
> > relevant functions to support dense hitmasks (no padding) and
@@ -15,18 +18,34 @@ class InteractiveShell(SingleActiveInteractiveShell):
+def _start_application(self, get_privileged_command: Callable[[str], str] |
None) -> None:
+"""Overrides :meth:`_start_application` in the parent class.
+
+Add a weakref finalize class after starting
Hi Brian,
Would be good to add a doc update too. Thanks.
Acked-by: Brian Dooley
Thanks for the suggestion! I totally missed that.
I've uploaded v2 with a doc update.
Thanks,
Jack
Add the option to create one session for the PMD, and share it across
all of the queue pairs. This may help to discover/debug concurrency
issues (both correctness and performance) that can occur when using this
configuration.
Signed-off-by: Jack Bond-Preston
Acked-by: Brian Dooley
Reviewed-by: W
On 13. 6. 2024 20:15, jspew...@iol.unh.edu wrote:
From: Jeremy Spewock
Interactive shells are managed in a way currently where they are closed
and cleaned up at the time of garbage collection. Due to there being no
guarantee of when this garbage collection happens in Python, there is no
way
Configuration to run vlan test suite
Signed-off-by: Dean Marx
---
dts/framework/config/conf_yaml_schema.json | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/dts/framework/config/conf_yaml_schema.json
b/dts/framework/config/conf_yaml_schema.json
index 4731f4511d..eca8244f27
Ported over the promisc and verbose mode functions from
v2 of the queue start/stop suite to use for the VLAN suite.
Tweaked some of the verification methods to be
more concise, changed some docstrings to be more specific.
Signed-off-by: Dean Marx
---
dts/framework/remote_session/testpmd_shell.
Tweaked logic on sending and verifying packets for
more concise code, added verbose and promisc
function calls from pmd shell module.
Signed-off-by: Dean Marx
---
dts/tests/TestSuite_vlan.py | 59 +
1 file changed, 27 insertions(+), 32 deletions(-)
diff --g
On Wed, Jun 12, 2024 at 04:00:51PM +0100, Anatoly Burakov wrote:
> From: Ian Stokes
>
> Replace enum ice_status with int. Also replcae ICE_SUCCESS instances with 0.
>
> Signed-off-by: Przemek Kitszel
> Signed-off-by: Ian Stokes
> ---
> drivers/net/ice/base/ice_common.c | 12 ++--
> drivers/ne
On Wed, Jun 12, 2024 at 04:00:49PM +0100, Anatoly Burakov wrote:
> From: Ian Stokes
>
> Code using ice_ptp_one_port_cmd() (or the device specific variants for E822
> and
> ETH56G) has a subtle bug where-in the executing of ice_ptp_exec_tmr_cmd() will
> cause all other ports to execute their last
On Wed, Jun 12, 2024 at 04:00:39PM +0100, Anatoly Burakov wrote:
> From: Ian Stokes
>
> If FW supports the corresponding functionality, the driver allows PFs to
> subscribe the same switch recipes. Then when the PF is done with a switch
> recipe, the PF can ask the FW to free that switch recipe.
The type argument is carried through malloc heap routines but
never used there. It is only used a rte_malloc for tracing.
Signed-off-by: Stephen Hemminger
---
lib/eal/common/eal_common_memzone.c | 6 ++---
lib/eal/common/malloc_heap.c| 39 -
lib/eal/common/ma
The string type is only used for tracing and not used as
documented by dump_stats.
Signed-off-by: Stephen Hemminger
---
lib/eal/include/rte_malloc.h | 27 +--
1 file changed, 13 insertions(+), 14 deletions(-)
diff --git a/lib/eal/include/rte_malloc.h b/lib/eal/include/rt
The type parameter for malloc is only used for tracing.
Fix documentation and don't pass through heap routines.
Stephen Hemminger (2):
rte_malloc: document that type is for tracing
eal: remove type argument from internal routines
v3 - drop event/sw patch (already fixed)
lib/eal/common/eal_c
When using ring based ethdev, mbuf->port is not set on received packets.
For applications that use the mbuf->port to identify the incoming port,
especially when eventdev RX adapter pulls the packet on a different core
and the application running on a worker core has no clue on the incoming
port. T
On Tue, May 14, 2024 at 10:52 PM Honnappa Nagarahalli
wrote:
> > On May 13, 2024, at 11:35 AM, Abdullah Ömer Yamaç
> > wrote:
> >
> > Set the maximum reclamation size to user provided value
> >
> > Fixes: 769b2de7fb52 ("hash: implement RCU resources reclamation")
> > Cc: sta...@dpdk.org
> >
> >
On Mon, May 20, 2024 at 8:51 PM Honnappa Nagarahalli
wrote:
> > On May 15, 2024, at 5:54 AM, Abdullah Ömer Yamaç
> > wrote:
> >
> > This patch adds a new feature to the hash library to allow the user to
> > reclaim the defer queue. This is useful when the user wants to force
> > reclaim resource
> From: Bruce Richardson [mailto:bruce.richard...@intel.com]
>
> On Mon, Jun 10, 2024 at 10:56:00AM +0200, Morten Brørup wrote:
> > PING (again) for review.
> >
> > Many applications use bursts of more than 32 packets,
> > and some applications buffer more than 512 packets.
> >
> > This patch upda
On Fri, May 10, 2024 at 10:40 AM Kiran Kumar Kokkilagadda
wrote:
> > When the order of graph destroy is not the reverse order of create, that is,
> > when it is destroyed at will, the newly created graph id will be the same as
> > the existing graph id, which is not the expected unique graph id. T
On Thu, Apr 25, 2024 at 5:28 PM Stephen Hemminger
wrote:
> > As per Intel, this is not supported, and the librte-bpf test fails on 32bit
> > x86
> > kernels, so disable the library and the pmd.
Cc: sta...@dpdk.org
> >
> > Signed-off-by: Luca Boccassi
> Acked-by: Stephen Hemminger
Applied, th
On Mon, Jun 10, 2024 at 10:56:00AM +0200, Morten Brørup wrote:
> PING (again) for review.
>
> Many applications use bursts of more than 32 packets,
> and some applications buffer more than 512 packets.
>
> This patch updates the mempool perf test accordingly.
>
> -Morten
>
> > From: Morten Brør
On Wed, Mar 20, 2024 at 6:11 PM Robin Jarry wrote:
>
> * Quote graph name to avoid parsing errors when it contains a dash.
> * Use fixed margin and smaller font for a more compact layout.
> * Use sans-serif font, the default is times new roman which is not the
> best choice for a packet processi
On Mon, Apr 1, 2024 at 10:37 PM Robin Jarry wrote:
>
> In rte_graph_cluster_stats_get, the walk model of the first graph is
> checked to determine if multi-core dispatch specific counters should be
> updated or not. This global list is accessed without any locks.
>
> If the global list is modified
On Tue, Jun 18, 2024 at 11:24 AM Robin Jarry wrote:
>
> The graph id is determined based on a global variable that is
> incremented every time a graph is created, and decremented every time
> a graph is destroyed. This only works if graphs are destroyed in the
> reverse order in which they have be
The rte_vhost_driver_unregister() vhost_user_read_cb()
vhost_user_client_reconnect() can be called at the same time by 3 threads.
when memory of vsocket is freed in rte_vhost_driver_unregister(),
then vhost_user_read_cb() maybe add vsocket to reconn_list,
the invalid memory of vsocket is accessed
On Mon, Jun 17, 2024 at 4:49 PM Stephen Hemminger
wrote:
>
> On Fri, 31 May 2024 10:41:18 -0700
> Stephen Hemminger wrote:
>
> > This reverts commit 628362c94a0b567a39a0177539c12c97d999.
> >
> > That commit broke use of dumpcap with vdev's and probably
> > other uses of secondary processes wi
On Tue, Jun 18, 2024 at 02:51:12PM +0200, David Marchand wrote:
> Similarly to libraries, when setting a list of desired applications,
> report if some application cannot be built because of a missing
> dependency.
>
> Signed-off-by: David Marchand
> ---
> app/meson.build | 10 ++
> 1 fi
> From: lihuisong (C) [mailto:lihuis...@huawei.com]
>
> Hi Morten,
>
> Thanks for your review.
>
>
> 在 2024/6/14 16:04, Morten Brørup 写道:
> >> +#define PM_QOS_SYSFILE_RESUME_LATENCY_US \
> >> + "/sys/devices/system/cpu/cpu%u/power/pm_qos_resume_latency_us"
> > Is it OK to access this path usi
Similarly to libraries, when setting a list of desired applications,
report if some application cannot be built because of a missing
dependency.
Signed-off-by: David Marchand
---
app/meson.build | 10 ++
1 file changed, 10 insertions(+)
diff --git a/app/meson.build b/app/meson.build
ind
> From: Gregory Etelson [mailto:getel...@nvidia.com]
>
> DPDK IPv6 header definition combined the `version`, `traffic class`
> and `flow label` header fields into a single 32 bits structure member
> `vtc_flow`.
>
> The patch expands IPv6 header definition with dedicated structure
> members for th
Re Robin,
On Wed, Mar 27, 2024 at 10:17 AM Robin Jarry wrote:
>
> In some cases, the node context data is used to store two pointers
> because the data is larger than the reserved 16 bytes. Having to define
> intermediate structures just to be able to cast is tedious. And without
> intermediate s
Hi Morten,
Thanks for your review.
在 2024/6/14 16:04, Morten Brørup 写道:
+#define PM_QOS_SYSFILE_RESUME_LATENCY_US \
+ "/sys/devices/system/cpu/cpu%u/power/pm_qos_resume_latency_us"
Is it OK to access this path using the lcore_id as CPU parameter to
open_core_sysfs_file(), or must
DPDK IPv6 header definition combined the `version`, `traffic class`
and `flow label` header fields into a single 32 bits structure member
`vtc_flow`.
The patch expands IPv6 header definition with dedicated structure
members for the `version`, `traffic class` and `flow label` fields.
The `traffic c
On Mon, May 27, 2024 at 07:19:21PM +0300, Igor Gutorov wrote:
> According to the Intel X710/XXV710/XL710 Datasheet, the maximum receive
> queue descriptor length is 0x1FE0 (8160 in base 10). This is specified
> as QLEN in table 8-12, page 1083.
>
> I've tested this change with an XXV710 NIC and it
> -Original Message-
> From: Slava Ovsiienko
> Sent: Tuesday, June 18, 2024 7:17 PM
> To: Dariusz Sosnowski ; Ori Kam
> ; Suanming Mou ; Matan
> Azrad ; Bing Zhao ;
> NBU-Contact-Thomas Monjalon (EXTERNAL)
> Cc: dev@dpdk.org
> Subject: RE: [PATCH] maintainers: update for networking mlx
Updated the latest support for cpfl pmd in MEV TS
firmware version which is 1.4.
Signed-off-by: Soumyadeep Hore
---
doc/guides/nics/cpfl.rst | 2 ++
1 file changed, 2 insertions(+)
diff --git a/doc/guides/nics/cpfl.rst b/doc/guides/nics/cpfl.rst
index 9b7a99c894..528c809819 100644
--- a/doc/gui
1 - 100 of 167 matches
Mail list logo
