Unverified Commit 4e29dd38 authored by Vadim Pasternak's avatar Vadim Pasternak Committed by Ilpo Järvinen
Browse files

Documentation/ABI: Add new attribute for mlxreg-io sysfs interfaces



Add documentation for the new attributes:
- Request and response for access to protetced flashes:
  "global_wp_request", "global_wp_response".
  Only for systems equipped with BMC - grant can be provided only by
  BMC in case its security policy allows to grant access.
- Request to unlock ASICs, which has been shutdown due-to ASIC thermal
  event: "shutdown_unlock".
- Data processor Units (DPU) boot progress: "boot_progress".
- DPU reset causes: "reset_aux_pwr_or_reload", "reset_dpu_thermal",
  "reset_from_main_board".
- Reset control for DPU components: "perst_rst", "phy_rst", "tpm_rst",
  "usbphy_rst".
- DPU Unified Fabric Manager upgrade - "ufm_upgrade".
- Hardware Id of Data Process Unit board - "dpu_id".

Reviewed-by: default avatarMichael Shych <michaelsh@nvidia.com>
Signed-off-by: default avatarVadim Pasternak <vadimp@nvidia.com>
Link: https://lore.kernel.org/r/20250504165507.9003-3-vadimp@nvidia.com


Reviewed-by: default avatarIlpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: default avatarIlpo Järvinen <ilpo.jarvinen@linux.intel.com>
parent 56b0bb7f
Loading
Loading
Loading
Loading
+98 −0
Original line number Diff line number Diff line
@@ -715,3 +715,101 @@ Description: This file shows 1 in case the system reset happened due to the
		switch board.

		The file is read only.

What:		/sys/devices/platform/mlxplat/mlxreg-io/hwmon/hwmon*/global_wp_request
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	This file when written 1 activates request to allow access to
		the write protected flashes. Such request can be performed only
		for system equipped with BMC (Board Management Controller),
		which can grant access to protected flashes. In case BMC allows
		access - it will respond with "global_wp_response". BMC decides
		regarding time window of granted access. After granted window is
		expired, BMC will change value back to 0.
		Default value is 0.

		The file is read/write.

What:		/sys/devices/platform/mlxplat/mlxreg-io/hwmon/hwmon*/global_wp_response
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	This file, when set 1, indicates that access to protected
		flashes have been granted to host CPU by BMC.
		Default value is 0.

		The file is read only.

What:		/sys/devices/platform/mlxplat/mlxreg-io/hwmon/hwmon*/shutdown_unlock
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	When ASICs are getting overheated, system protection
		hardware mechanism enforces system reboot. After system
		reboot ASICs come up in locked state. To unlock ASICs,
		this file should be written 1
		Default value is 0.

		The file is read/write.

What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/boot_progress
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	These files show the Data Process Unit board boot progress
		state. Valid states are:
		- 4 : OS starting.
		- 5 : OS running.
		- 6 : Low-Power Standby.

		The file is read only.

What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/dpu_id
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	This file shows hardware Id of Data Process Unit board.

		The file is read only.

What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/reset_aux_pwr_or_reload
What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/reset_dpu_thermal
What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/reset_from_main_board
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	These files expose the cause of the most recent reset of the Data
		Processing Unit (DPU) board. The possible causes are:
		- Power auxiliary outage or power reload.
		- Thermal shutdown.
		- Reset request from the main board.
		Value 1 in file means this is reset cause, 0 - otherwise. Only one of
		the above causes could be 1 at the same time, representing only last
		reset cause.

		The files are read only.

What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/perst_rst
What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/phy_rst
What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/tpm_rst
What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/usbphy_rst
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	These files allow to reset hardware components of Data Process
		Unit board. Respectively PCI, Ethernet PHY, TPM and USB PHY
		resets.
		Default values for all the attributes is 1. Writing 0 will
		cause reset of the related component.

		The files are read/write.

What:		/sys/devices/platform/mlxplat/i2c_mlxcpld.*/i2c-*/i2c-*/*-00**/mlxreg-io.*/hwmon/hwmon*/ufm_upgrade
Date:		May 2025
KernelVersion:	6.16
Contact:	Vadim Pasternak <vadimp@nvidia.com>
Description:	These files show status of Unified Fabric Manager upgrade.
		state. 0 - means upgrade is done, 1 - otherwise.

		The file is read only.