Skip to contentSkip to conversation

pwm: lpss: platform: Save/restore the ctrl register over a suspend/resume

PATCH Applied Hans de Goede 5 participants 9 messages Apr 26, 2018 – Jun 6, 2018

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/2 pwm: lpss: platform: Save/restore the ctrl register over a suspend/resume

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

---
Changes in v2:
-Add Cc: [email protected] to make sure this goes into stable
together with "ACPI / LPSS: Add missing prv_offset setting for byt/cht PWM devices" which depends on this

---
 drivers/pwm/pwm-lpss-platform.c |  5 +++++
 drivers/pwm/pwm-lpss.c          | 30 ++++++++++++++++++++++++++++++
 drivers/pwm/pwm-lpss.h          |  2 ++
 3 files changed, 37 insertions(+)
drivers/pwm/pwm-lpss-platform.c +5 -0
index 5d6ed1507d29..5561b9e190f8 100644
@@ -74,6 +74,10 @@ static int pwm_lpss_remove_platform(struct platform_device *pdev)
 	return pwm_lpss_remove(lpwm);
 }
 
+static SIMPLE_DEV_PM_OPS(pwm_lpss_platform_pm_ops,
+			 pwm_lpss_suspend,
+			 pwm_lpss_resume);
+
 static const struct acpi_device_id pwm_lpss_acpi_match[] = {
 	{ "80860F09", (unsigned long)&pwm_lpss_byt_info },
 	{ "80862288", (unsigned long)&pwm_lpss_bsw_info },
@@ -86,6 +90,7 @@ static struct platform_driver pwm_lpss_driver_platform = {
 	.driver = {
 		.name = "pwm-lpss",
 		.acpi_match_table = pwm_lpss_acpi_match,
+		.pm = &pwm_lpss_platform_pm_ops,
 	},
 	.probe = pwm_lpss_probe_platform,
 	.remove = pwm_lpss_remove_platform,
drivers/pwm/pwm-lpss.c +30 -0
index 8db0d40ccacd..4721a264bac2 100644
@@ -32,10 +32,13 @@
 /* Size of each PWM register space if multiple */
 #define PWM_SIZE			0x400
 
+#define MAX_PWMS			4
+
 struct pwm_lpss_chip {
 	struct pwm_chip chip;
 	void __iomem *regs;
 	const struct pwm_lpss_boardinfo *info;
+	u32 saved_ctrl[MAX_PWMS];
 };
 
 static inline struct pwm_lpss_chip *to_lpwm(struct pwm_chip *chip)
@@ -177,6 +180,9 @@ struct pwm_lpss_chip *pwm_lpss_probe(struct device *dev, struct resource *r,
 	unsigned long c;
 	int ret;
 
+	if (WARN_ON(info->npwm > MAX_PWMS))
+		return ERR_PTR(-ENODEV);
+
 	lpwm = devm_kzalloc(dev, sizeof(*lpwm), GFP_KERNEL);
 	if (!lpwm)
 		return ERR_PTR(-ENOMEM);
@@ -212,6 +218,30 @@ int pwm_lpss_remove(struct pwm_lpss_chip *lpwm)
 }
 EXPORT_SYMBOL_GPL(pwm_lpss_remove);
 
+int pwm_lpss_suspend(struct device *dev)
+{
+	struct pwm_lpss_chip *lpwm = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < lpwm->info->npwm; i++)
+		lpwm->saved_ctrl[i] = readl(lpwm->regs + i * PWM_SIZE + PWM);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pwm_lpss_suspend);
+
+int pwm_lpss_resume(struct device *dev)
+{
+	struct pwm_lpss_chip *lpwm = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < lpwm->info->npwm; i++)
+		writel(lpwm->saved_ctrl[i], lpwm->regs + i * PWM_SIZE + PWM);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pwm_lpss_resume);
+
 MODULE_DESCRIPTION("PWM driver for Intel LPSS");
 MODULE_AUTHOR("Mika Westerberg <[email protected]>");
 MODULE_LICENSE("GPL v2");
drivers/pwm/pwm-lpss.h +2 -0
index 98306bb02cfe..7a4238ad1fcb 100644
@@ -28,5 +28,7 @@ struct pwm_lpss_boardinfo {
 struct pwm_lpss_chip *pwm_lpss_probe(struct device *dev, struct resource *r,
 				     const struct pwm_lpss_boardinfo *info);
 int pwm_lpss_remove(struct pwm_lpss_chip *lpwm);
+int pwm_lpss_suspend(struct device *dev);
+int pwm_lpss_resume(struct device *dev);
 
 #endif	/* __PWM_LPSS_H */
Signature · 1 line
2.17.0
#1 lore raw

PATCH v2 2/2 ACPI / LPSS: Add missing prv_offset setting for byt/cht PWM devices

The LPSS PWM device on on Bay Trail and Cherry Trail devices has a set of private registers at offset 0x800, the current lpss_device_desc for them already sets the LPSS_SAVE_CTX flag to have these saved/restored over device-suspend, but the current lpss_device_desc was not setting the prv_offset field, leading to the regular device registers getting saved/restored instead.

This is causing the PWM controller to no longer work, resulting in a black screen, after a suspend/resume on systems where the firmware clears the APB clock and reset bits at offset 0x804.

This commit fixes this by properly setting prv_offset to 0x800 for the PWM devices.

Fixes e1c748179754 ("ACPI / LPSS: Add Intel BayTrail ACPI mode PWM")
Fixes 1bfbd8eb8a7f ("ACPI / LPSS: Add ACPI IDs for Intel Braswell")

---
Changes in v2:
-Add Fixes and Cc:stable tags

---
 drivers/acpi/acpi_lpss.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/acpi/acpi_lpss.c +2 -0
index 2bcffec8dbf0..c4ba9164e582 100644
@@ -229,11 +229,13 @@ static const struct lpss_device_desc lpt_sdio_dev_desc = {
 
 static const struct lpss_device_desc byt_pwm_dev_desc = {
 	.flags = LPSS_SAVE_CTX,
+	.prv_offset = 0x800,
 	.setup = byt_pwm_setup,
 };
 
 static const struct lpss_device_desc bsw_pwm_dev_desc = {
 	.flags = LPSS_SAVE_CTX | LPSS_NO_D3_DELAY,
+	.prv_offset = 0x800,
 	.setup = bsw_pwm_setup,
 };
 
Signature · 1 line
2.17.0
#2 lore raw
In reply to Hans de Goede
Quoted 20 lines · CEST Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

Quoted 8 lines

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

Andy, Thierry, any comments or concerns regarding this series?

Quoted 106 lines

---
Changes in v2:

(quoted diff) +0 -2
-Add Cc: [email protected] to make sure this goes into stable
 together with "ACPI / LPSS: Add missing prv_offset setting for byt/cht
 PWM devices" which depends on this
---
 drivers/pwm/pwm-lpss-platform.c |  5 +++++
 drivers/pwm/pwm-lpss.c          | 30 ++++++++++++++++++++++++++++++
 drivers/pwm/pwm-lpss.h          |  2 ++
 3 files changed, 37 insertions(+)
drivers/pwm/pwm-lpss-platform.c +5 -0
index 5d6ed1507d29..5561b9e190f8 100644
@@ -74,6 +74,10 @@ static int pwm_lpss_remove_platform(struct platform_device *pdev)
 	return pwm_lpss_remove(lpwm);
 }
 
+static SIMPLE_DEV_PM_OPS(pwm_lpss_platform_pm_ops,
+			 pwm_lpss_suspend,
+			 pwm_lpss_resume);
+
 static const struct acpi_device_id pwm_lpss_acpi_match[] = {
 	{ "80860F09", (unsigned long)&pwm_lpss_byt_info },
 	{ "80862288", (unsigned long)&pwm_lpss_bsw_info },
@@ -86,6 +90,7 @@ static struct platform_driver pwm_lpss_driver_platform = {
 	.driver = {
 		.name = "pwm-lpss",
 		.acpi_match_table = pwm_lpss_acpi_match,
+		.pm = &pwm_lpss_platform_pm_ops,
 	},
 	.probe = pwm_lpss_probe_platform,
 	.remove = pwm_lpss_remove_platform,
drivers/pwm/pwm-lpss.c +30 -0
index 8db0d40ccacd..4721a264bac2 100644
@@ -32,10 +32,13 @@
 /* Size of each PWM register space if multiple */
 #define PWM_SIZE			0x400
 
+#define MAX_PWMS			4
+
 struct pwm_lpss_chip {
 	struct pwm_chip chip;
 	void __iomem *regs;
 	const struct pwm_lpss_boardinfo *info;
+	u32 saved_ctrl[MAX_PWMS];
 };
 
 static inline struct pwm_lpss_chip *to_lpwm(struct pwm_chip *chip)
@@ -177,6 +180,9 @@ struct pwm_lpss_chip *pwm_lpss_probe(struct device *dev, struct resource *r,
 	unsigned long c;
 	int ret;
 
+	if (WARN_ON(info->npwm > MAX_PWMS))
+		return ERR_PTR(-ENODEV);
+
 	lpwm = devm_kzalloc(dev, sizeof(*lpwm), GFP_KERNEL);
 	if (!lpwm)
 		return ERR_PTR(-ENOMEM);
@@ -212,6 +218,30 @@ int pwm_lpss_remove(struct pwm_lpss_chip *lpwm)
 }
 EXPORT_SYMBOL_GPL(pwm_lpss_remove);
 
+int pwm_lpss_suspend(struct device *dev)
+{
+	struct pwm_lpss_chip *lpwm = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < lpwm->info->npwm; i++)
+		lpwm->saved_ctrl[i] = readl(lpwm->regs + i * PWM_SIZE + PWM);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pwm_lpss_suspend);
+
+int pwm_lpss_resume(struct device *dev)
+{
+	struct pwm_lpss_chip *lpwm = dev_get_drvdata(dev);
+	int i;
+
+	for (i = 0; i < lpwm->info->npwm; i++)
+		writel(lpwm->saved_ctrl[i], lpwm->regs + i * PWM_SIZE + PWM);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(pwm_lpss_resume);
+
 MODULE_DESCRIPTION("PWM driver for Intel LPSS");
 MODULE_AUTHOR("Mika Westerberg <[email protected]>");
 MODULE_LICENSE("GPL v2");
drivers/pwm/pwm-lpss.h +2 -0
index 98306bb02cfe..7a4238ad1fcb 100644
@@ -28,5 +28,7 @@ struct pwm_lpss_boardinfo {
 struct pwm_lpss_chip *pwm_lpss_probe(struct device *dev, struct resource *r,
 				     const struct pwm_lpss_boardinfo *info);
 int pwm_lpss_remove(struct pwm_lpss_chip *lpwm);
+int pwm_lpss_suspend(struct device *dev);
+int pwm_lpss_resume(struct device *dev);
 
 #endif	/* __PWM_LPSS_H */
reviewer Reviewed-by
#3 lore raw
Quoted 37 lines · Rafael J. Wysocki
Quoted 36 lines · CEST Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated
mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty- cycle
after this, we end up with a "PWM_SW_UPDATE was not cleared" error and
the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss- platform
code, which save/restore the ctrl register over a suspend/resume, fixing
this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable
bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been
covered up sofar by a bug in the acpi-lpss.c code which was save/restoring
the regular device registers instead of the lpss private registers due to
lpss_device_desc.prv_offset not being set. This is fixed by a later patch
in this series.

Quoted 4 lines
Quoted 2 lines

Andy, Thierry, any comments or concerns regarding this series?

Fine by me:

Signature · 2 lines
Andy Shevchenko <[email protected]>
Intel Finland Oy
#4 lore raw
In reply to Rafael J. Wysocki
Quoted 27 lines · Rafael J. Wysocki
Quoted 26 lines · CEST Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

Quoted 4 lines
Quoted 2 lines

Andy, Thierry, any comments or concerns regarding this series?

Hans said in the cover letter of the first version of this series that he preferred to merge both patches through the PWM tree because of the dependency. So I'm waiting for an Acked-by from you on the ACPI bits.

Thierry

#5 lore raw

Hi,

Quoted 28 lines · Thierry Reding
Quoted 27 lines · Rafael J. Wysocki
Quoted 26 lines · CEST Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Andy, Thierry, any comments or concerns regarding this series?

Hans said in the cover letter of the first version of this series that he preferred to merge both patches through the PWM tree because of the dependency. So I'm waiting for an Acked-by from you on the ACPI bits.

The other way around (both through ACPI) would work too, but yes it would be good to keep them together,

Regards,

Hans

#6 lore raw
In reply to Thierry Reding
Quoted 28 lines · Thierry Reding
Quoted 27 lines · Rafael J. Wysocki
Quoted 26 lines · CEST Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Andy, Thierry, any comments or concerns regarding this series?

Hans said in the cover letter of the first version of this series that he preferred to merge both patches through the PWM tree because of the dependency. So I'm waiting for an Acked-by from you on the ACPI bits.

OK, please feel free to add it to the patches, then.

Thanks,
Rafael

#7 lore raw

Hi Thierry,

Since Rafael has indicated that you can take both patches upstream with his ack, through the pwm tree:

Quoted 2 lines · Rafael J. Wysocki

On Mon, May 14, 2018 at 1:50 PM, Thierry Reding
<[email protected]> wrote:

<snip>

Quoted 7 lines
Quoted 3 lines

Hans said in the cover letter of the first version of this series that he preferred to merge both patches through the PWM tree because of the dependency. So I'm waiting for an Acked-by from you on the ACPI bits.

OK, please feel free to add it to the patches, then. > Thanks,
Rafael

So can we get these 2 bugfix patches merged now please ?

Regards,

Hans

#8 lore raw
In reply to Hans de Goede
Quoted 33 lines · Hans de Goede

On some devices the contents of the ctrl register get lost over a suspend/resume and the PWM comes back up disabled after the resume.

This is seen on some Bay Trail devices with the PWM in ACPI enumerated mode, so it shows up as a platform device instead of a PCI device.

If we still think it is enabled and then try to change the duty-cycle after this, we end up with a "PWM_SW_UPDATE was not cleared" error and the PWM is stuck in that state from then on.

This commit adds suspend and resume pm callbacks to the pwm-lpss-platform code, which save/restore the ctrl register over a suspend/resume, fixing this.

Note that:

1) There is no need to do this over a runtime suspend, since we only runtime suspend when disabled and then we properly set the enable bit and reprogram the timings when we re-enable the PWM.

2) This may be happening on more systems then we realize, but has been covered up sofar by a bug in the acpi-lpss.c code which was save/restoring the regular device registers instead of the lpss private registers due to lpss_device_desc.prv_offset not being set. This is fixed by a later patch in this series.

---
Changes in v2:
-Add Cc: [email protected] to make sure this goes into stable
together with "ACPI / LPSS: Add missing prv_offset setting for byt/cht PWM devices" which depends on this

Quoted 5 lines
---
 drivers/pwm/pwm-lpss-platform.c |  5 +++++
 drivers/pwm/pwm-lpss.c          | 30 ++++++++++++++++++++++++++++++
 drivers/pwm/pwm-lpss.h          |  2 ++
 3 files changed, 37 insertions(+)

Applied, thanks.

Thierry