| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
In some cases, we noticed that the ttyUSB device's bus number could
change, which sometimes caused the rules to fail. The new extra rules
fix this issue.
Tested:
We have verified that the rules work on real machines.
Change-Id: Ida6427990fd6414e03fda1af0dfe2e941d8f32cd
Signed-off-by: P.K. Lee <p.k.lee@quantatw.com>
|
|
- Match the downstream machine name of "fbdarwin"
- This way, we can keep a downstream "switch" that changes the build from current Meta-OpenBMC image to LFOpenBMC.
- We need the machine names to match so upstream layers and conf files can be pulled.
Change-Id: Iea79dc648ac9f5605680d44fcad170af5bb2584c
Signed-off-by: Robert Ercean Dima <robert3@meta.com>
|
|
After switching usb between from sled 1 to sled 4 muliple times, usb driver may failed to detect usb hub, thereore, can not detect mac mini.
There are 2 ways to fix them:
1. There are four gpio to determinne which sled switch to BMC, if we set those gpio sequentially (not set them at the same time), it may cause usb driver at unkonown stage while setting those gpio. Therefore, use devemem command to set those gpio at once.
2. If we rebind the usb driver after switching usb mux, it may help usb driver re-detect usb hub if usb driver crash.
Change-Id: I1993bdc80b338eb10f9467396860efb423c67cb2
Signed-off-by: Eric Chiang Quanta <eric.chiang@quantatw.com>
|
|
Add software version D-Bus objects for Rainbow-mmc to support firmware
version tracking.
Tested: Verified that Rainbow-mmc version objects are created correctly
and accessible via D-Bus interface.
Change-Id: I6fe84752a38d849d7af0b7be7622cdf0ffd44422
Signed-off-by: Freddie Jheng <Freddie.Jheng@quantatw.com>
|
|
Use LineName instead of ChipId and GpioNum for each GPIO monitor to
avoid issues caused by GPIO rearrangements.
Change-Id: I92c02aa3c44dab75a650994ffc27488b84f71aad
Signed-off-by: Eldin Lee <eldin.lee@quantatw.com>
|
|
As requested by the Meta and Thermal teams, the system must not power
off before the valve train is closed. Therefore, a delay has been added
to meet this requirement.
Depends-On: https://gerrit.openbmc.org/c/openbmc/entity-manager/+/82525
Change-Id: Ief23624f4ffa36f551b313e5ffdfd8907c821057
Signed-off-by: Jason Hsu <Jason-Hsu@quantatw.com>
Signed-off-by: Eldin Lee <eldin.lee@quantatw.com>
|
|
to Interposer
We used the Interposer board in the Clemente DVT design change,
so the CPLD is now on the Interposer.
Change-Id: I401a0745a426bb53ff8b5dcb57db12b23fa1c2b6
Signed-off-by: Kimi Chen <kimi.zy.chen@fii-foxconn.com>
|
|
Add software version D-Bus objects for swb PCIe switch to support firmware version tracking.
Tested: Verified that swb PCIe switch version objects are created correctly and accessible via D-Bus interface.
Change-Id: Idf63273c21b726cd9b01863735d959131a7154fc
Signed-off-by: Freddie Jheng <Freddie.Jheng@quantatw.com>
|
|
Description
Add the service of on and off fan board 0 or fan board 1.
Motivation
When replacing the fan board, we need to turn off the fan board power.
Test plan:
1. systemctl start FanboardPowerOff@0
2. systemctl start FanboardPowerOn@0
3. sled cycle
4. Check whether the fan speed is normal
mfg-tool fan-speed
5. systemctl start FanboardPowerOff@1
6. systemctl start FanboardPowerOn@1
7. sled cycle
8. Check whether the fan speed is normal
mfg-tool fan-speed
Change-Id: I4c54721b5447557247733c4fbc4c7240db34eb04
Signed-off-by: Andy-Hsu-wiwynn <andy.hsu.wiwynn@gmail.com>
|
|
The FRU inventory and MAC address
(xyz.openbmc_project.Inventory.Item.NetworkInterface) are now processed
by entity-manager, so remove the unused configuration.
Change-Id: I5b53b8db20dadaff4d4cfdc80e0580d46c2dc630
Signed-off-by: Cosmo Chou <cosmo.chou@quantatw.com>
|
|
Add cable monitor and update it in sync with
entity-manager/configurations/meta/ventura_cable.json to support
the following ports:
- RackDripPanPort, RackFloorLeakPort1, RackFloorLeakPort2, RackRightManifoldPort, RackLeftManifoldPort
- LeakPort1 to LeakPort27
Change-Id: I440133f706c78c35f35eb8e6996d66441f00294d
Signed-off-by: Eldin Lee <eldin.lee@quantatw.com>
|
|
Remove ${PN}-chassis-poweron-log dependency
Remove phosphor-reset-chassis-on@.service,
phosphor-reset-chassis-running@.service,
obmc-power-start@.service, and obmc-power-stop@.service.
These services call mapper-wait@-org-openbmc-control-power* services,
but no corresponding /org/openbmc/control/power1 D-Bus object exists.
Mapper-wait waits indefinitely causing startup hang.
Removing these services fixes the boot wait issue.
Change-Id: Id055baf7a8113e8647b7b9b065d67fcbcc90ab5d
Signed-off-by: Jason Hsu <Jason-Hsu@quantatw.com>
|
|
Network switches need different networking configuration than compute switches.
Introduce FBOSS generic network configuration and add compute and fboss yocto overrides.
Change-Id: Ib5464003dc972894535c9ed7435e02b925cfcb21
Signed-off-by: Robert Ercean Dima <robert3@meta.com>
|
|
read-only
Platforms could use this common script when hard power-cycling.
Remounting the filesystem as read-only helps in avoiding possible data corruption.
Test Plan:
- Add logic around using this script inside yosemite4.
- In yosemite4 a sled cycle is hard powercycle of the whole platform.
- See the filesystem being remounted as read-only when calling sled-cycle
...
Execute sled cycle
o "/xyz/openbmc_project/logging/entry/120914"
Starting sled cycle...
Attempting to remount /run/mnt-persist as read-only with a number of 3 retries
Succeeded to remount /run/mnt-persist as read-only
...
Change-Id: I6dc837e0441f3cdb1345ba84856bc64ea5205a13
Signed-off-by: Robert Ercean Dima <robert3@meta.com>
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
- Add 0.5s delay after SLED insertion to avoid driver binding failures.
- Clear LED service state on SLED removal.
Change-Id: Ia4cb4eba980eb70a7ea709e38c2fcf0c3e0a9e77
Signed-off-by: Cosmo Chou <cosmo.chou@quantatw.com>
|
|
Remove fan configuration from minerva-phosphor-multi-gpio-presence.json.
Since the file is now empty, this patch deletes it and removes its
references from the .bbapend file.
Tested:
Related PR:
https://gerrit.openbmc.org/c/openbmc/entity-manager/+/82819
Change-Id: I99e947c7df085426dd04e301e93bffc1de9607d4
Signed-off-by: William de Abreu Pinho <williamdapinho@gmail.com>
|
|
- Unbind PCA9539 on removal and rebind PCA9539 after SLED insertion to
ensure VBUS enable is correctly controlled.
Tested on bletchley/bletchley15:
(1) remove and then insert bletchley15 sled3
(2) check VBUS_BMC_EN (bit1 of ioexp) is 0 when host is off
```
root@bmc:~# i2cget -f -y 3 0x76 0x0 | awk '{print and(rshift($1,1),0x1)}'
0
```
(3) check VBUS_BMC_EN (bit1 of ioexp) is 1 when host is on
```
root@bmc:~# i2cget -f -y 3 0x76 0x0 | awk '{print and(rshift($1,1),0x1)}'
1
```
Change-Id: Ife7061839dcb6672229f70083f419431797c4012
Signed-off-by: Cosmo Chou <cosmo.chou@quantatw.com>
|
|
- Replace unavailable dbus-monitor with busctl monitor.
Tested on bletchley/bletchley15:
```
Aug 25 04:11:07 bmc systemd[1]: Started SLED3 Present Manager.
Aug 25 04:11:07 bmc busctl[765]: Monitoring bus message stream.
Aug 25 04:37:02 bmc detect-sled-present[770]: SLED3 Removal
Aug 25 04:37:07 bmc detect-sled-present[770]: SLED3 Insertion
Aug 25 04:40:21 bmc detect-sled-present[770]: SLED3 Removal
Aug 25 04:40:32 bmc detect-sled-present[770]: SLED3 Insertion
```
Change-Id: I50069fa4baa696a6bd55318e885d228ebee79f44
Signed-off-by: Cosmo Chou <cosmo.chou@quantatw.com>
|
|
The logfiles are using invalid console id. Fix it.
Ex: use correct console id ath_uart1 instead of ath-uart1.
Change-Id: I863d3601e0d2470bdccee0306d851309534f072c
Signed-off-by: Prabha Veerubhotla <vvlprabha@gmail.com>
|
|
Add logic in monitor-pldm-sensor to detect Sentinel Dome entity-manager
object paths that include a retimer, preventing them from being
misclassified as an unknown configuration.
Motivation:
Enable the monitor-pldm-sensor script for Sentinel Dome when the FRU
includes a retimer, ensuring phosphor-pid-control does not start before
the PLDM sensor objects are fully exported on D-Bus, preventing
phosphor-pid-control from unnecessarily spending time and resources on
reloading, which may prolong the time spent in fail-safe mode where fan
PWM is set to 100%.
Change-Id: Ie1b1dfd87b29f00a85c2059300a96bda75b68d67
Signed-off-by: Eric Yang <eric.yang.wiwynn@gmail.com>
|
|
- Use loop for blocking service check, remove code duplication
- Replace "-z" with `systemctl is-active` to cover unknown states
- Move power sync into power-cmd functions for consistency
- Add initial-poweron-device.service for i2c mux rebind
- Add Conflicts in host-poweron@.service to avoid race conditions
Change-Id: Idfc778000bf3dc880b6c589482856992b476ee0a
Signed-off-by: Peter Yin <peter.yin@quantatw.com>
|
|
Remove phosphor-reboot-host@.service from SYSTEMD_SERVICE for the host
package.
Motivation:
phosphor-reboot-host@.service was already removed from
HOST_DEFAULT_TARGETS, but not from SYSTEMD_SERVICE.
This left an extra set-host-transition-to-off@.service running,
inflating host DC power state event counts.
Test plan:
- Build Code: pass
- Performed 432 DC power cycles accross blade hosts with no extra host
DC power state events: pass
Change-Id: I9ea7bb00d8900dc3ef55491625fbe40656176b8b
Signed-off-by: Eric Yang <eric.yang.wiwynn@gmail.com>
|
|
Add renice -10 for higher scheduling priority, preventing motor
control delays. e.g. too late to stop motor.
Change-Id: Ic301126d1747d78589a88dfbbdef906c65514de3
Signed-off-by: Cosmo Chou <cosmo.chou@quantatw.com>
|
|
The GPIO should be renamed to power-12v-memory-good.
Referencing an incorrect GPIO may result in host power-on failure.
This change is related to the upcoming DTS update.
'''
Lore: https://lore.kernel.org/all/20250822061454.2610386-1-peteryin.openbmc@gmail.com/
'''
Change-Id: I58508b0fb9b488124c60bf5fb6f6a985ac95a28f
Signed-off-by: Peter Yin <peter.yin@quantatw.com>
|
|
yosemite4 was facing some performance issue so we add lock to prevent
multiple simultaneous chassis cycles.
However, that makes it quite time-consuming when we cycling all slot at
the same time.
Change to allow two slot do chassis-powercycle simultaneous to decrease
time amount of cycling all slot.
Signed-off-by: Allen Wang <Allen_Wang@quantatw.com>
Change-Id: I45888b5f44e51e26e4c6816ac3cbcbdb1c40133d
|
|
Monitor the GPIO pins related to VR faults and apply the standard event
log format.
Change-Id: I4cbdac9055645585e1cfc383e54c9d0275d1eb4a
Signed-off-by: Fred Chen <Fred.Chen@quantatw.com>
|
|
Task Description:
Enhance the rescan-cxl-eid script by adding a check for
host DC power status before performing CXL EID rescanning.
This ensures rescanning only happens when the host DC power is on.
Motivation:
Previously, the script would proceed even if the host DC power was
not on. This could cause the host-poweron service to remain in the
activating state and block execution of DC on. With this fix,
rescanning will only happen when host DC power is on, improving
service state transitions and system stability.
Design:
- Added the check_host_status() function to check DC power via GPIO
before rescanning.
- The main loop now calls check_host_status before EID checks.
Tested:
- Check power state after power control - pass
- When DC power is off, the script logs the message, outputs CXL
EID status, and exits as expected.
- When DC power is on, the EID rescan proceeds normally.
Change-Id: Ib18b2f3d40fed4058cf07155198d54c5030e386b
Signed-off-by: Liora Guo <liora.guo.wiwynn@gmail.com>
|
|
# Description
Fixed MCTP configuration: Changed MTU (Maximum Transmission Unit) from a fixed 68 bytes
to the default setting. According to the MCTP specification, if the MTU is not explicitly
set, it defaults to the MTU of the network card.
# Motivation
According to the DFT specification, the WF CXL update time meets the requirement of completing
within five minutes. This modification reduces the CXL firmware update duration to under five minutes.
Change-Id: Ice3e80618ec8a821871288c4814dd671008127f0
Signed-off-by: Carter Chen <carter.chen.wiwynn@gmail.com>
|
|
Description:
1. Add policy-chassis-poweron.service to ensure that power policy is executed only after AC ON is completed (obmc-chassis-poweron target is Active)
2. Modify the dependency order after sled cycle
Original:
multi-user.target is active -> Power Policy (discover-system-state service) -> DC ON (require AC ON)
After:
multi-user.target is active -> AC ON -> AC ON target is active -> Power Policy -> DC ON
Motivation:
When executing slot AC OFF -> AC ON or sled cycle, AC ON may be triggered twice ,resulting in unexpected entries in the SEL.
Tested:
- Check power state and SEL count after power control - pass
Change-Id: I9a5e2f56ab2314079dcf91524bce7baa52d020cc
Signed-off-by: Zoey YJ Chung <zoey.yj.chung.wiwynn@gmail.com>
|
|
check_valid_sgpio: replace hardcoded gpioget calls with
get_gpio by name
Updated sgpio validation to use descriptive get_gpio calls:
- power-card-enable
- SLOT_ID_BCB_3
- power-pvdd33-s5
Change-Id: I429e1d3c1d3a38f4f8d0ceeca927e6ee92daacf2
Signed-off-by: Peter Yin <peter.yin@quantatw.com>
|
|
Add recovery log creation for both power rail and voltage regulator
faults when deassert events occur.
Change-Id: I5474750b5fee9a4b7b4f76650b3d6c28d3c85e1e
Signed-off-by: Peter Yin <peter.yin@quantatw.com>
|
|
Task Description:
Replace the deprecated dbus-send command with busctl in the
rescan-cxl-eid script. The dbus-send utility is no longer available
on systems using VIRTUAL-RUNTIME_dbus = "dbus-broker". This update
ensures continued functionality for slot type checking
during CXL EID rescanning.
Motivation:
The dbus-send utility is no longer available in the system when
using dbus-broker. Using busctl is the recommended replacement.
Design:
The script now uses busctl to call the ObjectMapper's GetSubTreePaths
method, replacing the previous dbus-send invocation. No additional
logic or parameters were changed, and the new command matches the
required D-Bus call for slot type detection.
Test Log:
- Verified that the slot type is correctly detected and that the
bypass logic for T1/T2 system operates as expected.
- All behaviors matched expectations and no regressions were observed.
Change-Id: I3280a95fd9cde08dabf9d053d53eaf5836d23bf2
Signed-off-by: Liora Guo <liora.guo.wiwynn@gmail.com>
|
|
Add udev rules to create persistent symlinks for ttyUSB devices
on SITv3 and Rainbow platforms. This ensures consistent device
naming regardless of enumeration order, making it easier for
services and scripts to reference the correct serial ports.
Change-Id: I0c6a1e0d0b022c2fa550f28d98894059ae63b468
Signed-off-by: Henry Wu <Henry_Wu@quantatw.com>
|
|
Switch to the new event format for PowerRailFault and
VoltageRegulatorFault.
The following example of GPIO pins log format uses the new event format
with Redfish MessageID support.
- OpenBMC_StatePower.1.0.PowerRailFault
- OpenBMC_StatePower.1.0.PowerRailFaultRecovered
- PWRGD_MTIA_BLADE1_PWROK_N
```
{
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/2",
"@odata.type": "#LogEntry.v1_9_0.LogEntry",
"AdditionalDataURI": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/2/attachment",
"Created": "2025-08-07T07:08:13.127+00:00",
"EntryType": "Event",
"Id": "2",
"Message": "A fault has been detected on power rail '/xyz/openbmc_project/inventory/system/board/Minerva_MTIA_Blade_1/pwrgd'.",
"MessageArgs": [
"/xyz/openbmc_project/inventory/system/board/Minerva_MTIA_Blade_1/pwrgd"
],
"MessageId": "OpenBMC_StatePower.1.0.PowerRailFault",
"Modified": "2025-08-07T07:08:21.703+00:00",
"Name": "System Event Log Entry",
"Resolution": "None.",
"Resolved": true,
"Severity": "Warning"
},
{
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/3",
"@odata.type": "#LogEntry.v1_9_0.LogEntry",
"AdditionalDataURI": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/3/attachment",
"Created": "2025-08-07T07:08:21.746+00:00",
"EntryType": "Event",
"Id": "3",
"Message": "A fault has cleared on power rail '/xyz/openbmc_project/inventory/system/board/Minerva_MTIA_Blade_1/pwrgd'.",
"MessageArgs": [
"/xyz/openbmc_project/inventory/system/board/Minerva_MTIA_Blade_1/pwrgd"
],
"MessageId": "OpenBMC_StatePower.1.0.PowerRailFaultRecovered",
"Modified": "2025-08-07T07:08:21.746+00:00",
"Name": "System Event Log Entry",
"Resolution": "None.",
"Resolved": false,
"Severity": "OK"
}
```
- OpenBMC_StatePower.1.0.VoltageRegulatorFault
- OpenBMC_StatePower.1.0.VoltageRegulatorFaultRecovered
- P12V_INA230_ALERT
```
{
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/2",
"@odata.type": "#LogEntry.v1_9_0.LogEntry",
"AdditionalDataURI": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/2/attachment",
"Created": "2025-08-07T07:13:53.719+00:00",
"EntryType": "Event",
"Id": "2",
"Message": "A fault has been detected on voltage regulator '/xyz/openbmc_project/inventory/system/board/Minerva_CMM/p12v-ina230-alert'",
"MessageArgs": [
"/xyz/openbmc_project/inventory/system/board/Minerva_CMM/p12v-ina230-alert"
],
"MessageId": "OpenBMC_StatePower.1.0.VoltageRegulatorFault",
"Modified": "2025-08-07T07:13:57.809+00:00",
"Name": "System Event Log Entry",
"Resolution": "None.",
"Resolved": true,
"Severity": "Warning"
},
{
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/3",
"@odata.type": "#LogEntry.v1_9_0.LogEntry",
"AdditionalDataURI": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/3/attachment",
"Created": "2025-08-07T07:13:57.853+00:00",
"EntryType": "Event",
"Id": "3",
"Message": "A fault has cleared on voltage regulator '/xyz/openbmc_project/inventory/system/board/Minerva_CMM/p12v-ina230-alert'.",
"MessageArgs": [
"/xyz/openbmc_project/inventory/system/board/Minerva_CMM/p12v-ina230-alert"
],
"MessageId": "OpenBMC_StatePower.1.0.VoltageRegulatorFaultRecovered",
"Modified": "2025-08-07T07:13:57.853+00:00",
"Name": "System Event Log Entry",
"Resolution": "None.",
"Resolved": false,
"Severity": "OK"
}
```
Change-Id: Ie4ed9eab5fbf20640585fc01d5cc9303eae704f8
Signed-off-by: Yang Chen <yang.chen@quantatw.com>
|
|
Remove:
- phosphor-reset-chassis-on@.service
- phosphor-reset-chassis-running@.service
- obmc-power-start@.service
- obmc-power-stop@.service
1. These services call mapper-wait@-org-openbmc-control-power* services,
but no corresponding /org/openbmc/control/power1 D-Bus object exists.
Mapper-wait waits indefinitely causing startup hang.
2. On this system PGOOD only indicates chassis standby power good.
phosphor-reset-chassis-running@.service uses PGOOD to create
/run/openbmc/chassis@%i-on, which causes
phosphor-discover-system-state@%i.service to be skipped. We want
discover to always run at BMC boot to evaluate the power-restore policy
and trigger host power-on (always-on).
Removing these services fixes the boot issue.
Change-Id: I87aafd2acf58e2605e010e9aab7ed821387e79d3
Signed-off-by: Fred Chen <Fred.Chen@quantatw.com>
|
|
Add support in Settings Manager to record version information
for the secondary Bianca board, specifically HGX_gpu and
HGX_inforom_gpu.
Change-Id: I6f92fda92032045fe4b010892e4599f9082802ec
Signed-off-by: Alex Wang <alex.ts.wang@fii-foxconn.com>
|
|
In the past, trigger fan threshold assert do not turn on amber led , but always turn on blue led.
After this change, trigger fan threshold assert should turn on amber led , and turn off blue led.
Here's the command to trigger fan threshold assert :
```
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan0_fault xyz.openbmc_project.Led.Group Asserted
b false
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan1_fault xyz.openbmc_project.Led.Group Asserted
b false
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan2_fault xyz.openbmc_project.Led.Group Asserted
b false
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan3_fault xyz.openbmc_project.Led.Group Asserted
b false
root@bmc:~# systemctl stop phosphor-pid-control.service
root@bmc:~# busctl set-property xyz.openbmc_project.FanSensor /xyz/openbmc_project/sensors/fan_pwm/FAN0_PWM xyz.openbmc_project.Sensor.Value Value d 0
root@bmc:~# busctl set-property xyz.openbmc_project.FanSensor /xyz/openbmc_project/sensors/fan_pwm/FAN1_PWM xyz.openbmc_project.Sensor.Value Value d 0
root@bmc:~# busctl set-property xyz.openbmc_project.FanSensor /xyz/openbmc_project/sensors/fan_pwm/FAN2_PWM xyz.openbmc_project.Sensor.Value Value d 0
root@bmc:~# busctl set-property xyz.openbmc_project.FanSensor /xyz/openbmc_project/sensors/fan_pwm/FAN3_PWM xyz.openbmc_project.Sensor.Value Value d 0
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan0_fault xyz.openbmc_project.Led.Group Asserted
b true
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan1_fault xyz.openbmc_project.Led.Group Asserted
b true
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan2_fault xyz.openbmc_project.Led.Group Asserted
b true
root@bmc:~# busctl get-property xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/fan3_fault xyz.openbmc_project.Led.Group Asserted
b true
```
Change-Id: I715c6ccbcfb45b4c384f169ac81261afdf2e7730
Signed-off-by: Eric Chiang Quanta <eric.chiang@quantatw.com>
|
|
This change updates the Clemente FSC table and related sensor configs:
* Add new sensors:
- BOOTDRIVE_TEMP_C
- PDB_VR_P12V_N1_TEMP1_C
- PDB_VR_P12V_N2_TEMP1_C
- VIRTUAL_FIOBOARD_TEMP_C
* Replace FIOBOARD_TEMP_C with VIRTUAL_FIOBOARD_TEMP_C
- VIRTUAL_FIOBOARD_TEMP_C is computed based on fan duty table
and calibration offset values.
These changes align fan control behavior with the new FSC
table and ensure proper thermal management with the added sensors.
Change-Id: I05159e11acc3ebfb2c5b88525c520246dd51a67a
Signed-off-by: Leo Wang <leo.jt.wang@fii-foxconn.com>
|
|
In the previous architecture, when a tray leak event occurs,
the PDB is responsible for controlling the HPM to shut down the host,
and the BMC only needs to send an event message.
However, in the new hardware architecture,
the interposer cannot directly control the HPM to shut down the host,
and it must rely on the BMC to notify the SCM CPLD to control the HPM
for host shutdown.
Tested: Verified host shutdown after receiving the tray leak event.
Change-Id: Ie4234b9eccdc05997100f839b5ae84cdc5068578
Signed-off-by: Kimi Chen <kimi.zy.chen@fii-foxconn.com>
|
|
Add a systemd drop-in for serial-getty@.service to configure
the serial console baudrate via stty before agetty starts.
The new drop-in file is installed into:
${systemd_system_unitdir}/serial-getty@.service.d/serial-getty@.conf
This ensures that serial-getty instances (e.g. ttyS3) will
have the correct baudrate (57600) applied consistently
during boot.
Change-Id: I82215199b197068fc45a2566f3d7440c01cf89ac
Signed-off-by: Peter Yin <peter.yin@quantatw.com>
|
|
Correct the event log message to accurately reflect the failure reason
when host power-on fails due to RMC main power being disabled.
Change-Id: Ib48491c84e6523cb4ab6e6f398e707ded2346341
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
Enable ARM-SBMR support for boot progress (postcode) recording.
Change-Id: I7608b45e003acde492709c57e0f9b5a0e0d1dc8e
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
On the Clemente platform, the backend NIC becomes available only after
the host is powered on. The CX8 TMP421 driver fails to bind during the
initial BMC boot following an AC power cycle due to the NIC being
inactive.
This update adds a service to bind the CX8 TMP421 driver once the host
is powered, ensuring the driver is correctly loaded.
Test result:
- All node drivers are bind successfully by backend-nic-driver-bind
```
root@bmc:~# systemctl status backend-nic-driver-bind
○ backend-nic-driver-bind.service - Bind backend NIC thermal driver
Loaded: loaded (/usr/lib/systemd/system/backend-nic-driver-bind.service; enabled; preset: enabled)
Active: inactive (dead) since Fri 2025-08-01 01:19:44 PDT; 5min ago
Duration: 40.720s
Invocation: 17026d86b8474da9b0585b80a9e6639a
Process: 1232 ExecStart=/usr/libexec/plat-svc/backend-nic-driver-bind (code=exited, status=0/SUCCESS)
Main PID: 1232 (code=exited, status=0/SUCCESS)
Mem peak: 1.3M
CPU: 399ms
Aug 01 01:19:03 bmc systemd[1]: Started Bind backend NIC thermal driver.
Aug 01 01:19:44 bmc backend-nic-driver-bind[1232]: Attempt 1...
Aug 01 01:19:44 bmc backend-nic-driver-bind[1232]: All devices successfully bind
Aug 01 01:19:44 bmc systemd[1]: backend-nic-driver-bind.service: Deactivated successfully.
```
- All backend NIC temperature sensors are available
```
root@bmc:~# mfg-tool sensor-display 2>/dev/null | table-sensor-display | egrep "IOB.*NIC.*"
IOB0_NIC0_TEMP_C ok 38 DegreesC N/A N/A 5 105 108 N/A
IOB0_NIC1_TEMP_C ok 39 DegreesC N/A N/A 5 105 108 N/A
IOB1_NIC0_TEMP_C ok 38 DegreesC N/A N/A 5 105 108 N/A
IOB1_NIC1_TEMP_C ok 37 DegreesC N/A N/A 5 105 108 N/A
```
Change-Id: I1656439bdb05d8a2df8455a9ab223fb4dbab91e6
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
On the Clemente platform, the OSFP FRU EEPROM becomes accessible once
the module's standby power is enabled. Therefore, it is not necessary to
bind the EEPROM driver during host power-on.
This change removes the osfp-eeprom-driver-bind.service from Clemente
limiting, its use to the Catalina platform only.
Change-Id: If9183ad22759b57dba7e85b8640a7058e69ca7d4
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
Aligning with the Catalina platform behavior, this change ensures that
missing second FPGA ready does not block the overall system bring-up
process.
Change-Id: I69db27aa437eeca3fdc002010f857db5a2b9ca68
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
On the Clemente platform, the OSFP FRU EEPROM becomes accessible only
after standby power is enabled.
This change adds the missing OSFP FRU device binding step to ensure it
is properly bound once standby power is available.
Change-Id: I7a4e5f31bfb3a7fe41d904a7d1ca6dd32458faea
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
Accessing the I2C bus connected to the CX8 while the host is powered off
may cause the I2C mux to become unresponsive.
To address this, introduce a reset_i2c_mux() function before binding the
I2C mux to ensure the driver probed correctly.
Additionally, call reset_i2c_mux() again after the mux is bound to avoid
potential hang issues, as the CX8 TMP421 driver binding is automatically
triggered when the kernel probes downstream devicesdefined in the device
tree.
Change-Id: Ib5a7e54274499bf4419f8658a35f33c9b665cba2
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
Fix incorrect line name in reset_host_usb() to ensure proper USB device
detection.
Test result:
```
root@bmc:~# lsusb -t
unable to initialize usb spec/: Bus 001.Port 001: Dev 001, Class=root_hub, Driver=ehci-platform/1p, 480M
|__ Port 001: Dev 002, If 0, Class=[unknown], Driver=hub/2p, 480M
|__ Port 001: Dev 003, If 0, Class=[unknown], Driver=cdc_ether, 480M
|__ Port 001: Dev 003, If 1, Class=[unknown], Driver=cdc_ether, 480M
|__ Port 002: Dev 008, If 0, Class=[unknown], Driver=hub/7p, 480M
|__ Port 001: Dev 009, If 0, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 009, If 1, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 009, If 2, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 009, If 3, Class=[unknown], Driver=cp210x, 12M
|__ Port 007: Dev 010, If 0, Class=[unknown], Driver=hub/7p, 480M
|__ Port 001: Dev 011, If 0, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 011, If 1, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 011, If 2, Class=[unknown], Driver=cp210x, 12M
|__ Port 001: Dev 011, If 3, Class=[unknown], Driver=cp210x, 12M
```
Change-Id: I4eeb8fd729ed2a51bf7969056679b323fca50a47
Signed-off-by: Potin Lai <potin.lai@quantatw.com>
|
|
Potin Lai (1):
remove group extension commands
Change-Id: I39012e1c8fd0c7d41e06a839b534ac7c6f22ce33
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Task Description:
Add HOST_SELECTOR entry in def_gpio.json to support the OCP debug card.
Motivation:
Without this entry, the Buttons service can't export or poll these GPIOs,
causing HostSelector to fail and preventing the OCP debug card from working.
Change-Id: I5a0b859ed3eddef967aa1bb680e922f46e6f1e90
Signed-off-by: Liora Guo <liora.guo.wiwynn@gmail.com>
|
