summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2020-05-08iwlwifi: pcie: allocate much smaller byte-count tableJohannes Berg2-12/+19
The hardware needs a byte-count table with the size of each frame on the queue to build A-MPDUs, but: * newer generation no longer have the duplicated space at the end, they can deal with the wrap properly - and we don't even fill the dup anyway * we have a maximum queue size of 512 right now and don't use the theoretical hardware maximum of 65536. Together, this reduces the byte count table DMA allocation from 64KiB (65536*2 + 64*2 rounded up) to 1 KiB (though that might be rounded up to a full 4 KiB page by the allocator, not sure it can share the allocations.) Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200425130140.c263b787b5ab.I059507a9760b1ce1d45d84dcaa91629a5cfb58e0@changeid
2020-05-08iwlwifi: dump api version in yaml formatMordechay Goodstein1-4/+100
Used for debugging what FW API we are using to understand misalignment with API changes. The output looks like this as a yaml format fw_api_ver: 0x0001: name: MVM_ALIVE cmd_ver: 99 notif_ver: 4 0x0108: name: PHY_CONTEXT_CMD cmd_ver: 2 notif_ver: 0 ... Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.18bf540ab8e0.I6217488f1740f0e6accd0cecd09dfd46bad88426@changeid
2020-05-08iwlwifi: pcie: convert all AX101 devices to the device tablesLuca Coelho3-36/+37
Convert all Qu/Hr1 devices to the new device tables, by modifying the corresponding structures, adding a new name and generalizing the device recognition. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.ec0e04102d2c.Ia36f2c7bbf06cb6436424d40d6adb2376f2962ee@changeid
2020-05-08iwlwifi: pcie: remove occurrences of 22000 in the FW name definesLuca Coelho1-40/+32
We don't use the number 22000 for our devices anymore, so remove all occurrences of it in the FW name macros. While at it, add IWL_QU_B_HR_B to the list of firmwares used by the driver, which was missing. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.93cc41bdbb4d.Ib7599901888a2d050f851bd878a875f593f2e8e5@changeid
2020-05-08iwlwifi: pcie: convert QnJ with Hr to the device tableLuca Coelho3-28/+14
Convert all the occurrences of QnJ with Hr into a single entry in the PCI device table. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.21283ca79d82.I57a493c679a56fff11417afdaec76617711ff6a9@changeid
2020-05-08iwlwifi: pcie: remove mangling for iwl_ax101_cfg_qu_hrLuca Coelho1-23/+0
All devices that use iwl_ax101_cfg_qu_hr are recognized via the device info table, so the cfg will never be iwl_ax101_cfg_qu_hr. Remove the code that converts this into QuZ and Qu-C, since it's not needed anymore. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.30b6b4aae1c1.If5e583835c9f7f2771a50ba1b2f33bb85f25b383@changeid
2020-05-08iwlwifi: pcie: remove outdated comment about PCI RTPM referenceLuca Coelho1-10/+2
This comment was only related to the PCI RTPM implementation, which has been removed a while ago, and is not meaningless. Remove it. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.f362e4560a3b.I8d858c6c9a6c98b45d2195dfe28dabe0286c8a83@changeid
2020-05-08iwlwifi: mvm: add IML/ROM information to the assertion dumpsLuca Coelho1-0/+20
Dump the IML/ROM error code and data, which are read from some registers, when printing an assertion dump. This makes it easier to debug IML/ROM errors. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.a522161a7372.I2a65ee35a5e0242f8a0e106f126356dff81ef59d@changeid
2020-05-08iwlwifi: move iwl_set_soc_latency to iwl-drv to be used by other op_modesEmmanuel Grumbach3-47/+55
All the op_mode need to send this command as well. Instead of duplicating the code from mvm, put the code in a common place. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.7f30f977f9bf.I060b51d0d66d09b9d1ee512e7de8f2d695a52152@changeid
2020-05-08iwlwifi: yoyo: remove magic numberMordechay Goodstein1-1/+5
The for loop is iterating over active regions so iterate only over the len of the active regions buffer size. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.e10482b9eed7.I15da7bb25d9b9e3eef1c1b117dc585e703ce756a@changeid
2020-05-08iwlwifi: mvm: initialize iwl_dev_tx_power_cmd to zeroLuca Coelho1-4/+3
If the REDUCE_TX_POWER_CMD version is v4 or v5, we are not initializing some values before sending to the FW, which causes SAR not to work properly. Solve this by initializing the struct in the declaration. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.0dc957a264ff.I43cfd72d539c1287ccaaa454e95c673dac38214f@changeid
2020-05-08iwlwifi: dbg: set debug descriptor to NULL outside of iwl_fw_free_dump_descShahar S Matityahu1-9/+10
To avoid static analysis warning and to make the flow more readable, set the debug descriptor to NULL outside iwl_fw_free_dump_desc and only in the required places. Signed-off-by: Shahar S Matityahu <shahar.s.matityahu@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424194456.5d5c50750a52.I17e33fc268c2097b7c42877f86cef2aa163b913a@changeid
2020-05-08iwlwifi: bump FW API to 55 for AX devicesLuca Coelho1-3/+3
Start supporting API version 55 for AX devices. We skipped 54 because it won't be published. Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.5b4628e18203.I626d3b19d84214a83cd92dc4ddf291903b0983aa@changeid
2020-05-08iwlwifi: yoyo: add D3 resume timepointMordechay Goodstein1-0/+3
This timepoint is used for getting a clean log (if needed) after resume without any commands/interrupts from the driver to the FW. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.7cd450812977.Ibc9ddd6eae6af5ce499ac1e4f6c01853577d1e83@changeid
2020-05-08iwlwifi: mvm: set properly station flags in STA_HE_CTXT_CMDShaul Triebitz1-5/+9
For ACK_ENABLED and 32BIT_BA_BITMAP flags check the station capabilities rather than bss_conf.ack_enabled and bss_conf.multi_sta_back_32bit. These fields are stations capabilities and should not be in bss_conf. Also note that the bss_conf flags are set in station mode only. In the next patch I will remove ack_enabled and multi_sta_back_32bit from the bss_conf structure. Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.bc7230b74f93.I144f73cd6a797a7060429981fee62572861bc76b@changeid
2020-05-08iwlwifi: avoid debug max amsdu config overwriting itselfMordechay Goodstein2-11/+15
If we set amsdu_len one after another the second one overwrites the orig_amsdu_len so allow only moving from debug to non debug state. Also the TLC update check was wrong: it was checking that also the orig is smaller then the new updated size, which is not the case in debug amsdu mode. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Fixes: af2984e9e625 ("iwlwifi: mvm: add a debugfs entry to set a fixed size AMSDU for all TX packets") Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.e565446a4fce.I9729d8c520d8b8bb4de9a5cdc62e01eb85168aac@changeid
2020-05-08iwlwifi: mvm: remove iwlmvm's tfd_q_hang_detect module parameterEmmanuel Grumbach4-31/+13
This should be controlled by the firmware debugging mechanism and not by a module parameter. This has always been true. Remove it and assume it is set. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.b6e4982e62ae.I7f7352f79c40ada2f221bd4b41449a40821e833f@changeid
2020-05-08iwlwifi: mvm: stop supporting swcrypto and bt_coex_active module parametersEmmanuel Grumbach3-47/+33
Leave them active for iwldvm. We do not test this configuration and there is no reason nowadays to allow this. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.674a325b008b.Ifc925ca84500fb76c7b6f926a24a34ca777b4192@changeid
2020-05-08iwlwifi: mvm: add support for range request command version 9Avraham Stern2-22/+155
This version adds support for per responder calibrations. Currently the driver will use a single value for all responders and bandwidths. Signed-off-by: Avraham Stern <avraham.stern@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.5ce74a87009c.I9079332b21eef490bbdbf8d7d66e35d7d0c7882b@changeid
2020-05-08iwlwifi: yoyo: don't access TLV before verifying lenMordechay Goodstein2-25/+24
If we access the TLV memory with shorter len than the struct we access garbage data that was not given by the user. On the way rewrite the checker in a cleaner way. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Fixes: a9248de42464 ("iwlwifi: dbg_ini: add TLV allocation new API support") Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.54418c829390.I15d6b462a0e69a280b6c6cfbcb6bcb05bb5f79ee@changeid
2020-05-08iwlwifi: remove antenna_coupling module parameterEmmanuel Grumbach5-35/+7
This module parameter should not be mangled by users. This relates to a very old driver and I doubt people can really check the antenna coupling in a way that would make the BT Coexistence work better with a real value. Drop it. Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.6e566897ce0a.I8395a50c1c39522e542366064bff33a33009ce7b@changeid
2020-05-08iwlwifi: yoyo: use hweight_long instead of bit manipulatingMordechay Goodstein1-12/+13
Also we can only have one fid1 or fid2 set so no need to check if the fid2 is set in case fid1 wasn't set. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182644.34e74106bad3.Ic3a9f0a35fed47b02ebcd27c2dc2b50cb1e56bdf@changeid
2020-05-08iwlwifi: yoyo: add support for parsing SHARED_MEM_ALLOC version 4Mordechay Goodstein5-21/+52
The new version adds the information for RX2C FIFO addresses. Use the new addresses to parse the FIFO info when dumping. Signed-off-by: Mordechay Goodstein <mordechay.goodstein@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182643.97cc25d96b53.I65fd0400d80f505bd6d7eed442f12db24b25bbe3@changeid
2020-05-08iwlwifi: fw api: fix PHY data 2/3 positionJohannes Berg1-8/+6
In AX210 devices, the PHY data wasn't actually reported, but now that it's going to be reported it turns out that the position is supposed to be the other way around, fix that. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Luca Coelho <luciano.coelho@intel.com> Link: https://lore.kernel.org/r/iwlwifi.20200424182643.06de959301f5.I544c353a8a811f107bd66d168e37920237ecf071@changeid
2020-05-08net: relax SO_TXTIME CAP_NET_ADMIN checkEric Dumazet1-10/+18
Now sch_fq has horizon feature, we want to allow QUIC/UDP applications to use EDT model so that pacing can be offloaded to the kernel (sch_fq) or the NIC. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Willem de Bruijn <willemb@google.com> Acked-by: Willem de Bruijn <willemb@google.com> Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08Merge branch 'bonding-report-transmit-status-to-callers'David S. Miller10-82/+65
Eric Dumazet says: ==================== bonding: report transmit status to callers First patches cleanup netpoll, and make sure it provides tx status to its users. Last patch changes bonding to not pretend packets were sent without error. By providing more accurate status, TCP stack can avoid adding more packets if the slave qdisc is already full. This came while testing latest horizon feature in sch_fq, with very low pacing rate flows, but should benefit hosts under stress. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08bonding: propagate transmit statusEric Dumazet3-48/+32
Currently, bonding always returns NETDEV_TX_OK to its caller. It is worth trying to be more accurate : TCP for instance can have different recovery strategies if it can have more precise status, if packet was dropped by slave qdisc. This is especially important when host is under stress. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Jay Vosburgh <j.vosburgh@gmail.com> Cc: Veaceslav Falico <vfalico@gmail.com> Cc: Andy Gospodarek <andy@greyhouse.net> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08netpoll: accept NULL np argument in netpoll_send_skb()Eric Dumazet7-24/+17
netpoll_send_skb() callers seem to leak skb if the np pointer is NULL. While this should not happen, we can make the code more robust. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08netpoll: netpoll_send_skb() returns transmit statusEric Dumazet2-5/+8
Some callers want to know if the packet has been sent or dropped, to inform upper stacks. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08netpoll: move netpoll_send_skb() out of lineEric Dumazet2-10/+12
There is no need to inline this helper, as we intend to add more code in this function. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08netpoll: remove dev argument from netpoll_send_skb_on_dev()Eric Dumazet2-7/+8
netpoll_send_skb_on_dev() can get the device pointer directly from np->dev Rename it to __netpoll_send_skb() Following patch will move netpoll_send_skb() out-of-line. Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: phy: fix less than zero comparison with unsigned variable valColin Ian King1-1/+1
The unsigned variable val is being checked for an error by checking if it is less than zero. This can never occur because val is unsigned. Fix this by making val a plain int. Addresses-Coverity: ("Unsigned compared against zero") Fixes: bdbdac7649fa ("ethtool: provide UAPI for PHY master/slave configuration.") Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net/smc: remove set but not used variables 'del_llc, del_llc_resp'YueHaibing1-7/+1
Fixes gcc '-Wunused-but-set-variable' warning: net/smc/smc_llc.c: In function 'smc_llc_cli_conf_link': net/smc/smc_llc.c:753:31: warning: variable 'del_llc' set but not used [-Wunused-but-set-variable] struct smc_llc_msg_del_link *del_llc; ^ net/smc/smc_llc.c: In function 'smc_llc_process_srv_delete_link': net/smc/smc_llc.c:1311:33: warning: variable 'del_llc_resp' set but not used [-Wunused-but-set-variable] struct smc_llc_msg_del_link *del_llc_resp; ^ Signed-off-by: YueHaibing <yuehaibing@huawei.com> Signed-off-by: Karsten Graul <kgraul@linux.ibm.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08tcp: tcp_mark_head_lost is only valid for sack-tcpzhang kai1-25/+7
so tcp_is_sack/reno checks are removed from tcp_mark_head_lost. Signed-off-by: zhang kai <zhangkaiheb@126.com> Acked-by: Neal Cardwell <ncardwell@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: remove newlines in NL_SET_ERR_MSG_MODJacob Keller6-9/+84
The NL_SET_ERR_MSG_MOD macro is used to report a string describing an error message to userspace via the netlink extended ACK structure. It should not have a trailing newline. Add a cocci script which catches cases where the newline marker is present. Using this script, fix the handful of cases which accidentally included a trailing new line. I couldn't figure out a way to get a patch mode working, so this script only implements context, report, and org. Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Andy Whitcroft <apw@canonical.com> Cc: Joe Perches <joe@perches.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08Merge branch 'ti-am65x-cpts-follow-up-dt-bindings-update'David S. Miller6-25/+49
Grygorii Strashko says: ==================== net: ethernet: ti: am65x-cpts: follow up dt bindings update This series is follow update for TI A65x/J721E Common platform time sync (CPTS) driver [1] to implement DT bindings review comments from Rob Herring <robh@kernel.org> [2]. - "reg" and "compatible" properties are made required for CPTS DT nodes which also required to change K3 CPSW driver to use of_platform_device_create() instead of of_platform_populate() for proper CPTS and MDIO initialization - minor DT bindings format changes - K3 CPTS example added to K3 MCU CPSW bindings [1] https://lwn.net/Articles/819313/ [2] https://lwn.net/ml/linux-kernel/20200505040419.GA8509@bogus/ ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08arm64: dts: ti: k3-am65/j721e-mcu: update cpts nodeGrygorii Strashko2-2/+6
Update CPTS node following DT binding update: - add reg and compatible properties - fix node name Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08dt-binding: net: ti: am65x-cpts: make reg and compatible requiredGrygorii Strashko2-17/+23
This patch follows K3 CPTS review comments from Rob Herring <robh@kernel.org>. - "reg" and "compatible" properties are required now - minor format changes - K3 CPTS example added to K3 MCU CPSW bindings Cc: Rob Herring <robh@kernel.org> Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: ethernet: ti: am65-cpsw-nuss: use of_platform_device_create() for mdioGrygorii Strashko2-6/+20
The MCU CPSW expected to populate only MDIO device, but follow up patches will add "compatible" property to the MCU CPSW CPTS node which will cause creation of CPTS device and MCU CPSW init failure. Hence, switch to use of_platform_device_create() instead of of_platform_populate() for MDIO device population. Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08Merge branch 'hsr-hsr-code-refactoring'David S. Miller3-23/+47
Taehee Yoo says: ==================== hsr: hsr code refactoring There are some unnecessary routine in the hsr module. This patch removes these routines. The first patch removes incorrect comment. The second patch removes unnecessary WARN_ONCE() macro. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08dpaa2-eth: create a function to flush the XDP fdsIoana Ciornei2-22/+46
Create an independent function that takes a particular frame queue and an array of frame descriptors and tries to enqueue them until it hits the maximum number fo retries. The same function will be used in the next patch also on the XDP_TX path. Also, create the dpaa2_eth_xdp_fds structure to incorporate the array of FDs as well as the number of FDs already populated. Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08hsr: remove WARN_ONCE() in hsr_fill_frame_info()Taehee Yoo1-1/+1
When VLAN frame is being sent, hsr calls WARN_ONCE() because hsr doesn't support VLAN. But using WARN_ONCE() is overdoing. Using netdev_warn_once() is enough. Signed-off-by: Taehee Yoo <ap420073@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08soc: fsl: dpio: properly compute the consumer indexIoana Ciornei1-0/+1
Mask the consumer index before using it. Without this, we would be writing frame descriptors beyond the ring size supported by the QBMAN block. Fixes: 3b2abda7d28c ("soc: fsl: dpio: Replace QMAN array mode with ring mode enqueue") Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com> Acked-by: Li Yang <leoyang.li@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08Merge branch 'tc-gate-offload-for-SJA1105-DSA-switch'David S. Miller17-44/+1746
Vladimir Oltean says: ==================== tc-gate offload for SJA1105 DSA switch Expose the TTEthernet hardware features of the switch using standard tc-flower actions: trap, drop, redirect and gate. v1 was submitted at: https://patchwork.ozlabs.org/project/netdev/cover/20200503211035.19363-1-olteanv@gmail.com/ v2 was submitted at: https://patchwork.ozlabs.org/project/netdev/cover/20200503211035.19363-1-olteanv@gmail.com/ Changes in v3: Made sure there are no compilation warnings when CONFIG_NET_DSA_SJA1105_TAS or CONFIG_NET_DSA_SJA1105_VL are disabled. Changes in v2: Using a newly introduced dsa_port_from_netdev public helper. ==================== Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08docs: net: dsa: sja1105: document intended usage of virtual linksVladimir Oltean1-0/+116
Add some verbiage describing how the hardware features of the switch are exposed to users through tc-flower. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: dsa: sja1105: implement tc-gate using time-triggered virtual linksVladimir Oltean10-17/+759
Restrict the TTEthernet hardware support on this switch to operate as closely as possible to IEEE 802.1Qci as possible. This means that it can perform PTP-time-based ingress admission control on streams identified by {DMAC, VID, PCP}, which is useful when trying to ensure the determinism of traffic scheduled via IEEE 802.1Qbv. The oddity comes from the fact that in hardware (and in TTEthernet at large), virtual links always need a full-blown action, including not only the type of policing, but also the list of destination ports. So in practice, a single tc-gate action will result in all packets getting dropped. Additional actions (either "trap" or "redirect") need to be specified in the same filter rule such that the conforming packets are actually forwarded somewhere. Apart from the VL Lookup, Policing and Forwarding tables which need to be programmed for each flow (virtual link), the Schedule engine also needs to be told to open/close the admission gates for each individual virtual link. A fairly accurate (and detailed) description of how that works is already present in sja1105_tas.c, since it is already used to trigger the egress gates for the tc-taprio offload (IEEE 802.1Qbv). Key point here, we remember that the schedule engine supports 8 "subschedules" (execution threads that iterate through the global schedule in parallel, and that no 2 hardware threads must execute a schedule entry at the same time). For tc-taprio, each egress port used one of these 8 subschedules, leaving a total of 4 subschedules unused. In principle we could have allocated 1 subschedule for the tc-gate offload of each ingress port, but actually the schedules of all virtual links installed on each ingress port would have needed to be merged together, before they could have been programmed to hardware. So simplify our life and just merge the entire tc-gate configuration, for all virtual links on all ingress ports, into a single subschedule. Be sure to check that against the usual hardware scheduling conflicts, and program it to hardware alongside any tc-taprio subschedule that may be present. The following scenarios were tested: 1. Quantitative testing: tc qdisc add dev swp2 clsact tc filter add dev swp2 ingress flower skip_sw \ dst_mac 42:be:24:9b:76:20 \ action gate index 1 base-time 0 \ sched-entry OPEN 1200 -1 -1 \ sched-entry CLOSE 1200 -1 -1 \ action trap ping 192.168.1.2 -f PING 192.168.1.2 (192.168.1.2) 56(84) bytes of data. ............................. --- 192.168.1.2 ping statistics --- 948 packets transmitted, 467 received, 50.7384% packet loss, time 9671ms 2. Qualitative testing (with a phase-aligned schedule - the clocks are synchronized by ptp4l, not shown here): Receiver (sja1105): tc qdisc add dev swp2 clsact now=$(phc_ctl /dev/ptp1 get | awk '/clock time is/ {print $5}') && \ sec=$(echo $now | awk -F. '{print $1}') && \ base_time="$(((sec + 2) * 1000000000))" && \ echo "base time ${base_time}" tc filter add dev swp2 ingress flower skip_sw \ dst_mac 42:be:24:9b:76:20 \ action gate base-time ${base_time} \ sched-entry OPEN 60000 -1 -1 \ sched-entry CLOSE 40000 -1 -1 \ action trap Sender (enetc): now=$(phc_ctl /dev/ptp0 get | awk '/clock time is/ {print $5}') && \ sec=$(echo $now | awk -F. '{print $1}') && \ base_time="$(((sec + 2) * 1000000000))" && \ echo "base time ${base_time}" tc qdisc add dev eno0 parent root taprio \ num_tc 8 \ map 0 1 2 3 4 5 6 7 \ queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \ base-time ${base_time} \ sched-entry S 01 50000 \ sched-entry S 00 50000 \ flags 2 ping -A 192.168.1.1 PING 192.168.1.1 (192.168.1.1): 56 data bytes ... ^C --- 192.168.1.1 ping statistics --- 1425 packets transmitted, 1424 packets received, 0% packet loss round-trip min/avg/max = 0.322/0.361/0.990 ms And just for comparison, with the tc-taprio schedule deleted: ping -A 192.168.1.1 PING 192.168.1.1 (192.168.1.1): 56 data bytes ... ^C --- 192.168.1.1 ping statistics --- 33 packets transmitted, 19 packets received, 42% packet loss round-trip min/avg/max = 0.336/0.464/0.597 ms Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: dsa: sja1105: support flow-based redirection via virtual linksVladimir Oltean7-6/+437
Implement tc-flower offloads for redirect, trap and drop using non-critical virtual links. Commands which were tested to work are: # Send frames received on swp2 with a DA of 42:be:24:9b:76:20 to the # CPU and to swp3. This type of key (DA only) when the port's VLAN # awareness state is off. tc qdisc add dev swp2 clsact tc filter add dev swp2 ingress flower skip_sw dst_mac 42:be:24:9b:76:20 \ action mirred egress redirect dev swp3 \ action trap # Drop frames received on swp2 with a DA of 42:be:24:9b:76:20, a VID # of 100 and a PCP of 0. tc filter add dev swp2 ingress protocol 802.1Q flower skip_sw \ dst_mac 42:be:24:9b:76:20 vlan_id 100 vlan_prio 0 action drop Under the hood, all rules match on DMAC, VID and PCP, but when VLAN filtering is disabled, those are set internally by the driver to the port-based defaults. Because we would be put in an awkward situation if the user were to change the VLAN filtering state while there are active rules (packets would no longer match on the specified keys), we simply deny changing vlan_filtering unless the list of flows offloaded via virtual links is empty. Then the user can re-add new rules. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: dsa: sja1105: make room for virtual link parsing in flower offloadVladimir Oltean2-27/+112
Virtual links are a sja1105 hardware concept of executing various flow actions based on a key extracted from the frame's DMAC, VID and PCP. Currently the tc-flower offload code supports only parsing the DMAC if that is the broadcast MAC address, and the VLAN PCP. Extract the key parsing logic from the L2 policers functionality and move it into its own function, after adding extra logic for matching on any DMAC and VID. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: dsa: sja1105: add static tables for virtual linksVladimir Oltean4-0/+318
This patch adds the register definitions for the: - VL Lookup Table - VL Policing Table - VL Forwarding Table - VL Forwarding Parameters Table These are needed in order to perform TTEthernet operations: QoS classification, flow-based policing and/or frame redirecting with the switch. Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-05-08net: dsa: introduce a dsa_port_from_netdev public helperVladimir Oltean2-0/+10
As its implementation shows, this is synonimous with calling dsa_slave_dev_check followed by dsa_slave_to_port, so it is quite simple already and provides functionality which is already there. However there is now a need for these functions outside dsa_priv.h, for example in drivers that perform mirroring and redirection through tc-flower offloads (they are given raw access to the flow_cls_offload structure), where they need to call this function on act->dev. But simply exporting dsa_slave_to_port would make it non-inline and would result in an extra function call in the hotpath, as can be seen for example in sja1105: Before: 000006dc <sja1105_xmit>: { 6dc: e92d4ff0 push {r4, r5, r6, r7, r8, r9, sl, fp, lr} 6e0: e1a04000 mov r4, r0 6e4: e591958c ldr r9, [r1, #1420] ; 0x58c <- Inline dsa_slave_to_port 6e8: e1a05001 mov r5, r1 6ec: e24dd004 sub sp, sp, #4 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 6f0: e1c901d8 ldrd r0, [r9, #24] 6f4: ebfffffe bl 0 <dsa_8021q_tx_vid> 6f4: R_ARM_CALL dsa_8021q_tx_vid u8 pcp = netdev_txq_to_tc(netdev, queue_mapping); 6f8: e1d416b0 ldrh r1, [r4, #96] ; 0x60 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 6fc: e1a08000 mov r8, r0 After: 000006e4 <sja1105_xmit>: { 6e4: e92d4ff0 push {r4, r5, r6, r7, r8, r9, sl, fp, lr} 6e8: e1a04000 mov r4, r0 6ec: e24dd004 sub sp, sp, #4 struct dsa_port *dp = dsa_slave_to_port(netdev); 6f0: e1a00001 mov r0, r1 { 6f4: e1a05001 mov r5, r1 struct dsa_port *dp = dsa_slave_to_port(netdev); 6f8: ebfffffe bl 0 <dsa_slave_to_port> 6f8: R_ARM_CALL dsa_slave_to_port 6fc: e1a09000 mov r9, r0 u16 tx_vid = dsa_8021q_tx_vid(dp->ds, dp->index); 700: e1c001d8 ldrd r0, [r0, #24] 704: ebfffffe bl 0 <dsa_8021q_tx_vid> 704: R_ARM_CALL dsa_8021q_tx_vid Because we want to avoid possible performance regressions, introduce this new function which is designed to be public. Suggested-by: Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
generated by cgit v1.2.3 (git 2.39.0) at 2025-09-24 17:38:36 +0300