| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
| |
Add << and >> operators to the Stream class for use with a
std::vector<char>.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I22a08bb6518f9aed2385bcdce446a67ea9ea386d
|
|
|
|
|
|
|
|
| |
This class is used by a few other classes, so make it easier to get
data out of it.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I2feb4b83a09e6cb6056c9f1a903ca55e8cacf2ae
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This substructure is part of the callout subsection in the SRC
section of a PEL, and contains information about the FRU (Field
Replaceable Unit) being called out.
This includes:
* The specific type of FRU (see the flags field definitions)
* The FRU part number
* The FRU CCIN value (CCIN = a keyword in VPD).
* The FRU serial number
Instead of just calling out a FRU, this structure can instead be used to
call out a maintenance procedure, which is a string that is used as
a key into the service documentation that maps to a procedure to fix
the problem.
This commit only adds support for creating an object from a flattened PEL,
such as one that comes down from the host. A future commit will handle
creating it from scratch for BMC errors.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ic2b9489abea48084116bf2f450bd293c2d655979
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the SRC section of a PEL, there is a field called the 'ASCII string'.
This is the string of 32 characters that shows up on the panel when the
SRC function is chosen, and usually when people refer to an SRC, the
first 8 characters of this field is what they are referring to.
This new class handles that string. It will belong to the SRC section
object.
For BMC error SRCs, it looks like: BDSSRRRR
Where:
BD = "BD", indicating a BMC error SRC
SS = subsystem value from PEL spec
RRRR = reason code of the error
The remaining 24 characters are spaces (' ').
For example:
"BD8D1234 "
For BMC power* related errors, the value is:
"11001234 "
Where the difference is the "11" instead of "BD", and the following
2 bytes are always "00".
* 'power' means comes from the repository that monitors for power
faults. This is different purely to help keep field service
documentation the same as in previous IBM server generations.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I6e7292e7f5b501428999781b1a5ee5c243a63ac6
|
|
|
|
|
|
|
|
|
|
|
| |
This class is used for accessing the UserData section of a PEL.
This section contains free format data that can be identified by the
component ID, subtype, and version fields in the section header.
Signed-off-by: Aatir Manzur <aatrapps@gmail.com>
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I1223f84353e81202d1ff63c00f3d926cda4994e5
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Now that the Generic section object has been introduced so there are
objects for every section, a flatten can be done by flattening every
object inside the PEL and the previous workaround to save the original
raw data can be removed.
This also adds a test case that uses a real PEL from a previous
generation of systems to flatten to give some better coverage than just
using hand coded PEL sections.
A side affect of this is that the PEL constructors that take the raw
data cannot take a const vector of data, as the Stream class that will
be used to read from the vector cannot take a const. Testcases have
been updated to ensure this data is not modified.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I64ae1d1d4a742c80e14666d6b2a6e1e0efd5fd62
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When unflattening a PEL, create objects for every PEL section in the
log. It will use a factory method to choose which object type to create
based on the section ID in the section header. All of these object will
go into a vector of Section objects, which is the base class for every
PEL section class.
The factory will default to creating a Generic object when it doesn't
have any other type to create.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ief0e4df5c586a46cea66ca47b4479e3444815309
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This object will be created when unflattening a PEL when there is no
other class to use for that section. It just contains a vector<uint8_t>
for its data.
This is done so that the code can always have objects for all PEL
sections, which helps in validating (can at least ensure every section
has a valid header and size), printing (will always have an object to
get its data to at least hex dump), and re-flattening (no need to keep
around the original data buffer).
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I2b79feb4abc0f44179bdb8eab950f0d274e4e472
|
|
|
|
|
|
|
|
| |
Add << and >> operators to the Stream class for use with a
std::vector<uint8_t>.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I9656ee4e34840d311148e6a701131b3420e62d25
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a constructor to the PEL class so it can be built from the message
registry entry for that error along with the event log properties.
When support for new sections are added, they will be created here as
well along with the PrivateHeader and UserHeader section classes.
Eventually, this constructor will be called from the Manager class after
it is told a new event log has been created and after it looks up that
error's PEL message registry entry.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I2d8ca550736aab45a30ac3db9861723b33e6cd32
|
|
|
|
|
|
|
|
|
|
|
| |
Add a constructor to the UserHeader section class so it can be built
from the message registry entry for that error along with the event log
severity.
This will be used when creating PELs from OpenBMC event logs.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I7e432f59de7b3f0ba77c3e5887ed5ec3f442ed44
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a function to map the OpenBMC event log severity values to PEL
severity values.
When creating a PEL from an OpenBMC event log, the event log will have
its own severity property, and if the PEL message registry entry for
that error doesn't supply its own severity (it's optional), use this
function to come up with the PEL severity value to use.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I66aa001265d8acadb165de874e4ade03a8e28007
|
|
|
|
|
|
|
|
|
|
| |
Add a constructor to the PrivateHeader class so it can be built from the
creator's component ID, the OpenBMC log ID, and the creation timestamp.
This will be used when creating PELs from OpenBMC event logs.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I2d1217b9ff0317195366bee50fa72953fd5341dc
|
|
|
|
|
|
|
|
| |
The SRC (System Reference code) is a section in the PEL and several of
its fields are sourced from the message registry.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I4ca7d18a8225f2667b762015c6fd74bfb59d70ff
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The message registry is a JSON file that holds data required to create a
PEL out of an OpenBMC event log. It includes fields like 'subsystem',
'event type', 'action flags', 'SRC reason code', etc.
Many fields in the message registry are optional, and a very minimal
entry make look like:
{
"Name": "xyz.openbmc_project.Power.Error.Fault",
"Subsystem": "power_supply",
"ActionFlags": ["service_action", "report"],
"SRC":
{
"ReasonCode": "0x2030"
}
}
This commit adds support to look up a message registry entry based on an
OpenBMC event log's 'Message' property (i.e.
xyz.openbmc_project.Power.Error.Fault) and then fill in a structure with
the fields found. Future commits will fill in the SRC related fields,
as well as actually create the PEL.
The message registry file can be found on the BMC at:
/usr/share/phosphor-logging/pels/message_registry.json.
For testing, users can put their own message_registry.json in
/etc/phosphor-logging, and that will take precedence.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ie4195ed7e58ab6a231271f6b295e63b1d0a4cd78
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add tables that allow one to go between how a PEL field actually shows
up in the PEL (raw bytes) and how it shows up in the message registry (a
string enumeration). The tables also have a column to show a string
description of that value that can be used by the parser, though for now
those descriptions are all left at "TODO".
There only needs to be a table for a PEL field when there is a
corresponding message registry field that is a string enumeration, so
that when code looks up an error in the message registry it knows what
to fill in the PEL with.
Also provide APIs to look up a row in the table by either the PEL value
or the message registry value.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Iac849bcd2b0449a8d03fac7eb067484e91d28259
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This PEL section contains the Machine Type-Model and Serial number of
the enclosure and is required for BMC PELs. In the constructor that
creates the section from scratch, it gets those values from the
DataInterface class.
This commit doesn't hook this section into the PEL class as there are
some prerequisites that still need to be done first.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I24d679b57751afb00539691defef180191ea8fc7
|
|
|
|
|
|
|
|
|
| |
These will eventually need to be known to a piece of code that
unflattens a PEL into the appropriate section class objects based on the
section ID field in the PEL data.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I90b9d4be49b2e3807a620745fa663f94f7f4e62c
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There are PEL fields that contain information that must be obtained from
various places throughout the BMC, such as the machine type/model/SN
from VPD, a few types of codes levels, etc.
Create a DataInterface class that will provide the APIs for collecting
this information. It has an abstract base class so that its functions
can be mocked to return specific data in test cases.
This commit provides APIs to read and present the machine type-model and
machine serial number. These will be used in the FailingMTM and
ExtendedUserHeader PEL sections.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Iec41fea8d9510ba711475154f019bd59f0028d2e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This class represents the (M)achine (T)ype-(M)odel (S)erial number
structure for the PEL, where it is used in both the ExtendedUserHeader
and FailingMTMS sections.
It consists of an 8 byte machine type+model field of the form TTT-MMMM,
followed by a 12 byte machine serial number field. Unused bytes are set
to 0.
Note that this is not a PEL section itself. It's just used by other PEL
sections.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I15f9858e951a913ab2353cf93b7f20cc2709c502
|
|
|
|
|
|
|
|
|
| |
UserHeader::validate() had some extra return statements
that aren't needed. It is desired for the function to run
all the way through even on a validation failure.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I9b99970179cd7c84a0201b7d67e39b0bee209beb
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
To prepare for supporting PEL sections that can be in any order,
which will probably be stored in a std::vector<unique_ptr<Section>>,
add a pure virtual function in the Section base class so this list
of sections can just be iterated on and have every object in it
flattened.
This flatten() call replaces the operator<<(Stream&, <object>)
functions currently in use, so also convert the operator>> to
unflatten() to make things consistent.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Id68f16fe4197b389a8495c21539a64f9f583c800
|
|
|
|
|
|
|
|
| |
Add a new BCD time conversion function which will be used in a future
commit to convert an OpenBMC event log timestamp into BCD time.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I08f585d1663bf2d2f73d42b8716756d8c7e3559e
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add the getPELData() function on the Repository class to return
PEL data based on a PEL ID or OBMC event log ID.
The intended use for this will be a D-Bus method, mainly used for
debug via the REST interface, to get the PEL data off the BMC when
only the OpenBMC event log ID is known, which will be the case until
the Redfish APIs are ready.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ia1d8bff627992fae16be9136f2814f01ea69009e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When someone deletes an OpenBMC event log, the corresponding PEL should
also be deleted. This commit adds support for that by keeping an
internal map of the IDs to the files that contain the PEL data in the
repository in order to find the file to remove. This will then get
called in the phosphor-logging extension's delete hook.
The actual map key is a structure of both the PEL ID and OpenBMC log ID,
so either can be used to find a PEL in the repository, which will be
useful in other cases, for example for retrieving PEL data based on the
PEL ID.
As the map needs to match the actual repository file contents, it will
get built on startup, and modified when PELs are added and removed.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I33bb96da297c770e175c5c6b19705dda1c8766b6
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Create the Repository class that can save PELs in (and later retrieve
them from) the filesystem. It provides an add() method that can add
a PEL object to the repository.
Now, when the Manager class sees an OpenBMC event log created with the
RAWPEL metadata in the AdditionalData property that points at a file
that contains a PEL, it can save that PEL. Before the PEL is saved, the
log ID and commit timestamp fields in the PEL will be updated - the log
ID to a unique value, and the timestamp to the current time.
Change-Id: I8dbaddf0f155bcb6d40b933294ada83feb75ce53
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This class represents a Platform Event Log.
A PEL consists of sections, and this commit just adds support for the
only required sections - the Private Header and User Header, by
including those specific objects. More will be added in the future.
The only constructor provided in this commit is to construct the object
from an existing flattened PEL buffer. This is for use in the case when
a PEL is received from off the BMC, such as from the host. Future
commits will add support for creating PELs from OpenBMC event logs.
Since there aren't objects yet for every PEL section, the class cannot
make a full flattened PEL without still keeping around the original PEL
data it received in the constructor as mentioned above. So for now it
will keep that data and just overlay the sections it does support when
flattening. In the future, a fully formed PEL will be able to be
constructed from just flattening the section objects in the correct
order.
This commit provides a few public methods of note:
* data() - returns a flattened PEL
* assignID() - sets a unique ID in the log ID field in the Private
Header
* setCommitTime() - Sets the commit timestamp in the Private Header
to the current time
* valid() - Says if the PEL is properly formed
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I2a9d82df9cd096ce77ecca7b2f73b097b8368aa2
|
|
|
|
|
|
|
|
|
|
|
|
| |
Create generatePELID() to return a unique 4B PEL ID every time it is
called. It will start at a base value, and then increment by 1 each
time. It uses a file to save the next value to use.
This will be used by the PEL handling code to create unique values
for the error log ID field in the Private Header section.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I841a8dcc5dc48e2b663004be3dccfb114ba366f2
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The second section in a PEL is always the 'User Header' section. This
commit adds a class to represent that. Right now, the only constructor
available is filling in its data fields from a PEL stream.
Several of the fields in this section have predefined values that are
defined by the PEL specification. Defining any constants or enums for
those will be left to future commits where they will actually be used.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I8b5f856a4284d44c31b04e98a664f20cd8fa0cb6
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The first section in a PEL is always the 'Private Header' section. This
commit adds a class to represent that. Right now, the only constructor
available is filling in its data fields from a PEL stream.
The Section base class, which will be the base class of all PEL
sections, is also being introduced here. It contains the section header
structure, and a valid flag that derived classes can use.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ia5806017155fe1ef29ea57bf8ab202ff861bde2e
|
|
|
|
|
|
|
|
|
|
| |
A PEL is made up of sections, and every section has an 8B
section header. This commit adds a SectionHeader structure
that will represent that header. It will then be included in
all upcoming PEL sections.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ia5356560f49707e21aebca28f4a0b525aa24158d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A PEL stores time in BCD, with a byte each for:
* year MSB
* year LSB
* month
* day
* hour
* minutes
* seconds
* hundredths
This commit adds a structure to represent this, and functions to:
* Create a BCD structure from a std::chrono::time_point
* Convert any number to BCD
* Write the BCD structure into a Stream
* Extract a BCD structure from a Stream
Refresher: The BCD value of 32 is 0x32.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I09ea4098f3a3981931f595d11fb63aff31d9fb0d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This stream inserts data into and extracts data from the vector<uint8_t>
that it is given in its contructor. That vector is how PEL data is
stored. This object takes care of the endian conversion for fields that
require it, as PEL data is big endian.
On writes, it will expand the vector if necessary.
An exception will be thrown an invalid access is attempted, such as
trying to extract a value when at the end of the data.
It provides >> and << operators for common data types, as well as
read()/write() functions when using other types.
Example:
std::vector<uint8_t> data;
Stream stream{data};
uin32_t value = 0x12345678;
stream << value;
stream.offset(0);
uint32_t newValue;
stream >> newValue;
assert(value == newValue);
uint8_t buf[3000] = {0};
stream.write(buf, 3000);
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I8dc5566371749b45a260389a564836433323eef8
|
|
|
|
|
|
|
|
|
|
|
| |
This class will have the logic for how to handle PELs. It will
also eventually provide D-Bus interfaces.
This first commit has stubs, plus some basic code to find if a PEL
is being passed in with the OpenBMC event log.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I55a5da4d2239c688fded31c112895c3c92bab56d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The AdditionalData property on the xyz.openbmc_project.Logging.Entry
interface is a vector of strings of the form: "KEY=VALUE". The
PEL processing code will be interested in those keys and values, and
this class adds a way to get at those values based on a key without
having to do string parsing each time. It returns an
std::optional<std::string> value, and if the key isn't found, then the
std::optional value will be empty.
For Example:
AdditionalData ad{additionalDataPropertyValue};
// Get the value for the FOO key
std::optional<std::string> val = ad.getValue("FOO");
if (val)
std::cout << (*val).size();
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I6ba458840278784b1cc6a0ed88a7fece8794df7d
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When one uses the new `Create` D-Bus API to create an event
log that will be converted into a PEL, the log-manager daemon
will not know the PID of the creator, unlike when the 'commit'
or 'report' shared library APIs are used.
Document this method for allowing the creator to pass in their
PID, which they can find with getpid(), if they want their PID
in an event log and in the corresponding PEL.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I0fadaed301bcd87de1d4395f6c6ec0baacfe0b97
|
|
|
|
|
|
|
|
| |
Add a markdown file just for PEL specific documentation
starting with how to specify a raw PEL in an OpenBMC event log.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: I5bde74b3f271e334dc77f42fd38ca1fef84bd29a
|
|
The goal of extensions is to extend phosphor-logging's
`xyz.openbmc_project.Logging.Entry` log support to allow other log
formats to be created without incurring extra D-Bus call overhead.
The README.md change in this commit provides additional documentation on
how extensions work. The summary is that they allow code that resides
in this repository to provide functions that can be called at certain
points (startup, log creation/deletion) such that the code can then
create their own logs based on the contents of an OpenBMC log. A
specific extension's code is compiled in using a --enable configure
option, so platforms that did not use those log formats would incur no
performance/size penalties.
This commit provides the support for extensions, plus a basic OpenPower
PEL (Platform Event Log) extension as the first extension. PELs are
event logs used only on some OpenPower systems.
Signed-off-by: Matt Spinler <spinler@us.ibm.com>
Change-Id: Ifbb31325261c157678c29bbebc7f6d32d282582f
|