[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAJZ5v0gxNOWhC58PHeUhW_tgf6d1fGJVZ1x91zkDdht11yUv-A@mail.gmail.com>
Date: Thu, 29 Jun 2023 10:59:44 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Chungkai Yang <Chung-kai.Yang@...iatek.com>
Cc: rafael@...nel.org, len.brown@...el.com, pavel@....cz,
linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org,
ccj.yeh@...iatek.com
Subject: Re: [PATCH v2] PM: QoS: Restore support for default value on
frequency QoS
On Tue, Jun 27, 2023 at 9:18 AM Chungkai Yang
<Chung-kai.Yang@...iatek.com> wrote:
>
> PM_QOS_DEFAULT_VALUE case is not covered.
>
> Commit 8d36694245f2 ("PM: QoS: Add check to make sure CPU freq is
> non-negative") makes sure CPU freq is non-negative to avoid negative
> value converting to unsigned data type. However, when the value is
> PM_QOS_DEFAULT_VALUE, pm_qos_update_target specifically uses
> c->default_value which is set to FREQ_QOS_MIN/MAX_DEFAULT_VALUE when
> cpufreq_policy_alloc is executed, for this case handling.
>
> Adding check for PM_QOS_DEFAULT_VALUE to let default setting work will
> fix this problem.
>
> Signed-off-by: Chungkai Yang <Chung-kai.Yang@...iatek.com>
>
> ---
> V1 -> V2: Check both freq_qos_add/update_request.
>
> Link:
> https://lore.kernel.org/lkml/20230626035144.19717-1-Chung-kai.Yang@mediatek.com/
> ---
> kernel/power/qos.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/power/qos.c b/kernel/power/qos.c
> index af51ed6d45ef..260aca45c681 100644
> --- a/kernel/power/qos.c
> +++ b/kernel/power/qos.c
> @@ -531,7 +531,7 @@ int freq_qos_add_request(struct freq_constraints *qos,
> {
> int ret;
>
> - if (IS_ERR_OR_NULL(qos) || !req || value < 0)
> + if (IS_ERR_OR_NULL(qos) || !req || (value < 0 && value != PM_QOS_DEFAULT_VALUE))
> return -EINVAL;
>
> if (WARN(freq_qos_request_active(req),
> @@ -563,7 +563,7 @@ EXPORT_SYMBOL_GPL(freq_qos_add_request);
> */
> int freq_qos_update_request(struct freq_qos_request *req, s32 new_value)
> {
> - if (!req || new_value < 0)
> + if (!req || (new_value < 0 && new_value != PM_QOS_DEFAULT_VALUE))
This is functionally correct, but it's better to avoid duplicating the
value check, so please add a helper for it, for example:
static bool freq_qos_value_invalid(s32 value)
{
return value < 0 && value != PM_QOS_DEFAULT_VALUE;
}
and call it in both places.
> return -EINVAL;
>
> if (WARN(!freq_qos_request_active(req),
> --
Powered by blists - more mailing lists