[v4,1/2] ACPI: IPMI: Add helper to wait for when SMI is selected

On Dell servers, many APCI methods of acpi_power_meter module evaluate
variables inside IPMI region, so the region handler needs to be
installed. In addition to that, the handler needs to be fully
functional, and that depends on SMI being selected.

So add a helper to let acpi_power_meter know when the handler is
installed and ready to be used.

Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
---
v4:
 - Wording.
 - Define and comment on timeout value.
 - Move the completion to driver_data.
 - Remove the tenary operator.

v3:
 - New patch.

 drivers/acpi/acpi_ipmi.c | 23 ++++++++++++++++++++++-
 include/acpi/acpi_bus.h  |  5 +++++
 2 files changed, 27 insertions(+), 1 deletion(-)

On Tue, Jan 9, 2024 at 5:12 AM Kai-Heng Feng
<kai.heng.feng@canonical.com> wrote:
>
> On Dell servers, many APCI methods of acpi_power_meter module evaluate
> variables inside IPMI region, so the region handler needs to be
> installed. In addition to that, the handler needs to be fully
> functional, and that depends on SMI being selected.
>
> So add a helper to let acpi_power_meter know when the handler is
> installed and ready to be used.
>
> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>

Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

and I'm expecting this to be routed along with patch [2/2] that has
been posted elsewhere.

> ---
> v4:
>  - Wording.
>  - Define and comment on timeout value.
>  - Move the completion to driver_data.
>  - Remove the tenary operator.
>
> v3:
>  - New patch.
>
>  drivers/acpi/acpi_ipmi.c | 23 ++++++++++++++++++++++-
>  include/acpi/acpi_bus.h  |  5 +++++
>  2 files changed, 27 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/acpi/acpi_ipmi.c b/drivers/acpi/acpi_ipmi.c
> index 0555f68c2dfd..5fba4dab5d08 100644
> --- a/drivers/acpi/acpi_ipmi.c
> +++ b/drivers/acpi/acpi_ipmi.c
> @@ -22,6 +22,8 @@ MODULE_LICENSE("GPL");
>  /* the IPMI timeout is 5s */
>  #define IPMI_TIMEOUT                   (5000)
>  #define ACPI_IPMI_MAX_MSG_LENGTH       64
> +/* 2s should be suffient for SMI being selected */
> +#define ACPI_IPMI_SMI_SELECTION_TIMEOUT        (2 * HZ)
>
>  struct acpi_ipmi_device {
>         /* the device list attached to driver_data.ipmi_devices */
> @@ -54,6 +56,7 @@ struct ipmi_driver_data {
>          * to this selected global IPMI system interface.
>          */
>         struct acpi_ipmi_device *selected_smi;
> +       struct completion smi_selection_done;
>  };
>
>  struct acpi_ipmi_msg {
> @@ -463,8 +466,10 @@ static void ipmi_register_bmc(int iface, struct device *dev)
>                 if (temp->handle == handle)
>                         goto err_lock;
>         }
> -       if (!driver_data.selected_smi)
> +       if (!driver_data.selected_smi) {
>                 driver_data.selected_smi = ipmi_device;
> +               complete(&driver_data.smi_selection_done);
> +       }
>         list_add_tail(&ipmi_device->head, &driver_data.ipmi_devices);
>         mutex_unlock(&driver_data.ipmi_lock);
>
> @@ -578,6 +583,20 @@ acpi_ipmi_space_handler(u32 function, acpi_physical_address address,
>         return status;
>  }
>
> +int acpi_wait_for_acpi_ipmi(void)
> +{
> +       long ret;
> +
> +       ret = wait_for_completion_interruptible_timeout(&driver_data.smi_selection_done,
> +                                                       ACPI_IPMI_SMI_SELECTION_TIMEOUT);
> +
> +       if (ret <= 0)
> +               return -ETIMEDOUT;
> +
> +       return 0;
> +}
> +EXPORT_SYMBOL_GPL(acpi_wait_for_acpi_ipmi);
> +
>  static int __init acpi_ipmi_init(void)
>  {
>         int result;
> @@ -586,6 +605,8 @@ static int __init acpi_ipmi_init(void)
>         if (acpi_disabled)
>                 return 0;
>
> +       init_completion(&driver_data.smi_selection_done);
> +
>         status = acpi_install_address_space_handler(ACPI_ROOT_OBJECT,
>                                                     ACPI_ADR_SPACE_IPMI,
>                                                     &acpi_ipmi_space_handler,
> diff --git a/include/acpi/acpi_bus.h b/include/acpi/acpi_bus.h
> index 1216d72c650f..afa6e4d4bf46 100644
> --- a/include/acpi/acpi_bus.h
> +++ b/include/acpi/acpi_bus.h
> @@ -821,11 +821,16 @@ static inline void acpi_put_acpi_dev(struct acpi_device *adev)
>  {
>         acpi_dev_put(adev);
>  }
> +
> +int acpi_wait_for_acpi_ipmi(void);
> +
>  #else  /* CONFIG_ACPI */
>
>  static inline int register_acpi_bus_type(void *bus) { return 0; }
>  static inline int unregister_acpi_bus_type(void *bus) { return 0; }
>
> +static inline int acpi_wait_for_acpi_ipmi(void) { return 0; }
> +
>  #endif                         /* CONFIG_ACPI */
>
>  #endif /*__ACPI_BUS_H__*/
> --

On 1/9/24 11:50, Rafael J. Wysocki wrote:
> On Tue, Jan 9, 2024 at 5:12 AM Kai-Heng Feng
> <kai.heng.feng@canonical.com> wrote:
>>
>> On Dell servers, many APCI methods of acpi_power_meter module evaluate
>> variables inside IPMI region, so the region handler needs to be
>> installed. In addition to that, the handler needs to be fully
>> functional, and that depends on SMI being selected.
>>
>> So add a helper to let acpi_power_meter know when the handler is
>> installed and ready to be used.
>>
>> Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
> 
> Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> and I'm expecting this to be routed along with patch [2/2] that has
> been posted elsewhere.
> 

Note that linux-hwmon@ isn't copied on this patch and it therefore
does not show up in hwmon patchwork. To me that indicates that the
patch should _not_ be applied through hwmon. If hwmon is supposed
to pick it up, please copy the hwmon mailing list.

Guenter