| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Installs into bin instead of sbin per guidelines.
Signed-off-by: Patrick Venture <venture@google.com>
Change-Id: Ie7747c1a6581742b2256ba2ddaf985df303ec39f
|
|
|
|
|
|
|
|
|
| |
We need to conform to the std::variant interface for sdbusplus to
eventually switch away from mapbox. The .get() interface is only
provided by mapbox and needs to be converted.
Change-Id: Ia41808d6bab73b097fc631fd9eff549be9a22415
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
|
|
|
|
|
| |
Add clang-format to repo.
Change-Id: Ib3f81524b66414df3e6c93d8a6df6e22c43841e0
Signed-off-by: Patrick Venture <venture@google.com>
|
|
|
|
|
|
|
| |
Update configure.ac to choose the c++17 standard
Change-Id: I805dd7d6de5d8be529d48646cec0c815574c281d
Signed-off-by: Vernon Mauery <vernon.mauery@linux.intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The C++ implementation of the mapper will throw an exception
if the passed in interface was not found on any paths, which
is expected to happen when fru-fault-monitor checks for error
logs on startup and there aren't any. Catch that exception
and just continue on the path of there being no error logs.
Note this is made slightly more complicated by the fact that the
default constructor for sdbusplus::message is deleted, so one
cannot just do:
sdbusplus::message reply;
try
{
auto method = bus.new_method_call(...);
reply = bus.call(method);
}
...
reply.read(...);
Change-Id: Ic15c7112cbfe0f3df974748f90828f12255ec29a
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
|
|
|
|
| |
The upcoming C++ mapper implementation needs it this way.
Tested: With the new mapper and this fix, the call works.
Change-Id: I9b65cb8e12ff0369517ab82dca0be0fae5ae281c
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Handle exceptions from the sdbusplus method call_noreply API.
The phosphor-fru-fault-monitor app would fail with the following
message on Romulus with the latest sdbusplus changes:
sd_bus_call noreply: org.freedesktop.DBus.Error.UnknownObject:
Unknown object '/xyz/openbmc_project/led/groups/core1_fault'.
Add error handling.
Tested: The phosphor-fru-fault-monitor does not core dump on Romulus.
Change-Id: I2bec59c0e0f483e8bfc0cb20a343cb05670212b7
Signed-off-by: Adriana Kobylak <anoo@us.ibm.com>
|
|
|
|
|
|
|
|
|
| |
Will cause symbol collisions when this repository's local
error yaml is copied out during building so phosphor-logging
includes it in its elog-errors.hpp.
Change-Id: Ic467418106d45b6b49b6cabec304c4ae64a3d76b
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With the latest sdbusplus exception handling changes,
an exception was being thrown when trying to read a pair
out of the interfacesAdded signal message.
Change the code to read the path and map separately.
Also modified the name of the data structures to match
what the message actually returned.
Tested: The exception journal entries went away and LEDs
will turn on and off again.
Change-Id: I2c72000e0da365743715da966a0c9e571efd33d5
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
| |
Change-Id: If812954f4c391555a17dbb033e4cf8f2c1ef4c77
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
|
|
|
|
|
|
|
|
|
|
|
| |
Some of the messages we try and parse are not structured in the
format we expect when parsing. When error logging is enabled in
sdbusplus this causes the phosphor-fru-fault-manager to crash. Fix
this crashing by logging errors and ignoring the bad signals and
message responses.
Change-Id: Ie817beaae96961b33a343aabf1dbba9f31606d4b
Signed-off-by: William A. Kennington III <wak@google.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Change the printouts from cout to log<level::DEBUG>
The rest of led manager already logs to phosphor-logging
and only these assertion/deassertion messages were
going to cout.
Change-Id: I9b284221592ac6b48d749acd2f3dfe2d07d4e454
Signed-off-by: Alexander Soldatov <a.soldatov@yadro.com>
Signed-off-by: Alexander Amelkin <a.amelkin@yadro.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's possible that multiple services can provide the
/xyz/openbmc_project/logging paths, which means the
interfaces added handler can get multiple callbacks
for a single error.
That being the case, move down the trace that shows
the new error log object path until after the code
checks that the association interface is present, as
that interface will only be present on one object.
Another option would have been to remove the trace
completely, but it is fairly useful for debug.
Change-Id: I98659562ef024455762110e20109bf3079b96cbe
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
|
|
|
| |
Took a .gitignore from another repo and added a few repo
specific items.
Change-Id: I221f6af5662828be73fa1bbab9b9aa8ae24f2085
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The mapper will no longer emit properties-changed signals
when the last endpoints property is removed from the fault
association object. It will only emit an interfaces-removed
signal in that case, so change the code to subscribe to the
interfaces-removed signal instead.
Change-Id: I225376e550d21b82473015324cd9a261e16fe6db
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
|
|
|
|
|
|
|
|
|
| |
Spelling errors found using github.com/lucasdemarchi/codespell
A tool to fix common misspellings.
This tool is licensed under GNU General Public License, version 2.
Change-Id: I30fd1f3678a2e0fa1c24f6d5b2c0970903db4596
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>
|
|
|
|
|
| |
Change-Id: Idbc2808136f1589e7e4340c336613ece2d069da1
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
| |
An error was logged while starting the fru fault monitor
with no existing error logs, which is not correct.
Resolves openbmc/openbmc#2330
Change-Id: I6b222b45dd99b653443c58d9982d6576f8af0bfa
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
|
|
|
|
|
|
|
| |
Resolves openbmc/openbmc#1631
Change-Id: I1639c84cbb180dfe7f6494b5e80ca45514a3e730
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
| |
There was a race condition while calling mapper from led monitoring after
error logs started persisting, to avoid such errors the callout information is
now extracted from the message received as part of interface added signal.
Resolves openbmc/openbmc#1853
Change-Id: Ie913992bdcb3b1cb93677d64331db34202f8ae72
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
An LED can be part of any number of groups. It is a requirement that
the priority of that particular LED is same in all the groups where
that LED is a member.
If the input yaml file does not conform to this requirement, compilation
should fail.
Change-Id: I187730e835521d14564495cbe52eb139fc1bd264
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
| |
In some cases, it is needed that the Solid-ON action over rules
Blink and this patch adds that support.
Change-Id: Ib0a88b11142ccef3e39ef0a7c6eb3a037c878bc2
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
| |
* iteritems / iterkeys does not exist in python3, use
items / keys directly.
* Order of dicts cannot be relied on.
Change-Id: Ieaf5f58115d4015070d812814d98ca1e58d92ae2
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
|
|
|
| |
Change-Id: Id833073b9e45fae24d6bd0e7f9310ef0ba8658de
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
|
|
|
| |
Change-Id: I28378b155da15282b1b5bf0403075b55bda5b70d
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
|
|
|
|
|
|
|
|
|
| |
This application waits for new fault entries and resolution of
exiting faults to assert or deassert LED for the corresponding FRUs,
respectively.
Change-Id: I92ead0c8d3132dd7a6740b536231d6588ac42471
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
|
|
|
|
|
|
|
|
|
|
| |
When certain LED is first set to "Blink" and is set to
"ON" again, LED needs to stay "Blink" as opposed to
turning "ON". This is the default behavior needed on all
systems unless overridden.
Change-Id: Ied52bb60c538516802b05eb72c139a31499178f5
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
LED group consists of name and the associated LEDs that need to be
actioned on when the group is asserted or de-asserted.
However, its not a hard requirement that all the groups must have the
associated LEDs if the system does not have the usecase.
For example, it is required from the firmware standpoint that the
groups "bmc_booted" and "power_on" be present but not all systems
may have a mapping LED. So in those cases, the group itself can be
in asserted state but without doing anything underneath. Providing
this support will make sure that addition of LEDs into those groups
if needed in the future are handled seamlessly.
Change-Id: I345886a3edbabf11f1e9f5c9de4b9cf28a7c33f6
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Regardless of the machine type, there needs to be standard
set of LED groups that would get asserted / de-asserted under
these conditions:
1) BMC booted
2) Power On
3) Power Off
This makes sense on hardwares that have physical LEDs.
However, QEMU does not have physical LEDs associated with
the groups but the groups themselves can be in asserted state so
that if physical LEDs are implemented in future, they can be used.
Change-Id: Ie66f5ac59fcdbecb7ba79c4dbcbdc8fceaace42f
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This converts the LED group names and LED names from all formats
to lower_case_underscore.
Example : fan0IdentifyTest will become fan0_identify_test
Fixes openbmc/openbmc#1312
Change-Id: Iffa458d593239eb1cb6be0bcda749260cb01c80d
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previous code made a call to get the service name for every physical
LED and that was a little costly. This patch changes it to get the
service names of all the physical LED dbus paths and uses it.
This is done everytime a group of LEDs are to be manipulated.
Fixes openbmc/phosphor-led-manager#4
Change-Id: I5ce455f683a38eae7f9b383013f5729ec7dd7fae
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
| |
Today's code returned a hardcoded string for physical LED action
and this patch uses the generated string for the passed in action enum
Fixes openbmc/phosphor-led-manager#2
Change-Id: I960e2d4346966caee526e1a0e8713e18d34b428a
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
| |
Partial fix for openbmc/openbmc#1112.
Change-Id: Ie1995bd0a1d0fe32891ab814a7fe3c0e2aa47db0
Signed-off-by: Leonel Gonzalez <lgonzalez@us.ibm.com>
|
|
|
|
|
| |
Change-Id: I7db65c8bea0bc111e9b0810ffdc33a8ffc924f5e
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
| |
Change-Id: I582caca53aa84ba7037a61f7a2a947ca0b313cc2
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
| |
Moved from hyphen to underscore, changed to all lowercase
and removed some groups
Change-Id: I0e3137230003153c9a898f0afb9f4db63ee46936
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
| |
Change-Id: I3107c1d961c459379b77548a738533567eccf693
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
| |
Fixes openbmc/openbmc#1039
Change-Id: I995168828a187671072f761668c873c2b1936874
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
| |
The output is currently placed in the same directory as the
script. Default instead to cwd, and allow the user to specify
so out of tree builds can work.
Change-Id: Ia7b94f509e0239f7697634d7a52f22e38d7efc6a
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
|
|
|
| |
Change-Id: Ifce2040d750254227ebee7af4e46fca029820c83
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
|
|
|
|
|
|
| |
Somewhere along the line the repo stopped building. Fix a couple
minor build issues.
Change-Id: I17e24a6f7f1a9946b5c6b5b6544a0f6f172a585f
Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
|
|
|
|
|
|
|
| |
Define a variable in configure having the information about config files
Change-Id: Ic68c02189a07d7198a4ea1ba858ec90aca9126ae
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Existing parse_led.py had a hardcoded reference to the current
directory and the file name 'led.yaml'. This patch introduces
changes through which an arbitrary directory and filename can be
passed. What would still remain the older way is the generation
of led-gen.hpp into the current source directorty
Change-Id: I352dadd6aa99ef80192dfca3071357917d4593b8
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
| |
Previous code only considered the Name and Action for the LED but then
DutyOn and Period are also the properties that are associated with LED
and this patchset handles that.
Change-Id: Ifb57719ed3bcf01a4704618089d9ce676e4b951c
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
| |
This patchset enables gtest to allow different combinations
of LED set operations to be verified at build time.
Change-Id: I9c2ddf82c2e23be911233b23037ee44e3ce301db
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
| |
Previous commit used /usr/bin/python to parse a yaml file and that
resulted in CI failures and this patch corrects that by using the
one in toolchain.
Change-Id: I0f3efcab0c0f620a6ddb7b3500aca5694351fe10
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Consider the below scenario:
1) User turns on Identify_Fan_1, which results in Fan_1's LED to blink
2) Fan_1 is faulted resulting in Fan_1, EncFaultFront and EncFaultBack
to be asserted.
3) User turuns off Identify_Fan_1. This should result in Fan_1 going
back to SolidOn state since the Fan_1 is still faulted.
Current implementation handles #3 by first turning off Fan_1 LED and
then turning it back on. So this was adding extra dbus calls to the
actual LED manager and also this would mean that Group manager would
know the underlying LED implementation.
With this patch, Group Manager tells the physical manager to move to
SolidOn state directly from Blink and physical LED manager now has the
job of doing the actual turn off followed by turn on and thus hiding
the implementation locally.
Fixes openbmc/phosphor-led-manager#3
Change-Id: Ic961095d01e59ba39a841ce8b74bbcde8ddec1df
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
| |
This extends generated sdbusplus interface and provides implementation for
handling LED group operations.
Change-Id: I9e6f83f2f801de24d33937bc651228b1c0ccdc37
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
| |
This enables creating custom groups and participating LEDs so that
it can later be generated from MRW. For each of the group, a dbus object
is created which will announce LED actions. Corresponding groups are
asserted / de-asserted based on user input.
Change-Id: I7e64bea13767b8d083dd946f4cf3aeb37e62ff17
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This enables creating custom groups and participating LEDs so that it can later
be generated from MRW. For each of the group, a dbus object is created which
will announce LED actions.
Fixes openbmc/openbmc#550
Change-Id: I7a56d08755288dcfce45ee4c6d6b6c5e5aa454f7
Signed-off-by: Vishwanatha Subbanna <vishwa@linux.vnet.ibm.com>
|