一、mlxconfig 介绍
NVIDIA Firmware Tools (MFT)中包含很多工具，mlxconfig是其中之一。用户可以使用mlxconfig工具更改Mellaox网卡的配置，修改后的配置会固定在固件中，即使后续将网卡重置或更瓣固件，配置也保持不变。要使用mlxconfig首先需要在系统中安装NVIDIA Firmware Tools (MFT)并启动MST。这一步可以参见本站文章NVIDIA Mellanox网卡如何收集日志。
二、使用要求
- OFED/WinOF driver to be installed and enabled (for ConnectX-3 and ConnectX-3 Pro)
- Access to the device through the PCI interface (pciconf/pci_cr)
- For the adapter cards below, the following firmware versions are required:
- ConnectX®-3/ConnectX-3 Pro: v2.31.5000 or above
- Connect-IB®: v10.10.6000 or above
- Supported devices:
- Adapter cards: ConnectX-3/ConnectX-3 Pro/Connect-IB/ConnectX-4/ConnectX-4 Lx/ConnectX-5/ConnectX-5 Ex/ConnectX-6/ConnectX-6 Dx/ConnectX-6 Lx/NVIDIA BlueField®/NVIDIA BlueField-2
- Switches: Switch-IB/Switch-IB 2/Spectrum™/Spectrum-2/Spectrum-3/Quantum
- Changing device configurations enabled.
[!NOTE]
为了使配置成功后的更改生效，需要重新启动系统。
三、mlxconfig命令简介
1、基本命令格式。
# mlxconfig [Options] <commands> [Parameters]
2、常用参数。
| Parameters | Description |
|---|---|
| -a|–all_attrs | Show all attributes in the XML template |
| -d|–dev |
Performs operation for a specified mst device |
| -b|–db |
Use a specific database file. |
| -f|–file <conf.file> | Raw configuration file |
| -y|–yes | Answers yes in prompt |
| -e|-enable_verbosity | Show default and current configurations. Note: For 5th generation (Group II) devices, the –enable_verbosity option works with ConnectX-4 firmware v12.14.0016 and above for querying the default configurations, and with ConnectX-4 firmware v12.17.1010 and above for querying the current configurations. |
| -v|–version | Displays version info |
| -h|–help | Displays help message |
| -p|–private_key | pem file for private key |
| -u|–key_uuid | keypair uuid |
| -eng|–openssl_engine | OpenSSL engine name |
| -k|–open_ssl_key_id | OpenSSL key identifier |
| q[uery] | Queries the supported configurations. Note: Query command will query a single device if a device is specified. Otherwise, it will query all devices on the machine. |
| s[et] | Sets configurations to a specific device |
| set_raw | Sets raw configuration file (5th generation/Group II devices only) |
| get_raw | Gets raw configuration file (5th generation/Group II devices only) |
| i[show_confs] | Display information about all configurations |
| clear_semaphore | Clear the tool’s semaphore |
| r[eset] | Resets configurations to their default value |
| backup | Backs up configurations to a file (only 5th generation (Group II) devices). Use set_raw command to restore file. |
| gen_tlvs_file | Generate a List of all TLVs. TLVs output file name must be specified |
| g[en_xml_template] | Generate an XML template. TLVs input file name and XML output file name must be specified |
| xml2raw | Generate a Raw file from an XML file. XML input file name and raw output file name must be specified |
| raw2xml | Generate an XML file from a Raw file. raw input file name and XML output file name must be specified |
| xml2bin | Generate binary configuration dump file from XML file. XML input file name and bin output file name must be specified. |
| create_conf | Generate configuration file from XML file. XML input file name and bin output file name must be specified. |
| apply | Apply a configuration file, that was created with create_conf command. bin input file name must be specified. |
| -t|–device_type <switch/hca/linkx> | Specify the device type. |
| -s|–session_id | Specify the session id for token keep alive session. |
| -st|–session_time | Specify session time for token keep alive session. |
| -tkn|–token_type | Specify token type. |
| challenge_request | Send a token challenge request to the device. Token type must be specified. |
| remote_token_keep_alive | Start a remote token session for a specified time. session id must be specified. |
| token_supported | Query which tokens are supported. |
| query_token_session | Query the status of a token session. |
| end_token_session | End an active token session. |
3、命令演示，查询网卡当前所有配置信息。
[root@localhost ~]# mlxconfig -d /dev/mst/mt4127_pciconf0 q
Device #1:
----------
Device type: ConnectX6LX
Name: SN37A28486_Ax
Description: ThinkSystem Mellanox ConnectX-6 Lx 10/25GbE SFP28 2-port PCIe Ethernet Adapter
Device: /dev/mst/mt4127_pciconf0
Configurations: Next Boot
MEMIC_BAR_SIZE 0
MEMIC_SIZE_LIMIT _256KB(1)
HOST_CHAINING_MODE DISABLED(0)
HOST_CHAINING_CACHE_DISABLE False(0)
HOST_CHAINING_DESCRIPTORS Array[0..7]
HOST_CHAINING_TOTAL_BUFFER_SIZE Array[0..7]
FLEX_PARSER_PROFILE_ENABLE 0
PROG_PARSE_GRAPH False(0)
FLEX_IPV4_OVER_VXLAN_PORT 0
ROCE_NEXT_PROTOCOL 254
ESWITCH_HAIRPIN_DESCRIPTORS Array[0..7]
ESWITCH_HAIRPIN_TOT_BUFFER_SIZE Array[0..7]
PF_BAR2_SIZE 0
PF_NUM_OF_VF_VALID False(0)
NON_PREFETCHABLE_PF_BAR False(0)
VF_VPD_ENABLE False(0)
PF_NUM_PF_MSIX_VALID False(0)
PER_PF_NUM_SF False(0)
STRICT_VF_MSIX_NUM False(0)
VF_NODNIC_ENABLE False(0)
NUM_PF_MSIX_VALID True(1)
NUM_OF_VFS 8
NUM_OF_PF 2
PF_BAR2_ENABLE False(0)
SRIOV_EN True(1)
PF_LOG_BAR_SIZE 5
VF_LOG_BAR_SIZE 0
NUM_PF_MSIX 63
NUM_VF_MSIX 11
INT_LOG_MAX_PAYLOAD_SIZE AUTOMATIC(0)
PCIE_CREDIT_TOKEN_TIMEOUT 0
NETWORK_PORTS_NUMBER 0
RT_PPS_ENABLED_ON_POWERUP False(0)
PHY_COUNT_LINK_UP_DELAY DELAY_NONE(0)
ACCURATE_TX_SCHEDULER False(0)
PARTIAL_RESET_EN False(0)
RESET_WITH_HOST_ON_ERRORS False(0)
PCI_DOWNSTREAM_PORT_OWNER Array[0..15]
CQE_COMPRESSION BALANCED(0)
IP_OVER_VXLAN_EN False(0)
MKEY_BY_NAME False(0)
PRIO_TAG_REQUIRED_EN False(0)
UCTX_EN True(1)
REAL_TIME_CLOCK_ENABLE False(0)
RDMA_SELECTIVE_REPEAT_EN False(0)
PCI_ATOMIC_MODE PCI_ATOMIC_DISABLED_EXT_ATOMIC_ENABLED(0)
TUNNEL_ECN_COPY_DISABLE False(0)
LRO_LOG_TIMEOUT0 6
LRO_LOG_TIMEOUT1 7
LRO_LOG_TIMEOUT2 8
LRO_LOG_TIMEOUT3 13
LOG_TX_PSN_WINDOW 7
VF_MIGRATION_MODE DEVICE_DEFAULT(0)
LOG_MAX_OUTSTANDING_WQE 7
ROCE_ADAPTIVE_ROUTING_EN False(0)
TUNNEL_IP_PROTO_ENTROPY_DISABLE False(0)
PCC_INT_NP_RTT_DSCP 26
PCC_INT_NP_RTT_DSCP_EN False(0)
PCC_INT_NP_RTT_DATA_MODE RTT_V0(64)
PCC_INT_EN False(0)
PCC_INT_SYSTEM_RTT 0
ICM_CACHE_MODE DEVICE_DEFAULT(0)
TX_SCHEDULER_BURST 0
ROCE_CC_LEGACY_DCQCN False(0)
LOG_MAX_QUEUE 17
LARGE_MTU_TWEAK_64 False(0)
AES_XTS_TWEAK_INC_64 False(0)
CRYPTO_POLICY UNRESTRICTED(1)
LOG_DCR_HASH_TABLE_SIZE 11
MAX_PACKET_LIFETIME 0
DCR_LIFO_SIZE 16384
ROCE_CC_PRIO_MASK_P1 255
ROCE_CC_CNP_MODERATION_P1 DEVICE_DEFAULT(0)
ROCE_CC_SHAPER_COALESCE_P1 DEVICE_DEFAULT(0)
ROCE_CC_PRIO_MASK_P2 255
ROCE_CC_CNP_MODERATION_P2 DEVICE_DEFAULT(0)
ROCE_CC_SHAPER_COALESCE_P2 DEVICE_DEFAULT(0)
CLAMP_TGT_RATE_AFTER_TIME_INC_P1 True(1)
CLAMP_TGT_RATE_P1 False(0)
RPG_TIME_RESET_P1 300
RPG_BYTE_RESET_P1 32767
RPG_THRESHOLD_P1 1
RPG_MAX_RATE_P1 0
RPG_AI_RATE_P1 5
RPG_HAI_RATE_P1 50
RPG_GD_P1 11
RPG_MIN_DEC_FAC_P1 50
RPG_MIN_RATE_P1 1
RATE_TO_SET_ON_FIRST_CNP_P1 0
DCE_TCP_G_P1 1019
DCE_TCP_RTT_P1 1
RATE_REDUCE_MONITOR_PERIOD_P1 4
INITIAL_ALPHA_VALUE_P1 1023
MIN_TIME_BETWEEN_CNPS_P1 4
CNP_802P_PRIO_P1 6
CNP_DSCP_P1 48
CLAMP_TGT_RATE_AFTER_TIME_INC_P2 True(1)
CLAMP_TGT_RATE_P2 False(0)
RPG_TIME_RESET_P2 300
RPG_BYTE_RESET_P2 32767
RPG_THRESHOLD_P2 1
RPG_MAX_RATE_P2 0
RPG_AI_RATE_P2 5
RPG_HAI_RATE_P2 50
RPG_GD_P2 11
RPG_MIN_DEC_FAC_P2 50
RPG_MIN_RATE_P2 1
RATE_TO_SET_ON_FIRST_CNP_P2 0
DCE_TCP_G_P2 1019
DCE_TCP_RTT_P2 1
RATE_REDUCE_MONITOR_PERIOD_P2 4
INITIAL_ALPHA_VALUE_P2 1023
MIN_TIME_BETWEEN_CNPS_P2 4
CNP_802P_PRIO_P2 6
CNP_DSCP_P2 48
LLDP_NB_DCBX_P1 False(0)
LLDP_NB_RX_MODE_P1 OFF(0)
LLDP_NB_TX_MODE_P1 OFF(0)
LLDP_NB_DCBX_P2 False(0)
LLDP_NB_RX_MODE_P2 OFF(0)
LLDP_NB_TX_MODE_P2 OFF(0)
ROCE_RTT_RESP_DSCP_P1 0
ROCE_RTT_RESP_DSCP_MODE_P1 DEVICE_DEFAULT(0)
ROCE_RTT_RESP_DSCP_P2 0
ROCE_RTT_RESP_DSCP_MODE_P2 DEVICE_DEFAULT(0)
DCBX_IEEE_P1 True(1)
DCBX_CEE_P1 True(1)
DCBX_WILLING_P1 True(1)
DCBX_IEEE_P2 True(1)
DCBX_CEE_P2 True(1)
DCBX_WILLING_P2 True(1)
KEEP_ETH_LINK_UP_P1 True(1)
KEEP_IB_LINK_UP_P1 False(0)
KEEP_LINK_UP_ON_BOOT_P1 False(0)
KEEP_LINK_UP_ON_STANDBY_P1 False(0)
DO_NOT_CLEAR_PORT_STATS_P1 False(0)
AUTO_POWER_SAVE_LINK_DOWN_P1 False(0)
KEEP_ETH_LINK_UP_P2 True(1)
KEEP_IB_LINK_UP_P2 False(0)
KEEP_LINK_UP_ON_BOOT_P2 False(0)
KEEP_LINK_UP_ON_STANDBY_P2 False(0)
DO_NOT_CLEAR_PORT_STATS_P2 False(0)
AUTO_POWER_SAVE_LINK_DOWN_P2 False(0)
NUM_OF_VL_P1 _4_VLs(3)
NUM_OF_TC_P1 _8_TCs(0)
NUM_OF_PFC_P1 8
VL15_BUFFER_SIZE_P1 0
QOS_TRUST_STATE_P1 TRUST_PCP(1)
NUM_OF_VL_P2 _4_VLs(3)
NUM_OF_TC_P2 _8_TCs(0)
NUM_OF_PFC_P2 8
VL15_BUFFER_SIZE_P2 0
QOS_TRUST_STATE_P2 TRUST_PCP(1)
DUP_MAC_ACTION_P1 LAST_CFG(0)
MPFS_MC_LOOPBACK_DISABLE_P1 False(0)
MPFS_UC_LOOPBACK_DISABLE_P1 False(0)
UNKNOWN_UPLINK_MAC_FLOOD_P1 False(0)
SRIOV_IB_ROUTING_MODE_P1 LID(1)
IB_ROUTING_MODE_P1 LID(1)
DUP_MAC_ACTION_P2 LAST_CFG(0)
MPFS_MC_LOOPBACK_DISABLE_P2 False(0)
MPFS_UC_LOOPBACK_DISABLE_P2 False(0)
UNKNOWN_UPLINK_MAC_FLOOD_P2 False(0)
SRIOV_IB_ROUTING_MODE_P2 LID(1)
IB_ROUTING_MODE_P2 LID(1)
PHY_AUTO_NEG_P1 DEVICE_DEFAULT(0)
PHY_RATE_MASK_OVERRIDE_P1 False(0)
PHY_FEC_OVERRIDE_P1 DEVICE_DEFAULT(0)
PHY_AUTO_NEG_P2 DEVICE_DEFAULT(0)
PHY_RATE_MASK_OVERRIDE_P2 False(0)
PHY_FEC_OVERRIDE_P2 DEVICE_DEFAULT(0)
PF_TOTAL_SF 0
PF_SD_GROUP 0
PF_SF_BAR_SIZE 0
PF_NUM_PF_MSIX 63
ROCE_CONTROL ROCE_ENABLE(2)
PCI_WR_ORDERING per_mkey(0)
MULTI_PORT_VHCA_EN False(0)
PORT_OWNER True(1)
ALLOW_RD_COUNTERS True(1)
RENEG_ON_CHANGE True(1)
TRACER_ENABLE True(1)
IP_VER IPv4(0)
BOOT_UNDI_NETWORK_WAIT 0
UEFI_HII_EN True(1)
BOOT_DBG_LOG False(0)
UEFI_LOGS DISABLED(0)
BOOT_VLAN 1
LEGACY_BOOT_PROTOCOL PXE(1)
BOOT_INTERRUPT_DIS False(0)
BOOT_LACP_DIS True(1)
BOOT_VLAN_EN False(0)
BOOT_PKEY 0
P2P_ORDERING_MODE DEVICE_DEFAULT(0)
DYNAMIC_VF_MSIX_TABLE False(0)
EXP_ROM_UEFI_ARM_ENABLE True(1)
EXP_ROM_UEFI_x86_ENABLE True(1)
EXP_ROM_PXE_ENABLE True(1)
ADVANCED_PCI_SETTINGS False(0)
SAFE_MODE_THRESHOLD 10
SAFE_MODE_ENABLE True(1)
4、单独查询某一项设置。
# mlxconfig -d /dev/mst/mt4117_pciconf0 --enable_verbosity q
Device #1:
----------
Device type: ConnectX4LX
PCI device: /dev/mst/mt4117_pciconf0
Configurations: Default Current Next Boot
* NUM_OF_VFS 8 5 5
SRIOV_EN True(1) True(1) True(1)
The '*' shows parameters with next value different from default/current value.
5、修改网卡设置，使用set命令。
# mlxconfig -d /dev/mst/mt4099_pciconf0 set WOL_MAGIC_EN_P2=1 NUM_OF_VFS=24
Device type: ConnectX-3
PCI device: /dev/mst/mt4099_pciconf0
Configurations: Next Boot New
NUM_OF_VFS 16 24
WOL_MAGIC_EN_P2 False(0) True(1)
Apply new Configuration?(y/n) [n]: y
Applying... Done!
-I- Please reboot the system to load new configurations.
6、重置网卡（Load default），使用reset命令。
# mlxconfig -d /dev/mst/mt4099_pciconf0 reset
Reset configuration for device /dev/mst/mt4099_pciconf0? ? (y/n) [n] : y
Applying... Done!
-I- Please power-cycle device to load new configurations.
>mlxconfig -d /dev/mst/mt4099_pciconf0 query
Device 1:
----------
Device type: ConnectX-3
PCI Device: /dev/mst/mt4099_pciconf0
Configurations: Next Boot
SRIOV_EN True(1)
NUM_OF_VFS 8
WOL_MAGIC_EN_P1 False(0)
WOL_MAGIC_EN_P2 False(0)
7、网卡支持的配置和参数列表可以通过“mlxconfig -d
[root@localhost ~]# mlxconfig -d /dev/mst/mt4127_pciconf0 show_confs
List of configurations the device /dev/mst/mt4127_pciconf0 may support:
MODULE POWER:
MODULE_MAX_POWER_OVERRIDE_M0=<BINARY> Max consumption power allowed, from thermal perspective, for module cage.Please advise that change of this field value requires the suitable LFM capability.
0x0 - Maximum power is from devices default capability
Other - Maximum power consumption in multiples of 0.25W.
Note: Device has upper limit of max power consumption to prevent harmful configuration
MODULE_MAX_POWER_OVERRIDE_M1=<BINARY> Max consumption power allowed, from thermal perspective, for module cage.Please advise that change of this field value requires the suitable LFM capability.
0x0 - Maximum power is from devices default capability
Other - Maximum power consumption in multiples of 0.25W.
Note: Device has upper limit of max power consumption to prevent harmful configuration
MODULE SPLIT:
MODULE_SPLIT_M0=<BINARY> MAC number assignment for logical lane <i>.
MAC numbering starts from value of 1.
number of logical lanes assigned to the same MAC number will define the port width.
Value of 0 indicates lane is on default device mapping
Value of 0xFF indicates lane is unmapped
MODULE_SPLIT_M1=<BINARY> MAC number assignment for logical lane <i>.
MAC numbering starts from value of 1.
number of logical lanes assigned to the same MAC number will define the port width.
Value of 0 indicates lane is on default device mapping
Value of 0xFF indicates lane is unmapped
MEMIC CONF:
MEMIC_ATOMIC=<DEVICE_DEFAULT|MEMIC_ATOMIC_DISABLE|MEMIC_ATOMIC_ENABLE>Indicates whether Atomic operations address range is supported for MEMIC. When accessing this range Read and Write operation are translated into 'Atomic test-and-set' and 'Atomic add' respectively.
0x0: DEVICE_DEFAULT
0x1: MEMIC_ATOMIC_DISABLE
0x2: MEMIC_ATOMIC_ENABLE
other values are reserved
MEMIC_ATOMIC_ENDIANESS=<DEVICE_DEFAULT|MEMIC_ATOMIC_ENDIANNESS_BIG|MEMIC_ATOMIC_ENDIANNESS_LITTLE>Indicates what is the endianness of MEMIC for DM operations.
This configuration does not apply to Internal CPU.
0x0: DEVICE_DEFAULT
0x1: MEMIC_ATOMIC_ENDIANNESS_BIG
0x2: MEMIC_ATOMIC_ENDIANNESS_LITTLE
MEMIC_BAR_SIZE=<NUM> The amount of BAR size assigned for MEMIC. The size in bytes is memic_size_limit*2^log_memic_bar_size.
When activating MEMIC Atomic access by MEMIC_ATOMIC, the allocated BAR size for MEMIC is doubled.
MEMIC_SIZE_LIMIT=<DISABLED|_1024KB|_256KB|_512KB>The maximum amount of internal device memory that can be consumed by the MEMIC application.
0x0: DISABLED
0x1: _256KB
0x2: _512KB
0x3: _1024KB
HOST CHAINING CONF:
HOST_CHAINING_CACHE_DISABLE=<False|True>When TRUE, host chaining data is not cached on the device.
HOST_CHAINING_DESCRIPTORS=<NUM> Log(base 2) of the number of packets descriptors that should be allocated by the host for host chaining for a given IEEE802.1p priority i.
0 means that no descriptors are allocated for this priority and traffic with this priority will be dropped.
HOST_CHAINING_MODE=<BASIC|DISABLED> Enable and select host-chaining mode.
0x0: DISABLED
0x1: BASIC
other values are reserved
HOST_CHAINING_TOTAL_BUFFER_SIZE=<NUM> Log(base 2) of the buffer size (in bytes) allocated for host chaining for a given IEEE802.1p priority i.
0 means no buffer for this priority and traffic with this priority will be dropped.
FPGA CONF:
FPGA_AUTO_RELOAD=<False|True> When set, the FPGA bitstream is reloaded from flash as part of the ConnectX boot sequence.
INTERNAL CPU CONF:
INTERNAL_CPU_ESWITCH_MANAGER=<ECPF|EXT_HOST_PF>Defines the owner of Eth Embedded Switch responsibilities
0x0: ECPF
0x1: EXT_HOST_PF
Valid for INTERNAL_CPU_MODEL = EMBEDDED_CPU
INTERNAL_CPU_IB_VPORT0=<ECPF|EXT_HOST_PF>Defines the owner of IB Vport0 responsibilities
0x0: ECPF
0x1: EXT_HOST_PF
INTERNAL_CPU_MODEL=<EMBEDDED_CPU|SEPARATED_HOST>Select the model for the Internal CPU
0x0: SEPARATED_HOST - Supported only when NV_INTERNAL_CPU_CAP.separate_host_model_supported==1.
0x1: EMBEDDED_CPU - Supported only when NV_INTERNAL_CPU_CAP.embedded_cpu_model_supported==1.
other values are reserved
This NVCONFIG parameter can only be configured from the embedded CPU.
INTERNAL_CPU_OFFLOAD_ENGINE=<DISABLED|ENABLED>Defines whether the Internal CPU is used as an offload engine
0x0: ENABLED
0x1: DISABLED
Valid for INTERNAL_CPU_MODEL = EMBEDDED_CPU
INTERNAL_CPU_PAGE_SUPPLIER=<ECPF|EXT_HOST_PF>Defines the owner of providing ICM pages to the external host functions
0x0: ECPF
0x1: EXT_HOST_PF
Valid for INTERNAL_CPU_MODEL = EMBEDDED_CPU
FLEX PARSER CONF:
FLEX_IPV4_OVER_VXLAN_PORT=<NUM> The UDP port for incoming IPoVxLAN traffic (non-standard).
When set, also affects flex_vxlan_gpe_supported when enabled on same profile.
FLEX_PARSER_PROFILE_ENABLE=<NUM> Indicates which flex parser profile to enable. Each profile supports a set of protocols. The support indication and the set of protocols supported by profile 'x' reported in NC_FLEX_PARSER_CAP.flex_parser_profile_x_supported.
PROG_PARSE_GRAPH=<False|True> When TRUE, the device parse graph may be dynamically configured. The amount of programmable resources is reduced according to the amount of protocols already supported by FLEX_PARSER_PROFILE_ENABLE.
ROCE 1 5 CONF:
ROCE_NEXT_PROTOCOL=<NUM> The next protocol value set in the IPv4/IPv6 packets for RoCE v1.5.
INTERNAL HAIRPIN CONF:
ESWITCH_HAIRPIN_DESCRIPTORS=<NUM> Log(base 2) of the number of packets descriptors allocated internally for hairpin for a given IEEE802.1p priority i.
0 means that no descriptors are allocated for this priority and traffic with this priority will be dropped.
ESWITCH_HAIRPIN_TOT_BUFFER_SIZE=<NUM> Log(base 2) of the buffer size (in bytes) allocated internally for hairpin for a given IEEE802.1p priority i.
0 means no buffer for this priority and traffic with this priority will be dropped.
DPA AUTH:
DPA_AUTHENTICATION=<False|True> When TRUE, DPA code is authenticated before executed.
GLOBAL PCI CONF:
DPU_RESET_NOTIFICATION_ENABLED=<DISABLED|ENABLED>When set, the NIC can report PCIe errors through AER cap to all hosts it is connected to on DPU reset\panic. (If no error reporting is enabled, the DPU will not report any error.)
0x0: DISABLED
0x1: ENABLED
FPP_EN=<False|True> When this bit is cleared, the device exposes a single PCI function for both ports. When set, the device exposes one or more PCI functions for each port (this is the only mode supported by ConnectX-4 devices).
HIDE_PORT2_PF=<False|True> When TRUE, the device will not advertise the PFs associated with port 2, except for the Embedded CPU (ECPF) if exists. This configuration is available only when NV_PCI_CONF.ADVANCED_PCI_SETTINGS is TRUE.
INTERNAL_CPU_RSHIM=<DISABLED|ENABLED> Defines whether an RSHIM function will be exposed (when available) to the external host.
0x0: ENABLED
0x1: DISABLED
NON_PREFETCHABLE_PF_BAR=<False|True> When set, the PF BAR prefetchable bit is cleared.
Note: PCI switches and operation systems have dedicated quotas for non-prefetchable memory, hence, you may need to decrease log_pf_uar_bar_size to enable this feature.
NUM_OF_PF=<NUM> Total number of Network PCIe functions (PFs) exposed by the device. In case the number of PFs cannot be equally distributed between the number of ports, the remainder of PFs will be distribute between the ports with the lower numbers. For a Multi-Host device, this number is applied to each host individually.
Value 0 is only supported when an Emulated PCI Switch is present or another type of PF (e.g. emulated device).
NUM_OF_VFS=<NUM> The total number of Virtual Functions (VFs) that can be supported, for each PF.
Valid when PF_NUM_OF_VF_VALID is FALSE
NUM_PF_MSIX=<NUM> Number of MSI-X vectors and EQs per PF.
NUM_PF_MSIX_VALID=<False|True> When set, the num_pf_msix field is valid.
When cleared, the number of PF MSI-X is defined by PER_PF_NUM_MSIX
NUM_VF_MSIX=<NUM> Number of MSI-X vectors and EQs per VF.
PARTIAL_WRITE_CACHE_MODE=<ADDRESS_BASED|DEVICE_DEFAULT|DISABLED|FUNCTION_BASED>Specifies the mode of operation for PCI write cache, for partial cacheline writes:
0x0: DEVICE_DEFAULT - device default configuration
0x1: DISABLED
0x2: FUNCTION_BASED
0x3: ADDRESS_BASED
other values are reserved
PER_PF_NUM_SF=<False|True> When TRUE, the SFs configuration is defined by TOTAL_SF and SF_BAR_SIZE for each PF individually. In case they are not defined for a PF, device defaults are used.
When FALSE, the SFs configuration is defined by device defaults.
Valid only when PF_BAR2_ENABLE is set to FALSE.
PER_PF_NUM_VF_MSIX=<False|True> When set to TRUE, the VF MSI-X configuration is defined by NUM_VF_MSIX for each PF individually. In case they are not defined for a PF, device defaults are used for that PF.
When set to FALSE, the MSI-X configuration is defined by
NUM_VF_MSIX for all PFs.
PF_BAR2_ENABLE=<False|True> When TRUE, BAR2 is exposed on all external host PFs (but not on the embedded ARM PFs/ECPFs). The BAR2 size is defined by the log_pf_bar2_size.
When FALSE, the SFs and BAR2 configurations are defined by PER_PF_NUM_SF
PF_BAR2_SIZE=<NUM> Log (base 2) of the size of a PF's BAR2 size, given in MB.
PF_LOG_BAR_SIZE=<NUM> Log 2 of the size of a PF"s UAR BAR in MBs.
PF_NUM_OF_VF_VALID=<False|True> When TRUE, the number of VFs is defined for each networking PF individually in PF_NUM_OF_VF. In case they are not defined for a PF, device defaults are used instead.
When FALSE, the number of VFs is defined symmetrically for all PFs by NUM_OF_VFS
PF_NUM_PF_MSIX_VALID=<False|True> When TRUE, the MSI-X configuration is defined by PF_NUM_PF_MSIX for each PF individually. In case they are not defined for a PF, device defaults are used.
When FALSE, the MSI-X configuration is defined by device defaults
Valid only when NUM_PF_MSIX_VALID is set to FALSE.
SRIOV_EN=<False|True> Enable Single-Root I/O Virtualization (SR-IOV)
STRICT_VF_MSIX_NUM=<False|True> When set to TRUE, num_vf_msix defines a strict number (no calculation and roundup).
VF_LOG_BAR_SIZE=<NUM> Log 2 of the size of a VF"s UAR BAR in MBs.
VF_NODNIC_ENABLE=<False|True> When set to TRUE, VF has VSC Gateway exposed through PCI, and may access the NODNIC initialization segment
VF_VPD_ENABLE=<False|True> When set, VPD Capability is exposed to Virtual Functions.
TPT CONF:
INT_LOG_MAX_PAYLOAD_SIZE=<AUTOMATIC|_4KB>Sets the PCIe burst size for the NIC internal Translation and Protection (TPT) mechanism.
0x0: AUTOMATIC
0xC: _4KB
GLOBAL PCI CONF 2:
MAX_ACC_OUT_READ=<NUM> Maximum accumulated outstanding Read requests bytes. Value is given in units of 1K bytes.
The limit is applied for each PCI bus link individually.
Value 0x0 indicates the use of device defaults.
Configuration is available only when NV_PCI_CONF.ADVANCED_PCI_SETTINGS is TRUE.
PCIE_CREDIT_TOKEN_TIMEOUT=<NUM> PCIe credit timeout. In case a pending transaction has no credits longer than this timeout value, buffered transactions will be dropped. Value is given in milliseconds in the range of 10 to 4,000.
Value 0x0 indicates the device will adjust the timeout automatically.
Value 0xffff disables the feature.
PCI SWITCH EVB CONF:
PCI_SW_LANES_MAPPING=<BINARY> Evaluated from the LSB, each bit represents the mapping of two PCI lanes to independent PCI buses where a value of zero indicates that the corresponding lanes are the first lanes of a new bus, and a value of 1 indicates that the corresponding lanes belongs to the same bus as the previous lanes.
Examples:
11111110b => all 16 lanes are mapped to the host. The actual bus width is learned dynamically so it may also be 8 or 4 lanes.
00000000b => The host bus width is up to 2 lanes and there are additional 7 independent buses.
11101110b => The host bus is up to 8 lanes and there is an addition bus of up to 8 lanes.
PCI_SW_LANES_REMAPPING_EN=<False|True> When set allows remapping of the PCI switch lanes to slots according to the PCI_SW_LANES_MAPPING parameter.
POWER CONF:
ACCURATE_TX_SCHEDULER=<False|True> When TRUE, the device will optimize the transmit scheduler for high accuracy.
When False, the device defaults will apply for the scheduler.
ADVANCED_POWER_SETTINGS=<False|True> Show/hide additional power settings parameters.
Warning: Wrong power settings may cause physical damage.
DISABLE_SLOT_POWER_LIMITER=<False|True> When cleared, the device is not allowed to consume more than 25W from the PCIe power rails, unless a PCI slot power limit message statesthat a new power limit is received.
When set, the slot power limiter is disabled, and the device is allowed to consume more than 25W from the PCIe power rails.
Note: if the power limiter is active and there is not enough power, the device will shut down the network modules.
LAG_RESOURCE_ALLOCATION=<DEVICE_DEFAULT|PRE_ALLOCATION>Defines the way resources are allocated for LAG
0x0: DEVICE_DEFAULT
0x1: PRE_ALLOCATION - resources are preallocated regardless of LAG activation
using pre-allocation allows LAG activation regardless of functions state
Other values are reserved
NETWORK_PORTS_NUMBER=<NUM> Number of network ports enabled by the device, If the value exceeds the number of ports supported by the card, this field will be ignored. Value 0x0 indicates device defaults.
PARTIAL_RESET_EN=<False|True> When set, the partial reset flow is enabled. This reset skips on resetting GPIOs, loading PLL new parameters and performing self- tests. Partial reset will be triggered only if no exceptions requiring full reset were detected (for example, if any of the boot record"s parameter had changed)
PHY_COUNT_LINK_UP_DELAY=<DELAY_10_SEC|DELAY_20_SEC|DELAY_NONE>Defines a period after Link UP where phy counters activation is delayed
0x0: DELAY_NONE
0x1: DELAY_10_SEC
0x2: DELAY_20_SEC
Other values are reserved
RESET_WITH_HOST_ON_ERRORS=<False|True> When set, host reset will trigger a device reset, if the device has previously detected fatal errors.
RT_PPS_ENABLED_ON_POWERUP=<False|True> When TRUE the PPS_OUT will be activated on power up
Applicable only when REAL_TIME_CLOCK_ENABLE is TRUE
SW_RECOVERY_ON_ERRORS=<False|True> When set, SW will be instructed to perform a recovery flow when health buffer reports an error.
HOST SYNC CONF:
ICMD_SEM_LOCK_THRESHOLD=<NUM> Host will be expelled from VSEC Gateway usage if it holds the ICMD Global Semaphore for a period longer than this threshold. Value is given in units of 0.1sec
Value 0xFFFF indicates this threshold is infinite (disabled)
IDLE_ICMD_SEM_LOCK_THRESHOLD=<NUM> Host will be expelled from VSEC Gateway usage if it holds the ICMD Global Semaphore, without issuing any ICMD, for a period longer than this threshold. Value is given in units of 0.1sec
Value 0xFFFF indicates this threshold is infinite (disabled)
VSEC_EXPULSION_DURATION=<NUM> Host VSEC Gateway expulsion will persist for this period. Value is given in units of 0.1sec
EMULATION NVME CONF:
NVME_EMULATION_CLASS_CODE=<NUM> PCIe class_code register for the NVME emulated device
NVME_EMULATION_DEVICE_ID=<NUM> PCIe device_id config register for the NVME emulated device.
NVME_EMULATION_ENABLE=<False|True> When set to TRUE, NVME device emulation is enabled.
NVME_EMULATION_MAX_QUEUE_DEPTH=<NUM> Log (base 2) of the maximal queue depth of NVME Physical Functions.
Value 0 indicates device default.
NVME_EMULATION_NUM_MSIX=<NUM> Number of MSI-X Vectors assigned for each PF/VF of the NVME emulation device. Value 0x0 will use device defaults
NVME_EMULATION_NUM_PF=<NUM> Total number of PCIe functions (PFs) exposed by the device for NVME emulation.
NVME_EMULATION_NUM_VF=<NUM> The total number of Virtual Functions (VFs) that can be supported for each PF.
NVME_EMULATION_NUM_VF_MSIX=<NUM> Number of MSI-X vectors assigned to each VF associated with this type of PF. Value 0x0 indicates device defaults.
NVME_EMULATION_REVISION_ID=<NUM> PCIe revision_id register for the NVME emulated device
NVME_EMULATION_SUBSYSTEM_ID=<NUM> PCIe subsystem_id register for the NVME emulated device
NVME_EMULATION_SUBSYSTEM_VENDOR_ID=<NUM>PCIe subsystem_vendor_id register for the NVME emulated device
NVME_EMULATION_VENDOR_ID=<NUM> PCIe vendor_id config register for the NVME emulated device.
EMULATION PCI SWITCH CONF:
PCI_SWITCH_EMULATION_ENABLE=<False|True>When TRUE, the device will expose a PCI switch. All PF configurations are applied on the switch downstream ports. On such case, each PF will have a different PCI device on the emulated switch. This configuration allows to expose 0 network PFs toward host
PCI_SWITCH_EMULATION_NUM_PORT=<NUM> Number of emulated switch ports. Each port can hold a single PCI device (emulated or not).
VIRTIO_EMULATION_HOTPLUG_TRANS=<False|True>When TRUE, the device will support hotplug of EMulated transitional VIRTIO devices
EMULATION VIRTIO NET CONF:
VIRTIO_NET_EMULATION_ENABLE=<False|True>When set to TRUE, VIRTIO_NET device emulation is enabled.
VIRTIO_NET_EMULATION_NUM_MSIX=<NUM> Number of MSI-X Vectors assigned for each PF/VF of the VIRTIO_NET emulation device. Value 0x0 will use device defaults
VIRTIO_NET_EMULATION_NUM_PF=<NUM> Total number of PCIe functions (PFs) exposed by the device for VIRTIO_NET emulation.
VIRTIO_NET_EMULATION_NUM_VF=<NUM> The total number of Virtual Functions (VFs) that can be supported for each PF.
VIRTIO_NET_EMULATION_NUM_VF_MSIX=<NUM> Number of MSI-X vectors assigned to each VF associated with this type of PF. Value 0x0 indicates device defaults.
VIRTIO_NET_EMULATION_PF_PCI_LAYOUT=<VIRTIO_1_X|VIRTIO_TRANSITIONAL>Indicates which VIrtIO specification the PCI layout of the emulated Physical Function(s) will follow.
0x0: VIRTIO_1_X
0x1: VIRTIO_TRANSITIONAL
other values are reserved
VIRTIO_NET_EMULATION_SUBSYSTEM_ID=<NUM> PCIe subsystem_id register for the VIRTIO_NET emulated device
VIRTIO_NET_EMULATION_VF_PCI_LAYOUT=<VIRTIO_1_X|VIRTIO_TRANSITIONAL>Indicates which VIrtIO specification the PCI layout of the emulated Virtual Function(s) will follow.
0x0: VIRTIO_1_X
0x1: VIRTIO_TRANSITIONAL
other values are reserved
VIRTIO_NET_EMU_SUBSYSTEM_VENDOR_ID=<NUM>PCIe subsystem_vendor_id register for the VIRTIO_NET emulated device
EMULATION VIRTIO BLK CONF:
VIRTIO_BLK_EMULATION_ENABLE=<False|True>When set to TRUE, VIRTIO_BLK device emulation is enabled.
VIRTIO_BLK_EMULATION_NUM_MSIX=<NUM> Number of MSI-X Vectors assigned for each PF/VF of the VIRTIO_BLK emulation device. Value 0x0 will use device defaults
VIRTIO_BLK_EMULATION_NUM_PF=<NUM> Total number of PCIe functions (PFs) exposed by the device for VIRTIO_BLK emulation.
VIRTIO_BLK_EMULATION_NUM_VF=<NUM> The total number of Virtual Functions (VFs) that can be supported for each PF.
VIRTIO_BLK_EMULATION_NUM_VF_MSIX=<NUM> Number of MSI-X vectors assigned to each VF associated with this type of PF. Value 0x0 indicates device defaults.
VIRTIO_BLK_EMULATION_PF_PCI_LAYOUT=<VIRTIO_1_X|VIRTIO_TRANSITIONAL>Indicates which VIrtIO specification the PCI layout of the emulated Physical Function(s) will follow.
0x0: VIRTIO_1_X
0x1: VIRTIO_TRANSITIONAL
other values are reserved
VIRTIO_BLK_EMULATION_SUBSYSTEM_ID=<NUM> PCIe subsystem_id register for the VIRTIO_BLK emulated device
VIRTIO_BLK_EMULATION_VF_PCI_LAYOUT=<VIRTIO_1_X|VIRTIO_TRANSITIONAL>Indicates which VIrtIO specification the PCI layout of the emulated Virtual Function(s) will follow.
0x0: VIRTIO_1_X
0x1: VIRTIO_TRANSITIONAL
other values are reserved
VIRTIO_BLK_EMU_SUBSYSTEM_VENDOR_ID=<NUM>PCIe subsystem_vendor_id register for the VIRTIO_BLK emulated device
GLOBAL PCI CONF 3:
PCI_BUS0_RESTRICT=<False|True> When TRUE, PCI bus width, speed and ASPM will be restricted according to PCI_BUS_RESTRICT_WIDTH, PCI_BUS_RESTRICT_SPEED, PCI_BUS_RESTRICT_ASPM respectively
PCI_BUS0_RESTRICT_ASPM=<False|True> When FALSE, PCI bus will not have ASPM enabled.
Valid when PCI_BUS_RESTRICT is TRUE.
PCI_BUS0_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>Restricts the PCI speed to be smaller or equal to:
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
Valid when PCI_BUS_RESTRICT is TRUE.
PCI_BUS0_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>Restricts the PCI bus width to be smaller or equal to:
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
Valid when PCI_BUS_RESTRICT is TRUE.
PCI_BUS1_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS1_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS1_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS1_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS2_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS2_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS2_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS2_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS3_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS3_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS3_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS3_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS4_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS4_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS4_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS4_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS5_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS5_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS5_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS5_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS6_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS6_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS6_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS6_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_BUS7_RESTRICT=<False|True> See PCI_BUS0_RESTRICT
PCI_BUS7_RESTRICT_ASPM=<False|True> See PCI_BUS0_RESTRICT_ASPM
PCI_BUS7_RESTRICT_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5>See PCI_BUS0_RESTRICT_SPEED
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
PCI_BUS7_RESTRICT_WIDTH=<PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>See PCI_BUS0_RESTRICT_WIDTH
0x0: PCI_X1
0x1: PCI_X2
0x2: PCI_X4
0x3: PCI_X8
0x4: PCI_X16
PCI_DOWNSTREAM_PORT_OWNER=<DEVICE_DEFAULT|EMBEDDED_CPU|HOST_0|HOST_1|HOST_2|HOST_3|HOST_4|HOST_5|HOST_6|HOST_7>
Identifies the owner of the respective PCI Downstream Port.
0x0: DEVICE_DEFAULT
0x1: HOST_0
0x2: HOST_1
0x3: HOST_2
0x4: HOST_3
0x5: HOST_4
0x6: HOST_5
0x7: HOST_6
0x8: HOST_7
0xF: EMBEDDED_CPU
Other values are reserved
EMULATION VIRTIO FS CONF:
VIRTIO_FS_EMULATION_ENABLE=<False|True> When set to TRUE, VIRTIO_FS device emulation is enabled.
VIRTIO_FS_EMULATION_NUM_MSIX=<NUM> Number of MSI-X Vectors assigned for each PF/VF of the VIRTIO_FS emulation device. Value 0x0 will use device defaults
VIRTIO_FS_EMULATION_NUM_PF=<NUM> Total number of PCIe functions (PFs) exposed by the device for VIRTIO_FS emulation.
VIRTIO_FS_EMULATION_NUM_VF=<NUM> The total number of Virtual Functions (VFs) that can be supported for each PF.
VIRTIO_FS_EMULATION_NUM_VF_MSIX=<NUM> Number of MSI-X vectors assigned to each VF associated with this type of PF. Value 0x0 indicates device defaults.
VIRTIO_FS_EMULATION_SUBSYSTEM_ID=<NUM> PCIe subsystem_id register for the VIRTIO_FS emulated device
VIRTIO_FS_EMU_SUBSYSTEM_VENDOR_ID=<NUM> PCIe subsystem_vendor_id register for the VIRTIO_FS emulated device
GLOBAL PCI CONF 4:
OFF_BOARD_SERIALIZER=<False|True> If TRUE, an off-board Serializer for GPIO implementation is used
OFF_BOARD_SERIALIZER_VALID=<False|True> If TRUE, off_board_serializer field is valid
PCI_BUS00_ASPM=<False|True> When FALSE, PCI bus 00 will not have ASPM enabled.
PCI_BUS00_ENABLE=<False|True> When TRUE, PCI bus 00 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS00_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 00 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS00_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 00 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS00_SWITCH_INDEX=<NUM> Bus 00 Pointer to a PCIe Switch (Switch0, Switch1, Switch2)
Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS00_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 00 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS01_ASPM=<False|True> When FALSE, PCI bus 01 will not have ASPM enabled.
PCI_BUS01_ENABLE=<False|True> When TRUE, PCI bus 01 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS01_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 01 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS01_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 01 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS01_SWITCH_INDEX=<NUM> Bus 01 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS01_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 01 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS02_ASPM=<False|True> When FALSE, PCI bus 02 will not have ASPM enabled.
PCI_BUS02_ENABLE=<False|True> When TRUE, PCI bus 02 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS02_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 02 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS02_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 02 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS02_SWITCH_INDEX=<NUM> Bus 02 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS02_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 02 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS03_ASPM=<False|True> When FALSE, PCI bus 03 will not have ASPM enabled.
PCI_BUS03_ENABLE=<False|True> When TRUE, PCI bus 03 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS03_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 03 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS03_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 03 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS03_SWITCH_INDEX=<NUM> Bus 03 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS03_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 03 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS04_ASPM=<False|True> When FALSE, PCI bus 04 will not have ASPM enabled.
PCI_BUS04_ENABLE=<False|True> When TRUE, PCI bus 04 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS04_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 04 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS04_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 04 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS04_SWITCH_INDEX=<NUM> Bus 04 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS04_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 04 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS05_ASPM=<False|True> When FALSE, PCI bus 05 will not have ASPM enabled.
PCI_BUS05_ENABLE=<False|True> When TRUE, PCI bus 05 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS05_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 05 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS05_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 05 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS05_SWITCH_INDEX=<NUM> Bus 05 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS05_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 05 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS06_ASPM=<False|True> When FALSE, PCI bus 06 will not have ASPM enabled.
PCI_BUS06_ENABLE=<False|True> When TRUE, PCI bus 06 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS06_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 06 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS06_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 06 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS06_SWITCH_INDEX=<NUM> Bus 06 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS06_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 06 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS07_ASPM=<False|True> When FALSE, PCI bus 07 will not have ASPM enabled.
PCI_BUS07_ENABLE=<False|True> When TRUE, PCI bus 07 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS07_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 07 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS07_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 07 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS07_SWITCH_INDEX=<NUM> Bus 07 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS07_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 07 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS10_ASPM=<False|True> When FALSE, PCI bus 10 will not have ASPM enabled.
PCI_BUS10_ENABLE=<False|True> When TRUE, PCI bus 10 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS10_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 10 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS10_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 10 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS10_SWITCH_INDEX=<NUM> Bus 10 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS10_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 10 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_BUS11_ASPM=<False|True> When FALSE, PCI bus 11 will not have ASPM enabled.
PCI_BUS11_ENABLE=<False|True> When TRUE, PCI bus 11 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS11_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 11 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS11_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 11 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS11_SWITCH_INDEX=<NUM> Bus 11 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS11_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 11 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
--- CUT HERE ---
PCI_BUS27_ASPM=<False|True> When FALSE, PCI bus 27 will not have ASPM enabled.
PCI_BUS27_ENABLE=<False|True> When TRUE, PCI bus 27 parameters are enforced, otherwise the bus is not implemented and all other bus parameters are reserved
PCI_BUS27_HIERARCHY_TYPE=<PCIE_ENDPOINT|PCIE_EXTERNAL_HOST_SWITCH|PCIE_INTERNAL_HOST_SWITCH>Defines PCI Bus 27 connectivity hierarchy
0x0: PCIE_ENDPOINT - PCI link connected to external host, Device is EP
0x1: PCIE_EXTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to external host
0x2: PCIE_INTERNAL_HOST_SWITCH - PCI links connected to an internal switch, upstream link connected to internal host
PCI_BUS27_SPEED=<PCI_GEN_1|PCI_GEN_2|PCI_GEN_3|PCI_GEN_4|PCI_GEN_5|PCI_GEN_6>PCI bus 27 speed
0x0: PCI_GEN_1
0x1: PCI_GEN_2
0x2: PCI_GEN_3
0x3: PCI_GEN_4
0x4: PCI_GEN_5
0x5: PCI_GEN_6
PCI_BUS27_SWITCH_INDEX=<NUM> Bus 27 Pointer to a Pcie Switch (switch0, switch1, switch2) Valid only for Hierarchy Type PCIE_HIER_TYPE_EXTERNAL_HOST_SWITCH
PCI_BUS27_WIDTH=<PCI_INACTIVE|PCI_X1|PCI_X16|PCI_X2|PCI_X4|PCI_X8>PCI bus 27 width:
0x0: PCI_INACTIVE
0x1: PCI_X1
0x2: PCI_X2
0x3: PCI_X4
0x4: PCI_X8
0x5: PCI_X16
PCI_SWITCH0_ENABLE=<False|True> When TRUE PCI Switch0 is enabled, otherwise it is disabled
PCI_SWITCH0_UPSTRAEM_PORT_BUS=<NUM> Index of the bus (i.e. - the 1st logic lane in the PEX of the respective bus) on which the upstream port belonging to Switch0 resides
PCI_SWITCH0_UPSTRAEM_PORT_PEX=<NUM> The index of the PEX on which the upstream port belonging to Switch0 resides
PCI_SWITCH1_ENABLE=<False|True> When TRUE PCI Switch1 is enabled, otherwise it is disabled
PCI_SWITCH1_UPSTRAEM_PORT_BUS=<NUM> Index of the bus (i.e. - the 1st logic lane in the PEX of the respective bus) on which the upstream port belonging to Switch1 resides
PCI_SWITCH1_UPSTRAEM_PORT_PEX=<NUM> The index of the PEX on which the upstream port belonging to Switch1 resides
PCI_SWITCH2_ENABLE=<False|True> When TRUE PCI Switch2 is enabled, otherwise it is disabled
PCI_SWITCH2_UPSTRAEM_PORT_BUS=<NUM> Index of the bus (i.e. - the 1st logic lane in the PEX of the respective bus) on which the upstream port belonging to Switch2 resides
PCI_SWITCH2_UPSTRAEM_PORT_PEX=<NUM> The index of the PEX on which the upstream port belonging to Switch2 resides
PCORE0_REF_CLK=<PCI_BUS_TABLE|SINGLE_REF_CLK>PCORE0 Reference clock configuration:
0x0: SINGLE_REF_CLK - Single Ref Clock for PCOREx, rck_buf (BINVHS)
0x1: PCI_BUS_TABLE - according to the pre-determined clocks configuration for the given PCIe buses as defined by PCIe Buses Table
PCORE0_REF_CLK_VALID=<False|True> When TRUE, pcore0_refclk field is valid
PCORE0_REVERSAL=<False|True> If TRUE, PCORE0 shall use Lane Reversal. Note: This field must be coherent with HW straps for Primary PCORE
PCORE0_REVERSAL_VALID=<False|True> If TRUE, pcore0_reversal field is valid
PCORE1_REF_CLK=<PCI_BUS_TABLE|SINGLE_REF_CLK>PCORE1 Reference clock configuration:
0x0: SINGLE_REF_CLK - Single Ref Clock for PCOREx, rck_buf (BINVHS)
0x1: PCI_BUS_TABLE - according to the pre-determined clocks configuration for the given PCIe buses as defined by PCIe Buses Table
PCORE1_REF_CLK_VALID=<False|True> When TRUE, pcore1_refclk field is valid
PCORE1_REVERSAL=<False|True> If TRUE, PCORE1 shall use Lane Reversal. Note: This field must be coherent with HW straps for Primary PCORE
PCORE1_REVERSAL_VALID=<False|True> If TRUE, pcore1_reversal field is valid
PCORE2_REF_CLK=<PCI_BUS_TABLE|SINGLE_REF_CLK>PCORE2 Reference clock configuration:
0x0: SINGLE_REF_CLK - Single Ref Clock for PCOREx, rck_buf (BINVHS)
0x1: PCI_BUS_TABLE - according to the pre-determined clocks configuration for the given PCIe buses as defined by PCIe Buses Table
PCORE2_REF_CLK_VALID=<False|True> When TRUE, pcore2_refclk field is valid
PCORE2_REVERSAL=<False|True> If TRUE, PCORE2 shall use Lane Reversal. Note: This field must be coherent with HW straps for Primary PCORE
PCORE2_REVERSAL_VALID=<False|True> If TRUE, pcore2_reversal field is valid
STRAP_SD_OR_MH=<False|True> If TRUE, each bus is associated with a unique host (Multi-Host), otherwise it is a Socket-Direct confguration
EXTERNAL HOST PRIV FAST CONF:
HOST_PRIV_RSHIM=<DEVICE_DEFAULT|DISABLE|ENABLE>Enforce state of RSHIM PF towards external host
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
GLOBAL PCI GPIO CONF:
ARM_RC_PERST_GPIO=<NUM> Indicates the number of GPIO used for downstream PERST#, common for all DSP ports belonging to ARM Root Complex
ARM_RC_PERST_GPIO_VALID=<False|True> When TRUE, arm_rc_perst_gpio field is valid.
Shall not be enabled unless PCI_SWITCH0_ENABLE is TRUE
CPLD_SERIALIZER_CFG_VALID=<False|True> When TRUE, cpld_serializer_chain_pcie_offset field is valid. Applicable when an off-card CPLD is used.
CPLD_SERIALIZER_CHAIN_PCIE_OFFSET=<NUM> The offset of the Serializer where the PCIe control signals start
PCI_BUS00_CONTROL_EN=<False|True> If set, the location of the control signals for bus00 is determined by the off-board Serializer with offset CPLD_SERIALIZER_CHAIN_PCIE_OFFSET
Shall not be set unless the corresponding PCI_BUSxy_ configurations (i.e. PCI_BUSxy_WIDTH, PCI_BUSxy_SPEED, PCI_BUSxy_SWITCH_INDEX, PCI_BUSxy_NIERARCHY_TYPE) are set
PCI_BUS00_CPLD_GPIO_MASK=<NUM> The mask for the usage of the signals in the respective CPLD Slot, as indicated by bus00_device_type and Bus00_serializer_index
PCI_BUS00_DEVICE_TYPE=<GENERIC_PCIE_DEVICE|NVME_SSD_DEVICE>Device type connected to bus00
0x0: GENERIC_PCIE_DEVICE
0x1: NVME_SSD_DEVICE
Other values are reserved
PCI_BUS00_PERST_SOURCE=<DEFAULT_SRC|OFF_CHIP|ON_CHIP_GPIO|SERIALIZER>Selector of the Source of the PERST# signal for bus00
0x0: DEFAULT_SRC
0x1: SERIALIZER - PERST# through Serializer
0x2: ON_CHIP_GPIO - PERST# by on-chip GPIO, defined by SwitchX_PERST#_GPIO (x=0-2) or ARM_RC_PERST#_GPIO respectively
0x3: OFF_CHIP - PERST# by direct off-chip connection ("Bypass")
PCI_BUS00_SERIALIZER_INDEX=<NUM> The index of the Serializer slot for bus00.
For bus00_device_type GENERIC_PCIE_DEVICE - the index of the PCIe GPIOs Slot.
For bus00_device_type NVME_SSD_DEVICE - the index of the SSD Slot
-- CUT HERE --
PCI_BUS27_CONTROL_EN=<False|True> See PCI_BUS00_CONTROL_EN
PCI_BUS27_CPLD_GPIO_MASK=<NUM> See PCI_BUS00_CPLD_GPIO_MASK
PCI_BUS27_DEVICE_TYPE=<GENERIC_PCIE_DEVICE|NVME_SSD_DEVICE>Device type connected to bus27
0x0: GENERIC_PCIE_DEVICE
0x1: NVME_SSD_DEVICE
Other values are reserved
PCI_BUS27_PERST_SOURCE=<DEFAULT_SRC|OFF_CHIP|ON_CHIP_GPIO|SERIALIZER>Selector of the Source of the PERST# signal for bus27
0x0: DEFAULT_SRC
0x1: SERIALIZER - PERST# through Serializer
0x2: ON_CHIP_GPIO - PERST# by on-chip GPIO, defined by SwitchX_PERST#_GPIO (x=0-2) or ARM_RC_PERST#_GPIO respectively
0x3: OFF_CHIP - PERST# by direct off-chip connection ("Bypass")
PCI_BUS27_SERIALIZER_INDEX=<NUM> See PCI_BUS00_SERIALIZER_INDEX
SWITCH0_PERST_GPIO=<NUM> Indicates the number of GPIO used for downstream PERST#, common for all DSP ports belonging to Switch0.
SWITCH0_PERST_GPIO_VALID=<False|True> When TRUE, switch0_perst_gpio field is valid.
Shall not be enabled unless PCI_SWITCH0_ENABLE is TRUE
SWITCH1_PERST_GPIO=<NUM> Indicates the number of GPIO used for downstream PERST#, common for all DSP ports belonging to Switch1.
SWITCH1_PERST_GPIO_VALID=<False|True> When TRUE, switch1_perst_gpio field is valid.
Shall not be enabled unless PCI_SWITCH1_ENABLE is TRUE
SWITCH2_PERST_GPIO=<NUM> Indicates the number of GPIO used for downstream PERST#, common for all DSP ports belonging to Switch2.
SWITCH2_PERST_GPIO_VALID=<False|True> When TRUE, switch2_perst_gpio field is valid.
Shall not be enabled unless PCI_SWITCH2_ENABLE is TRUE
SW OFFLOAD CONF:
CQE_COMPRESSION=<AGGRESSIVE|BALANCED> Configure which algorithm should be used by the NIC in order to decide when to activate CQE compression based on PCIe bus condition. Note that the driver can enable compression on a per CQE basis
0x0: BALANCED
0x1: AGGRESSIVE
ESWITCH_IPV4_TTL_MODIFY_ENABLE=<False|True>When TRUE, the device will supports e-switch rules modifying TTL of packets from the uplink to a vport.
IP_OVER_VXLAN_EN=<False|True> This parameter is only supported if NV_SW_OFFLOAD_CAP.ip_over_vxlan_supported is set.
mlxconfig_desc: When set, non-standard IPoVxLAN offload is enabled for incoming UDP port specified in IP_OVER_VXLAN_PORT.
IP_OVER_VXLAN_PORT=<NUM> This parameter is only supported if NV_SW_OFFLOAD_CAP.ip_over_vxlan_supported is set.
mlxconfig_desc: The UDP port for incoming IPoVxLAN traffic (non- standard).
minval:1
LOG_MAX_OUTSTANDING_WQE=<NUM> Log2 of the maximal amount of outstanding (uncompleted) WQEs a single Transmit Work Queue may hold. Additional WQEs will be delayed until some of the outstanding WQEs are completed.
Value 0x0 indicates device default.
Value 0x1F indicates device maximal supported value
LOG_TX_PSN_WINDOW=<NUM> Log2 of the transmission PSN window for QPs with Connected Transport Service (RC, DC, XRC).
Value 0x0 indicates device defaults
Value 0x3F indicates device maximal supported value
Values exceeding the device capabilities will be considered as the device maximal supported value as well.
LRO_LOG_TIMEOUT0=<NUM> Log2 of Large Receive Offload (LRO) timeout #0, in microseconds. Driver can select one of the 4 configured LRO timeouts on a per Qp basis in run-time (lro_timeout_period_usecs field of the TIR context)
LRO_LOG_TIMEOUT1=<NUM> Log2 of Large Receive Offload (LRO) timeout #1, in microseconds. Driver can select one of the 4 configured LRO timeouts on a per Qp basis in run-time (lro_timeout_period_usecs field of the TIR context)
LRO_LOG_TIMEOUT2=<NUM> Log2 of Large Receive Offload (LRO) timeout #2, in microseconds. Driver can select one of the 4 configured LRO timeouts on a per Qp basis in run-time (lro_timeout_period_usecs field of the TIR context)
LRO_LOG_TIMEOUT3=<NUM> Log2 of Large Receive Offload (LRO) timeout #3, in microseconds. Driver can select one of the 4 configured LRO timeouts on a per Qp basis in run-time (lro_timeout_period_usecs field of the TIR context)
MKEY_BY_NAME=<False|True> When TRUE, the device supports allocating MKey numbers by the device driver
PCI_ATOMIC_MODE=<PCI_ATOMICS_ENABLED_EXT_ATOMICS_DISABLED|PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED|PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED_NONCOHERENT|PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED_SERIALIZED|PCI_ATOMIC_DISABLED_EXT_ATOMIC_ENABLED>PCI Atomic mode.
0x0: PCI_ATOMIC_DISABLED_EXT_ATOMIC_ENABLED
0x1: PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED_SERIALIZED - Atomic serialization will guarantee atomicity for multiple accesses to the same address
0x2: PCI_ATOMICS_ENABLED_EXT_ATOMICS_DISABLED
0x3: PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED_NONCOHERENT
0x4: PCI_ATOMICS_ENABLED_EXT_ATOMICS_ENABLED
PRIO_TAG_REQUIRED_EN=<False|True> When set to TRUE, the system requires to always have prio tag on the network. SW of e-Switch manager should set the required flow table rules to ensure that. This field is supported only when NV_SW_OFFLOAD_CAP.prio_tag_required==1 and is reported in HCA_CAP.prio_tag_required.
RDMA_SELECTIVE_REPEAT_EN=<False|True> When TRUE, Selective Repeat for RDMA QPs is supported.
REAL_TIME_CLOCK_ENABLE=<False|True> When enabled, real time clock is enabled on the device, allowing timestamps presented in real time instead of cycles.
ROCE_ADAPTIVE_ROUTING_EN=<False|True> When TRUE, Adaptive Routing for RDMA QPs is supported.
TUNNEL_ECN_COPY_DISABLE=<False|True> When TRUE, ECN field copy from inner to outer header during encapsulation is disabled.
TUNNEL_IP_PROTO_ENTROPY_DISABLE=<False|True>When TRUE, tunnel encapsulation offload will exclude IP.protocol from the UDP source port entropy calculation.
UCTX_EN=<False|True> When set to TRUE, the device supports creation of UCTX contexts.
VECTOR_CALC_DISABLE=<False|True> When TRUE, Vector-Calc operations are disabled
VF_MIGRATION_MODE=<DEVICE_DEFAULT|MIGRATION_DISABLED|MIGRATION_ENABLED>Defines support for VF migration
0x0: DEVICE_DEFAULT
0x1: MIGRATION_DISABLED
0x2: MIGRATION_ENABLED
Other values are reserved
PROG CC:
USER_PROGRAMMABLE_CC=<False|True> When set to TRUE, loading user programmed CC image is enabled. When FALSE, such image cannot be loaded. If an image already exists when this configuration is changed, the image is ignored. Valid only if ROCE_CC_LEGACY_DCQCN is FALSE.
PCC INT:
PCC_HANDLE_CORE_UTIL=<DEVICE_DEFAULT|DISABLED|HIGH|LOW|MEDIUM>Defines PCC handling core utilization
0x0: DEVICE_DEFAULT
0x1: DISABLED - no allocated cores, PCC disabled
0x2: LOW - low core utilization
0x3: MEDIUM - medium core utilization
0x4: HIGH - high core utilization
other values are reserved
PCC_INT_EN=<False|True> When set to TRUE, the device will support PCC CC based on INT stamped RTT packets. The PCC algorithm should be programmed/configured separately to this configuration.
PCC_INT_NP_RTT_DATA_MODE=<HPCC_LAST_HOP|HPCC_MOST_CONGESTED|INGRESS_BYTE|NO_DATA|NV_INT_0|RTT_V0>Defines how Notification Point (NP) generates the data for the RTT response, out of the telemetry (INT) information.
0x00: HPCC_LAST_HOP - Use the last telemetry entry in the packet Alibaba only
0x10: HPCC_MOST_CONGESTED - Use the telemetry entry indicating the most congested telemetry point Alibaba only
0x20: NV_INT_0 - Nvidia Telemetry
0x40: RTT_V0
0x48: NO_DATA - (RTT_V1)
0x49: INGRESS_BYTE - ingress port byte rate (RTT_v1)
Other values are reserved
PCC_INT_NP_RTT_DSCP=<NUM> Defines how DSCP is used in RTT response.
PCC_INT_NP_RTT_DSCP_EN=<False|True> Defines how DSCP is used in RTT response.
PCC_INT_SYSTEM_RTT=<NUM> Modifies RTT evaluation.
PERFORMANCE TUNING CONF:
HAIRPIN_DATA_BUFFER_LOCK=<False|True> When TRUE, using locked internal memory for hairpin RQ data buffer is enabled
HAIRPIN_LOG_WQE_NUM=<NUM> Log base 2 of the number of WQEs in a hairpin queues. Value 0x0 indicates device default. The implicit data per WQE is the total size (defined by SWITCH_HAIRPIN_TOT_BUFFER_SIZE) divided by the number of WQEs. This value should exceed the expected MTU.
ICM_CACHE_MODE=<DEVICE_DEFAULT|LARGE_SCALE_STEERING>Tuning ICM caching mode for different setups
0x0: DEVICE_DEFAULT
0x1: LARGE_SCALE_STEERING - very large scale packet classification logic, with small scale for other objects
All other values are reserved
MULTI_PCI_RESOURCE_SHARING=<BLOCKING_SHARED|DEVICE_DEFAULT|NON_BLOCK_FAIR|NON_BLOCK_SHARED>Defines how resources are distributed for devices configured for multi-pci interfaces (e.g. Multi-Host, Socket Direct)
0x0: DEVICE_DEFAULT
0x1: NON_BLOCK_FAIR - resources are fairly assigned to entities
0x2: NON_BLOCK_SHARED - resources are partially assigned and partially shared between entities
0x3: BLOCKING_SHARED - resources are shared between entities
other values are reserved
STEERING_CACHE_REFRESH=<NUM> Log (base 2) of the period of Steering cache refreshments. Each unit represents 25us. Value 0x0 indicates device default.
TLS_OPTIMIZE=<False|True> When TRUE, the device will optimize processing for TLS operations.
When FALSE, device defaults will apply for TLS processing.
TX_SCHEDULER_BURST=<NUM> Log (base2) of the transmission scheduler default burst size, given in bytes, Value 0x0 indicates using device defaults.
TX_SCHEDULER_FWS_REACTIVITY=<ALGORITHMIC|DEVICE_DEFAULT|DIRECT>Defines transmission scheduler reactivity to full utilization of the available BW
0x0: DEVICE_DEFAULT
0x1: DIRECT
0x2: ALGORITHMIC
other values are reserved
TX_SCHEDULER_LOCALITY_MODE=<ACCUMULATIVE|DEVICE_DEFAULT|STATIC_MODE>Defines transmission scheduler adaptation to locality.
0x0: DEVICE_DEFAULT
0x1: STATIC_MODE
0x2: ACCUMULATIVE
Other values are reserved
ZERO_TOUCH_TUNING_ENABLE=<False|True> when TRUE, Zero Touch Tuning routine will automatically tune device registers in order to achieve the maximum performance
GLOBAL ROCE CC CONF:
ROCE_CC_DCQCN_COMPATIBILITY_MODE=<DEVICE_DEFAULT|GEN5|GEN6>Sets the DCQCN CC algorithm to work in the same manner other device generations.
0x0: DEVICE_DEFAULT
0x1: GEN5
0x2: GEN6
ROCE_CC_LEGACY_DCQCN=<False|True> When TRUE, the device will only use legacy Congestion Control DCQCN algorithm
ROCE_CC_LEGACY_DCQCN_SW=<False|True> When TRUE, the device will only use legacy SW Congestion Control DCQCN algorithm. Applicable for Ethernet ports only.
ROCE_CC_STEERING_EXT=<DEVICE_DEFAULT|DISABLED|ENABLED>Defines PCC extension for packet steering
0x0: DEVICE_DEFAULT
0x1: DISABLED
0x2: ENABLED
Other values are reserved
NVMF TARGET CONF:
NVMF_PACER_MODE=<DEVICE_DEFAULT|PACER_OFF|PACER_ON>Control IOP scheduling pacer
0x0: DEVICE_DEFAULT
0x1: PACER_OFF
0x2: PACER_ON
NVMF_PACER_THRESHOLD_READ=<NUM> Log (base 2) of IOP pacer low threshold for READ operations size. Smaller or equal operations will bypass the pacer. Value given in Bytes. Valid only when NVMF_PACER_MODE is PACER_ON.
NVMF_PACER_THRESHOLD_WRITE=<NUM> Log (base 2) of IOP pacer low threshold for WRITE operations size. Smaller or equal operations will bypass the pacer. Value given in Bytes. Valid only when NVMF_PACER_MODE is PACER_ON.
SW ACCELERATE CONF:
DBR_LESS_SQP=<DEVICE_DEFAULT|DISABLE|ENABLE>Control SQ/QP without Doorbell-Record (for send operations)
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
other values are reserved
LOG_MAX_OUTSTANDING_READ_ATOMIC=<NUM> Maximum number of uncompleted read atomic operations that can be handled in parallel by a QP, expressed in log2 scale.
This parameter may affect ICM memory consumption. Value 0x0 Sets the configuration to the device default.
LOG_MAX_QUEUE=<NUM> Log (base 2) of the upper limit the number of supported work queues in a single namespace (e.g. PCI function). Value 0x0 indicates using device defaults. unsupported values are saturated to the nearest supported value.
MULTIPATH_DSCP=<DEVICE_DEFAULT|DISABLE|DSCP_0|DSCP_1|DSCP_2>Multipath on transmit, set the DSCP bit to hold the MP eligible info.
0x0: DEVICE_DEFAULT
0x1: DISABLE
0x2: DSCP_0 - DSCP[0]
0x3: DSCP_1 - DSCP[1]
0x4: DSCP_2 - DSCP[2]
other values are reserved
SWP_L4_CHECKSUM_MODE=<DEVICE_DEFAULT|FULL_CSUM|L4_ONLY>Defines how L4 checksum is calculated by the device when using hints for header locations (SWP):
0x0: DEVICE_DEFAULT
0x1: FULL_CSUM
0x2: L4_ONLY - pseudo header is excluded
other values are reserved
EMULATION UPT CONF:
UPT_EMULATION_ENABLE=<False|True> When set to TRUE, UPT device emulation is enabled.
CRYPTO CONF:
AES_XTS_TWEAK_INC_64=<False|True> Enables AES-XTS tweak increment by (1<<64) with every block (non AES-XTS standard).
CRYPTO_POLICY=<DEVICE_DEFAULT|FIPS_LEVEL_2|UNRESTRICTED>Device cryptographic policy
0x0: DEVICE_DEFAULT
0x1: UNRESTRICTED
0x2: FIPS_LEVEL_2
other values are reserved
LARGE_MTU_TWEAK_64=<False|True> Enables large mtu tweak 64 [internal] supports 8K MTU on force LB QPs.
HOST MNG CONF:
PCIE_IN_BAND_VDM_DISABLE=<False|True> When TRUE, the management processor will disable PCIe in-band VDM (MCTP over PCIe) interface.
PCIE_SMBUS_DISABLE=<False|True> When TRUE, the management processor will disable SMBUS (embedded on the PCIe connector) interface.
PLDM_FW_UPDATE_DISABLE=<False|True> When TRUE, PLDM FW update is disabled
RBT_DISABLE=<False|True> When TRUE, the management processor will disable RBT.
RDE_DISABLE=<False|True> When TRUE, RDE is disabled
HOST SM CONF:
SM_DISABLE=<False|True> When TRUE, the host is not allowed to serve as an SM/SMA and/or transmit SMP packets.
EXTERNAL HOST PRIV CONF:
HOST_PRIV_FLASH_ACCESS=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to perform any device flash access
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_FW_UPDATE=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to perform FW update
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_NIC_RESET=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to perform NIC Reset
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_NV_GLOBAL=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to access global NV parameters.
Note that this field has higher priority over other fields that affect global TLVs.SW should make sure other global fields in this TLV do not conflict with this one (when not DEFAULT).
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_NV_HOST=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to access host NV parameters
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_NV_INTERNAL_CPU=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to access NV_INTERNAL_CPU_CONF NV TLV
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_NV_PORT=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to access port NV parameters
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
HOST_PRIV_PCC_UPDATE=<DEVICE_DEFAULT|DISABLE|ENABLE>Defines Host privilege to perform PCC algorithm FW image update
0x0: DEVICE_DEFAULT
0x1: ENABLE
0x2: DISABLE
ADVANCED TESTABILITY:
ADVANCED_TESTABILITY=<False|True> When set, MPEINJ register is enabled.
IB DC CONF:
DCR_LIFO_SIZE=<NUM> The total amount of InfiniBand DCRs (Dynamically Connected Responders), minus one.
LOG_DCR_HASH_TABLE_SIZE=<NUM> InfiniBand DC (Dynamically Connected) hash table log2 size (for example: 7 means 128 entries)
MAX_PACKET_LIFETIME=<NUM> Maximum time for a packet to traverse any of the IB/RoCE paths. Time in microseconds is derived from 4.096 usec * 2^MAX_PACKET_LIFETIME.
Value 0x0 indicates device default
LS NV PUBLIC KEY 0:
LC_NV_PUB_KEY_EXP=<NUM> The public key exponent.
In most cases, the standard exponent (65537) will be used (per RFC4871).
LC_NV_PUB_KEY_UUID=<BYTES> UUID of this key.
The UUID is created by the sign server when it generates a new RSA key-pair.
LS NV PUBLIC KEY 1:
LC_NV_PUB_KEY_0_255=<BYTES> first half of 4096 bit public-key (bytes 0 to 255)
LS NV PUBLIC KEY 2:
LC_NV_PUB_KEY_256_511=<BYTES> Second half of 4096 bit public-key (bytes 256 to 511)
VPI LINK TYPE:
LINK_TYPE_P1=<ETH|IB> Select the link type (Ethernet or Infiniband) for the port
0x1: IB - Infiniband
0x2: ETH - Ethernet
LINK_TYPE_P2=<ETH|IB>
ROCE CC:
IB_CC_SHAPER_COALESCE_P1=<DEVICE_DEFAULT|REMOTE_LID>Select CC algorithm shaper coalescing for IB
0x0: DEVICE_DEFAULT
0x2: REMOTE_LID - shaper is selected according to remote lid (IB)
other values are reserved
IB_CC_SHAPER_COALESCE_P2=<DEVICE_DEFAULT|REMOTE_LID>
ROCE_CC_ALGORITHM_P1=<ECN|QCN> Select RDMA over Converged Ethernet (RoCE) algorithm
0x0: ECN
0x1: QCN
ROCE_CC_ALGORITHM_P2=<ECN|QCN>
ROCE_CC_CNP_MODERATION_P1=<DEVICE_DEFAULT|PER_FLOW|PER_PORT>Select CNP moderation mode used by the DCQCN algorithm.
0x0: DEVICE_DEFAULT
0x1: PER_PORT
0x2: PER_FLOW
other values are reserved
ROCE_CC_CNP_MODERATION_P2=<DEVICE_DEFAULT|PER_FLOW|PER_PORT>
ROCE_CC_PRIO_MASK_P1=<NUM> Each bit in this mask indicates if the RoCE should be enabled on the n-th IEEE priority.
ROCE_CC_PRIO_MASK_P2=<NUM>
ROCE_CC_SHAPER_COALESCE_P1=<DEST_IP|DEVICE_DEFAULT|SOURCE_QP|_5_TUPLE>Select CC algorithm shaper coalescing for ROCE
0x0: DEVICE_DEFAULT
0x1: DEST_IP - shaper is selected according to dest IP
0x2: SOURCE_QP - shaper is selected according to source QP
0x3: _5_TUPLE - shaper is selected according to 5-tuples
other values are reserved
ROCE_CC_SHAPER_COALESCE_P2=<DEST_IP|DEVICE_DEFAULT|SOURCE_QP|_5_TUPLE>
ROCE CC ECN:
CLAMP_TGT_RATE_AFTER_TIME_INC_P1=<False|True>When receiving a CNP, the target rate should be updated if the transmission rate was increased due to the timer, and not only due to the byte counter
CLAMP_TGT_RATE_AFTER_TIME_INC_P2=<False|True>
CLAMP_TGT_RATE_P1=<False|True> If set, whenever a CNP is processed, the target rate is updated to be the current rate.
CLAMP_TGT_RATE_P2=<False|True>
CNP_802P_PRIO_P1=<NUM> The 802.1p priority value of the generated CNP for this port
CNP_802P_PRIO_P2=<NUM>
CNP_DSCP_P1=<NUM> The DiffServ Code Point of the generated CNP for this port.
CNP_DSCP_P2=<NUM>
CNP_RES_PRIO_MODE_P1=<False|True> If TRUE, CNP packets for this port contain priority from a received request. If FALSE, CNP responses use value set by CNP_802P_PRIO.
CNP_RES_PRIO_MODE_P2=<False|True>
DCE_TCP_G_P1=<NUM> Used to update the congestion estimator (alpha) once every dce_tcp_rtt microseconds, according to the equation:
Alpha = (cnp_received * dceTcpG) + (1 - dceTcpG) * alpha .
dceTcpG is divided by 2^10.
cnp_received is set to one if a CNP was received for this flow during period since the previous update and the current update
DCE_TCP_G_P2=<NUM>
DCE_TCP_RTT_P1=<NUM> The time between updates of the alpha value, in microseconds.
DCE_TCP_RTT_P2=<NUM>
INITIAL_ALPHA_VALUE_P1=<NUM> The initial value of alpha to use when receiving the first CNP for a flow. Expressed in a fixed point fraction of 2^10.
INITIAL_ALPHA_VALUE_P2=<NUM>
MIN_TIME_BETWEEN_CNPS_P1=<NUM> Minimum time between sending CNPs from the port, in microseconds.
MIN_TIME_BETWEEN_CNPS_P2=<NUM>
RATE_REDUCE_MONITOR_PERIOD_P1=<NUM> The minimum time between 2 consecutive rate reductions for a single flow. Rate reduction will occur only if a CNP is received during the relevant time interval.
RATE_REDUCE_MONITOR_PERIOD_P2=<NUM>
RATE_TO_SET_ON_FIRST_CNP_P1=<NUM> The rate that is set for the flow when a rate limiter is allocated to it upon first CNP received, in Mbps (=Full Port Speed).
RATE_TO_SET_ON_FIRST_CNP_P2=<NUM>
RPG_AI_RATE_P1=<NUM> The rate, in megabits per second, used to increase rpTargetRate in the RPR_ACTIVE_INCREASE.
RPG_AI_RATE_P2=<NUM>
RPG_BYTE_RESET_P1=<NUM> Transmitted data between rate increases if no CNPs are received. Given in Bytes (0=DISABLED)
RPG_BYTE_RESET_P2=<NUM>
RPG_GD_P1=<NUM> If a CNP is received, the flow rate is reduced at the beginning of the next rate_reduce_monitor_period interval to (1- Alpha/Gd)*CurrentRate. rpg_gd is given as log2(Gd), where Gd may only be powers of 2.
RPG_GD_P2=<NUM>
RPG_HAI_RATE_P1=<NUM> The rate, in megabits per second, used to increase rpTargetRate in the RPR_HYPER_INCREASE state.
RPG_HAI_RATE_P2=<NUM>
RPG_MAX_RATE_P1=<NUM> The maximum rate, in Mbits per second, at which an RP can transmit. Once this limit is reached, the RP rate limited is released and the flow is not rate limited any more (0=Full Port Speed).
RPG_MAX_RATE_P2=<NUM>
RPG_MIN_DEC_FAC_P1=<NUM> The minimum factor by which the current transmit rate can be changed when processing a CNP. Value is given as a percentage (1-100).
RPG_MIN_DEC_FAC_P2=<NUM>
RPG_MIN_RATE_P1=<NUM> The minimum value, in megabits per second, for rate to limit.
RPG_MIN_RATE_P2=<NUM>
RPG_THRESHOLD_P1=<NUM> The number of times rpByteStage or rpTimeStage can count before the RP rate control state machine advances states.
RPG_THRESHOLD_P2=<NUM>
RPG_TIME_RESET_P1=<NUM> Time between rate increases if no CNPs are received. Given in microseconds.
RPG_TIME_RESET_P2=<NUM>
LLDP NB CONF:
LLDP_NB_DCBX_P1=<False|True> Enables DCBX (applicable when LLDP_NB_TX_MODE and LLDP_NB_RX_MODE are in ALL mode).
LLDP_NB_DCBX_P2=<False|True>
LLDP_NB_RX_MODE_P1=<ALL|MANDATORY|OFF> Enable the internal LLDP client, and define which TLV it will process.
0x0: OFF - Doesnot listen to incoming LLDP BPDU (incoming LLDP frames will be routed to the host)
0x1: MANDATORY - Listen to incoming LLDP frames, store only the mandatory LLDP BPDUs (1..3)
0x2: ALL - Receive and store all incoming LLDP BPDUs
LLDP_NB_RX_MODE_P2=<ALL|MANDATORY|OFF>
LLDP_NB_TX_MODE_P1=<ALL|MANDATORY|OFF> Select which LLDP TLV will be generated by the NIC
0x0: OFF - NIC internal LLDP client will not send LLDP frames
0x1: MANDATORY - Transmits only mandatory LLDP BPDU (ChassisID, PortID & TTL)
0x2: ALL - Transmits optional LLDP BPDU if configured
LLDP_NB_TX_MODE_P2=<ALL|MANDATORY|OFF>
ROCE CONF:
ROCE_RTT_RESP_DSCP_MODE_P1=<DEVICE_DEFAULT|FIXED_VALUE|RTT_REQUEST>Defines the method for setting IP.DSCP in RTT response packets
0x0: DEVICE_DEFAULT
0x1: FIXED_VALUE - taken from ROCE_RTT_RESP_DSCP
0x2: RTT_REQUEST - taken from the RTT request
other values are reserved
ROCE_RTT_RESP_DSCP_MODE_P2=<DEVICE_DEFAULT|FIXED_VALUE|RTT_REQUEST>
ROCE_RTT_RESP_DSCP_P1=<NUM> The DiffServ Code Point of the generated RTT response for this port. If not set, RTT request value will be used. Overrides PCC_INT_NP_RTT_DSCP
ROCE_RTT_RESP_DSCP_P2=<NUM>
LLDP NB DCBX:
DCBX_CEE_P1=<False|True> Enable DCBX in CEE mode
DCBX_CEE_P2=<False|True>
DCBX_IEEE_P1=<False|True> Enable DCBX in IEEE mode
DCBX_IEEE_P2=<False|True>
DCBX_WILLING_P1=<False|True> Allow the NIC to accept DCBX configuration from the remote peer.
DCBX_WILLING_P2=<False|True>
KEEP LINK UP:
AUTO_POWER_SAVE_LINK_DOWN_P1=<False|True>When set to TRUE, the port will automatically power down when it senses the link is disconnected.
AUTO_POWER_SAVE_LINK_DOWN_P2=<False|True>
DO_NOT_CLEAR_PORT_STATS_P1=<False|True> When set, the port statistic counters are not cleared on first host init. When cleared the port statistics are cleared on first host init
DO_NOT_CLEAR_PORT_STATS_P2=<False|True>
KEEP_ETH_LINK_UP_P1=<False|True> When set, the NIC keeps the link up as long as the server is not in standby mode (Ethernet only).
KEEP_ETH_LINK_UP_P2=<False|True>
KEEP_IB_LINK_UP_P1=<False|True> When set, the NIC keeps the link up as long as the server is not in standby mode (InfiniBand only).
KEEP_IB_LINK_UP_P2=<False|True>
KEEP_LINK_UP_ON_BOOT_P1=<False|True> When set, the NIC keeps the link up as long as the server is not in standby mode and a driver is not initialized.
KEEP_LINK_UP_ON_BOOT_P2=<False|True>
KEEP_LINK_UP_ON_STANDBY_P1=<False|True> When set, the NIC keeps the link up from power-up until the server is turned on, and when the server is in standby
mode (PCI L2/3 state).
Note: The link will be kept up only if the server can provide enough power during standby (via PCI rails or AUX power).
KEEP_LINK_UP_ON_STANDBY_P2=<False|True>
QOS CONF:
ETS_SCHED_MODE_P1=<device_default|per_host|per_physical_port>Select the ETS scheduler mode:
0x0: device_default
0x1: per_physical_port
0x2: per_host
ETS_SCHED_MODE_P2=<device_default|per_host|per_physical_port>
NUM_OF_PFC_P1=<NUM> Number of IEEE priorities that may simultaneously support flow control.
(See pfc_cap_admin DCBX_PARAM Register)
NUM_OF_PFC_P2=<NUM>
NUM_OF_TC_P1=<_1_TC|_2_TCs|_3_TCs|_4_TCs|_5_TCs|_6_TCs|_7_TCs|_8_TCs>Number of traffic classes, when DCB-X is enabled, this is the maximum number of TC that can negotiate with the remote peer.
0x0: _8_TCs
0x1: _1_TC
0x2: _2_TCs
0x3: _3_TCs
0x4: _4_TCs
0x5: _5_TCs
0x6: _6_TCs
0x7: _7_TCs
NUM_OF_TC_P2=<_1_TC|_2_TCs|_3_TCs|_4_TCs|_5_TCs|_6_TCs|_7_TCs|_8_TCs>
NUM_OF_VL_P1=<_15_VLs|_1_VL|_2_VLs|_4_VLs|_8_VLs>Number of Infiniband Virtual Lanes for this port
0x1: _1_VL - VL0 only
0x2: _2_VLs - VL0, VL1
0x3: _4_VLs - VL0..VL3
0x4: _8_VLs - VL0..VL7
0x5: _15_VLs --VL0, VL14 (VL15 is always present for MADs)
NUM_OF_VL_P2=<_15_VLs|_1_VL|_2_VLs|_4_VLs|_8_VLs>
QOS_TRUST_STATE_P1=<TRUST_DSCP|TRUST_DSCP_PCP|TRUST_PCP|TRUST_PORT>Identifies the layer trusted to carry QoS/Flow Control information
0x0: TRUST_PORT - use transmit port
0x1: TRUST_PCP - use VLAN.PCP field
0x2: TRUST_DSCP - use IP.DSCP field
0x3: TRUST_DSCP_PCP - use VLAN.PCP when IP.DSCP is not present
QOS_TRUST_STATE_P2=<TRUST_DSCP|TRUST_DSCP_PCP|TRUST_PCP|TRUST_PORT>
VL15_BUFFER_SIZE_P1=<NUM> Log (base 2) of the VL15 receive port buffer, given in units of 512B. Value 0x0 indicates device defaults.
Valid only for Infiniband links.
VL15_BUFFER_SIZE_P2=<NUM>
MPFS CONF:
DUP_MAC_ACTION_P1=<LAST_CFG|LOAD_BALANCE>Defines the forwarding behavior in MPFS for MACs which are duplicated in more than one PF.
0x0: LAST_CFG - last PF which added the MAC will receive the traffic.
0x1: LOAD_BALANCE - MPFS will load-balance ipv4 traffic for packets with destination MAC addresses that belong to more than one PF between all PFs that have this MAC
DUP_MAC_ACTION_P2=<LAST_CFG|LOAD_BALANCE>
IB_ROUTING_MODE_P1=<GID|LID> Select the routing mode for Infiniband among the PFs (and hosts, if applicable)
0x0: GID - each PF has a unique GID
0x1: LID - each PF has a unique LID
IB_ROUTING_MODE_P2=<GID|LID>
MPFS_MC_LOOPBACK_DISABLE_P1=<False|True>When TRUE, MC traffic from PFs/Hosts will be sent to uplink and not duplicated to any of the other PFs/Hosts
MPFS_MC_LOOPBACK_DISABLE_P2=<False|True>
MPFS_UC_LOOPBACK_DISABLE_P1=<False|True>When TRUE, UC traffic from PFs/Hosts will be sent to uplink regardless of the destination address
MPFS_UC_LOOPBACK_DISABLE_P2=<False|True>
SRIOV_IB_ROUTING_MODE_P1=<GID|LID> Select the IB routing mode for Virtual Functions
0x0: GID
0x1: LID
SRIOV_IB_ROUTING_MODE_P2=<GID|LID>
UNKNOWN_UPLINK_MAC_FLOOD_P1=<False|True>Defines the forwarding behavior in MPFS for packets arriving from the network (uplink) with destination MAC address that does not appear in the MPFS FDB.
When TRUE, these packets are flooded to all local MPFS ports.
When FALSE, these packets are dropped.
UNKNOWN_UPLINK_MAC_FLOOD_P2=<False|True>
LINK PHY CONF:
PHY_AUTO_NEG_P1=<AUTO_NEG_DISABLED|AUTO_NEG_ENABLED|DEVICE_DEFAULT>Control link Auto Negotiation operation
0x0: DEVICE_DEFAULT
0x1: AUTO_NEG_ENABLED
0x2: AUTO_NEG_DISABLED
other values are reserved.
PHY_AUTO_NEG_P2=<AUTO_NEG_DISABLED|AUTO_NEG_ENABLED|DEVICE_DEFAULT>
PHY_FEC_OVERRIDE_P1=<DEVICE_DEFAULT|MODE_1|MODE_2|MODE_3|MODE_4>Customizes FEC configuration of the port
0x0: DEVICE_DEFAULT
0x1: MODE_1 - disabled DME, disabled FEC for 25G/50G/100G
0x2: MODE_2 - disabled DME, RS FEC for 25G/50G/100G
0x3: MODE_3 - disabled DME, FC FEC for 25G/50G or RS FEC for 100G
0x4: MODE_4 - disable non_dme (enable Auto Negotiation)
Other values are reserved
PHY_FEC_OVERRIDE_P2=<DEVICE_DEFAULT|MODE_1|MODE_2|MODE_3|MODE_4>
PHY_IB_NDR_FEC_OVERRIDE_P1=<DEVICE_DEFAULT|ELL_RS_FEC_272_257_PLR|RS_FEC_544_514|RS_FEC_544_514_PLR>Customizes FEC configuration of the port for IB NDR link speed
0x0: DEVICE_DEFAULT
0x1: RS_FEC_544_514
0x2: RS_FEC_544_514_PLR
0x3: ELL_RS_FEC_272_257_PLR
Other values are reserved
PHY_IB_NDR_FEC_OVERRIDE_P2=<DEVICE_DEFAULT|ELL_RS_FEC_272_257_PLR|RS_FEC_544_514|RS_FEC_544_514_PLR>
PHY_RATE_MASK_OVERRIDE_P1=<False|True> When TRUE, the supported link speeds can be customized by PHY_RATE_MASK
PHY_RATE_MASK_OVERRIDE_P2=<False|True>
PHY_RATE_MASK_P1=<NUM> Enables the supported rates for the link. Each bit represents a rate, set bit indicates the speed may be used, if supported by the device, Valid when PHY_RATE_MASK_OVERRIDE is TRUE.
Bit 0 - SGMII_100M
Bit 1 - 1000BASE-X / SGMII
Bit 3 - 5GBASE-R
Bit 4 - XFI / XAUI-1 - 10G
Bit 5 - XLAUI-4/XLPPI-4 - 40G
Bit 6 - 25GAUI-1/ 25GBASE-CR / KR
Bit 7 - 50GAUI-2 / LAUI-2/ 50GBASE-CR2/KR2
Bit 8 - 50GAUI-1 /50GBASE-CR / KR
Bit 9 - CAUI-4 / 100GBASE-CR4 / KR4
Bit 10 - 100GAUI-2 / 100GBASE-CR2 / KR2
Bit 11 - 100GAUI-1 / 100GBASE-CR / KR
Bit 12 - 200GAUI-4 / 200GBASE-CR4/KR4
Bit 13 - 200GAUI-2 / 200GBASE-CR2/KR2
Bit 15 - 400GAUI-8/ 400GBASE-CR8
Bit 16 - 400GAUI-4/ 400GBASE-CR4
Bit 19 - 800GAUI-8/ 800GBASE-CR8
Bit 31 - SGMII_10M
PHY_RATE_MASK_P2=<NUM>
WOL CONF:
WOL_MAGIC_EN=<False|True> Enables server Wake-on-LAN upon reception of WOL magic packet.
PF PCI CONF:
PF_DEVICE_ID=<NUM> The PCIe device ID used by this function.
PF_DEVICE_ID_ENABLE=<False|True> If FALSE, PF_DEVICE_ID is ignored
PF_NUM_OF_VF=<NUM> The total number of Virtual Functions (VFs) that will be exposed, for this PF. Value 0x0 indicates SR-IOV will be disabled.
Valid only when PF_NUM_OF_VF_VALID is set to TRUE
PF_NUM_PF_MSIX=<NUM> Number of MSI-X vectors assigned for this PF
Value 0x0 indicates device defaults.
Valid only when PF_NUM_PF_MSIX_VALID is set to TRUE
PF_NUM_VF_MSIX=<NUM> Number of MSI-X vectors assigned for each VF associated with the PF.
Value 0x0 indicates device defaults
valid only when PER_PF_NUM_VF_MSIX is set to TRUE.
PF_SD_GROUP=<NUM> The Socket Direct group index of the matching PF.
Value 0x0 indicates no grouping.
PF_SF_BAR_SIZE=<NUM> Log (base 2) of the BAR size of a single SF, given in KB. Valid only when PF_TOTAL_SF is non-zero and PER_PF_NUM_SF is set to TRUE.
PF_TOTAL_SF=<NUM> The total number of Sub Function partitions (SFs) that can be sup
ported, for this PF.
Valid only when PER_PF_NUM_SF is set to TRUE
HCA CONF:
MKEY_BY_NAME_RANGE=<DEVICE_DEFAULT|MKEY_RANGE_DISABLE|MKEY_RANGE_ENABLE>Control support for reserved Mkey range for Mkeys created by name. See MKEY_BY_NAME.
0x0: DEVICE_DEFAULT
0x1: MKEY_RANGE_DISABLE
0x2: MKEY_RANGE_ENABLE
other values are reserved
MULTI_PORT_VHCA_EN=<False|True> When set, the PF reports the multi-port-vhca capability.
PCI_WR_ORDERING=<force_relax|per_mkey> Controls the PCI ordering attribute for write accesses from the device
0x0: per_mkey - Each MKey may be configured differently for Relaxed Ordering
0x1: force_relax - All MKeys will be set to use Relaxed Ordering at creation time
show this setting only when NV_PCI_CONF. advanced_pci_settings is true.
ROCE_CONTROL=<DEVICE_DEFAULT|ROCE_DISABLE|ROCE_ENABLE>Control support for RDMA over Converged Ethernet (RoCE)
0x0: DEVICE_DEFAULT
0x1: ROCE_DISABLE
0x2: ROCE_ENABLE
other values are reserved
SILENT_MODE=<False|True> When TRUE, the networking function is disconnected from the network. It cannot transmit/received traffic directly to/from the network. Any communication must be done within the function, or by directing the traffic through another function.
ECPF CONF:
ECPF_ESWITCH_MANAGER=<ECPF|EXTERNAL_HOST_PF>This parameter is deprecated for INTERNAL_CPU_ESWITCH_MANAGER
Select who owns the E-switch.
0x0: EXTERNAL_HOST_PF
0x1: ECPF
ECPF_PAGE_SUPPLIER=<ECPF|EXTERNAL_HOST_PF>This parameter is deprecated for INTERNAL_CPU_PAGE_SUPPLIER
Select who supplies ICM pages of the external host functions to device.
0x0: EXTERNAL_HOST_PF - External host supplies required ICM pages to handle it.
0x1: ECPF - ECPF supplies required ICM pages to handle the external host.
EXTERNAL PORT CTRL:
ALLOW_RD_COUNTERS=<False|True> If set, indicates the host function which is allowed to read counters of external physical port.
PORT_OWNER=<False|True> If set, indicates the function of the host's own the external physical port.
RENEG_ON_CHANGE=<False|True> When cleared, a link-down/up sequence from the driver (PAOS) triggers re-negotiation of the link speed and parameters with the remote peer. When set, a link down/up sequence will trigger re- negotiation only if any link parameters were changed by a driver since last link negotiation. This parameter is applicable when the port_owner parameter is set for this PF
TRACER_ENABLE=<False|True> If set, indicates the function may own the Device Tracer.
ROM BOOT CONF2:
BOOT_UNDI_NETWORK_WAIT=<NUM> The number of seconds Flexboot waits after issuing a UNDI open commands until it starts using the interface.
IP_VER=<IPv4|IPv4_IPv6|IPv6|IPv6_IPv4> Select which IP protocol version will be used by Flexboot. If both are configured, Flexboot will try to boot with the 2nd protocol only if DHCP parameters for the first protocol are not available, or if booting with the first protocol has failed.
0x0: IPv4
0x1: IPv6
0x2: IPv4_IPv6
0x3: IPv6_IPv4
ROM UEFI CONF:
UEFI_HII_EN=<False|True>
ROM DEBUG LEVEL:
BOOT_DBG_LOG=<False|True>
BOOT_DBG_LOG_ARP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's Address resolution Protocol (ARP)
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_DHCP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's DHCP
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_DHCP6=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's DHCPv6
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_DRIVER_SETTINGS=<ADVANCED|ALL|BASIC|DISABLE>Debug level for non volatile configuration management.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_DRV=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s driver (init, teardown, poll CQ etc)
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_IP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s IPv4 stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_IPV6=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s IPv6 stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NDP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s IPv6 Neighbor Discovery Protocol
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NDRV=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s NODNIC driver (init, teardown, poll CQ etc)
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NDRV_CMD=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s NODNIC Driver interface
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NDRV_DEV=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s NODNIC Driver tear-down and bring ups operations.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NDRV_PORT=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s NODNIC Driver port management.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NEIGHBOR=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's Neighbor table (which is filled by ARP and NDP)
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_NETDEVICE=<ADVANCED|ALL|BASIC|DISABLE>Network Device Link status and traffic
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_PXE_UNDI=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s UNDI interface
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_ROMPREFIX=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's driver bootstrap code
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_STATUS=<ADVANCED|ALL|BASIC|DISABLE>Debug level for status update interface.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_STP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot's Spanning Tree Protocol
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_TCP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s TCP stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_TCPIP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s TCP/IP stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_TFTP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s TFTP stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_UDP=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s UDP stack.
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
BOOT_DBG_LOG_URI=<ADVANCED|ALL|BASIC|DISABLE>Debug level for Flexboot"s URI parsing code
0x0: DISABLE
0x1: BASIC
0x2: ADVANCED
0x3: ALL
ROM UEFI DEBUG LEVEL:
UEFI_DEBUG_LOG_BLKIO=<False|True> Enable UEFI debug level logs for BlkIo Driver.
UEFI_DEBUG_LOG_BM=<False|True> Enable UEFI debug level logs for Boot Manager.
UEFI_DEBUG_LOG_CACHE=<False|True> Enable UEFI debug level logs for Memory range cachebility changes.
UEFI_DEBUG_LOG_DISPATCH=<False|True> Enable UEFI debug level logs for PEI/DXE/SMM Dispatchers.
UEFI_DEBUG_LOG_EVENTS=<False|True> Enable UEFI debug level logs of Event.
UEFI_DEBUG_LOG_FS=<False|True> Enable UEFI debug level logs for EFI file system accesses.
UEFI_DEBUG_LOG_GCD=<False|True> Enable UEFI debug level logs for Global Coherency Database changes.
UEFI_DEBUG_LOG_INIT=<False|True> Enable UEFI debug level initialization log.
UEFI_DEBUG_LOG_LOAD=<False|True> Enable UEFI debug level logs for load events
UEFI_DEBUG_LOG_PAGE=<False|True> Enable UEFI debug level logs for page Alloc & Free.
UEFI_DEBUG_LOG_POOL=<False|True> Enable UEFI debug level logs for pool Alloc & Free.
UEFI_DEBUG_LOG_SNI=<False|True> Enable UEFI debug level logs SNI Driver.
UEFI_DEBUG_LOG_UNDI=<False|True> Enable UEFI debug level logs UNDI Driver.
UEFI_LOGS=<DISABLED|STDERR|STDOUT> Select the output device logs generated by the NIC/HBA UEFI expansion ROM.
0x0: DISABLED
0x1: STDOUT
0x2: STDERR
UEFI_LOG_MIN_SEVERITY=<DEBUG|ERROR|INFO|WARNING>NIC/HBA UEFI expansion ROM
0x0: ERROR
0x1: WARNING
0x2: INFO
0x3: DEBUG
Note - Enabling of debug logs for specific modules is possible after setting the severity to DEBUG.
UEFI_LOG_VERBOSE=<False|True> Detailed debug messages that may significantly impact boot
performance debug messages.
ROM BOOT CONF1:
BOOT_INTERRUPT_DIS=<False|True> When set to TRUE, legacy interrupts should not be used for receive/transmit indication. Polling should be used instead.
Supported only if boot_legacy_interrupt_disable_supported =TRUE.
BOOT_LACP_DIS=<False|True> Disable IEEE 802.3ad Link Aggregation Control Protocol (LACP) for legacy BIOS expansion ROM (Flexboot).
BOOT_RETRY_CNT1=<NONE|UNLIMITED|_1|_2|_3|_4|_5|_6>Number of retries to attempt in case of boot failure (FlexBoot)
0x0: NONE
0x1: _1
0x2: _2
0x3: _3
0x4: _4
0x5: _5
0x6: _6
0x7: UNLIMITED
BOOT_VLAN=<NUM> VLAN ID for network boot.
BOOT_VLAN_EN=<False|True> Enable VLAN mode for network boot.
LEGACY_BOOT_PROTOCOL=<ISCSI|ISCSI_WO_FAIL_PXE|NONE|PXE|PXE_WO_FAIL_ISCSI>Select boot protocol for legacy BIOS expansion ROM (Flexboot)
0x0: NONE - The expansion ROM will not add this PCI PF as a boot target during the POST stage.
0x1: PXE - The expansion ROM will use PXE boot.
0x2: ISCSI - The expansion ROM will use iSCSI boot.
0x4: PXE_WO_FAIL_ISCSI - PXE without fail-over to ISCSI. Supported when legacy_boot_wo_failover_supported==1.
0x5: ISCSI_WO_FAIL_PXE - ISCSI without fail-over to PXE.Supported when legacy_boot_wo_failover_supported==1.
ROM IB BOOT CONF:
BOOT_PKEY=<NUM> Infiniband P_Key to be used by PXE boot
PCI CONF:
ADVANCED_PCI_SETTINGS=<False|True> Show advanced PCI settings.
ATS_ENABLED=<False|True> When set to TRUE, the device will support ATS.
BAR_PAGE_ALIGNMENT=<ALIGN_4KB|ALIGN_64KB|DEVICE_DEFAULT>Controls PCI BAR page alignment
0x0: DEVICE_DEFAULT
0x1: ALIGN_4KB
0x2: ALIGN_64KB
Other values are reserved
DYNAMIC_VF_MSIX_TABLE=<False|True> When set to TRUE, the device will support allocation of MSI-Xs to VFs by the PF, after SR-IOV is enabled (while the VF is down).
EXP_ROM_NVME_UEFI_ARM_ENABLE=<False|True>Enable Expansion ROM option for UEFI for ARM based host for NVMe functions
EXP_ROM_NVME_UEFI_x86_ENABLE=<False|True>Enable Expansion ROM option for UEFI for x86 based host for NVMe functions
EXP_ROM_PXE_ENABLE=<False|True> Enable Expansion ROM option for PXE
EXP_ROM_UEFI_ARM_ENABLE=<False|True> Enable Expansion ROM option for UEFI for ARM based host
EXP_ROM_UEFI_x86_ENABLE=<False|True> Enable Expansion ROM option for UEFI for x86 based host
EXP_ROM_VIRTIO_BLK_UEFI_ARM_ENABLE=<False|True>Enable Expansion ROM option for UEFI for ARM based host for VirtIO-BLK functions
EXP_ROM_VIRTIO_BLK_UEFI_x86_ENABLE=<False|True>Enable Expansion ROM option for UEFI for x86 based host for VirtIO-BLK functions
EXP_ROM_VIRTIO_NET_PXE_ENABLE=<False|True>Enable Expansion ROM option for PXE for VirtIO-NET functions
EXP_ROM_VIRTIO_NET_UEFI_ARM_ENABLE=<False|True>Enable Expansion ROM option for UEFI for ARM based host for VirtIO-NET functions
EXP_ROM_VIRTIO_NET_UEFI_x86_ENABLE=<False|True>Enable Expansion ROM option for UEFI for x86 based host for VirtIO-NET functions
FORCE_ETH_PCI_SUBCLASS=<False|True> Force the PCI function identifier with Ethernet subclass (00h). Supported only when ADVANCED_PCI_SETTINGS is set.
IBM_AS_NOTIFY_EN=<False|True> Enable IBM's AS Notify feature. Supported only when NV_PCI_CAP.ibm_as_notify_supported is set.
IBM_CAPI_EN=<False|True> Enable IBM's Coherent Accelerator Processor Interface (CAPI) mode. Supported only when ADVANCED_PCI_SETTINGS is set.
IBM_TUNNELED_ATOMIC_EN=<False|True> Enable IBM's Tunneled Atomic feature. Supported only when NV_PCI_CAP.ibm_tunneled_atomic_supported is set.
MANAGEMENT_PF_MODE=<DEVICE_DEFAULT|MNG_PF_DISABLE|MNG_PF_ENABLE>Controls exposing a management networking PF connected directly to the BMC.
0x0: DEVICE_DEFAULT
0x1: MNG_PF_ENABLE
0x2: MNG_PF_DISABLE
Other values are reserved
P2P_ORDERING_MODE=<DEVICE_DEFAULT|SECURE_ALL|SECURE_NONE|SECURE_TRUST>PCI Peer-to-Peer ordering security mode, between read requests completions and writes.
0x0: DEVICE_DEFAULT
0x1: SECURE_NONE - no protection/no peer-to-peer
0x2: SECURE_TRUST - protect untrusted functions
0x3: SECURE_ALL - protect all functions
SAFE MODE CONF:
SAFE_MODE_ENABLE=<False|True> Enable Safe Mode activation after crossing a threshold of consecutive failed boots (no driver load).
SAFE_MODE_THRESHOLD=<NUM> Default value of the threshold for failed driver loads (driver load was not detected between reboots). When crossed, the device will load in Safe Mode with some default configurations, suspected as possible causes for driver load fail.
HOST PRIV CONF:
HOST_DISABLE_PORT_COUNTER=<False|True> When TRUE, the host will not be allowed to read Physical port counters
HOST_DISABLE_PORT_OWNER=<False|True> When TRUE, the host will not be allowed to be Port Owner
HOST_DISABLE_RSHIM=<False|True> When TRUE, the host does not have an RSHIM function to access the embedded CPU registers
HOST_DISABLE_TRACER_OWNER=<False|True> When TRUE, the host will not be allowed to own the Tracer
HOST_PRIV_LEVEL=<LIMITED|PRIVILEGED> Privilege level of the host
0x0: PRIVILEGED - host has all supported privileges
0x1: LIMITED - host is not allowed to modify global/per port/parameters or access other hosts parameters
四、案例分享
1、问题描述
用户将Mellanox网卡的两个端口配置成Bond模式(Active/Backup模式)，在测试过程中发现如果在OS中关闭活动端口（使用ip link set down命令），Bond不能立即切换到backup端口。需要等待120秒左右才会切换。这导致网络出现120秒的中断。
2、问题分析
2.1、测试直接拔掉网线，端口会立即切换到backup端口，网络不受影响。
2.2、通过观察发现，在系统中使用命令关闭活动端口后，网卡与交换机的物理链路还是UP状态，网卡与交换机的端口LED灯还是点亮状态。这说明使用ip link set down命令只是在系统逻辑层面将端口关闭，实际上物理层面的链接并没有被关闭，这导致端口没有进行failover切换。
3、解决办法
3.1、需要通过修改Mellanox网卡的端口配置来解决。Mellanox网卡固件中有一个功能叫做“KEEP_ETH_LINK_UP”。这将始终保持网卡端口物理链路处于活动状态，即使在OS逻辑层面已经将端口shutdown。这个功能默认是开启的，这是由于一些功能和服务需要网卡物理链路始终处于活动状态，例如socket-direct、multi-host、share-lan等功能。解决客户遇到的问题就是将该功能设置为“关闭”。关闭该功能后，在OS层面关闭端口，物理链路也会关闭。
3.2、参考命令如下，这个功能是基于端口的，所以要在需要关闭此功能的端口上分别进行操作，例如下面的命令是将同一张卡上的两个端口都进行设置。
sudo mlxconfig -d mt4117_pciconf0 set KEEP_ETH_LINK_UP_P1=0
sudo mlxconfig -d mt4117_pciconf0.1 set KEEP_ETH_LINK_UP_P2=0
