graph: add more infos to node dump

The function node_dump add some printing of node information.

Signed-off-by: Huichao cai <chcchc88@163.com>
Acked-by: Kiran Kumar Kokkilagadda <kirankumark@marvel

graph: add more infos to node dump

The function node_dump add some printing of node information.

Signed-off-by: Huichao cai <chcchc88@163.com>
Acked-by: Kiran Kumar Kokkilagadda <kirankumark@marvell.com>

show more ...

graph: improve node layout

The members "dispatch" and "xstat_off" of the structure "rte_node"
can be min cache aligned to make room for future expansion and to
make sure have better performance. Add

graph: improve node layout

The members "dispatch" and "xstat_off" of the structure "rte_node"
can be min cache aligned to make room for future expansion and to
make sure have better performance. Add corresponding comments.

Signed-off-by: Huichao Cai <chcchc88@163.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

show more ...

graph: fix memory leak in node clone

Free memory allocated for the node when xstats memory allocation fails.

Coverity issue: 445529
Fixes: 070db97e017b ("graph: support node xstats")

Signed-off-by

graph: fix memory leak in node clone

Free memory allocated for the node when xstats memory allocation fails.

Coverity issue: 445529
Fixes: 070db97e017b ("graph: support node xstats")

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...

graph: support node xstats

Add ability for nodes to advertise xstat counters during registration
and increment them in fastpath.
Add support for retrieving/printing stats for node specific counters

graph: support node xstats

Add ability for nodes to advertise xstat counters during registration
and increment them in fastpath.
Add support for retrieving/printing stats for node specific counters
using rte_graph_cluster_stats_get().
Add rte_node_xstat_increment() API to increment node specific counters.

Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>
Reviewed-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>

show more ...

graph: expose node context as pointers

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 structu

graph: expose node context as pointers

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 structures, casting to opaque pointers is hard without
violating strict aliasing rules.

Add an unnamed union to allow storing opaque pointers in the node
context. Unfortunately, aligning an unnamed union that contains an array
produces inconsistent results between C and C++. To preserve ABI/API
compatibility in both C and C++, move all fast-path area fields into an
unnamed struct which is itself cache aligned. Use __rte_cache_aligned to
preserve existing alignment on architectures where cache lines are 128
bytes.

Add a static assert to ensure that the fast path area does not grow
beyond a 64 bytes cache line.

Signed-off-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...

graph: make graphviz export more readable

Change the color of arrows leading to sink nodes to dark orange. Remove
the node oval shape around the sink nodes and make their text dark
orange. This resu

graph: make graphviz export more readable

Change the color of arrows leading to sink nodes to dark orange. Remove
the node oval shape around the sink nodes and make their text dark
orange. This results in a much more readable output for large graphs.
See the link below for an example.

Link: https://f.jarry.cc/rte-graph-dot/ipv6.svg
Signed-off-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...

use C linkage where appropriate in headers

Assure that 'extern "C" { /../ }' do not cover files included from a
particular header file, and address minor issues resulting from this
change of order.

use C linkage where appropriate in headers

Assure that 'extern "C" { /../ }' do not cover files included from a
particular header file, and address minor issues resulting from this
change of order.

Dealing with C++ should delegate to the individual include file level,
rather than being imposed by the user of that file. For example,
forcing C linkage prevents __Generic macros being replaced with
overloaded static inline functions in C++ translation units.

Eliminate 'extern "C"' from files which do not declare any symbols
(e.g., only macros or struct types).

On the other hand, the headers check is too naive in assuming that all
headers must contain a 'extern "C"'. Such a check was added in commit
1ee492bdc4ff ("buildtools/chkincs: check missing C++ guards").
Since this current change results in many headers not containing such
a token, remove the check for 'extern "C"' until we have a better
implementation.

Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>

show more ...

version: 24.11-rc0

Start a new release cycle with empty release notes.

The ABI version becomes 25.0.
The map files are updated to the new ABI major number (25).
The ABI exceptions are dropped and C

version: 24.11-rc0

Start a new release cycle with empty release notes.

The ABI version becomes 25.0.
The map files are updated to the new ABI major number (25).
The ABI exceptions are dropped and CI ABI checks are disabled because
compatibility is not preserved.

Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>

show more ...

graph: fix mcore dispatch walk

Head move happens before the core id check, which will cause the last
source node be executed even core id is not correct. This patch changes
head check to less than 1

graph: fix mcore dispatch walk

Head move happens before the core id check, which will cause the last
source node be executed even core id is not correct. This patch changes
head check to less than 1 instead of 0 to fix this issue.

Fixes: 35dfd9b9fd85 ("graph: introduce graph walk by cross-core dispatch")
Cc: stable@dpdk.org

Signed-off-by: Jingjing Wu <jingjing.wu@intel.com>
Acked-by: Zhirun Yan <zhirun.yan@intel.com>

show more ...

graph: enhance graphviz export

* 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 i

graph: enhance graphviz export

* 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 processing generated graph.
* Force bold blue border and arrows for source nodes.
* Force dark orange borders for sink nodes (nodes without next nodes).

Link: https://f.jarry.cc/rte-graph-dot/before.svg
Link: https://f.jarry.cc/rte-graph-dot/after.svg

Signed-off-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...

graph: fix stats retrieval while destroying a graph

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
updat

graph: fix stats retrieval while destroying a graph

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 by another thread while
rte_graph_cluster_stats_get is called, it can result in undefined
behaviour.

Adding a lock would make it impossible to call
rte_graph_cluster_stats_get in packet processing code paths. Avoid
accessing the global list instead by storing a bool field in the private
rte_graph_cluster_stats structure.

Also update the default callback to avoid accessing the global list and
use a different default callback depending on the graph model.

Fixes: 358ff83fe88c ("graph: add stats for mcore dispatch model")
Cc: stable@dpdk.org

Signed-off-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...

graph: fix ID collisions

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

graph: fix ID collisions

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 been created.

The following code produces duplicate graph IDs which can lead to
use-after-free bugs and other undefined behaviours:

a = rte_graph_create(...); // id=0 graph_id=1
b = rte_graph_create(...); // id=1 graph_id=2
rte_graph_destroy(a); // graph_id=1
c = rte_graph_create(...); // id=1 graph_id=2 (duplicate with b)
rte_graph_destroy(c); // frees memory still used by b

Remove the global counter. Make sure that the graph list is always
ordered by increasing graph ids. When creating a new graph, pick a free
id which is not allocated.

Update unit tests to ensure it works as expected.

Fixes: a91fecc19c5c ("graph: implement create and destroy")
Cc: stable@dpdk.org

Signed-off-by: Robin Jarry <rjarry@redhat.com>
Acked-by: Kiran Kumar K <kirankumark@marvell.com>

show more ...