Re: [PATCH v2 1/2] drm/xe/xe_ras: Add types and commands for RAS GPU health indicator

["Tauro, Riana" <riana.tauro@intel.com>]

On 4/23/2026 11:09 PM, Soham Purkait wrote:
> Add System Controller GPU health status values, RAS response codes
> and mailbox payload types for querying and updating GPU health.
>
> GPU health states are encoded as numeric values defined by
> enum xe_ras_health_status for use in System Controller commands.
> The GET_HEALTH command fetches the current health state, while
> the SET_HEALTH command updates it through the System Controller
> mailbox.
>
> struct xe_ras_health_get_input and struct xe_ras_health_get_response
> describe the GET_HEALTH request and response payloads, respectively,
> while struct xe_ras_health_set_input and
> struct xe_ras_health_set_response describe the SET_HEALTH request
> and response payloads, including the operation status and current
> health state in the response.

This is extremely detailed commit message. Avoid using structs in commit 
message

My suggestion is to split patches into get/set health and add structures 
in the patch they
are being used. But upto you.

>
> v2:
>   - Add enum for health status instead of xe_ras_health_status_t.
>                                                           (Andi), (Rodrigo)
>
> Signed-off-by: Soham Purkait <soham.purkait@intel.com>
> ---
>   drivers/gpu/drm/xe/xe_ras_types.h             | 83 +++++++++++++++++++
>   drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h | 15 ++++
>   2 files changed, 98 insertions(+)
>   create mode 100644 drivers/gpu/drm/xe/xe_ras_types.h
>
> diff --git a/drivers/gpu/drm/xe/xe_ras_types.h b/drivers/gpu/drm/xe/xe_ras_types.h
> new file mode 100644
> index 000000000000..5f884d6e24de
> --- /dev/null
> +++ b/drivers/gpu/drm/xe/xe_ras_types.h
> @@ -0,0 +1,83 @@
> +/* SPDX-License-Identifier: MIT */
> +/*
> + * Copyright © 2026 Intel Corporation
> + */
> +
> +#ifndef _XE_RAS_TYPES_H_
> +#define _XE_RAS_TYPES_H_
> +
> +#include <linux/types.h>
> +
> +/* RAS response status codes */
> +enum xe_ras_response_status {
> +	XE_RAS_STATUS_SUCCESS = 0,
> +	XE_RAS_STATUS_INVALID_PARAM,
> +	XE_RAS_STATUS_OP_NOT_SUPPORTED,
> +	XE_RAS_STATUS_TIMEOUT,
> +	XE_RAS_STATUS_HARDWARE_FAILURE,
> +	XE_RAS_STATUS_INSUFFICIENT_RESOURCES,
> +	XE_RAS_STATUS_UNKNOWN_ERROR,
> +};
> +
> +/**
> + * enum xe_ras_health_status - Device health status values
> + *
> + * Health indicator status denoted by numeric values to be used in system
> + * controller mailbox commands.
> + *
> + * @XE_RAS_HEALTH_STATUS_OK: The device is healthy and operating within normal
> + * parameters.
> + * @XE_RAS_HEALTH_STATUS_WARNING: The device is experiencing minor issues but is
> + * still operational.
> + * @XE_RAS_HEALTH_STATUS_CRITICAL: The device is in a critical state and may not
> + * be operational.
> + */
> +enum xe_ras_health_status {
> +	XE_RAS_HEALTH_STATUS_OK = 0,
> +	XE_RAS_HEALTH_STATUS_WARNING,
> +	XE_RAS_HEALTH_STATUS_CRITICAL,
> +};
> +
> +/**
> + * struct xe_ras_health_get_input - Input for XE_SYSCTRL_CMD_GET_HEALTH
> + */
> +struct xe_ras_health_get_input {
> +	/** @reserved: Reserved for future use, must be 0 */
> +	u32 reserved[2];
> +} __packed;
> +
> +/**
> + * struct xe_ras_health_get_response - Response for XE_SYSCTRL_CMD_GET_HEALTH
> + */
> +struct xe_ras_health_get_response {
> +	/** @current_health: Current health status (OK/WARNING/CRITICAL) */
> +	u8 current_health;
> +	/** @reserved: Reserved for alignment */
Future use in case any other fields are added
> +	u8 reserved[3];
> +} __packed;
> +
> +/**
> + * struct xe_ras_health_set_input - Input for XE_SYSCTRL_CMD_SET_HEALTH
> + */
> +struct xe_ras_health_set_input {
> +	/** @new_health: New health status to set */
> +	u8 new_health;
> +	/** @reserved: Reserved for alignment */
> +	u8 reserved[3];
> +} __packed;
> +
> +/**
> + * struct xe_ras_health_set_response - Response for XE_SYSCTRL_CMD_SET_HEALTH
> + */
> +struct xe_ras_health_set_response {
> +	/** @operation_status: Status of set operation (RAS_STATUS) */
use enum or XE_RAS_STATUS
> +	u32 operation_status;
> +	/** @current_health: Health status after this change */
> +	u8 current_health;
> +	/** @reserved: Reserved for alignment */
> +	u8 reserved[3];
> +	/** @reserved_2: Reserved for future expansion */
> +	u32 reserved_2[2];

Nit: reserved1

> +} __packed;
> +
> +#endif /* _XE_RAS_TYPES_H_ */
> diff --git a/drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h b/drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h
> index 89456aec6097..fc73e02c3202 100644
> --- a/drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h
> +++ b/drivers/gpu/drm/xe/xe_sysctrl_mailbox_types.h
> @@ -10,6 +10,21 @@
>   
>   #include "abi/xe_sysctrl_abi.h"
>   
> +/**
> + * enum xe_sysctrl_mailbox_command_id - RAS Command ID's for GFSP group
> + *
> + * @XE_SYSCTRL_CMD_GET_HEALTH: Get current health status
> + * @XE_SYSCTRL_CMD_SET_HEALTH: Set new health status
> + */
> +enum xe_sysctrl_mailbox_command_id {
> +	XE_SYSCTRL_CMD_GET_HEALTH = 0x0B,
> +	XE_SYSCTRL_CMD_SET_HEALTH = 0x0C

Add comma so that following patches do not have to add.

> +};
> +
> +enum xe_sysctrl_group {
> +	XE_SYSCTRL_GROUP_GFSP = 1

Use 0x01. Also make this consistent with the correctable series by Raag.

Thanks
Riana

> +};
> +
>   /**
>    * struct xe_sysctrl_mailbox_command - System Controller mailbox command
>    */