Thank you, Stephen, I've got it.
On 2025/8/30 4:22, Stephen Hemminger wrote:
> On Fri, 29 Aug 2025 16:24:42 +0800
> "Renyong Wan" wrote:
>
>> This series introduces multiple updates to the Yunsilicon xsc PMD,
>> including new feature support, optimizations, bug fixes.
>> Highlights include:
>>
>>
Currently, when retrieving the L4 protocol from the L3 header,
the case of IPv6 with extension headers is not handled correctly.
This patch fixes it.
Fixes: 76730c7b9b5a ("app/testpmd: use packet type parsing API")
Cc: sta...@dpdk.org
Signed-off-by: Dengdui Huang
---
app/test-pmd/csumonly.c | 3
On Fri, Aug 29, 2025 at 03:04:10PM +0200, David Marchand wrote:
> Hello,
>
> On Tue, 5 Aug 2025 at 22:00, Andre Muezerie
> wrote:
> >
> > This patch fixes some issues which were preventing this test to be
> > built on Windows:
> >
> > - Remove VLAs (not supported by msvc).
> > - Replace strsep()
On Thu, 21 Aug 2025 11:02:33 +0530
wrote:
> From: Sunil Kumar Kori
>
> Adding link type parameter to provide the type
> of port like twisted pair, fibre etc.
>
> Also added an API to convert the RTE_ETH_LINK_CONNECTOR_XXX
> to a readable string.
>
> Signed-off-by: Nithin Dabilpuram
> Signed-
On Fri, 29 Aug 2025 16:24:42 +0800
"Renyong Wan" wrote:
> This series introduces multiple updates to the Yunsilicon xsc PMD,
> including new feature support, optimizations, bug fixes.
> Highlights include:
>
> * Firmware version query support
> * TSO and FEC support
> * EEPROM dump and promiscuo
On Fri, 29 Aug 2025 22:49:52 +0800
Yang Ming wrote:
> The current implementation hardcodes the socket file path to
> /var/tmp, which has two issues:
>
> 1. Hardcoding absolute paths is not good practice.
> 2. /var/tmp may not be writable in containerized or restricted
>environments (e.g. whe
https://bugs.dpdk.org/show_bug.cgi?id=1783
Bug ID: 1783
Summary: crash in rollback_expand_heap
Product: DPDK
Version: 24.11
Hardware: All
OS: All
Status: UNCONFIRMED
Severity: normal
Priority: No
The idea is to split off framework and public API. All calls in the tests
should go through the public API. This will allow us to change the
framework internals without breaking the tests and also to provide a
stable API for test writers.
More importantly this will make it easier to non DTS develo
Adjust imports and calls in all tests to use the new public API.
Signed-off-by: Paul Szczepanek
---
dts/tests/TestSuite_blocklist.py | 11 +++--
dts/tests/TestSuite_checksum_offload.py | 44 ++-
dts/tests/TestSuite_dual_vlan.py | 12 ++---
dts/test
On Wed, Jul 23, 2025 at 05:19:58PM +0100, Bruce Richardson wrote:
> The ultimate of this patchset is to make it easier to run on systems
> with large numbers of cores, by simplifying the process of using core
> numbers >RTE_MAX_LCORE. The new EAL args "--lcores-remapped", also
> shortened to just "
Hi Ciara,
On Fri, 29 Aug 2025, Ciara Loftus wrote:
Allow the user to configure the state a link should be in after
the device has been stopped/closed. Make this configurable
through the new experimental rte_eth_dev_set_link_state_on_close
API. Three states are allowed:
1. down: bring (or keep)
Hi Experts,
is there any chance to review and accept this patch?
Brs,
Yang Ming
The current implementation always tries to open debug dump files
under /var/log, which may not be writable in containerized or
restricted environments (e.g. when the filesystem is mounted as
read-only).
This patch introduces an OS-specific helper function
mlx5_os_debug_dump_file_open() to unify th
The current implementation hardcodes the socket file path to
/var/tmp, which has two issues:
1. Hardcoding absolute paths is not good practice.
2. /var/tmp may not be writable in containerized or restricted
environments (e.g. when the filesystem is mounted read-only).
This patch replaces the h
> > > rte_free(vec.pt.val);
> > > -
> > > + vec.pt.val = NULL;
> >
> > Same could be the case for fips_mct_shake_test also.
> >
> Good catch, probably the same could happen with shake.
> I currently don’t have a way to test shake but I'm happy to make the change as
> part of this patch if you can
Allow the user to configure link state on close behaviour via the
link_state_on_close argument. Three options are allowed:
1. down: bring (or keep) the link down
2. up: bring (or keep) the link up
3. initial: restore the link to the state it was in when the device was
started.
Signed-off-by: Ciara
Implement the callback that allows for configuration of the link state
when the device is closed or stopped. The default behaviour remains the
same, that being that the link is restored to its original state on
device closure.
Signed-off-by: Ciara Loftus
---
drivers/net/intel/ice/ice_ethdev.c |
Allow the user to configure the state a link should be in after
the device has been stopped/closed. Make this configurable
through the new experimental rte_eth_dev_set_link_state_on_close
API. Three states are allowed:
1. down: bring (or keep) the link down
2. up: bring (or keep) the link up
3. ini
Most drivers bring the port link down on device close (eg. intel i40e), however
some others restore the link to its original state when the device was started
(eg. intel ice). It may be useful to be able to configure this behaviour which
is the motivation for this RFC.
This RFC proposes a way to m
On Mon, 28 Jul 2025 at 17:19, Bruce Richardson
wrote:
>
> We no longer support versions of linux before 4.19[1], therefore
> all supported linux versions support F_ADD_SEALS (Linux 3.17+) and
> MFD_HUGETLB (Linux 4.14+), so no need to have #ifdefs and fallbacks to
> support systems which do not su
On Tue, 19 Aug 2025 at 13:25, Bruce Richardson
wrote:
>
> DPDK now recommends compiler versions GCC 8+, and Clang 7+, so remove
> checks and workarounds for versions old than these.
>
> Signed-off-by: Bruce Richardson
> Acked-by: Stephen Hemminger
> Reviewed-by: David Marchand
Series applied,
On 8/29/25 12:26 PM, Bathija, Pravin wrote:
Answers inline.
Internal Use - Confidential
-Original Message-
From: Maxime Coquelin
Sent: Friday, August 29, 2025 2:18 AM
To: Bathija, Pravin ; Bruce Richardson
; David Marchand
; Chenbo Xia
Cc: dev@dpdk.org; pravin.m.bathija@gmail.
> -Original Message-
> From: Gowrishankar Muthukrishnan
> Sent: Friday 29 August 2025 13:17
> To: Finn, Emma
> Cc: dev@dpdk.org; Ji, Kai
> Subject: RE: [EXTERNAL] [PATCH] examples/fips_validation: fix dangling pointer
> for SHA test
> Importance: High
>
> > SHAKE is not supported by
Hello,
On Tue, 5 Aug 2025 at 22:00, Andre Muezerie
wrote:
>
> This patch fixes some issues which were preventing this test to be
> built on Windows:
>
> - Remove VLAs (not supported by msvc).
> - Replace strsep() (which is not natively available on Windows)
> with strtok_r().
> - Remove the "th
On Mon, 21 Jul 2025 at 20:18, Andre Muezerie
wrote:
>
> Header termios.h was not found on Windows, which was probably the
> reason this test was originally excluded from the Windows build.
> However, ir turns out this header is not needed for this test, so it
> can be removed and then test can be
On Mon, 21 Jul 2025 at 20:31, Andre Muezerie
wrote:
>
> This patch fixes a few issues which were preventing this test from
> getting compiled on Windows:
>
> 1) VLAs were removed: Variable Length Arrays are not supported by MSVC.
> Where possible constant C arrays were used. Where that approach wa
On 8/28/25 2:14 AM, Stephen Hemminger wrote:
With shadow warnings enabled there were a couple of cases
to fix. One was simple reuse of 'ret', other was overlap
between type and variable name.
Signed-off-by: Stephen Hemminger
Acked-by: Andrew Rybchenko
On Wed, Aug 20, 2025 at 10:45:21AM +0200, Dariusz Sosnowski wrote:
> This patch fixes a bug in
> mlx5_flow_dv_discover_counter_offset_support()
> uncovered by a warning reported by MSVC:
>
> ../drivers/net/mlx5/mlx5_flow_dv.c(19636): warning C5287:
> operands are different enum types '
On Wed, Aug 27, 2025 at 11:15:57AM +0200, David Marchand wrote:
> On Wed, 20 Aug 2025 at 10:47, Dariusz Sosnowski wrote:
> >
> > This patch fixes a bug in
> > mlx5_flow_dv_discover_counter_offset_support()
> > uncovered by a warning reported by MSVC:
> >
> > ../drivers/net/mlx5/mlx5_flow_dv.c(
> SHAKE is not supported by these hence I did not catch it.
OpenSSL PMD supports SHA and SHAKE. Could that help verifying both lines of fix
?.
Thanks,
Gowrishankar.
The title is not consistent with other commits in this library.
On 8/12/25 4:33 AM, Pravin M Bathija wrote:
- modify data structures and add functions to support
add and remove memory regions/slots
- define VHOST_MEMORY_MAX_NREGIONS & modify function
vhost_user_set_mem_table accordingly
-
Answers inline.
Internal Use - Confidential
> -Original Message-
> From: Maxime Coquelin
> Sent: Friday, August 29, 2025 2:18 AM
> To: Bathija, Pravin ; Bruce Richardson
> ; David Marchand
> ; Chenbo Xia
> Cc: dev@dpdk.org; pravin.m.bathija@gmail.com; Thomas Monjalon
>
> Subject: R
Looks good to me.
Reviewed-by: Luca Vizzarro
Reviewed-by: Luca Vizzarro
> -Original Message-
> From: Gowrishankar Muthukrishnan
> Sent: Friday 29 August 2025 11:23
> To: Finn, Emma
> Cc: dev@dpdk.org; Ji, Kai
> Subject: RE: [EXTERNAL] [PATCH] examples/fips_validation: fix dangling pointer
> for SHA test
> Importance: High
>
> > > > rte_free(vec.pt.
> -Original Message-
> From: Gowrishankar Muthukrishnan
> Sent: Thursday 28 August 2025 10:42
> To: Finn, Emma
> Cc: dev@dpdk.org; Ji, Kai
> Subject: RE: [EXTERNAL] [PATCH] examples/fips_validation: fix dangling
> pointer for SHA test
>
> > Set vector pointer to NULL after free to avoid
>
> On Thu, 28 Aug 2025 10:23:05 +
> Ciara Loftus wrote:
>
> > Introduce the boolean restore_link_state devarg for the ice driver. When
> > set, when the ice port is stopped, the physical link will be restored to
> > the state it was in when the device was started.
> >
> > Prior to this patc
Introduce the device argument link_state_on_close for the ice driver.
When the device is stopped or closed, the link state will be configured
as follows for each of the three options:
1. down: bring (or keep) the link down
2. up: bring (or keep) the link up
3. initial: restore the link to the state
Previously, mac_own was a bit-field indicating the MAC addresses owned by
the PMD, specifically marking those added to VFs by the PMD and allowing
them to be flushed during exit.
With this commit, the bit-field now applies to all MAC addresses added by
the PMD, including those for PF, VF, and SFs.
A bug was introduced by commit 2d0665a7f771 ("net/mlx5: align PF and
VF/SF MAC address handling"). Flows for multicast MAC addresses would
not be created anymore.
This patch series are to fix the issue.
Minggang Li (Gavin) (2):
net/mlx5: update how MAC address bit-fields are used
net/mlx5: ad
Hi Pravin,
On 8/29/25 11:11 AM, Bathija, Pravin wrote:
Dear Maxime, Bruce, David, Chenbo,
Could you please help review the code ? The goal is to merge the changes before
September 18 when the merge window closes for SPDK v25.09.
I will help review the code, but there are no guarantee it will
Dear Maxime, Bruce, David, Chenbo,
Could you please help review the code ? The goal is to merge the changes before
September 18 when the merge window closes for SPDK v25.09.
Regards,
Pravin
Internal Use - Confidential
> -Original Message-
> From: Bathija, Pravin
> Sent: Tuesday, August
Device multicast MAC addresses are managed using the mac_addr_add and
mac_addr_remove APIs.
In the mlx5_dev_spawn function, devices such as PF, VFs, and SFs obtain
the MAC addresses configured in netdev via netlink and store them in the
PMD device data, which also includes multicast MAC addresses.
Hi Dean,
This patch has now been rolled back due to a breaking doc building
issue. This is now pending v4.
Best,
Luca
This series introduces multiple updates to the Yunsilicon xsc PMD,
including new feature support, optimizations, bug fixes.
Highlights include:
* Firmware version query support
* TSO and FEC support
* EEPROM dump and promiscuous mode
* Link status and event support
* Optimizations for RSS, QP/CQ a
CQEs are not cache line aligned, which can cause cache conflicts when
polling Rx CQs under high-PPS load.
This patch checks CQ producer/consumer indices before processing packets,
reducing cache conflicts. The added MMIO read overhead is minor
compared to the latency caused by cache conflicts, imp
Add recent updates of xsc PMD to 25.11 release notes.
Signed-off-by: Renyong Wan
---
doc/guides/rel_notes/release_25_11.rst | 13 +
1 file changed, 13 insertions(+)
diff --git a/doc/guides/rel_notes/release_25_11.rst
b/doc/guides/rel_notes/release_25_11.rst
index 122029b3c8..f1cb43
Previously, the PCT manager was shared across all ports, which prevented
two ports from operating independently in separate DPDK processes. This
patch makes the PCT manager per-port, allowing each port to manage its
resources separately.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
d
Refactor the stop and close paths for xsc PMD.
Previously, stopping or closing queues could leave partially freed
resources and relied on flags to track queue initialization.
This patch ensures all queue resources are properly released, improves
code clarity, and makes stop/close operations safer
Coverity reported this issue: Using uninitialized value mac when calling
xsc_ethdev_mac_addr_add.
Coverity issue: 456592
Fixes: 0b0f9c161e ("net/xsc: initialize representors")
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
drivers/net/xsc/xsc_ethdev.c | 2 +-
1 file changed, 1 inserti
Add VFIO MSI-X interrupt support and handle Link Status Change events.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_cmd.h| 29
drivers/net/xsc/xsc_defs.h | 2 +
drivers/net/xsc/xsc_dev.c| 27
Refactor memory allocation for XSC QP and CQ to use the local NUMA node.
Previously, allocations used SOCKET_ID_ANY, which could lead to remote
memory accesses. This patch ensures PAS and CQE memory are allocated
on the same NUMA node as the device, improving locality and performance.
Signed-off-b
Refactor RSS RX queue creation for the XSC PMD.
This patch introduces proper QPN allocation and per-QP info setup,
improving memory management and error handling. It makes RSS
queue creation more robust, avoiding partial resource leaks in
case of failures.
Signed-off-by: Rong Qian
Signed-off-by:
Add functions to get and set FEC mode for xsc PMD.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_cmd.h| 24 ++
drivers/net/xsc/xsc_dev.c| 75
drivers/net/xsc/xsc_dev.h
Add support for getting and setting the link state.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_cmd.h| 73 +++-
drivers/net/xsc/xsc_dev.c| 21 ++-
drivers/net/xsc/xsc_dev.h|
Introduce TCP segmentation offload (TSO) support for xsc PMD.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_ethdev.c | 3 +--
drivers/net/xsc/xsc_tx.c | 12 +++-
drivers/net/xsc/xsc_tx.h | 3 +--
Add support for enabling and disabling promiscuous mode on XSC NIC ports.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_defs.h | 1 +
drivers/net/xsc/xsc_ethdev.c | 58 ++--
drivers/net
Add support for querying and reading the EEPROM of SFP/QSFP modules.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_cmd.h| 15 +++
drivers/net/xsc/xsc_dev.c| 185 +++
drivers/net/
Add support for querying the firmware version of the xsc device.
Signed-off-by: Rong Qian
Signed-off-by: Renyong Wan
---
doc/guides/nics/features/xsc.ini | 1 +
drivers/net/xsc/xsc_dev.c| 13 +
drivers/net/xsc/xsc_dev.h| 4
drivers/net/xsc/xsc_ethdev.c |
59 matches
Mail list logo
