Age | Commit message (Collapse) | Author | Files | Lines |
|
The NVMe SSDs on the ruler drives are already found by FruDevice.
Enable nvmesensor to read temperatures as soon as the x86 is
powered on.
Change-Id: I2191f909ab1b9e4bc79ae0d24de9042fe54291eb
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
After rebasing on upstream the config wasn't properly updated
to use genesis3 custom YAML config. Update the config to actually
read the MAC from EEPROM and set it on Dbus.
Tested: MACAddress field is set on DBus ans phosphor-networkd
applies the MAC address to the network interface.
Change-Id: Ideaffc68e75a005c351edd3b53860bfcc30dfde4
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
After rebasing on upstream the config wasn't properly updated
to use SBP1 custom YAML config. Update the config to actually
read the MAC from EEPROM and set it on Dbus.
Tested: MACAddress field is set on DBus ans phosphor-networkd
applies the MAC address to the network interface.
Change-Id: I44fa3f3350e5b532ab56f5317e4228b36a4a5a14
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
IBM observed consistent hangs in the `uart_otp` tool when programming
the OTP image into the SoC. This was root-caused by Aspeed:
> Message-ID: <TYZPR06MB677027C95FCFABCDA6F81C4D800DA@TYZPR06MB6770.apcprd06.prod.outlook.com>
> From: Neal Liu <neal_liu@aspeedtech.com>
> Subject: Server Management technical issue by Rose.Drehmel@us.ibm.com
>
> Hi Andrew,
>
> It’s okay, I found the problem.
>
> The utility timeout is because it programs key retire bits in
> OTPCFG4[7:0] without enabling secure boot.
>
> In the case of secure boot is enabled, the hardware would detect the
> current boot up key number #id, and it can only retire the number
> which is smaller than the current boot up key #id.
> If not, the OTP status will keep busy, and the utility stocks in
> polling loop.
>
> If you still want to disable key #0 without enabling secure boot,
> OTPCFG0[5] can be another option for you.
>
> I also provide a new programmer.bin to fix this infinite loop problem
> in case user thought BMC is crashed.
>
> You can try it with this command:
>
> $ uart_otp -s 2600 -p ast2600_otp_programmer.bin /dev/ttyUSBx
>
> Thanks
>
> Best Regards,
>
> -Neal
In discussion with Chris we determined that we were not intentionally
attempting to retire the development / low-security key, rather were
just trying to be complete in the specification of our configuration.
Neal responded to our request of how to avoid programming a key
retirement in the configuration file:
> Message-ID: <TYZPR06MB67700B238DB429A51E048E328010A@TYZPR06MB6770.apcprd06.prod.outlook.com>
> From: Neal Liu <neal_liu@aspeedtech.com>
> Subject: Server Management technical issue by Rose.Drehmel@us.ibm.com
>
> Hi Andrew,
>
> Just delete line #72 as unspecified value.
>
> Thanks
>
> Best Regards,
>
> -Neal
>>
>> From: Andrew Jeffery <andrewrj@au1.ibm.com>
>> To: Neal Liu <neal_liu@aspeedtech.com>
>> Subject: Re: Server Management technical issue by Rose.Drehmel@us.ibm.com
>>
>> Hi Neal,
>>
>> I've discussed your findings with Chris Engel, who is our platform
>> security person. We determined that we do not want to mark the low
>> security key as retired in the OTP as we're handling that via the
>> FWSPIMISO strapping pin.
>>
>> What change should I make to our OTP configuration so that we don't
>> retire key 0 during programming?
>>
>> Andrew
Line 72 in this case refers to our OTP configuration file:
https://github.com/openbmc/openbmc/blob/2a25492c13e2b768f94b864a51f84e82e4238aef/meta-ibm/recipes-bsp/u-boot/u-boot-aspeed-sdk/p10bmc/ibm.json#L72
Leave "Keys Retire ID" unspecified to avoid leaving the OTP engine busy.
Cc: Chris Engel <cjengel@us.ibm.com>
Cc: Rose Drehmel <Rose.Drehmel@us.ibm.com>
Cc: Briana Foxworth <befoxwor@us.ibm.com>
Cc: Nicole Nett <nschwart@us.ibm.com>
Change-Id: Ib6b75a40f5debd5ba1166f0f69a07114b76d9c34
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
This file is used to know which location codes to call out in error logs
when I2C and other accesses fail.
There are still a few missing callouts, but most are there.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ic0c61e362c5dcab4d374168da84bb7128d0bb22e
|
|
There is no in-tree user of this package. Remove it so we can archive
the repository.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I777ac0e0ffaf14508a49d5160831e59d3044fe5e
|
|
The phosphor-gpio-util that was being used was a custom tool that can be
replaced by gpioset. It was also just removed from its repo.
The service now has to be a oneshot because otherwise it isn't allowed
to have multiple ExecStart lines.
Tested:
Powered on multiple times, ensured the power supply input history was
cleared on each power on.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ib181c67b70e2feb109c1a2aabba8d046e7dfe380
|
|
Makes SOL and obmc-console-client working again.
Change-Id: I3d98319c2278794ba3db9ace4c89932c305d66d8
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Add phosphor-ipmi-fru and ipmi-fru-properties configurations
to parse the MAC address from board IPMI EEPROM.
Add config for phosphor-network to sync MAC to u-boot env.
Change-Id: Ib470d71df06e75f9936f31efa3b86708dc226d57
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Change-Id: I2a3e8bb54e07fc9d3e7b9bbc400c98c44f8d2810
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Change-Id: I55bc9fc8195b5954492d6ba31d9a2cbc10ddbfba
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Add static inventory devices that have no FRU EEPROM and
are known to be present.
entity-manager doesn't support creating those inventory items,
so add them here.
The following commits will add more functionality around the
created inventory items. Each item has a presence GPIO and
additional actions are triggered when a presence is detected.
Change-Id: I267b05765d408c88a50e005a3e5c2f4505c00f42
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
IBM systems do not always configure both eth0 and eth1 ports. This
causes this service to fail and hold up the boot of the BMC by the 2
minute timeout.
Override this service file to have IBM systems only wait for eth0 OR
eth1 to become active.
The SBP1 BMC has two NICs. Be happy with a link up on one of them.
TEST=Build for SBP1 & make sure with waits for either of the two
NIC to be active.
Change-Id: I2ed633805f7e71021c010b9bc53e076f2c55732e
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Add support for the IBM SBP1 board.
It's an evaluation platform based on Aspeed AST2600 SoC powering
a 4 socket Intel Sapphire Rapids platform.
The main difference to regular platforms is the non-existence of a
CPLD/FPGA. Everything on this board is directly connected to and
controlled by BMC including power sequence for PCH/CPU.
The power-sequencing makes use of in kernel PMBUS drivers that are
currently upstreamed and needs to be backported to OpenBMC kernel fork.
Change-Id: I34a0b70e1f7f6c1ab2be38e8d66d5912950e0d82
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Signed-off-by: Naresh Solanki <Naresh.Solanki@9elements.com>
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
currently admin user created with home directory '/' and there is an
error while deleting admin user.
this commit updates admin user home directory to /home/admin.
Tested by: Notice admin home directory and Delete admin user
Change-Id: I32210435ac0a94555d2f369e226df7e04f49d88c
Signed-off-by: Ravi Teja <raviteja28031990@gmail.com>
|
|
The reset-fan-watchdog service needs to toggle a GPIO in order to unlock
the fan watchdog circuit so fans can be controlled after the watchdog
has expired and locked them at high speed. This gets called on every
power on to bring the fans under control again in case the watchdog had
expired on the previous boot.
This was done previously using a custom application which used libgpiod
under the covers, which was written before gpioset was around.
Now that gpioset is available, just use it instead to toggle the GPIO so
there's less custom code we have to worry about maintaining.
The actual minimum timing requirement of the toggle is around 5ns, so
the 1us being used by all the config files meets that.
Tested:
Expired the watchdog by stopping the fan-watchdog service, then powered
off and back on and made sure the fans can still be controlled.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ibd36ff17e83b57d521364f0092dfa1f5abf34856
|
|
Add phosphor-ipmi-fru and ipmi-fru-properties configurations
to parse the MAC address from board IPMI EEPROM.
Add config for phosphor-network to sync MAC to u-boot env.
Change-Id: I76c294f8822da11da2fedef5fdd842ce6d10efd0
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Enable the fan control PID loop. It depends on the D-Bus
configuration posted here:
https://gerrit.openbmc.org/c/openbmc/entity-manager/+/58421
Change-Id: I32979f5705a745dcf0abdab5cce9d62340c33fbf
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Change-Id: I5be5a644de26648d194b66ae70148201b4dafd58
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Change-Id: Ic6a5bc3547e6416897a2a27eb59c0484a9abf67e
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Change-Id: Iffad7afabba59e0c50d2d7151847f166f9af0b9d
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
IBM systems do not always configure both eth0 and eth1 ports. This
causes this service to fail and hold up the boot of the BMC by the 2
minute timeout.
Override this service file to have IBM systems only wait for eth0 OR
eth1 to become active.
The Genesis3 BMC has two NICs. Be happy with a link up on one of them.
TEST=Build for Genesis3 & make sure with waits for either of the two
NIC to be active.
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
Change-Id: I0ffd652fabd5e998b14e23621d89046a9eed5455
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
Modify ACBEL PSU configuration file to read output voltage from sysfs
in2_label and to read the output current from curr2_label.
Testing: Placed the modified config file in
/etc//default/obmc/hwmon/ahb/apb/bus@1e78a000/i2c-bus@200/power-supply@5a.conf
then verified the output voltage and current output shows up on the BMC
sensors web page and redfish.
Change-Id: Iacf1469295b99f6a7363d26983786f9c7ed8826a
Signed-off-by: Faisal Awada <faisal@us.ibm.com>
|
|
Currently same subnet IPv4 configuration on both ethernet interfaces
does not work.
This commit sets kernel ARP settings which allow same subnet routes on
both ethernet interfaces at the same time.
Tested by:
Configure same subnet IP addresses on both interfaces and network
configuration works
Change-Id: I01962b8bf066ccf7006915e77f07287b6e5544d4
Signed-off-by: Ravi Teja <raviteja28031990@gmail.com>
|
|
Adjust meta-ibm to add support for the new machine.
Change-Id: I2eb61f3d9c85c23b16f026f8f9eb4506c78e97ad
Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com>
|
|
Dbus timeout for PLDM is set to 5 secs. After reset reload BMC is
too busy and cannot return in 5 secs. Increasing the dbus timeout
value to 10 seconds to address this issue.
Change-Id: I1f513062a617d53061ec3e0a4aca03c996f11d0e
Signed-off-by: vkaverap@in.ibm.com <vkaverap@in.ibm.com>
|
|
Patrick Williams (6):
clang-format: copy latest and re-format
writefrudata: remove vla
strgfnhandler: suppress unused parameter warnings
frup: suppress stringop-truncation warning
scripts: accept absolute paths to YAML
build: enable meson builds
build: install strgfnhandler into ipmi-providers dir
Vernon Mauery (1):
Set new maintainer as Patrick Williams
Willy Tu (1):
fru-parser: Remove the use of mktime
Switch to Meson.
Change-Id: Iec9c41cc17a72f84248b8cd75f70353e053f17e2
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
There is a desire to add non-openpower systems to the meta-ibm layer. To
do that, need to ensure the bbappend files that are appending to
openpower recipes are only included if the machine being built includes
the meta-openpower layer.
Tested:
- Confirmed a system not including meta-openpower builds ok with this
- Confirmed p10bmc still uses the bbappend files
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I32a08e9f20a20152d51c77cdfca6bbc15b5e93d4
|
|
Tell phosphor-buttons the proper power button profile to use for IBM
p10bmc systems so it meets IBM's requirements.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I48337c9b2d31c47c9e3aa0bfaa5f8ebda80e8577
|
|
Define three categories of ABI configuration via PACKAGECONFIG:
1. abi-production: Only deprecated and stable symbols are exposed
2. abi-maintenance: Only stable and testing symbols are exposed
3. abi-development: All deprecated, stable and testing symbols are
exposed
This makes it relatively straight-forward to perform broad testing for
use of deprecated functions by setting
`PACKAGECONFIG = "abi-maintenance"` in a `libpldm_%.bbappend` in the
relevant meta layer.
However, build for abi-production by default.
Change-Id: I9df824e2543829dbe66a7203240e5367ffded3e3
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
The first dtb listed in the KERNEL_DEVICETREE list is the default used
for the BMC's genesis boot. The Bonnell device tree specifies an eeprom
that's larger in size than other p10bmc systems. Therefore the device
driver currently truncates the data on this eeprom when doing a genesis
boot on Bonnell since the default device tree is for Rainier.
Move Bonnell to the top of the list to be the default since it has the
larger eeprom size.
Tested: Booted Rainier and Everest on simulation to test a genesis BMC
boot. Verified they first booted with the Bonnell's device tree, then
the BMC automatically rebooted after the machine type was identified and
the device tree corresponding to the system was loaded.
Change-Id: Icb9abe7d8b04d42befd9eb906d095f4fa046ab3c
Signed-off-by: Adriana Kobylak <anoo@us.ibm.com>
|
|
Clearly define the optional features and their defaults within the PSM
repo.
Change-Id: Ibf466d9319b4176c816ef22c4cdbfd8cd693a90b
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
The switch enables AF_MCTP and adds the `mctp` tool and `mctpd` so we
can setup and manage an in-kernel MCTP network.
Enable it for p10bmc and also add CONFIG_TUN so we can use a tun device
to transport packets between the kernel and the astlpc interface
implemented in libmctp.
Change-Id: I21d0af9e751f91975c760b4b4708b3b401b95417
Signed-off-by: Rashmica Gupta <rashmica@linux.ibm.com>
|
|
The documentation for systemctl's `disable` verb[1] states:
> This command implicitly reloads the system manager configuration after
> completing the operation. Note that this command does not implicitly
> stop the units that are being disabled. If this is desired, either
> combine this command with the --now switch, or invoke the stop command
> with appropriate arguments later.
[1]: https://www.man7.org/linux/man-pages/man1/systemctl.1.html
In simulation, reloads get expensive as they involve a bunch of IO.
During boot, under qemu, such a reload can take on the order of 15
seconds.
Further, Andrew Geissler states (no reference, private slack):
> The service-config-manager is responsible for enabling and disabling
> IPMI. On startup, it looks for the service and socket files via
> systemd d-bus queries and caches in a file, /etc/srvcfg-mgr.json. If
> you don’t run the services once on a fresh startup,
> service-config-manager never sees them and therefore never provides a
> way to enable/disable them. So on a fresh factory reset boot, you let
> them start, service-config-manager sees them and caches them, and then
> you disable them from there on out.
As evidenced by the `After` statements in the unit, the expectation is
that the units we're disabling have already been started.
The documentation for systemctl's `mask` verb[1] provides a useful
perspective:
> Mask one or more units, as specified on the command line. This will
> link these unit files to /dev/null, making it impossible to start
> them. This is a stronger version of disable, since it prohibits all
> kinds of activation of the unit, including enablement and manual
> activation.
The key insight here is the `disable` keyword tends to best be used to
prevent the unit from being started as part of a given target, e.g. the
default target.
Given that we would only need to reload the systemd configuration to
prevent the units from being started as part of a default target, and
the fact that they must have already been started in accordance with
`After` directive, there's no need to force an immediate reload of the
systemd configuration upon disabling the units. Further, it's possible
to combine the stop and the disable operation into one with `--now` as
demonstrated in the `disable` documentation above. And finally, as the
disable verb takes a unit PATTERN and not a singular unit, we can
compress the operation down to a single invocation.
Tested:
1. Booted up fresh p10bmc (factory reset), verified IPMI disabled by
default via Redfish API
2. Enabled IPMI via Redfish API, verified bmcweb indicated IPMI enabled
and services running
3. Rebooted BMC, verified IPMI was still enabled after BMC reboot
4. Disabled IPMI over Redfish, rebooted BMC, confirmed IPMI was still
disabled
Change-Id: I0926e9d16a56c2f022e415f4f40c35695dd155b8
Tested-by: Andrew Geissler <geissonator@yahoo.com>
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
The [Service] section specifies:
```
ExecStart=/bin/touch /var/lib/ipmi-net-disable-one-time
```
However, the [Unit] section specifies the following condition:
```
ConditionFileNotEmpty=!/var/lib/ipmi-net-disable-one-time
```
The documentation for ConditionFileNotEmpty[1] states:
> ConditionFileNotEmpty= is similar to ConditionPathExists= but verifies
> that a certain path exists and refers to a regular file with a
> non-zero size.
[1]: https://www.freedesktop.org/software/systemd/man/systemd.unit.html#ConditionFileNotEmpty=
And the documentation for `touch`[2] states:
> A FILE argument that does not exist is created empty, unless -c or -h
> is supplied.
[2]: https://www.man7.org/linux/man-pages/man1/touch.1.html
Neither `-c` nor `-h` are supplied, therefore the ExecStart= directive
will create an empty file. As the ConditionFileNotEmpty= directive is
negated the condition will always pass: Either the file doesn't exist,
or it does exist and is empty. Therefore, as configured, this unit
always runs.
Always running the unit is counter to its 'one-time' name. Switch to
`ConditionPathExists=`[3] to enforce the one-time nature.
[3]: https://www.freedesktop.org/software/systemd/man/systemd.unit.html#ConditionPathExists=
Change-Id: I1a8418a03c4fb9c6754f72f8651384b2f7e17715
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
IBM systems do not want to allow boot operations unless the BMC is in
the Ready state
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: Ic28cb6df9e7f803998a3fb380b3236ed31d6d4ec
|
|
As of 9a8f30ec5b58 ("obmc-console: Introduce console-id, deprecate
socket-id") in obmc-console `console-id` is preferred over `socket-id`.
Convert all in-tree configurations to `console-id` to enable removal of
support for the deprecated `socket-id` configuration option.
Change-Id: Ifd50603c378cfed5eb793b829b8296b7964e4194
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I3792c2aa7f4bcd92254717c5bd00bc438f9101be
|
|
LED endpoints with Identifying, Identified_by for System was missing in
Association json,as a result of that LocationIndicatorActive for system
was null in redfish GET output.
Also, from GUI, System identify LED was not functioning.
If we try to set it, it goes away once we refresh the page,and no impact
on led group.
So this has been addressed for P10bmc Rainier and Everest system.
Tested:-
-Toggling reflects on LocationIndicatorActive as well as LED state of
associated LED group.
-It Works fine with redfish curl command as well as GUI button.
$ curl -k -H "X-Auth-Token: $bmc_token" -X PATCH https://${bmc}/redfish/v1/Systems/system -d '{"LocationIndicatorActive": true}'
$
$ curl -k -H "X-Auth-Token: $bmc_token" -X GET https://${bmc}/redfish/v1/Systems/system |grep "LocationIndicatorActive"
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 6262 100 6262 0 0 42026 0 --:--:-- --:--: "LocationIndicatorActive": true,
- It shows up the result on hardware as below-
~# busctl introspect xyz.openbmc_project.LED.GroupManager /xyz/openbmc_project/led/groups/enclosure_identify
NAME TYPE SIGNATURE RESULT/VALUE FLAGS
org.freedesktop.DBus.Introspectable interface - - -
.Introspect method - s -
org.freedesktop.DBus.Peer interface - - -
.GetMachineId method - s -
.Ping method - - -
org.freedesktop.DBus.Properties interface - - -
.Get method ss v -
.GetAll method s a{sv} -
.Set method ssv - -
.PropertiesChanged signal sa{sv}as - -
xyz.openbmc_project.Led.Group interface - - -
.Asserted property b true emits-change writable
~# cat /var/lib/phosphor-led-manager/savedGroups
{
"value0": [
"/xyz/openbmc_project/led/groups/bmc_booted",
"/xyz/openbmc_project/led/groups/enclosure_identify"
]
}#
Change-Id: I20dd57cc0c4edb0e28039ca9a276075914adc9f6
Signed-off-by: Alpana Kumari <alpankum@in.ibm.com>
|
|
This feature helps to support a system where multiple
BMC's will be a part of a single composed system.
etcd is a strongly consistent, distributed key-value store
that provides a reliable way to store data that needs to be
accessed by a distributed system or cluster of machines.
Enabled "etcd" package to install etcd binaries and
the required systemd units to start etcd in cluster configuration.
Tested: verified image.
Installed binaries and systemd unit files.
rootfs/etc/etcd.d/etcd-existing.conf
rootfs/etc/etcd-new.conf.sample
rootfs/etc/systemd/system/multi-user.target.wants/etcd-new.service
rootfs/etc/systemd/system/multi-user.target.wants/etcd.service
rootfs/lib/systemd/system/etcd-new.service
rootfs/lib/systemd/system/etcd.service
rootfs/usr/bin/etcd
rootfs/usr/bin/etcdctl
Change-Id: Iad30df3475b861491492168f975d7b347a1481a6
Signed-off-by: ojayanth <ojayanth@in.ibm.com>
|
|
This new machine will be based on the p10bmc machine. It will be
utilized to prototype and develop next generation distributed OpenBMC
technology.
Change-Id: Id9cef6fae52867a8eabf9e38a210c26dfa04ed25
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
Would like to make a variant of the p10bmc machine that includes some of
the new etcd function. To do that, the search paths need to not include
the MACHINE variable. bitbake is smart enough to look under the
sub-directory and find the appropriate files.
Change-Id: I542b4568c18210870b3e621461585a4c7eee3121
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
Andrew Geissler (1):
systemd: no installation in templated targets
Includes needed service installation.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I22023bee9be60928a1ab990d627dc3d5553f30ad
|
|
Similar to previous generation of the POWER processor, introduce a
p10.inc that can be included on P10 systems.
Without this, the needed op-proc-control services and functions are not
brought in.
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
Change-Id: I58490e95e69ef8b4b31b0f4bc9213787058f0bba
|
|
Witherspoon uses UBI instead of static partitioning for flash
management, so make sure the qemu helper script can find the flash
image.
Tested: Reached a login prompt with `runqemu nographic slirp`.
Change-Id: I513d0a5b878c553456d70bef25adedfa25bf9c90
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Andrew Geissler (1):
systemd: no installation in templated targets
Includes changes required to ensure services are properly installed.
Tested:
-Confirmed services were installed into correct targets when building
p10bmc machine
Change-Id: I2d675ba083b2901ad1c26485d2f8b075fd6b3a25
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
This reverts commit ec7cab9378f548e94ee4af83946b2583d1fc7194.
ec7cab9378f5 ("Add socket-id for the first console") was only a partial
fixup of the obmc-console configurations found in the tree. For
historical reasons not all platforms that support multiple consoles
install a client configuration file for all consoles. Instead they relied
on some default behaviour that was removed in obmc-console's 4e7186918599
("Fixed broken dbus interface for multiple consoles"). Applying the
configuration changes in the manner of ec7cab9378f5 ("Add socket-id for
the first console") resulted in an asymmetric configuration between the
client and server, breaking some SSH SOL instances.
As of ae2460d0b8e8 ("obmc-console: Provide a default value for
`console-id`.") in obmc-console the requirement to specify
`socket-id`[^1] is lifted. Instead, the configuration can choose to
override a default value. This restores the original behaviour and
unbreaks SSH SOL.
[^1]: Now an alias of the `console-id` configuration key
Websocket and IPMI consoles remain broken. A fix is under development
that also takes a proper approach to supporting multiple consoles in
bmcweb, and the same solution pattern can be applied in ipmid.
More discussion of the problems involved can be found here:
https://amboar.github.io/notes/2023/05/08/happenings-in-obmc-console.html
Change-Id: I274284c791758f336da6d6301dc523bac2b5dd69
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
Upstream yocto introduced a change via e510222 (systemd-systemctl:
fix instance template WantedBy symlink construction).
This fixes a bug that we in OpenBMC had been taking advantage of in that
we were able to document our templated target dependencies without it
actually doing anything. The real installation of services within
targets occurs in our bitbake recipes due to the complexity of chassis
and host instances on a per machine basis.
Leave the dependency information in the service files but comment them
out. It's useful to be able to look at a service and understand which
targets it's going to be installed into by the bitbake recipes.
Change-Id: I1bf3bc6ce7a79d6c5847f5f1b4e6a97ca67332f8
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
Upstream yocto introduced a change via e510222 (systemd-systemctl:
fix instance template WantedBy symlink construction).
This fixes a bug that we in OpenBMC had been taking advantage of in that
we were able to document our templated target dependencies without it
actually doing anything. The real installation of services within
targets occurs in our bitbake recipes due to the complexity of chassis
and host instances on a per machine basis.
Leave the dependency information in the service files but comment them
out. It's useful to be able to look at a service and understand which
targets it's going to be installed into by the bitbake recipes.
Change-Id: I9c485cbd19eb2804efe9916d4080ba3c662f65e5
Signed-off-by: Andrew Geissler <geissonator@yahoo.com>
|
|
Patrick Williams (1):
clang-format: copy latest and re-format
Change-Id: Ib3888ded06c8d0766b158293420b64644d600d7e
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|