Avoid writing same data to IoExpander and ClockBuffer

During initialization hsbp-manager sets zero at bit position of clock buffer register and IO output register that has a NVMe drive in it.

Code

Avoid writing same data to IoExpander and ClockBuffer

During initialization hsbp-manager sets zero at bit position of clock buffer register and IO output register that has a NVMe drive in it.

Code changes are made to make sure that if the bit positions are
already set to zero, we can avoid the write to clock buffer register
and IO output register.

Tested:
- Made sure the working of hsbp manager is not affected.
- Made sure that write to clock buffer register and IO output register
does not happen if we restart hsbp manager service, as bits are
already set to zeros.

Change-Id: Iaa08be8eba93ecc51ef70cc4bd9fb371de6474a1
Signed-off-by: Arun Lal K M <arun.lal@intel.com>

show more ...

remove dbus interfaces on Backplane destruction

activationIface(dbus_interface) attribute of Backplane(struct) was not
getting removed from dbus on destruction of Backplane object. Later as
Backplan

remove dbus interfaces on Backplane destruction

activationIface(dbus_interface) attribute of Backplane(struct) was not
getting removed from dbus on destruction of Backplane object. Later as
Backplane object gets reconstructed for same Backplane add_interface API
throws exception since interface is still there.
Remove activation dbus_interface on destruction of Backplane object.

Led associated with a Backplane were not getting destructed on
destruction of Backplane object. This was because Led is created as
shared_ptr and Led object itself keeps one reference causing use_count
to never go to zero.
Use weak reference in Led object. This causes Led object destructor to
get called on destruction of Backplane.

Change-Id: I3499b3617810ca39c99ecd1a8de7f56c93a2a586
Signed-off-by: Vikash Chandola <vikash.chandola@intel.com>

show more ...

Add missing include

list was included through a dependency on sdbusplus which no longer
provides it, so it needs to be explicitly included here.

Change-Id: Ia39a1cd6b41c1837503ea9c0a8d272739589235f

Add missing include

list was included through a dependency on sdbusplus which no longer
provides it, so it needs to be explicitly included here.

Change-Id: Ia39a1cd6b41c1837503ea9c0a8d272739589235f
Signed-off-by: Jason M. Bills <jason.m.bills@intel.com>

show more ...

Revamp HSBP clock enabling code flow

This commit enhances the existing HSBP clock enabling logic
by eliminating most of the hardcoded register addresses, mapping
tables etc. The revamped code pulls

Revamp HSBP clock enabling code flow

This commit enhances the existing HSBP clock enabling logic
by eliminating most of the hardcoded register addresses, mapping
tables etc. The revamped code pulls all the configuration from the
Dbus object exposed by entity-manager, i.e. all the required hardware
information is moved into a configuration file, thereby reducing
hardcoding of values. This approch also makes it flexible to add
any number/type of clock buffers and also have flexibility in
mapping the drives (i.e. in case the hardware maps the HSBP drives
differently, only a configuratioon change is all that is needed
without having to touch the code).
Below is a template of configuration that needs to be appended
into the baseboard configuration file:
{
"RootI2cBus": 4,
"HsbpSupported": [
"HSBP_1",
"HSBP_2",
"HSBP_3"
],
"HSBP_1": [
"CPU0_NVME1",
"CPU0_NVME2",
"CPU0_NVME3",
"CPU0_NVME4",
"CPU0_NVME5",
"CPU0_NVME6",
"CPU0_NVME7",
"CPU0_NVME8"
],
"HSBP_2": [
"CPU1_NVME1",
"CPU1_NVME2",
"CPU1_NVME3",
"CPU1_NVME4",
"CPU1_NVME5",
"CPU1_NVME6",
"CPU1_NVME7",
"CPU1_NVME8"
],
"HSBP_3": [
"-",
"-",
"-",
"-",
"-",
"-",
"-",
"-"
],
"ClockBuffer": [
"abcd"
],
"IoExpander": [
"efgh"
],
"Name": "HSBP Config",
"Type": "HSBPConfiguration"
},
{
"Address": "0x12",
"Bus": 4,
"Mode": "SMBus",
"OutCtrlBaseAddr": "0x81",
"OutCtrlByteCount": 2,
"Byte0": [
"CPU0_NVME5",
"CPU0_NVME2",
"CPU0_NVME8",
"CPU0_NVME4",
"CPU0_NVME1",
"CPU0_NVME6",
"CPU0_NVME7",
"CPU0_NVME3"
],
"Byte1": [
"-",
"-",
"-",
"-",
"-",
"-",
"-",
"-"
],
"Name": "CPU0 Clock Buffer",
"Type": "abcd"
},
{
"Address": "0x34",
"Bus": 4,
"Mode": "IO",
"OutCtrlBaseAddr": "0x81",
"OutCtrlByteCount": 2,
"Byte0": [
"CPU1_NVME1",
"CPU1_NVME2",
"CPU1_NVME3",
"CPU1_NVME4",
"CPU1_NVME5",
"CPU1_NVME6",
"CPU1_NVME7",
"CPU1_NVME8"
],
"Byte1": [
"-",
"-",
"-",
"-",
"-",
"-",
"-",
"-"
],
"Name": "CPU1 Clock Buffer",
"Type": "abcd"
},
{
"Address": "0x56",
"Bus": 4,
"ConfIORegAddr": "0x06",
"OutCtrlBaseAddr": "0x02",
"OutCtrlByteCount": 2,
"IO0": [
"CPU1_NVME1",
"CPU1_NVME2",
"CPU1_NVME3",
"CPU1_NVME4",
"CPU1_NVME5",
"CPU1_NVME6",
"CPU1_NVME7",
"CPU1_NVME8"
],
"IO1": [
"-",
"-",
"-",
"-",
"-",
"-",
"-",
"-"
],
"Name": "IO Expander",
"Type": "efgh"
}

TESTED:
- On BMC boot, the connected HSBP drives were detected and
respective clock was enabled
- Hot Swapped couple of drives and confirmed the respective
clocks being enabled (The clocks were disabled for the Drives
which were removed)
- Verified by i2cdump of clock buffer(s) and checking respective
registers

Change-Id: Ice2b3b4f9d16df6d572ab34c8ea0cfddf657554c
Signed-off-by: P Dheeraj Srujan Kumar <p.dheeraj.srujan.kumar@intel.com>

show more ...

sdbusplus: use shorter type aliases

The sdbusplus headers provide shortened aliases for many types.
Switch to using them to provide better code clarity and shorter
lines. Possible replacements are

sdbusplus: use shorter type aliases

The sdbusplus headers provide shortened aliases for many types.
Switch to using them to provide better code clarity and shorter
lines. Possible replacements are for:
* bus_t
* exception_t
* manager_t
* match_t
* message_t
* object_t
* slot_t

Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: Idf165747d0fd6079ba217992f2c46a388ce2d906

show more ...

Fix to populate all HSBP inventory

Issue: if backplanes map cleared inside of the async_method_call,
system is not populating all HSBP inventory items in RedFish.

Fix: So moved

Fix to populate all HSBP inventory

Issue: if backplanes map cleared inside of the async_method_call,
system is not populating all HSBP inventory items in RedFish.

Fix: So moved backplanes.clear() to outside of the async_method_call.
Now RedFish populating all the HSBP inventory items.

Tested:
Verified from Redfish, shown both HSBP inventory items.

Signed-off-by: Jayaprakash Mutyala <mutyalax.jayaprakash@intel.com>
Change-Id: Idd8ca45b19e7d124a9cfbbe9bbbfb698fc289a6f

show more ...

Enable/Disable PCIe clocks on NVMe drive hot-add/remove

When the BMC detects the hot-add/hot-removal of an NVMe drive on HSBP,
the BMC shall enable/disable the clock to the NVMe drive on

Enable/Disable PCIe clocks on NVMe drive hot-add/remove

When the BMC detects the hot-add/hot-removal of an NVMe drive on HSBP,
the BMC shall enable/disable the clock to the NVMe drive on the
DB2000 clock buffer to reduce EMI.
In addition fix the issue of addition/removal events not getting
logged in System Event Log.

Tested:
1. DB2000 clock enabling/disabling on hot addition/removal of PCIe
drives tested with debug cpld image which emulates hot
addition/removal of drives. Addition/removal verified via ITP.
DB2000 clock enabling/disabling verified independently via
i2c-tools from BMC also.
2. SEL log entries on each hot-add/removal verified via redfish
Signed-off-by: Rohit Chandel <rohit.k.chandel@intel.com>
Change-Id: I5c6f8d6cf0b8fb737006b6ca14f2cfcd6a3de27d

show more ...

Remove asset when drive removed

This removes the asset interface when a drive is
removed so the old serial number and fru fields
don't show up in Redfish.

Tested: Had presen

Remove asset when drive removed

This removes the asset interface when a drive is
removed so the old serial number and fru fields
don't show up in Redfish.

Tested: Had present flip to not present after
predfined period, when not present, no fru fields
in redfish

Change-Id: Iea90509bdca26758472105c55001abac951128d3
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Log insertion/removal of all Hard Disk Drives

Insertion and removal events were only being logged for NVMe
drives. The D-Bus event system could only be enabled for NVMe
drives. The D

Log insertion/removal of all Hard Disk Drives

Insertion and removal events were only being logged for NVMe
drives. The D-Bus event system could only be enabled for NVMe
drives. The D-Bus method allowed the NVMe serial number to be added to
the insertion/removal event. SAS/SATA insertion/removal was not
detected. The only time SAS/SATA drives were detected was during HSPB
power on.

The changes here log the insertion and removal of all drives.

Tested:
Created an in code temporary test that toggled the Drive 1 Present bit
during every UpdateDrives call. This was done because physically
inserting/removing the drive is impractical.
Confirmed after each toggle there was a journal event for "Iventory
Added" and "Inventory Removed", and that the extended data included
the drive number.

Change-Id: I48ce1e0dc15cf8a42a498bf8dbe93a88ebd30a58
Signed-off-by: Johnathan Mantey <johnathanx.mantey@intel.com>

show more ...

Make hsbp-manager in charge of logging drive events

The logic was getting too complicated to get EM to log
all the drive adds and removes for some drives and not
others. This moves a

Make hsbp-manager in charge of logging drive events

The logic was getting too complicated to get EM to log
all the drive adds and removes for some drives and not
others. This moves all the logging to hsbp-manager.

This also changes the backplane object to use a
shared_ptr to make sure destruction happens correctly.

Tested: Tested using code branch that flips the bit,
sending to validation for more testing

Change-Id: I305d01374579b95dcfa16c21db0c9c70a98e8181
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Fix extra logs for NVMe drives

NVMe drives have FRUs. Because of this entity-manager
handles the add/remove messages. Make sure hsbp-manager
doesn't log for these, and that we rescan

Fix extra logs for NVMe drives

NVMe drives have FRUs. Because of this entity-manager
handles the add/remove messages. Make sure hsbp-manager
doesn't log for these, and that we rescan when they
change.

Tested: Extra logs went away, got add / remove messages
when status changed

Change-Id: Ie9b6d980be9cb557f44da1ffaed79ecc867c2d98
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add Activation interface to HSBP version object

Add activation interface to HSBP version object to represent the
'active' state of firmware.

Tested:
Flashed and check HSBP o

Add Activation interface to HSBP version object

Add activation interface to HSBP version object to represent the
'active' state of firmware.

Tested:
Flashed and check HSBP object for activation status.

Tested:
1. Verified in Redfish using below URL.
GET: https://<BMC-IP>/redfish/v1/UpdateService/FirmwareInventory/
J85894_HSBP_1
Response:
{
"@odata.context": "/redfish/v1/$metadata#SoftwareInventory.
SoftwareInventory",
"@odata.id": "/redfish/v1/UpdateService/FirmwareInventory/
J85894_HSBP_1",
"@odata.type": "#SoftwareInventory.v1_1_0.SoftwareInventory",
"Description": "HSBP image",
"Id": "J85894_HSBP_1",
"Name": "Software Inventory",
"Status": {
"Health": "OK",
"HealthRollup": "OK",
"State": "Enabled"
},
"Updateable": false,
"Version": "00.02.01"
}

Change-Id: I28e2c36594e2d1c889b3a31391f57212b045a741
Signed-off-by: jayaprakash Mutyala <mutyalax.jayaprakash@intel.com>

show more ...

Fix Software Version Interface

Should be /xyz/openbmc_project/software based on
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/28238

Tested:
Still available over redf

Fix Software Version Interface

Should be /xyz/openbmc_project/software based on
https://gerrit.openbmc-project.xyz/c/openbmc/bmcweb/+/28238

Tested:
Still available over redfish

Change-Id: I0a9823029d8ea030d1709328d893e65fbd916228
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add Asset and StorageController interface

This makes it so we can populate the storage controller
redfish schema.

Tested: Redfish validator passed

Change-Id: Ie3f5ae7e5

Add Asset and StorageController interface

This makes it so we can populate the storage controller
redfish schema.

Tested: Redfish validator passed

Change-Id: Ie3f5ae7e5abf0e0fd766491ea14939c7498828c7
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add drive logging

This adds logging for drive failures and inventory
changes.

Tested:

{
"@odata.context": "/redfish/v1/$metadata#LogEntry.LogEnt

Add drive logging

This adds logging for drive failures and inventory
changes.

Tested:

{
"@odata.context": "/redfish/v1/$metadata#LogEntry.LogEntry",
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/1565674864_1",
"@odata.type": "#LogEntry.v1_4_0.LogEntry",
"Created": "2019-08-13T05:41:04+00:00",
"EntryType": "Event",
"Id": "1565674864_1",
"Message": "Drive 2 with serial number N/A was installed.",
"MessageArgs": [
"Drive",
"2",
"N/A"
],
"MessageId": "OpenBMC.0.1.InventoryAdded",
"Name": "System Event Log Entry",
"Severity": "OK"
},
{
"@odata.context": "/redfish/v1/$metadata#LogEntry.LogEntry",
"@odata.id": "/redfish/v1/Systems/system/LogServices/EventLog/Entries/1565674987",
"@odata.type": "#LogEntry.v1_4_0.LogEntry",
"Created": "2019-08-13T05:43:07+00:00",
"EntryType": "Event",
"Id": "1565674987",
"Message": "Drive Error Occurred: Drive 1.",
"MessageArgs": [
"Drive 1"
],
"MessageId": "OpenBMC.0.1.DriveError",
"Name": "System Event Log Entry",
"Severity": "Warning"
}

Change-Id: Ia23e6f48b753cbeb36c3176c39a28f687ad7061c
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Update status in Redfish

Create associations when failed to change the
status to Warning.

Tested:

{
"@odata.context": "/redfish/v1/$metadata#Drive.Drive",

Update status in Redfish

Create associations when failed to change the
status to Warning.

Tested:

{
"@odata.context": "/redfish/v1/$metadata#Drive.Drive",
"@odata.id": "/redfish/v1/Systems/system/Storage/1/Drive/Drive_1",
"@odata.type": "#Drive.v1_7_0.Drive",
"Status": {
"Health": "Warning",
"HealthRollup": "Warning",
"State": "Enabled"
}
}

Change-Id: I3e5f87dc1253e8f24396a924e79d5474b0e5754a
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Make hsbp-manager expose all redfish properties

Hsbp manager needs to query entity-manager for nvme
fields to populate all driver interfaces. This is needed
so indexing works correct

Make hsbp-manager expose all redfish properties

Hsbp manager needs to query entity-manager for nvme
fields to populate all driver interfaces. This is needed
so indexing works correctly. This also exposes all M.2
drive interfaces at the end of the hsbp range. This allows
correct numbering. When we find a drive in entity-manager,
we backtrack from the virtual bus on the mux, to find out
what bus the mux is on, this tells us the hsbp. We then
use the virtual bus to identify the channel, this tells us
the drive index.

Tested: Saw 9 drives with a 1 hsbp and 1 M.2 configuration

Change-Id: I8e730a6a4229b3816991c629439ed4639710d114
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Handle power state changes

We need to check if power is on as the hsbp goes away
during power off. This copies the power logic from
dbus-sensors and uses it to avoid scanning when po

Handle power state changes

We need to check if power is on as the hsbp goes away
during power off. This copies the power logic from
dbus-sensors and uses it to avoid scanning when power
is off.

Tested: Warnings go away

Change-Id: Ie949e7a98eb70592318710f3eecdfe766f71f110
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add LED hooks to hsbp manager

This adds the led group interface so that they can
be activiated by NVMe daemon.

Tested: Set using dbus, went and saw drive flashing

Chang

Add LED hooks to hsbp manager

This adds the led group interface so that they can
be activiated by NVMe daemon.

Tested: Set using dbus, went and saw drive flashing

Change-Id: I2e1aa3f9338609370d9758a7c4874428dac5d4ea
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

hsbp-manager: Add drive interface to fru-less drives

Some drives don't have frus, but should still show up
in redfish. For this reason we need to populate the drive
interface when it

hsbp-manager: Add drive interface to fru-less drives

Some drives don't have frus, but should still show up
in redfish. For this reason we need to populate the drive
interface when it is not available in entity-manager.

Tested: All drives showed up in redfish

Change-Id: If3bfa1e3e9c03146a77c896a5c983fbdaff3f4f5
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Create associations for drives with inventory

This creates associations so that redfish can know which
drives map to which configuration items.

Tested:

root@intel-obmc:

Create associations for drives with inventory

This creates associations so that redfish can know which
drives map to which configuration items.

Tested:

root@intel-obmc:~# busctl introspect xyz.openbmc_project.HsbpManager /xyz/openbmc_project/inventory/item/drive/Drive_2 --no-pager
NAME TYPE SIGNATURE RESULT/VALUE FLAGS
org.freedesktop.DBus.Introspectable interface - - -
.Introspect method - s -
org.freedesktop.DBus.Peer interface - - -
.GetMachineId method - s -
.Ping method - - -
org.freedesktop.DBus.Properties interface - - -
.Get method ss v -
.GetAll method s a{sv} -
.Set method ssv - -
.PropertiesChanged signal sa{sv}as - -
xyz.openbmc_project.Association.Definitions interface - - -
.Associations property a(sss) 1 "inventory" "drive" "/xyz/openbmc_p... emits-change
xyz.openbmc_project.Inventory.Item interface - - -
.Present property b true emits-change
.PrettyName property s "Drive 2" emits-change
xyz.openbmc_project.State.Decorator.OperationalStatus interface - - -
.Functional property b true emits-change
xyz.openbmc_project.State.Drive interface - - -
.Rebuilding property b false emits-change

Also show up in mapper

Change-Id: Ia050316ac55faa89aad86567c93f9a74594e9180
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add rebuilding to hsbp manager

Interface is pushed here:
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-dbus-interfaces/+/25671

Tested: available on dbus

Change-

Add rebuilding to hsbp manager

Interface is pushed here:
https://gerrit.openbmc-project.xyz/c/openbmc/phosphor-dbus-interfaces/+/25671

Tested: available on dbus

Change-Id: I79a73bb0e62951b09f4eb8bf0594aa39916b5dd9
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

hsbp-manager: add drive presence and software version

This adds support for reading the sw version and drive
presence from the hsbp cpld.

Tested:

Version is available i

hsbp-manager: add drive presence and software version

This adds support for reading the sw version and drive
presence from the hsbp cpld.

Tested:

Version is available in redfish

root@intel-obmc:~# busctl tree xyz.openbmc_project.HsbpManager --no-pager
`-/xyz
`-/xyz/openbmc_project
|-/xyz/openbmc_project/inventory
| `-/xyz/openbmc_project/inventory/item
| |-/xyz/openbmc_project/inventory/item/drive
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_1
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_2
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_3
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_4
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_5
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_6
| | |-/xyz/openbmc_project/inventory/item/drive/Drive_7
| | `-/xyz/openbmc_project/inventory/item/drive/Drive_8
| `-/xyz/openbmc_project/inventory/item/hsbp
| `-/xyz/openbmc_project/inventory/item/hsbp/J85894_HSBP_1

root@intel-obmc:~# busctl introspect xyz.openbmc_project.HsbpManager /xyz/openbmc_project/inventory/item/hsbp/J85894_HSBP_1 --no-pager
NAME TYPE SIGNATURE RESULT/VALUE FLAGS
org.freedesktop.DBus.Introspectable interface - - -
.Introspect method - s -
org.freedesktop.DBus.Peer interface - - -
.GetMachineId method - s -
.Ping method - - -
org.freedesktop.DBus.Properties interface - - -
.Get method ss v -
.GetAll method s a{sv} -
.Set method ssv - -
.PropertiesChanged signal sa{sv}as - -
xyz.openbmc_project.Inventory.Item interface - - -
.Present property b true emits-change
.PrettyName property s "J85894 HSBP 1" emits-change
xyz.openbmc_project.Software.Version interface - - -
.Version property s "00.02.01" emits-change
.Purpose property s "xyz.openbmc_project.Software.Version... emits-change

Change-Id: I73fd2669f4a0e9b8e499107a960f2169f63873e3
Signed-off-by: James Feist <james.feist@linux.intel.com>

show more ...

Add HSBP Manager Daemon

This is the initial commit for the hsbp-manager
daemon. Currently it reads a json configuration
and detects whether the cpld is available. More
feaures wi

Add HSBP Manager Daemon

This is the initial commit for the hsbp-manager
daemon. Currently it reads a json configuration
and detects whether the cpld is available. More
feaures will be added in small patches.

Tested:

busctl introspect --no-pager xyz.openbmc_project.HsbpManager /xyz/openbmc_project/inventory/item/hsbp/J85894_HSBP_1
NAME TYPE SIGNATURE RESULT/VALUE FLAGS
org.freedesktop.DBus.Introspectable interface - - -
.Introspect method - s -
org.freedesktop.DBus.Peer interface - - -
.GetMachineId method - s -
.Ping method - - -
org.freedesktop.DBus.Properties interface - - -
.Get method ss v -
.GetAll method s a{sv} -
.Set method ssv - -
.PropertiesChanged signal sa{sv}as - -
xyz.openbmc_project.Inventory.Item interface - - -
.Present property b true emits-change
.PrettyName property s "J85894 HSBP 1" emits-change

Change-Id: Ibbce272e045f8643af74b778f273168442623919
Signed-off-by: Feist, James <james.feist@linux.intel.com>

show more ...