Alexis Lothoré (eBPF Foundation) says:
====================
this series aims to bring test_tcp_check_syncookie.sh scope into
test_progs to make sure that the corresponding tests are also run
automatically in CI. This script tests for bpf_tcp_{gen,check}_syncookie
and bpf_skc_lookup_tcp, in different contexts (ipv4, v6 or dual, and
with tc and xdp programs).
Some other tests like btf_skc_cls_ingress have some overlapping tests with
test_tcp_check_syncookie.sh, so this series moves the missing bits from
test_tcp_check_syncookie.sh into btf_skc_cls_ingress, which is already
integrated into test_progs.
- the first three commits bring some minor improvements to
btf_skc_cls_ingress without changing its testing scope
- fourth and fifth commits bring test_tcp_check_syncookie.sh features
into btf_skc_cls_ingress
- last commit removes test_tcp_check_syncookie.sh
The only topic for which I am not sure for this integration is the
necessity or not to run the tests with different program types:
test_tcp_check_syncookie.sh runs tests with both tc and xdp programs, but
btf_skc_cls_ingress currently tests those helpers only with a tc
program. Would it make sense to also make sure that btf_skc_cls_ingress
is tested with all the programs types supported by those helpers ?
The series has been tested both in CI and in a local x86_64 qemu
environment:
# ./test_progs -a btf_skc_cls_ingress
#38/1 btf_skc_cls_ingress/conn_ipv4:OK
#38/2 btf_skc_cls_ingress/conn_ipv6:OK
#38/3 btf_skc_cls_ingress/conn_dual:OK
#38/4 btf_skc_cls_ingress/syncookie_ipv4:OK
#38/5 btf_skc_cls_ingress/syncookie_ipv6:OK
#38/6 btf_skc_cls_ingress/syncookie_dual:OK
#38 btf_skc_cls_ingress:OK
Summary: 1/6 PASSED, 0 SKIPPED, 0 FAILED
---
Changes in v2:
- fix initial test author mail in Cc
- Fix default cases in switches: indent, action
- remove unneeded initializer
- remove duplicate interface bring-up
- remove unnecessary check and return in bpf program
- Link to v1: https://lore.kernel.org/r/20241016-syncookie-v1-0-3b7a0de12153@bootlin.com
====================
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
btf_skc_cls_ingress test currently checks that syncookie and
bpf_sk_assign/release helpers behave correctly in multiple scenarios,
but only with ipv6 socket.
Increase those helpers coverage by adding testing support for IPv4
sockets and IPv4/IPv6 sockets. The rework is mostly based on features
brought earlier in test_tcp_check_syncookie.sh to cover some fixes
performed on those helpers, but test_tcp_check_syncookie.sh is not
integrated in test_progs. The most notable changes linked to this are:
- some rework in the corresponding eBPF program to support both types of
traffic
- the switch from start_server to start_server_str to allow to check
some socket options
- the introduction of new subtests for ipv4 and ipv4/ipv6
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@bootlin.com>
Link: https://lore.kernel.org/r/20241020-syncookie-v2-4-2db240225fed@bootlin.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
There are a few global variables in btf_skc_cls_ingress.c, which are not
really used by different tests. Get rid of those global variables, by
performing the following updates:
- make srv_sa6 local to the main runner function
- make skel local to the main function, and propagate it through
function arguments
- get rid of duration by replacing CHECK macros with the ASSERT_XXX
macros. While updating those assert macros:
- do not return early on asserts performing some actual tests, let the
other tests run as well (keep the early return for parts handling
test setup)
- instead of converting the CHECK on skel->bss->linum, just remove it,
since there is already a call to print_err_line after the test to
print the failing line in the bpf program
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@bootlin.com>
Link: https://lore.kernel.org/r/20241020-syncookie-v2-3-2db240225fed@bootlin.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
btf_skc_cls_ingress.c currently runs two subtests, and create a
dedicated network namespace for each, but never cleans up the created
namespace once the test has ended.
Add missing namespace cleanup after each subtest to avoid accumulating
namespaces for each new subtest. While at it, switch namespace
management to netns_{new,free}
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@bootlin.com>
Link: https://lore.kernel.org/r/20241020-syncookie-v2-2-2db240225fed@bootlin.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
btf_skc_cls_ingress currently describe two tests, both running a simple
tcp server and then initializing a connection to it. The sole difference
between the tests is about the tcp_syncookie configuration, and some
checks around this feature being enabled/disabled.
Share the common code between those two tests by moving the code into a
single runner, parameterized by a "gen_cookies" argument. Split the
performed checks accordingly.
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@bootlin.com>
Link: https://lore.kernel.org/r/20241020-syncookie-v2-1-2db240225fed@bootlin.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Zijian Zhang says:
====================
Function msg_verify_data should have context of bytes_cnt and k instead of
assuming they are zero. Otherwise, test_sockmap with data integrity test
will report some errors. I also fix the logic related to size and index j
1/ 6 sockmap::txmsg test passthrough:FAIL
2/ 6 sockmap::txmsg test redirect:FAIL
7/12 sockmap::txmsg test apply:FAIL
10/11 sockmap::txmsg test push_data:FAIL
11/17 sockmap::txmsg test pull-data:FAIL
12/ 9 sockmap::txmsg test pop-data:FAIL
13/ 1 sockmap::txmsg test push/pop data:FAIL
...
Pass: 24 Fail: 52
After fixing msg_verify_data, some of the errors are solved, but for push
pull and pop, we may need more fixes to msg_verify_data, added a TODO
10/11 sockmap::txmsg test push_data:FAIL
11/17 sockmap::txmsg test pull-data:FAIL
12/ 9 sockmap::txmsg test pop-data:FAIL
...
Pass: 37 Fail: 15
Besides, added a custom errno EDATAINTEGRITY for msg_verify_data, we
shall not ignore the error in txmsg_cork case, and fixed the txmsg_redir
in test_txmsg_pull "Test pull + redirect" case.
====================
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Function msg_verify_data should have context of bytes_cnt and k instead of
assuming they are zero. Otherwise, test_sockmap with data integrity test
will report some errors. I also fix the logic related to size and index j
1/ 6 sockmap::txmsg test passthrough:FAIL
2/ 6 sockmap::txmsg test redirect:FAIL
7/12 sockmap::txmsg test apply:FAIL
10/11 sockmap::txmsg test push_data:FAIL
11/17 sockmap::txmsg test pull-data:FAIL
12/ 9 sockmap::txmsg test pop-data:FAIL
13/ 1 sockmap::txmsg test push/pop data:FAIL
...
Pass: 24 Fail: 52
After applying this patch, some of the errors are solved, but for push,
pull and pop, we may need more fixes to msg_verify_data, added a TODO
10/11 sockmap::txmsg test push_data:FAIL
11/17 sockmap::txmsg test pull-data:FAIL
12/ 9 sockmap::txmsg test pop-data:FAIL
...
Pass: 37 Fail: 15
Besides, added a custom errno EDATAINTEGRITY for msg_verify_data, we
shall not ignore the error in txmsg_cork case.
Fixes: 753fb2ee09 ("bpf: sockmap, add msg_peek tests to test_sockmap")
Fixes: 16edddfe3c ("selftests/bpf: test_sockmap, check test failure")
Acked-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Zijian Zhang <zijianzhang@bytedance.com>
Link: https://lore.kernel.org/r/20241012203731.1248619-2-zijianzhang@bytedance.com
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Simon Wunderlich says:
====================
This cleanup patchset includes the following patches:
- bump version strings, by Simon Wunderlich
- Add flex array to struct batadv_tvlv_tt_data, by Erick Archer
- Use string choice helper to print booleans, by Sven Eckelmann
- replace call_rcu by kfree_rcu for simple kmem_cache_free callback,
by Julia Lawall
* tag 'batadv-next-pullrequest-20241015' of git://git.open-mesh.org/linux-merge:
batman-adv: replace call_rcu by kfree_rcu for simple kmem_cache_free callback
batman-adv: Use string choice helper to print booleans
batman-adv: Add flex array to struct batadv_tvlv_tt_data
batman-adv: Start new development cycle
====================
Link: https://patch.msgid.link/20241015073946.46613-1-sw@simonwunderlich.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Daniel Borkmann says:
====================
pull-request: bpf-next 2024-10-14
The following pull-request contains BPF updates for your *net-next* tree.
We've added 21 non-merge commits during the last 18 day(s) which contain
a total of 21 files changed, 1185 insertions(+), 127 deletions(-).
The main changes are:
1) Put xsk sockets on a struct diet and add various cleanups. Overall, this helps
to bump performance by 12% for some workloads, from Maciej Fijalkowski.
2) Extend BPF selftests to increase coverage of XDP features in combination
with BPF cpumap, from Alexis Lothoré (eBPF Foundation).
3) Extend netkit with an option to delegate skb->{mark,priority} scrubbing to
its BPF program, from Daniel Borkmann.
4) Make the bpf_get_netns_cookie() helper available also to tc(x) BPF programs,
from Mahe Tardy.
5) Extend BPF selftests covering a BPF program setting socket options per MPTCP
subflow, from Geliang Tang and Nicolas Rybowski.
bpf-next-for-netdev
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (21 commits)
xsk: Use xsk_buff_pool directly for cq functions
xsk: Wrap duplicated code to function
xsk: Carry a copy of xdp_zc_max_segs within xsk_buff_pool
xsk: Get rid of xdp_buff_xsk::orig_addr
xsk: s/free_list_node/list_node/
xsk: Get rid of xdp_buff_xsk::xskb_list_node
selftests/bpf: check program redirect in xdp_cpumap_attach
selftests/bpf: make xdp_cpumap_attach keep redirect prog attached
selftests/bpf: fix bpf_map_redirect call for cpu map test
selftests/bpf: add tcx netns cookie tests
bpf: add get_netns_cookie helper to tc programs
selftests/bpf: add missing header include for htons
selftests/bpf: Extend netkit tests to validate skb meta data
tools: Sync if_link.h uapi tooling header
netkit: Add add netkit scrub support to rt_link.yaml
netkit: Simplify netkit mode over to use NLA_POLICY_MAX
netkit: Add option for scrubbing skb meta data
bpf: Remove unused macro
selftests/bpf: Add mptcp subflow subtest
selftests/bpf: Add getsockopt to inspect mptcp subflow
...
====================
Link: https://patch.msgid.link/20241014211110.16562-1-daniel@iogearbox.net
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Timestamp values are read using pointers to 64-bit big endian values.
But the type of these pointers is u64 *, host byte order.
Use __be64 * instead.
Flagged by Sparse:
.../gianfar.c:2212:60: warning: cast to restricted __be64
.../gianfar.c:2475:53: warning: cast to restricted __be64
Introduced by
commit cc772ab7cd ("gianfar: Add hardware RX timestamping support").
Compile tested only.
No functional change intended.
Signed-off-by: Simon Horman <horms@kernel.org>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Link: https://patch.msgid.link/20241011-gianfar-be64-v1-1-a77ebe972176@kernel.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
YNL specs can use string expressions for limits, like s32-min
or u16-max. We convert all of those into their numeric values
when generating the code, which isn't always helpful. Try to
retain the string representations in the output. Any sort of
calculations still need the integers.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Joe Damato <jdamato@fastly.com>
Link: https://patch.msgid.link/20241010151248.2049755-1-kuba@kernel.org
[pabeni@redhat.com: regenerated netdev-genl-gen.c]
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Use newly defined 'active-high' property to set the
VEND1_GLOBAL_LED_DRIVE_VDD bit and let 'active-low' clear that bit. This
reflects the technical reality which was inverted in the previous
description in which the 'active-low' property was used to actually set
the VEND1_GLOBAL_LED_DRIVE_VDD bit, which means that VDD (ie. supply
voltage) of the LED is driven rather than GND.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/86a413b4387c42dcb54f587cc2433a06f16aae83.1728558223.git.daniel@makrotopia.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Wei Fang says:
====================
make PHY output RMII reference clock
The TJA11xx PHYs have the capability to provide 50MHz reference clock
in RMII mode and output on REF_CLK pin. Therefore, add the new property
"nxp,rmii-refclk-output" to support this feature. This property is only
available for PHYs which use nxp-c45-tja11xx driver, such as TJA1103,
TJA1104, TJA1120 and TJA1121.
====================
Link: https://patch.msgid.link/20241010061944.266966-1-wei.fang@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
For TJA11xx PHYs, they have the capability to output 50MHz reference
clock on REF_CLK pin in RMII mode, which is called "revRMII" mode in
the PHY data sheet.
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Per the RMII specification, the REF_CLK is sourced from MAC to PHY
or from an external source. But for TJA11xx PHYs, they support to
output a 50MHz RMII reference clock on REF_CLK pin. Previously the
"nxp,rmii-refclk-in" was added to indicate that in RMII mode, if
this property present, REF_CLK is input to the PHY, otherwise it
is output. This seems inappropriate now. Because according to the
RMII specification, the REF_CLK is originally input, so there is
no need to add an additional "nxp,rmii-refclk-in" property to
declare that REF_CLK is input.
Unfortunately, because the "nxp,rmii-refclk-in" property has been
added for a while, and we cannot confirm which DTS use the TJA1100
and TJA1101 PHYs, changing it to switch polarity will cause an ABI
break. But fortunately, this property is only valid for TJA1100 and
TJA1101. For TJA1103/TJA1104/TJA1120/TJA1121 PHYs, this property is
invalid because they use the nxp-c45-tja11xx driver, which is a
different driver from TJA1100/TJA1101. Therefore, for PHYs using
nxp-c45-tja11xx driver, add "nxp,rmii-refclk-out" property to
support outputting RMII reference clock on REF_CLK pin.
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Try to rebuild YNL if either user added a new family or the specs
of the families have changed. Stanislav's ncdevmem cause a false
positive build failure in NIPA because libynl.a isn't rebuilt
after ethtool is added to YNL_GENS.
Note that sha1sum is already used in other parts of the build system.
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20241011230311.2529760-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Joe Damato says:
====================
Add support for per-NAPI config via netlink
Greetings:
Welcome to v6. Minor changes from v5 [1], please see changelog below.
There were no explicit comments from reviewers on the call outs in my
v5, so I'm retaining them from my previous cover letter just in case :)
A few important call outs for reviewers:
1. This revision seems to work (see below for a full walk through). I
think this is the behavior we talked about, but please let me know if
a use case is missing.
2. Re a previous point made by Stanislav regarding "taking over a NAPI
ID" when the channel count changes: mlx5 seems to call napi_disable
followed by netif_napi_del for the old queues and then calls
napi_enable for the new ones. In this RFC, the NAPI ID generation is
deferred to napi_enable. This means we won't end up with two of the
same NAPI IDs added to the hash at the same time.
Can we assume all drivers will napi_disable the old queues before
napi_enable the new ones?
- If yes: we might not need to worry about a NAPI ID takeover
function.
- If no: I'll need to make a change so that the NAPI ID generation
is deferred only for drivers which have opted into the config
space via calls to netif_napi_add_config
3. I made the decision to remove the WARN_ON_ONCE that (I think?)
Jakub previously suggested in alloc_netdev_mqs (WARN_ON_ONCE(txqs
!= rxqs);) because this was triggering on every kernel boot with my
mlx5 NIC.
4. I left the "maxqs = max(txqs, rxqs);" in alloc_netdev_mqs despite
thinking this is a bit strange. I think it's strange that we might
be short some number of NAPI configs, but it seems like most people
are in favor of this approach, so I've left it.
I'd appreciate thoughts from reviewers on the above items, if at all
possible.
Now, on to the implementation.
Firstly, this implementation moves certain settings to napi_struct so that
they are "per-NAPI", while taking care to respect existing sysfs
parameters which are interface wide and affect all NAPIs:
- NAPI ID
- gro_flush_timeout
- defer_hard_irqs
Furthermore:
- NAPI ID generation and addition to the hash is now deferred to
napi_enable, instead of during netif_napi_add
- NAPIs are removed from the hash during napi_disable, instead of
netif_napi_del.
- An array of "struct napi_config" is allocated in net_device.
IMPORTANT: The above changes affect all network drivers.
Optionally, drivers may opt-in to using their config space by calling
netif_napi_add_config instead of netif_napi_add.
If a driver does this, the NAPI being added is linked with an allocated
"struct napi_config" and the per-NAPI settings (including NAPI ID) are
persisted even as hardware queues are destroyed and recreated.
To help illustrate how this would end up working, I've added patches for
3 drivers, of which I have access to only 1:
- mlx5 which is the basis of the examples below
- mlx4 which has TX only NAPIs, just to highlight that case. I have
only compile tested this patch; I don't have this hardware.
- bnxt which I have only compiled tested. I don't have this
hardware.
NOTE: I only tested this on mlx5; I have no access to the other hardware
for which I provided patches. Hopefully other folks can help test :)
Here's how it works when I test it on my mlx5 system:
$ ethtool -l eth4 | grep Combined | tail -1
Combined: 2
First, output the current NAPI settings:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 0,
'gro-flush-timeout': 0,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 0,
'gro-flush-timeout': 0,
'id': 344,
'ifindex': 7,
'irq': 327}]
Now, set the global sysfs parameters:
$ sudo bash -c 'echo 20000 >/sys/class/net/eth4/gro_flush_timeout'
$ sudo bash -c 'echo 100 >/sys/class/net/eth4/napi_defer_hard_irqs'
Output current NAPI settings again:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Now set NAPI ID 345, via its NAPI ID to specific values:
$ sudo ./tools/net/ynl/cli.py \
--spec Documentation/netlink/specs/netdev.yaml \
--do napi-set \
--json='{"id": 345,
"defer-hard-irqs": 111,
"gro-flush-timeout": 11111}'
None
Now output current NAPI settings again to ensure only NAPI ID 345
changed:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 111,
'gro-flush-timeout': 11111,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Now, increase gro-flush-timeout only:
$ sudo ./tools/net/ynl/cli.py \
--spec Documentation/netlink/specs/netdev.yaml \
--do napi-set --json='{"id": 345,
"gro-flush-timeout": 44444}'
None
Now output the current NAPI settings once more:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 111,
'gro-flush-timeout': 44444,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Now set NAPI ID 345 to have gro_flush_timeout of 0:
$ sudo ./tools/net/ynl/cli.py \
--spec Documentation/netlink/specs/netdev.yaml \
--do napi-set --json='{"id": 345,
"gro-flush-timeout": 0}'
None
Check that NAPI ID 345 has a value of 0:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 111,
'gro-flush-timeout': 0,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Change the queue count, ensuring that NAPI ID 345 retains its settings:
$ sudo ethtool -L eth4 combined 4
Check that the new queues have the system wide settings but that NAPI ID
345 remains unchanged:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 347,
'ifindex': 7,
'irq': 529},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 346,
'ifindex': 7,
'irq': 528},
{'defer-hard-irqs': 111,
'gro-flush-timeout': 0,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Now reduce the queue count below where NAPI ID 345 is indexed:
$ sudo ethtool -L eth4 combined 1
Check the output:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Re-increase the queue count to ensure NAPI ID 345 is re-assigned the same
values:
$ sudo ethtool -L eth4 combined 2
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[{'defer-hard-irqs': 111,
'gro-flush-timeout': 0,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Create new queues to ensure the sysfs globals are used for the new NAPIs
but that NAPI ID 345 is unchanged:
$ sudo ethtool -L eth4 combined 8
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[...]
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 346,
'ifindex': 7,
'irq': 528},
{'defer-hard-irqs': 111,
'gro-flush-timeout': 0,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 100,
'gro-flush-timeout': 20000,
'id': 344,
'ifindex': 7,
'irq': 327}]
Last, but not least, let's try writing the sysfs parameters to ensure
all NAPIs are rewritten:
$ sudo bash -c 'echo 33333 >/sys/class/net/eth4/gro_flush_timeout'
$ sudo bash -c 'echo 222 >/sys/class/net/eth4/napi_defer_hard_irqs'
Check that worked:
$ ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/netdev.yaml \
--dump napi-get --json='{"ifindex": 7}'
[...]
{'defer-hard-irqs': 222,
'gro-flush-timeout': 33333,
'id': 346,
'ifindex': 7,
'irq': 528},
{'defer-hard-irqs': 222,
'gro-flush-timeout': 33333,
'id': 345,
'ifindex': 7,
'irq': 527},
{'defer-hard-irqs': 222,
'gro-flush-timeout': 33333,
'id': 344,
'ifindex': 7,
'irq': 327}]
[1]: https://lore.kernel.org/20241009005525.13651-1-jdamato@fastly.com
v5: https://lore.kernel.org/20241009005525.13651-1-jdamato@fastly.com
rfcv4: https://lore.kernel.org/lkml/20241001235302.57609-1-jdamato@fastly.com
rfcv3: https://lore.kernel.org/20240912100738.16567-8-jdamato@fastly.com
rfcv2: https://lore.kernel.org/20240908160702.56618-1-jdamato@fastly.com
====================
Link: https://patch.msgid.link/20241011184527.16393-1-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a persistent NAPI config area for NAPI configuration to the core.
Drivers opt-in to setting the persistent config for a NAPI by passing an
index when calling netif_napi_add_config.
napi_config is allocated in alloc_netdev_mqs, freed in free_netdev
(after the NAPIs are deleted).
Drivers which call netif_napi_add_config will have persistent per-NAPI
settings: NAPI IDs, gro_flush_timeout, and defer_hard_irq settings.
Per-NAPI settings are saved in napi_disable and restored in napi_enable.
Co-developed-by: Martin Karsten <mkarsten@uwaterloo.ca>
Signed-off-by: Martin Karsten <mkarsten@uwaterloo.ca>
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20241011184527.16393-6-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Allow per-NAPI gro_flush_timeout setting.
The existing sysfs parameter is respected; writes to sysfs will write to
all NAPI structs for the device and the net_device gro_flush_timeout
field. Reads from sysfs will read from the net_device field.
The ability to set gro_flush_timeout on specific NAPI instances will be
added in a later commit, via netdev-genl.
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://patch.msgid.link/20241011184527.16393-4-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add defer_hard_irqs to napi_struct in preparation for per-NAPI
settings.
The existing sysfs parameter is respected; writes to sysfs will write to
all NAPI structs for the device and the net_device defer_hard_irq field.
Reads from sysfs show the net_device field.
The ability to set defer_hard_irqs on specific NAPI instances will be
added in a later commit, via netdev-genl.
Signed-off-by: Joe Damato <jdamato@fastly.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://patch.msgid.link/20241011184527.16393-2-jdamato@fastly.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Eric Dumazet says:
====================
tcp: add skb->sk to more control packets
Currently, TCP can set skb->sk for a variety of transmit packets.
However, packets sent on behalf of a TIME_WAIT sockets do not
have an attached socket.
Same issue for RST packets.
We want to change this, in order to increase eBPF program
capabilities.
This is slightly risky, because various layers could
be confused by TIME_WAIT sockets showing up in skb->sk.
v2: audited all sk_to_full_sk() users and addressed Martin feedback.
====================
Link: https://patch.msgid.link/20241010174817.1543642-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
ip_send_unicast_reply() send orphaned 'control packets'.
These are RST packets and also ACK packets sent from TIME_WAIT.
Some eBPF programs would prefer to have a meaningful skb->sk
pointer as much as possible.
This means that TCP can now attach TIME_WAIT sockets to outgoing
skbs.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Brian Vazquez <brianvv@google.com>
Link: https://patch.msgid.link/20241010174817.1543642-6-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
tcp_v6_send_response() send orphaned 'control packets'.
These are RST packets and also ACK packets sent from TIME_WAIT.
Some eBPF programs would prefer to have a meaningful skb->sk
pointer as much as possible.
This means that TCP can now attach TIME_WAIT sockets to outgoing
skbs.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Brian Vazquez <brianvv@google.com>
Link: https://patch.msgid.link/20241010174817.1543642-5-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
TCP stack is not attaching skb to TIME_WAIT sockets yet,
but we would like to allow this in the future.
Add sk_listener_or_tw() helper to detect the three states
that FQ needs to take care.
Like NEW_SYN_RECV, TIME_WAIT are not full sockets and
do not contain sk->sk_pacing_status, sk->sk_pacing_rate.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Brian Vazquez <brianvv@google.com>
Link: https://patch.msgid.link/20241010174817.1543642-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Address byte-order miss-matches flagged by Sparse.
In tg3_load_firmware_cpu() and tg3_get_device_address()
this is done using appropriate types to store big endian values.
In the cases of tg3_test_nvram(), where buf is an array which
contains values of several different types, cast to __le32
before converting values to host byte order.
Reported by Sparse as:
.../tg3.c:3745:34: warning: cast to restricted __be32
.../tg3.c:13096:21: warning: cast to restricted __le32
.../tg3.c:13096:21: warning: cast from restricted __be32
.../tg3.c:13101:21: warning: cast to restricted __le32
.../tg3.c:13101:21: warning: cast from restricted __be32
.../tg3.c:17070:63: warning: incorrect type in argument 3 (different base types)
.../tg3.c:17070:63: expected restricted __be32 [usertype] *val
.../tg3.c:17070:63: got unsigned int *
dr.../tg3.c:17071:63: warning: incorrect type in argument 3 (different base types)
.../tg3.c:17071:63: expected restricted __be32 [usertype] *val
.../tg3.c:17071:63: got unsigned int *
Also, address white-space issues on lines modified for the above.
And, for consistency, lines adjacent to them.
Compile tested only.
No functional change intended.
Signed-off-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20241009-tg3-sparse-v1-1-6af38a7bf4ff@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
