Age | Commit message (Collapse) | Author | Files | Lines |
|
Add a sysctl configuration file to direct sysctl to enable the
Linux kernel's panic on OOM option.
Tested on Everest hardware:
root@ever10bmc:~# cat /proc/sys/vm/panic_on_oom
2
Signed-off-by: Eddie James <eajames@linux.ibm.com>
Change-Id: Iaf6462c8911ee15deafe5a4fc4bc51f9db864afd
|
|
Don't use /usr/bin/env, it's an antipattern.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: If388f5427bb2ca45718d0ca99452bf3f147861dd
|
|
/usr/bin/env is now an antipattern.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I7e8488b984db0258f68107b23a0e3a7ed149f763
|
|
It's time to require this for p10 based machines now
Tested:
- Confirmed by default it is now "Required"
root@p10bmc:~# pldmtool bios GetBIOSAttributeCurrentValueByHandle -a hb_tpm_required
{
"CurrentValue": "Required"
}
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I0959ffad3d4a38b542087be4a6ca6949ff1b4356
|
|
The dbus sink actions will cause phosphor-dump-manager to take a full
BMC dump if the debug action is received on the trigger source.
Change-Id: If57612a75ca7f68ed2636c9b23dc04972d99626d
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
For the same reason as supply-monitor (a dependency on org.open_power
dbus interfaces).
Signed-off-by: Zev Weiss <zev@bewilderbeest.net>
Change-Id: Ibcf52fc180da744cced04d4720036d8bc84bf89c
|
|
This is the recommended keyword to use and it appears to fix a subtle
bug of some sort where the new secure-check is overriding other
features.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ic724ec7611741d6286926e96cb0b61530cdcf638
|
|
IBM manufacturing has requested this new feature be enabled on p10bmc
machines.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I47fcc24667dfcccc4314548b29772227280eef4c
|
|
Adriana has been a long time contributor to IBM and will make a great
co-owner.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I5542dc1e454709a5d89f899da3ba50bb30f8eb11
|
|
This configures the distro name printed at system boot to indicate that
p10bmc is a combination of OpenBMC and custom components.
Change-Id: I130ce058244bd989a36855000cb4412215a74015
Signed-off-by: Joel Stanley <joel@jms.id.au>
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I746b04e1e36abaf2ff3852beb8a04535692442c1
|
|
Patrick Williams (1):
build: enable C++20
Change-Id: Ib2bc73d1f7227f9a8494598be123eac66c2f8ace
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
This was added with cde0f094f for libseccomp, but shortly after that
upstream moved libseccomp from meta-security to core (241c7d2e6). As
such, meta-security is no longer used or required.
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Change-Id: I371e54b11f6336720dfc6edf0ef733d22b7fb4f2
|
|
phosphor-dbus-monitor still uses autoconf and has a hard dependency on
phosphor-snmp, causing phosphor-snmp to be brought in on systems that
have no interest in using SNMP. The vast majority of systems do not
supply any config for phosphor-dbus-monitor, so the usage of it does not
seem wide enough to enable the IMAGE_FEATURE by default.
- Remove 'obmc-dbus-monitor' from IMAGE_FEATURE by default.
- Clean up all systems that manually remove the dbus-monitor feature or
package.
- Explicitly add the 'obmc-dbus-monitor' to the following systems:
- meta-ibm: all
- meta-inspur: fp5280g2
- meta-quanta: s6q
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Iad6c4104466e21d1b9843dac964a634019eb6935
|
|
This commit enables the hypervisor n/w configuration service
for the IBM systems and when it is enabled, the service file
gets created that starts after the bios manager service.
Tested By:
Service file is created for IBM machines and the service
is up and running when the image is flashed onto the system
Signed-off-by: Asmitha Karunanithi <asmitk01@in.ibm.com>
Change-Id: Iac6d4a52df2a85fa3c8b60cc42fa6905337769de
|
|
1) Move ibm specific plugins from recipe to tools/dreport.d/ibm.d/plugins.d
in the repository
2) Move openpower specific plugins from recipe to
tools/dreport/opepower.d/plugins.d in the repository.
4) Modified to install the plugins from the machine specific
plugin folder from the repository
Reference links
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-debug-collector/+/52852/
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-debug-collector/+/52479
Includes:
phosphor-debug-collector: srcrev bump 02634e5531..124d31f09c
Marri Devender Rao (3):
openpower:dreport: add guarded resolved and ephemeral records to BMC dump
ibm:dreport add vpd persistent data to BMC dump
dreport: Fix network and timedate plugin failures during BMC dump generation
Signed-off-by: Marri Devender Rao <devenrao@in.ibm.com>
Change-Id: I0bd350b9b564c4e98a6765902de76274f90500e4
|
|
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I2887152b9451e8f0d812ba6cc5cf20f0ab61f38a
|
|
Per [1][2], Yocto is starting to deprecate ABORT and has replaced
the "ABORT" action in BB_DISKMON_DIRS entries with "HALT".
1. https://wiki.yoctoproject.org/wiki/Inclusive_language
2. https://git.yoctoproject.org/poky/commit/?id=4f77505d94a8f6260933f457e9848d1d2fa98ce5
Tested:
Built obmc-phosphor-image successfully and eliminate the following
warnings:
```
WARNING: The BB_DISKMON_DIRS "ABORT" action has been renamed to
"HALT", update configuration
```
Signed-off-by: George Liu <liuxiwei@inspur.com>
Change-Id: If57d0ded9fac41e23b31b01e2a0e309ac7388148
|
|
Patrick Williams (1):
sdbusplus: object: don't use 'bool' argument constructor
Change-Id: I59cee1db5bd10776d6f49ad58056e008d7409fa7
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
phosphor-state-manager: srcrev bump 1ab2b6cea6..211d972d75
Adriana Kobylak (1):
chassis: Monitor the PowerSystemInputs for power status
Allen.Wang (2):
host-state-manager: Add multi-host support
chassis-state-manager: Correct poweroff state Unit
Andrew Geissler (2):
ensure bmc-state-manager stays in quiesce state
generate bmc dump on monitored systemd target fail
Brandon Wyman (1):
Correct pinhole reset logic for chassis state
Mike Capps (1):
Set Critical severity for Blackout errors
Patrick Williams (7):
filesystem: use non-experimental version
meson: remove stdc++fs dependency
meson: add subproject wraps for current dependencies
sdbusplus: object: don't use 'bool' argument constructor
meson: add missing dependencies and wraps
crit-service: update defaults for templated Host/Chassis
sched-host-tran: support multi-host
Potin Lai (3):
chassis-state-manager: Add multi-chassis support
discover-system-state: add multi-host support
host-state-manager: fix host0 failure with reading wrong object path
shamim ali (1):
power-on-hours: ensure not incremented when system off
Some commits went into phosphor-state-manager support that starts the
framework to support multiple chassis and hosts within a system. This
commit templates both the host and chassis services.
Change-Id: I19764aa0c625d4d28d8d0e672658f1a3910335af
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
Since P9 systems need only to stop the link failover, and P10 systems
need only to set the gigabit link speed, split the nsci-netlink
service into two and make it per-interface. Use the systemd network
check to make sure the link is up first. Use a helper script to fetch
the ifindex for the network interface
Signed-off-by: Eddie James <eajames@us.ibm.com>
Change-Id: I6558787a2234575b528864bb71705964a63c0498
|
|
The powercap range is currently hardcoded by mrw-override-settings.yaml
This commit will remove the hardcoded range and allow the full range of
values to be set.
The actual limits are available via dbus which can be used to validate
the user power cap.
Change-Id: I71e6009d5bc288c0d8480db26ce9e6782bdbd167
Signed-off-by: Chris Cain <cjcain@us.ibm.com>
|
|
New systems should be using the 'use-json' model by default, but it
currently requires a bbappend in order to enable it. Automatically
detect based on the PREFERRED_PROVIDER of virtual/...-config-native if
the intended use is JSON or YAML.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Ie674d1b901395b4fd41838b6e978ca1ff00b0277
|
|
Palmetto and Romulus were using phosphor-webui. phosphor-webui built on
AngularJS, and AngularJS goes End of Life December 31, 2021.
So Move Plametto and Romulus to use webui-vue.
webui-vue is a modern replacement for phoshor-webui. Weui-vue is built
on vue.js, uses Redfish, and has additional features like
- Ability to easily theme to meet brand guidelines
- Accessibility
- Full Redfish
- Improved user experience based on user feedback
- Language translation-ready
- Modern front-end framework with an active community and future
development roadmap
Tested: manually tested on Palmetto and Romulus system. Able to log in
into the system using the latest webui-vue. Able to retrieve
dump log and another log, all basic functionality is working.
Signed-off-by: Abhishek Patel <Abhishek.Patel@ibm.com>
Change-Id: I883a84552383f3d7f94d9227a11fc7410bc43ec5
|
|
Add the power supply temperature readings to the hwmon configuration and
the Rainier and Everest assocation files.
Signed-off-by: Brandon Wyman <bjwyman@gmail.com>
Change-Id: I00e7fde4a58da553b9374733b22b8608d0c9da41
|
|
This takes the default critical services list from meta-phosphor and
appends on extra services IBM is interested in.
Ideally, there would be an easy way to just append an extra file to the
command line but this is going to need some more thought. For now, just
override the entire file.
Tested:
- Booted witherspoon up in QEMU and verified new json file was loaded
and that stopping one of the new critical services resulted in a
dump, error log, and BMC Quiesced state.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Icf5db5331072cbcbd0359e0b789b664e31423124
|
|
This commit adds p10bmc specific configuration files to display
CPU and DIMM FRU information. Without these configuration files
the default phosphor recipe configuration files are used, and we
do not get FRU info for all configured CPUs and DIMMs.
The changes are meta-ibm specific, the image was built and tested
against our simulation model.
Signed-off-by: Shantappa Teekappanavar <sbteeks@yahoo.com>
Change-Id: I60dd8ce05941105dfaf6b5e67fa8c7b0bb06534a
|
|
* Deprecate N-1 release (hardknott).
* Enable N+1 release (kirkstone).
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I39e027e02dab64b4390b46ffbd9c299c858f403e
|
|
The power_mode and power_ips dbus objects are being consolidated and
moved under the org.open_power.OCC.Control service.
Their default values will now be set via entity-manager.
Verified on Everest hardware.
Change-Id: Ibfa1e55126862db3521855ad0743a1423e555dac
Signed-off-by: Chris Cain <cjcain@us.ibm.com>
|
|
Providing `opkg` in the image allows us to easily install packages
after-the-fact, usually to debug systems in the lab.
This is not a mechanism for providing firmware updates.
No specific configuration is supplied to enable its use, and hosting of
package archives is left as an exercise to the user. That said, bitbake
generates its own package archive in the course of creating the output
image. It also provides a `package-index` target, which after execution,
allows you to expose the archive with no more work than:
```
$ python3 -m http.server --directory tmp/deploy/ipk 8080`
```
The addition of opkg with the chosen options increases the uncompressed
filesystem size by 3855360 bytes (~3.68MiB).
Change-Id: I7731a4ebe08249e2216e8ca1984f8e8d87558687
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
The MAINTAINERS files are no longer used and are out of date (references
to IRC still exist in most of them). Remove them and rely on the OWNERS
exclusively.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I0f5d2719ad97e218ce03c4503efe1b1c92ac455e
|
|
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I7a4e86c61417f006e5ba218e3e5862dce3cd9e9a
|
|
e.g.
WARNING: /home/peter/openbmc/meta-ibm/meta-romulus/recipes-phosphor/occ/romulus-occ-control-config-native.bb: QA Issue: romulus-occ-control-config-native: native/nativesdk class is not inherited last, this can result in unexpected behaviour. Classes inherited after native/nativesdk: openpower-occ-control.bbclass [native-last]
WARNING: /home/peter/openbmc/meta-ibm/meta-romulus/recipes-phosphor/logging/romulus-phosphor-logging-callouts-native.bb: QA Issue: romulus-phosphor-logging-callouts-native: native/nativesdk class is not inherited last, this can result in unexpected behaviour. Classes inherited after native/nativesdk: phosphor-logging.bbclass [native-last]
Signed-off-by: Peter Foley <pefoley@google.com>
Change-Id: I31be11e264e1701215d2eadd28dc2249153a5d61
|
|
For PowerVM, Mutual TLS has no current use case. Disabling mTLS auth
by setting meson build option to disabled.
Note, bmcweb can still use mTLS auth to connect with LDAP server;
however for this use case the bmcweb user must auth in another way.
Signed-off-by: Ali Ahmed <ama213000@gmail.com>
Change-Id: I35adaeb1b6318d5f1de619509e05ae49970d8486
|
|
Match on everything under meta-ibm/recipes-kernel.
Change-Id: I56a2b578d1cb05ab7a2f3be9bb434a695e57b4f3
Signed-off-by: Joel Stanley <joel@jms.id.au>
|
|
It's been a pleasure working with the community on these bitbake layers
over the years. The other maintainers have been doing all the work for
awhile now so it's time to make it offical.
Change-Id: I0f8390417294cb82575aba87b5c0e7bd7528d4c1
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
As mentioned by I970e06ca6f9d0a9792af122eb25da1bb9a06f058, append and
remove directives with '+=' are no longer valid Yocto syntax and raise a
warning.
See https://git.yoctoproject.org/poky/tree/documentation/migration-guides/migration-3.5.rst#n40
One interesting caveat has to do with the difference between "append"
and "+=".
foo:append = "a"
foo:append = "b"
Results in `foo = "ab"`, but
foo += "a"
foo += "b"
Results in `foo = "a b"`. When `:append +=` is used it has behavior
like the `+=` operator. Therefore, in some cases we need to insert
additional whitespace after the `:append = "` to ensure concatenation
behaves as expected. I've manually reviewed the results to ensure there
is no ill side-effects from an additional space potentially being added.
Fix up the entire tree with the following one-liner (and similar for
`remove`):
git ls-files -- ':!:poky/**' ':!:meta-security/**' \
':!:meta-raspberrypi/**' ':!:meta-openembedded/**' \
| grep "\.bb" | xargs grep -l ":append.*+=" \
| xargs sed -i 's/:append\(.*\)+=\([^"]*\)" */:append\1=\2" /'
or ...
| xargs grep -l ":remove.*+=" \
| xargs sed -i "s/:remove\(.*\)+=/:remove\1=/"
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Iea6235e7c76e252f4d473652957c2925f6b8845a
|
|
In order to disambiguate from other uses of the FMT variable name,
rename to OBMC_CONTROL_POWER_FMT in obmc-op-control-power bit bake
files. Change was suggested in the following review:
https://gerrit.openbmc-project.xyz/c/openbmc/openbmc/+/50765
Signed-off-by: Jim Wright <jlwright@us.ibm.com>
Change-Id: If74dc281b8dfa9c06b86012f010d5b5a40d1a0f6
|
|
Some addition development details must be managed in this bbappend due
to co-development of the AST2600 and IBM p10bmc designs. IBM did bringup
of secure-boot on pre-production AST2600 silicon and this shaped how the
platform's OTP was configured.
The PEM files represent the public portion of the IBM signing key-pairs
for p10bmc systems. These are included to provide a canonical location
for the production OTP image artifact.
Change-Id: I7caa6cfd5848b1d671ef95f8031b76088673900a
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Don't support the hardware root-of-trust on Tacoma boards.
Change-Id: I144d3bf445063f6341ae1958611873f79c9c00c7
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Make the power supply sensors visible via redfish, chassis, all_sensors.
Input voltage, input power, output voltage, output current.
Signed-off-by: Brandon Wyman <bjwyman@gmail.com>
Change-Id: I79c9a07c849ad8015fb9fa42242aeab932b755c9
|
|
Add the power-supply configuration files for the ibm-cffps power
supplies used in the Rainier and Everest systems in order to get
telemetry data added to D-Bus via phosphor-hwmon.
Rainier power supplies on bus 3 at 0x68, 0x69, 0x6a, and 0x6b.
Everest power supplies on bus 3 at 0x68, 0x69, 0x6d, and 0x6b. Older
hardware has power supply at 0x6a instead of 0x6d.
Input voltage, input power, output voltage, and output current.
No critical or warning low values, to avoid logging errors when AC is
pulled, or other fault conditions handled by the PSU monitor. AC loss,
brownout, over-voltage, over-current, and other conditions should be
handled by the PSU monitor application.
No critical/warning high values, for similar doubled up fault logging.
Change-Id: I9f39e41e2dd0a7b77910207322506d8376f98772
Signed-off-by: Brandon Wyman <bjwyman@gmail.com>
|
|
Add bb files to support phosphor-power-control application in p10bmc
image. Remove skeleton power control application and phosphor-power
power sequencer application from image.
Signed-off-by: Jim Wright <jlwright@us.ibm.com>
Change-Id: Ib1c3cb5ba6880ea73265332107503b61def1a2ef
|
|
Add the associations from the sensors on the panel and the sensors on
the motherboard to their parent FRUs in the inventory.
The chassis association isn't needed here because hwmontempsensor
already creates it.
The association will be used for sensor callouts.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I4292646b0b101608f2bcd4c2d0f65e13f16d3d87
|
|
Tested on Everest HW
Change-Id: I667434f8701b0a635a7d3ca8077b47e4ae297f45
Signed-off-by: Chris Cain <cjcain@us.ibm.com>
|
|
The PL2303 is a serial-to-USB converter chip, needed to communicate
with a UPS. These are commonly attached via USB1.1, meaning that a
UHCI driver is required to communicate with them, so enable that too.
Signed-off-by: Eddie James <eajames@linux.ibm.com>
Change-Id: I3a7550a51d084af498e7a5aab5c6b6e75002a4e3
|
|
This fixes power supply addresses.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I36cfa1097590d01ea26953c3d32f4aec8d3aff7d
|
|
Per [1], Yocto is going to start requiring both branch names and
`protocol=https` in GitHub SRC_URI variables. Ran the upstream tool to
automatically fix our URIs.
1. https://git.yoctoproject.org/poky/tree/documentation/migration-guides/migration-3.5.rst#n10
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I5f9619f7ff707050b469ae74c8aa16846568c5e4
|
|
This package does not currently support openssl v3:
| TpmToOsslMath.h:79:5: error: #error Untested OpenSSL version
| 79 | # error Untested OpenSSL version
| | ^~~~~
For now, remove this package so we can make forward progress with
openssl v3 in openbmc.
The following issue appears to be tracking support:
https://sourceforge.net/p/ibmswtpm2/tickets/18/
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I2249c323b8219525dcef856d8ceab63de417e299
|
|
To support the watchdog application being able to restart after expiring
the watchdog (by being killed) multiple times per BMC reboot, a manual
magic close must be issued by writing a 'V' into the watchdog device
before starting the watchdog. Without that, the underlying hardware
watchdog wouldn't actually start running when the watchdog app was
started again, such as on the next boot.
Most watchdogs simply reboot the system upon expiration, but in this
case it's being used to trip a GPIO that the fan hardware latches on to
go to full speed so must support being restarted multiple times.
The busybox watchdog application used to support this, but it was
removed since it apparently caused some undesireable print statements
from the kernel on other platforms.
An alternative may be to fix this in the ASPEED watchdog driver, though
I'm not sure what all that would entail or if it even makes sense.
Tested: Expire and start the watchdog multiple times and the fans now go
to full speed each time the watchdog expires, unlike before.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I770c4fd7cca15949abd7de9dc07975562080f032
|