1.. SPDX-License-Identifier: GPL-2.0 2 3==================== 4mlx5 devlink support 5==================== 6 7This document describes the devlink features implemented by the ``mlx5`` 8device driver. 9 10Parameters 11========== 12 13.. list-table:: Generic parameters implemented 14 15 * - Name 16 - Mode 17 - Validation 18 * - ``enable_roce`` 19 - driverinit 20 - Type: Boolean 21 22 If the device supports RoCE disablement, RoCE enablement state controls 23 device support for RoCE capability. Otherwise, the control occurs in the 24 driver stack. When RoCE is disabled at the driver level, only raw 25 ethernet QPs are supported. 26 * - ``io_eq_size`` 27 - driverinit 28 - The range is between 64 and 4096. 29 * - ``event_eq_size`` 30 - driverinit 31 - The range is between 64 and 4096. 32 * - ``max_macs`` 33 - driverinit 34 - The range is between 1 and 2^31. Only power of 2 values are supported. 35 36The ``mlx5`` driver also implements the following driver-specific 37parameters. 38 39.. list-table:: Driver-specific parameters implemented 40 :widths: 5 5 5 85 41 42 * - Name 43 - Type 44 - Mode 45 - Description 46 * - ``flow_steering_mode`` 47 - string 48 - runtime 49 - Controls the flow steering mode of the driver 50 51 * ``dmfs`` Device managed flow steering. In DMFS mode, the HW 52 steering entities are created and managed through firmware. 53 * ``smfs`` Software managed flow steering. In SMFS mode, the HW 54 steering entities are created and manage through the driver without 55 firmware intervention. 56 57 SMFS mode is faster and provides better rule insertion rate compared to 58 default DMFS mode. 59 * - ``fdb_large_groups`` 60 - u32 61 - driverinit 62 - Control the number of large groups (size > 1) in the FDB table. 63 64 * The default value is 15, and the range is between 1 and 1024. 65 * - ``esw_multiport`` 66 - Boolean 67 - runtime 68 - Control MultiPort E-Switch shared fdb mode. 69 70 An experimental mode where a single E-Switch is used and all the vports 71 and physical ports on the NIC are connected to it. 72 73 An example is to send traffic from a VF that is created on PF0 to an 74 uplink that is natively associated with the uplink of PF1 75 76 Note: Future devices, ConnectX-8 and onward, will eventually have this 77 as the default to allow forwarding between all NIC ports in a single 78 E-switch environment and the dual E-switch mode will likely get 79 deprecated. 80 81 Default: disabled 82 * - ``esw_port_metadata`` 83 - Boolean 84 - runtime 85 - When applicable, disabling eswitch metadata can increase packet rate up 86 to 20% depending on the use case and packet sizes. 87 88 Eswitch port metadata state controls whether to internally tag packets 89 with metadata. Metadata tagging must be enabled for multi-port RoCE, 90 failover between representors and stacked devices. By default metadata is 91 enabled on the supported devices in E-switch. Metadata is applicable only 92 for E-switch in switchdev mode and users may disable it when NONE of the 93 below use cases will be in use: 94 1. HCA is in Dual/multi-port RoCE mode. 95 2. VF/SF representor bonding (Usually used for Live migration) 96 3. Stacked devices 97 98 When metadata is disabled, the above use cases will fail to initialize if 99 users try to enable them. 100 101 Note: Setting this parameter does not take effect immediately. Setting 102 must happen in legacy mode and eswitch port metadata takes effect after 103 enabling switchdev mode. 104 * - ``hairpin_num_queues`` 105 - u32 106 - driverinit 107 - We refer to a TC NIC rule that involves forwarding as "hairpin". 108 Hairpin queues are mlx5 hardware specific implementation for hardware 109 forwarding of such packets. 110 111 Control the number of hairpin queues. 112 * - ``hairpin_queue_size`` 113 - u32 114 - driverinit 115 - Control the size (in packets) of the hairpin queues. 116 117The ``mlx5`` driver supports reloading via ``DEVLINK_CMD_RELOAD`` 118 119Info versions 120============= 121 122The ``mlx5`` driver reports the following versions 123 124.. list-table:: devlink info versions implemented 125 :widths: 5 5 90 126 127 * - Name 128 - Type 129 - Description 130 * - ``fw.psid`` 131 - fixed 132 - Used to represent the board id of the device. 133 * - ``fw.version`` 134 - stored, running 135 - Three digit major.minor.subminor firmware version number. 136 137Health reporters 138================ 139 140tx reporter 141----------- 142The tx reporter is responsible for reporting and recovering of the following three error scenarios: 143 144- tx timeout 145 Report on kernel tx timeout detection. 146 Recover by searching lost interrupts. 147- tx error completion 148 Report on error tx completion. 149 Recover by flushing the tx queue and reset it. 150- tx PTP port timestamping CQ unhealthy 151 Report too many CQEs never delivered on port ts CQ. 152 Recover by flushing and re-creating all PTP channels. 153 154tx reporter also support on demand diagnose callback, on which it provides 155real time information of its send queues status. 156 157User commands examples: 158 159- Diagnose send queues status:: 160 161 $ devlink health diagnose pci/0000:82:00.0 reporter tx 162 163.. note:: 164 This command has valid output only when interface is up, otherwise the command has empty output. 165 166- Show number of tx errors indicated, number of recover flows ended successfully, 167 is autorecover enabled and graceful period from last recover:: 168 169 $ devlink health show pci/0000:82:00.0 reporter tx 170 171rx reporter 172----------- 173The rx reporter is responsible for reporting and recovering of the following two error scenarios: 174 175- rx queues' initialization (population) timeout 176 Population of rx queues' descriptors on ring initialization is done 177 in napi context via triggering an irq. In case of a failure to get 178 the minimum amount of descriptors, a timeout would occur, and 179 descriptors could be recovered by polling the EQ (Event Queue). 180- rx completions with errors (reported by HW on interrupt context) 181 Report on rx completion error. 182 Recover (if needed) by flushing the related queue and reset it. 183 184rx reporter also supports on demand diagnose callback, on which it 185provides real time information of its receive queues' status. 186 187- Diagnose rx queues' status and corresponding completion queue:: 188 189 $ devlink health diagnose pci/0000:82:00.0 reporter rx 190 191.. note:: 192 This command has valid output only when interface is up. Otherwise, the command has empty output. 193 194- Show number of rx errors indicated, number of recover flows ended successfully, 195 is autorecover enabled, and graceful period from last recover:: 196 197 $ devlink health show pci/0000:82:00.0 reporter rx 198 199fw reporter 200----------- 201The fw reporter implements `diagnose` and `dump` callbacks. 202It follows symptoms of fw error such as fw syndrome by triggering 203fw core dump and storing it into the dump buffer. 204The fw reporter diagnose command can be triggered any time by the user to check 205current fw status. 206 207User commands examples: 208 209- Check fw heath status:: 210 211 $ devlink health diagnose pci/0000:82:00.0 reporter fw 212 213- Read FW core dump if already stored or trigger new one:: 214 215 $ devlink health dump show pci/0000:82:00.0 reporter fw 216 217.. note:: 218 This command can run only on the PF which has fw tracer ownership, 219 running it on other PF or any VF will return "Operation not permitted". 220 221fw fatal reporter 222----------------- 223The fw fatal reporter implements `dump` and `recover` callbacks. 224It follows fatal errors indications by CR-space dump and recover flow. 225The CR-space dump uses vsc interface which is valid even if the FW command 226interface is not functional, which is the case in most FW fatal errors. 227The recover function runs recover flow which reloads the driver and triggers fw 228reset if needed. 229On firmware error, the health buffer is dumped into the dmesg. The log 230level is derived from the error's severity (given in health buffer). 231 232User commands examples: 233 234- Run fw recover flow manually:: 235 236 $ devlink health recover pci/0000:82:00.0 reporter fw_fatal 237 238- Read FW CR-space dump if already stored or trigger new one:: 239 240 $ devlink health dump show pci/0000:82:00.1 reporter fw_fatal 241 242.. note:: 243 This command can run only on PF. 244 245vnic reporter 246------------- 247The vnic reporter implements only the `diagnose` callback. 248It is responsible for querying the vnic diagnostic counters from fw and displaying 249them in realtime. 250 251Description of the vnic counters: 252 253- total_error_queues 254 number of queues in an error state due to 255 an async error or errored command. 256- send_queue_priority_update_flow 257 number of QP/SQ priority/SL update events. 258- cq_overrun 259 number of times CQ entered an error state due to an overflow. 260- async_eq_overrun 261 number of times an EQ mapped to async events was overrun. 262- comp_eq_overrun 263 number of times an EQ mapped to completion events was 264 overrun. 265- quota_exceeded_command 266 number of commands issued and failed due to quota exceeded. 267- invalid_command 268 number of commands issued and failed dues to any reason other than quota 269 exceeded. 270- nic_receive_steering_discard 271 number of packets that completed RX flow 272 steering but were discarded due to a mismatch in flow table. 273- generated_pkt_steering_fail 274 number of packets generated by the VNIC experiencing unexpected steering 275 failure (at any point in steering flow). 276- handled_pkt_steering_fail 277 number of packets handled by the VNIC experiencing unexpected steering 278 failure (at any point in steering flow owned by the VNIC, including the FDB 279 for the eswitch owner). 280 281User commands examples: 282 283- Diagnose PF/VF vnic counters:: 284 285 $ devlink health diagnose pci/0000:82:00.1 reporter vnic 286 287- Diagnose representor vnic counters (performed by supplying devlink port of the 288 representor, which can be obtained via devlink port command):: 289 290 $ devlink health diagnose pci/0000:82:00.1/65537 reporter vnic 291 292.. note:: 293 This command can run over all interfaces such as PF/VF and representor ports. 294