RE: [PATCH] cpufreq: amd-pstate: Don't create attributes when registration fails
From: Yuan, Perry
Date: Sun Jun 23 2024 - 23:32:39 EST
[AMD Official Use Only - AMD Internal Distribution Only]
Hi Mario,
> -----Original Message-----
> From: Limonciello, Mario <Mario.Limonciello@xxxxxxx>
> Sent: Monday, June 24, 2024 4:09 AM
> To: Yuan, Perry <Perry.Yuan@xxxxxxx>; Shenoy, Gautham Ranjal
> <gautham.shenoy@xxxxxxx>
> Cc: linux-pm@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx; Limonciello,
> Mario <Mario.Limonciello@xxxxxxx>; alex.s.cochran@xxxxxxxxx; Peter
> Jung <peterjung1337@xxxxxxxxx>
> Subject: [PATCH] cpufreq: amd-pstate: Don't create attributes when
> registration fails
>
> If driver registration fails then immediately return the failure instead of
> continuing to register attributes.
>
> This fixes issues of falling back from amd-pstate to other drivers when
> cpufreq init has failed for any reason.
>
> Reported-by: alex.s.cochran@xxxxxxxxx
> Signed-off-by: Mario Limonciello <mario.limonciello@xxxxxxx>
> ---
> Cc: Peter Jung <peterjung1337@xxxxxxxxx> drivers/cpufreq/amd-pstate.c |
> 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> index 05e80d2f8803..7195dfa98cc5 100644
> --- a/drivers/cpufreq/amd-pstate.c
> +++ b/drivers/cpufreq/amd-pstate.c
> @@ -2030,8 +2030,10 @@ static int __init amd_pstate_init(void)
> }
>
> ret = cpufreq_register_driver(current_pstate_driver);
> - if (ret)
> + if (ret) {
> pr_err("failed to register with return %d\n", ret);
> + goto disable_driver;
> + }
>
> dev_root = bus_get_dev_root(&cpu_subsys);
> if (dev_root) {
> @@ -2047,6 +2049,8 @@ static int __init amd_pstate_init(void)
>
> global_attr_free:
> cpufreq_unregister_driver(current_pstate_driver);
> +disable_driver:
> + amd_pstate_enable(false);
> return ret;
> }
> device_initcall(amd_pstate_init);
> --
> 2.43.0
In the previous failure behavior, I saw most of time, the driver will fallback to acpi_cpufreq driver when pstate driver failed to load.
I agree to disable it explicitly with this patch.
Reviewed-by: Perry Yuan <Perry.Yuan@xxxxxxx>