From: Tudor Ambarus <tudor.ambarus@linaro.org>
To: Alexey Klimov <alexey.klimov@linaro.org>,
Daniel Lezcano <daniel.lezcano@kernel.org>,
"Rafael J. Wysocki" <rafael@kernel.org>,
Zhang Rui <rui.zhang@intel.com>,
Lukasz Luba <lukasz.luba@arm.com>, Rob Herring <robh@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
Conor Dooley <conor+dt@kernel.org>,
Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>,
Krzysztof Kozlowski <krzk@kernel.org>
Cc: Mateusz Majewski <m.majewski2@samsung.com>,
Thomas Turner <tturner@lineageos.org>,
Peter Griffin <peter.griffin@linaro.org>,
linux-pm@vger.kernel.org, linux-samsung-soc@vger.kernel.org,
devicetree@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 2/3] thermal: samsung: acpm-tmu: refactor SoC-specific operations into driver data
Date: Thu, 19 Nov 2026 14:17:23 +0200 [thread overview]
Message-ID: <29209163-a01a-4862-9824-90616da73f49@linaro.org> (raw)
In-Reply-To: <20261118-acpm-tmu-e850-support-v2-2-19848c0d97e5@linaro.org>
On 11/18/26 3:51 PM, Alexey Klimov wrote:
> The current ACPM TMU driver assumes that all TMU operations (init,
> threshold updates, and interrupt clearing) are handled exclusively
> via ACPM IPC commands, which is true for the GS101.
>
> However, other Exynos platforms (like Exynos850) lack complete firmware
> support for these operations and require direct MMIO register access
> instead for missing ACPM TMU IPC calls.
>
> Prepare the driver for multi-SoC support by moving the hardcoded GS101
> IPC calls into SoC-specific callbacks: "tz_control", "tmu_init",
> "tmu_update_thresholds" and "tmu_check_and_clear_irqs" within the
> acpm_tmu_driver_data struct. Additionally, save the regmap pointer in
> the private structure so that these callbacks can utilize it for other
> SoCs.
>
> Signed-off-by: Alexey Klimov <alexey.klimov@linaro.org>
> ---
> drivers/thermal/samsung/acpm-tmu.c | 113 +++++++++++++++++++++++++++----------
> 1 file changed, 82 insertions(+), 31 deletions(-)
>
> diff --git a/drivers/thermal/samsung/acpm-tmu.c b/drivers/thermal/samsung/acpm-tmu.c
> index 4799ef518a7a..0def0ec09084 100644
> --- a/drivers/thermal/samsung/acpm-tmu.c
> +++ b/drivers/thermal/samsung/acpm-tmu.c
> @@ -74,6 +74,8 @@ struct acpm_tmu_sensor {
> };
>
> struct acpm_tmu_priv {
> + const struct acpm_tmu_driver_data *data;
> + struct regmap *regmap;
> struct regmap_field *regmap_fields[REG_INTPEND_COUNT];
> struct acpm_handle *handle;
> struct device *dev;
> @@ -86,9 +88,18 @@ struct acpm_tmu_priv {
>
> struct acpm_tmu_driver_data {
> const struct reg_field *reg_fields;
> + const struct regmap_config *regmap_config;
> const struct acpm_tmu_sensor_group *sensor_groups;
> unsigned int num_sensor_groups;
> unsigned int mbox_chan_id;
> +
> + /* SoC-specific TMU routines and values */
> + int (*tz_control)(struct acpm_tmu_sensor *sensor, bool on);
on e850 this sets CORE_EN, what this callback do is to enable or
disable a sensor group
what do you think about:
int (*enable_sensor)(struct acpm_tmu_sensor *sensor, bool on);
> + int (*tmu_check_and_clear_irqs)(struct acpm_tmu_sensor *sensor,
> + bool *pending_irq);
drop tmu prefixes, the struct is already acpm_tmu
> + int (*tmu_init)(struct acpm_tmu_priv *priv);
> + int (*tmu_update_thresholds)(struct acpm_tmu_sensor *sensor,
> + u8 thresholds[2], u8 inten);
> };
How about doing the following:
struct acpm_tmu_soc_ops {
int (*init)(...);
int (*enable_sensor)(...);
int (*set_thresholds)(...);
bool (*clear_pending_irq)(...);
};
struct acpm_tmu_driver_data {
...
const struct acpm_tmu_soc_ops *ops;
};
static const struct acpm_tmu_soc_ops gs101_tmu_ops = {
.init = gs101_init,
.enable_sensor = gs101_enable_sensor,
.set_thresholds = gs101_set_thresholds,
.clear_pending_irq = gs101_clear_pending_irq,
};
>
> #define ACPM_TMU_SENSOR_GROUP(_mask, _id) \
> @@ -131,24 +142,32 @@ static const struct regmap_config gs101_regmap_config = {
> .max_register = GS101_REG_INTPEND(15),
> };
>
> -static const struct acpm_tmu_driver_data acpm_tmu_gs101 = {
> - .reg_fields = gs101_reg_fields,
> - .sensor_groups = gs101_sensor_groups,
> - .num_sensor_groups = ARRAY_SIZE(gs101_sensor_groups),
> - .mbox_chan_id = 9,
> -};
> +static int gs101_tz_control(struct acpm_tmu_sensor *sensor, bool on)
> +{
> + struct acpm_tmu_priv *priv = sensor->priv;
> + struct acpm_handle *handle = priv->handle;
> +
> + return handle->ops->tmu.tz_control(handle, priv->mbox_chan_id,
> + sensor->group->id, on);
> +}
> +
> +static int gs101_tmu_init(struct acpm_tmu_priv *priv)
> +{
> + struct acpm_handle *handle = priv->handle;
> +
> + return handle->ops->tmu.init(handle, priv->mbox_chan_id);
> +}
>
> static int acpm_tmu_op_tz_control(struct acpm_tmu_sensor *sensor, bool on)
> {
> struct acpm_tmu_priv *priv = sensor->priv;
> - struct acpm_handle *handle = priv->handle;
> - const struct acpm_tmu_ops *ops = &handle->ops->tmu;
> int ret;
>
> - ret = ops->tz_control(handle, priv->mbox_chan_id, sensor->group->id,
> - on);
> - if (ret)
> - return ret;
> + if (priv->data->tz_control) {
> + ret = priv->data->tz_control(sensor, on);
> + if (ret)
> + return ret;
> + }
>
Both SoCs implement all the callbacks, and tmu_check_and_clear_irqs()
is already called unconditionally. Please make all of them mandatory
and drop the NULL checks. As it is, a missing tz_control() would still
flip sensor->enabled.
> WRITE_ONCE(sensor->enabled, on);
>
> @@ -227,8 +246,8 @@ static int acpm_tmu_get_temp(struct thermal_zone_device *tz, int *temp)
> return 0;
> }
>
> -static int acpm_tmu_update_thresholds(struct acpm_tmu_sensor *sensor,
> - u8 thresholds[2], u8 inten)
> +static int gs101_tmu_update_thresholds(struct acpm_tmu_sensor *sensor,
> + u8 thresholds[2], u8 inten)
This moves the SoC-independent part (sensor->lock, disabling the sensor
and restoring its state) into the GS101 callback, and patch 3 then
duplicates it for Exynos850. Please keep that part in the core and have
the SoC callback only program the thresholds.
> {
> struct acpm_tmu_priv *priv = sensor->priv;
> struct acpm_handle *handle = priv->handle;
> @@ -292,7 +311,8 @@ static int acpm_tmu_set_trips(struct thermal_zone_device *tz, int low, int high)
> if (ret)
> return ret;
>
> - ret = acpm_tmu_update_thresholds(sensor, thresholds, inten);
> + if (priv->data->tmu_update_thresholds)
> + ret = priv->data->tmu_update_thresholds(sensor, thresholds, inten);
>
> pm_runtime_put_autosuspend(dev);
>
> @@ -331,12 +351,12 @@ static const struct thermal_zone_device_ops acpm_tmu_sensor_ops = {
> .set_trips = acpm_tmu_set_trips,
> };
>
> -static bool acpm_tmu_clear_pending_irq(struct acpm_tmu_sensor *sensor)
> +static int gs101_handle_irqs(struct acpm_tmu_sensor *sensor, bool *pending_irq)
why not keeping the bool, the ret code is not used anyway.
> {
> struct acpm_tmu_priv *priv = sensor->priv;
> const struct acpm_tmu_ops *ops = &priv->handle->ops->tmu;
> unsigned long mask = sensor->group->mask;
> - bool pending_irq = false;
> + bool pending = false;
> int i, ret;
> u32 val;
>
> @@ -345,16 +365,16 @@ static bool acpm_tmu_clear_pending_irq(struct acpm_tmu_sensor *sensor)
> for_each_set_bit(i, &mask, EXYNOS_TMU_SENSORS_MAX_COUNT) {
> ret = regmap_field_read(priv->regmap_fields[i], &val);
> if (ret)
> - return false;
> + return ret;
>
> if (val) {
> - pending_irq = true;
> + pending = true;
> break;
> }
> }
>
> - if (!pending_irq)
> - return false;
> + if (!pending)
> + return 0;
>
> ret = ops->clear_tz_irq(priv->handle, priv->mbox_chan_id,
> sensor->group->id);
> @@ -362,10 +382,12 @@ static bool acpm_tmu_clear_pending_irq(struct acpm_tmu_sensor *sensor)
> dev_err_ratelimited(priv->dev,
> "Sensor %u: failed to clear IRQ (%d)\n",
> sensor->group->id, ret);
> - return false;
> + return ret;
> }
>
> - return true;
> + *pending_irq = true;
> +
> + return ret;
> }
>
> static irqreturn_t acpm_tmu_thread_fn(int irq, void *id)
> @@ -383,8 +405,13 @@ static irqreturn_t acpm_tmu_thread_fn(int irq, void *id)
>
> for (i = 0; i < priv->num_sensors; i++) {
> struct acpm_tmu_sensor *sensor = &priv->sensors[i];
> + bool pending_irq = false;
>
> - if (!sensor->tzd || !acpm_tmu_clear_pending_irq(sensor))
> + if (!sensor->tzd)
> + continue;
> +
> + ret = priv->data->tmu_check_and_clear_irqs(sensor, &pending_irq);
> + if (ret || !pending_irq)
> continue;
>
> handled = true;
> @@ -398,15 +425,30 @@ static irqreturn_t acpm_tmu_thread_fn(int irq, void *id)
> return handled ? IRQ_HANDLED : IRQ_NONE;
> }
>
> +static const struct acpm_tmu_driver_data acpm_tmu_gs101 = {
> + .reg_fields = gs101_reg_fields,
> + .regmap_config = &gs101_regmap_config,
> + .sensor_groups = gs101_sensor_groups,
> + .num_sensor_groups = ARRAY_SIZE(gs101_sensor_groups),
> + .mbox_chan_id = 9,
> + .tz_control = gs101_tz_control,
> + .tmu_check_and_clear_irqs = gs101_handle_irqs,
> + .tmu_init = gs101_tmu_init,
> + .tmu_update_thresholds = gs101_tmu_update_thresholds,
> +};
> +
> static const struct of_device_id acpm_tmu_match[] = {
> - { .compatible = "google,gs101-tmu-top" },
> + {
> + .compatible = "google,gs101-tmu-top",
> + .data = &acpm_tmu_gs101
> + },
> { /* sentinel */ },
> };
> MODULE_DEVICE_TABLE(of, acpm_tmu_match);
>
> static int acpm_tmu_probe(struct platform_device *pdev)
> {
> - const struct acpm_tmu_driver_data *data = &acpm_tmu_gs101;
> + const struct acpm_tmu_driver_data *data;
> struct acpm_handle *acpm_handle;
> struct device *dev = &pdev->dev;
> struct acpm_tmu_priv *priv;
> @@ -414,6 +456,10 @@ static int acpm_tmu_probe(struct platform_device *pdev)
> void __iomem *base;
> int i, ret;
>
> + data = of_device_get_match_data(dev);
> + if (!data)
> + return dev_err_probe(dev, -ENODEV, "No matching driver data found\n");
plain return -ENODEV would do too, dev_err_probe brings nothing here.
Cheers,
ta
next prev parent reply other threads:[~2026-11-19 12:17 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-18 13:51 [PATCH v2 0/3] thermal: samsung: acpm-tmu: Add Exynos850 support Alexey Klimov
2026-11-18 13:51 ` [PATCH v2 1/3] dt-bindings: thermal: gs101-tmu: add Exynos850 compatible Alexey Klimov
2026-11-18 14:45 ` Tudor Ambarus
2026-11-18 13:51 ` [PATCH v2 2/3] thermal: samsung: acpm-tmu: refactor SoC-specific operations into driver data Alexey Klimov
2026-11-19 12:17 ` Tudor Ambarus [this message]
2026-11-18 13:51 ` [PATCH v2 3/3] thermal: samsung: acpm-tmu: add Exynos850 support Alexey Klimov
2026-11-19 12:21 ` Tudor Ambarus
2026-11-19 11:41 ` [PATCH v2 0/3] thermal: samsung: acpm-tmu: Add " Tudor Ambarus
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=29209163-a01a-4862-9824-90616da73f49@linaro.org \
--to=tudor.ambarus@linaro.org \
--cc=alexey.klimov@linaro.org \
--cc=bzolnier@gmail.com \
--cc=conor+dt@kernel.org \
--cc=daniel.lezcano@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=krzk+dt@kernel.org \
--cc=krzk@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=linux-samsung-soc@vger.kernel.org \
--cc=lukasz.luba@arm.com \
--cc=m.majewski2@samsung.com \
--cc=peter.griffin@linaro.org \
--cc=rafael@kernel.org \
--cc=robh@kernel.org \
--cc=rui.zhang@intel.com \
--cc=tturner@lineageos.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®