linux-kernel - Re: [PATCH] cpufreq/amd-pstate: Fix min_limit perf and freq updation for performance governor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <2489c84a-580b-4cc3-acf0-110b5d1309d2@amd.com>
Date: Mon, 7 Apr 2025 08:25:18 -0500
From: Mario Limonciello <mario.limonciello@....com>
To: Dhananjay Ugwekar <dhananjay.ugwekar@....com>, gautham.shenoy@....com
Cc: linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] cpufreq/amd-pstate: Fix min_limit perf and freq updation
 for performance governor

On 4/7/2025 3:19 AM, Dhananjay Ugwekar wrote:
> The min_limit perf and freq values can get disconnected with performance
> governor, as we only modify the perf value in the special case. Fix that
> by modifying the perf and freq values together
> 
> Fixes: 009d1c29a451 ("cpufreq/amd-pstate: Move perf values into a union")
> Signed-off-by: Dhananjay Ugwekar <dhananjay.ugwekar@....com>

Reviewed-by: Mario Limonciello <mario.limonciello@....com>

Thanks!  I'll get this added to bleeding-edge and move it to fixes in a 
few days.

> ---
>   drivers/cpufreq/amd-pstate.c | 11 +++++++----
>   1 file changed, 7 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> index 024d33d5e367..76aa4a3698c3 100644
> --- a/drivers/cpufreq/amd-pstate.c
> +++ b/drivers/cpufreq/amd-pstate.c
> @@ -607,13 +607,16 @@ static void amd_pstate_update_min_max_limit(struct cpufreq_policy *policy)
>   	union perf_cached perf = READ_ONCE(cpudata->perf);
>   
>   	perf.max_limit_perf = freq_to_perf(perf, cpudata->nominal_freq, policy->max);
> -	perf.min_limit_perf = freq_to_perf(perf, cpudata->nominal_freq, policy->min);
> +	WRITE_ONCE(cpudata->max_limit_freq, policy->max);
>   
> -	if (cpudata->policy == CPUFREQ_POLICY_PERFORMANCE)
> +	if (cpudata->policy == CPUFREQ_POLICY_PERFORMANCE) {
>   		perf.min_limit_perf = min(perf.nominal_perf, perf.max_limit_perf);
> +		WRITE_ONCE(cpudata->min_limit_freq, min(cpudata->nominal_freq, cpudata->max_limit_freq));
> +	} else {
> +		perf.min_limit_perf = freq_to_perf(perf, cpudata->nominal_freq, policy->min);
> +		WRITE_ONCE(cpudata->min_limit_freq, policy->min);
> +	}
>   
> -	WRITE_ONCE(cpudata->max_limit_freq, policy->max);
> -	WRITE_ONCE(cpudata->min_limit_freq, policy->min);
>   	WRITE_ONCE(cpudata->perf, perf);
>   }
>