[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0f8ccc3d-46ca-4632-9ecf-5da596ad5a1a@arm.com>
Date: Fri, 3 May 2024 09:42:02 +0200
From: Pierre Gondois <pierre.gondois@....com>
To: liwei <liwei728@...wei.com>
Cc: linux-acpi@...r.kernel.org, rafael@...nel.org,
linux-kernel@...r.kernel.org, lenb@...nel.org, liwei391@...wei.com,
viresh.kumar@...aro.org
Subject: Re: [PATCH] cpufreq/cppc: fix perf_to_khz conversion exception
Hello Liwei,
Thanks for the fix. I think there might be a similar issue
in cppc_khz_to_perf(),
Regards,
Pierre
On 4/30/24 12:12, liwei wrote:
> When the nominal_freq recorded by the kernel is equal to lowest_freq and
> the frequency reduction operation is triggered externally, there is a
> conversion logic error in cppc_perf_to_khz(), causing the kernel to be
> unable to feedback the true frequency.
>
> Fix this by adding the branch processing logic when nominal_freq is equal
> to lowest_freq.
>
> Fixes: ec1c7ad47664 ("cpufreq: CPPC: Fix performance/frequency conversion")
> Signed-off-by: liwei <liwei728@...wei.com>
> ---
> drivers/acpi/cppc_acpi.c | 10 ++++++++--
> 1 file changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
> index a40b6f3946ef..92aac6974e0e 100644
> --- a/drivers/acpi/cppc_acpi.c
> +++ b/drivers/acpi/cppc_acpi.c
> @@ -1869,9 +1869,15 @@ unsigned int cppc_perf_to_khz(struct cppc_perf_caps *caps, unsigned int perf)
> u64 mul, div;
>
> if (caps->lowest_freq && caps->nominal_freq) {
> - mul = caps->nominal_freq - caps->lowest_freq;
> + /* Avoid the special case when nominal_freq is equal to lowest_freq */
> + if (caps->lowest_freq == caps->nominal_freq) {
> + mul = caps->nominal_freq;
> + div = caps->nominal_perf;
> + } else {
> + mul = caps->nominal_freq - caps->lowest_freq;
> + div = caps->nominal_perf - caps->lowest_perf;
> + }
> mul *= KHZ_PER_MHZ;
> - div = caps->nominal_perf - caps->lowest_perf;
> offset = caps->nominal_freq * KHZ_PER_MHZ -
> div64_u64(caps->nominal_perf * mul, div);
> } else {
Powered by blists - more mailing lists