From: Thomas Richard <thomas.richard@bootlin.com>
To: Daniel Lezcano <daniel.lezcano@linaro.org>, rafael@kernel.org
Cc: rui.zhang@intel.com, lukasz.luba@arm.com,
linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org,
gregory.clement@bootlin.com, thomas.petazzoni@bootlin.com,
theo.lebrun@bootlin.com, u-kumar1@ti.com,
Keerthy <j-keerthy@ti.com>
Subject: Re: [PATCH v3] thermal: k3_j72xx_bandgap: implement suspend/resume support
Date: Wed, 24 Apr 2024 11:25:09 +0200 [thread overview]
Message-ID: <22278de4-ade3-45fc-8cd5-dd35043ed6c8@bootlin.com> (raw)
In-Reply-To: <d8c1d7a4-60d0-4940-b604-ee62591fa3e8@linaro.org>
On 4/22/24 17:46, Daniel Lezcano wrote:
> On 17/04/2024 14:03, Thomas Richard wrote:
>> From: Théo Lebrun <theo.lebrun@bootlin.com>
>>
>> This add suspend-to-ram support.
>>
>> The derived_table is kept-as is, so the resume is only about
>> pm_runtime_* calls and restoring the same registers as the probe.
>>
>> Extract the hardware initialization procedure to a function called at
>> both probe-time & resume-time.
>>
>> The probe-time loop is split in two to ensure doing the hardware
>> initialization before registering thermal zones. That ensures our
>> callbacks cannot be called while in bad state.
>>
>> Signed-off-by: Théo Lebrun <theo.lebrun@bootlin.com>
>> Acked-by: Keerthy <j-keerthy@ti.com>
>> Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
>> ---
>>
>> v3:
>> - Remove __maybe_unused attributes and use the magic of PTR_IF()
>
> Where are these changes ?
- I removed __maybe_unused attribute for k3_j72xx_bandgap_suspend() and
k3_j72xx_bandgap_resume()
- The DEFINE_SIMPLE_DEV_PM_OPS macro is used to declare
k3_j72xx_bandgap_pm_ops.
- The pm_sleep_ptr macro is used.
>
>> v2:
>> - Fix warnings/errors reported by kernel test robot
>>
>> drivers/thermal/k3_j72xx_bandgap.c | 112 ++++++++++++++++++++---------
>> 1 file changed, 79 insertions(+), 33 deletions(-)
>>
>> diff --git a/drivers/thermal/k3_j72xx_bandgap.c
>> b/drivers/thermal/k3_j72xx_bandgap.c
>> index c74094a86982..a7244e05ed5c 100644
>> --- a/drivers/thermal/k3_j72xx_bandgap.c
>> +++ b/drivers/thermal/k3_j72xx_bandgap.c
>> @@ -178,6 +178,7 @@ struct k3_j72xx_bandgap {
>> void __iomem *base;
>> void __iomem *cfg2_base;
>> struct k3_thermal_data *ts_data[K3_VTM_MAX_NUM_TS];
>> + int cnt;
>> };
>> /* common data structures */
>> @@ -338,24 +339,53 @@ static void print_look_up_table(struct device
>> *dev, int *ref_table)
>> dev_dbg(dev, "%d %d %d\n", i, derived_table[i],
>> ref_table[i]);
>> }
>> +static void k3_j72xx_bandgap_init_hw(struct k3_j72xx_bandgap *bgp)
>> +{
>> + struct k3_thermal_data *data;
>> + int id, high_max, low_temp;
>> + u32 val;
>> +
>> + for (id = 0; id < bgp->cnt; id++) {
>> + data = bgp->ts_data[id];
>> + val = readl(bgp->cfg2_base + data->ctrl_offset);
>> + val |= (K3_VTM_TMPSENS_CTRL_MAXT_OUTRG_EN |
>> + K3_VTM_TMPSENS_CTRL_SOC |
>> + K3_VTM_TMPSENS_CTRL_CLRZ | BIT(4));
>> + writel(val, bgp->cfg2_base + data->ctrl_offset);
>> + }
>> +
>> + /*
>> + * Program TSHUT thresholds
>> + * Step 1: set the thresholds to ~123C and 105C WKUP_VTM_MISC_CTRL2
>> + * Step 2: WKUP_VTM_TMPSENS_CTRL_j set the MAXT_OUTRG_EN bit
>> + * This is already taken care as per of init
>> + * Step 3: WKUP_VTM_MISC_CTRL set the ANYMAXT_OUTRG_ALERT_EN bit
>> + */
>> + high_max = k3_j72xx_bandgap_temp_to_adc_code(MAX_TEMP);
>> + low_temp = k3_j72xx_bandgap_temp_to_adc_code(COOL_DOWN_TEMP);
>> +
>> + writel((low_temp << 16) | high_max, bgp->cfg2_base +
>> K3_VTM_MISC_CTRL2_OFFSET);
>> + mdelay(100);
>
> 100ms ?
>
> Ok, I know you did not introduce this delay. But as it is a suspend /
> resume, this delay will be called at each resume which is not acceptable
> given its duration.
>
> Please, investigate a way to get rid of this delay
I talked to Keerthy, this delay was initially added to be sure the
thresholds are programmed before enabling the interrupt.
But for him, this delay is useless.
I did the test, I removed the delay and I didn't see any thermal events.
So we can remove it.
I'll send a new version without the delay, and I'll mention it in the
commit message.
>
>
>> + writel(K3_VTM_ANYMAXT_OUTRG_ALERT_EN, bgp->cfg2_base +
>> K3_VTM_MISC_CTRL_OFFSET);
>> +}
>> +
>> struct k3_j72xx_bandgap_data {
>> const bool has_errata_i2128;
>> };
>> static int k3_j72xx_bandgap_probe(struct platform_device *pdev)
>> {
>> - int ret = 0, cnt, val, id;
>> - int high_max, low_temp;
>> - struct resource *res;
>> + const struct k3_j72xx_bandgap_data *driver_data;
>> + struct thermal_zone_device *ti_thermal;
>> struct device *dev = &pdev->dev;
>> + bool workaround_needed = false;
>> struct k3_j72xx_bandgap *bgp;
>> struct k3_thermal_data *data;
>> - bool workaround_needed = false;
>> - const struct k3_j72xx_bandgap_data *driver_data;
>> - struct thermal_zone_device *ti_thermal;
>> - int *ref_table;
>> struct err_values err_vals;
>> void __iomem *fuse_base;
>> + int ret = 0, val, id;
>> + struct resource *res;
>> + int *ref_table;
>
> No related to the changes but the patch is showing signs the driver
> needs a code cleanup AFAICT
>
>> const s64 golden_factors[] = {
>> -490019999999999936,
>> @@ -422,10 +452,10 @@ static int k3_j72xx_bandgap_probe(struct
>> platform_device *pdev)
>> /* Get the sensor count in the VTM */
>> val = readl(bgp->base + K3_VTM_DEVINFO_PWR0_OFFSET);
>> - cnt = val & K3_VTM_DEVINFO_PWR0_TEMPSENS_CT_MASK;
>> - cnt >>= __ffs(K3_VTM_DEVINFO_PWR0_TEMPSENS_CT_MASK);
>> + bgp->cnt = val & K3_VTM_DEVINFO_PWR0_TEMPSENS_CT_MASK;
>> + bgp->cnt >>= __ffs(K3_VTM_DEVINFO_PWR0_TEMPSENS_CT_MASK);
>> - data = devm_kcalloc(bgp->dev, cnt, sizeof(*data), GFP_KERNEL);
>> + data = devm_kcalloc(bgp->dev, bgp->cnt, sizeof(*data), GFP_KERNEL);
>> if (!data) {
>> ret = -ENOMEM;
>> goto err_alloc;
>> @@ -449,8 +479,8 @@ static int k3_j72xx_bandgap_probe(struct
>> platform_device *pdev)
>> else
>> init_table(3, ref_table, pvt_wa_factors);
>> - /* Register the thermal sensors */
>> - for (id = 0; id < cnt; id++) {
>> + /* Precompute the derived table & fill each thermal sensor struct */
>> + for (id = 0; id < bgp->cnt; id++) {
>> data[id].bgp = bgp;
>> data[id].ctrl_offset = K3_VTM_TMPSENS0_CTRL_OFFSET + id * 0x20;
>> data[id].stat_offset = data[id].ctrl_offset +
>> @@ -470,13 +500,13 @@ static int k3_j72xx_bandgap_probe(struct
>> platform_device *pdev)
>> else if (id == 0 && !workaround_needed)
>> memcpy(derived_table, ref_table, TABLE_SIZE * 4);
>> - val = readl(data[id].bgp->cfg2_base + data[id].ctrl_offset);
>> - val |= (K3_VTM_TMPSENS_CTRL_MAXT_OUTRG_EN |
>> - K3_VTM_TMPSENS_CTRL_SOC |
>> - K3_VTM_TMPSENS_CTRL_CLRZ | BIT(4));
>> - writel(val, data[id].bgp->cfg2_base + data[id].ctrl_offset);
>> -
>> bgp->ts_data[id] = &data[id];
>> + }
>> +
>> + k3_j72xx_bandgap_init_hw(bgp);
>> +
>> + /* Register the thermal sensors */
>> + for (id = 0; id < bgp->cnt; id++) {
>> ti_thermal = devm_thermal_of_zone_register(bgp->dev, id,
>> &data[id],
>> &k3_of_thermal_ops);
>> if (IS_ERR(ti_thermal)) {
>> @@ -486,21 +516,7 @@ static int k3_j72xx_bandgap_probe(struct
>> platform_device *pdev)
>> }
>> }
>> - /*
>> - * Program TSHUT thresholds
>> - * Step 1: set the thresholds to ~123C and 105C WKUP_VTM_MISC_CTRL2
>> - * Step 2: WKUP_VTM_TMPSENS_CTRL_j set the MAXT_OUTRG_EN bit
>> - * This is already taken care as per of init
>> - * Step 3: WKUP_VTM_MISC_CTRL set the ANYMAXT_OUTRG_ALERT_EN bit
>> - */
>> - high_max = k3_j72xx_bandgap_temp_to_adc_code(MAX_TEMP);
>> - low_temp = k3_j72xx_bandgap_temp_to_adc_code(COOL_DOWN_TEMP);
>> -
>> - writel((low_temp << 16) | high_max, data[0].bgp->cfg2_base +
>> - K3_VTM_MISC_CTRL2_OFFSET);
>> - mdelay(100);
>> - writel(K3_VTM_ANYMAXT_OUTRG_ALERT_EN, data[0].bgp->cfg2_base +
>> - K3_VTM_MISC_CTRL_OFFSET);
>> + platform_set_drvdata(pdev, bgp);
>> print_look_up_table(dev, ref_table);
>> /*
>> @@ -527,6 +543,35 @@ static void k3_j72xx_bandgap_remove(struct
>> platform_device *pdev)
>> pm_runtime_disable(&pdev->dev);
>> }
>> +static int k3_j72xx_bandgap_suspend(struct device *dev)
>> +{
>> + pm_runtime_put_sync(dev);
>> + pm_runtime_disable(dev);
>> + return 0;
>> +}
>> +
>> +static int k3_j72xx_bandgap_resume(struct device *dev)
>> +{
>> + struct k3_j72xx_bandgap *bgp = dev_get_drvdata(dev);
>> + int ret;
>> +
>> + pm_runtime_enable(dev);
>> + ret = pm_runtime_get_sync(dev);
>> + if (ret < 0) {
>> + pm_runtime_put_noidle(dev);
>> + pm_runtime_disable(dev);
>> + return ret;
>> + }
>> +
>> + k3_j72xx_bandgap_init_hw(bgp);
>> +
>> + return 0;
>> +}
>> +
>> +static DEFINE_SIMPLE_DEV_PM_OPS(k3_j72xx_bandgap_pm_ops,
>> + k3_j72xx_bandgap_suspend,
>> + k3_j72xx_bandgap_resume);
>> +
>> static const struct k3_j72xx_bandgap_data
>> k3_j72xx_bandgap_j721e_data = {
>> .has_errata_i2128 = true,
>> };
>> @@ -554,6 +599,7 @@ static struct platform_driver
>> k3_j72xx_bandgap_sensor_driver = {
>> .driver = {
>> .name = "k3-j72xx-soc-thermal",
>> .of_match_table = of_k3_j72xx_bandgap_match,
>> + .pm = pm_sleep_ptr(&k3_j72xx_bandgap_pm_ops),
>> },
>> };
>>
>
--
Thomas Richard, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
prev parent reply other threads:[~2024-04-24 9:25 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-17 12:03 [PATCH v3] thermal: k3_j72xx_bandgap: implement suspend/resume support Thomas Richard
2024-04-22 15:46 ` Daniel Lezcano
2024-04-24 9:25 ` Thomas Richard [this message]
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=22278de4-ade3-45fc-8cd5-dd35043ed6c8@bootlin.com \
--to=thomas.richard@bootlin.com \
--cc=daniel.lezcano@linaro.org \
--cc=gregory.clement@bootlin.com \
--cc=j-keerthy@ti.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=lukasz.luba@arm.com \
--cc=rafael@kernel.org \
--cc=rui.zhang@intel.com \
--cc=theo.lebrun@bootlin.com \
--cc=thomas.petazzoni@bootlin.com \
--cc=u-kumar1@ti.com \
/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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).