service-deps: remove logging as dependency

With the introduction of the following best practice:
https://github.com/openbmc/docs/commit/43421bdd06dece37b2acc4c656d7839bfedcc681

It was determined th

service-deps: remove logging as dependency

With the introduction of the following best practice:
https://github.com/openbmc/docs/commit/43421bdd06dece37b2acc4c656d7839bfedcc681

It was determined that phosphor-logging should be a d-bus activated
service. That is, this service will be activated when called so there is
no reason to specifically list it as a dependency in service files.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I8d871768bd3e515741825287bed4646a51c9b996

show more ...

scheduled-host-transition: make backward compatible

Similar to what was done with the host and chassis code, ensure the old
d-bus service name, xyz.openbmc_project.State.ScheduledHostTransition,
is

scheduled-host-transition: make backward compatible

Similar to what was done with the host and chassis code, ensure the old
d-bus service name, xyz.openbmc_project.State.ScheduledHostTransition,
is registered when the host instance is 0.

This ensures any code or tools that had hard coded the service name
prior to the instance based support continue to work

Tested:
- Confirmed with this change that both
xyz.openbmc_project.State.ScheduledHostTransition and
xyz.openbmc_project.State.ScheduledHostTransition0 work

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I1fb5ca622db01c931d70461d0ab4847c890b6df7

show more ...

doc: update README with some misc items

- Utilize abbreviation (PSM) when not starting a sentence with it
- Add details of new supported states
- Specify targets utilized to control each objects sta

doc: update README with some misc items

- Utilize abbreviation (PSM) when not starting a sentence with it
- Add details of new supported states
- Specify targets utilized to control each objects state

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ib9ffb107f4e86b1fe34f07b73116f26eec3cb9d3

show more ...

markdownlint: fix all warnings

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I254f27e1b0a7840b1f119d589d71a4cf4b166425

prettier: re-format

Prettier is enabled in openbmc-build-scripts on Markdown, JSON, and YAML
files to have consistent formatting for these file types. Re-run the
formatter on the whole repository.

prettier: re-format

Prettier is enabled in openbmc-build-scripts on Markdown, JSON, and YAML
files to have consistent formatting for these file types. Re-run the
formatter on the whole repository.

Change-Id: Id3397e8757554401055d3a5c921ce9d3eee3ee82
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>

show more ...

beautysh: re-format

beautysh is enabled in the openbmc-build-scripts on Bash/Zsh/POSIX-sh
files to have a consistent formatting. Re-run the formatter on the
whole repository.

Signed-off-by: Patric

beautysh: re-format

beautysh is enabled in the openbmc-build-scripts on Bash/Zsh/POSIX-sh
files to have a consistent formatting. Re-run the formatter on the
whole repository.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I8c4f46426087652161d7cbeab489bfe32d25f17e

show more ...

treewide: Remove uses of bind

Bind is less compatible with function wrappers like function2 and more
terse to read.

Change-Id: I34474b71758db1ffee301729c585c4d555df2064
Signed-off-by: William A. Ke

treewide: Remove uses of bind

Bind is less compatible with function wrappers like function2 and more
terse to read.

Change-Id: I34474b71758db1ffee301729c585c4d555df2064
Signed-off-by: William A. Kennington III <wak@google.com>

show more ...

Support chassis power cycle

A new system is being designed which will utilize phosphor-state-manager
and it requires the chassis system power cycle feature.

Test plan:
Change property "RequestedPow

Support chassis power cycle

A new system is being designed which will utilize phosphor-state-manager
and it requires the chassis system power cycle feature.

Test plan:
Change property "RequestedPowerTransition" to trigger
chassis power cycle target.

Change-Id: I30b0efd7c0fc72a9fb378f48e47c4c1af2b22ad6
Signed-off-by: Bonnie Lo <Bonnie_Lo@wiwynn.com>

show more ...

improve new chassis power services

A few miscellaneous changes to improve descriptions and update where
these services are actually installed by the bitbake recipes.

Signed-off-by: Andrew Geissler

improve new chassis power services

A few miscellaneous changes to improve descriptions and update where
these services are actually installed by the bitbake recipes.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ic4471edcc589cd28f153a708d53f5a9b53fa2913

show more ...

move chassis power services to phosphor-state-manager

Move the generic services which are used to power on and off the chassis
into phosphor-state-manager. This will make the use of
phosphor-state-m

move chassis power services to phosphor-state-manager

Move the generic services which are used to power on and off the chassis
into phosphor-state-manager. This will make the use of
phosphor-state-manager for chassis power control more straightforward.

These services are directly copied from openbmc/openbmc/
meta-phosphor/recipes-phosphor/chassis/obmc-op-control-power/ and
renamed to match with the existing naming conventions.

The only change to these files was if a Wants/Before/After service name
was changing as a part of this.

A follow on commit will improve some of the Description fields.

The following email has more details on this change:
https://lists.ozlabs.org/pipermail/openbmc/2022-November/032457.html

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I35d555139897cf58a26e36d1558ab7f5225d34bd

show more ...

update host state documentation

Add the new transitioning to on and off states that are now supported by
the host state software.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I

update host state documentation

Add the new transitioning to on and off states that are now supported by
the host state software.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ifced2625546f365f1dd4f5888b8c702e314f2758

show more ...

hypervisor: add to documentation

The optional hypervisor package was added recently to
phosphor-state-manager. Add some basic documentation on it and what it
supports.

Signed-off-by: Andrew Geissle

hypervisor: add to documentation

The optional hypervisor package was added recently to
phosphor-state-manager. Add some basic documentation on it and what it
supports.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I8fb31a8f49599c6c664493856de7da7d246a8bfb

show more ...

obmc-host-graceful-quiesce: ensure starts synch targets

As seen in other scenarios, the power off synchronization targets do not
always start in time when the obmc-host-graceful-quiesce@.target is
s

obmc-host-graceful-quiesce: ensure starts synch targets

As seen in other scenarios, the power off synchronization targets do not
always start in time when the obmc-host-graceful-quiesce@.target is
started. This can result the host being stopped before it has had the
chance to gracefully shut itself down.

Ensure the synchronization targets are started by adding them into the
target with 'Wants' relationships.

Tested:
- Booted system into hostboot, manually started the target and verified
via the journal that the pldm-softpoweroff application completed prior
to instructions stopping.

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Iede5194c24ad05467235419eaf0f343bccdffc1a

show more ...

only quiesce in reboot on host crash

This script was introduced to ensure a system properly halted in
scenarios where the host has crashed (obmc-host-crash@.target). There
was a loophole where the s

only quiesce in reboot on host crash

This script was introduced to ensure a system properly halted in
scenarios where the host has crashed (obmc-host-crash@.target). There
was a loophole where the standard reboot attempt logic check was not
running in this specific path.

Unfortunately, that original change was not done quite right. This
script is actually called on all host reboots so we have run into
situations where the host is being put into quiesce before being given
one last try to boot. That last try in non-host crash scenarios is where
certain recovery actions (like trying alternate boot sources for some
devices) is done.

This commit modifies the script to ensure it only moves the host to the
quiesce state when the cause of the host reboot was a host crash.

Tested:
- Verified in normal reboot path, even if AttemptsLeft is 0, the reboot
is still done
- Verified that if AttemptsLeft is 0 and reboot reason is HostCrash that
host was put in quiesce

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I18e21cc33b3c38334ad760148a0754e2bd7d0780

show more ...

call new obmc-chassis-blackout@.target on blackout

The following commit introduced a new systemd target to be called when
OpenBMC firmware detects a power blackout has occurred (system
unexpectedly

call new obmc-chassis-blackout@.target on blackout

The following commit introduced a new systemd target to be called when
OpenBMC firmware detects a power blackout has occurred (system
unexpectedly lost all power to system):
https://gerrit.openbmc.org/c/openbmc/docs/+/54204

This target is now in place and the appropriate services have been
installed into it. Switch the state-manager code over to call this new
target.

Tested:
- Verified in simulation that the target was called when the system
experienced a blackout.

From a powered off state.
root@p10bmc: systemctl stop xyz.openbmc_project.State.Chassis.service

Manipulated file to indicate that the power was on before service ended
root@p10bmc: vi /var/lib/chassStateChangeTime

root@p10bmc: systemctl start xyz.openbmc_project.State.Chassis.service
Jun 22 17:05:37 p10bmc phosphor-chassis-state-manager[1155]:
The system suffered a power blackout.

root@p10bmc: systemctl status obmc-chassis-blackout@0.target
Active: active since Wed 2022-06-22 17:05:37 UTC; 8s ago

root@p10bmc: systemctl status phosphor-reset-state-sensors@0.service
Jun 22 17:05:37 p10bmc systemd[1]: Finished Reset host sensors.

root@p10bmc: systemctl status obmc-led-group-stop@power_on.service
Active: active (exited) since Wed 2022-06-22 17:05:37 UTC; 8s ago

Signed-off-by: Corey Hardesty <corey.hardesty@icloud.com>
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: If95f8da52ec024e699204f4b26784c7457da5abf

show more ...

power-restore: always honor AlwaysOn policy

An internal bug was opened recently noting that when the
only-run-apr-on-power-loss config option was set, the AlwaysOn policy
was not executed. This didn

power-restore: always honor AlwaysOn policy

An internal bug was opened recently noting that when the
only-run-apr-on-power-loss config option was set, the AlwaysOn policy
was not executed. This didn't make sense because the AlwaysOn policy is
specifically saying to always power the system on. This commit changes
the logic to always power on, no matter what, if the user has set the
policy to AlwaysOn.

This change fixes a subtle bug in the one-time logic as well. Prior
to this change, the one-time setting would always be executed even if
the only-run-apr-on-power-loss config option was set. The reasoning
behind the bug was that the one-time did need to always be executed,
when it was set to AlwaysOn. The one-time is used in situations where
host or BMC firmware needs to reboot the BMC but ensure the system boots
back up after the BMC reboot (inband code update, brownout conditions).
This commit ensures we continue to execute that behavior but also
ensures we do not run the one-time otherwise when
only-run-apr-on-power-loss is set and there was not an AC loss.

Tested:
- Verified AlwaysOn always powered on system
- Verified power was only "Restored" when there was an AC loss

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ic5ba84cf2d4fe80fc60ebf9440520a80d91c9ed9

show more ...

host-reboot: add host instance in get_reboot_count()

Add multi-host support to get_reboot_count()

Signed-off-by: Potin Lai <potin.lai@quantatw.com>
Change-Id: I77c00273bfcf4a45a6e24ac8e1a8540f2d609

host-reboot: add host instance in get_reboot_count()

Add multi-host support to get_reboot_count()

Signed-off-by: Potin Lai <potin.lai@quantatw.com>
Change-Id: I77c00273bfcf4a45a6e24ac8e1a8540f2d609e04

show more ...

MAINTAINERS: remove file

The MAINTAINERS file is deprecated in favor of OWNERS.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Id93e95a004abcf04838c7512e8c1de9b92b2b0ab

Ensure POR reset is logged

In situations where the last reboot cause is not changed to POR,
specifically after unexpected AC losses, bmc-state-manager fails to
update the property. Therefore after a

Ensure POR reset is logged

In situations where the last reboot cause is not changed to POR,
specifically after unexpected AC losses, bmc-state-manager fails to
update the property. Therefore after a power cycle the last reboot
cause property is set to unknown. Thankfully, chassis-state-manager
generates a file indicating an AC loss occurred and that can be used
to identify that the last reboot cause was a power-on-cycle. It was
necessary to add an "After xyz.openbmc_project.State.Chassis"
dependency to bmc-state-managers service file to ensure the AC loss
was generated prior to bmc-state-manager executing.

Tested:

- Ensured that when the system was power cycled that the last reboot
cause was set to "POR" when the chassis lost power file was present.

- Verified that when the file was not present the last reboot
cause was set accordingly (typically unknown)

Change-Id: Ifa6d5d6f89c221477ee2f2fa1ed15adade359e45
Signed-off-by: NodeMan97 <corey.hardesty@icloud.com>

show more ...

Only run power restore on AC losses

Added a meson build option that allows the customer to enable automatic
power restore policy only for AC losses where the chassis power was
on prior to a BMC rese

Only run power restore on AC losses

Added a meson build option that allows the customer to enable automatic
power restore policy only for AC losses where the chassis power was
on prior to a BMC reset(and off when it comes back online).

The default is set to 'false' which provides the standard behavior of
APR which is to always run APR on an BMC reboot.
If the option 'only-run-apr-on-power-loss' is 'true' then APR is only
run in the case of the chassis losing power. AC power losses trigger the
creation of a file called 'chassis@0-lost-power' that identifies an
AC power loss has occurred. This generated file can be found in the
directory '/run/openbmc/'.

Tested:
- Verified that when built with option 'only-run-apr-on-power-loss'
set to false(default) that APR was ran.
- Verified that when built with option 'only-run-apr-on-power-loss'
set to true that apr was not ran when a power loss did not occur.
- Verified that when built with option 'only-run-apr-on-power-loss'
set to true that apr was only ran when a power loss occured.

Jun 08 01:55:13 p10bmc phosphor-chassis-state-manager[3009]:
The system suffered a power blackout.

root@p10bmc:/lib/systemd/system# ls /run/openbmc/
chassis@0-lost-power

Signed-off-by: Corey Hardesty <corey.hardesty@icloud.com>
Change-Id: Iac37761416d699e3a469936e00868a8d9b6dfdb1

show more ...

Utilization of power restore delay for APR

There was an issue where APR would start too early in the boot
process causing conflicts between the BMC and host firmware on
the i2c buses. Waiting to run

Utilization of power restore delay for APR

There was an issue where APR would start too early in the boot
process causing conflicts between the BMC and host firmware on
the i2c buses. Waiting to run APR ensures the BMC reaches a
READY state and that the vpd collection has completed.
This commit adds a delay based off of the value of
'PowerRestoreDelay' a Dbus property hosted by the RestorePolicy
interface. The system owners may override this value which is
initially defaulted to 0 via an override in their meta layer.

Tested:
- Ran an AC cycle of the system and restarted phosphor-discover-
system-state@0.service.

- Ensure that the bmc reboot was delayed by the value
specified by PowerRestoreDelay.

- Verified through on real hardware. Simulation environment
has system noise that create unpredictable delay results
with delays over 15 seconds. The delays in simulation always
adhere to the property, but take additional time to boot.

- PowerRestoreDelay at 30 seconds (ran 5 separate times)
systemctl stop phosphor-discover-system-state@0.service

Jul 22 17:47:23 p10bmc phosphor-discover-system-state[2224]:
power_policy=Restore, powering host on (30s delay)

Jul 22 17:47:53 p10bmc systemd[1]: Finished Reboot If Enabled

5/5 tests were delayed exactly 30 seconds.


- PowerRestoreDelay at 100 seconds

Jul 22 17:52:45 p10bmc phosphor-discover-system-state[2224]:
power_policy=Restore, powering host on (100s delay)

Jul 22 17:54:25 p10bmc systemd[1]: Finished Reboot If Enabled.

1/1 test(s) were delayed exactly 100 seconds.


- PowerRestoreDelay at 0 seconds

Jul 22 18:02:51 p10bmc phosphor-discover-system-state[2271]:
power_policy=Restore, powering host on (30s delay)

Jul 22 18:02:51 p10bmc systemd[1]: Finished Reboot If Enabled.


Signed-off-by: Corey Hardesty <corey.hardesty@icloud.com>
Change-Id: I69f7f9f3a9e638defcbee2bbb8dda6d9c4b7f0ae

show more ...

Power restore from last state after AC Loss

In situations where the power restore policy was set to Restore from
the last known state (Restore) and the requested host transition was
not denoted as "

Power restore from last state after AC Loss

In situations where the power restore policy was set to Restore from
the last known state (Restore) and the requested host transition was
not denoted as "On" would lead to the host never powering back on.
For example, this occurs during a fresh boot of a system or after
completing a system update where other RequestedHostTransition
properties may be set such as 'GracefulWarmReboot'. This lead to the
old logic in discover system state to never power the host back on
which means there was a potential that users would have to manually
power back on the host. This change inverts the logic, essentially
checking to ensure the 'RequestedHostTransition' was not set to off
and therefore indicating that power should be restored in all other
cases (per the interface documentation).

Tested:
- All testing was with PowerRestorePolicy as 'Restore'
The corresponding gui option is 'LastState'.

- Applied changes and ensured that power was restored to
the host with 'GracefulWarmReboot' set as the desired
host transition property.

- Set 'RequestedHostTransition' property to "Reboot",
"GraceFulWarmReboot", "ForcedWarmReboot" and verified
that chassis power state was restored to 'On' after
a power cycle.

- Confirmed that with any of the transition properties
listed above set that the 'RequestedHostTransition'
state switched to on after reboot.

- Verified that power state remained off when the
'RequestedHostTransition' property was set to 'Off'.
Also that this property remained 'Off' after reboot.


Signed-off-by: Corey Hardesty <corey.hardesty@icloud.com>
Change-Id: Ie582bf16b51a4cbc0c3c8b3016e3ed9e5f59cb80

show more ...

OWNERS: switch 'matches' to 'matchers'

The original OWNERS template had a mistake which used 'matches' instead
of the field supported by the Gerrit plugin 'matchers'. Update the
OWNERS file to have

OWNERS: switch 'matches' to 'matchers'

The original OWNERS template had a mistake which used 'matches' instead
of the field supported by the Gerrit plugin 'matchers'. Update the
OWNERS file to have the correct field.

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I8f9504a48c331e344127d8cf7ace2aba957eda63

show more ...

sdbusplus: use shorter type aliases

The sdbusplus headers provide shortened aliases for many types.
Switch to using them to provide better code clarity and shorter
lines. Possible replacements are

sdbusplus: use shorter type aliases

The sdbusplus headers provide shortened aliases for many types.
Switch to using them to provide better code clarity and shorter
lines. Possible replacements are for:
* bus_t
* exception_t
* manager_t
* match_t
* message_t
* object_t
* slot_t

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I41103c892db258a85640aa1442acd9a295c8a847

show more ...

host-reboot: enforce host boot count on host crash

The host reboot path can be triggered via paths outside of the normal
RequestedHostTransition property set. For example, if the host crashes,
it ju

host-reboot: enforce host boot count on host crash

The host reboot path can be triggered via paths outside of the normal
RequestedHostTransition property set. For example, if the host crashes,
it just triggers an automatic reboot using the systemd targets, which
means the check to prevent continuous host reboots is not run.

This commit ensures we decrement the reboot counter when a host crash
occurs, and it also ensures we check that count when deciding if a
reboot should occur.

While testing, it was found when we end in the host Quiesce state (after
all reboot attempts have expired), the host stayed in Quiesce, even
after issuing a power off. The fix for that was to ensure the
obmc-host-quiesce@.target conflicted with obmc-host-stop@.target to
ensure it would properly run on the power off request and update the
host state to Off.

Tested:
- Verified that AttemptsLeft was properly decremented on host crashes
- Verified that when AttempsLeft reached 0, the reboot was halted
and the host state was moved to Quiesced
- Verified that once in Quiesced, system could be powered off, and a
boot worked fine (and AttempsLeft was back to 3)

Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I8d101b987517e160becc712ed70950f1d3204397

show more ...