Both legs of the loop share the same logic: the common parts are about
reserving and filling both address and length into the description.
This is moved into reserve_and_fill().
Bugzilla ID: 1440
Suggested-by: Maryam Tahhan
Signed-off-by: Ariel Otilibili
---
drivers/net/af_xdp/rte_eth_af_xdp.c
Hello,
The series addresses Bugzilla ID 1440 in two steps;
1. Fix use after free.
2. Refactor af_xdp_tx_zc().
Thank you,
---
v6
* added missing credits to Maryam Tahhan in v5
v5
(https://inbox.dpdk.org/dev/20250131183439.909831-1-ariel.otilib...@6wind.com/)
* fix style issues in the signature
tx_bytes is computed after both legs are tested. This might
produce a use after memory free.
The computation is now moved into each leg.
Bugzilla ID: 1440
Fixes: d8a210774e1d ("net/af_xdp: support unaligned umem chunks")
Signed-off-by: Ariel Otilibili
Acked-by: Stephen Hemminger
---
.mailmap
Looks good. Just some very small comments on polishing.
Reviewed-by: Luca Vizzarro
On 31/01/2025 19:38, Nicholas Pratte wrote:
Forwarding restarts in the run-time MTU adjustment test case have been
explicitly added, given that the 'requires_forwarding_restart' decorator
from a previous patch w
Reviewed-by: Luca Vizzarro
While I initially thought that this would improve the cleanliness of
future test suite writing, because of the various nuances of each
testpmd run-time command, having the framework managing forwarding when
needed introduces far too many additional challenges, such that, it
simply is not worth the
This implementation conflicts with other bug fixes implemented in
earlier patches. The decorator was devised as a mean to make suite
writing a bit simpler, semantically, but the addition of this adds too
many complications for it to be worth implementing.
Signed-off-by: Nicholas Pratte
---
dts/f
Forwarding restarts in the run-time MTU adjustment test case have been
explicitly added, given that the 'requires_forwarding_restart' decorator
from a previous patch was removed.
Signed-off-by: Nicholas Pratte
---
dts/tests/TestSuite_mtu.py | 8 +++-
1 file changed, 7 insertions(+), 1 deleti
DTS was pinned to use the version 2 of Fabric which was still backwards
compatible with Python 2. Many of the backwards-compatible modules have
been officially dropped in recent versions of Python, e.g. imp was
dropped in Python 3.12.
Given DTS does not support any version older than 3.10, bump up
Starting from Python 3.13 functools.partial is started to be treated as
a method descriptor. As a consequence, any class attributes whose value
is functools.partial is no longer treated as an attribute, but a class
method instead.
NicCapability is an enum that expects to have methods as values, an
When autodoc-pydantic is not installed in the system, autodoc has some
trouble parsing Pydantic's field_validator decorators. As a consequence
some unwanted warnings are printed when building the docs.
This commit removes any use of field validators in favour of "after"
model validators. Therefore
Hi everyone,
The first patch will finally get rid of any warnings in the doc build.
Whereas the other two patches are required to build docs or even run DTS
in Python 3.12 and 3.13.
Python 3.13 has introduced a new breaking change without warning, which
broke the way we used a specific stdlib fu
On Fri, Jan 31, 2025 at 12:55 AM Stephen Hemminger <
step...@networkplumber.org> wrote:
> Need space after xdp_desc to follow style used elsewhere in DPDK.
>
There you are, Stephen:
https://inbox.dpdk.org/dev/20250131183439.909831-3-ariel.otilib...@6wind.com/
Both legs of the loop share the same logic: the common parts are about
reserving and filling both address and length into the description.
This is moved into reserve_and_fill().
Bugzilla ID: 1440
Signed-off-by: Ariel Otilibili
---
drivers/net/af_xdp/rte_eth_af_xdp.c | 62 ---
tx_bytes is computed after both legs are tested. This might
produce a use after memory free.
The computation is now moved into each leg.
Bugzilla ID: 1440
Fixes: d8a210774e1d ("net/af_xdp: support unaligned umem chunks")
Signed-off-by: Ariel Otilibili
Acked-by: Stephen Hemminger
---
.mailmap
Hello,
The series addresses Bugzilla ID 1440 in two steps;
1. Fix use after free.
2. Refactor af_xdp_tx_zc().
Thank you,
---
v5
* fix style issues in the signature of reserve_and_fill() (Stephen Hemminger)
v4
(https://inbox.dpdk.org/dev/20250130221853.789366-1-ariel.otilib...@6wind.com/)
* red
Hello Maxime,
On Friday, January 31, 2025 14:09 CET, Maxime Coquelin
wrote:
> This patch fixes a regression in vhost_driver_register()
> causing VDUSE devices registration to fail systematically
> because the return value was initialized to -1 and not
> changed later on for this type of devices
On Thu, Jan 30, 2025 at 07:26:17PM +, Vladimir Medvedkin wrote:
> Fix a typo in avx512 rx function, where accidentally the _mm_load_si128
> argument was passed by value instead of a pointer.
>
> Fixes: 43fd3624fdfe ("drivers: replace GCC pragma with cast")
> Cc: andre...@linux.microsoft.com
>
On Fri, Jan 31, 2025 at 08:54:24AM +0100, Morten Brørup wrote:
> > From: Andre Muezerie [mailto:andre...@linux.microsoft.com]
> > Sent: Thursday, 30 January 2025 22.55
> >
> > As per guidance technical board meeting 2024/04/17. This series
> > removes the use of VLAs from code built for Windows fo
On Thu, Jan 30, 2025 at 05:26:25PM +, Bruce Richardson wrote:
> Add in two extra PCI device id's for future HW versions to be
> supported by idxd driver. This aligns with the definitions present
> in kernel idxd driver [1].
>
> [1]
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/li
The requirement to have DPDK in the DSA wq name is given in the idxd
chapter section 5.3.3 [1]. However, that information should also be
given in section 5.3.1 [2] where we list out the fields to be set and
what values they should have.
[1]
https://doc.dpdk.org/guides-24.11/dmadevs/idxd.html#devi
On Fri, Jan 31, 2025 at 2:12 PM Bruce Richardson
wrote:
>
> On Fri, Jan 31, 2025 at 12:58:13PM +, Anatoly Burakov wrote:
> > Intel IGC and E1000 drivers are distinct, but they are actually generated
> > from the same base code. This patchset will merge together all e1000-derived
> > drivers in
On Fri, Jan 31, 2025 at 12:58:13PM +, Anatoly Burakov wrote:
> Intel IGC and E1000 drivers are distinct, but they are actually generated
> from the same base code. This patchset will merge together all e1000-derived
> drivers into one common base, with three different ethdev driver
> frontends
This patch fixes a regression in vhost_driver_register()
causing VDUSE devices registration to fail systematically
because the return value was initialized to -1 and not
changed later on for this type of devices.
Fixes: 4d2aa150769b ("vhost: remove check around mutex init")
Signed-off-by: Maxime
Intel IGC and E1000 drivers are distinct, but they are actually generated
from the same base code. This patchset will merge together all e1000-derived
drivers into one common base, with three different ethdev driver
frontends (EM, IGB, and IGC).
After the merge is done, base code is also updated t
Update README with more up-to-date information.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/README | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/net/intel/e1000/base/README
b/drivers/net/intel/e1000/base/README
index b84ee5ad6e..2a0696f3
From: Piotr Kubaj
Rename version to nvm_version to avoid shadowing global variables on some
systems.
Signed-off-by: Piotr Kubaj
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_i210.c | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drive
From: Loktionov Aleksandr
In e1000_hash_mc_addr_generic() the expression:
"mc_addr[4] >> 8 - bit_shift", right shifting "mc_addr[4]"
shift by more than 7 bits always yields zero, so hash becomes not so
different. Add initialization with bit_shift = 1, and add a loop condition
to ensure
From: Barbara Skobiej
One of the bit shifts in MAC hash calculation triggers a static analysis
warning about a potential overflow. Fix the data type to avoid this.
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Barbara Skobiej
Signed-off-by: Anatoly Burakov
--
From: Dima Ruinskiy
Static analysis has detected a write that is not checked for errors,
leading to ignored error return value. Add a check.
Fixes: edcdb3c5f71b ("e1000/base: fix link flap on 82579")
Cc: sta...@dpdk.org
Signed-off-by: Dima Ruinskiy
Signed-off-by: Anatoly Burakov
---
drivers/
From: Barbara Skobiej
Fix setting device reset status bit in e1000_reset_hw_82580() function
for 82580 by first reading the register value, and then setting the
device reset bit.
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Barbara Skobiej
Signed-off-by: Anat
From: Jakub Buchocki <101559614+jbuch...@users.noreply.github.com>
Coverity has detected potential usage of uninitialized values inside the
array. Fix the warning by initializing array.
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Jakub Buchocki
Signed-off-by:
From: Przemysław Ciesielski
There are some static analysis warnings due to wrong data types being used
for various operations. Fix them with explicit typecasts.
Fixes: 5a32a257f957 ("e1000: more NICs in base driver")
Fixes: 38db3f7f50bd ("e1000: update base driver")
Cc: sta...@dpdk.org
Signed-o
From: Amir Avivi
Fix static analysis warning about comparison between types of
incompatible width, which might lead to an infinite loop due to overflow.
Fixes: af75078fece3 ("first public release")
Cc: sta...@dpdk.org
Signed-off-by: Amir Avivi
Signed-off-by: Anatoly Burakov
---
drivers/net/i
From: Nir Efrati
On some customer platforms it was observed that invalid PHY ID (0x) was
returned on first try. This patch introduces PHY ID retry mechanism to
perform retries on PHY ID read for up to a second. Because the original
code was intentionally not checking PHY ID due to a different
From: Marcin Jurczak
This patch removes non-inclusive language from the e1000 shared code and
comments.
Signed-off-by: Marcin Jurczak
Signed-off-by: Anatoly Burakov
---
.../net/intel/e1000/base/e1000_80003es2lan.c | 8 ++--
drivers/net/intel/e1000/base/e1000_82571.c| 4 +-
drivers/net/
From: Nir Efrati
Sporadic MDI errors are observed on some platforms. To address them, a
read/write retry mechanism is introduced. Set the initial retry counter
to 3 with an option to disable/enable the mechanism for some specific
flows that errors are expected.
Signed-off-by: Nir Efrati
Signed-
From: Dima Ruinskiy
PHY page and register notation traditionally uses decimal numbering. This
patch aligns the debug prints to this notation, as well as moves prints
to after we get a value of interest to us.
Signed-off-by: Dima Ruinskiy
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1
From: Sasha Neftin
When reading NVM checksum, we may encounter the following scenarios:
- Checksum may be invalid, and can be updated
- Checksum may be invalid but cannot be updated because NVM is read-only
For the latter case, we should just ignore invalid checksum and not attempt
to update it
From: Nir Efrati
Remove workaround for hardware synchronization before reset. The original
bug that was fixed by this workaround was root caused later and fixed in
other ways.
Signed-off-by: Nir Efrati
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_ich8lan.c | 17 --
From: Dima Ruinskiy
Current implementation of e1000_get_bus_info_ich8lan first attempts to
call e1000_get_bus_info_pcie_generic, and hard codes the bus width if the
latter does not return a valid value.
However, e1000_get_bus_info_pcie_generic will always fail on an ICH
device, since no PCIe cap
From: Adam Ludkiewicz
Define constants for WoL control on i210 NICs.
Signed-off-by: Adam Ludkiewicz
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_i210.h | 4
1 file changed, 4 insertions(+)
diff --git a/drivers/net/intel/e1000/base/e1000_i210.h
b/drivers/net/int
From: Dima Ruinskiy
Mark e1000_access_phy_wakeup_reg_bm as non-static.
Signed-off-by: Dima Ruinskiy
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_phy.c | 6 ++
drivers/net/intel/e1000/base/e1000_phy.h | 2 ++
2 files changed, 4 insertions(+), 4 deletions(-)
diff -
From: Nir Efrati
Add some definitions related to ULP WoL, and set longer ULP exit timeout on
more MAC types.
Signed-off-by: Nir Efrati
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_hw.h | 1 +
drivers/net/intel/e1000/base/e1000_ich8lan.c | 2 +-
drivers/net/intel/
From: Nir Efrati
Add Extended Firmware Semaphore (EXFWSM) register definition.
Signed-off-by: Nir Efrati
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 3 ++-
drivers/net/intel/e1000/base/e1000_regs.h| 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
From: Piotr Pietruszewski
This change rearranges the ICH8LAN code to have less indentation levels.
Signed-off-by: Piotr Pietruszewski
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_ich8lan.c | 84 ++--
1 file changed, 43 insertions(+), 41 deletions(-)
d
From: Sasha Neftin
Add new fields in structure to indicate if EEE LPI entries have been
observed on Tx and Rx path.
Signed-off-by: Sasha Neftin
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_hw.h | 2 ++
drivers/net/intel/e1000/base/e1000_mac.c | 2 ++
2 files changed,
From: Sasha Neftin
Some tools require device profile information to be present, so add it.
Signed-off-by: Sasha Neftin
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_hw.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/intel/e1000/base/e1000_hw.h
b/driv
From: Evgeny Efimov
Add enable/disable EEE API to base code.
Signed-off-by: Evgeny Efimov
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_api.c | 16
drivers/net/intel/e1000/base/e1000_api.h | 1 +
drivers/net/intel/e1000/base/e1000_hw.h | 1 +
dri
From: Sasha Neftin
Add definitions for address select mask (selects source address to be
used in the address filtering), as well as queue select (indicates which
Rx queue should get the packet matching this MAC address).
Signed-off-by: Sasha Neftin
Signed-off-by: Anatoly Burakov
---
drivers/n
From: Nir Efrati
On some MAC types, packet loss is observed due to wake DMA clock gating.
Disable wake DMA clock for some MAC types to avoid packet loss.
Signed-off-by: Nir Efrati
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_ich8lan.c | 7 +++
drivers/net/intel/e1
From: Vitaly Lifshits
Starting from MTP PHY type there is a synchronization issue between the
PHY's clock and PCH's synopsis PHY. This causes Tx packets corruptions.
The workaround to this issue is to change P0s power down state from
P0/P0s to P1 prior to disabling K1.
Signed-off-by: Vitaly Lif
From: Carolyn Wyborny
On 82575, driver is not supposed to check if managegement control is
enabled, and flush the Rx FIFO unconditionally.
Signed-off-by: Carolyn Wyborny
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_base.c | 3 ++-
1 file changed, 2 insertions(+), 1 de
The original code had incorrect indentation, and the fix was applied to
follow the indentation, i.e. adding brackets making the indentation valid.
However, the actual issue was incorrect indentation and not missing braces,
so the fix was incorrect.
This fix restores the original logic and corrects
There are some mismatches between what's in the base driver, and what is
in DPDK version of it. This patch fixes the following issues:
- Unneeded #ifndef clause
- Missing comments
- Duplicate definitions
- Missed definitions
- Extraneous fall-through comments
- Function comments not matching funct
From: Vitaly Lifshits
Add support for new hardware ID's from the I219 series.
Signed-off-by: Lifshits, Vitaly
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_api.c | 32 +
drivers/net/intel/e1000/base/e1000_hw.h | 36 +---
drivers
Some hardware device ID's from i210 and i219 series were added over the
years, but they were never added to the table of support PCI ID's for EM
and IGB drivers respectively. This commit adds the missing device ID's to
the table.
Signed-off-by: Anatoly Burakov
---
Notes:
This commit is not c
Add missing i225 devices to the list of supported devices by IGC driver,
as well as resort the device list to match sorting order of definitions
in the base code.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/igc_ethdev.c | 14 +-
1 file changed, 9 insertions(+), 5 delet
Add in various miscellaneous definitions (Rx/Tx, RSS, etc.) for i225
series NICs. This is mostly identical to IGC driver, with some minor
differences due to the way IGC driver was originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 107
Enable Latency Tolerance Reporting (LTR) code and add its associated
definitions for i225. This is mostly identical to IGC driver, with some
minor differences due to the way IGC driver was originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 31
Add EEE advertising support for i225 series NICs. This is mostly
identical to IGC driver, with some minor differences due to the way IGC
driver was originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 9 +++
drivers/net/intel/e1000/base/e1000_i
Add support for NVM/EEPROM operations for i225 series NICs. This is
mostly identical to IGC driver, with some minor differences due to the
way IGC driver was originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 40 +
drivers/net/intel/e1000/bas
Add LED blink support for i225 series NICs. This is mostly identical to
IGC driver, with some minor differences due to the way IGC driver was
originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_i225.c | 44 +++
drivers/net/intel/e1000/b
Add link bringup support for i225 series NICs. This is mostly identical
to IGC driver, with some minor differences due to the way IGC driver was
originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_defines.h | 16 +++
drivers/net/intel/e1000/base/e1000_i22
Add basic init, and PHY/MAC type support for i225 series NICs. This is
mostly identical to IGC driver, with some minor differences due to the
way IGC driver was originally generated.
Signed-off-by: Anatoly Burakov
---
drivers/net/intel/e1000/base/e1000_api.c | 20 ++
drivers/net/intel/e1000
From: Pawel Malinowski
According to datasheet software ownership of SWSM.SWESMBI bit should not
exceed 100ms. Current implementation caused incorrect timeout counter
values, where each iteration equals 50us delay. Because of that driver was
allowed to wait for semaphore even for 1,5s. This might
On Thu, Jan 30, 2025 at 04:14:42PM +, Vladimir Medvedkin wrote:
> Currently ICE PMD uses group attribute to select the appropriate HW engine
> to offload the flow. This behavior violates the rte_flow API, existing
> documentation/examples, and reveals hardware specific details.
>
> This patch
On Fri, Jan 31, 2025 at 11:43:53AM +0100, David Marchand wrote:
> On Thu, Jan 30, 2025 at 8:26 PM Vladimir Medvedkin
> wrote:
> >
> > Fix a typo in avx512 rx function, where accidentally the _mm_load_si128
> > argument was passed by value instead of a pointer.
> >
> > Fixes: 43fd3624fdfe ("drivers
On Thu, Jan 30, 2025 at 8:26 PM Vladimir Medvedkin
wrote:
>
> Fix a typo in avx512 rx function, where accidentally the _mm_load_si128
> argument was passed by value instead of a pointer.
>
> Fixes: 43fd3624fdfe ("drivers: replace GCC pragma with cast")
>
> Signed-off-by: Vladimir Medvedkin
Indee
Hi,
> -Original Message-
> From: Stephen Hemminger
> Sent: Tuesday, January 28, 2025 2:04 AM
> To: dev@dpdk.org
> Cc: Stephen Hemminger ; Xu, Rosen
>
> Subject: [RFC 7/7] raw/ifpga: use EAL version of TAILQ_FOREACH_SAFE
>
> Prefer the EAL version over local version of macro.
>
> Signed
CN20K does not have IE engines, hence change the default eng group
for cn20k and use legacy for cn10k or older version.
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/roc_cpt.c | 6 +++---
drivers/common/cnxk/roc_cpt.h | 10 +++---
drivers/common/cnxk/roc_nix_inl.c
From: Satha Rao
This change refactors the interrupt handling to be platform-specific.
Some platforms directly call ioctls, while others provide a library API
for the same functionality. Moving the interrupt handling to
platform-specific implementations enhances clarity and maintainability.
Signe
From: Satha Rao
Return -ENOSPC when there is no space to update the complete MC address
list, without flushing the existing list of addresses.
Signed-off-by: Satha Rao
---
drivers/net/cnxk/cnxk_ethdev_ops.c | 32 ++
1 file changed, 24 insertions(+), 8 deletions(-)
From: Srujana Challa
This patch updates the error log to a debug log since
it is not needed.
Signed-off-by: Srujana Challa
---
drivers/common/cnxk/roc_nix_inl_dev.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/common/cnxk/roc_nix_inl_dev.c
b/drivers/common/cnxk/
From: Monendra Singh Kushwaha
This patch adds support to reset inline device stats.
Signed-off-by: Monendra Singh Kushwaha
---
drivers/common/cnxk/roc_nix_inl.h | 1 +
drivers/common/cnxk/roc_nix_inl_dev.c | 27 +++
drivers/common/cnxk/version.map | 1 +
3 f
Make sure the w2 in inbound SA is set for inline IPsec have
L3 header on errors.
Fixes: 350b7a536a51 ("common/cnxk: enable L3 header write back in SA")
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/cnxk_security.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/dr
From: Rahul Bhansali
For small packets less than 55 bytes, SQ error interrupts are
observed.
When checksum offload flag is enabled and mbuf ol_flags are not
set, then default L3 type will be set to IPv6 in vector
processing. Based on this, HW will still validate for minimum
header size and genera
From: Monendra Singh Kushwaha
This patch enables allmulti mode on rpm/cgx vf devices.
Signed-off-by: Monendra Singh Kushwaha
---
.mailmap | 1 +
drivers/common/cnxk/roc_mbox.h| 1 +
drivers/common/cnxk/roc_nix_npc.c | 10 +++---
3 files changed, 9 insertions(
From: Rahul Bhansali
Inline IPsec Rx support for cn20k
Signed-off-by: Rahul Bhansali
---
drivers/event/cnxk/cn20k_worker.h | 111 --
drivers/net/cnxk/cn20k_rx.h | 5 +-
2 files changed, 109 insertions(+), 7 deletions(-)
diff --git a/drivers/event/cnxk/cn20k
From: Rahul Bhansali
Inline IPsec Rx support for cn20k
Signed-off-by: Rahul Bhansali
---
drivers/event/cnxk/cn20k_worker.h | 4 +-
drivers/net/cnxk/cn20k_rx.h | 737 --
2 files changed, 689 insertions(+), 52 deletions(-)
diff --git a/drivers/event/cnxk/cn20
From: Rahul Bhansali
Store the pool buffer size in lookup memory to calculate
mbuf start address for reassembly case in fastpath.
Also, restructured lookup memory data per port.
Signed-off-by: Rahul Bhansali
---
drivers/net/cnxk/cn20k_ethdev.c | 17 +
drivers/net/cnxk/cn20k_rxtx.h
From: Rahul Bhansali
Sync in CPT_RXC_TIME_CFG mbox as per new fields added for
cn20k and restructure to support it.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/roc_mbox.h| 2 ++
drivers/common/cnxk/roc_nix_inl.c | 55 +--
drivers/common/cnxk/roc_nix_i
From: Rahul Bhansali
Add action2 config for IPsec rule.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/hw/nix.h | 13 +++--
drivers/common/cnxk/roc_mbox.h | 1 +
drivers/common/cnxk/roc_npc.h | 1 +
drivers/common/cnxk/roc_npc_mcam.c | 1 +
4 files changed, 14
Add support for cn20k inline IPsec session create/destroy.
Signed-off-by: Nithin Dabilpuram
---
drivers/net/cnxk/cn10k_ethdev_sec.c |7 -
drivers/net/cnxk/cn20k_ethdev.c | 11 +
drivers/net/cnxk/cn20k_ethdev.h | 17 +
drivers/net/cnxk/cn20k_ethdev_sec.c | 1182 +++
Use UCAST_CPT in cn20k as opposed to UCAST_IPSEC in cn10k
for inline IPsec rule.
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/hw/nix.h| 1 +
drivers/common/cnxk/roc_npc.c | 15 +++
drivers/common/cnxk/roc_npc_mcam.c | 7 ---
drivers/common
From: Rahul Bhansali
For Inline device, add new roc API to get IPsec and
reassembly profile id.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/roc_nix_inl.c | 60 +++
drivers/common/cnxk/roc_nix_inl.h | 2 ++
2 files changed, 62 insertions(+)
diff --git a/d
From: Rahul Bhansali
Reassembly profile configuration for nix inline path.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/roc_features.h | 6 +
drivers/common/cnxk/roc_ie_ow.c| 22 +++
drivers/common/cnxk/roc_ie_ow.h| 2 +
drivers/common/cnxk/roc_nix.h
In CN20k, since we have 16 Inline inbound queues possible, add
support to attach inline inbound queue directly to the application
instead of getting it attached to CPT PF.
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/roc_features.h | 12 +
drivers/common/cnxk/roc_mbox.h
From: Rahul Bhansali
Support in NIX inline device for inbound and outbound SA init.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/roc_cpt.h | 1 +
drivers/common/cnxk/roc_nix_inl.c | 62 ++-
drivers/common/cnxk/roc_nix_inl_dev.c | 22 ++
dri
Add mbox support for global inline profile allocation.
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/roc_mbox.h| 45 ++
drivers/common/cnxk/roc_nix_inl.c | 53 +++
2 files changed, 85 insertions(+), 13 deletions(-)
diff --git a/
From: Rahul Bhansali
Features updated for cn20k platform.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/roc_features.h | 22 +++---
1 file changed, 11 insertions(+), 11 deletions(-)
diff --git a/drivers/common/cnxk/roc_features.h
b/drivers/common/cnxk/roc_features.h
i
From: Rahul Bhansali
Add SA context invalidate support for cn20k.
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/hw/cpt.h | 11 -
drivers/common/cnxk/roc_nix.h | 1 +
drivers/common/cnxk/roc_nix_inl.c | 37 ++-
3 files changed, 47 insertions
From: Rahul Bhansali
Adds below structures for cn20k,
- cpt_parse_hdr_s
- cpt_rxc_sg_s
- cpt_rxc_ptr_info_s
Signed-off-by: Rahul Bhansali
---
drivers/common/cnxk/hw/cpt.h| 211 +---
drivers/common/cnxk/roc_cpt.h | 2 +-
drivers/common/cnxk/roc_cpt_debug.
Add support for cn20k IPsec session create/destroy.
Signed-off-by: Nithin Dabilpuram
---
drivers/common/cnxk/cnxk_security.c | 546 +++-
drivers/common/cnxk/cnxk_security.h | 12 +-
drivers/common/cnxk/version.map | 2 +
3 files changed, 557 insertions(+), 3 deleti
From: Anoob Joseph
Add 20K defines for IPsec.
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/meson.build | 1 +
drivers/common/cnxk/roc_api.h | 1 +
drivers/common/cnxk/roc_ie_ow.c | 41 +++
drivers/common/cnxk/roc_ie_ow.h | 537
drivers/common/cnxk
From: Anoob Joseph
Add CPT cn20k device enumeration.
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/roc_constants.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/common/cnxk/roc_constants.h
b/drivers/common/cnxk/roc_constants.h
index 67cd74b28a..ac492651de 100644
--- a/dr
From: Anoob Joseph
add CPT LMT defines
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/roc_cpt.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/common/cnxk/roc_cpt.h b/drivers/common/cnxk/roc_cpt.h
index c8cf9354da..ac27479371 100644
--- a/drivers/common/cnxk/roc_cpt.h
+++
From: Anoob Joseph
IE engines are present only with cn9k and cn10k.
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/roc_cpt.c | 43 +++
drivers/common/cnxk/roc_cpt.h | 3 +++
2 files changed, 41 insertions(+), 5 deletions(-)
diff --git a/drivers/common/cnx
From: Anoob Joseph
CTX defines are common for all cases using CPT CTX.
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/cnxk_security.h | 1 +
drivers/common/cnxk/roc_cpt.h | 16
drivers/common/cnxk/roc_ie_ot.h | 16
drivers/net/cnxk/cn10k_rxtx.h
From: Anoob Joseph
9k would need special handling compared to 10k & 20k. Update the check
to reflect the same.
Signed-off-by: Anoob Joseph
---
drivers/common/cnxk/roc_cpt.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/common/cnxk/roc_cpt.c b/drivers/common/
1 - 100 of 104 matches
Mail list logo