Fabric: add support for PCIe Switch Port URI

This patch enable support for following properties for Port of a PCIe
Switch. [1]
- PortProtocol
- PortType
- CurrentSpeedGbps
- ActiveWidth

One of the

Fabric: add support for PCIe Switch Port URI

This patch enable support for following properties for Port of a PCIe
Switch. [1]
- PortProtocol
- PortType
- CurrentSpeedGbps
- ActiveWidth

One of the devices that gets enabled with this patch is Nvidia ConnectX
devices, which are network cards featuring an integrated PCIe switch.
These devices combine both PCIe ports and network ports in a single
unit. Since such devices don't strictly qualify as Fabric Adapters, the
Switch URI is used instead of the FabricAdapter URI.

Port schema only allows certain URIs as Port URI. URI
/redfish/v1/Fabrics/{FabricId}/Switches/{SwitchId}/Ports/{PortId} seems
most appropriate choice for PCIe Switch Port. [1]

The Fabric resource is modeled similarly to the System resource, meaning
that only one Fabric resource will exist for each BMC. Route handler for
collections and each individual components are added in this patch for
each URI resource under /redfish/v1/Fabrics.

DBus Interface "xyz.openbmc_project.Inventory.Item.PCIeSwitch" is used
to identify the Switch resources. Association between Switch and Port is
`connecting` and `connected_to`.

Feature like Port Metrics properties (for PCIe Error Counters) can be
added in future at Port Metric URI.

dbus-sensors patches -
https://gerrit.openbmc.org/c/openbmc/dbus-sensors/+/84079
https://gerrit.openbmc.org/c/openbmc/dbus-sensors/+/83202

Tested: Build an image for nvl32-obmc machine with the following patch
cherry picked.

https://gerrit.openbmc.org/c/openbmc/dbus-sensors/+/84079
https://gerrit.openbmc.org/c/openbmc/openbmc/+/85490

The openbmc patch cherry-picks the following patches that are currently
under review.

```
1. device tree
https://lore.kernel.org/all/aRbLqH8pLWCQryhu@molberding.nvidia.com/
2. mctpd patches
https://github.com/CodeConstruct/mctp/pull/85
3. u-boot changes
https://lore.kernel.org/openbmc/20251121-msx4-v1-0-fc0118b666c1@nvidia.com/T/#t
4. kernel changes as specified in the openbmc patch (for espi)
5. entity-manager changes
https://gerrit.openbmc.org/c/openbmc/entity-manager/+/85455
6. platform-init changes
https://gerrit.openbmc.org/c/openbmc/platform-init/+/85456
7. spi changes
https://lore.kernel.org/all/20251121-w25q01jv_fixup-v1-1-3d175050db73@nvidia.com/
```

redfish service validator is passing.

```
$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/
{
"@odata.id": "/redfish/v1/Fabrics",
"@odata.type": "#FabricCollection.FabricCollection",
"Members": [
{
"@odata.id": "/redfish/v1/Fabrics/fabric"
}
],
"Members@odata.count": 1,
"Name": "Fabric Collection"
}%

$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/fabric/
{
"@odata.id": "/redfish/v1/Fabrics/fabric",
"@odata.type": "#Fabric.v1_2_0.Fabric",
"Id": "fabric",
"Name": "fabric Fabric",
"Switches": {
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches"
}
}%

$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/fabric/Switches/
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches",
"@odata.type": "#SwitchCollection.SwitchCollection",
"Members": [
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0"
},
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_1"
},
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_2"
},
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_3"
}
],
"Members@odata.count": 4,
"Name": "fabric Switch Collection"
}%

$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0",
"@odata.type": "#Switch.v1_7_0.Switch",
"Id": "Nvidia_ConnectX_0",
"Name": "Nvidia_ConnectX_0",
"Ports": {
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports"
},
"Status": {
"Health": "OK",
"State": "Enabled"
}
}%

$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports",
"@odata.type": "#PortCollection.PortCollection",
"Members": [
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/DOWN_0"
},
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/DOWN_1"
},
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/UP_0"
}
],
"Members@odata.count": 3,
"Name": "Nvidia_ConnectX_0 Port Collection"
}%

$ curl -k -u 'root:0penBmc' https://${bmc_ip}/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/UP_0/
{
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/UP_0",
"@odata.type": "#Port.v1_4_0.Port",
"ActiveWidth": 8,
"CurrentSpeedGbps": 32.0,
"Id": "UP_0",
"Metrics": {
"@odata.id": "/redfish/v1/Fabrics/fabric/Switches/Nvidia_ConnectX_0/Ports/UP_0/Metrics"
},
"Name": "Nvidia_ConnectX_0 UP_0 Port",
"PortProtocol": "PCIe",
"PortType": "UpstreamPort",
"Status": {
"Health": "OK",
"State": "Enabled"
}
}%
```

[1]: https://redfish.dmtf.org/schemas/v1/Port_v1.xml

Change-Id: I52f4ca62b4953f6196c589e340602a0d7885d9c1
Signed-off-by: Harshit Aghera <haghera@nvidia.com>

show more ...

Do not allow data beyond the trailer

There is nothing in the multipart spec[1] that states that a parser
should allow any bytes after a multipart payload.

Several unit tests have a \r\n after their

Do not allow data beyond the trailer

There is nothing in the multipart spec[1] that states that a parser
should allow any bytes after a multipart payload.

Several unit tests have a \r\n after their boundary condition that
previously the parser just ignored. Testing shows this is fairly
normal, so handle both cases still, but if any other characters show up,
fail the parse.

Unit test is also simplified to be more clear.

Tested: Unit test coverage

[1] https://datatracker.ietf.org/doc/html/rfc7578#section-4.1

Change-Id: I16643c61867708886cc87c236447ec1c19bf934f
Signed-off-by: Ed Tanous <etanous@nvidia.com>

show more ...

Don't use at for field access in tests

at() throws an exception when a field doesn't exist. This is somewhat
paradoxical when using EXPECT_EQ, as that is not supposed to stop the
test on a failure.

Don't use at for field access in tests

at() throws an exception when a field doesn't exist. This is somewhat
paradoxical when using EXPECT_EQ, as that is not supposed to stop the
test on a failure. Convert calls to use operator[] which does not
throw.

Tested: Unit test

Change-Id: I9f1af1732294f2755f0beb422cffe0eb379f4e76
Signed-off-by: Ed Tanous <etanous@nvidia.com>

show more ...

Remove old login routines

At one point the bmcweb /login route has supported a bunch of
different login mechanisms:
1. Sending username and password via headers, mirroring a non-openbmc
implement

Remove old login routines

At one point the bmcweb /login route has supported a bunch of
different login mechanisms:
1. Sending username and password via headers, mirroring a non-openbmc
implementation that has never been used in openbmc, and predates this
repo being called bmcweb.
2. Sending username and password under a "data" object, mirroring
phosphor-rest json webserver.
3. Sending username and password as an http multipart as part of the
initial multipart parser patch.
4. Sending a json payload as {"username": <>, "password": <>}

This commit removes all but the final login mechanism. Redfish login
has been used exclusively for many many years, and only the webui used
4 above up until a few years ago in commit
1ff8e89fd2397c468ab0237158e5aeeff2692413

Keeping one viable login mechanism that's not Redfish is useful in the
future if there are those that want to disable Redfish. All others
could potentially be security issues, so keeping the code simple and
conscise here is ideal.

This commit does not attempt to make a backwards compatibility path,
under the assumption that other than 4, none of the other code was ever
used in any real capacity. If we find that to not be true, we can add
back the portions where we need compatibility, but this seems unlikely.

Tested: WIP

Change-Id: I04b4968836f0f824f46b3dff180ad92feb16967c
Signed-off-by: Ed Tanous <etanous@nvidia.com>

show more ...

memory.hpp: Correct spelling of volatileRegionMaxSizeInKiB

Correct the misspelled variable name 'voltaileRegionMaxSizeInKib' to
'volatileRegionMaxSizeInKiB' for consistency and clarity. The typo
was

memory.hpp: Correct spelling of volatileRegionMaxSizeInKiB

Correct the misspelled variable name 'voltaileRegionMaxSizeInKib' to
'volatileRegionMaxSizeInKiB' for consistency and clarity. The typo
was missing the letter 't' in 'volatile' and had inconsistent
capitalization ('Kib' vs 'KiB')

Tested: code compile done.No functional change

Change-Id: Ie5b49e6541f4cf0000a5176669017c187c6822d2
Signed-off-by: Ramya Sivakumar <sramya@ami.com>

show more ...

test: Add Cables and JsonSchemas to service root test

Add test assertions for Cables and JsonSchemas endpoints in the
ServiceRoot response. These verify that both endpoints are properly
exposed with

test: Add Cables and JsonSchemas to service root test

Add test assertions for Cables and JsonSchemas endpoints in the
ServiceRoot response. These verify that both endpoints are properly
exposed with their respective @odata.id URIs at /redfish/v1/Cables
and /redfish/v1/JsonSchemas

Change-Id: Icec5d2c5a99872f6a44a6c82f0d5ab02490d9339
Signed-off-by: Ramya Sivakumar <sramya@ami.com>

show more ...

Remove usages of nlohmann::json::begin()

nlohmann::json::begin() throws an uncaught exception.

Tested: Redfish service validator passes.

Signed-off-by: Ed Tanous <ed@tanous.net>
Change-Id: I08244b

Remove usages of nlohmann::json::begin()

nlohmann::json::begin() throws an uncaught exception.

Tested: Redfish service validator passes.

Signed-off-by: Ed Tanous <ed@tanous.net>
Change-Id: I08244b0787cd4d6e592b0731196490a5160aba62

show more ...

Move clang-off / clang-on to catch scope

As a comment in https://gerrit.openbmc.org/c/openbmc/bmcweb/+/86868
suggested "You might also consider moving this up a line so you catch
the scope on both s

Move clang-off / clang-on to catch scope

As a comment in https://gerrit.openbmc.org/c/openbmc/bmcweb/+/86868
suggested "You might also consider moving this up a line so you catch
the scope on both sides."

This enforces just the slightest bit more clang-format.

Rerun script.

Tested: Builds. Manual changes to script only. Rest generated.

Change-Id: I8ff01befc56c576716f5d83bd90763354b1ff0c5
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

Rename switch namespace

switch is a reserved keyword in C++.. Therefore "namespace switch" is
illegal C++ code.

Add a couple lines of python code to check for keywords and then rename
with a rf_ at

Rename switch namespace

switch is a reserved keyword in C++.. Therefore "namespace switch" is
illegal C++ code.

Add a couple lines of python code to check for keywords and then rename
with a rf_ at the front, e.g. namespace rf_switch.

Rerun the script update_schemas.py script.

This showed downstream, but probably would have upstream too.

```
switch.hpp:8:11: error: expected identifier or '{' [clang-diagnostic-error]
8 | namespace switch
| ^
.../bmcweb/redfish-core/include/generated/enums/switch.hpp:8:11: error: expected unqualified-id [clang-diagnostic-error]

...

FAILED: meson-internal__clang-tidy-fix
```

Change-Id: I58be67fc0df6e5056e63da5302724e6509b7114b
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

Sensors: Add unit tests

An earlier commit [1] refactored parts of objectPropertiesToJson() into
sub-functions to make the code easier to follow. The review of that
change requested unit tests for th

Sensors: Add unit tests

An earlier commit [1] refactored parts of objectPropertiesToJson() into
sub-functions to make the code easier to follow. The review of that
change requested unit tests for these new functions. [2] This commit
adds them.

[1] https://gerrit.openbmc.org/c/openbmc/bmcweb/+/85835
[2] https://gerrit.openbmc.org/c/openbmc/bmcweb/+/85835/comments/408617a1_474b4062

Tested: Compiles and unit tests pass

Change-Id: I65e06deccb4b7194603f2749a977986bcb673cd1
Signed-off-by: Janet Adkins <janeta@us.ibm.com>

show more ...

Fix variable naming convention in afterGetUUID

Rename uUID to uuid to follow standard C++ camelCase naming
conventions for local variables. This improves code consistency
and readability

Tested:

Fix variable naming convention in afterGetUUID

Rename uUID to uuid to follow standard C++ camelCase naming
conventions for local variables. This improves code consistency
and readability

Tested:
- Built successfully with no compilation errors
- No behavioral changes, purely a refactoring commit

Change-Id: I9d72a184138dd5e4123579d04c1d5e25ed1f2b09
Signed-off-by: Ramya Sivakumar <sramya@ami.com>

show more ...

PCIe: Fix typo in variable name pciecDeviceName

Rename variable from 'pciecDeviceName' to 'pcieDeviceName' to correct
the typo (extra 'c'). This improves code readability and follows
proper naming c

PCIe: Fix typo in variable name pciecDeviceName

Rename variable from 'pciecDeviceName' to 'pcieDeviceName' to correct
the typo (extra 'c'). This improves code readability and follows
proper naming conventions

No functional change

Change-Id: Iaae7c6a0b3ac725fd393cc7db4a0f7ddd9457198
Signed-off-by: Ramya Sivakumar <sramya@ami.com>

show more ...

Move to Redfish 2025.4

One line change and rerun the script. 2025.4 includes new properties.
One use case is for Redundancy. For the complete overview see [1].

Tested: Visual and build only. In the

Move to Redfish 2025.4

One line change and rerun the script. 2025.4 includes new properties.
One use case is for Redundancy. For the complete overview see [1].

Tested: Visual and build only. In the past these have not broken things.

[1]: https://www.dmtf.org/sites/default/files/Redfish_Release_2025.4_Overview.pdf

Change-Id: Icaf710eaa99816264993a964ef9dd8a7f5e7722a
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

Modify enum generation for clang

Clang-format was unhappy with the new
redfish-core/include/generated/enums/switch.hpp in 2025.4

```
diff --git a/redfish-core/include/generated/enums/switch.hpp b/r

Modify enum generation for clang

Clang-format was unhappy with the new
redfish-core/include/generated/enums/switch.hpp in 2025.4

```
diff --git a/redfish-core/include/generated/enums/switch.hpp b/redfish-core/include/generated/enums/switch.hpp
index 6acf6e28..cabc8562 100644
--- a/redfish-core/include/generated/enums/switch.hpp
+++ b/redfish-core/include/generated/enums/switch.hpp
@@ -5,7 +5,7 @@

namespace switch
{
-// clang-format off
+ // clang-format off
```

Move turning clang-format off before the namespace fixes.
¯\_(ツ)_/¯
Do that in generate_schema_enums.py and rerun update_schemas.py.

Adding 4 spaces before the clang-format off broke other generated
enums. Broke out from 2025.4.

Tested: Visual and build only.

Change-Id: Ic03aa558b405957f15035570d376c46c545906c0
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

Bump date of redfish-use-hardcoded-system-location-indicator

The meson option redfish-use-hardcoded-system-location-indicator was
added to retain the old behavior of using hard-coded D-Bus paths for

Bump date of redfish-use-hardcoded-system-location-indicator

The meson option redfish-use-hardcoded-system-location-indicator was
added to retain the old behavior of using hard-coded D-Bus paths for
getting and setting the LEDs for the Systems. [1] It was added with the
default state of enabled with the documented plan to alter the default
state in October 2025.

The support in entity-manager for the LED associations needs to be added
before a change to the default option can be made. The date is being
bumped to provide time for the necessary entity-manager changes.

Tested:
- Compiled

[1] https://github.com/openbmc/bmcweb/commit/eb261e1f5455575234246764e73ef07d4a50031b

Change-Id: If9d0610b7fb2bdc61d02bc73c0d56eb719982692
Signed-off-by: Janet Adkins <janeta@us.ibm.com>

show more ...

Fix typo: retryAttemps -> retryAttempts

Fixed incorrect variable name 'retryAttemps' to 'retryAttempts'
in the EventService PATCH handler. This typo was present in the
variable declaration and all u

Fix typo: retryAttemps -> retryAttempts

Fixed incorrect variable name 'retryAttemps' to 'retryAttempts'
in the EventService PATCH handler. This typo was present in the
variable declaration and all usages within the DeliveryRetryAttempts
parameter handling logic

Tested: Built successfully, no functional changes

Change-Id: I93699c9954c33c94d93bcc05d00c621c08aa3679
Signed-off-by: Ramya Sivakumar <sramya@ami.com>

show more ...

Use boost::urls::format

boost::urls::format is specifically meant for URL construction. It
handles encoding like percentage encoding which std::format does not.

Change-Id: I16c1467c856e36f072bd4248

Use boost::urls::format

boost::urls::format is specifically meant for URL construction. It
handles encoding like percentage encoding which std::format does not.

Change-Id: I16c1467c856e36f072bd4248186b9819e40ae228
Signed-off-by: Brad Bishop <bradbish@qti.qualcomm.com>

show more ...

Use boost::urls::format

boost::urls::format is specifically meant for URL construction. It
handles encoding like percentage encoding which std::format does not.

Change-Id: Ief7d574d1652096c2701979a

Use boost::urls::format

boost::urls::format is specifically meant for URL construction. It
handles encoding like percentage encoding which std::format does not.

Change-Id: Ief7d574d1652096c2701979a9f8fe9897957f492
Signed-off-by: Brad Bishop <bradbish@qti.qualcomm.com>

show more ...

Fix CI: Reformat using Black 26.1.0

CI is failing due to the formatting of scripts/parse_registries.py. Run
Black.

black -l 79 <filename>

Current:
Running black (black, 26.1.0 (compiled: no))

Fix CI: Reformat using Black 26.1.0

CI is failing due to the formatting of scripts/parse_registries.py. Run
Black.

black -l 79 <filename>

Current:
Running black (black, 26.1.0 (compiled: no))
reformatted scripts/parse_registries.py

Before:
Running black (black, 25.12.0 (compiled: no))

Tested: Visual only.

Change-Id: I08051019760994a095eeedab2ae0d8c91984e979
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

DEVELOPING.md: Clean up CMake reference

Switched to Meson years ago, this CMakeLists.txt got missed.

Tested: Doc change only. Copied into https://stackedit.io/ to make sure
it formatted okay.

Chan

DEVELOPING.md: Clean up CMake reference

Switched to Meson years ago, this CMakeLists.txt got missed.

Tested: Doc change only. Copied into https://stackedit.io/ to make sure
it formatted okay.

Change-Id: I66c3ac5f307b33bc65b92028a0b4c7d254a18e0c
Signed-off-by: Gunnar Mills <gmills@us.ibm.com>

show more ...

Fix Bmcweb crash due to HTTP2 connection reset

Changes added : handle dangling pointer in response completion handler

Problem : When high number of connections made to BMC and suddenly
reset those

Fix Bmcweb crash due to HTTP2 connection reset

Changes added : handle dangling pointer in response completion handler

Problem : When high number of connections made to BMC and suddenly
reset those connections, crash was observed in bmcweb.

Fix : Since the connection already torn but bmcweb trying to close those
connections in response completion handler causing the crash. Added
code to handle the closed or reset connections

Tested :

made around 200 connections and reset them immediately, repeated these
steps around 15 times no crash observed

Change-Id: I03b5fd563275c4138a6eb004e09d277547af9692
Signed-off-by: Chandra Harkude <chandramohan.harkude@gmail.com>

show more ...

Lower severity of HTTP logs from INFO to DEBUG

Couple of these logs are heavy hitters and fill up the journal
especially with external entities constantly hitting the
redfish endpoints. This change

Lower severity of HTTP logs from INFO to DEBUG

Couple of these logs are heavy hitters and fill up the journal
especially with external entities constantly hitting the
redfish endpoints. This change moves the largest two logs
we saw (on a yosemite4 machine) from INFO to DEBUG.

Change-Id: Iac7337e3040e0348e41bba79c8200320e8ba12ef
Signed-off-by: Amithash Prasad <amithash@meta.com>

show more ...

Renaming of eventlog variables and functions

Following a comment on [1], fix some of the variable and function names,
to match redfish terms.

Tested: Unit tests pass. Validation succeeded.

[1] htt

Renaming of eventlog variables and functions

Following a comment on [1], fix some of the variable and function names,
to match redfish terms.

Tested: Unit tests pass. Validation succeeded.

[1] https://gerrit.openbmc.org/c/openbmc/bmcweb/+/83354/13

Change-Id: Ie52309126c798b17136be99fd4a3d6650721257e
Signed-off-by: Oliver Brewka <oliver.brewka@9elements.com>

show more ...

Enable TCP keepalives for HTTP connections

This fix enables TCP keepalives at the OS layer. It also enables a 15
minute deadline timer at the bmcweb level when waiting on an idle HTTP
keepalive conn

Enable TCP keepalives for HTTP connections

This fix enables TCP keepalives at the OS layer. It also enables a 15
minute deadline timer at the bmcweb level when waiting on an idle HTTP
keepalive connection.

Tested: romulus image running bmcweb, start connections with keepalive,
block incoming connections `iptables -P INPUT DROP`, validate that
sockets eventually die and are tracked with keepalives `ss -nto`

Change-Id: I8f5040440348c060dae1d0516ec202a0e4dc349e
Signed-off-by: Joey Berkovitz <joey@berkovitz.us>

show more ...

Improve the efficiency of the chassis

Removed redundant inner for loop of chassis.

Tested: I watched webui still see UUID and Location
Code in chassis.

Change-Id: I4abd5bf7a46eb997ade6537c198cccde

Improve the efficiency of the chassis

Removed redundant inner for loop of chassis.

Tested: I watched webui still see UUID and Location
Code in chassis.

Change-Id: I4abd5bf7a46eb997ade6537c198cccde1abf3196
Signed-off-by: Bowei Yu <yubowei0982@phytium.com.cn>

show more ...