Age | Commit message (Collapse) | Author | Files | Lines |
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I85272779478b66452acd11be93f5fea99e4c3a34
|
|
The old name is incorrect, it should be gbmcncsidhcp
Change-Id: I28fb55e3587fcfbeda4cfa333317cd0b62883d9c
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
We provide a new variable called `GBMC_NCSI_DHCP_RELAY` to indicate
whether or not our image should support DHCP relaying. By default as a
heuristic, we use the FLASH_SIZE for the value as only 64MB and bigger
images will need it in practice.
In the future it will be tuned per-platform to be enabled only where
needed and eventually disabled by default.
Change-Id: I08bc07074cada79fcf7b18522d1219ffcc14d1be
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
We never used this and we never intend to pick it back up.
Change-Id: Iaa23d6b9d87129eca8e76d67b8173941342ea2e9
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Willy Tu (2):
fmt: Remove fmtlib dependencies
Add wrapper script to setup delay powercycle
Change-Id: I0830a9788a3b7fb6405de707963719b52dc2d62d
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
This allows us to easily map in other NICs from a machine config.
Change-Id: Ib1aede5af75bfda30d11fd5f310b3ff096af86bb
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Most platforms want to use 1 as has been the implicit standard. However,
this is not required and might be changed for some machines. This makes
it easy to change via a machine configuration variable.
Change-Id: I5dff5dfb1c97bab531042fd364114f320d3ebe42
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Our configuration assumes the NCSI interface in slot 1 is eth0,
let's update it to rewrite the interface name appropriately.
Change-Id: Iecaaeac2e0a2ebc33b3b19d456de85f4e9ec48de
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
We want to be able to enable any slot for potentially multiple NICs.
Follow on changes will reuse this logic.
We can now enable arbitrary ethernet channels by calling
and the relevant ipmi json mappings will be established. The resulting
file looks something like
```
{
"0": {
"name": "IPMB",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "ipmb",
"protocol_type": "ipmb-1.0",
"session_supported": "session-less",
"is_ipmi": true
}
},
"1": {
"name": "ncsi",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "lan-802.3",
"protocol_type": "ipmb-1.0",
"session_supported": "multi-session",
"is_ipmi": true
}
},
"2": {
"name": "front",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "lan-802.3",
"protocol_type": "ipmb-1.0",
"session_supported": "multi-session",
"is_ipmi": true
}
},
"3": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"4": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"5": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"6": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"7": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"8": {
"name": "INTRABMC",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "oem",
"protocol_type": "oem",
"session_supported": "session-less",
"is_ipmi": true
}
},
"9": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"10": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"11": {
"name": "gbmcbr",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "lan-802.3",
"protocol_type": "ipmb-1.0",
"session_supported": "multi-session",
"is_ipmi": true
}
},
"12": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"13": {
"name": "RESERVED",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "reserved",
"protocol_type": "na",
"session_supported": "session-less",
"is_ipmi": true
}
},
"14": {
"name": "SELF",
"is_valid": false,
"active_sessions": 0,
"channel_info": {
"medium_type": "ipmb",
"protocol_type": "ipmb-1.0",
"session_supported": "session-less",
"is_ipmi": true
}
},
"15": {
"name": "ipmi_kcs3",
"is_valid": true,
"active_sessions": 0,
"channel_info": {
"medium_type": "system-interface",
"protocol_type": "kcs",
"session_supported": "session-less",
"is_ipmi": true
}
}
}
```
Change-Id: I361edd658491ea85b79b60d3107fab6dd5e192e2
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
On many machines it's safe to purge all configurations that do not match
the NCSI interface, to keep a tidy config directory and purge
potentially problematic configs.
Change-Id: I5f714516eb3718544f0787ac38eb69d485086b42
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Sometimes we have machines that switch between interface names when
upgrading versions, we want to keep the configs to the best of our
ability when migrating forward.
Change-Id: I58fb962b19431114dc342a3252ce2f834c160194
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Michael Shen (1):
all: Use stdplus::print over std::fprintf
Change-Id: I71977d404fd66eeadf131137fa7e22d02efc4ae1
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Patrick Williams (1):
clang-format: copy latest and re-format
Change-Id: I436ff65bf944d90f79481ce939e036ed7756e0fd
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Patrick Williams (1):
clang-format: copy latest and re-format
Change-Id: Idaa34228786fcc5e0b4ad098ba83d93c80eb121e
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
With rng-tools being removed by default, it should not be present in
any meta-google machine. Remove the configuration for it which would
be unused.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I47b849b55b65f0fb5c8963a7e2d6319ebc44b35f
|
|
This allows for MAC assignments like `0c:c4:13:4d:d4:ff` of `2`,
where the second address requires a rollover to `0c:c4:13:4d:d5:00`
Change-Id: Ief1489e4f99b27c9af7d4b5f1826dccf36863e0f
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
The packagegroup-obmc-apps bbappend attempts to remove `bmcweb` from
a package group that it isn't even part of, meaning the code has no
affect.
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I94216b012993b96e433f2f8a5349d7ac840a6bee
|
|
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I515d5f732b02038b20a8ae45afc8b8d93d5d278d
|
|
Willy Tu (1):
ncsid: Fix if_packet warnings
Change-Id: I4eaa632f786d12869cd06ea7d6d0aaddece50725
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Blocks reads, whos values are greated than 128, will be formated as
negative integers, causing cascading errors in the checksum calculation.
Format using unsigned specifier instead to prevent this.
Tested:
Before:
```
dd if=/sys/bus/i2c/devices/13-0050/eeprom bs=1 count=8 skip=8 | \
hexdump -v -e '1/1 "%d "'
1 1 32 36 -107 47 -115 -121
```
After:
```
dd if=/sys/bus/i2c/devices/13-0050/eeprom bs=1 count=8 skip=8 | \
hexdump -v -e '1/1 "%u "'
1 1 32 36 149 47 141 135
```
Change-Id: I187f0918a25005310bb900022047e03fe68483d4
Signed-off-by: Vlad Sytchenko <vsytch@google.com>
|
|
Instead of individual filenames, adding the containing directory
itself.
Adding the directory used by the gRPC receiver of SMBIOS blobs from
remote hosts, so they can stick around after the smbios-mdr service is
restarted, without requiring hosts to send them again.
As for the gRPC receiver itself, see this patch stack:
https://gerrit.openbmc.org/c/openbmc/smbios-mdr/+/65048
Change-Id: Ic73ab81e3a9761a101b6040a445a5cf042cbb4b0
Signed-off-by: Josh Lehan <krellan@google.com>
|
|
The previous change introduced a bashism, but yocto expects that all
shell scripting is POSIX compatible.
Fixes: bf4e27c3408955c214000108001e92c308c1861f
Change-Id: Ifc3cce4d625a876b224853b3b04b1b84b95327b2
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
These addresses should not be used by default for traffic on the
interface, as we may have multiple BMCs using the same hardcoded address
WRT legacy applications. We want multiple BMCs of the same type to be
able to coexist if they have unique addresses and right now they are not
correctly preferring these addresses.
Change-Id: I0d02cbf63b54f2464ae889ea773e032da0e9b650
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
glome is now in the meta-security layer and shouldn't(*) be included
here.
glome-config is also removed; downstreams should bbappend glome to
overwrite /etc/glome/config.
(*) https://lore.kernel.org/all/Ys6zBXR+8AP3wjYG@heinlein.stwcx.org.github.beta.tailscale.net/
Tested:
Added meta-security layer to a build and tested downstream BMC build.
Link: https://gerrit.openbmc.org/id/I8e96a5c15a277c343ca38c32ad5b987944642008
Change-Id: I8e96a5c15a277c343ca38c32ad5b987944642008
Signed-off-by: Luke Granger-Brown <lukegb@google.com>
|
|
If dhcp-done is running already we want a fresh start, so return error
to trigger a powercycle.
Change-Id: I146054adacb0ea62de54ec525915cecd27d3e85b
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
In rsyslog, the logs are read via the journal files like with
`journalctl`. The socket used by ForwardToSyslog is not used at all
(`run/systemd/journal/syslog`), so we will disable the forwarding to
avoid unnecessary work.
With the logs spamming the CPU usage of systemd-journald went down
after this change.
Before:
Always top 1 or two in CPU usage,
```
PID USER PRI NI VIRT RES SHR S CPU%-MEM% TIME+ Command
30892 root 20 0 23520 10108 9720 R 84.3 1.0 0:37.72 /usr/lib/systemd/systemd-journald
```
After:
Less CPU usage overall.
```
PID USER PRI NI VIRT RES SHR S CPU%-MEM% TIME+ Command
30976 root 20 0 23520 7608 7216 R 40.8 0.8 2:32.36 /usr/lib/systemd/systemd-journald
```
Reference:
- https://www.freedesktop.org/software/systemd/man/journald.conf.html#Forwarding%20to%20traditional%20syslog%20daemons
Change-Id: I50cf868d63697fc9b72529863bb83d249f3275c7
Signed-off-by: Willy Tu <wltu@google.com>
|
|
George Liu (1):
meson_options.txt: Support for reading options from meson.options
Change-Id: I1a69373c9b6eaa2089cb8371fc3c72f4025a301d
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Follow advice described in meta-tpm README to avoid warnings on each
bitbake invocation saying that the layer is not configured.
Change-Id: Ic7578ce4fe558bcae5204bb42b30f2247eff1c54
Signed-off-by: Vlad Sytchenko <vsytch@google.com>
|
|
The SRCREV now matches
https://github.com/openbmc/openbmc/blob/b3415dd342accf1b4b75d2aa5b041d24cc0c7783/meta-security/meta-tpm/recipes-tpm1/hoth/libhoth_git.bb#L4
and the override is not needed anymore.
Change-Id: I785497b05c5c224c6b224fb1e6180a0bfdea50fa
Signed-off-by: Willy Tu <wltu@google.com>
|
|
We have seen that smartnic workaround is running after dhcp done.
If we see dhcp-done service running, no need to restart dhcp.
Change-Id: I83ad8bc376f6727ec98a4ce4b2e545d00a2b9c82
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
Move gbmc ramoops collector into private repo.
This is to reduce the customized systemd script in public openbmc.
Change-Id: I9e8861fc2261852eb86b9b222749cb340ac66dad
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
Allow the state file for rsyslog as configured by the default
rsyslog.conf. This prevent us from reading the the whole logs again (if
the logs persists across reboots).
Reference:
- https://gerrit.openbmc.org/c/openbmc/openbmc/+/64418
Change-Id: Ia3650627a26e2dbdcc62f1c464e389e4e668e3a8
Signed-off-by: Willy Tu <wltu@google.com>
|
|
Switch to meson.
Patrick Williams (3):
build: add basic meson config
build: add meson for unit-tests
build: remove autotools support
William A. Kennington III (1):
build: Some minor refactoring
Change-Id: I3697ef580954f5423af0620df5c25e496de7e1c2
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
Pmsg type will generate pmsg-, which will be used by kdump, this is to
let this service not processing the core file.
Change-Id: Ida4f8038a07ccea4be659cd734acc1a7d5c86e59
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
Kdump preprocessing needs to read, modify, and repack a initfs, this is to enable cpio
-O option.
Change-Id: I3f59da1ad0094bbf4a0ef58094a26dfac3423f1d
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
William A. Kennington III (1):
ncsid: Fix build issue
Change-Id: I564484dc83c986b9ef6a92fb631f1dabca53dd67
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Signed-off-by: Harvey Wu <Harvey.Wu@quantatw.com>
Change-Id: I2dffda1e08e8c536a7f9b27a936dcebb4674deb7
|
|
python3-core is taking a lot of space causing the image to be too large
in some systems. Remove the python package config to remove the
dependencies. This is not done in bbappend because it didn't work when
testing it out.
Tested:
Built with gbs and it doesn't have python3-core when tested with
```
INHERIT += "buildhistory"
BUILDHISTORY_COMMIT = "1"
```
to get the package sizes.
Change-Id: I5040cfb46ad618d622855891a462d53a887433fd
Signed-off-by: Willy Tu <wltu@google.com>
|
|
We have platforms using fd1x instead of fd00, this will make sure we
pick the correct ip address.
Change-Id: I9c385a43adb539cfbfdd80651c59b0c0eba05082
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|
|
Hao Zhou (1):
File System: Dependency Injection and Mock
John Wedig (1):
Add IPMI OEM command LinuxBootDone
Patrick Williams (1):
build: upgrade to C++23
Change-Id: I2d203a8ca9c5e0947b2d70f9d3aec3831c31fa64
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
Signed-off-by: John Wedig <johnwedig@google.com>
|
|
John Wedig (1):
espi-control: Add program to disable eSPI
Change-Id: I1eaaf766531cad5a0855a92f87805336cb0eeb77
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Willy Tu (1):
Add dhcp-done to base meson.build
Change-Id: Ic42ce388597835a95efb48eb604e94bc143245f2
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Willy Tu (1):
iwyu: Cleanup include headers
Change-Id: Ia2fefc046a550df78ad043ec56dc263d66c16cdb
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Willy Tu (1):
cleanup fmt::format for std::format
Change-Id: Ic422c068fc99a96267f9f17cd4a3fc0a81cc1c7a
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
The default rsyslog.conf uses `imjournal`, so we should enable it by
default to make sure that the config is valid.
Change-Id: Idba361d3a374a88207e7825216f393522647d11d
Signed-off-by: Willy Tu <wltu@google.com>
|
|
Patrick Williams (1):
build: upgrade to C++23
Change-Id: I687dcbb1ff35a4e767770e40bc2c9f65f5e905d8
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
This supports extending what is unpacked from the tarball and allows
arbitrary hooks to execute on the unpacked files.
Tested: Added some code to one of our platforms that consumes the
tarball with multiple data files and correctly processes different
subsets of the files.
Change-Id: Id48a5c9327bafa367efd231b3907c305fe6994fe
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
Hao Zhou (1):
Add flags for bmc modes transitions
Change-Id: I23170671607f45e9e90d168051ee3afe2e56dfff
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
Willy Tu (1):
build: Fix build warnings
Change-Id: Ifa7f7075b8c6c2cdc6fe1b8f19118daff085edac
Signed-off-by: Andrew Geissler <openbmcbump-github@yahoo.com>
|
|
networkd wait-online services will wait for all ethernet interfaces to
be up by default. Some interface will never be ready because it is
unused Currently only rsyslog is waiting for this service so it should
be fine to just remove this service.
Change-Id: I21b8b2e00b75fedcc2497ca40a405ed598371b53
Signed-off-by: Yuxiao Zhang <yuxiaozhang@google.com>
|