Re: [PATCH v2 5/5] clk: implement sync_state support

[Ulf Hansson]

On Tue, Jun 16, 2026 at 11:09 PM Brian Masney <bmasney@xxxxxxxxxx> wrote:
>
> The existing support for disabling unused clks runs in the late initcall
> stage, and it has been known for a long time that this is broken since
> it runs too early in the boot up process. It doesn't work for kernel
> modules, and it also doesn't work if all of the consumers haven't fully
> probed yet. Folks have long recommended to boot certain platforms with
> clk_ignore_unused to work around issues with disabling unused clks.
>
> Let's go ahead and add a framework-level sync_state callback for the clk
> subsystem. If a driver doesn't have a sync_state callback configured,
> which is the 99+% use case today, then let's set it up to use the
> clk_sync_state() introduced in this commit so that no driver changes
> are needed.
>
> At the time of this writing, there are currently only 7 clk drivers that
> implement sync_state, and all are Qualcomm SoCs where they interact with
> the interconnect framework via icc_sync_state(). A shared helper has
> been created for this platform that calls clk_sync_state(). It is
> expected that any new clk drivers that want to implement their own
> sync_state will also need to call clk_sync_state() at the end of their
> custom sync_state callback.
>
> There will be several stages of disabling unused clks:
>
> - The first phase will be executed at late_initcall and it will only
>   disable unused clks that do not have a struct dev.
> - The sync_state callback will be invoked for each clk driver once all
>   consumers have probed.
>
> This is based on previous attempts by Saravana Kannan and Abel Vesa
> that are linked below.
>
> This change was tested on a Thinkpad x13s laptop.
>
>     [    0.308051] clk: Disabling unused clocks not associated with a device
>     [    6.541069] qcom_aoss_qmp c300000.power-management: clk: Disabling unused clocks
>     [    6.843310] qcom-qmp-pcie-phy 1c24000.phy: clk: Disabling unused clocks
>     [    7.604556] qcom-qmp-pcie-phy 1c14000.phy: clk: Disabling unused clocks
>     [    8.446161] qcom-qmp-usb-phy 88f1000.phy: clk: Disabling unused clocks
>     [    8.446293] qcom-qmp-usb-phy 88ef000.phy: clk: Disabling unused clocks
>     [    8.546067] qcom-qmp-combo-phy 88eb000.phy: clk: Disabling unused clocks
>     [    8.546203] qcom-qmp-combo-phy 8903000.phy: clk: Disabling unused clocks
>     [    8.546254] qcom-edp-phy aec5a00.phy: clk: Disabling unused clocks
>     [   15.436834] qcom-cpufreq-hw 18591000.cpufreq: clk: Disabling unused clocks
>     [   15.436953] clk-rpmh 18200000.rsc:clock-controller: clk: Disabling unused clocks
>     [   15.723348] qcom-qmp-pcie-phy 1c06000.phy: clk: Disabling unused clocks
>     [   21.063241] q6prm-lpass-clock 3000000.remoteproc:glink-edge:gpr:service@2:clock-controller: clk: Disabling unused clocks
>     [   21.081996] va_macro 3370000.codec: clk: Disabling unused clocks
>     [   21.092740] rx_macro 3200000.rxmacro: clk: Disabling unused clocks
>     [   21.118261] wsa_macro 3240000.codec: clk: Disabling unused clocks
>     [   21.128758] tx_macro 3220000.txmacro: clk: Disabling unused clocks
>
> Tested-by: Jens Glathe <jens.glathe@xxxxxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Brian Masney <bmasney@xxxxxxxxxx>
> Link: https://www.youtube.com/watch?v=tXYzM8yLIQA
> Link: https://lore.kernel.org/all/20210407034456.516204-1-saravanak@xxxxxxxxxx/
> Link: https://lore.kernel.org/all/20221227204528.1899863-1-abel.vesa@xxxxxxxxxx/

For future revisions, please add ulfh@xxxxxxxxxx on to/cc.

> ---
>  drivers/clk/clk.c | 72 +++++++++++++++++++++++++++++++++++++++++++++----------
>  1 file changed, 59 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/clk/clk.c b/drivers/clk/clk.c
> index 9cb2b42d1be4..7a15cceec620 100644
> --- a/drivers/clk/clk.c
> +++ b/drivers/clk/clk.c

[...]

>  void clk_sync_state(struct device *dev)
>  {
> -       /* Will fill in */
> +       __clk_disable_unused(dev);
>  }
>  EXPORT_SYMBOL_GPL(clk_sync_state);
>
> @@ -4345,8 +4382,17 @@ __clk_register(struct device *dev, struct device_node *np, struct clk_hw *hw)
>         core->dev = dev;
>         clk_pm_runtime_init(core);
>         core->of_node = np;
> -       if (dev && dev->driver)
> +       if (dev && dev->driver) {
>                 core->owner = dev->driver->owner;
> +
> +               /*
> +                * If a clk provider sets their own sync_state, then it needs to
> +                * also call clk_sync_state(). dev_set_drv_sync_state() won't
> +                * overwrite the sync_state callback, and this call will fail
> +                * with -EBUSY.
> +                */
> +               dev_set_drv_sync_state(dev, clk_sync_state);

We have cases where a device node represents a provider for multiple
types of resources, like clocks, power-domains (genpds), resets, etc,
as in the qcom case, for example.

For power-domain provider drivers (genpd) we also try to assign the
->sync_state() callback, see of_genpd_add_provider_simple() and
of_genpd_add_provider_simple(). This means the above doesn't play well
with how genpd behaves, so we need to figure out a way to manage these
cases.

In this regard, we also have of_genpd_sync_state(), which allows a
genpd provider driver to explicitly call genpd's sync state function,
if/when needed.

Unfortunately I am not able to suggest a detailed solution for how to
move this forward at this point, as it requires some more thinking and
I am heading for some vacation very soon.

> +       }
>         core->hw = hw;
>         core->flags = init->flags;
>         core->num_parents = init->num_parents;
>
> --
> 2.54.0

Kind regards
Uffe