| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
[1] disabled redfish-allow-deprecated-power-thermal by default in
bmcweb and set a June 2026 date for its removal.
Follow bmcweb and disable redfish-allow-deprecated-power-thermal.
Cleanup places that already disabled it.
It has been 5 years since Redfish put out the improved PowerSubsystem
and ThermalSubsystem. bmcweb has had those enabled by default for 1.5
years. Cleaning these up allows for removing
redfish-allow-deprecated-power-thermal easier next year.
[1]: https://gerrit.openbmc.org/c/openbmc/bmcweb/+/83023
Change-Id: Ic1fa0ba346a0572f9c5cdc3e9d58b6e9fd4084ba
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>
|
|
Change-Id: I1d53510329749ff122c8a65157a72c8ff88bd306
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
For some of the most commonly set bmcweb options, change them into
packageconfig entries.
Systems that used these config options have been moved to the
PACKAGECONFIG. Some systems that had CMAKE build options still had
those directly removed.
Change-Id: I741f60a725905d67f7a8ef698f95b25895088db4
Signed-off-by: Ed Tanous <etanous@nvidia.com>
|
|
This commit updates latest change to Ampere platforms pwquality.conf
which:
1. Disable checking passwords against user name
2. Disable checking passwords against CrackList [1]
[1]: https://linuxsoft.cern.ch/cern/slc6X/x86_64/yum/updates/repoview/cracklib.html
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: I183a8c32b108b98dd113acb602a497acbfd9f2f0
|
|
Ampere flash bios script will find host UEFI partition by grep pattern
"pnor-uefi" in /proc/mtd and expect only one result. Which will fail if
this pattern appears in multiple partition.
This patch matches exact MTD partition name when finding MTD device to
flash host firmware.
Change-Id: Id674b544b5219347cd6e87a0be4d03983e26a04e
Signed-off-by: Tam Nguyen <tamnguyenchi@os.amperecomputing.com>
|
|
Prevent APR feature is run when BMC has been rebooted.
Tested:
1. Set power policy is always-on
$ipmitool chassis policy always-on
2. Turn off Host
$ipmitool power off
3. Reboot BMC
$reboot
4. Wait until BMC is booted.
BMC does not run APR feture as expectation.
Change-Id: Id811ee023e82e762ca242faaa40c79d2bf98457a
Signed-off-by: Thang Tran <thuutran@amperecomputing.com>
|
|
We are using openssl command to create the certificate fingerprint [1],
but there is no dependency listed in the recipe.
This adds RDEPENDS on the openssl binary.
Long term solution should be linking to the openssl library directly,
but as we intend to replace this current implementation with something
else different [2], this change should suffice in the meanwhile.
Tested:
- openssl command executes successfully:
root@mtmitchell-dcscm:~# openssl -v
OpenSSL 3.4.0 22 Oct 2024 (Library: OpenSSL 3.4.0 22 Oct 2024)
- Get Manager Certificate Fingerprint via IPMI inband successfully:
[root@fedora ~]# ipmitool raw 0x2c 0x01 0x52 0x01
52 01 d4 58 4d 68 1c 78 8f 94 3d 48 98 a1 fa 62
38 bf d1 7e 85 9a bf 7a 3f dc 80 a3 16 eb ed 6d
d5 f5
[1] https://github.com/ampere-openbmc/ampere-ipmi-oem/blob/f92b584da2253797d17ba8c4af8c6f1cbe28dad2/src/redfishhostinterface.cpp#L264-L271
[2] https://gerrit.openbmc.org/c/openbmc/docs/+/79327
Change-Id: I1292ddbe96ac7ca08a85d81686115fde83516723
Signed-off-by: Khang D Nguyen <khangng@os.amperecomputing.com>
|
|
Correct OWNERS file configuration to use Email instead of user name in
CLA.
Change-Id: I99e6ea131f362b12a6a522c06e0f7943e18609aa
Signed-off-by: Thu Nguyen <thu@os.amperecomputing.com>
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I6291eeaf0ca7d4bd674bf0717f1234bfd757f6fe
|
|
Thang Q. Nguyen (1):
OWNERS: update maintainer information
Change-Id: I3b9ca189da8b43e1afa6009a56d644dcd6771731
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
As I will change to another assignment soon, remove me from the OWNER
list.
Note that Thu Nguyen still maintains meta-ampere.
Change-Id: I9688fef7a623e51d9dff5d82fbc6efa66e5b95fd
Signed-off-by: Thang Q. Nguyen <thang@os.amperecomputing.com>
|
|
When CPU temp fluctuates around the hightemp threshold point, 2
hightemp assertion/deassertion services are triggered by gpio events
continuously, which can reach the default rate limit (5 starts/sec)
of the systemd services and cause missing of logs. When a service
reaches that, it won't be able to start anymore. This happens
throughout all hightemp/overtemp test cases and gives an illusion of log
dupplications.
This commit adds [Unit] StartLimitIntervalSec=0 to
ampere_hightemp_start@.service and ampere_hightemp_stop@.service to
disable the rate limit on these services.
Tested: Force CPU temp to reach hightemp (by slowering fan speeds to 10%
and run stress-ng script from Host console)
=> When CPU fluctuates around the hightemp, as long as
phosphor-multi-gpio-monitor service triggers a gpio event on this gpio
line, the corresponding service will be called.
NOTE: This only ensures systemd service calls are sufficiently executed
upon events from phosphor-multi-gpio-monitor service. This does not
guarantee the preciseness of events triggered by the mentioned service.
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: Ieb47858e94b7e18e1c999aad29a38546a3b13f36
|
|
This commit refactors phosphor-fan configurations, overriding services
and ampere_fanctrl.sh script for Mt.Jade, Mt.Mitchell and Mt.Jefferson:
1. phosphor-fan overriding service files, under 3 platforms are moved to
meta-common/
2. ampere_fanctrl.sh is beautified and added 2 commands for getting
speeds of all fans and setting speeds for all fans.
3. phosphor-fan config files for Mt.Jade are moved to a folder ahead
instead of being placed in mtjade/.
4. ampere_fanctrl.sh script is added for Mt.Jade but only supports
"setspeed all" to replace the ampere_set_fan_max_speed.sh. This
ensures a constant behavior among platforms.
This commit also lets platform_init script use ampere_fanctrl.sh to set
initial fan speed.
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: I4a8f9590ce52aac28fe21b1c6545d53933da747b
|
|
Currently fanctl is not being used by phosphor-fan. However, it is
included in the image by default as a development tool.
This commit removes fanctl from final image
and reduces squashfs-xz image size by ~200KB.
Tested: fanctl is removed from rootfs.
Co-authored-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: I33ab7fe473f05a761a51177dde700aea6395fd63
Signed-off-by: Khang Nguyen <khangng@os.amperecomputing.com>
Signed-off-by: Nhat Hieu Le <lhieu@os.amperecomputing.com>
|
|
We tried to remove phosphor-hwmon in packagegroup-obmc-apps a long time
ago, but phosphor-hwmon is still being built because it is the default
value of VIRTUAL-RUNTIME_obmc-sensors-hwmon and we did not override
the value properly. This can be verified by checking the dependencies
graph of the final image:
$ bitbake -g obmc-phosphor-image
...
$ oe-depends-dot -k phosphor-hwmon -w ./task-depends.dot
Because: obmc-phosphor-image packagegroup-obmc-apps
obmc-phosphor-image -> packagegroup-obmc-apps -> phosphor-hwmon
This removes the recipe by reverting previous change and
overriding the config properly.
Tested:
$ oe-depends-dot -k phosphor-hwmon -w ./task-depends.dot
ERROR: Can't find key phosphor-hwmon in ./task-depends.dot
Change-Id: Id0000000703bbe24f6624275f826a7b84ecb53e6
Signed-off-by: Khang D Nguyen <khangng@os.amperecomputing.com>
|
|
In the past, we tried to remove phosphor-ipmi-fru, but only the dynamic
library was removed. Some binaries from the recipe are still included
(phosphor-read-eeprom).
By default, phosphor-ipmi-fru is included in
VIRTUAL-RUNTIME_phosphor-ipmi-providers variable.
This empties the variable and prevents phosphor-ipmi-fru
from being built on Ampere platforms.
Tested on Mitchell, its phosphor-read-eeprom binary is removed.
ipmitool fru print works normally.
Change-Id: Id0000000d728abcd11ad4d47ab0bf48ed713070a
Co-authored-by: Thang Tran <thuutran@os.amperecomputing.com>
Signed-off-by: Khang D Nguyen <khangng@os.amperecomputing.com>
|
|
This change aims to adapt to the modifications from upstream [1]. It
allows updating value to D-Bus with the input value being a byte array.
Tested:
1. Do power cycle HOST
2. Check `BootProgress` is change when `obmcutil status`
3. Read value
busctl get-property xyz.openbmc_project.State.Boot.Raw \
/xyz/openbmc_project/state/boot/raw0 \
xyz.openbmc_project.State.Boot.Raw Value
3. Check data
busctl call xyz.openbmc_project.State.Boot.PostCode0 \
/xyz/openbmc_project/State/Boot/PostCode0 \
xyz.openbmc_project.State.Boot.PostCode GetPostCodes q 1 -j
4. Redfish check:
curl -X GET --user root:0penBmc --insecure -H Content-Type: \
application/json https://<bmc_ip>/redfish/v1/Systems/\
system/LogServices/PostCodes/Entries
5. WebUI login go to Logs -> POST Code logs
The POST code is display
[1]: https://gerrit.openbmc.org/q/topic:%22primary-postcode-new-intf%22
Change-Id: Id00000004e2535a597559c668b03b7bc647a84c3
Signed-off-by: Khang D Nguyen <khangng@os.amperecomputing.com>
|
|
There are no records on why we included util-linux in our images,
as well as no dependencies from other recipes.
We do use the logger utility from util-linux, but it is in its own
util-linux-logger recipe now.
This removes util-linux from our final images.
Tested on Mitchell:
- Does not affect the release regression test.
- logger works.
Change-Id: Id00000009f8153e6732637c2aec30d1eeaad5be2
Signed-off-by: Khang D Nguyen <khangng@os.amperecomputing.com>
|
|
Change I69b9651ff36ecc0f0769b46fad94477d6f883786 eliminated the
non-functional "example" entity-map and replaced it with an empty
default. Since multiple meta-layers already use an empty default,
remove the extra bitbake code and instead rely on the empty default
in meta-phosphor.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I1a29671727e6c58357b4371b0a66356410700ba0
|
|
When BMC reboot, it can not control the leds, therefore, it should
deassert all of the Leds to inform to CPLD that BMC goes to BMC-less
mode.
Tested:
1. Request to flash/reboot BMC.
2. All off the Leds are turned off as expectation.
Change-Id: I0a992343c66873edd8efe4e05ffaa0fe6e7dfa4c
Signed-off-by: Thang Tran <thuutran@amperecomputing.com>
|
|
The overtemp script uses bash in its shbang and thus depends on it.
Without explicitly requiring it, bitbake may fail because of a QA
failure.
Tested: Ran 'bitbake obmc-phosphor-image' for mtjade successfully.
Change-Id: I062a9d266b9bb4389015a8758157b80981834a71
Signed-off-by: Tan Siewert <tan@siewert.io>
|
|
Add ampere-mctp-i2c-binding service to add/remove MCTP endpoints.
When turning the Host ON, the service will use mctp commands to
configure Host MCTP endpoints.
When turning OFF the Host, the service will automatically remove MCTP
endpoints.
Tested:
1. Check the MCTP Endpoints D-Bus object paths in
xyz.openbmc_project.MCTP service when the host is on in 1P/2P
systems.
```
|-/xyz/openbmc_project/mctp/1/20
|-/xyz/openbmc_project/mctp/1/22
```
Change-Id: I69a23f3f60584c08f72c540cec86ab0f7aee9c1e
Signed-off-by: Thu Nguyen <thu@os.amperecomputing.com>
|
|
Enable check-fwupdate-before-do-transition option to check updating
status when users request to do BMC/Chassis/Host transition. When
firmware being updated, BMC/Chassis/Host transition requests are
rejected to ensure system is safe.
Tested:
1. Turn off the Host.
2. Request to flash new Host firmware via WebUi/Redfish.
3. Turn on the host.
$ipmitool power on
4. BMC rejects request.
5. Wait until flashing done.
6. Request to turn on the Host.
$ipmitool power on
7. Host is turned on.
Change-Id: I7e99aaa03f3cee3a07d87c01b9656d0738c482ac
Signed-off-by: Thang Tran <thuutran@amperecomputing.com>
|
|
This commit corrects all MessageID strings used in the repository for
Redfish Event Log. The MessageID is parsed to be mapped with Redfish
Registries format in bmcweb, and it should only has 4 fields
<RegistryName.MajorVersion.MinorVersion.MessageKey>. Any redundant
string field added to that will cause a parsing failure and the log
won't appear the Redfish Event Log entries.
Signed-off-by: Chaul Ly <chaul@amperecomputing.com>
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Change-Id: I4ff2cf83538b0cde6f4af001ba253825c1691332
|
|
Signed-off-by: hieul.ampere <hieul@amperecomputing.com>
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Change-Id: I73975d53f8f95c3163c431965939b276c98a792b
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I2665e9e7e37242d546eafdb4d1c65dd0369ca635
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I8f74fcbb7bf66bb24fde1571ef09fac1095c7e59
|
|
Jayanth Othayoth (2):
clang-tidy: Replace NULL with nullptr
clang-tidy: Initial commit
Change-Id: I1c62898d790c8844c69a31e83f72685b76c4dfbb
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
As hardware design, the BMC_GPIO18B7_S01_UART1_SEL GPIO should be set
to 1 to select the Mpro0.
Tested:
1. Open SOLSSH console for Mpro0
$ ssh -p 2201 root@<BMC_IP>
2. Verify that Mpro0 console can show logs when Host power on
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I87cfa49a19896f75b239d664c875068f53cd5f84
|
|
Issue: When users request OOB command "ipmitool sol activate", if
host's console show any data then users can not request any OOB
ipmi commands.
Root cause: After the @acb8d90 commit, the rmcp-ping option effects
to message handling, it makes the netipmid fail to send/receive
message to/from "sol client".
This commit disables rmcp-ping option to make the netipmid can work
normally when users request OOB "ipmitool sol activate" command.
Tested:
1. Request access Host's console via OOB command.
$ ipmitool sol activate
2. Request power cycle.
$ ipmitool power cycle
3. The Host's console show the logs of Host.
4. Request another OOB command.
$ ipmitool chassis status
5. The chassis status is shown.
Change-Id: I4f42c816471343e398779858536f861d0af98ae7
Signed-off-by: Thang Tran <thuutran@amperecomputing.com>
|
|
The latest yocto release(styhead,5.1) has made a significant change[1]
to the use of the WORKDIR parameter.
This is not the complete change required but gets us going in the right
direction.
[1]: https://docs.yoctoproject.org/next/migration-guides/migration-5.1.html#workdir-changes
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I4f184b361be69e7704a0657a4a00cdad981aff8a
|
|
Input GPIOs don't need to be configured as input. Applications that
poll the GPIO should configure it to input direction before checking.
This commit removes codes to configure input direction for GPIOs.
Signed-off-by: Thu Nguyen <thu@os.amperecomputing.com>
Change-Id: Iace6173efae9a2c8f3ccd9f3f4ae0bed5239955d
|
|
There was no way for phosphor-fan to know if it host boots with 1P or
2P, so when it detects absence of any of the thermal sensor used for
thermal control on Socket 1, it will consider as a sensor failure and
set fan PWM to 100%. However, the effort to support GPIO presence
detection in phosphor-fan just to decide whether to put one sensor
into consideration is big.
There's a need to support 2 different config set for mtmitchell 1P
and 2P
Their common config files will be placed in
/usr/share/phosphor-fan-presence/<daemon>/mtmitchell/ together with
the 2 different files groups_1p.json and groups_2p.json in
/usr/share/phosphor-fan-presence/control/mtmitchell/. These two files
contains the thermal sensors needed for thermal control in 1P and 2P
system which is the only different point between 1P and 2P. The file
to be loaded which is groups.json will have the same content as
groups_2p.json at build time.
phosphor-fan will read the PrettyName property under the
xyz.openbmc_projects.Inventory.Item interface to look for the
sub-folder if any config can't be found in the default location.
The property is configured in Entity Manager config so it can be
probed together with the Motherboard with the value "mtmitchell".
When BMC boots and ampere-platform-init service detects CPU1
presence, it will copy groups_2p.json into groups.json. Otherwise,
it will copy groups_1p.json into groups.json.
That way, phosphor-fan can used a different set of thermal sensors
for 1P.
Tested:
BMC boots on 1P system, in normal condition, fans are not driven to
the max speed.
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: If67efc4792b3d3295478a892ab73bdd9f3a41fa3
|
|
`Type` has been removed from phosphor-virtual-sensors [1], so this
commit updates the JSON configuration file and deletes all `Type`.
[1]: https://github.com/openbmc/phosphor-virtual-sensor/commit/06f8874f26adff78f960c8937ce11b723ad31e87
Signed-off-by: George Liu <liuxiwei@ieisystem.com>
Change-Id: I9dd1bc85779095ae5f034be527a5bd30b2ba5b2a
|
|
This filters NVMe sensors from IPMI sensor report by adding a list of
service to filter to /usr/share/ipmi-providers/sensor_filter.json.
Tested: check if nvmeX sensors are not reported by IPMI
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: Ied8c7c3a6d57e89fc1398c2e01ebfb1812e7a845
|
|
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I1d050abbacc58e3b5cfa3d2bfb480dff61124524
|
|
This adds fan_fault, psu_fault, gpio_fault, overtemp_fault, and
ras_ue_fault to Led group.
Tested: verify that these leds created under
xyz.openbmc_project.LED.GroupManager
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: Ief92516a249b61c9ed828bfd8106368a0b5e18bf
|
|
This commit configures the bus and address of the BMC FRU device
so that IPMI OEM commands can identify such.
Tested:
1. Request to update MAC address.
$ ipmitool raw 0x3c 0x01 0x70 0xe2 0x84 0x86 0x76 0xc0
2. Reboot then check MAC address.
$ ip a
3. MAC address is updated as step 1.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I236da1b233cd4d6b04bd5c2fdcfd888de4e4fa17
|
|
This commit supports initial automatic fan control with minimized
control table following Sx_SOC_Pkg_Temp.
Tested: BMC is on, fans run following the table
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: If99787b273347bd57ae88918403a76da42c06ae4
|
|
Add blacklist configuration for Mt.Jefferson to prevent
entity-manager scanning to unexpected FRU devices.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I3c199da965d55c2580f95a101a21dd4e41b64173
|
|
Add blacklist configuration for Mt.Mitchell to prevent entity-manager
scanning to unexpected FRU devices.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I0eac341f3fa2ebe22a129fa862ba20edda536cdd
|
|
Add blacklist configuration for Mt.Jade to prevent entity-manager
scanning to unexpected FRU devices.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: If306eb3509bbcb4cd5c4650cb886b68b57893bdb
|
|
This enables IPMI dynamic sensor to support on-board sensors report
via IPMI.
Tested: the on-board sensors report via ipmitool sdr command.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: Ie9473d6d8245da48c4730a8891f9063dc16c9bfd
|
|
This supports fan controller init and set the default speed to 60%.
Tested: verify that fan speed is as 60%.
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I1c81e3a394056acf0cb897e6175b63fcd64e5bda
|
|
This commit updates changes to phosphor-fan configuration files of
Ampere's Mt.Jade platform. This includes control methods by CPU sensors
which are DIMM and CPU temperature sensors.
Signed-off-by: Chau Ly <chaul@amperecomputing.com>
Change-Id: I7e6d5b2261cc31358549ff04d642c8971d94603c
|
|
Support Host power control for Mt.Jefferson platform.
Tested: Host power control via IPMI, Webui, and Redfish
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: Ib6ab3bf04ac4b61adc2578807975de1ae1472a92
|
|
This commit supports phosphor-hostlogger in STREAM mode, to forward
the byte stream from obmc-console to rsyslog via the imuxsock module.
The logs can be obtained in /var/log/. The logs are rotated by
logrotate. The size threshold of cpu console is 256KB and the others
are 64KB. Each log is polled every 1 minute to check if its file size
exceeds the threshold to rotate it to the new file. Each console can
be saved in up to 2 log files.
Tested (CPU console):
1. Connect to CPU SOL console
$ ssh root@<BMC_IP> -p 2200
2. Execute commands in the CPU SOL console
=> The console output on CPU SOL console will all be logged to
/var/log/obmc-console-cpu.log
=> The log will be rotated to /var/log/obmc-console-cpu.log.1 if the
original log file size exceeds 256KB every 1min poll
Signed-off-by: Hieu Huynh <hieuh@os.amperecomputing.com>
Change-Id: I2c22415bc2928732e7d7733e89fa2eb8faa13b9d
|
|
This commit adds the SOLSSH support for HOST console.
- BMC_UART1 : Boot S0 console
- BMC_UART2 : S0 Mpro console
Tested:
1. Connect to S0 Boot console
ssh -p 2200 root@bmc_ip
2. Connect to S0 Mpro console
ssh -p 2201 root@bmc_ip
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Change-Id: Ia95c81e1ca2072ee55af5eee173e67843f715a18
|
|
Support Host firmware update for Host EEPROM, MB1 FRU, MB2 FRU,
CIO FRU, BMC FRU, MB CPLD, BMC CPLD via BMC console and Redfish.
Signed-off-by: Chanh Nguyen <chanh@os.amperecomputing.com>
Change-Id: I987c4ee6808719778ae893f4cab80b10a2e7cd83
|
|
The virtual LAN over USB (usb0) connection between BMC and HOST is
locally utilized by the Redfish Host Interface. It is configured with
a static IPv4 address '192.168.0.10' by the network configuration
file [1] '00-bmc-usb0.network'.
Returning to the Ampere Altra Mt. Jade platform, an AC power cycle
caused the usb0 IP address to be lost [2]. The `ipmi lan print 3`
command gives a zero IP address on which HOST has not yet initiated
the PCIe USB device, which was caused by the usb0 link not being
ready. The solution was determined and well-tested. However, the same
issue with the upstream driver in edk2 is still present today.
Therefore, a clean solution ought to be used to fix it.
This patch sets ConfigureWithoutCarrier to 'yes' to allows usb0
interface to retain the static IP address even if the physical
carrier link is down. This helps HOST UEFI firmware to retrieve the
IP address information using the `IPMI lan print 3` command to
populate the SMBIOS Type 42 for the subsequent redfish discovery even
when the usb0 link is not yet established.
Tested with the following steps:
1. Check the usb0 IP address by the command `ip a`. The IP
192.168.0.10 should be shown in the command output.
2. Check the `ipmi lan print 3`. Since the IP address is available,
the IP address shown in the command output should be idential with
the one in step 1.
3. `networkctl down usb0` - Set the usb0 link down.
4. Run again step 1 & 2. The usb0 IP address should be retained.
Without this patch, it returns all zeros.
[1] https://www.freedesktop.org/software/systemd/man/latest/systemd.network.html
[2] https://den-bugzilla.amperecomputing.com/show_bug.cgi?id=92830
Signed-off-by: Nhi Pham <nhi@os.amperecomputing.com>
Change-Id: I1b1375082ea70510c1ea708947d7f796f6ba7624
|
