> -Original Message-
> From: Thomas Monjalon
> Sent: Wednesday, July 7, 2021 2:32 PM
> To: Ruifeng Wang
> Cc: Jan Viktorin ; jer...@marvell.com; Bruce
> Richardson ; dev@dpdk.org;
> david.march...@redhat.com; nd ; Honnappa Nagarahalli
>
> Subject: Re: [PATCH v4 3/3] build: add option to
Hi Jiayu,
On 7/7/21 1:18 PM, Jiayu Hu wrote:
> No need to use bitfields in the structure rte_vhost_async_features.
> This patch reworks the structure to improve code readability. In
> addition, add preserved padding fields on the structure for future
> usage.
>
> Signed-off-by: Jiayu Hu
> ---
>
This patch simplify the pattern of flow rules of FDIR for IP fragment.
Flow rule can be created by the following command:
1. flow create 0 ingress pattern eth /
ipv4 fragment_offset spec 0x2000 fragment_offset mask 0x2000 /
end
2. flow create 0 ingress pattern eth / ipv6 /
ipv6_frag_ext
This patch simplify the pattern of flow rules of FDIR for IP fragment.
Flow rule can be created by the following command:
1. flow create 0 ingress pattern eth /
ipv4 fragment_offset spec 0x2000 fragment_offset mask 0x2000 /
end
2. flow create 0 ingress pattern eth / ipv6 /
ipv6_frag_ext
This patch simplify the pattern of flow rules of FDIR for IP fragment.
Flow rule can be created by the following command:
1. flow create 0 ingress pattern eth /
ipv4 fragment_offset spec 0x2000 fragment_offset mask 0x2000 /
end
2. flow create 0 ingress pattern eth / ipv6 /
ipv6_frag_ext
07/07/2021 07:48, Ruifeng Wang:
> Introduce a meson option 'use_wfe' to select wait until equal method.
> The default is disable. Traditional polling loop is used.
> When enabled, architecture specific mechanism is relied on to do the
> wait.
Why do we need an option?
Can it be automatic to enable
Update 21.08 release note for GTPoGRE FDIR and RSS.
Signed-off-by: Lingyu Liu
---
doc/guides/rel_notes/release_21_08.rst | 2 ++
1 file changed, 2 insertions(+)
diff --git a/doc/guides/rel_notes/release_21_08.rst
b/doc/guides/rel_notes/release_21_08.rst
index cd02820e68..df719420b8 100644
---
Support AVF RSS for inner most header of GTPoGRE packet. It supports
RSS based on inner most IP src + dst address and TCP/UDP src + dst
port.
Signed-off-by: Lingyu Liu
---
drivers/net/iavf/iavf_hash.c | 142 +--
1 file changed, 118 insertions(+), 24 deletions(-)
Support AVF FDIR for inner header of GTPoGRE tunnel packet.
Only patterns without inner most L3,L4 header support outer L3 src/dst
and TEID,QFI FDIR.
++---+
|Pattern |Input Set |
+-
Add GTPoGRE pattern support for AVF FDIR and RSS.
Signed-off-by: Lingyu Liu
---
drivers/net/iavf/iavf_generic_flow.c | 684 +++
drivers/net/iavf/iavf_generic_flow.h | 88
2 files changed, 772 insertions(+)
diff --git a/drivers/net/iavf/iavf_generic_flow.c
b/driver
Support AVF RSS and FDIR for GTPoGRE packet.
Lingyu Liu (4):
net/iavf: support flow pattern for GTPoGRE
net/iavf: support AVF FDIR for GTPoGRE tunnel packet
net/iavf: support AVF RSS for GTPoGRE packet
doc: update iavf driver FDIR/RSS for GTPoGRE
doc/guides/rel_notes/release_21_08.rst |
Hi,
> -Original Message-
> From: Maxime Coquelin
> Sent: Tuesday, July 6, 2021 5:32 PM
> To: Burakov, Anatoly ; Ding, Xuan
> ; Xia, Chenbo ; Thomas
> Monjalon ; David Marchand
>
> Cc: dev@dpdk.org; Hu, Jiayu ; Pai G, Sunil
> ; Richardson, Bruce ; Van
> Haaren, Harry ; Liu, Yong ;
> Ma, W
On Tue, Jul 6, 2021 at 1:58 PM Conor Walsh wrote:
>
> This patch aligns the l3fwd EM code with the changes made to LPM in
> commit 74fb854a3de6 ("examples/l3fwd: remove useless reloads in LPM
> main loop").
> This change ensures the compiler knows that the lcore config variables
> are constant val
Introduce a meson option 'use_wfe' to select wait until equal method.
The default is disable. Traditional polling loop is used.
When enabled, architecture specific mechanism is relied on to do the
wait.
Signed-off-by: Ruifeng Wang
---
config/arm/meson.build | 2 +-
meson_options.txt | 2 ++
Instead of polling for tail to be updated, use wfe instruction.
Signed-off-by: Gavin Hu
Signed-off-by: Ruifeng Wang
Reviewed-by: Steve Capper
Reviewed-by: Ola Liljedahl
Reviewed-by: Honnappa Nagarahalli
Acked-by: Konstantin Ananyev
Acked-by: Jerin Jacob
---
lib/ring/rte_ring_c11_pvt.h
From: Gavin Hu
In acquiring a spinlock, cores repeatedly poll the lock variable.
This is replaced by rte_wait_until_equal API.
Running the micro benchmarking and the testpmd and l3fwd traffic tests
on ThunderX2, Ampere eMAG80 and Arm N1SDP, everything went well and no
notable performance gain no
The rte_wait_until_equal_xxx APIs abstract the functionality of 'polling
for a memory location to become equal to a given value'[1].
Use the API for the rte spinlock and ring implementations.
With the wait until equal APIs being stable, changes will not impact ABI.
[1] http://patches.dpdk.org/cov
Introduce a meson option 'use_wfe' to select wait until equal method.
The default is disable. Traditional polling loop is used.
When enabled, architecture specific mechanism is relied on to do the
wait.
Signed-off-by: Ruifeng Wang
---
config/arm/meson.build | 2 +-
meson_options.txt | 2 ++
Instead of polling for tail to be updated, use wfe instruction.
Signed-off-by: Gavin Hu
Signed-off-by: Ruifeng Wang
Reviewed-by: Steve Capper
Reviewed-by: Ola Liljedahl
Reviewed-by: Honnappa Nagarahalli
---
lib/ring/rte_ring_c11_pvt.h | 4 ++--
lib/ring/rte_ring_generic_pvt.h | 3 +--
2
From: Gavin Hu
In acquiring a spinlock, cores repeatedly poll the lock variable.
This is replaced by rte_wait_until_equal API.
Running the micro benchmarking and the testpmd and l3fwd traffic tests
on ThunderX2, Ampere eMAG80 and Arm N1SDP, everything went well and no
notable performance gain no
The rte_wait_until_equal_xxx APIs abstract the functionality of 'polling
for a memory location to become equal to a given value'[1].
Use the API for the rte spinlock and ring implementations.
With the wait until equal APIs being stable, changes will not impact ABI.
[1] http://patches.dpdk.org/cov
https://bugs.dpdk.org/show_bug.cgi?id=723
Peng,Zhihong (zhihongx.p...@intel.com) changed:
What|Removed |Added
Status|UNCONFIRMED |RESOLVED
Resoluti
This patch allows to check the amount of in-flight packets
for vhost queues which register async channel acceleration.
Signed-off-by: Jiayu Hu
---
v2:
* use DEBUG log level
* return -1 when async is not registered
doc/guides/prog_guide/vhost_lib.rst | 5 +
lib/vhost/rte_vhost_async.h
> -Original Message-
> From: Conor Walsh
> Sent: Tuesday, July 6, 2021 7:58 PM
> To: david.march...@redhat.com; konstantin.anan...@intel.com;
> vladimir.medved...@intel.com; Ruifeng Wang ;
> jer...@marvell.com
> Cc: dev@dpdk.org; paulis.grib...@intel.com; Conor Walsh
>
> Subject: [PATCH]
This patch adds thread unsafe version for async register and
unregister functions.
Signed-off-by: Jiayu Hu
---
doc/guides/prog_guide/vhost_lib.rst | 12 +++
lib/vhost/rte_vhost_async.h | 39 +
lib/vhost/version.map | 4 +
lib/vhost/vhost.c | 15
No need to use bitfields in the structure rte_vhost_async_features.
This patch reworks the structure to improve code readability. In
addition, add preserved padding fields on the structure for future
usage.
Signed-off-by: Jiayu Hu
---
examples/vhost/main.c | 4 ++--
lib/vhost/rte_vhost_as
The vhost notifies the application of device readiness via
vhost_user_notify_queue_state(), but calling this function
is not protected by the lock. This patch is to make this
function call lock protected.
Fixes: d0fcc38f5fa4 ("vhost: improve device readiness notifications")
Cc: sta...@dpdk.org
Sig
Lock protection is needed during the vhost notifies the application of
device readiness, so the first patch adds lock protection. In addition,
the second patch reworks async feature structure to improve readability.
After performing locking, existed async vhost registration functions will
cause dea
> -Original Message-
> From: Andrew Rybchenko
> Sent: Tuesday, July 6, 2021 4:05 PM
> To: Zhang, Qi Z ; Zhang, AlvinX
> ; ajit.khapa...@broadcom.com
> Cc: dev@dpdk.org
> Subject: Re: [PATCH v3] ethdev: add IPv4 and L4 checksum RSS offload types
>
> On 7/6/21 10:18 AM, Zhang, Qi Z wrote:
LGTM, thanks
And I'am prepare dmadev V2, include:
a) Fix code review comments (e.g. multiple-process support, doxygen, comments,
typo)
b) Flatten device abstraction to two layer: dmadev <> vchan
c) Public API use dev_id and vchan_id to locate one vchan
d) Using the flags parameter instead of the
On Mon, Jul 5, 2021 at 2:24 PM Thomas Monjalon wrote:
> 13/06/2021 02:06, Ajit Khaparde:
> > From: Farah Smith
> >
> > - Translate Truflow action types for Thor to HCAPI RM
> > resource defined SRAM banks.
> > - move module type enum definitions to tf_core API
> > - Switch to subtype concept f
On Mon, Jul 5, 2021 at 2:27 PM Thomas Monjalon wrote:
> Please make the title start with an imperative verb.
>
During the merge, modified the title to:
add shared TCAM region support.
> One more nit below.
>
> 13/06/2021 02:06, Ajit Khaparde:
> > From: Farah Smith
> >
> > - switch to single s
在 2021/7/6 16:36, Andrew Rybchenko 写道:
@David, could you take a look at the ABI breakage warnings for
the patch. May we ignore it since ABI looks backward
compatible? Or should be marked as a minor change ABI
which is backward compatible with DPDK_21?
On 7/6/21 7:10 AM, Huisong Li wrote:
Curr
Add a new testpmd pattern field 'last_rsvd' that supports the
last 8-bits matching of VXLAN header.
The examples for the "last_rsvd" pattern field are as below:
1. ...pattern eth / ipv4 / udp / vxlan last_rsvd is 0x80 / end ...
This flow will exactly match the last 8-bits to be 0x80.
2. ...patt
This update adds support for the VXLAN header last 8-bits
matching when creating steering rules. At the PCIe probe
stage, we create a dummy VXLAN matcher using misc5 to check
rdma-core library's capability.
The logic is, group 0 depends on HCA_CAP to enable misc or misc5
for VXLAN matching while g
This update adds support for the VXLAN last 8-bits reserved field
matching when creating sw steering rules.
Add a new testpmd pattern field 'last_rsvd' that supports the last
8-bits matching of VXLAN header.
rongwei liu (2):
drivers: add VXLAN header the last 8-bits matching support
app/testp
From: Farah Smith
- switch to single slice management on Wh+
- Support of shared session WC_TCAM_HIGH and WC_TCAM_LOW regions
- Enable/disable using TF_TCAM_SHARED flag in tf_core.h
- Fix empty session module DBs in the case that none are
allocated for a given module type
Signed-off-by: Farah
From: Farah Smith
- Translate Truflow action types for Thor to HCAPI RM
resource defined SRAM banks.
- move module type enum definitions to tf_core API
- Switch to subtype concept for RM.
- alloc/free working for Thor SRAM table type for full AR.
Signed-off-by: Farah Smith
Signed-off-by: Rand
IPV6_FRAG_EXT item is missed for RSS expansion which causes flows like
[1] are wrongly expanded.
Different from other items, IPV6_FRAG_EXT hasn't next field because HW
only support to do hash of UDP/TCP for non-fragment.
This MLX5_EXPANSION_IPV6_FRAG_EXT node in RSS expansion graph only helps
RSS
Some RSS expandable items are missing which leads to the expanded
rte flow rules with wrong patterns.
Fix by adding missed items.
Fixes: d91093b9a2af ("net/mlx5: fix RSS pattern expansion")
Cc: sta...@dpdk.org
Signed-off-by: Xiaoyu Min
---
drivers/net/mlx5/mlx5_flow.c | 2 ++
1 file changed, 2
This patch set fixs the missing items in RSS expansion.
Xiaoyu Min (2):
net/mlx5: fix missing RSS expandable items
net/mlx5: fix missing RSS expansion of IPv6 frag
drivers/net/mlx5/mlx5_flow.c | 10 +-
1 file changed, 9 insertions(+), 1 deletion(-)
--
2.32.0
> -Original Message-
> From: Wang, Haiyue
> Sent: Tuesday, July 6, 2021 11:07 PM
> To: David Marchand ; dev@dpdk.org
> Cc: sta...@dpdk.org; Yang, Qiming ; Zhang, Qi Z
> ; Stillwell Jr, Paul M
> Subject: RE: [PATCH v2] net/ice: fix memzone leak when firmware is missing
>
> > -Origi
> -Original Message-
> From: Thomas Monjalon
> Sent: Wednesday, July 7, 2021 4:00 AM
> To: Zhang, Qi Z
> Cc: Wu, Wenjun1 ; dev@dpdk.org; Wu, Jingjing
> ; Xing, Beilei ; Yigit, Ferruh
>
> Subject: Re: [dpdk-dev] [PATCH v2 0/4] support AVF RSS and FDIR for GRE
> tunnel packet
>
> 21/06
> -Original Message-
> From: Thomas Monjalon
> Sent: Wednesday, July 7, 2021 6:11 AM
> To: Zhang, Qi Z
> Cc: Liu, Lingyu ; dev@dpdk.org; Xing, Beilei
> ; Wu, Jingjing ; Guo,
> Junfeng ; Yigit, Ferruh
> Subject: Re: [dpdk-dev] [PATCH v2 0/3] Support outer layer FDIR for
> GTPoGRE packe
On Tue, 6 Jul 2021 13:40:56 -0700
David Christensen wrote:
> On 6/15/21 1:12 AM, zhihongx.p...@intel.com wrote:
> > From: Zhihong Peng
> >
> > AddressSanitizer (ASan) is a google memory error detect
> > standard tool. It could help to detect use-after-free and
> > {heap,stack,global}-buffer ove
06/07/2021 04:40, Zhang, Qi Z:
> From: Liu, Lingyu
> > Add support for outer most L3 layer FDIR for GTPoGRE packet. GTPU TEID FDIR
> > is supported for GTPoGRE flow pattern without inner L3,L4 fields. GTPU
> > enhance header is supported.
> >
> > Lingyu Liu (3):
> > net/iavf: add flow pattern f
On 6/15/21 1:12 AM, zhihongx.p...@intel.com wrote:
From: Zhihong Peng
AddressSanitizer (ASan) is a google memory error detect
standard tool. It could help to detect use-after-free and
{heap,stack,global}-buffer overflow bugs in C/C++ programs,
print detailed error information when error happ
Add in some basic dmadev unit tests for testing drivers and the library
itself.
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 2 +
app/test/test_dmadev.c | 320 +
2 files changed, 322 insertions(+)
create mode 100644 app/test/test_dmadev.
Signed-off-by: Bruce Richardson
---
drivers/dma/meson.build | 11 +++
1 file changed, 11 insertions(+)
create mode 100644 drivers/dma/meson.build
diff --git a/drivers/dma/meson.build b/drivers/dma/meson.build
new file mode 100644
index 0..986b28be5
--- /dev/null
+++ b/drivers/dm
Drop the failed enqueue count since that is best tracked by the
application so that retries of the same job can be counted as desired by
the app developer. Since the "doorbell" function is separate from the
actual functions to enqueue descriptors, track a separate stat for jobs
which were submitted
Allow the user to skip passing the "out" parameters to the
rte_dmadev_completed() API call, by using local replacements in the
inline function. This simplifies drivers, and compilers should be able
to remove the branches at compile time in many cases.
Signed-off-by: Bruce Richardson
---
lib/dmad
Rather than having a special type for the index values used in dmadev,
just use regular int types, with appropriate return value notifications.
Signed-off-by: Bruce Richardson
---
lib/dmadev/rte_dmadev.h | 59
lib/dmadev/rte_dmadev_core.h | 12
remove the xstats function calls, as they are not needed for this class
as-yet.
Signed-off-by: Bruce Richardson
---
lib/dmadev/rte_dmadev.c | 63 --
lib/dmadev/rte_dmadev.h | 89 -
lib/dmadev/rte_dmadev_pmd.h | 19
lib/
a dump() function to print the state of a device to a file (e.g. sterr
or stdout) is very useful for debugging drivers.
Signed-off-by: Bruce Richardson
---
lib/dmadev/rte_dmadev.c | 17 +
lib/dmadev/rte_dmadev.h | 19 +++
lib/dmadev/rte_dmadev_pmd.h | 5 +
Export the rte_dmadevices array and the allocate and release functions
which are needed by PMDs.
Signed-off-by: Bruce Richardson
---
drivers/meson.build | 1 +
lib/dmadev/rte_dmadev.c | 2 ++
lib/dmadev/version.map | 3 +++
3 files changed, 6 insertions(+)
diff --git a/drivers/meson.build
For 32-bit builds, iova's are 64-bit still, so to ensure we can still
use PA mode on 32-bit we need to convert all enqueue "void *" parameters
to rte_iova_t
Signed-off-by: Bruce Richardson
---
lib/dmadev/rte_dmadev.h | 8
lib/dmadev/rte_dmadev_core.h | 4 ++--
2 files changed, 6 in
This patchset contains a series of changes to dmadev based on work being done to
port over our drivers to test this new infrastructure. Some of these are bug
fixes to enable compilation e.g. missing exports or meson.build files, while
others are suggested changes to enhance the API. All these patch
> Added support for crypto adapter OP_FORWARD mode.
>
> As OcteonTx CPT crypto completions could be out of order, each crypto op
> is enqueued to CPT, dequeued from CPT and enqueued to SSO one-by-one.
>
> Signed-off-by: Shijith Thotton
> ---
Acked-by: Akhil Goyal
> Set crypto adapter event device slow-path call backs.
>
> Signed-off-by: Shijith Thotton
> ---
> drivers/crypto/octeontx/meson.build | 1 +
> drivers/crypto/octeontx/otx_cryptodev.c | 4 ++
> drivers/crypto/octeontx/otx_cryptodev.h | 4 --
> .../crypto/octeontx/otx_cry
21/06/2021 04:54, Zhang, Qi Z:
> Acked-by: Qi Zhang
>
> Applied to dpdk-next-net-intel.
GRE should have been added to the new flow features table
in doc/guides/nics/features/iavf.ini
Will fix when pulling.
For future pulls, please run devtools/check-doc-vs-code.sh
Hi Stephen,
>
> You are right, but without this we cannot rule out the chance that the
> user downloaded DPDK, reverted to any version earlier than 20.11 (where
> make
> and meson co-existed) and use make to compile - also they happened to
> have
> intel-ipsec-mb version 1.0 or newer installed bef
> Hi Akhil,
>
> This is a required feature from our customer. So could you help us merging
> it?
>
> And we do not plan to remove this PMD - just put it to a new folder with
> common code shared for all SW Intel-ipsec-mb based PMDs.
> From the user point of view everything will be the same includ
> 10/05/2021 11:44, Anoob Joseph:
> > Add checks to catch overflow of any offsets. Offset control word
> > specifies,
> >
> > 1. 16 bits encryption offset
> > 2. 8 bits IV offset
> > 3. 8 bits auth offset
> >
> > Signed-off-by: Anoob Joseph
Acked-by: Akhil Goyal
>
> Should it be merged in DPDK
> Currently, there is a hard limitation on the PMD power management
> support that only allows it to support a single queue per lcore. This is
> not ideal as most DPDK use cases will poll multiple queues per core.
>
> The PMD power management mechanism relies on ethdev Rx callbacks, so it
> is v
06/07/2021 04:57, Zhang, Qi Z:
> > Acked-by: Leyi Rong
>
> Applied to dpdk-next-net-intel.
During your copy/paste of the ack, you transformed it into
Acked-by: Reyi Rong
and also
Acked-by: Leyi Rong
I fix it to
Acked-by: Leyi Rong
as original above.
Hi,
We have tested lib/librte_table/rte_table_hash_ext.c: rte_table_hash_ext_lookup
using dpdk 19.11
and it appears that the function is not reentrant safe causing incorrect hash
lookup misses at very small rate (< 0.005%)
when invoked by multiple threads in parallel.
The reason for reentr
On 7/6/21 2:08 AM, Bruce Richardson wrote:
On Tue, Jul 06, 2021 at 08:56:37AM +, Juraj Linkeš wrote:
-Original Message-
From: Bruce Richardson
Sent: Tuesday, June 29, 2021 1:29 PM
To: Juraj Linkeš
Cc: tho...@monjalon.net; david.march...@redhat.com;
honnappa.nagaraha...@arm.co
> Currently, if dev_configure is not called or fails to be called, users
> can still call dev_start successfully. So it is necessary to have a flag
> which indicates whether the device is configured, to control whether
> dev_start can be called and eliminate dependency on user invocation order.
>
On Tue, Jul 6, 2021 at 9:51 PM Thomas Monjalon wrote:
>
> 25/06/2021 16:01, Jerin Jacob:
> > On Wed, Jun 23, 2021 at 10:17 AM Nithin Dabilpuram
> > wrote:
> > >
> > > This patchset adds support for Marvell CN106XX SoC based on 'common/cnxk'
> > > driver. In future, CN9K a.k.a octeontx2 will also
25/06/2021 16:01, Jerin Jacob:
> On Wed, Jun 23, 2021 at 10:17 AM Nithin Dabilpuram
> wrote:
> >
> > This patchset adds support for Marvell CN106XX SoC based on 'common/cnxk'
> > driver. In future, CN9K a.k.a octeontx2 will also be supported by same
> > driver when code is ready and 'net/octeontx2
This patch fixed some unreasonable error check. Move all checks into one
helper function before configuring. Skip the check for DCF (VF0).
Fixes: 3a5a5bfc618b ("net/ice: support QoS config VF bandwidth in DCF")
Cc: sta...@dpdk.org
Signed-off-by: Ting Xu
---
drivers/net/ice/ice_dcf_sched.c | 65
Hi all,
I have rerun the failing unit test. It also recreated the report, so
that category should be passing now.
Currently, I am looking more into the ABI test that is failing on
Arch, as well as the failures with DTS tests.
I will keep this thread updated.
Thanks,
Brandon
On Mon, Jul 5, 2021 a
Hi,
> -Original Message-
> From: dev On Behalf Of Xiaoyu Min
> Sent: Friday, July 2, 2021 11:35 AM
> Cc: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH 0/3] MPLS RSS fixs
>
> This patch set fixs several problems of MPLS RSS:
>
> - removed MPLSoIP(HW doesn't support) related code
> - support
> -Original Message-
> From: David Marchand
> Sent: Tuesday, July 6, 2021 22:13
> To: dev@dpdk.org
> Cc: Wang, Haiyue ; sta...@dpdk.org; Yang, Qiming
> ; Zhang,
> Qi Z ; Stillwell Jr, Paul M
>
> Subject: [PATCH v2] net/ice: fix memzone leak when firmware is missing
>
> Caught by our QE
On Tue, Jul 6, 2021 at 1:50 PM David Marchand wrote:
>
> On Mon, Jul 5, 2021 at 9:28 AM Olivier Matz wrote:
> >
> > Seen with address sanitizer.
> >
> > rte_mempool_virt2iova() can only be used on mempool elements. In this case,
> > it is incorrect, and rte_mem_virt2iova() has to be used.
> >
> >
On Thu, Jun 24, 2021 at 1:53 PM David Marchand
wrote:
>
> This hack was needed with the make build system.
> With meson, any private header from a library is visible as long as a
> dependency to this library is expressed.
>
> Signed-off-by: David Marchand
Acked-by: Bruce Richardson
Acked-by: Ty
Introduce an internal firmware loading helper to remove code duplication
in our drivers and handle xz compressed firmwares by calling libarchive.
This helper tries to look for .xz suffixes so that drivers are not aware
the firmwares have been compressed.
libarchive is set as an optional dependenc
Both "normal" and "dcf" inits have their copy of some firmware loading
code.
The DSN query is moved in specific parts for the "normal" and "dcf" init.
A common helper ice_load_pkg is then introduced and takes an adapter
pointer as its main input.
This helper takes care of finding the right firmw
Fedora 34 only provides compressed firmwares.
Introduce an internal driver helper to handle transparently compression.
I chose libarchive for decompressing as it seems widely available and
DPDK had used it in the past.
Windows support only matters for net/ice and firmware loading was skipped
in
Hi Guys,
Sorry for missing this mail, for some reason it was missed in my inbox,
This is the link to this rfc:
https://www.3gpp.org/ftp/Specs/archive/38_series/38.415/38415-g30.zip
Kindest regards,
Raslan Darawsheh
> -Original Message-
> From: dev On Behalf Of Andrew Rybchenko
> Sent:
https://bugs.dpdk.org/show_bug.cgi?id=752
Bug ID: 752
Summary: Leak in ice init if ice_flow_init or
ice_reset_fxp_resource fails
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: U
From: Ivan Ilchenko
Report Rx scatter offload capability depending on VIRTIO_NET_F_MRG_RXBUF.
If Rx scatter is not requested, ensure that provided Rx buffers on
each Rx queue are big enough to fit Rx packets up to configured MTU.
Fixes: ce17eddefc20 ("ethdev: introduce Rx queue offloads API")
C
Caught by our QE.
When the firmware is missing, memzones were not released.
$ dpdk-testpmd -c 0x1f -a 0:0:0.0 -- -i
...
testpmd> dump_memzone
...
Zone 6: name:, len:0x15040, virt:0x1661b24c0, socket_id:0,
flags:0
physical segments used:
addr: 0x14000 iova: 0x14000 len: 0x4000 pagesz
On 6/15/21 4:15 PM, Cheng Jiang wrote:
> Applications need to stop DMA transfers and finish all the in-flight
> pkts when in VM memory hot-plug case and async vhost is used. This
> patch is to provide an unsafe API to drain in-flight pkts which are
> submitted to DMA engine in vhost async data p
> > > > > For Tx inline processing, when RTE_SECURITY_TX_OLOAD_NEED_MDATA is
> > > > > set, rte_security_set_pkt_metadata() needs to be called for pkts
> > > > > to associate a Security session with a mbuf before submitting
> > > > > to Ethdev Tx. This is apart from setting PKT_TX_SEC_OFFLOAD in
Hi Honnappa,
I think this cancelled all of the following events (or this was the last
one). Do you want to throw a marker on the calendar for next week (July
14)?
Cheers,
Lincoln
On Mon, Jul 5, 2021 at 3:42 PM Honnappa Nagarahalli <
honnappa.nagaraha...@arm.com> wrote:
> Hello,
>
>
On Tue, Jul 6, 2021 at 2:53 PM Wang, Haiyue wrote:
> > In such a case a single label should be enough, right?
>
> Yes, since pf->proto_xtr = NULL can be 'freed' safely.
Ah ok, I thought proto_xtr was initialised earlier and I had missed
the ice_pf_sw_init() call.
I'll go with your suggestion.
Tha
01/07/2021 09:15, Viacheslav Ovsiienko:
> There is the compilation flag HAVE_MLX5DV_DR_DEVX_PORT
> that depends on presence of mlx5dv_query_devx_port routine
> in rdma-core library.
The english syntax is difficult here.
You probably mean:
"The compilation flag HAVE_MLX5DV_DR_DEVX_PORT depends on t
01/07/2021 09:15, Viacheslav Ovsiienko:
> In order to get E-Switch vport identifiers the mlx5 PMD relies
> on two approaches:
> - use port query API if it is provided by rdma-core library
> - otherwise, deduce vport ids from the related VF index
> The latter is not reliable and may not work wit
> -Original Message-
> From: David Marchand
> Sent: Tuesday, July 6, 2021 21:28
> To: Wang, Haiyue
> Cc: dev@dpdk.org; sta...@dpdk.org; Yang, Qiming ;
> Zhang, Qi Z
> ; Stillwell Jr, Paul M
> Subject: Re: [PATCH] net/ice: fix memzone leak when firmware is missing
>
> On Tue, Jul 6, 202
This commit adds the multiple-thread flow insertion optimization
description.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
doc/guides/nics/mlx5.rst | 5 +
doc/guides/rel_notes/release_21_08.rst | 6 ++
2 files changed, 11 insertions(+)
diff --git a/doc/guides/nic
As hrxq struct has the indirect table pointer, while matching the
hrxq, better to use the hrxq indirect table instead of searching
from the list.
This commit optimizes the hrxq indirect table matching.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5_rxq.c | 18 +
This commit changes the index pool memory release configuration
to 0 when memory reclaim mode is not required.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
in
Currently, all the hash list tables are allocated during start up.
Since different applications may only use dedicated limited actions,
optimized the hash list table allocate on demand will save initial
memory.
This commit optimizes hash list table allocate on demand.
Signed-off-by: Suanming Mou
This commit enables the tag and header modify action index pool
per-core cache in non-reclaim memory mode.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5.c | 4 +++-
drivers/net/mlx5/mlx5.h | 1 +
drivers/net/mlx5/mlx5_flow_dv.c | 3 ++-
3 files chan
With the new per core optimization to the list, the hash bucket size
can be tuned to a more accurate number.
This commit adjusts the hash bucket size.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
drivers/net/mlx5/linux/mlx5_os.c | 2 +-
drivers/net/mlx5/mlx5.c | 2 +-
drivers
From: Matan Azrad
Modify header actions are allocated by mlx5_malloc which has a big
overhead of memory and allocation time.
One of the action types under the modify header object is SET_TAG,
The SET_TAG action is commonly not reused by the flows and each flow has
its own value.
Hence, the mlx
This commit supports the list non-lcore operations with
an extra sub-list and lock.
Signed-off-by: Suanming Mou
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_common_utils.c | 92 +
drivers/common/mlx5/mlx5_common_utils.h | 9 ++-
2 files changed, 71 insertions(+),
Currently, hash list uses the cache list as bucket list. The list
in the buckets have the same name, ctx and callbacks. This wastes
the memory.
This commit abstracts all the name, ctx and callback members in the
list to a constant struct and others to the inconstant struct, uses
the wrapper functi
From: Matan Azrad
Using the mlx5 list utility object in the hlist buckets.
This patch moves the list utility object to the common utility, creates
all the clone operations for all the hlist instances in the driver.
Also adjust all the utility callbacks to be generic for both list and
hlist.
Si
1 - 100 of 227 matches
Mail list logo
