- 29 Jul, 2021 5 commits
-
-
Paolo Abeni authored
This change leverages the infrastructure introduced by the previous patches to allow soft devices passing to the GRO engine owned skbs without impacting the fast-path. It's up to the GRO caller ensuring the slow_gro bit validity before invoking the GRO engine. The new helper skb_prepare_for_gro() is introduced for that goal. On slow_gro, skbs are aggregated only with equal sk. Additionally, skb truesize on GRO recycle and free is correctly updated so that sk wmem is not changed by the GRO processing. rfc-> v1: - fixed bad truesize on dev_gro_receive NAPI_FREE - use the existing state bit Signed-off-by:
Paolo Abeni <pabeni@redhat.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Paolo Abeni authored
After the previous patches, at GRO time, skb->slow_gro is usually 0, unless the packets comes from some H/W offload slowpath or tunnel. We can optimize the GRO code assuming !skb->slow_gro is likely. This remove multiple conditionals in the most common path, at the price of an additional one when we hit the above "slow-paths". Signed-off-by:
-
Paolo Abeni authored
Similar to the previous one, but tracking the active_extensions field status. Signed-off-by:
-
Paolo Abeni authored
Similar to the previous patch, but covering the dst field: the slow_gro flag is additionally set when a dst is attached to the skb RFC -> v1: - use the existing flag instead of adding a new one Signed-off-by:
-
Paolo Abeni authored
The new flag tracks if any state field is set, so that GRO requires 'unusual'/slow prepare steps. Set such flag when a ct entry is attached to the skb, and never clear it. The new bit uses an existing hole into the sk_buff struct RFC -> v1: - use a single state bit, never clear it - avoid moving the _nfct field Signed-off-by:
-
- 28 Jul, 2021 27 commits
-
-
Hu Haowen authored
Append ioam6-sysctl to toctree in order to get rid of building warnings. Signed-off-by:
Hu Haowen <src.res@email.cn> Signed-off-by:
-
Vladimir Oltean authored
Currently there are issues when adding a bridge FDB entry as VLAN-aware and deleting it as VLAN-unaware, or vice versa. However this is an unneeded complication, since the bridge always installs its default FDB entries in VLAN 0 to match on VLAN-unaware ports, and in the default_pvid (VLAN 1) to match on VLAN-aware ports. So instead of trying to outsmart the bridge, just install all entries it gives us, and they will start matching packets when the vlan_filtering mode changes. Signed-off-by:
Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by:
-
David S. Miller authored
Vladimir Oltean says: ==================== Plug the last 2 holes in the switchdev notifiers for local FDB entries The work for trapping local FDB entries to the CPU in switchdev/DSA started with the "RX filtering in DSA" series: https://patchwork.kernel.org/project/netdevbpf/cover/20210629140658.2510288-1-olteanv@gmail.com/ and was continued with further improvements such as "Fan out FDB entries pointing towards the bridge to all switchdev member ports": https://patchwork.kernel.org/project/netdevbpf/cover/20210719135140.278938-1-vladimir.oltean@nxp.com/ https://patchwork.kernel.org/project/netdevbpf/cover/20210720173557.999534-1-vladimir.oltean@nxp.com/ There are only 2 more issues left to be addressed (famous last words), and these are: - dynamically learned FDB entries towards interfaces foreign to DSA need to be replayed too - adding/deleting a VLAN on a port causes the local FDB entries in that VLAN to be prematurely deleted This patch series addresses both, and patch 2 depends on 1 to work properly. ==================== Signed-off-by:
-
Vladimir Oltean authored
Currently the following script: 1. ip link add br0 type bridge vlan_filtering 1 && ip link set br0 up 2. ip link set swp2 up && ip link set swp2 master br0 3. ip link set swp3 up && ip link set swp3 master br0 4. ip link set swp4 up && ip link set swp4 master br0 5. bridge vlan del dev swp2 vid 1 6. bridge vlan del dev swp3 vid 1 7. ip link set swp4 nomaster 8. ip link set swp3 nomaster produces the following output: [ 641.010738] sja1105 spi0.1: port 2 failed to delete 00:1f:7b:63:02:48 vid 1 from fdb: -2 [ swp2, swp3 and br0 all have the same MAC address, the one listed above ] In short, this happens because the number of FDB entry additions notified to switchdev is unbalanced with the number of deletions. At step 1, the bridge has a random MAC address. At step 2, the br_fdb_replay of swp2 receives this initial MAC address. Then the bridge inherits the MAC address of swp2 via br_fdb_change_mac_address(), and it notifies switchdev (only swp2 at this point) of the deletion of the random MAC address and the addition of 00:1f:7b:63:02:48 as a local FDB entry with fdb->dst == swp2, in VLANs 0 and the default_pvid (1). During step 7: del_nbp -> br_fdb_delete_by_port(br, p, vid=0, do_all=1); -> fdb_delete_local(br, p, f); br_fdb_delete_by_port() deletes all entries towards the ports, regardless of vid, because do_all is 1. fdb_delete_local() has logic to migrate local FDB entries deleted from one port to another port which shares the same MAC address and is in the same VLAN, or to the bridge device itself. This migration happens without notifying switchdev of the deletion on the old port and the addition on the new one, just fdb->dst is changed and the added_by_user flag is cleared. In the example above, the del_nbp(swp4) causes the "addr 00:1f:7b:63:02:48 vid 1" local FDB entry with fdb->dst == swp4 that existed up until then to be migrated directly towards the bridge (fdb->dst == NULL). This is because it cannot be migrated to any of the other ports (swp2 and swp3 are not in VLAN 1). After the migration to br0 takes place, swp4 requests a deletion replay of all FDB entries. Since the "addr 00:1f:7b:63:02:48 vid 1" entry now point towards the bridge, a deletion of it is replayed. There was just a prior addition of this address, so the switchdev driver deletes this entry. Then, the del_nbp(swp3) at step 8 triggers another br_fdb_replay, and switchdev is notified again to delete "addr 00:1f:7b:63:02:48 vid 1". But it can't because it no longer has it, so it returns -ENOENT. There are other possibilities to trigger this issue, but this is by far the simplest to explain. To fix this, we must avoid the situation where the addition of an FDB entry is notified to switchdev as a local entry on a port, and the deletion is notified on the bridge itself. Considering that the 2 types of FDB entries are completely equivalent and we cannot have the same MAC address as a local entry on 2 bridge ports, or on a bridge port and pointing towards the bridge at the same time, it makes sense to hide away from switchdev completely the fact that a local FDB entry is associated with a given bridge port at all. Just say that it points towards the bridge, it should make no difference whatsoever to the switchdev driver and should even lead to a simpler overall implementation, will less cases to handle. This also avoids any modification at all to the core bridge driver, just what is reported to switchdev changes. With the local/permanent entries on bridge ports being already reported to user space, it is hard to believe that the bridge behavior can change in any backwards-incompatible way such as making all local FDB entries point towards the bridge. Signed-off-by:
-
Vladimir Oltean authored
Currently when a switchdev port joins a bridge, we replay all FDB entries pointing towards that port or towards the bridge. However, this is insufficient in certain situations: (a) DSA, through its assisted_learning_on_cpu_port logic, snoops dynamically learned FDB entries on foreign interfaces. These are FDB entries that are pointing neither towards the newly joined switchdev port, nor towards the bridge. So these addresses would be missed when joining a bridge where a foreign interface has already learned some addresses, and they would also linger on if the DSA port leaves the bridge before the foreign interface forgets them. None of this happens if we replay the entire FDB when the port joins. (b) There is a desire to treat local FDB entries on a port (i.e. the port's termination MAC address) identically to FDB entries pointing towards the bridge itself. More details on the reason behind this in the next patch. The point is that this cannot be done given the current structure of br_fdb_replay() in this situation: ip link set swp0 master br0 # br0 inherits its MAC address from swp0 ip link set swp1 master br0 What is desirable is that when swp1 joins the bridge, br_fdb_replay() also notifies swp1 of br0's MAC address, but this won't in fact happen because the MAC address of br0 does not have fdb->dst == NULL (it doesn't point towards the bridge), but it has fdb->dst == swp0. So our current logic makes it impossible for that address to be replayed. But if we dump the entire FDB instead of just the entries with fdb->dst == swp1 and fdb->dst == NULL, then the inherited MAC address of br0 will be replayed too, which is what we need. A natural question arises: say there is an FDB entry to be replayed, like a MAC address dynamically learned on a foreign interface that belongs to a bridge where no switchdev port has joined yet. If 10 switchdev ports belonging to the same driver join this bridge, one by one, won't every port get notified 10 times of the foreign FDB entry, amounting to a total of 100 notifications for this FDB entry in the switchdev driver? Well, yes, but this is where the "void *ctx" argument for br_fdb_replay is useful: every port of the switchdev driver is notified whenever any other port requests an FDB replay, but because the replay was initiated by a different port, its context is different from the initiating port's context, so it ignores those replays. So the foreign FDB entry will be installed only 10 times, once per port. This is done so that the following 4 code paths are always well balanced: (a) addition of foreign FDB entry is replayed when port joins bridge (b) deletion of foreign FDB entry is replayed when port leaves bridge (c) addition of foreign FDB entry is notified to all ports currently in bridge (c) deletion of foreign FDB entry is notified to all ports currently in bridge Signed-off-by: -
David S. Miller authored
Michael Chan says: ==================== bnxt_en: PTP enhancements This series adds two PTP enhancements. This first one is to register the PHC during probe time and keep it registered whether it is in ifup or ifdown state. It will get unregistered and possibly reregistered if the firmware PTP capability changes after firmware reset. The second one is to add the 1PPS (one pulse per second) feature to support input/output of the 1PPS signal. ==================== Signed-off-by: -
Pavan Chebbi authored
FW can report to driver via ASYNC event if it encountered an invalid signal on any TSIO PIN. Driver will log this event for the user to take corrective action. Reviewed-by:
Somnath Kotur <somnath.kotur@broadcom.com> Reviewed-by:
Arvind Susarla <arvind.susarla@broadcom.com> Reviewed-by:
Edwin Peer <edwin.peer@broadcom.com> Signed-off-by:
Pavan Chebbi <pavan.chebbi@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Signed-off-by:
-
Pavan Chebbi authored
Once the PPS pins are configured, the FW can report PPS values using ASYNC event. This patch adds the ASYNC event handler and subsequent reporting of the events to kernel. Signed-off-by:
-
Pavan Chebbi authored
Application will send ioctls to set/clear PPS pin functions based on user input. This patch implements the driver callbacks that will configure the TSIO pins using firmware commands. After firmware reset, the TSIO pins will be reconfigured again. Reviewed-by:
-
Pavan Chebbi authored
1PPS (One Pulse Per Second) is a signal generated either by the NIC PHC or an external timing source. Integrating the support to configure and use 1PPS using the TSIO pins along with PTP timestamps will add Grand Master capability to the 5750X family chipsets. This patch initializes the driver data structures and registers the 1PPS with kernel, based on the TSIO pins' capability in the hardware. This will create a /dev/ppsX device which applications can use to receive PPS events. Later patches will define functions to configure and use the pins. Reviewed-by:
-
Michael Chan authored
During error recovery or hot firmware upgrade, the chip may be under reset and the PHC register read cycles may cause completion timeouts. Check that the chip is not under reset condition before proceeding to read the PHC by checking the flag BNXT_STATE_IN_FW_RESET. We also need to take the ptp_lock before we set this flag to prevent race conditions. We need this logic because the PHC now will stay registered after bnxt_close(). Reviewed-by:
-
Michael Chan authored
It was pointed out by Richard Cochran that registering the PHC during probe is better than during ifup, so move bnxt_ptp_init() back to bnxt_init_one(). In order to work correctly after firmware reset which may result in PTP config. changes, we modify bnxt_ptp_init() to return if the PHC has been registered earlier. If PTP is no longer supported by the new firmware, we will unregister the PHC and clean up. This partially reverts: d7859afb ("bnxt_en: Move bnxt_ptp_init() to bnxt_open()") Reviewed-by:
-
David S. Miller authored
Joakim Zhang says: ==================== net: fec: add support for i.MX8MQ and i.MX8QM This patch set adds supports for i.MX8MQ and i.MX8QM, both of them extend new features. ChangeLogs: V1->V2: * rebase on schema binding, and update dts compatible string. * use generic ethernet controller property for MAC internal RGMII clock delay rx-internal-delay-ps and tx-internal-delay-ps ==================== Signed-off-by: -
Joakim Zhang authored
Add "fsl,imx8qm-fec" compatible string for FEC to support new feature (RGMII delayed clock). Signed-off-by:
Joakim Zhang <qiangqing.zhang@nxp.com> Signed-off-by:
-
Joakim Zhang authored
Add "fsl,imx8mq-fec" compatible string for FEC to support new feature (IEEE 802.3az EEE standard). Signed-off-by:
-
Fugang Duan authored
i.MX8QM ENET IP version support timing specification that MAC integrate clock delay in RGMII mode, the delayed TXC/RXC as an alternative option to work well with various PHYs. Signed-off-by:
Fugang Duan <fugang.duan@nxp.com> Signed-off-by:
-
Fugang Duan authored
The i.MX8MQ ENET version support IEEE802.3az eee mode, add eee mode tx lpi enable to support ethtool interface. usage: 1. set sleep and wake timer to 5ms: ethtool --set-eee eth0 eee on tx-lpi on tx-timer 5000 2. check the eee mode: ~# ethtool --show-eee eth0 EEE Settings for eth0: EEE status: enabled - active Tx LPI: 5000 (us) Supported EEE link modes: 100baseT/Full 1000baseT/Full Advertised EEE link modes: 100baseT/Full 1000baseT/Full Link partner advertised EEE link modes: 100baseT/Full Note: For realtime case and IEEE1588 ptp case, it should disable EEE mode. Signed-off-by: -
Fugang Duan authored
The ENET of imx8mq and imx8qm are basically the same as imx6sx, but they have new features support based on imx6sx, like: - imx8mq: supports IEEE 802.3az EEE standard. - imx8qm: supports RGMII mode delayed clock. Signed-off-by:
-
Joakim Zhang authored
Add RGMII internal clock delay for FEC controller. Signed-off-by:
-
Joakim Zhang authored
Add more compatible items for i.MX8/8M platforms. Signed-off-by:
-
Peilin Ye authored
Recently we added a new option, SKBMOD_F_ECN, to tc-skbmod(8). Add a control-plane selftest for it. Depends on kernel patch "net/sched: act_skbmod: Add SKBMOD_F_ECN option support", as well as iproute2 patch "tc/skbmod: Introduce SKBMOD_F_ECN option". Reviewed-by:
Cong Wang <cong.wang@bytedance.com> Signed-off-by:
Peilin Ye <peilin.ye@bytedance.com> Signed-off-by:
-
Peilin Ye authored
Currently, when doing rate limiting using the tc-police(8) action, the easiest way is to simply drop the packets which exceed or conform the configured bandwidth limit. Add a new option to tc-skbmod(8), so that users may use the ECN [1] extension to explicitly inform the receiver about the congestion instead of dropping packets "on the floor". The 2 least significant bits of the Traffic Class field in IPv4 and IPv6 headers are used to represent different ECN states [2]: 0b00: "Non ECN-Capable Transport", Non-ECT 0b10: "ECN Capable Transport", ECT(0) 0b01: "ECN Capable Transport", ECT(1) 0b11: "Congestion Encountered", CE As an example: $ tc filter add dev eth0 parent 1: protocol ip prio 10 \ matchall action skbmod ecn Doing the above marks all ECT(0) and ECT(1) packets as CE. It does NOT affect Non-ECT or non-IP packets. In the tc-police scenario mentioned above, users may pipe a tc-police action and a tc-skbmod "ecn" action together to achieve ECN-based rate limiting. For TCP connections, upon receiving a CE packet, the receiver will respond with an ECE packet, asking the sender to reduce their congestion window. However ECN also works with other L4 protocols e.g. DCCP and SCTP [2], and our implementation does not touch or care about L4 headers. The updated tc-skbmod SYNOPSIS looks like the following: tc ... action skbmod { set SETTABLE | swap SWAPPABLE | ecn } ... Only one of "set", "swap" or "ecn" shall be used in a single tc-skbmod command. Trying to use more than one of them at a time is considered undefined behavior; pipe multiple tc-skbmod commands together instead. "set" and "swap" only affect Ethernet packets, while "ecn" only affects IPv{4,6} packets. It is also worth mentioning that, in theory, the same effect could be achieved by piping a "police" action and a "bpf" action using the bpf_skb_ecn_set_ce() helper, but this requires eBPF programming from the user, thus impractical. Depends on patch "net/sched: act_skbmod: Skip non-Ethernet packets". [1] https://datatracker.ietf.org/doc/html/rfc3168 [2] https://en.wikipedia.org/wiki/Explicit_Congestion_NotificationReviewed-by: -
Yang Yingliang authored
If nfp_tunnel_add_ipv6_off() fails, it should return error code in nfp_fl_ct_add_offload(). Fixes: 5a2b9304 ("nfp: flower-ct: compile match sections of flow_payload") Reported-by:
Hulk Robot <hulkci@huawei.com> Signed-off-by:
Yang Yingliang <yangyingliang@huawei.com> Signed-off-by:
Louis Peens <louis.peens@corigine.com> Reviewed-by:
Simon Horman <simon.horman@corigine.com> Signed-off-by:
-
David S. Miller authored
Leon Romanovsky says: ==================== Remove duplicated devlink registration check Changelog: v1: * Added two new patches that remove registration field from mlx5 and ti drivers. v0: https://lore.kernel.org/lkml/ed7bbb1e4c51dd58e6035a058e93d16f883b09ce.1627215829.git.leonro@nvidia.com -------------------------------------------------------------------- Both registered flag and devlink pointer are set at the same time and indicate the same thing - devlink/devlink_port are ready. Instead of checking ->registered use devlink pointer as an indication. ==================== Signed-off-by:
-
Leon Romanovsky authored
Both registered flag and devlink pointer are set at the same time and indicate the same thing - devlink/devlink_port are ready. Instead of checking ->registered use devlink pointer as an indication. Signed-off-by:
Leon Romanovsky <leonro@nvidia.com> Signed-off-by:
-
Leon Romanovsky authored
Devlink is an integral part of mlx5 driver and all flows ensure that devlink_*_register() will success. That makes the ->registered check an obsolete. Signed-off-by:
-
Leon Romanovsky authored
The commit that introduced devlink support released devlink resources in wrong order, that made an unwind flow to be asymmetrical. In addition, the am65-cpsw-nuss used internal to devlink core field - registered. In order to fix the unwind flow and remove such access to the registered field, rewrite the code to call devlink_port_unregister only on registered ports. Fixes: 58356eb3 ("net: ti: am65-cpsw-nuss: Add devlink support") Signed-off-by:
-
- 27 Jul, 2021 8 commits
-
-
David S. Miller authored
Alex Elder says: ==================== net: ipa: add clock references This series continues preparation for implementing runtime power management for IPA. We need to ensure that the IPA core clock and interconnects are operational whenever IPA hardware is accessed. And in particular this means that any external entry point that can lead to accessing IPA hardware must guarantee the hardware is "up" when it is accessed. The first four patches in this series take IPA clock references when needed by such external entry points, dropping those references in those same functions when they are no longer required. The last patch is a bit different, though it too prepares for enabling runtime power management. It avoids suspending/resuming endpoints if setup is not complete. ==================== Signed-off-by: -
Alex Elder authored
Until we complete the setup stage of initialization, GSI is not initialized and therefore endpoints aren't usable. So avoid suspending endpoints during system suspend unless setup is complete. Clear the setup_complete flag at the top of ipa_teardown() to reflect the fact that things are no longer in setup state. Get rid of a misplaced (and superfluous) comment. Signed-off-by:
Alex Elder <elder@linaro.org> Signed-off-by:
-
Alex Elder authored
The IPA network device can be opened at any time, and an opened network device can be stopped any time. Both of these callback functions require access to the hardware, and therefore they need the IPA clock to be operational. Take an IPA clock reference in both the ->open and ->stop callback functions, dropping the reference when they are done accessing hardware. The ->start_xmit callback requires a little different handling, and that will be added separately. Signed-off-by:
-
Alex Elder authored
The remoteproc SSR callback function for the modem requires hardware access when handling a modem crash or shutdown. Take and later release an IPA clock reference in ipa_modem_crashed(), to ensure the hardware is operational. Signed-off-by:
-
Alex Elder authored
Two places call ipa_setup(). The first, ipa_probe(), holds an IPA clock reference when calling ipa_setup() (if the AP is responsible for IPA firmware loading). But if the modem is loading IPA firmware, ipa_smp2p_modem_setup_ready_isr() calls ipa_setup() after the modem has signaled the hardware is ready. This can happen at any time, and there is no guarantee the hardware is active. Have ipa_smp2p_modem_setup() take an IPA clock reference before it calls ipa_setup(), and release it once setup is complete. Signed-off-by:
-
Alex Elder authored
Any entry point that leads to IPA hardware access must ensure the hardware is operational (clocked). Currently we ensure this by taking an extra clock reference during setup that is not released until we receive a system suspend request. But this extra reference will soon go away. When the platform driver ->probe function is called, we first need hardware access in ipa_config(). Although ipa_config() takes an IPA clock reference, it the special reference taken to prevent suspending the hardware. Have ipa_probe() take a reference before calling ipa_config(), so that the "no-suspend" reference can eventually go away. Drop this reference before ipa_probe() returns. Similarly, the driver ->remove function can be called at any time. Take an IPA clock reference at the beginning of that function, and drop it again after the deconfig stage has completed (at which point hardware access is no longer needed). Signed-off-by:
-
David S. Miller authored
Alex Elder says: ==================== net: ipa: IPA interrupt cleanup The first patch in this series makes all IPA interrupt handling be done in a threaded context. The remaining ones refactor some code to simplify that threaded handler function. ==================== Signed-off-by: -
Alex Elder authored
Now that ipa_isr_thread() is a simple wrapper that gets a clock reference around ipa_interrupt_process_all(), get rid of the called function and just open-code it in ipa_isr_thread(). Signed-off-by:
-
