[PATCH v2 0/3] occ: Add support for P10

[PATCH v2 0/3] occ: Add support for P10

[Joel Stanley]

Hi Guenter, here's v2 of this series on behalf of Eddie. I made the
change to the compatible string that we spoke about in v2, and I'm happy
for these to go through the hwmon tree.

v1: https://lore.kernel.org/linux-hwmon/20200501150833.5251-1-eajames@linux.ibm.com/

The OCC in the P10 has a number of differences from the P9. Add some logic to
handle the differences in accessing the OCC from the service processor, and
support the new temperature sensor type.

Eddie James (3):
  dt-bindings: fsi: Add P10 OCC device documentation
  fsi: occ: Add support for P10
  hwmon: (occ) Add new temperature sensor type

 .../devicetree/bindings/fsi/ibm,p9-occ.txt    |  12 +-
 drivers/fsi/fsi-occ.c                         | 125 +++++++++++++-----
 drivers/hwmon/occ/common.c                    |  75 +++++++++++
 3 files changed, 172 insertions(+), 40 deletions(-)

-- 
2.29.2

[PATCH v2 1/3] dt-bindings: fsi: Add P10 OCC device documentation

[Joel Stanley]

From: Eddie James <eajames@linux.ibm.com>

Add the P10 compatible string.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Joel Stanley <joel@jms.id.au>
---
 Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt b/Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt
index 99ca9862a586..e73358075a90 100644
--- a/Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt
+++ b/Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt
@@ -1,13 +1,13 @@
-Device-tree bindings for FSI-attached POWER9 On-Chip Controller (OCC)
----------------------------------------------------------------------
+Device-tree bindings for FSI-attached POWER9/POWER10 On-Chip Controller (OCC)
+-----------------------------------------------------------------------------
 
-This is the binding for the P9 On-Chip Controller accessed over FSI from a
-service processor. See fsi.txt for details on bindings for FSI slave and CFAM
+This is the binding for the P9 or P10 On-Chip Controller accessed over FSI from
+a service processor. See fsi.txt for details on bindings for FSI slave and CFAM
 nodes. The OCC is not an FSI slave device itself, rather it is accessed
-through the SBE fifo.
+through the SBE FIFO.
 
 Required properties:
- - compatible = "ibm,p9-occ"
+ - compatible = "ibm,p9-occ" or "ibm,p10-occ"
 
 Examples:
 
-- 
2.29.2

Re: [PATCH v2 1/3] dt-bindings: fsi: Add P10 OCC device documentation

[Rob Herring]

On Fri, 20 Nov 2020 11:33:13 +1030, Joel Stanley wrote:
> From: Eddie James <eajames@linux.ibm.com>
> 
> Add the P10 compatible string.
> 
> Signed-off-by: Eddie James <eajames@linux.ibm.com>
> Reviewed-by: Joel Stanley <joel@jms.id.au>
> Signed-off-by: Joel Stanley <joel@jms.id.au>
> ---
>  Documentation/devicetree/bindings/fsi/ibm,p9-occ.txt | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
> 

Acked-by: Rob Herring <robh@kernel.org>

[PATCH v2 2/3] fsi: occ: Add support for P10

[Joel Stanley]

From: Eddie James <eajames@linux.ibm.com>

The P10 OCC has a different SRAM address for the command and response
buffers. In addition, the SBE commands to access the SRAM have changed
format. Add versioning to the driver to handle these differences.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Joel Stanley <joel@jms.id.au>
---
v2: Fix sparse warning with a cast as Guneter suggested
---
 drivers/fsi/fsi-occ.c | 125 ++++++++++++++++++++++++++++++------------
 1 file changed, 91 insertions(+), 34 deletions(-)

diff --git a/drivers/fsi/fsi-occ.c b/drivers/fsi/fsi-occ.c
index 9eeb856c8905..10ca2e290655 100644
--- a/drivers/fsi/fsi-occ.c
+++ b/drivers/fsi/fsi-occ.c
@@ -14,6 +14,7 @@
 #include <linux/mutex.h>
 #include <linux/fsi-occ.h>
 #include <linux/of.h>
+#include <linux/of_device.h>
 #include <linux/platform_device.h>
 #include <linux/sched.h>
 #include <linux/slab.h>
@@ -24,8 +25,13 @@
 #define OCC_CMD_DATA_BYTES	4090
 #define OCC_RESP_DATA_BYTES	4089
 
-#define OCC_SRAM_CMD_ADDR	0xFFFBE000
-#define OCC_SRAM_RSP_ADDR	0xFFFBF000
+#define OCC_P9_SRAM_CMD_ADDR	0xFFFBE000
+#define OCC_P9_SRAM_RSP_ADDR	0xFFFBF000
+
+#define OCC_P10_SRAM_CMD_ADDR	0xFFFFD000
+#define OCC_P10_SRAM_RSP_ADDR	0xFFFFE000
+
+#define OCC_P10_SRAM_MODE	0x58	/* Normal mode, OCB channel 2 */
 
 /*
  * Assume we don't have much FFDC, if we do we'll overflow and
@@ -37,11 +43,14 @@
 #define OCC_TIMEOUT_MS		1000
 #define OCC_CMD_IN_PRG_WAIT_MS	50
 
+enum versions { occ_p9, occ_p10 };
+
 struct occ {
 	struct device *dev;
 	struct device *sbefifo;
 	char name[32];
 	int idx;
+	enum versions version;
 	struct miscdevice mdev;
 	struct mutex occ_lock;
 };
@@ -235,29 +244,43 @@ static int occ_verify_checksum(struct occ_response *resp, u16 data_length)
 	return 0;
 }
 
-static int occ_getsram(struct occ *occ, u32 address, void *data, ssize_t len)
+static int occ_getsram(struct occ *occ, u32 offset, void *data, ssize_t len)
 {
 	u32 data_len = ((len + 7) / 8) * 8;	/* must be multiples of 8 B */
-	size_t resp_len, resp_data_len;
-	__be32 *resp, cmd[5];
-	int rc;
+	size_t cmd_len, resp_len, resp_data_len;
+	__be32 *resp, cmd[6];
+	int idx = 0, rc;
 
 	/*
 	 * Magic sequence to do SBE getsram command. SBE will fetch data from
 	 * specified SRAM address.
 	 */
-	cmd[0] = cpu_to_be32(0x5);
+	switch (occ->version) {
+	default:
+	case occ_p9:
+		cmd_len = 5;
+		cmd[2] = cpu_to_be32(1);	/* Normal mode */
+		cmd[3] = cpu_to_be32(OCC_P9_SRAM_RSP_ADDR + offset);
+		break;
+	case occ_p10:
+		idx = 1;
+		cmd_len = 6;
+		cmd[2] = cpu_to_be32(OCC_P10_SRAM_MODE);
+		cmd[3] = 0;
+		cmd[4] = cpu_to_be32(OCC_P10_SRAM_RSP_ADDR + offset);
+		break;
+	}
+
+	cmd[0] = cpu_to_be32(cmd_len);
 	cmd[1] = cpu_to_be32(SBEFIFO_CMD_GET_OCC_SRAM);
-	cmd[2] = cpu_to_be32(1);
-	cmd[3] = cpu_to_be32(address);
-	cmd[4] = cpu_to_be32(data_len);
+	cmd[4 + idx] = cpu_to_be32(data_len);
 
 	resp_len = (data_len >> 2) + OCC_SBE_STATUS_WORDS;
 	resp = kzalloc(resp_len << 2, GFP_KERNEL);
 	if (!resp)
 		return -ENOMEM;
 
-	rc = sbefifo_submit(occ->sbefifo, cmd, 5, resp, &resp_len);
+	rc = sbefifo_submit(occ->sbefifo, cmd, cmd_len, resp, &resp_len);
 	if (rc)
 		goto free;
 
@@ -287,20 +310,21 @@ static int occ_getsram(struct occ *occ, u32 address, void *data, ssize_t len)
 	return rc;
 }
 
-static int occ_putsram(struct occ *occ, u32 address, const void *data,
-		       ssize_t len)
+static int occ_putsram(struct occ *occ, const void *data, ssize_t len)
 {
 	size_t cmd_len, buf_len, resp_len, resp_data_len;
 	u32 data_len = ((len + 7) / 8) * 8;	/* must be multiples of 8 B */
 	__be32 *buf;
-	int rc;
+	int idx = 0, rc;
+
+	cmd_len = (occ->version == occ_p10) ? 6 : 5;
 
 	/*
 	 * We use the same buffer for command and response, make
 	 * sure it's big enough
 	 */
 	resp_len = OCC_SBE_STATUS_WORDS;
-	cmd_len = (data_len >> 2) + 5;
+	cmd_len += data_len >> 2;
 	buf_len = max(cmd_len, resp_len);
 	buf = kzalloc(buf_len << 2, GFP_KERNEL);
 	if (!buf)
@@ -312,11 +336,23 @@ static int occ_putsram(struct occ *occ, u32 address, const void *data,
 	 */
 	buf[0] = cpu_to_be32(cmd_len);
 	buf[1] = cpu_to_be32(SBEFIFO_CMD_PUT_OCC_SRAM);
-	buf[2] = cpu_to_be32(1);
-	buf[3] = cpu_to_be32(address);
-	buf[4] = cpu_to_be32(data_len);
 
-	memcpy(&buf[5], data, len);
+	switch (occ->version) {
+	default:
+	case occ_p9:
+		buf[2] = cpu_to_be32(1);	/* Normal mode */
+		buf[3] = cpu_to_be32(OCC_P9_SRAM_CMD_ADDR);
+		break;
+	case occ_p10:
+		idx = 1;
+		buf[2] = cpu_to_be32(OCC_P10_SRAM_MODE);
+		buf[3] = 0;
+		buf[4] = cpu_to_be32(OCC_P10_SRAM_CMD_ADDR);
+		break;
+	}
+
+	buf[4 + idx] = cpu_to_be32(data_len);
+	memcpy(&buf[5 + idx], data, len);
 
 	rc = sbefifo_submit(occ->sbefifo, buf, cmd_len, buf, &resp_len);
 	if (rc)
@@ -356,21 +392,35 @@ static int occ_putsram(struct occ *occ, u32 address, const void *data,
 static int occ_trigger_attn(struct occ *occ)
 {
 	__be32 buf[OCC_SBE_STATUS_WORDS];
-	size_t resp_len, resp_data_len;
-	int rc;
+	size_t cmd_len, resp_len, resp_data_len;
+	int idx = 0, rc;
 
-	BUILD_BUG_ON(OCC_SBE_STATUS_WORDS < 7);
+	BUILD_BUG_ON(OCC_SBE_STATUS_WORDS < 8);
 	resp_len = OCC_SBE_STATUS_WORDS;
 
-	buf[0] = cpu_to_be32(0x5 + 0x2);        /* Chip-op length in words */
+	switch (occ->version) {
+	default:
+	case occ_p9:
+		cmd_len = 7;
+		buf[2] = cpu_to_be32(3); /* Circular mode */
+		buf[3] = 0;
+		break;
+	case occ_p10:
+		idx = 1;
+		cmd_len = 8;
+		buf[2] = cpu_to_be32(0xd0); /* Circular mode, OCB Channel 1 */
+		buf[3] = 0;
+		buf[4] = 0;
+		break;
+	}
+
+	buf[0] = cpu_to_be32(cmd_len);		/* Chip-op length in words */
 	buf[1] = cpu_to_be32(SBEFIFO_CMD_PUT_OCC_SRAM);
-	buf[2] = cpu_to_be32(0x3);              /* Mode: Circular */
-	buf[3] = cpu_to_be32(0x0);              /* Address: ignore in mode 3 */
-	buf[4] = cpu_to_be32(0x8);              /* Data length in bytes */
-	buf[5] = cpu_to_be32(0x20010000);       /* Trigger OCC attention */
-	buf[6] = 0;
+	buf[4 + idx] = cpu_to_be32(8);		/* Data length in bytes */
+	buf[5 + idx] = cpu_to_be32(0x20010000);	/* Trigger OCC attention */
+	buf[6 + idx] = 0;
 
-	rc = sbefifo_submit(occ->sbefifo, buf, 7, buf, &resp_len);
+	rc = sbefifo_submit(occ->sbefifo, buf, cmd_len, buf, &resp_len);
 	if (rc)
 		goto error;
 
@@ -429,7 +479,7 @@ int fsi_occ_submit(struct device *dev, const void *request, size_t req_len,
 
 	/* Extract the seq_no from the command (first byte) */
 	seq_no = *(const u8 *)request;
-	rc = occ_putsram(occ, OCC_SRAM_CMD_ADDR, request, req_len);
+	rc = occ_putsram(occ, request, req_len);
 	if (rc)
 		goto done;
 
@@ -440,7 +490,7 @@ int fsi_occ_submit(struct device *dev, const void *request, size_t req_len,
 	/* Read occ response header */
 	start = jiffies;
 	do {
-		rc = occ_getsram(occ, OCC_SRAM_RSP_ADDR, resp, 8);
+		rc = occ_getsram(occ, 0, resp, 8);
 		if (rc)
 			goto done;
 
@@ -476,8 +526,7 @@ int fsi_occ_submit(struct device *dev, const void *request, size_t req_len,
 	/* Grab the rest */
 	if (resp_data_length > 1) {
 		/* already got 3 bytes resp, also need 2 bytes checksum */
-		rc = occ_getsram(occ, OCC_SRAM_RSP_ADDR + 8,
-				 &resp->data[3], resp_data_length - 1);
+		rc = occ_getsram(occ, 8, &resp->data[3], resp_data_length - 1);
 		if (rc)
 			goto done;
 	}
@@ -517,6 +566,7 @@ static int occ_probe(struct platform_device *pdev)
 	if (!occ)
 		return -ENOMEM;
 
+	occ->version = (uintptr_t)of_device_get_match_data(dev);
 	occ->dev = dev;
 	occ->sbefifo = dev->parent;
 	mutex_init(&occ->occ_lock);
@@ -575,7 +625,14 @@ static int occ_remove(struct platform_device *pdev)
 }
 
 static const struct of_device_id occ_match[] = {
-	{ .compatible = "ibm,p9-occ" },
+	{
+		.compatible = "ibm,p9-occ",
+		.data = (void *)occ_p9
+	},
+	{
+		.compatible = "ibm,p10-occ",
+		.data = (void *)occ_p10
+	},
 	{ },
 };
 
-- 
2.29.2

[PATCH v2 3/3] hwmon: (occ) Add new temperature sensor type

[Joel Stanley]

From: Eddie James <eajames@linux.ibm.com>

The latest version of the On-Chip Controller (OCC) has a different
format for the temperature sensor data. Add a new temperature sensor
version to handle this data.

Signed-off-by: Eddie James <eajames@linux.ibm.com>
Reviewed-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Joel Stanley <joel@jms.id.au>
---
 drivers/hwmon/occ/common.c | 75 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 75 insertions(+)

diff --git a/drivers/hwmon/occ/common.c b/drivers/hwmon/occ/common.c
index a71777990d49..7a5e539b567b 100644
--- a/drivers/hwmon/occ/common.c
+++ b/drivers/hwmon/occ/common.c
@@ -41,6 +41,14 @@ struct temp_sensor_2 {
 	u8 value;
 } __packed;
 
+struct temp_sensor_10 {
+	u32 sensor_id;
+	u8 fru_type;
+	u8 value;
+	u8 throttle;
+	u8 reserved;
+} __packed;
+
 struct freq_sensor_1 {
 	u16 sensor_id;
 	u16 value;
@@ -307,6 +315,60 @@ static ssize_t occ_show_temp_2(struct device *dev,
 	return snprintf(buf, PAGE_SIZE - 1, "%u\n", val);
 }
 
+static ssize_t occ_show_temp_10(struct device *dev,
+				struct device_attribute *attr, char *buf)
+{
+	int rc;
+	u32 val = 0;
+	struct temp_sensor_10 *temp;
+	struct occ *occ = dev_get_drvdata(dev);
+	struct occ_sensors *sensors = &occ->sensors;
+	struct sensor_device_attribute_2 *sattr = to_sensor_dev_attr_2(attr);
+
+	rc = occ_update_response(occ);
+	if (rc)
+		return rc;
+
+	temp = ((struct temp_sensor_10 *)sensors->temp.data) + sattr->index;
+
+	switch (sattr->nr) {
+	case 0:
+		val = get_unaligned_be32(&temp->sensor_id);
+		break;
+	case 1:
+		val = temp->value;
+		if (val == OCC_TEMP_SENSOR_FAULT)
+			return -EREMOTEIO;
+
+		/*
+		 * VRM doesn't return temperature, only alarm bit. This
+		 * attribute maps to tempX_alarm instead of tempX_input for
+		 * VRM
+		 */
+		if (temp->fru_type != OCC_FRU_TYPE_VRM) {
+			/* sensor not ready */
+			if (val == 0)
+				return -EAGAIN;
+
+			val *= 1000;
+		}
+		break;
+	case 2:
+		val = temp->fru_type;
+		break;
+	case 3:
+		val = temp->value == OCC_TEMP_SENSOR_FAULT;
+		break;
+	case 4:
+		val = temp->throttle * 1000;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return snprintf(buf, PAGE_SIZE - 1, "%u\n", val);
+}
+
 static ssize_t occ_show_freq_1(struct device *dev,
 			       struct device_attribute *attr, char *buf)
 {
@@ -745,6 +807,10 @@ static int occ_setup_sensor_attrs(struct occ *occ)
 		num_attrs += (sensors->temp.num_sensors * 4);
 		show_temp = occ_show_temp_2;
 		break;
+	case 0x10:
+		num_attrs += (sensors->temp.num_sensors * 5);
+		show_temp = occ_show_temp_10;
+		break;
 	default:
 		sensors->temp.num_sensors = 0;
 	}
@@ -844,6 +910,15 @@ static int occ_setup_sensor_attrs(struct occ *occ)
 			attr->sensor = OCC_INIT_ATTR(attr->name, 0444,
 						     show_temp, NULL, 3, i);
 			attr++;
+
+			if (sensors->temp.version == 0x10) {
+				snprintf(attr->name, sizeof(attr->name),
+					 "temp%d_max", s);
+				attr->sensor = OCC_INIT_ATTR(attr->name, 0444,
+							     show_temp, NULL,
+							     4, i);
+				attr++;
+			}
 		}
 	}
 
-- 
2.29.2

Re: [PATCH v2 0/3] occ: Add support for P10

[Guenter Roeck]

On Fri, Nov 20, 2020 at 11:33:12AM +1030, Joel Stanley wrote:
> Hi Guenter, here's v2 of this series on behalf of Eddie. I made the
> change to the compatible string that we spoke about in v2, and I'm happy
> for these to go through the hwmon tree.
> 

I'll be happy to do that, as soon as we get a Reviewed-by: tag for the DT
change.

Thanks,
Guenter

> v1: https://lore.kernel.org/linux-hwmon/20200501150833.5251-1-eajames@linux.ibm.com/
> 
> The OCC in the P10 has a number of differences from the P9. Add some logic to
> handle the differences in accessing the OCC from the service processor, and
> support the new temperature sensor type.
> 
> Eddie James (3):
>   dt-bindings: fsi: Add P10 OCC device documentation
>   fsi: occ: Add support for P10
>   hwmon: (occ) Add new temperature sensor type
> 
>  .../devicetree/bindings/fsi/ibm,p9-occ.txt    |  12 +-
>  drivers/fsi/fsi-occ.c                         | 125 +++++++++++++-----
>  drivers/hwmon/occ/common.c                    |  75 +++++++++++
>  3 files changed, 172 insertions(+), 40 deletions(-)
> 
> -- 
> 2.29.2
> 

Re: [PATCH v2 0/3] occ: Add support for P10

[Joel Stanley]

On Sat, 21 Nov 2020 at 18:55, Guenter Roeck <linux@roeck-us.net> wrote:
>
> On Fri, Nov 20, 2020 at 11:33:12AM +1030, Joel Stanley wrote:
> > Hi Guenter, here's v2 of this series on behalf of Eddie. I made the
> > change to the compatible string that we spoke about in v2, and I'm happy
> > for these to go through the hwmon tree.
> >
>
> I'll be happy to do that, as soon as we get a Reviewed-by: tag for the DT
> change.

Rob has sent that through now, so this should be good to go in.

Thank you.

>
> Thanks,
> Guenter
>
> > v1: https://lore.kernel.org/linux-hwmon/20200501150833.5251-1-eajames@linux.ibm.com/
> >
> > The OCC in the P10 has a number of differences from the P9. Add some logic to
> > handle the differences in accessing the OCC from the service processor, and
> > support the new temperature sensor type.
> >
> > Eddie James (3):
> >   dt-bindings: fsi: Add P10 OCC device documentation
> >   fsi: occ: Add support for P10
> >   hwmon: (occ) Add new temperature sensor type
> >
> >  .../devicetree/bindings/fsi/ibm,p9-occ.txt    |  12 +-
> >  drivers/fsi/fsi-occ.c                         | 125 +++++++++++++-----
> >  drivers/hwmon/occ/common.c                    |  75 +++++++++++
> >  3 files changed, 172 insertions(+), 40 deletions(-)
> >
> > --
> > 2.29.2
> >

Re: [PATCH v2 0/3] occ: Add support for P10

[Guenter Roeck]

Hi Joel,

On 12/7/20 6:41 PM, Joel Stanley wrote:
> On Sat, 21 Nov 2020 at 18:55, Guenter Roeck <linux@roeck-us.net> wrote:
>>
>> On Fri, Nov 20, 2020 at 11:33:12AM +1030, Joel Stanley wrote:
>>> Hi Guenter, here's v2 of this series on behalf of Eddie. I made the
>>> change to the compatible string that we spoke about in v2, and I'm happy
>>> for these to go through the hwmon tree.
>>>
>>
>> I'll be happy to do that, as soon as we get a Reviewed-by: tag for the DT
>> change.
> 
> Rob has sent that through now, so this should be good to go in.
> 

Thanks for the reminder. Series applied to hwmon-next.

Thanks,
Genter

> Thank you.
> 
>>
>> Thanks,
>> Guenter
>>
>>> v1: https://lore.kernel.org/linux-hwmon/20200501150833.5251-1-eajames@linux.ibm.com/
>>>
>>> The OCC in the P10 has a number of differences from the P9. Add some logic to
>>> handle the differences in accessing the OCC from the service processor, and
>>> support the new temperature sensor type.
>>>
>>> Eddie James (3):
>>>   dt-bindings: fsi: Add P10 OCC device documentation
>>>   fsi: occ: Add support for P10
>>>   hwmon: (occ) Add new temperature sensor type
>>>
>>>  .../devicetree/bindings/fsi/ibm,p9-occ.txt    |  12 +-
>>>  drivers/fsi/fsi-occ.c                         | 125 +++++++++++++-----
>>>  drivers/hwmon/occ/common.c                    |  75 +++++++++++
>>>  3 files changed, 172 insertions(+), 40 deletions(-)
>>>
>>> --
>>> 2.29.2
>>>