| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Extract the long onmessage() websocket lambda in requestRoutes() into a
named function, handleDbusMonitorMessage(), per the <10 line lambda
coding standard in docs/COMMON_ERRORS.md.
Tested:
- No other functional change.
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings
introduced.
Change-Id: I340df11dfbc626e26dab931f0ca0581cbd17e102
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
In order to silence clang-tidy, check payload value.
Tested: Code compiles. clang-tidy runs through.
Change-Id: Ib9db2552b0c9f8974728b8afe909a138d948b278
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>
|
|
Instead of registering each route separately, register all routes in a
single function call for systems dump, manager dump and manager
faultlog.
Tested: Code compiles. Unit tests pass. Validator passed.
Change-Id: I335ef22873137743eade4d3c0119ad38c6a68a83
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>
|
|
Similar to how other log services got broken out into own header files,
do the same for dump / faultlog.
A new dump utility header has been created, request handler and route
register function went into manager / systems log service header files
respectively.
Code logic stays untouched in this patch.
Tested: Code compiles. Unit test runs through.
Validator was run on romulus in qemu with phosphor-debug-collector ibm
extension enabled to get access to /xyz/openbmc_project/dump/system.
Via the 'CreateDump' method empty dump entries were manually created
for:
/xyz/openbmc_project/dump/system
/xyz/openbmc_project/dump/bmc
/xyz/openbmc_project/dump/faultlog
```
busctl call xyz.openbmc_project.Dump.Manager \
/xyz/openbmc_project/dump/bmc \
xyz.openbmc_project.Dump.Create \
CreateDump a{sv} 0
> o "/xyz/openbmc_project/dump/bmc/entry/1"
busctl tree xyz.openbmc_project.Dump.Manager
`- /xyz
`- /xyz/openbmc_project
`- /xyz/openbmc_project/dump
|- /xyz/openbmc_project/dump/bmc
| `- /xyz/openbmc_project/dump/bmc/entry
| |- /xyz/openbmc_project/dump/bmc/entry/1
| |- /xyz/openbmc_project/dump/bmc/entry/2
| |- /xyz/openbmc_project/dump/bmc/entry/3
| |- /xyz/openbmc_project/dump/bmc/entry/4
| |- /xyz/openbmc_project/dump/bmc/entry/5
| `- /xyz/openbmc_project/dump/bmc/entry/6
|- /xyz/openbmc_project/dump/faultlog
| `- /xyz/openbmc_project/dump/faultlog/entry
| `- /xyz/openbmc_project/dump/faultlog/entry/1
|- /xyz/openbmc_project/dump/resource
`- /xyz/openbmc_project/dump/system
`- /xyz/openbmc_project/dump/system/entry
`- /xyz/openbmc_project/dump/system/entry/1
```
Overwrite OperationStatus of /xyz/openbmc_project/dump/system/entry/1
for the entry to be added to the collection:
```
busctl set-property xyz.openbmc_project.Dump.Manager \
/xyz/openbmc_project/dump/system/entry/1 \
xyz.openbmc_project.Common.Progress Status \
s xyz.openbmc_project.Common.Progress.OperationStatus.Completed \
```
Systems/Dump
Validating /redfish/v1/Systems/system/LogServices/Dump...
- Pass: 14, Warn: 0, Fail: 0, Skip: 9
Validating /redfish/v1/Systems/system/LogServices/Dump/Entries...
- Pass: 17, Warn: 0, Fail: 0, Skip: 30
Validating /redfish/v1/Systems/system/LogServices/Dump/Entries/1...
- Pass: 10, Warn: 0, Fail: 0, Skip: 26
+--------------+--------------+--------------+--------------+
| PASS | WARN | FAIL | NOT TESTED |
+--------------+--------------+--------------+--------------+
| 41 | 0 | 0 | 65 |
+--------------+--------------+--------------+--------------+
Managers/Dump
Validating /redfish/v1/Managers/bmc/LogServices/Dump...
- Pass: 14, Warn: 0, Fail: 0, Skip: 9
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries...
- Pass: 70, Warn: 0, Fail: 0, Skip: 162
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/1...
- Pass: 11, Warn: 0, Fail: 0, Skip: 25
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/2...
- Pass: 11, Warn: 0, Fail: 0, Skip: 25
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/3...
- Pass: 9, Warn: 0, Fail: 0, Skip: 27
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/4...
- Pass: 9, Warn: 0, Fail: 0, Skip: 27
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/5...
- Pass: 9, Warn: 0, Fail: 0, Skip: 27
Validating /redfish/v1/Managers/bmc/LogServices/Dump/Entries/6...
- Pass: 9, Warn: 0, Fail: 0, Skip: 27
+--------------+--------------+--------------+--------------+
| PASS | WARN | FAIL | NOT TESTED |
+--------------+--------------+--------------+--------------+
| 142 | 0 | 0 | 329 |
+--------------+--------------+--------------+--------------+
Managers/FaultLog
Validating /redfish/v1/Managers/bmc/LogServices/FaultLog...
- Pass: 12, Warn: 0, Fail: 0, Skip: 9
Validating /redfish/v1/Managers/bmc/LogServices/FaultLog/Entries...
- Pass: 13, Warn: 0, Fail: 0, Skip: 34
Validating /redfish/v1/Managers/bmc/LogServices/FaultLog/Entries/1...
- Pass: 6, Warn: 0, Fail: 0, Skip: 30
+--------------+--------------+--------------+--------------+
| PASS | WARN | FAIL | NOT TESTED |
+--------------+--------------+--------------+--------------+
| 31 | 0 | 0 | 73 |
+--------------+--------------+--------------+--------------+
Change-Id: I6379f877149c545b3b3bfcda8510c6ee1085dcf9
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>
|
|
Several functions in processor.hpp and virtual_media.hpp return
ProcessorArchitecture, InstructionSet, and TransferProtocolType
values as std::string literals, which lacks compile-time type
safety. This can lead to typos or invalid values that are only
caught at runtime.
Convert these to use processor::ProcessorArchitecture,
processor::InstructionSet, and virtual_media::TransferProtocolType
enums instead. These enums are already defined in their
respective generated header files with NLOHMANN_JSON_SERIALIZE_ENUM
mappings.
The enums automatically serialize to the same JSON strings. This
provides compile-time type checking and prevents invalid property
values.
Tested:
- Verified properties serialize to identical JSON values.
- Redfish Service Validator: PASS: 6082, WARN: 353, FAIL: 0,
NOT TESTED: 5245
Change-Id: Ia8c0a80e418770c39c159d66d30655b621e488b2
Signed-off-by: Vinothkumar Shanmugavel <vinothkumars@ami.com>
|
|
Cover dbusToRfPowerState() and the callbacks introduced by the Fabric
Switch PowerState change: afterGetSwitchPowerState and
afterGetSwitchPowerStateService. The tests assert each branch's
externally visible behavior: known PowerState decorator values map to
the matching resource::PowerState and unmapped values yield nullopt; a
property read error or an unmapped value leaves PowerState omitted
without failing the request; EBADR, io_error and an empty mapper
object leave PowerState omitted; any other mapper error sets an
internal server error.
Change-Id: Id488f88b8d7784a4f195fcf7c832fbecfdb0b513
Signed-off-by: JY Voon <jvoon@nvidia.com>
|
|
Extend handleFabricSwitchPathSwitchGet() in fabric.hpp to read
xyz.openbmc_project.State.Decorator.PowerState from the resolved Switch
D-Bus path and surface the value as the Redfish PowerState property on
/redfish/v1/Fabrics/{FabricId}/Switches/{SwitchId}.
Tested: Build an image for nvl32-obmc machine with the following
patches cherry-picked:
1. Align with upstream u-boot dts tree:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/89932
2. mctpd configuration:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/87390
3. Enable nvidia-gpu sensor:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/89933
4. nvidia-gpu: add PowerState on ConnectX PCIeDevice:
https://gerrit.openbmc.org/c/openbmc/dbus-sensors/+/90170
$ curl -sk https://{BMC_IP}/redfish/v1/Fabrics/fabric/Switches/\
Nvidia_ConnectX_24_PCIe
{
"@odata.id":
"/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_24_PCIe",
"@odata.type": "#Switch.v1_7_0.Switch",
"Id": "Nvidia_ConnectX_24_PCIe",
"Name": "Nvidia_ConnectX_24_PCIe",
"Ports": {
"@odata.id":
".../Switches/Nvidia_ConnectX_24_PCIe/Ports"
},
"PowerState": "On",
"Status": {
"Health": "OK",
"State": "Enabled"
}
}
$ busctl get-property xyz.openbmc_project.GpuSensor \
/xyz/openbmc_project/inventory/Nvidia_ConnectX_24_PCIe \
xyz.openbmc_project.State.Decorator.PowerState PowerState
s "xyz.openbmc_project.State.Decorator.PowerState.State.On"
```
Redfish Service Validator:
Summary - PASS: 9752, WARN: 373, FAIL: 0, NOT TESTED: 13355
Validating /redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_24_PCIe...
- Pass: 9, Warn: 0, Fail: 0, Skip: 32
Validating /redfish/v1/Fabrics/fabric/Switches/Terminus_18_PCIeSwitch_1_100...
- Pass: 8, Warn: 0, Fail: 0, Skip: 33
```
Change-Id: I4bc636664ae74280b72eab13219e5bf5b1e79a85
Signed-off-by: JY Voon <jvoon@nvidia.com>
|
|
When phosphor-ipmi-net or any other protocol service is not registered
with the D-Bus object mapper, handleProtocolEnabled() was returning
HTTP 500 InternalError to the client.
The mapper returns EIO (value=5, generic category) when the requested
service object path does not exist. Handle this specific error code
and return PropertyNotWritable, consistent with the existing fallback
for missing backend entries in the same function.
Tested: Toggled IPMI/ProtocolEnabled on EVB-AST2600 where
phosphor-ipmi-net is absent. Confirmed response changed from
500 InternalError to 405 PropertyNotWritable.
Change-Id: I1da8eb5658985512faeb6603d8a9be6fb2295aae
Signed-off-by: Vijaysankar Ravi <vijaysankarr@ami.com>
|
|
Add optional support for UUID property for PCIeDevice.
Tested: Build an image for nvl32-obmc machine with the following patch
cherry picked.
```
1. Align with upstream u-boot dts tree:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/89932
2. mctpd configuration:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/87390
3. Enable nvidia-gpu sensor:
https://gerrit.openbmc.org/c/openbmc/openbmc/+/89933
```
PCIeDevice with Common.UUID published (GPU path):
$ busctl get-property xyz.openbmc_project.GpuSensor \
/xyz/openbmc_project/inventory/Nvidia_GPU_10 \
xyz.openbmc_project.Common.UUID UUID
s "68174e99-daf4-02d1-ba39-4ed2bf8a0f21"
$ curl -sk https://{BMC_IP}/redfish/v1/Systems/system/\
PCIeDevices/Nvidia_GPU_10
{
"@odata.id": "/redfish/v1/Systems/system/PCIeDevices/Nvidia_GPU_10",
"@odata.type": "#PCIeDevice.v1_19_0.PCIeDevice",
"Id": "Nvidia_GPU_10",
"Manufacturer": "NVIDIA",
"Model": "RTXPRO6000BlackwellDC",
"Name": "PCIe Device",
"PartNumber": "900-2G153-0000-000",
"SerialNumber": "1792425045093",
"Status": {
"Health": "OK",
"State": "Enabled"
},
"UUID": "68174e99-daf4-02d1-ba39-4ed2bf8a0f21"
}
PCIeDevice without Common.UUID published (ConnectX path):
$ busctl get-property xyz.openbmc_project.GpuSensor \
/xyz/openbmc_project/inventory/Nvidia_ConnectX_24_PCIe \
xyz.openbmc_project.Common.UUID UUID
Failed to get property UUID on interface
xyz.openbmc_project.Common.UUID: Unknown interface
xyz.openbmc_project.Common.UUID or property UUID.
$ curl -sk https://{BMC_IP}/redfish/v1/Systems/system/\
PCIeDevices/Nvidia_ConnectX_24_PCIe | jq 'has("UUID")'
false
Change-Id: I38cf972f03f0dd0299b708ea8203872c73c8e224
Signed-off-by: JY Voon <jvoon@nvidia.com>
|
|
Extended Cable's Status.State to use Available mapping and added
Status.Health to Cable which it did not have before
Utilize resource_util's getResourceHealth and getResourceState to
standardize Redfish Status.State and Status.Health
Tested: on QEMU, which has no host and so no real cable inventory, with
a cable object injected into phosphor-inventory-manager.
Functional true -> "Health": "OK", "State": "Enabled"
Functional false -> "Health": "Critical", "State": "Enabled"
Where State can be "Enabled", "Absent", and "UnavailableOffline" if the
xyz.openbmc_project.State.Decorator.Availability interface is
implemented
Change-Id: I9f99f7cfe348545506753cc4eedd81570fbf3fba
Signed-off-by: Akshay Gaitonde <a.g@utexas.edu>
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
Add unit tests for the security header helper.
Cover headers that are always added, preserve existing Cache-Control
values, and verify the HTML-only header path.
Also validate representative Content-Security-Policy and
Permissions-Policy values so future header regressions are caught by
unit test coverage.
Tested:
Passed local CI docker run for bmcweb unit tests
Change-Id: I069d203f6bc35155d17d8488ab5511aa38218f3a
Signed-off-by: Joel P J <joelpj@ami.com>
|
|
- Map `Available` to Redfish `UnavailableOffline` for resource which are
present but not available
- Utilize getResourceState and getResourceHealth utility functions to
standardize Status.State and Status.Health
Tested:
- Redfish Service Validator passes
- Request
```
curl -k -v https://${bmc}/redfish/v1/Systems/system/Processors/cpu0/
```
```
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0",
"@odata.type": "#Processor.v1_18_0.Processor",
"Id": "cpu0",
...
"LocationIndicatorActive": false,
"MaxSpeedMHz": 0,
"Name": "Processor",
"ProcessorType": "CPU",
"Socket": "",
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
},
...
}
```
Change-Id: I8e08c15eb6e10251d3e1fe0902150a76f9c63bd2
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
getLocationCode() used an inline lambda for the D-Bus async response
callback. The lambda body was long enough to trigger the long-lambda
ast-grep rule, which required an inline suppression comment
(// ast-grep-ignore: long-lambda).
Extract the lambda into a named afterGetLocationCode() function and
bind it with std::bind_front() instead. This matches the existing
bmcweb pattern for D-Bus response callbacks, improves readability,
and removes the need for the ast-grep suppression.
Also change the EBADR log message from BMCWEB_LOG_ERROR to
BMCWEB_LOG_WARNING.
Tested:
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings.
Change-Id: I3ffa51b1f6bb4bc7bd31e0b2f361294c8aa36978
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
Convert the inline DELETE subscription lambda into a named
handleEventDestinationDelete() bound via std::bind_front, matching the
handler style used elsewhere in this file.
Tested:
- No other functional change.
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings
introduced.
Change-Id: I92b00d312548cbd38a3d1a821cea7839764b199d
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
checkForQuiesced() reads the ActiveState property of the
obmc-bmc-service-quiesce@0.target systemd unit using an unescaped
object path. systemd derives unit object paths by escaping every
character outside [A-Za-z0-9] as an underscore plus two hex digits,
so '-' becomes _2d, '@' becomes _40 and '.' becomes _2e.
The unescaped form is not merely a mismatch. '-', '@' and '.' are
not valid characters in a D-Bus object path, so the call fails on
the bus before it ever reaches systemd. The error branch then falls
through to Health "OK" and State "Enabled", so the failure is
silent and a quiesced BMC reports itself as healthy and enabled
over Redfish.
The correctly escaped form is already used elsewhere in the tree:
manager_diagnostic_data.hpp queries bmcweb_2eservice.
Users will now see Status/State "Quiesced" and Status/Health
"Critical" on /redfish/v1/Managers/<id> when the BMC has quiesced.
Previously that condition always reported "Enabled" and "OK".
Clients that implicitly relied on Managers never reporting Quiesced
will see the new value. This is the behavior the existing code
already intends and what the Redfish Resource schema specifies, so
it is treated as a bug fix rather than a breaking change.
Tested: Built obmc-phosphor-image for romulus (bmcweb 92b625bc) and
ran it under QEMU (qemu-system-arm 11.0.2, -M romulus-bmc).
D-Bus layer, on the running BMC:
'''
# busctl call org.freedesktop.systemd1 /org/freedesktop/systemd1 \
org.freedesktop.systemd1.Manager GetUnit s \
obmc-bmc-service-quiesce@0.target
o "/org/freedesktop/systemd1/unit/obmc_2dbmc_2dservice_2dquiesce_400_2etarget"
# busctl get-property org.freedesktop.systemd1 \
/org/freedesktop/systemd1/unit/obmc-bmc-service-quiesce@0.target \
org.freedesktop.systemd1.Unit ActiveState
Invalid object path:
/org/freedesktop/systemd1/unit/obmc-bmc-service-quiesce@0.target
# busctl get-property org.freedesktop.systemd1 \
/org/freedesktop/systemd1/unit/obmc_2dbmc_2dservice_2dquiesce_400_2etarget \
org.freedesktop.systemd1.Unit ActiveState
s "active"
'''
Before this patch, with obmc-bmc-service-quiesce@0.target active,
GET /redfish/v1/Managers/bmc still returned:
'''
"Status": { "Health": "OK", "State": "Enabled" }
'''
After this patch, same image and machine, with the quiesce target
inactive and CurrentBMCState Ready:
'''
"Status": { "Health": "OK", "State": "Enabled" }
'''
and then after systemctl start obmc-bmc-service-quiesce@0.target,
with CurrentBMCState Quiesced:
'''
"Status": { "Health": "Critical", "State": "Quiesced" }
'''
No unit test was added: bmcweb has no D-Bus mocking in test/ and
there is no test file for managers.hpp.
Change-Id: I676c4f58c90e9619603b27bad59940e382f57f13
Signed-off-by: Xinyuan Wang <wang.x.6f@gmail.com>
|
|
Silent clang-tidy warnings:
- remove unused header includes
- switch <systemd/sd_bus.h> to <systemd/sd-bus-protocol> for
sd_bus_error usage
- change str.find usage to str.contains
Tested: Code compiles.
Change-Id: I814239d47e1586ebc28bccc00c08e41f711b9beb
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>
|
|
Report invalid ResetType values with
Base.1.19.ActionParameterValueNotInList instead of
Base.1.19.ActionParameterUnknown.
ResetType is a recognized action parameter, so an unsupported value
must be reported as an invalid enumeration value rather than an unknown
parameter name. This gives clients the correct error semantics and
actionable response details.
Tested:
The invalid-value response now identifies ResetType as a known
parameter and reports the supplied value as unsupported.
POST /redfish/v1/Systems/system/Actions/ComputerSystem.Reset with
{"ResetType":"Nope"} returns HTTP 400 and
Base.1.19.ActionParameterValueNotInList.
POST /redfish/v1/Systems/system/Actions/ComputerSystem.Reset with {}
continues to return HTTP 400 and Base.1.19.PropertyMissing.
Change-Id: Ieb71f89b0bedf5ceeaaa8c8c6d89c05ae4a4aac6
Signed-off-by: Joel Pullokaran Jesin <joelpj@ami.com>
|
|
Replace the dump entry response assignments for `EntryType` and
`DiagnosticDataType` with generated `LogEntry` enum values.
No functional changes
Change-Id: I5b2ce19dfe2cb7070494d4e315605946ecc7bcba
Signed-off-by: Joel Pullokaran Jesin <joelpj@ami.com>
|
|
The VirtualMedia handlers compared the manager path segment against
a hardcoded "bmc" instead of BMCWEB_REDFISH_MANAGER_URI_NAME. That
constant comes from the redfish-manager-uri-name meson option,
which defaults to "bmc" but may be set to any value.
When the option is set to another value, Manager still advertises
VirtualMedia at /redfish/v1/Managers/<name>/VirtualMedia, but the
VirtualMedia handlers rejected that URI with resourceNotFound.
The collection GET, the resource GET, and the InsertMedia and
EjectMedia actions were all unreachable, so Manager published a
dangling link.
Compare against BMCWEB_REDFISH_MANAGER_URI_NAME instead, matching
how managers.hpp, network_protocol.hpp, and update_service.hpp
already validate the manager segment. Behavior is unchanged for
the default option value.
Tested:
Built bmcweb with -Dredfish-manager-uri-name=manager.
Verified GET /redfish/v1/Managers/manager/VirtualMedia works.
Verified the default manager URI behavior remains unchanged.
Change-Id: I2375398d6a2780ef403870061d6583b3ae699551
Signed-off-by: Rython Cai <rython_cai@jabil.com>
|
|
ossf maintains a list of best practice compiler flags for hardened
services[1] implement them.
Tested: Unit tests pass. No code changes.
[1] https://github.com/ossf/wg-best-practices-os-developers/blob/main/docs/Compiler-Hardening-Guides/Compiler-Options-Hardening-Guide-for-C-and-C%2B%2B.md
Change-Id: I0c83c30aa2d80317d9e016544d736a60ad894806
Signed-off-by: Ed Tanous <etanous@nvidia.com>
|
|
Change ResetType to std::string so readJsonAction() automatically
returns ActionParameterMissing when the field is absent, removing
the redundant explicit check. Also replace propertyValueNotInList
with actionParameterValueNotInList for unsupported ResetType values,
consistent with the Redfish schema semantics for action parameters.
Tested:
- No functionality changes, updated the validation fix.
- Verified property missing error by testing locally in AST2600
platform.
- Service Validator Passed.
+--------------+--------------+--------------+--------------+
| PASS | WARN | FAIL | NOT TESTED |
+--------------+--------------+--------------+--------------+
| 5820 | 353 | 0 | 4906 |
+--------------+--------------+--------------+--------------+
Change-Id: I286afe08a66f0fc07c518cd5b227e5312fe5a524
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
- Utilized resource util's getResourceState and getResourceHealth to
get the Status.State and Status.Health of the power supply resource
Tested:
- Unit tests passed
- Redfish Service Validator passed
- Response as expected:
```
curl -k -v https://${bmc}/redfish/v1/Chassis/chassis/PowerSubsystem/PowerSupplies/powersupply0
```
Results in
```
{
"@odata.id": "/redfish/v1/Chassis/chassis/PowerSubsystem/PowerSupplies/powersupply0",
...
"SerialNumber": "Y YL106 63F0",
"SparePartNumber": "0 ",
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
...
}
```
Change-Id: I9fed94f7800fe01b6cb6de1a768da381bc116155
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
The getSwMinimumVersion and getSwStatus functions each contained an
inline lambda that triggered the long-lambda ast-grep rule, requiring
an inline suppression (// ast-grep-ignore: long-lambda).
This commit extracts each lambda into a separate named function,
afterGetSwMinimumVersion and afterGetSwStatus respectively, and uses
std::bind_front to handle the asynchronous callback. This aligns the
code with the preferred bmcweb pattern for D-Bus callbacks, improves
readability, and allows the removal of the static analysis
suppressions.
Tested:
- No other functional change.
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings
introduced.
Change-Id: I5557501b1af6506ebd23301110899b324b94d867
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
The getAssetInfo function contained an inline lambda that triggered
the long-lambda ast-grep rule, requiring an inline suppression
(// ast-grep-ignore: long-lambda).
This commit extracts the lambda into a separate afterGetAssetInfo
function and uses std::bind_front to handle the asynchronous
callback. This aligns the code with the preferred bmcweb pattern for
D-Bus callbacks, improves readability, and allows the removal of the
static analysis suppression.
Tested:
- No other functional change.
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings
introduced.
Change-Id: I471ab2402be00c81e629e965994a650024e120c4
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
The getManagedHostProperty function contained an inline lambda that
triggered the long-lambda ast-grep rule, requiring an inline
suppression (// ast-grep-ignore: long-lambda).
This commit extracts the lambda into a separate
afterGetManagedHostProperty function and uses std::bind_front to
handle the asynchronous callback. This aligns the code with the
preferred bmcweb pattern for D-Bus callbacks, improves readability,
and allows the removal of the static analysis suppression.
Tested:
- No other functional change.
- Tested on an AST2600 SoC platform.
- Redfish Service Validator passed with no new errors or warnings
introduced.
Change-Id: I8e1fe0aae26b07a4f84c1d41aef536627f7051ac
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
The translateMemoryTypeToRedfish() function returns memory device
types as std::string literals, which lacks compile-time type
safety. This can lead to typos or invalid values that are only
caught at runtime.
Convert this function to return memory::MemoryDeviceType enum
instead. This enum is already defined in generated/enums/memory.hpp
with NLOHMANN_JSON_SERIALIZE_ENUM mapping and provides proper type
checking.
The enum automatically serializes to the same JSON strings, so
there is no functional change to the Redfish API. This provides
compile-time type checking and better error handling through the
Invalid enum value for unknown or unsupported memory types.
Tested:
- No functional change.
- Verified MemoryDeviceType property serializes to
identical JSON values.
- Redfish Service Validator: PASS: 6002, WARN: 353, FAIL: 0,
NOT TESTED: 5227
Change-Id: I34087487ce88ab4767a27d93ff85448b946a6528
Signed-off-by: Vinothkumar Shanmugavel <vinothkumars@ami.com>
|
|
- Map `Available` to Redfish `UnavailableOffline` for Status.State when
resource is present but not available
Tested:
- Unit tests passed
- Redfish Service Validator passed
- Request
```
curl -k -v https://${bmc}/redfish/v1/Systems/system/Memory/dimm3
```
Results in
```
{
"@odata.id": "/redfish/v1/Systems/system/Memory/dimm3",
"@odata.type": "#Memory.v1_12_0.Memory",
...
"Status": {
"Health": "OK",
"State": "Enabled"
}
...
```
`UnavailableOffline` if not available but present
```
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
```
Change-Id: I163095f013bc233a8eb7de3894ef950a75d93379
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- Utilize getResourceState and getResourceHealth utility function for
Assembly Status.State and Status.Health
- Map `Available` to Redfish `UnavailableOffline`
Tested:
- Unit tests passed
- Redfish Service Validator passed
- Request expected
```
curl -k -v https://${bmc}/redfish/v1/Chassis/chassis/Assembly
```
```
{
"@odata.id": "/redfish/v1/Chassis/chassis/Assembly",
"@odata.type": "#Assembly.v1_6_0.Assembly",
"Assemblies": [
{
"@odata.id": "/redfish/v1/Chassis/chassis/Assembly#/Assemblies/0",
"@odata.type": "#Assembly.v1_6_0.AssemblyData",
"Location": {
"PartLocation": {
"ServiceLabel": "Ufcs-N00-D0"
}
},
"LocationIndicatorActive": false,
"MemberId": "0",
"Name": "base_op_panel_blyth",
"Status": {
"Health": "OK",
"State": "Absent"
}
},
{
"@odata.id": "/redfish/v1/Chassis/chassis/Assembly#/Assemblies/1",
"@odata.type": "#Assembly.v1_6_0.AssemblyData",
"Location": {
"PartLocation": {
"ServiceLabel": "U78DA.N00.1234567-D1"
}
},
"LocationIndicatorActive": false,
"MemberId": "1",
"Model": "6B86",
"Name": "lcd_op_panel_hill",
"PartNumber": "PN12345",
"SerialNumber": "YL6B86010000",
"SparePartNumber": "F191014",
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
}
],
"Assemblies@odata.count": 2,
"Id": "Assembly",
"Name": "Assembly Collection"
}
```
Change-Id: If7f06b27dbdfa914a9db4480ba856d9dd25c8795
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- Utilize getResourceState and getResourceHealth utility function for
resource's Status.State and Status.Health
- Map `Available` to Redfish `UnavailableOffline`
Tested:
- Unit tests passed
- Redfish Service Validator passed
- Request expected:
```
curl -k -v https://${bmc}/redfish/v1/Systems/system/FabricAdapters/disk_backplane0/Ports/dp0_connector1
```
Results in
```
{
"@odata.id": "/redfish/v1/Systems/system/FabricAdapters/disk_backplane0/Ports/dp0_connector1",
"@odata.type": "#Port.v1_11_0.Port",
"Id": "dp0_connector1",
...
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
}
```
if the resource is present but not available
Change-Id: I9220a9df2f602ed0712b10142782eeaa6f949ac9
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- Utilized resource util's getResourceState and getResourceHealth to
get the Status.State and Status.Health of the fabric adapters resource
- Map `Available` to Redfish `UnavailableOffline`
Tested:
- Unit tests passed
- Redfish Service Validator passed
```
curl -k -v https://${bmc}/redfish/v1/Systems/system/FabricAdapters/pcie_card0
```
Results in
```
{
"@odata.id": "/redfish/v1/Systems/system/FabricAdapters/pcie_card0",
"@odata.type": "#FabricAdapter.v1_4_0.FabricAdapter",
...
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
...
}
```
If the resource is functional and present but not available
Change-Id: I045191029ba0f08d1eccb2bf4406cd4925289459
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- Utilize getResourceState and getResourceHealth utility function for
PCIe Status.State and Status.Health
- Map `Available` to Redfish `UnavailableOffline` for resources that are
present but unavailable
Tested:
- Unit tests passed
- Redfish Service Validator
- Request
```
curl -k -v https://${bmc}/redfish/v1/Systems/system/PCIeDevices/pcie_card12
```
- Results in
```
{
"@odata.id": "/redfish/v1/Systems/system/PCIeDevices/pcie_card12",
"@odata.type": "#PCIeDevice.v1_19_0.PCIeDevice",
"Id": "pcie_card12",
"Name": "PCIe Device",
"PCIeFunctions": {
"@odata.id": "/redfish/v1/Systems/system/PCIeDevices/pcie_card12/PCIeFunctions"
},
...
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
}
```
Change-Id: I97c40b8800dcbd902a7eb5d9bc1902a738ddfd6b
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- Utilized resource util's getResourceState() and getResourceHealth()
to get Status.State and Status.Health of the fan resource
- Added state mapping for `Available` to Redfish `UnavailableOffline`
Tested:
- Unit tests passed
- Redfish Service Validator passed
- Response as expected:
```
curl -k -v https://${bmc}/redfish/v1/Chassis/chassis/ThermalSubsystem/Fans/fan0
```
Results in
```
{
"@odata.id": "/redfish/v1/Chassis/chassis/ThermalSubsystem/Fans/fan0",
"@odata.type": "#Fan.v1_3_0.Fan",
"Id": "fan0",
"Location": {
"PartLocation": {
"ServiceLabel": "U78DA.N00.WZS003H-A0"
}
},
"Manufacturer": "Delta",
"Model": "7B5F",
"Name": "fan0",
"PartNumber": "02YK323",
"SerialNumber": "YL12JP1C1234",
"SparePartNumber": "02YK323",
"Status": {
"Health": "OK",
"State": "UnavailableOffline"
}
```
Where the state can be `Enabled`, `UnavailableOffline`, or `Absent`
Change-Id: I21b257c39735b2f3d519549aa52a8c6a0cc48ac9
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
- State `Available` will be mapped to `UnavailableOffline` for
Status.State for resources that are present but unavailable
- Added a utility function to reduce code duplication for retrieving
resource's Status.State (`getResourceState`) and Status.Health
(`getResourceHealth`)
Tested:
- Compiles and builds
- Unit tests pass
Change-Id: I1164a9412cbd4e1ab8938366ffc7ea9c7fe58786
Signed-off-by: Justin Nguyen <justinnanguyen@gmail.com>
|
|
The getFanPaths function contained an inline lambda that triggered
the long-lambda ast-grep rule, requiring an inline suppression
(// ast-grep-ignore: long-lambda).
This commit extracts the lambda into a separate afterGetFanPaths
function and uses std::bind_front to handle the asynchronous
callback. This aligns the code with the preferred bmcweb pattern for
D-Bus callbacks, improves readability, and allows the removal of the
static analysis suppression.
Change-Id: Ibb72a939590077f1f8b4f19dfc5809d3411d0742
Signed-off-by: Yuvakumar Selvamani <yuvakumars@ami.com>
|
|
This commit makes two minor cleanups to the Thermal resource:
1. Removes an obsolete `TODO` comment regarding Chassis Redundancy
information in the GET handler. The redundancy data is already
being correctly populated for Thermal resources via the
`populateFanRedundancy()` function called within `sensors.hpp`.
2. Fixes a copy-paste error in the PATCH handler. If an empty payload
was provided, the `resourceNotFound` message incorrectly referenced
"Voltages" (which belongs to the Power schema) instead of "Fans".
Change-Id: Icb35f1e85fe56895dcc47c6241a556da3a3238d9
Signed-off-by: Shivraj Toravi <shivrajnt@ami.com>
|
|
Read FullDuplex from D-Bus and expose it in the Redfish
EthernetInterface GET response.
Tested:
Get /redfish/v1/Managers/bmc/EthernetInterfaces/eth0
{
"@odata.id": "/redfish/v1/Managers/bmc/EthernetInterfaces/eth0",
"@odata.type": "#EthernetInterface.v1_9_0.EthernetInterface",
"DHCPv4": {
"DHCPEnabled": true,
"UseDNSServers": true,
"UseDomainName": true,
"UseNTPServers": true
},
"DHCPv6": {
"OperatingMode": "Enabled",
"UseDNSServers": true,
"UseDomainName": true,
"UseNTPServers": true
},
"Description": "Management Network Interface",
"EthernetInterfaceType": "Physical",
"FQDN": "evb-ast2600",
"FullDuplex": true,
"HostName": "evb-ast2600",
"IPv4Addresses": [
{
"Address": "x.x.x.x",
"AddressOrigin": "DHCP",
"Gateway": "x.x.x.x",
"SubnetMask": "255.255.255.0"
}
],
"IPv4StaticAddresses": [],
"IPv6AddressPolicyTable": [],
"IPv6Addresses": [
{
"Address": "fe80::xxxx:xxxx:xxxx:xxxx",
"AddressOrigin": "LinkLocal",
"PrefixLength": 64
}
],
"IPv6DefaultGateway": "0:0:0:0:0:0:0:0",
"IPv6StaticAddresses": [],
"IPv6StaticDefaultGateways": [],
"Id": "eth0",
"InterfaceEnabled": true,
"LinkStatus": "LinkUp",
"MACAddress": "xx:xx:xx:xx:xx:xx",
"MTUSize": 1500,
"Name": "Manager Ethernet Interface",
"NameServers": [
"x.x.x.x"
],
"SpeedMbps": 1000,
"StatelessAddressAutoConfig": {
"IPv6AutoConfigEnabled": true
},
"StaticNameServers": [],
"Status": {
"State": "Enabled"
}
- Redfish Service Validator v3.1.6 run against
tree: /redfish/v1/Managers/bmc/EthernetInterfaces
PASS: 199, WARN: 0, FAIL: 0, NOT TESTED: 98
Change-Id: I3892a75b6a023fcd435f06074011e5a199d11c08
Signed-off-by: Rashid MP <rashidmp@ami.com>
|
|
Several utility functions return State and Health values as
std::string literals, which lacks compile-time type safety.
This can lead to typos or invalid values that are only caught
at runtime.
Convert these functions to return resource::State and
resource::Health enums instead. These enums are already
defined in resource.hpp with NLOHMANN_JSON_SERIALIZE_ENUM
mappings and are used extensively throughout the codebase.
The enums automatically serialize to the same JSON strings, so
there is no functional change. This provides compile-time type
checking.
Tested:
- No functional change.
- Verified State and Health properties serialize to identical
JSON values.
- Redfish Service Validator: PASS: 5788, WARN: 353, FAIL: 0,
NOT TESTED: 4878
Change-Id: I590fc3dddac6f126c26814b549d2fea961409170
Signed-off-by: Vinothkumar Shanmugavel <vinothkumars@ami.com>
|
|
The handleThermalMetricsHead function contained an inline lambda that
triggered the long-lambda ast-grep rule, requiring an inline
suppression (// ast-grep-ignore: long-lambda).
This commit extracts the lambda into a separate doThermalMetricsHead
function and uses std::bind_front to handle the asynchronous callback.
This aligns the code with the preferred bmcweb pattern for D-Bus
callbacks, improves readability, and allows the removal of the static
analysis suppression.
Change-Id: I656e9d03e0780ac252d8778c825c12def9eac89e
Signed-off-by: Shivraj Toravi <shivrajnt@ami.com>
|
|
afterSnmpClientCreate() had no handler for ObjectAlreadyExists from
the D-Bus Network.Client.Create method. When phosphor-snmp detected a
duplicate SNMP client (same address and port), it threw
ObjectAlreadyExists (errno EEXIST). Since EEXIST != EBADR, the
existing fallthrough logic incorrectly returned
PropertyValueOutOfRange with "SNMPv2c"/"Protocol" instead of
reporting the actual cause.
Add an explicit check for ObjectAlreadyExists before the EBADR check
and map it to the correct Redfish response (resourceAlreadyExists).
Related:
https://gerrit.openbmc.org/c/openbmc/phosphor-dbus-interfaces/+/90890
https://gerrit.openbmc.org/c/openbmc/phosphor-snmp/+/92218
Change-Id: Ia038795366bf7bf6acf9adcd09ae24d4c56e7172
Signed-off-by: Vijaysankar Ravi <vijaysankarr@ami.com>
|
|
redfish-validator complained that there was a link in ComputerSystem
resource which is 404, the Log Services.
```
/redfish/v1/Systems/Tyan_S8030_Baseboard
Warnings
Warning: Schema not found for HostedStorageServices.HostedStorageServices
Errors
Entity Get Error: GET of resource LogServices at URI /redfish/v1/Systems/Tyan_S8030_Baseboard/LogServices returned HTTP error. Check URI.
Property Validation Error: LogServices
/redfish/v1/Systems/Tyan_S8030_Baseboard/LogServices
Errors
Get URI Error: URI did not return resource /redfish/v1/Systems/Tyan_S8030_Baseboard/LogServices
```
Tested: on Tyan S8030
set `-Dexperimental-redfish-multi-computer-system=enabled`.
Before patching, there is a link to nonexistent LogServices [1]
`"/redfish/v1/Systems/Tyan_S8030_Baseboard/LogServices"`:
```
curl --insecure --user root:root https://${bmc}/redfish/v1/Systems/Tyan_S8030_Baseboard
{
"@odata.id": "/redfish/v1/Systems/Tyan_S8030_Baseboard",
"@odata.type": "#ComputerSystem.v1_22_0.ComputerSystem",
"Actions": {
"#ComputerSystem.Reset": {
"@Redfish.ActionInfo": "/redfish/v1/Systems/Tyan_S8030_Baseboard/ResetActionInfo",
"target": "/redfish/v1/Systems/Tyan_S8030_Baseboard/Actions/ComputerSystem.Reset"
}
},
"Boot": {
"AutomaticRetryConfig": "RetryAttempts",
"AutomaticRetryConfig@Redfish.AllowableValues": [
"Disabled",
"RetryAttempts"
],
"BootSourceOverrideEnabled": "Disabled",
"BootSourceOverrideMode": "UEFI",
"BootSourceOverrideMode@Redfish.AllowableValues": [
"Legacy",
"UEFI"
],
"BootSourceOverrideTarget": "None",
"BootSourceOverrideTarget@Redfish.AllowableValues": [
"None",
"Pxe",
"Hdd",
"Cd",
"Diags",
"BiosSetup",
"Usb"
],
"StopBootOnFault": "Never",
"TrustedModuleRequiredToBoot": "Disabled"
},
"Description": "Computer System",
"GraphicalConsole": {
"ConnectTypesSupported": [
"KVMIP"
],
"MaxConcurrentSessions": 4,
"ServiceEnabled": true
},
"HostWatchdogTimer": {
"FunctionEnabled": false,
"Status": {
"State": "Enabled"
},
"TimeoutAction": "PowerCycle"
},
"Id": "Tyan_S8030_Baseboard",
"LastResetTime": "2026-01-26T02:23:22+00:00",
"Links": {
"ManagedBy": [
{
"@odata.id": "/redfish/v1/Managers/bmc"
}
]
},
"LocationIndicatorActive": false,
"LogServices": {
"@odata.id": "/redfish/v1/Systems/Tyan_S8030_Baseboard/LogServices"
},
"MemorySummary": {
"TotalSystemMemoryGiB": 0.0
},
"Name": "Tyan_S8030_Baseboard",
"PowerRestorePolicy": "AlwaysOff",
"PowerState": "Off",
"ProcessorSummary": {
"Count": 0
},
"SerialConsole": {
"IPMI": {
"ServiceEnabled": true
},
"MaxConcurrentSessions": 15,
"SSH": {
"HotKeySequenceDisplay": "Press ~. to exit console",
"Port": 2200,
"ServiceEnabled": true
}
},
"Status": {
"Health": "OK",
"State": "Disabled"
},
"SystemType": "Physical"
}
```
After patching, the invalid link is gone:
```
curl --insecure --user root:root https://${bmc}/redfish/v1/Systems/Tyan_S8030_Baseboard
{
"@odata.id": "/redfish/v1/Systems/Tyan_S8030_Baseboard",
"@odata.type": "#ComputerSystem.v1_22_0.ComputerSystem",
"Actions": {
"#ComputerSystem.Reset": {
"@Redfish.ActionInfo": "/redfish/v1/Systems/Tyan_S8030_Baseboard/ResetActionInfo",
"target": "/redfish/v1/Systems/Tyan_S8030_Baseboard/Actions/ComputerSystem.Reset"
}
},
"Boot": {
"AutomaticRetryConfig": "RetryAttempts",
"AutomaticRetryConfig@Redfish.AllowableValues": [
"Disabled",
"RetryAttempts"
],
"BootSourceOverrideEnabled": "Disabled",
"BootSourceOverrideMode": "UEFI",
"BootSourceOverrideMode@Redfish.AllowableValues": [
"Legacy",
"UEFI"
],
"BootSourceOverrideTarget": "None",
"BootSourceOverrideTarget@Redfish.AllowableValues": [
"None",
"Pxe",
"Hdd",
"Cd",
"Diags",
"BiosSetup",
"Usb"
],
"StopBootOnFault": "Never",
"TrustedModuleRequiredToBoot": "Disabled"
},
"Description": "Computer System",
"GraphicalConsole": {
"ConnectTypesSupported": [
"KVMIP"
],
"MaxConcurrentSessions": 4,
"ServiceEnabled": true
},
"HostWatchdogTimer": {
"FunctionEnabled": false,
"Status": {
"State": "Enabled"
},
"TimeoutAction": "PowerCycle"
},
"Id": "Tyan_S8030_Baseboard",
"LastResetTime": "2026-01-26T02:23:22+00:00",
"Links": {
"ManagedBy": [
{
"@odata.id": "/redfish/v1/Managers/bmc"
}
]
},
"LocationIndicatorActive": false,
"MemorySummary": {
"TotalSystemMemoryGiB": 0.0
},
"Name": "Tyan_S8030_Baseboard",
"PowerRestorePolicy": "AlwaysOff",
"PowerState": "Off",
"ProcessorSummary": {
"Count": 0
},
"SerialConsole": {
"IPMI": {
"ServiceEnabled": true
},
"MaxConcurrentSessions": 15,
"SSH": {
"HotKeySequenceDisplay": "Press ~. to exit console",
"Port": 2200,
"ServiceEnabled": true
}
},
"Status": {
"Health": "OK",
"State": "Disabled"
},
"SystemType": "Physical"
}
```
References:
[1] https://github.com/openbmc/bmcweb/blob/ac69e77ea938e8c2a5bd3dad258cdafd9d3182b7/redfish-core/lib/log_services.hpp#L910
Change-Id: I60d1e5f6e7d3f0c293c389f90b1d45203143c08c
Signed-off-by: Alexander Hansen <alexander.hansen@9elements.com>
|
|
The ClearLog action returns a JSON Redfish response, but it checks the
Accept header for application/octet-stream. This causes requests with
Accept: application/json to be marked as Bad Request.
The handler does not return after setting the error status, so it still
clears the log and returns a JSON success message. Remove the invalid
check to return the correct status.
Change-Id: Ica492e3185f970d39dd1655165ba623c5e7e29a4
Signed-off-by: Rex Fu <rex.fu@amd.com>
|
|
The afterGetUUID and afterGetInventory handlers set UUID and
SubModel properties without validating D-Bus returned values are
non-empty. Per Redfish schema, UUID must be a valid RFC4122
identifier and empty strings are invalid. This adds empty string
checks to prevent setting invalid property values in the Redfish
response.
When D-Bus returns empty strings for these properties (which can
occur during early boot or with certain hardware configurations),
they will now be omitted from the response rather than returned
as empty values.
Tested:
- Verified properties are omitted when D-Bus returns empty string
- Verified properties are correctly set when valid values present
- No functional change when valid non-empty values are returned
Change-Id: I80b4f8588cb065d25097f4fd3712987adef820c0
Signed-off-by: Vinothkumar Shanmugavel <vinothkumars@ami.com>
|
|
Currently the BiosReset action is only supported on openpower platforms,
as the implementation calls into an org.openpower owned dbus service.
As part of this patch series [1], a new cmos-reset service is added to
platform-init that also should be requested via said Redfish action,
independent of the platform.
To unify, call into a xyz.openbmc_project owned service. Accordingly,
adpapt the busname in openpower-pnor-code-mgmt [2].
Tested: In order to test the openpower case, built a romulus image
based on this patch and the patch in openpower-pnor-code-mgmt.
Made a POST request and checked the journal that the request triggered
the service.
Limitation of this testing is that the dbus call errors out as expected
with "Failed to reset bios: No route to host".
I don't have access to an openpower platform, but if real hw testing is
needed I could ask around on discord if someone could run the code on
their machine.
All testing for the patch series in platform-init has already been
tested through Redfish.
[1] https://gerrit.openbmc.org/c/openbmc/platform-init/+/88369
[2] https://gerrit.openbmc.org/c/openbmc/openpower-pnor-code-mgmt/+/88228
Change-Id: I5a52f138848063814e08b152756b8d5d8c31e267
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>
|
|
For some reason, validator run at the recent bmcweb bump[1] wasn't able
to download the schemas and hit:
```
/redfish/v1/Fabrics
FabricCollection
Pass: 5
Failure: 1
Show Results
Show Payload
NameValueResult
Schema Error: Unable to locate the schema definition for the 'Fabric.Fabric' type.FAILMembers@odata.count1PASSMembers@odata.nextLink[Not Present]SKIPNameFabric CollectionPASSOem[Not Present]SKIP
```
Fabric was added at 3132dacead0.
Long term it would be nice to get rid of this only install certain
schemas.. Separate discussion.
[1]: https://gerrit.openbmc.org/c/openbmc/openbmc/+/93063
Commands ran:
cd redfish-core/schema/dmtf/json-schema-installed
ln -fs ../json-schema/Fabric.v1_4_0.json Fabricv1_4_0.json
cd redfish-core/schema/dmtf/installed
ln -fs ../csdl/Fabric_v1.xml Fabric_v1.xml
Tested: Inspection only.
Change-Id: I8a36b978b7228238f7011beefd7d03f24a7839b1
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>
|
|
Added early-return checks in AccountService PATCH handling to avoid
continuing processing after an error response is already set.
Earlier, when no auth-method fields were present in the request,
the auth patch flow could still contribute success semantics.
In error scenarios, this could produce mixed outcomes in one response
(for example, Success with PropertyNotWritable for MaxPasswordLength)
and could also cause PATCH requests for valid writable properties to
appear to hang or time out after an earlier failure path.
With this change:
- Auth-method processing returns immediately when no auth fields are
present.
- HTTP 200 is set only when status is still OK.
- Processing stops immediately after validation/helper errors.
- Read-only property updates such as MaxPasswordLength return only
PropertyNotWritable, as expected.
No Redfish schema or property additions in this commit.
Behavior-only change in AccountService PATCH handling.
Tested:
- EVB2600
PATCH /redfish/v1/AccountService with {"MaxPasswordLength":30}
response contains only PropertyNotWritable.
- EVB2600
PATCH with valid writable properties, for example:
{"AccountLockoutDuration":301,"AccountLockoutThreshold":32}
request completes normally without hang/abrupt client close.
- Redfish-Service-Validator: PASS.
Change-Id: Ie6e66c432d38ef29f4aadd70067440d15a93a95f
Signed-off-by: Sandeep Akant Patil <sandeepap@ami.com>
|
|
The SubProcessors core is a collection under the processor
schema. The association objects, (containing, contained_by), are used
to link the processor core.
The association between processor and core have been documented in
phosphor-dbus-interfaces [1]
[1] https://github.com/openbmc/phosphor-dbus-interfaces/commit/8c79b1dc0270d01c0b713a345c8ec39533c542e4
Tested: Redfish Validator Passed
```
curl -k -X GET https://${bmc}/redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core0
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core0",
"@odata.type": "#Processor.v1_18_0.Processor",
"Id": "core0",
"Name": "SubProcessor",
"ProcessorType": "Core"
}
```
Verified that below return a link header
```
- GET /redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core0
- HEAD /redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core0
```
Change-Id: I8cee9909ce20fc0bfdd56fb4fc992163546be180
Signed-off-by: George Liu <liuxiwei@inspur.com>
Signed-off-by: Myung Bae <myungbae@us.ibm.com>
|
|
The SubProcessors is a collection under the processor collection
schema. The association objects, (containing, contained_by), are
used to link the processor.
The association between processor and core have been documented in
phosphor-dbus-interfaces [1]
[1] https://github.com/openbmc/phosphor-dbus-interfaces/commit/8c79b1dc0270d01c0b713a345c8ec39533c542e4
Tested:
- GET cpu and cpu subprocessors
```
curl -k -X GET https://${bmc}/redfish/v1/Systems/system/Processors/cpu0
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0",
"@odata.type": "#Processor.v1_18_0.Processor",
"Id": "cpu0",
...
"SubProcessors": {
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0/SubProcessors"
},
...
}
```
```
curl -k -X GET https://${bmc}/redfish/v1/Systems/system/Processors/cpu0/SubProcessors
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0/SubProcessors",
"@odata.type": "#ProcessorCollection.ProcessorCollection",
"Members": [
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core0"
},
{
"@odata.id": "/redfish/v1/Systems/system/Processors/cpu0/SubProcessors/core1"
},
...
],
"Members@odata.count": 4,
"Name": "SubProcessor Collection"
}
```
- Verified that below return a link header
- GET /redfish/v1/Systems/system/Processors/cpu0/SubProcessors
- HEAD /redfish/v1/Systems/system/Processors/cpu0/SubProcessors
- Redfish Validator Passed
Change-Id: If155b97b0c782d82541c00ecf5ee70cb0180f71f
Signed-off-by: George Liu <liuxiwei@ieisystem.com>
Signed-off-by: Nikhil Namjoshi <nikhilnamjoshi@google.com>
Signed-off-by: Myung Bae <myungbae@us.ibm.com>
|
|
Copy the latest format file from the docs repository and apply using
clang-format-21.
See I795e88ada8e676c242b5a18888ce9c08afdedc93 for clang-21 enablement.
Change-Id: I08e1dad55a6dd1bc6b56fc47d2ddef7c9bbebdd7
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
Enforce the policy[1] which requires that subproject maintainers are
only list developers who have been active in the project.
Removed developers that have not been active per that policy from
OWNERS.
[1]: https://github.com/openbmc/docs/commit/1b531cd0d861d9c1d9ff81fb782e43f15c3c48d9
Change-Id: I752a375a9f080ce4809387405c7ac8b245825145
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
|
|
https://gerrit.openbmc.org/c/openbmc/bmcweb/+/92936 fixed the missing
return and corrected the property name spelling from 'HttpBasicAuth' to
'HTTPBasicAuth' but incorrectly swapped the arguments to
propertyValueNotInList.
The signature is:
propertyValueNotInList(res, arg1=<value>, arg2=<property name>)
Per the Redfish Base registry:
Message: "The value '%1' for the property %2 is not in the list of
acceptable values."
ArgDescriptions:
- %1: The value provided for the property.
- %2: The name of the property.
https://redfish.dmtf.org/registries/Base.1.23.0.json
Tested: Inspection only.
Change-Id: I5bbf80291abe1150fb6bd83722c6bcf9152b8472
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>
|
