- 08 Apr, 2024 20 commits
-
-
Geliang Tang authored
This patch exports ip_mptcp into mptcp_lib.sh as a public variable, named MPTCP_LIB_IP_MPTCP. Add a helper mptcp_lib_set_ip_mptcp() to set it, and a helper mptcp_lib_is_ip_mptcp() to test whether it is set. Use these two helpers in mptcp_join.sh. This patch is prepared for coming commits. Signed-off-by:
Geliang Tang <tanggeliang@kylinos.cn> Reviewed-by:
Matthieu Baerts (NGI0) <matttbe@kernel.org> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Geliang Tang authored
'delay 1' in tc-netem is confusing, not sure if it's a delay of 1 second or 1 millisecond. This patch explicitly adds millisecond units to make these commands clearer. Signed-off-by:
-
Geliang Tang authored
tc are used in some test scripts: mptcp_connect.sh, mptcp_join.sh and simult_flows.sh. It makes sense to check if tc is installed before running these scripts, just like other tools. So this patch add 'tc' check for mptcp_lib_check_tools(), and check it in these test scripts. Signed-off-by:
-
Eric Dumazet authored
tp->recvmsg_inq is used from tcp recvmsg() thus should be in tcp_sock_read_rx group. tp->tcp_clock_cache and tp->tcp_mstamp are written both in rx and tx paths, thus are better placed in tcp_sock_write_txrx group. Signed-off-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
Jose Ignacio Tornos Martinez authored
If the mac address can not be read from the device registers or the devicetree, a random address is generated, but this was already done from usbnet_probe, so it is not necessary to call eth_hw_addr_random from here again to generate another random address. Indeed, when reset was also executed from bind, generate another random mac address invalidated the check from usbnet_probe to configure if the assigned mac address for the interface was random or not, because it is comparing with the initial generated random address. Now, with only a reset from open operation, it is just a harmless simplification. Signed-off-by:
Jose Ignacio Tornos Martinez <jtornosm@redhat.com> Signed-off-by:
-
Michal Swiatkowski authored
From Arnd comments: "The memcpy() in the ip_tunnel_info_opts_set() causes a string.h fortification warning, with at least gcc-13: In function 'fortify_memcpy_chk', inlined from 'ip_tunnel_info_opts_set' at include/net/ip_tunnels.h:619:3, inlined from 'pfcp_encap_recv' at drivers/net/pfcp.c:84:2: include/linux/fortify-string.h:553:25: error: call to '__write_overflow_field' declared with attribute warning: detected write beyond size of field (1st parameter); maybe use struct_group()? [-Werror=attribute-warning] 553 | __write_overflow_field(p_size_field, size);" It is a false-positivie caused by ambiguity of the union. However, as Arnd noticed, copying here is unescessary. The code can be simplified to avoid calling ip_tunnel_info_opts_set(), which is doing copying, setting flags and options_len. Set correct flags and options_len directly on tun_info. Fixes: 6dd514f4 ("pfcp: always set pfcp metadata") Reported-by:Arnd Bergmann <arnd@arndb.de> Closes: https://lore.kernel.org/netdev/701f8f93-f5fb-408b-822a-37a1d5c424ba@app.fastmail.com/Acked-by:
Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Reviewed-by:
Simon Horman <horms@kernel.org> Signed-off-by:
-
David S. Miller authored
Jakub Kicinski says: ==================== selftests: net: groundwork for YNL-based tests Currently the options for writing networking tests are C, bash or some mix of the two. YAML/Netlink gives us the ability to easily interface with Netlink in higher level laguages. In particular, there is a Python library already available in tree, under tools/net. Add the scaffolding which allows writing tests using this library. The "scaffolding" is needed because the library lives under tools/net and uses YAML files from under Documentation/. So we need a small amount of glue code to find those things and add them to TEST_FILES. This series adds both a basic SW sanity test and driver test which can be run against netdevsim or a real device. When I develop core code I usually test with netdevsim, then a real device, and then a backport to Meta's kernel. Because of the lack of integration, until now I had to throw away the (YNL-based) test script and netdevsim code. Running tests in tree directly: $ ./tools/testing/selftests/net/nl_netdev.py KTAP version 1 1..2 ok 1 nl_netdev.empty_check ok 2 nl_netdev.lo_check # Totals: pass:2 fail:0 xfail:0 xpass:0 skip:0 error:0 in tree via make: $ make -C tools/testing/selftests/ TARGETS=net \ TEST_PROGS=nl_netdev.py TEST_GEN_PROGS="" run_tests [ ... ] and installed externally, all seem to work: $ make -C tools/testing/selftests/ TARGETS=net \ install INSTALL_PATH=/tmp/ksft-net $ /tmp/ksft-net/run_kselftest.sh -t net:nl_netdev.py [ ... ] For driver tests I followed the lead of net/forwarding and get the device name from env and/or a config file. v3: - fix up netdevsim C - various small nits in other patches (see changelog in patches) v2: https://lore.kernel.org/all/20240403023426.1762996-1-kuba@kernel.org/ - don't add to TARGETS, create a deperate variable with deps - support and use with - support and use passing arguments to tests v1: https://lore.kernel.org/all/20240402010520.1209517-1-kuba@kernel.org/ ==================== Signed-off-by:
-
Jakub Kicinski authored
Add a very simple test to make sure drivers report expected stats. Drivers which implement FEC or pause configuration should report relevant stats. Qstats must be reported, at least packet and byte counts, and they must match total device stats. Tested with netdevsim, bnxt, in-tree and installed. Signed-off-by:
Jakub Kicinski <kuba@kernel.org> Signed-off-by:
-
Jakub Kicinski authored
Add drivers/net as a target for mixed-use tests. The setup is expected to work similarly to the forwarding tests. Since we only need one interface (unlike forwarding tests) read the target device name from NETIF. If not present we'll try to run the test against netdevsim. Reviewed-by:
Petr Machata <petrm@nvidia.com> Signed-off-by:
-
Jakub Kicinski authored
Real devices should implement qstats. Devices which support pause or FEC configuration should also report the relevant stats. nsim was missing FEC stats completely, some of the qstats and pause stats required toggling a debugfs knob. Note that the tests which used pause always initialize the setting so they shouldn't be affected by the different starting value. Signed-off-by:
-
Jakub Kicinski authored
Add a trivial test using YNL. $ ./tools/testing/selftests/net/nl_netdev.py KTAP version 1 1..2 ok 1 nl_netdev.empty_check ok 2 nl_netdev.lo_check Instantiate the family once, it takes longer than the test itself. Reviewed-by:
-
Jakub Kicinski authored
Add glue code for accessing the YNL library which lives under tools/net and YAML spec files from under Documentation/. Automatically figure out if tests are run in tree or not. Since we'll want to use this library both from net and drivers/net test targets make the library a target as well, and automatically include it when net or drivers/net are included. Making net/lib a target ensures that we end up with only one copy of it, and saves us some path guessing. Add a tiny bit of formatting support to be able to output KTAP from the start. Reviewed-by:
-
git://git.open-mesh.org/linux-mergeDavid S. Miller authored
Simon Wunderlich says: ==================== This cleanup patchset includes the following patches: - bump version strings, by Simon Wunderlich - prefer kfree_rcu() over call_rcu() with free-only callbacks, by Dmitry Antipov - bypass empty buckets in batadv_purge_orig_ref(), by Eric Dumazet ==================== Signed-off-by: -
Andy Shevchenko authored
Replace open coded variant of device_is_compatible(). Signed-off-by:
Andy Shevchenko <andriy.shevchenko@linux.intel.com> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Eric Dumazet authored
Commit 6025b913 ("net: dqs: add NIC stall detector based on BQL") added three sysfs files. Use the recommended sysfs_emit() helper. Signed-off-by:
-
Alexander Lobakin authored
Structures which are about to be copied to userspace shouldn't have uninitialized fields or paddings. memset() the whole &ip_tunnel_parm in ip_tunnel_parm_to_user() before filling it with the kernel data. The compilers will hopefully combine writes to it. Fixes: 117aef12 ("ip_tunnel: use a separate struct to store tunnel params in the kernel") Reported-by:
Dan Carpenter <dan.carpenter@linaro.org> Closes: https://lore.kernel.org/netdev/5f63dd25-de94-4ca3-84e6-14095953db13@moroto.mountainSigned-off-by:
Alexander Lobakin <aleksander.lobakin@intel.com> Reviewed-by:
-
Eric Dumazet authored
No longer hold RTNL while calling ip6addrlbl_dump() ("ip addrlabel show") ip6addrlbl_dump() was already mostly relying on RCU anyway. Add READ_ONCE()/WRITE_ONCE() annotations around net->ipv6.ip6addrlbl_table.seq Note that ifal_seq value is currently ignored in iproute2, and a bit weak. We might user later cb->seq / nl_dump_check_consistent() protocol if needed. Also change return value for a completed dump, so that NLMSG_DONE can be appended to current skb, saving one recvmsg() system call. v2: read net->ipv6.ip6addrlbl_table.seq once, (David Ahern) Signed-off-by:David Ahern <dsahern@kernel.org> Signed-off-by:
-
Eric Dumazet authored
fqdir_free_fn() is using very expensive rcu_barrier() When one netns is dismantled, we often call fqdir_exit() multiple times, typically lauching fqdir_free_fn() twice. Delaying by one second fqdir_free_fn() helps to reduce the number of rcu_barrier() calls, and lock contention on rcu_state.barrier_mutex. Signed-off-by:
-
Breno Leitao authored
Commit 3e2f544d ("net: get stats64 if device if driver is configured") moved the callback to dev_get_tstats64() to net core, so, unless the driver is doing some custom stats collection, it does not need to set .ndo_get_stats64. Since this driver is now relying in NETDEV_PCPU_STAT_TSTATS, then, it doesn't need to set the dev_get_tstats64() generic .ndo_get_stats64 function pointer. Signed-off-by:
Breno Leitao <leitao@debian.org> Reviewed-by:
-
Breno Leitao authored
With commit 34d21de9 ("net: Move {l,t,d}stats allocation to core and convert veth & vrf"), stats allocation could be done on net core instead of in this driver. With this new approach, the driver doesn't have to bother with error handling (allocation failure checking, making sure free happens in the right spot, etc). This is core responsibility now. Remove the allocation in the ip6_vti and leverage the network core allocation instead. Signed-off-by:
-
- 06 Apr, 2024 20 commits
-
-
David S. Miller authored
Maxime Chevallier says: ==================== Introduce PHY listing and link_topology tracking This is V11 for the link topology addition, allowing to track all PHYs that are linked to netdevices. This V11 addresses the various netlink-related issues that were raised by Jakub, and fixes some typos in the documentation. As a remainder, here's what the PHY listings would look like : - eth0 has a 88x3310 acting as media converter, and an SFP module with an embedded 88e1111 PHY - eth2 has a 88e1510 PHY PHY for eth0: PHY index: 1 Driver name: mv88x3310 PHY device name: f212a600.mdio-mii:00 Downstream SFP bus name: sfp-eth0 PHY id: 0 Upstream type: MAC PHY for eth0: PHY index: 2 Driver name: Marvell 88E1111 PHY device name: i2c:sfp-eth0:16 PHY id: 21040322 Upstream type: PHY Upstream PHY index: 1 Upstream SFP name: sfp-eth0 PHY for eth2: PHY index: 1 Driver name: Marvell 88E1510 PHY device name: f212a200.mdio-mii:00 PHY id: 21040593 Upstream type: MAC Ethtool patches : https://github.com/minimaxwell/ethtool/tree/link-topo-v6 ==================== Signed-off-by:
-
Maxime Chevallier authored
Some netlink commands are target towards ethernet PHYs, to control some of their features. As there's several such commands, add the ability to pass a PHY index in the ethnl request, which will populate the generic ethnl_req_info with the relevant phydev when the command targets a PHY. Signed-off-by:
Maxime Chevallier <maxime.chevallier@bootlin.com> Reviewed-by:
-
Maxime Chevallier authored
Knowing the bus name is helpful when we want to expose the link topology to userspace, add a helper to return the SFP bus name. Signed-off-by:
-
Maxime Chevallier authored
There are a few PHY drivers that can handle SFP modules through their sfp_upstream_ops. Introduce Phylib helpers to keep track of connected SFP PHYs in a netdevice's namespace, by adding the SFP PHY to the upstream PHY's netdev's namespace. By doing so, these SFP PHYs can be enumerated and exposed to users, which will be able to use their capabilities. Signed-off-by:
-
Maxime Chevallier authored
Pass the phy_device as a parameter to the sfp upstream .disconnect_phy operation. This is preparatory work to help track phy devices across a net_device's link. Signed-off-by:
-
Maxime Chevallier authored
Link topologies containing multiple network PHYs attached to the same net_device can be found when using a PHY as a media converter for use with an SFP connector, on which an SFP transceiver containing a PHY can be used. With the current model, the transceiver's PHY can't be used for operations such as cable testing, timestamping, macsec offload, etc. The reason being that most of the logic for these configuration, coming from either ethtool netlink or ioctls tend to use netdev->phydev, which in multi-phy systems will reference the PHY closest to the MAC. Introduce a numbering scheme allowing to enumerate PHY devices that belong to any netdev, which can in turn allow userspace to take more precise decisions with regard to each PHY's configuration. The numbering is maintained per-netdev, in a phy_device_list. The numbering works similarly to a netdevice's ifindex, with identifiers that are only recycled once INT_MAX has been reached. This prevents races that could occur between PHY listing and SFP transceiver removal/insertion. The identifiers are assigned at phy_attach time, as the numbering depends on the netdevice the phy is attached to. The PHY index can be re-used for PHYs that are persistent. Signed-off-by:
-
Pawel Dembicki authored
The same implementation is also valid for 88E1145. VCT in 88E1111 is similar to the 88E609x family. The main difference lies in register organization and required workarounds. It utilizes the same fields in registers but requires a simpler implementation. Signed-off-by:
Pawel Dembicki <paweldembicki@gmail.com> Reviewed-by:
-
Jakub Kicinski authored
devlink_compat_running_version() sticks out when running netdevsim tests and watching dropped skbs. Add nlmsg_consume() for cases were we want to free a netlink skb but it is expected, rather than a drop. af_netlink code uses consume_skb() directly, which is fine, but some may prefer the symmetry of nlmsg_new() / nlmsg_consume(). Signed-off-by:
-
Jakub Kicinski authored
The ->decrypted bit can be reused for other crypto protocols. Remove the direct dependency on TLS, add helpers to clean up the ifdefs leaking out everywhere. Signed-off-by:
-
Jakub Kicinski authored
Hangbin Liu says: ==================== ynl: rename array-nest to indexed-array rename array-nest to indexed-array and add un-nest sub-type support ==================== Link: https://lore.kernel.org/r/20240404063114.1221532-1-liuhangbin@gmail.comSigned-off-by:
-
Hangbin Liu authored
Add binary and integer sub-type support for indexed-array to display bond arp and ns targets. Here is what the result looks like: # ip link add bond0 type bond mode 1 \ arp_ip_target 192.168.1.1,192.168.1.2 ns_ip6_target 2001::1,2001::2 # ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/rt_link.yaml \ --do getlink --json '{"ifname": "bond0"}' --output-json | jq '.linkinfo' "arp-ip-target": [ "192.168.1.1", "192.168.1.2" ], [...] "ns-ip6-target": [ "2001::1", "2001::2" ], Signed-off-by:Hangbin Liu <liuhangbin@gmail.com> Link: https://lore.kernel.org/r/20240404063114.1221532-3-liuhangbin@gmail.comSigned-off-by:
-
Hangbin Liu authored
Some implementations, like bonding, has nest array with same attr type. To support all kinds of entries under one nest array. As discussed[1], let's rename array-nest to indexed-array, and assuming the value is a nest by passing the type via sub-type. [1] https://lore.kernel.org/netdev/20240312100105.16a59086@kernel.org/Suggested-by:
-
Eric Dumazet authored
tp->window_clamp can be read locklessly, add READ_ONCE() and WRITE_ONCE() annotations. Signed-off-by:
Jason Xing <kerneljasonxing@gmail.com> Link: https://lore.kernel.org/r/20240404114231.2195171-1-edumazet@google.comSigned-off-by:
-
Jakub Kicinski authored
Rahul Rameshbabu says: ==================== ethtool HW timestamping statistics The goal of this patch series is to introduce a common set of ethtool statistics for hardware timestamping that a driver implementer can hook into. The statistics counters added are based on what I believe are common patterns/behaviors found across various hardware timestamping implementations seen in the kernel tree today. The mlx5 family of devices is used as the PoC for this patch series. Other vendors are more than welcome to chime in on this series. Link: https://lore.kernel.org/netdev/20240402205223.137565-1-rrameshbabu@nvidia.com/ Link: https://lore.kernel.org/netdev/20240309084440.299358-1-rrameshbabu@nvidia.com/ Link: https://lore.kernel.org/netdev/20240223192658.45893-1-rrameshbabu@nvidia.com/Signed-off-by:
Rahul Rameshbabu <rrameshbabu@nvidia.com> ==================== Link: https://lore.kernel.org/r/20240403212931.128541-1-rrameshbabu@nvidia.comSigned-off-by:
-
Rahul Rameshbabu authored
Print the nested stats attribute containing timestamping statistics when the --show-time-stamping flag is used. [root@binary-eater-vm-01 linux-ethtool-ts]# ./tools/net/ynl/ethtool.py --show-time-stamping mlx5_1 Time stamping parameters for mlx5_1: Capabilities: hardware-transmit hardware-receive hardware-raw-clock PTP Hardware Clock: 0 Hardware Transmit Timestamp Modes: off on Hardware Receive Filter Modes: none all Statistics: tx-pkts: 8 tx-lost: 0 tx-err: 0 Signed-off-by:Dragos Tatulea <dtatulea@nvidia.com> Link: https://lore.kernel.org/r/20240403212931.128541-8-rrameshbabu@nvidia.comSigned-off-by:
-
Jakub Kicinski authored
Recent changes added header flags to the spec. Use an enum instead of defines for more seamless codegen. [Jakub: drop the already applied parts and rewrite message] Signed-off-by:
-
Rahul Rameshbabu authored
Feed driver statistics counters related to hardware timestamping to standardized ethtool hardware timestamping statistics group. Signed-off-by:
-
Rahul Rameshbabu authored
Count number of transmitted packets that were hardware timestamped at the device DMA layer. Signed-off-by:
-
Rahul Rameshbabu authored
Track the number of times a CQE was expected to not be delivered on PTP Tx port timestamping CQ. A CQE is expected to not be delivered if a certain amount of time passes since the corresponding CQE containing the DMA timestamp information has arrived. Increment the late_cqe counter when such a CQE does manage to be delivered to the CQ. Signed-off-by:
Tariq Toukan <tariqt@nvidia.com> Reviewed-by:
Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20240403212931.128541-3-rrameshbabu@nvidia.comSigned-off-by:
-
Rahul Rameshbabu authored
Multiple network devices that support hardware timestamping appear to have common behavior with regards to timestamp handling. Implement common Tx hardware timestamping statistics in a tx_stats struct_group. Common Rx hardware timestamping statistics can subsequently be implemented in a rx_stats struct_group for ethtool_ts_stats. Signed-off-by:
-
