[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241008193418.34573-1-kuurtb@gmail.com>
Date: Tue, 8 Oct 2024 16:34:18 -0300
From: Kurt Borja <kuurtb@...il.com>
To: ilpo.jarvinen@...ux.intel.com
Cc: Dell.Client.Kernel@...l.com,
hdegoede@...hat.com,
kuurtb@...il.com,
linux-kernel@...r.kernel.org,
platform-driver-x86@...r.kernel.org
Subject: Re: [PATCH v2] alienware-wmi: Dell AWCC platform_profile support
Removed.
> > +static int platform_profile_get(struct platform_profile_handler *pprof,
> > + enum platform_profile_option *profile)
> > +{
> > + acpi_status status;
> > + u32 in_args = 0x0B;
>
> Use a named define instead of a magic number.
>
Done
> > + u32 out_data;
> > +
> > + status = alienware_wmax_command(&in_args, sizeof(in_args),
> > + WMAX_METHOD_THERMAL_INFORMATION, (u32 *) &out_data);
>
> Casting to the same type??
>
> Also, alienware_wmax_command() inputs int * which makes the cast look even
> more odd?!?
>
> I can see there are pre-existing bogus (u32 *) casts too which would be
> nice to clean up in another patch.
>
Yes I thought it was odd but added them just in case. I can submit a
patch cleaning those up after this one.
> > +
> > + if (ACPI_FAILURE(status) || out_data < 0)
>
> u32 cannot be < 0??
>
> Is this an indication of a problem in the error handling?
>
Yes it was, thank you for catching it.
> > + return -EOPNOTSUPP;
> > +
> > + switch (out_data) {
> > + case WMAX_THERMAL_LOW_POWER:
> > + *profile = PLATFORM_PROFILE_LOW_POWER;
> > + break;
> > + case WMAX_THERMAL_QUIET:
> > + *profile = PLATFORM_PROFILE_QUIET;
> > + break;
> > + case WMAX_THERMAL_BALANCED:
> > + *profile = PLATFORM_PROFILE_BALANCED;
> > + break;
> > + case WMAX_THERMAL_BALANCED_PERFORMANCE:
> > + *profile = PLATFORM_PROFILE_BALANCED_PERFORMANCE;
> > + break;
> > + case WMAX_THERMAL_PERFORMANCE:
> > + case WMAX_THERMAL_GMODE:
> > + *profile = PLATFORM_PROFILE_PERFORMANCE;
> > + break;
> > + default:
> > + return -ENODATA;
> > + }
> > +
> > + return 0;
> > +}
> > +
> > +#define SET_MASK(prof) ((prof << 8) | 1)
>
> Name these with two defines. One define for naming that magic 1 and
> another for the profile field, use GENMASK() + FIELD_PREP() for it.
>
> Also, there's no need for this to be macro so change it into a static
> function.
>
Done.
> > +static int platform_profile_set(struct platform_profile_handler *pprof,
> > + enum platform_profile_option profile)
> > +{
> > + acpi_status status;
> > + u32 in_args;
> > + u32 out_data;
> > +
> > + switch (profile) {
> > + case PLATFORM_PROFILE_LOW_POWER:
> > + in_args = SET_MASK(WMAX_THERMAL_LOW_POWER);
> > + break;
> > + case PLATFORM_PROFILE_QUIET:
> > + in_args = SET_MASK(WMAX_THERMAL_QUIET);
> > + break;
> > + case PLATFORM_PROFILE_BALANCED:
> > + in_args = SET_MASK(WMAX_THERMAL_BALANCED);
> > + break;
> > + case PLATFORM_PROFILE_BALANCED_PERFORMANCE:
> > + in_args = SET_MASK(WMAX_THERMAL_BALANCED_PERFORMANCE);
> > + break;
> > + case PLATFORM_PROFILE_PERFORMANCE:
> > + in_args = SET_MASK(WMAX_THERMAL_PERFORMANCE);
> > + break;
> > + default:
> > + return -EOPNOTSUPP;
> > + }
> > +
> > + status = alienware_wmax_command(&in_args, sizeof(in_args),
> > + WMAX_METHOD_THERMAL_CONTROL, (u32 *) &out_data);
>
> Cast to same type.
>
> > +
> > + if (ACPI_FAILURE(status) || out_data < 0)
>
> u32 cannot be < 0.
>
> > + return -EOPNOTSUPP;
> > +
> > + return 0;
> > +}
> > +
> > +static int gmode_platform_profile_set(struct platform_profile_handler *pprof,
> > + enum platform_profile_option profile)
> > +{
> > + acpi_status status;
> > + u32 in_args;
> > + u32 out_data;
> > +
> > + switch (profile) {
> > + case PLATFORM_PROFILE_LOW_POWER:
> > + in_args = SET_MASK(WMAX_THERMAL_LOW_POWER);
> > + break;
> > + case PLATFORM_PROFILE_QUIET:
> > + in_args = SET_MASK(WMAX_THERMAL_QUIET);
> > + break;
> > + case PLATFORM_PROFILE_BALANCED:
> > + in_args = SET_MASK(WMAX_THERMAL_BALANCED);
> > + break;
> > + case PLATFORM_PROFILE_BALANCED_PERFORMANCE:
> > + in_args = SET_MASK(WMAX_THERMAL_BALANCED_PERFORMANCE);
> > + break;
> > + case PLATFORM_PROFILE_PERFORMANCE:
> > + in_args = SET_MASK(WMAX_THERMAL_GMODE);
> > + break;
> > + default:
> > + return -EOPNOTSUPP;
> > + }
> > +
> > + status = alienware_wmax_command(&in_args, sizeof(in_args),
>
> > + WMAX_METHOD_THERMAL_CONTROL, (u32 *) &out_data);
> > + if (ACPI_FAILURE(status) || out_data < 0)
>
> The same two issues here with the types.
>
> > + return -EOPNOTSUPP;
> > +
> > + return 0;
> > +}
> > +
> > +static int create_platform_profile(void)
> > +{
> > + pp_handler.profile_get = platform_profile_get;
> > +
> > + if (quirks->gmode > 0)
> > + pp_handler.profile_set = gmode_platform_profile_set;
> > + else
> > + pp_handler.profile_set = platform_profile_set;
> > +
> > + set_bit(PLATFORM_PROFILE_LOW_POWER, pp_handler.choices);
> > + set_bit(PLATFORM_PROFILE_QUIET, pp_handler.choices);
> > + set_bit(PLATFORM_PROFILE_BALANCED, pp_handler.choices);
> > + set_bit(PLATFORM_PROFILE_BALANCED_PERFORMANCE, pp_handler.choices);
> > + set_bit(PLATFORM_PROFILE_PERFORMANCE, pp_handler.choices);
> > +
> > + return platform_profile_register(&pp_handler);
> > +}
> > +
> > static int __init alienware_wmi_init(void)
> > {
> > int ret;
> > @@ -807,6 +987,12 @@ static int __init alienware_wmi_init(void)
> > goto fail_prep_deepsleep;
> > }
> >
> > + if (quirks->thermal > 0) {
> > + ret = create_platform_profile();
> > + if (ret)
> > + goto fail_prep_thermal_profile;
> > + }
> > +
> > ret = alienware_zone_init(platform_device);
> > if (ret)
> > goto fail_prep_zones;
> > @@ -817,6 +1003,7 @@ static int __init alienware_wmi_init(void)
> > alienware_zone_exit(platform_device);
> > fail_prep_deepsleep:
> > fail_prep_amplifier:
> > +fail_prep_thermal_profile:
> > fail_prep_hdmi:
> > platform_device_del(platform_device);
> > fail_platform_device2:
> > @@ -836,6 +1023,7 @@ static void __exit alienware_wmi_exit(void)
> > remove_hdmi(platform_device);
> > platform_device_unregister(platform_device);
> > platform_driver_unregister(&platform_driver);
> > + platform_profile_remove();
>
> IIRC, I once checked that this will lead to a crash if it wasn't created
> first (which in your driver isn't always the case).
>
Fixed.
> --
> i.
Thank you for your feedback. I will submit v3 with the corrections soon.
Kurt
Powered by blists - more mailing lists