From: Nandini Persad
This document was created to assist contributors in creating DPDK drivers
and provides suggestions and guidelines on how to upstream effectively.
Signed-off-by: Nandini Persad
---
v2 - text should be i file not patch
doc/guides/contributing/index.rst | 1 +
doc/gui
Calling memcpy with a null pointer even if zero length is
undefined, so check if data_length is zero.
Problem reported by Gcc analyzer.
Signed-off-by: Stephen Hemminger
---
drivers/net/tap/tap_netlink.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/net/tap/tap_net
https://bugs.dpdk.org/show_bug.cgi?id=1521
Bug ID: 1521
Summary: Potential null dereference in mlx5
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: UNCONFIRMED
Severity: normal
P
> From: Stephen Hemminger [mailto:step...@networkplumber.org]
>
> On Tue, 13 Aug 2024 22:53:55 +0200
> Morten Brørup wrote:
>
> > Stephen,
> >
> > Kernel 4.19 will reach EOL in December 2024 [1], so we should skip
> ahead to 5.4 (projected EOL December 2025) or 5.10 (projected EOL
> December 202
On Tue, 13 Aug 2024 22:53:55 +0200
Morten Brørup wrote:
> Stephen,
>
> Kernel 4.19 will reach EOL in December 2024 [1], so we should skip ahead to
> 5.4 (projected EOL December 2025) or 5.10 (projected EOL December 2026).
>
> For reference, the CIP project provides a SLTS (Super Long-Time Stab
On Tue, 13 Aug 2024 22:53:55 +0200
Morten Brørup wrote:
> Stephen,
>
> Kernel 4.19 will reach EOL in December 2024 [1], so we should skip ahead to
> 5.4 (projected EOL December 2025) or 5.10 (projected EOL December 2026).
>
> For reference, the CIP project provides a SLTS (Super Long-Time Stab
Stephen,
Kernel 4.19 will reach EOL in December 2024 [1], so we should skip ahead to 5.4
(projected EOL December 2025) or 5.10 (projected EOL December 2026).
For reference, the CIP project provides a SLTS (Super Long-Time Stable) kernel
based on version 5.10 [2]. Their previous SLTS was based o
This document was created to assist contributors in
creating DPDK drivers, providing suggestions and
guidelines on how to upstream effectively.
Signed-off-by: Ferruh Yigit ,
Thomas Monjalon ,
Nandini Persad
---
0001-doc-add-new-driver-guidelines.patch | 206 +++
doc/guides/co
Reviewed-by: Dean Marx
Reviewed-by: Dean Marx
Good catch!
Reviewed-by: Nicholas Pratte
On Tue, Aug 13, 2024 at 1:17 PM Dean Marx wrote:
>
> correct docstring error in conf.yaml showing incorrect
> example pci address for TG nodes
>
> Signed-off-by: Dean Marx
> ---
> dts/conf.yaml | 8
> 1 file changed, 4 insertions(+), 4 deletio
The next DPDK Summit will be hosted in Montreal, Canada on September 24-25.
We will talk about the future of DPDK, the best userland networking libraries
having an incredible hardware support from our large community.
It will be an opportunity to connect, learn and collaborate with developers
fro
correct docstring error in conf.yaml showing incorrect
example pci address for TG nodes
Signed-off-by: Dean Marx
---
dts/conf.yaml | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/dts/conf.yaml b/dts/conf.yaml
index 7d95016e68..ca5e87636e 100644
--- a/dts/conf.yaml
+++
This API was used only for cfgfile tests and was never built
after the conversion to meson. Will be replaced by simpler
method of doing cfgfile tests.
Signed-off-by: Stephen Hemminger
Acked-by: Bruce Richardson
---
app/test/meson.build | 2 -
app/test/resource.c | 276 ---
These tests were not built since the conversion to meson.
Instead of using embedded resource functions, put data in include
file and generate temporary file before the test.
The changes to app/test/meson.build are to handle auto-generated
files (resources) differently. Don't scan these files to lo
Simple script to read a file and make it into a initialized
C string in a header file.
Signed-off-by: Stephen Hemminger
Acked-by: Bruce Richardson
---
buildtools/gen-header.py | 36
buildtools/meson.build | 2 +-
2 files changed, 37 insertions(+), 1 delet
The input files don't need to be in a separate subdirectory.
Signed-off-by: Stephen Hemminger
Acked-by: Bruce Richardson
---
app/test/test_cfgfiles/{etc => }/empty.ini| 0
app/test/test_cfgfiles/{etc => }/empty_key_value.ini | 0
app/test/test_cfgfiles/{etc => }/invalid_section.ini
The cfgfile tests did not get built since conversion to meson
and they used an awkward way to manage the test data.
This patchset converts the tests to use a helper to take
text file and make it into a C header. Then use the C header
to generate temporary files as needed.
v6 - rearrange patch ord
Thanks,
I have addressed the documentation changes with patch set v8. If it all looks
good let me know.
add dataclass for passing in flow rule creation arguments, as well as a
__str__ method for converting to a sendable testpmd command. Add
flow_create method to TestPmdShell class for initializing flow rules.
Signed-off-by: Dean Marx
---
dts/framework/remote_session/testpmd_shell.py | 57 +
add dataclass for passing in flow rule creation arguments, as well as a
__str__ method for converting to a sendable testpmd command. Add
flow_create method to TestPmdShell class for initializing flow rules.
Signed-off-by: Dean Marx
---
dts/framework/remote_session/testpmd_shell.py | 57 +
https://bugs.dpdk.org/show_bug.cgi?id=1520
Bug ID: 1520
Summary: Some drivers still use RTE_LIBRTE_ETHDEV_DEBUG instead
of RTE_ETHDEV_DEBUG_RX/TX
Product: DPDK
Version: 24.07
Hardware: All
OS: All
> From: Stephen Hemminger [mailto:step...@networkplumber.org]
>
> Redefining memcpy as rte_memcpy has no performance gain on
> current compilers, and introduced bugs like this one where
> rte_memcpy() will be detected as referencing past the destination.
>
> Bugzilla ID: 1510
> Fixes: 142778b3702
在 2024/8/13 5:56, Stephen Hemminger 写道:
Only in one place does the power library need un-constify a pointer
and this can be done by casting to uintptr_t first. Better, to have
the warning enabled across the rest of the code.
Signed-off-by: Stephen Hemminger
---
Acked-by: Huisong Li
On Mon, Aug 12, 2024 at 02:56:09PM -0700, Stephen Hemminger wrote:
> Only in one place does the power library need un-constify a pointer
> and this can be done by casting to uintptr_t first. Better, to have
> the warning enabled across the rest of the code.
>
> Signed-off-by: Stephen Hemminger
>
My bad, please ignore the first comment, I've confused the - and + on
the left side.
sorry for the misunderstanding.
For the second I still think would make more sense to reset the
statistics at
when a queue is started rather than stopped.
On 13/08/2024 09:24, Mattia Milani wrote:
Dear all,
Dear all,
I'm a DPDK newbie, but I hope anyway you are gonna consider my
comments or help me to understand the code a little bit better.
1) Aren't you updating the variable two times? for example, in
'acc100_enqueue_enc_cb',
you update both 'enqueued_count' and 'enqueue_err_count' and then
Series-
Acked-by: Hemant Agrawal
On 13-08-2024 05:11, Nicolas Chautru wrote:
v2: update to ABI doc suggested by Maxime.
These series include introducing a new paramter in the queue stat
which can be used to monitor the number of available enqueue
still possible.
The acc PMD is then refacto
28 matches
Mail list logo
