OMAP2+: PM: omap device: API's for handling mstandby mode

PATCH Superseded G, Manjunath Kondaiah 3 participants 9 messages Dec 2, 2010 – Jan 26, 2011

Subsystems (4)

Arm (32-bit) Arm sub-architectures OMAP2+ support OMAP hwmod support

lore

Share & tools

lore

G, Manjunath Kondaiah

author

2010-12-02 13:59 UTC #0 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Certain errata in OMAP2+ processors will require forcing master standby to "no standby" mode before completing on going operation. Without this, the results will be unpredictable.

Since current implementation of PM run time framework does not support changing sysconfig settings during middle of the on going operation, these API's will support the same. One API will force the device's sysconfig mstandby mode settings to "no standby" and other API will release "no standby" mode and sets it to "smart standby" or "no standby? depending on HWMOD_SWSUP_MSTANDBY value.

The hwmod API "omap_hwmod_set_master_standbymode" will use no_stdby_cnt(introduced in omap_hwmod structure) for controlling access to sysconfig register settings in case of overlapping request/release API's are called. It also disables interrupts during syconfig register access.

These API's should be used by device drivers only incase of erratum applicable to their modules if there is no other methods to resolve.

These API's are required for multiple DMA errata which require putting DMA controller in no mstandby mode before stopping dma.

The applicable errata:
1. Erratum ID: i557(Applicable for omap36xx all ES versions) The channel hangs when the Pause bit (DMA4_CDPi [7] ) is cleared through config port while in Standby.

2. Erratum ID: i541
sDMA FIFO draining does not finish. Applicable to all omap2+ except omap4.

3. Erratum ID:i88
The sDMA to be put in no mstandby mode before disabling the channel after completing the data transfer operation.
Applicable only for OMAP3430 ES1.0

Also fixes typo HWMOD_SWSUP_MSTDBY to HWMOD_SWSUP_MSTANDBY in omap_hwmod.h

Signed-off-by G, Manjunath Kondaiah <[email protected]>

Cc Kevin Hilman <[email protected]>

Cc Paul Walmsley <[email protected]>

Cc [email protected]

---
v3: Review comments incorporated for:
https://patchwork.kernel.org/patch/282212/

v4: added mutex changes
https://patchwork.kernel.org/patch/338611/

v5: typo fixes for errata and erratum
https://patchwork.kernel.org/patch/352481/

v6: fixed oh increment bug and also mutex(missing in v5)
https://patchwork.kernel.org/patch/372231/

v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.

This patch has dependency on the patch series:
http://www.spinics.net/lists/linux-omap/msg40188.html
(OMAP2+: hwmod core upgrades for 2.6.38: first set) Apply the above set of patches before using this patch.

 arch/arm/mach-omap2/omap_hwmod.c              |   70 +++++++++++++++++++++++++
 arch/arm/plat-omap/include/plat/omap_device.h |    2 +
 arch/arm/plat-omap/include/plat/omap_hwmod.h  |    3 +
 arch/arm/plat-omap/omap_device.c              |   60 +++++++++++++++++++++
 4 files changed, 135 insertions(+), 0 deletions(-)

arch/arm/mach-omap2/omap_hwmod.c +70 -0

index 8a9847e..287818b 100644
@@ -1474,6 +1474,76 @@ int omap_hwmod_set_slave_idlemode(struct omap_hwmod *oh, u8 idlemode)
 }
 
 /**
+ * omap_hwmod_set_master_standbymode - set the hwmod's OCP mstandby mode
+ * @oh: struct omap_hwmod *
+ * @midlemode: flag to set mstandby to either "no standby" or "smart standby"
+ *
+ * Sets the IP block's OCP mstandby mode in hardware, and updates our
+ * local copy.  Intended to be used by drivers that have some erratum
+ * that requires direct manipulation of the MIDLEMODE bits.  Returns
+ * -EINVAL if @oh is null, or passes along the return value from
+ * _set_master_standbymode().
+ *
+ * Any users of this function should be scrutinized carefully.
+ */
+int omap_hwmod_set_master_standbymode(struct omap_hwmod *oh, u8 idlemode)
+{
+	u32 v;
+	u8 sf;
+	int retval = -1;
+	unsigned int long flags;
+
+	if (!oh)
+		return -EINVAL;
+
+	if (!oh->class->sysc)
+		return -EINVAL;
+
+	spin_lock_irqsave(&oh->_lock, flags);
+
+	v = oh->_sysc_cache;
+	sf = oh->class->sysc->sysc_flags;
+
+	if (!(sf & SYSC_HAS_MIDLEMODE)) {
+		spin_unlock_irqrestore(&oh->_lock, flags);
+		return -EINVAL;
+	}
+
+	if (idlemode) {
+		if (oh->no_stdby_cnt) {
+			atomic_inc(oh->no_stdby_cnt);
+			goto in_nostandby_mode;
+		}
+		atomic_inc(oh->no_stdby_cnt);
+		idlemode = HWMOD_IDLEMODE_NO;
+	} else {
+		if (oh->no_stdby_cnt == 1) {
+			atomic_dec(oh->no_stdby_cnt);
+			idlemode = (oh->flags & HWMOD_SWSUP_MSTANDBY) ?
+				HWMOD_IDLEMODE_NO : HWMOD_IDLEMODE_SMART;
+		} else if (oh->no_stdby_cnt > 1) {
+			goto out_nostandby_mode;
+		} else {
+			goto in_nostandby_mode;
+		}
+	}
+	retval = _set_master_standbymode(oh, idlemode, &v);
+
+	if (!retval)
+		_write_sysconfig(v, oh);
+
+	return retval;
+
+out_nostandby_mode:
+	atomic_dec(oh->no_stdby_cnt);
+
+in_nostandby_mode:
+	spin_unlock_irqrestore(&oh->_lock, flags);
+
+	return retval;
+}
+
+/**
  * omap_hwmod_register - register a struct omap_hwmod
  * @oh: struct omap_hwmod *
  *

arch/arm/plat-omap/include/plat/omap_device.h +2 -0

index 28e2d1a..4fbf7c1 100644
@@ -116,6 +116,8 @@ int omap_device_enable_hwmods(struct omap_device *od);
 int omap_device_disable_clocks(struct omap_device *od);
 int omap_device_enable_clocks(struct omap_device *od);
 
+int omap_device_require_no_mstandby(struct platform_device *pdev);
+int omap_device_release_no_mstandby(struct platform_device *pdev);
 
 /*
  * Entries should be kept in latency order ascending

arch/arm/plat-omap/include/plat/omap_hwmod.h +3 -0

index 62bdb23..9d591fb 100644
@@ -508,6 +508,7 @@ struct omap_hwmod {
 	u8				masters_cnt;
 	u8				slaves_cnt;
 	u8				hwmods_cnt;
+	u8				no_stdby_cnt;
 	u8				_int_flags;
 	u8				_state;
 	u8				_postsetup_state;

@@ -537,6 +538,8 @@ int omap_hwmod_disable_clocks(struct omap_hwmod *oh);
 
 int omap_hwmod_set_slave_idlemode(struct omap_hwmod *oh, u8 idlemode);
 
+int omap_hwmod_set_master_standbymode(struct omap_hwmod *oh, u8 idlemode);
+
 int omap_hwmod_reset(struct omap_hwmod *oh);
 void omap_hwmod_ocp_barrier(struct omap_hwmod *oh);

arch/arm/plat-omap/omap_device.c +60 -0

66 lines not on this page · raw message

Signature · 1 line

1.7.1

G, Manjunath Kondaiah

author

2010-12-02 14:23 UTC #1 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Quoted 38 lines · G, Manjunath Kondaiah

Certain errata in OMAP2+ processors will require forcing master standby to "no standby" mode before completing on going operation. Without this, the results will be unpredictable.
Since current implementation of PM run time framework does not support changing sysconfig settings during middle of the on going operation, these API's will support the same. One API will force the device's sysconfig mstandby mode settings to "no standby" and other API will release "no standby" mode and sets it to "smart standby" or "no standby? depending on HWMOD_SWSUP_MSTANDBY value.
The hwmod API "omap_hwmod_set_master_standbymode" will use no_stdby_cnt(introduced in omap_hwmod structure) for controlling access to sysconfig register settings in case of overlapping request/release API's are called. It also disables interrupts during syconfig register access.
These API's should be used by device drivers only incase of erratum applicable to their modules if there is no other methods to resolve.
These API's are required for multiple DMA errata which require putting DMA controller in no mstandby mode before stopping dma.
The applicable errata:
1. Erratum ID: i557(Applicable for omap36xx all ES versions) The channel hangs when the Pause bit (DMA4_CDPi [7] ) is cleared through config port while in Standby.
2. Erratum ID: i541
sDMA FIFO draining does not finish. Applicable to all omap2+ except omap4.
3. Erratum ID:i88
The sDMA to be put in no mstandby mode before disabling the channel after completing the data transfer operation.
Applicable only for OMAP3430 ES1.0

Quoted 7 lines

Also fixes typo HWMOD_SWSUP_MSTDBY to HWMOD_SWSUP_MSTANDBY in omap_hwmod.h
Signed-off-by G, Manjunath Kondaiah <[email protected]>
Cc Kevin Hilman <[email protected]>
Cc Paul Walmsley <[email protected]>
Cc [email protected]

Pls ignore this patch. I have resent this patch as:
[PATCH v7 RESEND] OMAP2+: PM: omap device: API's for handling mstandby mode

Sorry for noise.

-Manjunath

[...]

Cousson, Benoit

2010-12-03 08:38 UTC #2 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

In reply to G, Manjunath Kondaiah

Quoted 38 lines · Manjunath Kondaiah

Certain errata in OMAP2+ processors will require forcing master standby to "no standby" mode before completing on going operation. Without this, the results will be unpredictable.
Since current implementation of PM run time framework does not support changing sysconfig settings during middle of the on going operation, these API's will support the same. One API will force the device's sysconfig mstandby mode settings to "no standby" and other API will release "no standby" mode and sets it to "smart standby" or "no standby? depending on HWMOD_SWSUP_MSTANDBY value.
The hwmod API "omap_hwmod_set_master_standbymode" will use no_stdby_cnt(introduced in omap_hwmod structure) for controlling access to sysconfig register settings in case of overlapping request/release API's are called. It also disables interrupts during syconfig register access.
These API's should be used by device drivers only incase of erratum applicable to their modules if there is no other methods to resolve.
These API's are required for multiple DMA errata which require putting DMA controller in no mstandby mode before stopping dma.
The applicable errata:
1. Erratum ID: i557(Applicable for omap36xx all ES versions) The channel hangs when the Pause bit (DMA4_CDPi [7] ) is cleared through config port while in Standby.
2. Erratum ID: i541
sDMA FIFO draining does not finish. Applicable to all omap2+ except omap4.
3. Erratum ID:i88
The sDMA to be put in no mstandby mode before disabling the channel after completing the data transfer operation.
Applicable only for OMAP3430 ES1.0

Quoted 7 lines

Also fixes typo HWMOD_SWSUP_MSTDBY to HWMOD_SWSUP_MSTANDBY in omap_hwmod.h
Signed-off-by G, Manjunath Kondaiah<[email protected]>
Cc Kevin Hilman<[email protected]>
Cc Paul Walmsley<[email protected]>
Cc linux-arm-kernel at lists.infradead.org

You have to CC lakml during send-email, but it should not be in the changelog.

On the other hand, it is a good practice to add all the authors of the file you change in CC.

Quoted 10 lines

---
v3: Review comments incorporated for:
https://patchwork.kernel.org/patch/282212/
v4: added mutex changes
https://patchwork.kernel.org/patch/338611/
v5: typo fixes for errata and erratum
https://patchwork.kernel.org/patch/352481/

Quoted 6 lines

v6: fixed oh increment bug and also mutex(missing in v5)
https://patchwork.kernel.org/patch/372231/
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.

I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
These hwmod APIs are rather low level and should never be used except for workaround. So I'd prefer keeping this API simple and not sexy at all in order to prevent people to abuse it.

Regards,
Benoit

G, Manjunath Kondaiah

author

2010-12-03 09:47 UTC #3 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Quoted 66 lines · Cousson, Benoit

Quoted 38 lines · Manjunath Kondaiah
Certain errata in OMAP2+ processors will require forcing master standby to "no standby" mode before completing on going operation. Without this, the results will be unpredictable.
Since current implementation of PM run time framework does not support changing sysconfig settings during middle of the on going operation, these API's will support the same. One API will force the device's sysconfig mstandby mode settings to "no standby" and other API will release "no standby" mode and sets it to "smart standby" or "no standby? depending on HWMOD_SWSUP_MSTANDBY value.
The hwmod API "omap_hwmod_set_master_standbymode" will use no_stdby_cnt(introduced in omap_hwmod structure) for controlling access to sysconfig register settings in case of overlapping request/release API's are called. It also disables interrupts during syconfig register access.
These API's should be used by device drivers only incase of erratum applicable to their modules if there is no other methods to resolve.
These API's are required for multiple DMA errata which require putting DMA controller in no mstandby mode before stopping dma.
The applicable errata:
1. Erratum ID: i557(Applicable for omap36xx all ES versions) The channel hangs when the Pause bit (DMA4_CDPi [7] ) is cleared through config port while in Standby.
2. Erratum ID: i541
sDMA FIFO draining does not finish. Applicable to all omap2+ except omap4.
3. Erratum ID:i88
The sDMA to be put in no mstandby mode before disabling the channel after completing the data transfer operation.
Applicable only for OMAP3430 ES1.0
Quoted 7 lines
Also fixes typo HWMOD_SWSUP_MSTDBY to HWMOD_SWSUP_MSTANDBY in omap_hwmod.h
Signed-off-by G, Manjunath Kondaiah<[email protected]>
Cc Kevin Hilman<[email protected]>
Cc Paul Walmsley<[email protected]>
Cc [email protected]
You have to CC lakml during send-email, but it should not be in the changelog.
On the other hand, it is a good practice to add all the authors of the file you change in CC.
Quoted 13 lines
---
v3: Review comments incorporated for:
https://patchwork.kernel.org/patch/282212/
v4: added mutex changes
https://patchwork.kernel.org/patch/338611/
v5: typo fixes for errata and erratum
https://patchwork.kernel.org/patch/352481/
v6: fixed oh increment bug and also mutex(missing in v5)
https://patchwork.kernel.org/patch/372231/

Quoted 6 lines

Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.

Are you referring to spin lock or usage count?

-Manjunath

Cousson, Benoit

2010-12-03 12:19 UTC #4 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Quoted 1 line · Manjunath Kondaiah

* Cousson, Benoit<[email protected]> [2010-12-03 09:38:35 +0100]:

[...]

Quoted 8 lines

Quoted 6 lines
Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
Are you referring to spin lock or usage count?

The spinlock is needed, I was referring to the usage count.

That being said, the API proposed by Paul (request/release ) sound like a get/put, so maybe he had that kind of usage in mind.

I'm still not convince it should be done at hwmod API level.

Paul,
Any thoughts on that?

Regards,
Benoit

G, Manjunath Kondaiah

author

2010-12-14 01:48 UTC #5 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Paul/Benoit,

Quoted 16 lines · Benoit

Quoted 1 line · Manjunath Kondaiah
* Cousson, Benoit<[email protected]> [2010-12-03 09:38:35 +0100]:
[...]
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
Are you referring to spin lock or usage count?
The spinlock is needed, I was referring to the usage count.

Quoted 8 lines

That being said, the API proposed by Paul (request/release ) sound like a get/put, so maybe he had that kind of usage in mind.
I'm still not convince it should be done at hwmod API level.
Paul,
Any thoughts on that?

How do we proceed further?

-Manjunath

G, Manjunath Kondaiah

author

2011-01-24 09:33 UTC #6 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Paul/Benoit,

Quoted 22 lines · Manjunath Kondaiah

Paul/Benoit,
Quoted 19 lines · Benoit
Quoted 1 line · Manjunath Kondaiah
* Cousson, Benoit<[email protected]> [2010-12-03 09:38:35 +0100]:
[...]
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
Are you referring to spin lock or usage count?
The spinlock is needed, I was referring to the usage count.
That being said, the API proposed by Paul (request/release ) sound like a get/put, so maybe he had that kind of usage in mind.

Quoted 7 lines

Quoted 5 lines
I'm still not convince it should be done at hwmod API level.
Paul,
Any thoughts on that?
How do we proceed further?

Gentle reminder!

Can we please align on this so that DMA sysconfig patches can be upstreamed?

Discussion on this topic can be accessed at:
https://patchwork.kernel.org/patch/372231/
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg39728.html

-Manjunath

G, Manjunath Kondaiah

author

2011-01-26 00:41 UTC #7 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Quoted 32 lines · Manjunath Kondaiah

Paul/Benoit,
Quoted 22 lines · Manjunath Kondaiah
Paul/Benoit,
Quoted 19 lines · Benoit
Quoted 1 line · Manjunath Kondaiah
* Cousson, Benoit<[email protected]> [2010-12-03 09:38:35 +0100]:
[...]
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
Are you referring to spin lock or usage count?
The spinlock is needed, I was referring to the usage count.
That being said, the API proposed by Paul (request/release ) sound like a get/put, so maybe he had that kind of usage in mind.
Quoted 7 lines
Quoted 5 lines
I'm still not convince it should be done at hwmod API level.
Paul,
Any thoughts on that?
How do we proceed further?

Quoted 8 lines

Gentle reminder!
Can we please align on this so that DMA sysconfig patches can be upstreamed?
Discussion on this topic can be accessed at:
https://patchwork.kernel.org/patch/372231/
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg39728.html

As there is no response from paul on this topic, I will go ahead with usage count logic proposed by Adrian Hunter <[email protected]> at:
https://patchwork.kernel.org/patch/366831/
Above logic has got:

Acked-by Santosh Shilimkar <[email protected]>

Acked-by Peter Ujfalusi <[email protected]>

-Manjunath

Paul Walmsley

maintainer

2011-01-26 19:05 UTC #8 lore raw

PATCH v7 1/1 OMAP2+: PM: omap device: API's for handling mstandby mode

Quoted 42 lines · Manjunath Kondaiah

Quoted 32 lines · Manjunath Kondaiah
Paul/Benoit,
Quoted 22 lines · Manjunath Kondaiah
Paul/Benoit,
Quoted 19 lines · Benoit
Quoted 1 line · Manjunath Kondaiah
* Cousson, Benoit<[email protected]> [2010-12-03 09:38:35 +0100]:
[...]
Quoted 8 lines
Quoted 6 lines
Quoted 3 lines
v7: replaced mutex lock with spin lock. Added use count for controlling access to sysconfig registers in case if overlapping request/release API's are used.
I'm not sure it should be done here. I'd rather keep that code in the DMA, since this is the only user of that feature.
Are you referring to spin lock or usage count?
The spinlock is needed, I was referring to the usage count.
That being said, the API proposed by Paul (request/release ) sound like a get/put, so maybe he had that kind of usage in mind.
Quoted 7 lines
Quoted 5 lines
I'm still not convince it should be done at hwmod API level.
Paul,
Any thoughts on that?
How do we proceed further?
Quoted 8 lines
Gentle reminder!
Can we please align on this so that DMA sysconfig patches can be upstreamed?
Discussion on this topic can be accessed at:
https://patchwork.kernel.org/patch/372231/
http://www.mail-archive.com/linux-omap at vger.kernel.org/msg39728.html

Quoted 6 lines

As there is no response from paul on this topic, I will go ahead with usage count logic proposed by Adrian Hunter <[email protected]> at:
https://patchwork.kernel.org/patch/366831/
Above logic has got:
Acked-by Santosh Shilimkar <[email protected]>
Acked-by Peter Ujfalusi <[email protected]>

Yes, that's probably a good idea for right now. There's a wider spectrum of opinion than I had thought on whether this type of thing belongs in the hwmod code. I'd like to have a greater degree of alignment on that before we merge something into the hwmod code.

So, unless there is some reason why it will cause problems not to have this in the hwmod code for right now, please just go ahead and deal with it in your DMA code, and then we will move it to the hwmod code later as needed.

- Paul