2024-03-08 21:05:57

by Armin Wolf

[permalink] [raw]
Subject: [PATCH v4 2/2] platform/x86: wmi: Avoid returning AE_OK upon unknown error

If an error code other than EINVAL, ENODEV or ETIME is returned
by ec_read()/ec_write(), then AE_OK is wrongly returned.

Fix this by only returning AE_OK if the return code is 0, and
return AE_ERROR otherwise.

Tested on a Dell Inspiron 3505 and a Asus Prime B650-Plus.

Reviewed-by: Hans de Goede <[email protected]>
Reviewed-by: Ilpo Järvinen <[email protected]>
Signed-off-by: Armin Wolf <[email protected]>
---
drivers/platform/x86/wmi.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
index d9bf6d452b3a..84d1ccf6bc14 100644
--- a/drivers/platform/x86/wmi.c
+++ b/drivers/platform/x86/wmi.c
@@ -1218,8 +1218,10 @@ acpi_wmi_ec_space_handler(u32 function, acpi_physical_address address,
return AE_NOT_FOUND;
case -ETIME:
return AE_TIME;
- default:
+ case 0:
return AE_OK;
+ default:
+ return AE_ERROR;
}
}

--
2.39.2



Subject: Re: [PATCH v4 2/2] platform/x86: wmi: Avoid returning AE_OK upon unknown error


On 3/8/24 1:05 PM, Armin Wolf wrote:
> If an error code other than EINVAL, ENODEV or ETIME is returned
> by ec_read()/ec_write(), then AE_OK is wrongly returned.
>
> Fix this by only returning AE_OK if the return code is 0, and
> return AE_ERROR otherwise.
>
> Tested on a Dell Inspiron 3505 and a Asus Prime B650-Plus.
>
> Reviewed-by: Hans de Goede <[email protected]>
> Reviewed-by: Ilpo Järvinen <[email protected]>
> Signed-off-by: Armin Wolf <[email protected]>
> ---
> drivers/platform/x86/wmi.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
> index d9bf6d452b3a..84d1ccf6bc14 100644
> --- a/drivers/platform/x86/wmi.c
> +++ b/drivers/platform/x86/wmi.c
> @@ -1218,8 +1218,10 @@ acpi_wmi_ec_space_handler(u32 function, acpi_physical_address address,
> return AE_NOT_FOUND;
> case -ETIME:
> return AE_TIME;
> - default:
> + case 0:
> return AE_OK;
> + default:
> + return AE_ERROR;
> }

After checking the callers of acpi_wmi_ec_space_handler() it looks like there is no benefit in returning different ACPI status per error values. It is not being used. why no just return for result < 0 AE_ERROR and return for other cases?
> }
>
> --
> 2.39.2
>
>
--
Sathyanarayanan Kuppuswamy
Linux Kernel Developer


2024-03-09 19:10:58

by Armin Wolf

[permalink] [raw]
Subject: Re: [PATCH v4 2/2] platform/x86: wmi: Avoid returning AE_OK upon unknown error

Am 09.03.24 um 18:41 schrieb Kuppuswamy Sathyanarayanan:

> On 3/8/24 1:05 PM, Armin Wolf wrote:
>> If an error code other than EINVAL, ENODEV or ETIME is returned
>> by ec_read()/ec_write(), then AE_OK is wrongly returned.
>>
>> Fix this by only returning AE_OK if the return code is 0, and
>> return AE_ERROR otherwise.
>>
>> Tested on a Dell Inspiron 3505 and a Asus Prime B650-Plus.
>>
>> Reviewed-by: Hans de Goede <[email protected]>
>> Reviewed-by: Ilpo Järvinen <[email protected]>
>> Signed-off-by: Armin Wolf <[email protected]>
>> ---
>> drivers/platform/x86/wmi.c | 4 +++-
>> 1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
>> index d9bf6d452b3a..84d1ccf6bc14 100644
>> --- a/drivers/platform/x86/wmi.c
>> +++ b/drivers/platform/x86/wmi.c
>> @@ -1218,8 +1218,10 @@ acpi_wmi_ec_space_handler(u32 function, acpi_physical_address address,
>> return AE_NOT_FOUND;
>> case -ETIME:
>> return AE_TIME;
>> - default:
>> + case 0:
>> return AE_OK;
>> + default:
>> + return AE_ERROR;
>> }
> After checking the callers of acpi_wmi_ec_space_handler() it looks like there is no benefit in returning different ACPI status per error values. It is not being used. why no just return for result < 0 AE_ERROR and return for other cases?

Hi,

those handler functions are being called in acpi_ev_address_space_dispatch(), which uses the return value to print error messages.
So it makes sense to return different ACPI error values here.

Thanks,
Armin Wolf

>> }
>>
>> --
>> 2.39.2
>>
>>

Subject: Re: [PATCH v4 2/2] platform/x86: wmi: Avoid returning AE_OK upon unknown error

On Sat, Mar 9, 2024 at 11:10 AM Armin Wolf <[email protected]> wrote:
>
> Am 09.03.24 um 18:41 schrieb Kuppuswamy Sathyanarayanan:
>
> > On 3/8/24 1:05 PM, Armin Wolf wrote:
> >> If an error code other than EINVAL, ENODEV or ETIME is returned
> >> by ec_read()/ec_write(), then AE_OK is wrongly returned.
> >>
> >> Fix this by only returning AE_OK if the return code is 0, and
> >> return AE_ERROR otherwise.
> >>
> >> Tested on a Dell Inspiron 3505 and a Asus Prime B650-Plus.
> >>
> >> Reviewed-by: Hans de Goede <[email protected]>
> >> Reviewed-by: Ilpo Järvinen <[email protected]>
> >> Signed-off-by: Armin Wolf <[email protected]>
> >> ---

Got it.

Reviewed-by: Kuppuswamy Sathyanarayanan
<[email protected]>

> >> drivers/platform/x86/wmi.c | 4 +++-
> >> 1 file changed, 3 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
> >> index d9bf6d452b3a..84d1ccf6bc14 100644
> >> --- a/drivers/platform/x86/wmi.c
> >> +++ b/drivers/platform/x86/wmi.c
> >> @@ -1218,8 +1218,10 @@ acpi_wmi_ec_space_handler(u32 function, acpi_physical_address address,
> >> return AE_NOT_FOUND;
> >> case -ETIME:
> >> return AE_TIME;
> >> - default:
> >> + case 0:
> >> return AE_OK;
> >> + default:
> >> + return AE_ERROR;
> >> }
> > After checking the callers of acpi_wmi_ec_space_handler() it looks like there is no benefit in returning different ACPI status per error values. It is not being used. why no just return for result < 0 AE_ERROR and return for other cases?
>
> Hi,
>
> those handler functions are being called in acpi_ev_address_space_dispatch(), which uses the return value to print error messages.
> So it makes sense to return different ACPI error values here.
>
> Thanks,
> Armin Wolf
>
> >> }
> >>
> >> --
> >> 2.39.2
> >>
> >>

Subject: Re: [PATCH v4 2/2] platform/x86: wmi: Avoid returning AE_OK upon unknown error


On 3/8/24 1:05 PM, Armin Wolf wrote:
> If an error code other than EINVAL, ENODEV or ETIME is returned
> by ec_read()/ec_write(), then AE_OK is wrongly returned.
>
> Fix this by only returning AE_OK if the return code is 0, and
> return AE_ERROR otherwise.
>
> Tested on a Dell Inspiron 3505 and a Asus Prime B650-Plus.
>
> Reviewed-by: Hans de Goede <[email protected]>
> Reviewed-by: Ilpo Järvinen <[email protected]>
> Signed-off-by: Armin Wolf <[email protected]>
> ---

Looks good to me.

Reviewed-by: Kuppuswamy Sathyanarayanan <[email protected]>

> drivers/platform/x86/wmi.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/platform/x86/wmi.c b/drivers/platform/x86/wmi.c
> index d9bf6d452b3a..84d1ccf6bc14 100644
> --- a/drivers/platform/x86/wmi.c
> +++ b/drivers/platform/x86/wmi.c
> @@ -1218,8 +1218,10 @@ acpi_wmi_ec_space_handler(u32 function, acpi_physical_address address,
> return AE_NOT_FOUND;
> case -ETIME:
> return AE_TIME;
> - default:
> + case 0:
> return AE_OK;
> + default:
> + return AE_ERROR;
> }
> }
>
> --
> 2.39.2
>
>
--
Sathyanarayanan Kuppuswamy
Linux Kernel Developer