Skip to contentSkip to conversation

PTP_CLK pin configuration for SJA1105 DSA driver

PATCH In review Vladimir Oltean 3 participants 7 messages Mar 23, 2020 – Mar 24, 2020

Subsystems (4)
lore
Share & tools
lore

Vladimir Oltean

author maintainer
#0 lore raw

cover PTP_CLK pin configuration for SJA1105 DSA driver

From: Vladimir Oltean <[email protected]>

This series adds support for the PTP_CLK pin on SJA1105 to be configured via the PTP subsystem, in the "periodic output" and "external timestamp input" modes.

Vladimir Oltean (4):
net: dsa: sja1105: unconditionally set DESTMETA and SRCMETA in AVB
table
net: dsa: sja1105: make future_base_time a common helper
net: dsa: sja1105: make the AVB table dynamically reconfigurable
net: dsa: sja1105: configure the PTP_CLK pin as EXT_TS or PER_OUT

 drivers/net/dsa/sja1105/sja1105.h             |   5 +
 .../net/dsa/sja1105/sja1105_dynamic_config.c  |  24 +-
 drivers/net/dsa/sja1105/sja1105_main.c        |  40 +++
 drivers/net/dsa/sja1105/sja1105_ptp.c         | 285 +++++++++++++++---
 drivers/net/dsa/sja1105/sja1105_ptp.h         |  31 ++
 drivers/net/dsa/sja1105/sja1105_spi.c         |   5 +
 .../net/dsa/sja1105/sja1105_static_config.c   |   5 +-
 .../net/dsa/sja1105/sja1105_static_config.h   |   1 +
 drivers/net/dsa/sja1105/sja1105_tas.c         |  27 --
 9 files changed, 355 insertions(+), 68 deletions(-)
Signature · 1 line
2.17.1
author maintainer
#1 lore raw

PATCH 1/4 net: dsa: sja1105: unconditionally set DESTMETA and SRCMETA in AVB table

From: Vladimir Oltean <[email protected]>

These fields configure the destination and source MAC address that the switch will put in the Ethernet frames sent towards the CPU port that contain RX timestamps for PTP.

These fields do not enable the feature itself, that is configured via SEND_META0 and SEND_META1 in the General Params table.

The implication of this patch is that the AVB Params table will always be present in the static config. Which doesn't really hurt.

This is needed because in a future patch, we will add another field from this table, CAS_MASTER, for configuring the PTP_CLK pin function. That can be configured irrespective of whether RX timestamping is enabled or not, so always having this table present is going to simplify things a bit.

---
 drivers/net/dsa/sja1105/sja1105_main.c | 32 ++++++++++++++++++++++
 drivers/net/dsa/sja1105/sja1105_ptp.c  | 38 --------------------------
 2 files changed, 32 insertions(+), 38 deletions(-)
drivers/net/dsa/sja1105/sja1105_main.c +32 -0
index afafe2ecf248..4d182293ca1f 100644
@@ -479,6 +479,35 @@ static int sja1105_init_general_params(struct sja1105_private *priv)
 	return 0;
 }
 
+static int sja1105_init_avb_params(struct sja1105_private *priv)
+{
+	struct sja1105_avb_params_entry *avb;
+	struct sja1105_table *table;
+
+	table = &priv->static_config.tables[BLK_IDX_AVB_PARAMS];
+
+	/* Discard previous AVB Parameters Table */
+	if (table->entry_count) {
+		kfree(table->entries);
+		table->entry_count = 0;
+	}
+
+	table->entries = kcalloc(SJA1105_MAX_AVB_PARAMS_COUNT,
+				 table->ops->unpacked_entry_size, GFP_KERNEL);
+	if (!table->entries)
+		return -ENOMEM;
+
+	table->entry_count = SJA1105_MAX_AVB_PARAMS_COUNT;
+
+	avb = table->entries;
+
+	/* Configure the MAC addresses for meta frames */
+	avb->destmeta = SJA1105_META_DMAC;
+	avb->srcmeta  = SJA1105_META_SMAC;
+
+	return 0;
+}
+
 #define SJA1105_RATE_MBPS(speed) (((speed) * 64000) / 1000)
 
 static void sja1105_setup_policer(struct sja1105_l2_policing_entry *policing,
@@ -567,6 +596,9 @@ static int sja1105_static_config_load(struct sja1105_private *priv,
 	if (rc < 0)
 		return rc;
 	rc = sja1105_init_general_params(priv);
+	if (rc < 0)
+		return rc;
+	rc = sja1105_init_avb_params(priv);
 	if (rc < 0)
 		return rc;
 
drivers/net/dsa/sja1105/sja1105_ptp.c +0 -38
index a836fc38c4a4..a07aaf55068f 100644
@@ -44,39 +44,6 @@ enum sja1105_ptp_clk_mode {
 #define ptp_data_to_sja1105(d) \
 		container_of((d), struct sja1105_private, ptp_data)
 
-static int sja1105_init_avb_params(struct sja1105_private *priv,
-				   bool on)
-{
-	struct sja1105_avb_params_entry *avb;
-	struct sja1105_table *table;
-
-	table = &priv->static_config.tables[BLK_IDX_AVB_PARAMS];
-
-	/* Discard previous AVB Parameters Table */
-	if (table->entry_count) {
-		kfree(table->entries);
-		table->entry_count = 0;
-	}
-
-	/* Configure the reception of meta frames only if requested */
-	if (!on)
-		return 0;
-
-	table->entries = kcalloc(SJA1105_MAX_AVB_PARAMS_COUNT,
-				 table->ops->unpacked_entry_size, GFP_KERNEL);
-	if (!table->entries)
-		return -ENOMEM;
-
-	table->entry_count = SJA1105_MAX_AVB_PARAMS_COUNT;
-
-	avb = table->entries;
-
-	avb->destmeta = SJA1105_META_DMAC;
-	avb->srcmeta  = SJA1105_META_SMAC;
-
-	return 0;
-}
-
 /* Must be called only with priv->tagger_data.state bit
  * SJA1105_HWTS_RX_EN cleared
  */
@@ -86,17 +53,12 @@ static int sja1105_change_rxtstamping(struct sja1105_private *priv,
 	struct sja1105_ptp_data *ptp_data = &priv->ptp_data;
 	struct sja1105_general_params_entry *general_params;
 	struct sja1105_table *table;
-	int rc;
 
 	table = &priv->static_config.tables[BLK_IDX_GENERAL_PARAMS];
 	general_params = table->entries;
 	general_params->send_meta1 = on;
 	general_params->send_meta0 = on;
 
-	rc = sja1105_init_avb_params(priv, on);
-	if (rc < 0)
-		return rc;
-
 	/* Initialize the meta state machine to a known state */
 	if (priv->tagger_data.stampable_skb) {
 		kfree_skb(priv->tagger_data.stampable_skb);
Signature · 1 line
2.17.1
author maintainer
#2 lore raw

PATCH 2/4 net: dsa: sja1105: make future_base_time a common helper

In reply to Vladimir Oltean

From: Vladimir Oltean <[email protected]>

Because the PTP_CLK pin starts toggling only at a time higher than the current PTP clock, this helper from the time-aware shaper code comes in handy here as well. We'll use it to transform generic user input for the perout request into valid input for the sja1105 hardware.

---
 drivers/net/dsa/sja1105/sja1105_ptp.h | 27 +++++++++++++++++++++++++++
 drivers/net/dsa/sja1105/sja1105_tas.c | 27 ---------------------------
 2 files changed, 27 insertions(+), 27 deletions(-)
drivers/net/dsa/sja1105/sja1105_ptp.h +27 -0
index 6f4a19eec709..119e345b40fc 100644
@@ -21,6 +21,33 @@ static inline s64 sja1105_ticks_to_ns(s64 ticks)
 	return ticks * SJA1105_TICK_NS;
 }
 
+/* Calculate the first base_time in the future that satisfies this
+ * relationship:
+ *
+ * future_base_time = base_time + N x cycle_time >= now, or
+ *
+ *      now - base_time
+ * N >= ---------------
+ *         cycle_time
+ *
+ * Because N is an integer, the ceiling value of the above "a / b" ratio
+ * is in fact precisely the floor value of "(a + b - 1) / b", which is
+ * easier to calculate only having integer division tools.
+ */
+static inline s64 future_base_time(s64 base_time, s64 cycle_time, s64 now)
+{
+	s64 a, b, n;
+
+	if (base_time >= now)
+		return base_time;
+
+	a = now - base_time;
+	b = cycle_time;
+	n = div_s64(a + b - 1, b);
+
+	return base_time + n * cycle_time;
+}
+
 struct sja1105_ptp_cmd {
 	u64 ptpstrtsch;		/* start schedule */
 	u64 ptpstopsch;		/* stop schedule */
drivers/net/dsa/sja1105/sja1105_tas.c +0 -27
index fa6750d973d7..77e547b4cd89 100644
@@ -28,33 +28,6 @@ static s64 sja1105_delta_to_ns(s64 delta)
 	return delta * 200;
 }
 
-/* Calculate the first base_time in the future that satisfies this
- * relationship:
- *
- * future_base_time = base_time + N x cycle_time >= now, or
- *
- *      now - base_time
- * N >= ---------------
- *         cycle_time
- *
- * Because N is an integer, the ceiling value of the above "a / b" ratio
- * is in fact precisely the floor value of "(a + b - 1) / b", which is
- * easier to calculate only having integer division tools.
- */
-static s64 future_base_time(s64 base_time, s64 cycle_time, s64 now)
-{
-	s64 a, b, n;
-
-	if (base_time >= now)
-		return base_time;
-
-	a = now - base_time;
-	b = cycle_time;
-	n = div_s64(a + b - 1, b);
-
-	return base_time + n * cycle_time;
-}
-
 static int sja1105_tas_set_runtime_params(struct sja1105_private *priv)
 {
 	struct sja1105_tas_data *tas_data = &priv->tas_data;
Signature · 1 line
2.17.1
author maintainer
#3 lore raw

PATCH 3/4 net: dsa: sja1105: make the AVB table dynamically reconfigurable

In reply to Vladimir Oltean

From: Vladimir Oltean <[email protected]>

The AVB table contains the CAS_MASTER field (to be added in the next patch) which decides the direction of the PTP_CLK pin.

Reconfiguring this field dynamically is highly preferable to having to reset the switch and upload a new static configuration, so we add support for exactly that.

---
 drivers/net/dsa/sja1105/sja1105.h             |  2 ++
 .../net/dsa/sja1105/sja1105_dynamic_config.c  | 24 ++++++++++++++++++-
 .../net/dsa/sja1105/sja1105_static_config.c   |  4 ++--
 3 files changed, 27 insertions(+), 3 deletions(-)
drivers/net/dsa/sja1105/sja1105.h +2 -0
index 4c40f2d51a54..5bc40175ee0d 100644
@@ -214,5 +214,7 @@ size_t sja1105_vlan_lookup_entry_packing(void *buf, void *entry_ptr,
 					 enum packing_op op);
 size_t sja1105pqrs_mac_config_entry_packing(void *buf, void *entry_ptr,
 					    enum packing_op op);
+size_t sja1105pqrs_avb_params_entry_packing(void *buf, void *entry_ptr,
+					    enum packing_op op);
 
 #endif
drivers/net/dsa/sja1105/sja1105_dynamic_config.c +23 -1
index 25381bd65ed7..bf9b36ff35bf 100644
@@ -124,6 +124,9 @@
 #define SJA1105ET_SIZE_GENERAL_PARAMS_DYN_CMD			\
 	SJA1105_SIZE_DYN_CMD
 
+#define SJA1105PQRS_SIZE_AVB_PARAMS_DYN_CMD			\
+	(SJA1105_SIZE_DYN_CMD + SJA1105PQRS_SIZE_AVB_PARAMS_ENTRY)
+
 #define SJA1105_MAX_DYN_CMD_SIZE				\
 	SJA1105PQRS_SIZE_MAC_CONFIG_DYN_CMD
 
@@ -481,6 +484,18 @@ sja1105et_general_params_entry_packing(void *buf, void *entry_ptr,
 	return 0;
 }
 
+static void
+sja1105pqrs_avb_params_cmd_packing(void *buf, struct sja1105_dyn_cmd *cmd,
+				   enum packing_op op)
+{
+	u8 *p = buf + SJA1105PQRS_SIZE_AVB_PARAMS_ENTRY;
+	const int size = SJA1105_SIZE_DYN_CMD;
+
+	sja1105_packing(p, &cmd->valid,   31, 31, size, op);
+	sja1105_packing(p, &cmd->errors,  30, 30, size, op);
+	sja1105_packing(p, &cmd->rdwrset, 29, 29, size, op);
+}
+
 #define OP_READ		BIT(0)
 #define OP_WRITE	BIT(1)
 #define OP_DEL		BIT(2)
@@ -610,7 +625,14 @@ struct sja1105_dynamic_table_ops sja1105pqrs_dyn_ops[BLK_IDX_MAX_DYN] = {
 		.addr = 0x38,
 	},
 	[BLK_IDX_L2_FORWARDING_PARAMS] = {0},
-	[BLK_IDX_AVB_PARAMS] = {0},
+	[BLK_IDX_AVB_PARAMS] = {
+		.entry_packing = sja1105pqrs_avb_params_entry_packing,
+		.cmd_packing = sja1105pqrs_avb_params_cmd_packing,
+		.max_entry_count = SJA1105_MAX_AVB_PARAMS_COUNT,
+		.access = (OP_READ | OP_WRITE),
+		.packed_size = SJA1105PQRS_SIZE_AVB_PARAMS_DYN_CMD,
+		.addr = 0x8003,
+	},
 	[BLK_IDX_GENERAL_PARAMS] = {
 		.entry_packing = sja1105et_general_params_entry_packing,
 		.cmd_packing = sja1105et_general_params_cmd_packing,
drivers/net/dsa/sja1105/sja1105_static_config.c +2 -2
index 63d2311817c4..fccc9c4d6b35 100644
@@ -102,8 +102,8 @@ static size_t sja1105et_avb_params_entry_packing(void *buf, void *entry_ptr,
 	return size;
 }
 
-static size_t sja1105pqrs_avb_params_entry_packing(void *buf, void *entry_ptr,
-						   enum packing_op op)
+size_t sja1105pqrs_avb_params_entry_packing(void *buf, void *entry_ptr,
+					    enum packing_op op)
 {
 	const size_t size = SJA1105PQRS_SIZE_AVB_PARAMS_ENTRY;
 	struct sja1105_avb_params_entry *entry = entry_ptr;
Signature · 1 line
2.17.1
author maintainer
#4 lore raw

PATCH 4/4 net: dsa: sja1105: configure the PTP_CLK pin as EXT_TS or PER_OUT

In reply to Vladimir Oltean

From: Vladimir Oltean <[email protected]>

The SJA1105 switch family has a PTP_CLK pin which emits a signal with fixed 50% duty cycle, but variable frequency and programmable start time.

On the second generation (P/Q/R/S) switches, this pin supports even more functionality. The use case described by the hardware documents talks about synchronization via oneshot pulses: given 2 sja1105 switches, arbitrarily designated as a master and a slave, the master emits a single pulse on PTP_CLK, while the slave is configured to timestamp this pulse received on its PTP_CLK pin (which must obviously be configured as input). The difference between the timestamps then exactly becomes the slave offset to the master.

The only trouble with the above is that the hardware is very much tied into this use case only, and not very generic beyond that:
- When emitting a oneshot pulse, instead of being told when to emit it, the switch just does it "now" and tells you later what time it was, via the PTPSYNCTS register. [ Incidentally, this is the same register that the slave uses to collect the ext_ts timestamp from, too. ]
- On the sync slave, there is no interrupt mechanism on reception of a new extts, and no FIFO to buffer them, because in the foreseen use case, software is in control of both the master and the slave pins, so it "knows" when there's something to collect.

These 2 problems mean that:
- We don't support (at least yet) the quirky oneshot mode exposed by the hardware, just normal periodic output.
- We abuse the hardware a little bit when we expose generic extts. Because there's no interrupt mechanism, we need to poll at double the frequency we expect to receive a pulse. Currently that means a non-configurable "twice a second".

---
 drivers/net/dsa/sja1105/sja1105.h             |   3 +
 drivers/net/dsa/sja1105/sja1105_main.c        |   8 +
 drivers/net/dsa/sja1105/sja1105_ptp.c         | 247 ++++++++++++++++++
 drivers/net/dsa/sja1105/sja1105_ptp.h         |   4 +
 drivers/net/dsa/sja1105/sja1105_spi.c         |   5 +
 .../net/dsa/sja1105/sja1105_static_config.c   |   1 +
 .../net/dsa/sja1105/sja1105_static_config.h   |   1 +
 7 files changed, 269 insertions(+)
drivers/net/dsa/sja1105/sja1105.h +3 -0
index 5bc40175ee0d..a358fc89a6db 100644
@@ -38,10 +38,13 @@ struct sja1105_regs {
 	u64 config;
 	u64 sgmii;
 	u64 rmii_pll1;
+	u64 ptppinst;
+	u64 ptppindur;
 	u64 ptp_control;
 	u64 ptpclkval;
 	u64 ptpclkrate;
 	u64 ptpclkcorp;
+	u64 ptpsyncts;
 	u64 ptpschtm;
 	u64 ptpegr_ts[SJA1105_NUM_PORTS];
 	u64 pad_mii_tx[SJA1105_NUM_PORTS];
drivers/net/dsa/sja1105/sja1105_main.c +8 -0
index 4d182293ca1f..e0c99bb63cdf 100644
@@ -504,6 +504,14 @@ static int sja1105_init_avb_params(struct sja1105_private *priv)
 	/* Configure the MAC addresses for meta frames */
 	avb->destmeta = SJA1105_META_DMAC;
 	avb->srcmeta  = SJA1105_META_SMAC;
+	/* On P/Q/R/S, configure the direction of the PTP_CLK pin as input by
+	 * default. This is because there might be boards with a hardware
+	 * layout where enabling the pin as output might cause an electrical
+	 * clash. On E/T the pin is always an output, which the board designers
+	 * probably already knew, so even if there are going to be electrical
+	 * issues, there's nothing we can do.
+	 */
+	avb->cas_master = false;
 
 	return 0;
 }
drivers/net/dsa/sja1105/sja1105_ptp.c +247 -0

297 lines not on this page · raw message

drivers/net/dsa/sja1105/sja1105_ptp.h +4 -0
index 119e345b40fc..43480b24f1f0 100644
@@ -49,6 +49,8 @@ static inline s64 future_base_time(s64 base_time, s64 cycle_time, s64 now)
 }
 
 struct sja1105_ptp_cmd {
+	u64 startptpcp;		/* start toggling PTP_CLK pin */
+	u64 stopptpcp;		/* stop toggling PTP_CLK pin */
 	u64 ptpstrtsch;		/* start schedule */
 	u64 ptpstopsch;		/* stop schedule */
 	u64 resptp;		/* reset */
@@ -57,12 +59,14 @@ struct sja1105_ptp_cmd {
 };
 
 struct sja1105_ptp_data {
+	struct delayed_work extts_work;
 	struct sk_buff_head skb_rxtstamp_queue;
 	struct ptp_clock_info caps;
 	struct ptp_clock *clock;
 	struct sja1105_ptp_cmd cmd;
 	/* Serializes all operations on the PTP hardware clock */
 	struct mutex lock;
+	u64 ptpsyncts;
 };
 
 int sja1105_ptp_clock_register(struct dsa_switch *ds);
drivers/net/dsa/sja1105/sja1105_spi.c +5 -0
index 45da162ba268..fef2c50cd3f6 100644
@@ -458,6 +458,8 @@ static struct sja1105_regs sja1105et_regs = {
 	.rmii_ext_tx_clk = {0x100018, 0x10001F, 0x100026, 0x10002D, 0x100034},
 	.ptpegr_ts = {0xC0, 0xC2, 0xC4, 0xC6, 0xC8},
 	.ptpschtm = 0x12, /* Spans 0x12 to 0x13 */
+	.ptppinst = 0x14,
+	.ptppindur = 0x16,
 	.ptp_control = 0x17,
 	.ptpclkval = 0x18, /* Spans 0x18 to 0x19 */
 	.ptpclkrate = 0x1A,
@@ -491,10 +493,13 @@ static struct sja1105_regs sja1105pqrs_regs = {
 	.qlevel = {0x604, 0x614, 0x624, 0x634, 0x644},
 	.ptpegr_ts = {0xC0, 0xC4, 0xC8, 0xCC, 0xD0},
 	.ptpschtm = 0x13, /* Spans 0x13 to 0x14 */
+	.ptppinst = 0x15,
+	.ptppindur = 0x17,
 	.ptp_control = 0x18,
 	.ptpclkval = 0x19,
 	.ptpclkrate = 0x1B,
 	.ptpclkcorp = 0x1E,
+	.ptpsyncts = 0x1F,
 };
 
 struct sja1105_info sja1105e_info = {
drivers/net/dsa/sja1105/sja1105_static_config.c +1 -0
index fccc9c4d6b35..bbfe034910a0 100644
@@ -108,6 +108,7 @@ size_t sja1105pqrs_avb_params_entry_packing(void *buf, void *entry_ptr,
 	const size_t size = SJA1105PQRS_SIZE_AVB_PARAMS_ENTRY;
 	struct sja1105_avb_params_entry *entry = entry_ptr;
 
+	sja1105_packing(buf, &entry->cas_master, 126, 126, size, op);
 	sja1105_packing(buf, &entry->destmeta,   125,  78, size, op);
 	sja1105_packing(buf, &entry->srcmeta,     77,  30, size, op);
 	return size;
drivers/net/dsa/sja1105/sja1105_static_config.h +1 -0
index f4a5c5c04311..8afafb6aef12 100644
@@ -230,6 +230,7 @@ struct sja1105_l2_policing_entry {
 };
 
 struct sja1105_avb_params_entry {
+	u64 cas_master;
 	u64 destmeta;
 	u64 srcmeta;
 };
Signature · 1 line
2.17.1
Acked-by on v1 4/4
#5 lore raw

Re: [PATCH net-next 4/4] net: dsa: sja1105: configure the PTP_CLK pin as EXT_TS or PER_OUT

Quoted 33 lines · Vladimir Oltean

From: Vladimir Oltean <[email protected]>

The SJA1105 switch family has a PTP_CLK pin which emits a signal with fixed 50% duty cycle, but variable frequency and programmable start time.

On the second generation (P/Q/R/S) switches, this pin supports even more functionality. The use case described by the hardware documents talks about synchronization via oneshot pulses: given 2 sja1105 switches, arbitrarily designated as a master and a slave, the master emits a single pulse on PTP_CLK, while the slave is configured to timestamp this pulse received on its PTP_CLK pin (which must obviously be configured as input). The difference between the timestamps then exactly becomes the slave offset to the master.

The only trouble with the above is that the hardware is very much tied into this use case only, and not very generic beyond that:
- When emitting a oneshot pulse, instead of being told when to emit it, the switch just does it "now" and tells you later what time it was, via the PTPSYNCTS register. [ Incidentally, this is the same register that the slave uses to collect the ext_ts timestamp from, too. ]
- On the sync slave, there is no interrupt mechanism on reception of a new extts, and no FIFO to buffer them, because in the foreseen use case, software is in control of both the master and the slave pins, so it "knows" when there's something to collect.

These 2 problems mean that:
- We don't support (at least yet) the quirky oneshot mode exposed by the hardware, just normal periodic output.
- We abuse the hardware a little bit when we expose generic extts. Because there's no interrupt mechanism, we need to poll at double the frequency we expect to receive a pulse. Currently that means a non-configurable "twice a second".

Quoted 1 line