linux-kernel - Re: [PATCH v2] pmdomain: ti_sci: Handle wakeup constraint if device has pinctrl wakeup state

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: <CAPDyKFrOjFeZDOs84egaCAmRwtnXcj5bZLniOb1gkDkDH0214g@mail.gmail.com>
Date: Tue, 21 Oct 2025 14:22:13 +0200
From: Ulf Hansson <ulf.hansson@...aro.org>
To: Kendall Willis <k-willis@...com>, khilman@...libre.com
Cc: Nishanth Menon <nm@...com>, Tero Kristo <kristo@...nel.org>, 
	Santosh Shilimkar <ssantosh@...nel.org>, linux-arm-kernel@...ts.infradead.org, 
	linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org, d-gole@...com, 
	vishalm@...com, sebin.francis@...com, msp@...libre.com, a-kaur@...com
Subject: Re: [PATCH v2] pmdomain: ti_sci: Handle wakeup constraint if device
 has pinctrl wakeup state

On Wed, 10 Sept 2025 at 23:16, Kendall Willis <k-willis@...com> wrote:
>
> For TI K3 SoCs with IO daisy-chaining using pinctrl wakeup state, avoid
> sending wakeup constraints to the PM co-processor. This allows the SoC to
> enter deeper low power states while still maintaining wakeup capability
> for peripherals using IO daisy-chain wakeup via pinctrl wakeup state,
> similar to the existing wake IRQ mechanism added in commit b06bc47279919
> ("pmdomain: ti_sci: handle wake IRQs for IO daisy chain wakeups").
>
> Detect the pinctrl wakeup state in the suspend path, and if it exists,
> skip sending the constraint.
>
> Signed-off-by: Kendall Willis <k-willis@...com>
> ---
> Implementation
> --------------
> This patch is intended to be implemented along with the following
> series. This patch has no dependencies on any of the other series:
>
> 1. "pmdomain: ti_sci: Handle wakeup constraint if device has pinctrl
>    wakeup state": (this patch) skips setting constraints for wakeup
>    sources that use pinctrl state 'wakeup'.
>
> 2. "serial: 8250: omap: Add wakeup support": Implements wakeup from
>    the UARTs for TI K3 SoCs
>    https://github.com/kwillis01/linux/tree/b4/uart-daisy-chain-8250-omap
>
> 3. "arm64: dts: ti: k3-am62: Support Main UART wakeup": Implements the
>    functionality to wakeup the system from the Main UART
>    https://github.com/kwillis01/linux/tree/b4/uart-daisy-chain-dts
>
> Testing
> -------
> Tested on a AM62P SK EVM board with all series and dependencies
> implemented. Suspend/resume verified with the Main UART wakeup source
> by entering a keypress on the console.
>
> This github branch has all the necessary patches to test the series
> using linux-next:
> https://github.com/kwillis01/linux/tree/uart-daisy-chain
>
> Version History
> ---------------
> Changes from v1 to v2:
>  - Reworded commit message to be concise and to reference commit
>    b06bc47279919
>
> v1: https://lore.kernel.org/all/20250904211607.3725897-1-k-willis@ti.com/
> ---
>  drivers/pmdomain/ti/ti_sci_pm_domains.c | 16 ++++++++++++++++
>  1 file changed, 16 insertions(+)
>
> diff --git a/drivers/pmdomain/ti/ti_sci_pm_domains.c b/drivers/pmdomain/ti/ti_sci_pm_domains.c
> index 82df7e44250bb64f9c4a2108b5e97bd782a5976d..884905fd0686c1b94aba68c03da038e577428adf 100644
> --- a/drivers/pmdomain/ti/ti_sci_pm_domains.c
> +++ b/drivers/pmdomain/ti/ti_sci_pm_domains.c
> @@ -10,6 +10,7 @@
>  #include <linux/err.h>
>  #include <linux/module.h>
>  #include <linux/of.h>
> +#include <linux/pinctrl/consumer.h>
>  #include <linux/platform_device.h>
>  #include <linux/pm_domain.h>
>  #include <linux/pm_qos.h>
> @@ -84,9 +85,24 @@ static inline void ti_sci_pd_set_wkup_constraint(struct device *dev)
>         struct generic_pm_domain *genpd = pd_to_genpd(dev->pm_domain);
>         struct ti_sci_pm_domain *pd = genpd_to_ti_sci_pd(genpd);
>         const struct ti_sci_handle *ti_sci = pd->parent->ti_sci;
> +       struct pinctrl *pinctrl = devm_pinctrl_get(dev);
> +       struct pinctrl_state *pinctrl_state_wakeup;
>         int ret;
>
>         if (device_may_wakeup(dev)) {
> +               /*
> +                * If device can wakeup using pinctrl wakeup state,
> +                * we do not want to set a constraint
> +                */
> +               if (!IS_ERR_OR_NULL(pinctrl)) {
> +                       pinctrl_state_wakeup = pinctrl_lookup_state(pinctrl, "wakeup");
> +                       if (!IS_ERR_OR_NULL(pinctrl_state_wakeup)) {
> +                               dev_dbg(dev, "%s: has wake pinctrl wakeup state, not setting " \
> +                                               "constraints\n", __func__);
> +                               return;
> +                       }
> +               }

Relying on the above condition and the wakeirq check that was added in
b06bc47279919, seems fragile and doesn't really seem like the best
approach to me.

I would rather think that we should rely on the driver for the
consumer device to successfully have completed its parts, by enabling
the out-band system-wakeup IRQ. In other words, make the consumer
driver to call device_set_out_band_wakeup() and then use
device_out_band_wakeup() instead of the pinctrl+wakeirq check above,
to understand if QoS constraints shall be sent to the FW or not.

> +
>                 /*
>                  * If device can wakeup using IO daisy chain wakeups,
>                  * we do not want to set a constraint.
>
> ---
> base-commit: 5f540c4aade9f1d58fb7b9490b4b7d5214ec9746
> change-id: 20250910-uart-daisy-chain-pmdomain-b83b2db2c3cc
>
> Best regards,
> --
> Kendall Willis <k-willis@...com>
>

Kind regards
Uffe