[PATCH net-next v3 0/9] Support external snapshots on dwmac1000

[PATCH net-next v3 0/9] Support external snapshots on dwmac1000

[Maxime Chevallier]

Hi,

This V3 is just a rebase a V2 on top of net-next to address some minor
conflicts. No changes were made.

This series is another take on the pervious work [1] done by
Alexis Lothoré, that fixes the support for external snapshots
timestamping in GMAC3-based devices.

Details on why this is needed are mentionned on the cover [2] from V1.

This V2 addresses multiple issues found in V1 :

 - The PTP_TCR register is configured from multiple places, as reported
   by Alexis, so we need to make sure that the extts configuration
   doesn't interfere with the hwtstamp configuration.

 - The interrupt management in V1 was incomplete, as the interrupt
   wasn't correctly acked.

 - This series also makes so that we only enable the extts interrupt
   when necessary.

[1]: https://lore.kernel.org/netdev/20230616100409.164583-1-alexis.lothore@bootlin.com/
[2]: https://lore.kernel.org/netdev/20241029115419.1160201-1-maxime.chevallier@bootlin.com/

Thanks Alexis for laying the groundwork for this,

Best regards,

Maxime

Link to V1: https://lore.kernel.org/netdev/20241029115419.1160201-1-maxime.chevallier@bootlin.com/
Link to V2: https://lore.kernel.org/netdev/20241104170251.2202270-1-maxime.chevallier@bootlin.com/

Maxime Chevallier (9):
  net: stmmac: Don't modify the global ptp ops directly
  net: stmmac: Use per-hw ptp clock ops
  net: stmmac: Only update the auto-discovered PTP clock features
  net: stmmac: Introduce dwmac1000 ptp_clock_info and operations
  net: stmmac: Introduce dwmac1000 timestamping operations
  net: stmmac: Enable timestamping interrupt on dwmac1000
  net: stmmac: Don't include dwmac4 definitions in stmmac_ptp
  net: stmmac: Configure only the relevant bits for timestamping setup
  net: stmmac: dwmac_socfpga: This platform has GMAC

 drivers/net/ethernet/stmicro/stmmac/common.h  |   4 +
 .../ethernet/stmicro/stmmac/dwmac-socfpga.c   |   1 +
 .../net/ethernet/stmicro/stmmac/dwmac1000.h   |  12 +++
 .../ethernet/stmicro/stmmac/dwmac1000_core.c  | 101 ++++++++++++++++++
 drivers/net/ethernet/stmicro/stmmac/hwif.c    |  15 ++-
 .../ethernet/stmicro/stmmac/stmmac_hwtstamp.c |  26 ++++-
 .../net/ethernet/stmicro/stmmac/stmmac_ptp.c  |  38 +++++--
 .../net/ethernet/stmicro/stmmac/stmmac_ptp.h  |  10 ++
 8 files changed, 196 insertions(+), 11 deletions(-)

-- 
2.47.0

[PATCH net-next v3 1/9] net: stmmac: Don't modify the global ptp ops directly

[Maxime Chevallier]

The stmmac_ptp_clock_ops are copied into the stmmac_priv structure
before being registered to the PTP core. Some adjustments are made prior
to that, such as the number of snapshots or max adjustment parameters.

Instead of modifying the global definition, then copying into the local
private data, let's first copy then modify the local parameters.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
index a6b1de9a251d..11ab1d6b916a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
@@ -298,20 +298,21 @@ void stmmac_ptp_register(struct stmmac_priv *priv)
 		priv->pps[i].available = true;
 	}
 
-	if (priv->plat->ptp_max_adj)
-		stmmac_ptp_clock_ops.max_adj = priv->plat->ptp_max_adj;
-
 	/* Calculate the clock domain crossing (CDC) error if necessary */
 	priv->plat->cdc_error_adj = 0;
 	if (priv->plat->has_gmac4 && priv->plat->clk_ptp_rate)
 		priv->plat->cdc_error_adj = (2 * NSEC_PER_SEC) / priv->plat->clk_ptp_rate;
 
-	stmmac_ptp_clock_ops.n_per_out = priv->dma_cap.pps_out_num;
-	stmmac_ptp_clock_ops.n_ext_ts = priv->dma_cap.aux_snapshot_n;
+	priv->ptp_clock_ops = stmmac_ptp_clock_ops;
+
+	priv->ptp_clock_ops.n_per_out = priv->dma_cap.pps_out_num;
+	priv->ptp_clock_ops.n_ext_ts = priv->dma_cap.aux_snapshot_n;
+
+	if (priv->plat->ptp_max_adj)
+		priv->ptp_clock_ops.max_adj = priv->plat->ptp_max_adj;
 
 	rwlock_init(&priv->ptp_lock);
 	mutex_init(&priv->aux_ts_lock);
-	priv->ptp_clock_ops = stmmac_ptp_clock_ops;
 
 	priv->ptp_clock = ptp_clock_register(&priv->ptp_clock_ops,
 					     priv->device);
-- 
2.47.0

[PATCH net-next v3 2/9] net: stmmac: Use per-hw ptp clock ops

[Maxime Chevallier]

The auxiliary snapshot configuration was found to differ depending on
the dwmac version. To prepare supporting this, allow specifying the
ptp_clock_info ops in the hwif array

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/common.h     |  2 ++
 drivers/net/ethernet/stmicro/stmmac/hwif.c       | 11 +++++++++++
 drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c |  4 +---
 3 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index 684489156dce..4a0a1708c391 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -551,6 +551,8 @@ struct mac_device_info;
 extern const struct stmmac_hwtimestamp stmmac_ptp;
 extern const struct stmmac_mode_ops dwmac4_ring_mode_ops;
 
+extern const struct ptp_clock_info stmmac_ptp_clock_ops;
+
 struct mac_link {
 	u32 caps;
 	u32 speed_mask;
diff --git a/drivers/net/ethernet/stmicro/stmmac/hwif.c b/drivers/net/ethernet/stmicro/stmmac/hwif.c
index cfc50289aed6..47458cbcbc94 100644
--- a/drivers/net/ethernet/stmicro/stmmac/hwif.c
+++ b/drivers/net/ethernet/stmicro/stmmac/hwif.c
@@ -113,6 +113,7 @@ static const struct stmmac_hwif_entry {
 	const void *dma;
 	const void *mac;
 	const void *hwtimestamp;
+	const void *ptp;
 	const void *mode;
 	const void *tc;
 	const void *mmc;
@@ -134,6 +135,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac100_dma_ops,
 		.mac = &dwmac100_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
 		.mmc = &dwmac_mmc_ops,
@@ -152,6 +154,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac1000_dma_ops,
 		.mac = &dwmac1000_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
 		.mmc = &dwmac_mmc_ops,
@@ -171,6 +174,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac4_dma_ops,
 		.mac = &dwmac4_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = NULL,
 		.tc = &dwmac4_tc_ops,
 		.mmc = &dwmac_mmc_ops,
@@ -192,6 +196,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac4_dma_ops,
 		.mac = &dwmac410_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = &dwmac4_ring_mode_ops,
 		.tc = &dwmac510_tc_ops,
 		.mmc = &dwmac_mmc_ops,
@@ -213,6 +218,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac410_dma_ops,
 		.mac = &dwmac410_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = &dwmac4_ring_mode_ops,
 		.tc = &dwmac510_tc_ops,
 		.mmc = &dwmac_mmc_ops,
@@ -234,6 +240,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac410_dma_ops,
 		.mac = &dwmac510_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = &dwmac4_ring_mode_ops,
 		.tc = &dwmac510_tc_ops,
 		.mmc = &dwmac_mmc_ops,
@@ -256,6 +263,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwxgmac210_dma_ops,
 		.mac = &dwxgmac210_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = NULL,
 		.tc = &dwxgmac_tc_ops,
 		.mmc = &dwxgmac_mmc_ops,
@@ -278,6 +286,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwxgmac210_dma_ops,
 		.mac = &dwxlgmac2_ops,
 		.hwtimestamp = &stmmac_ptp,
+		.ptp = &stmmac_ptp_clock_ops,
 		.mode = NULL,
 		.tc = &dwxgmac_tc_ops,
 		.mmc = &dwxgmac_mmc_ops,
@@ -362,6 +371,8 @@ int stmmac_hwif_init(struct stmmac_priv *priv)
 		priv->fpe_cfg.reg = entry->regs.fpe_reg;
 		priv->ptpaddr = priv->ioaddr + entry->regs.ptp_off;
 		priv->mmcaddr = priv->ioaddr + entry->regs.mmc_off;
+		memcpy(&priv->ptp_clock_ops, entry->ptp,
+		       sizeof(struct ptp_clock_info));
 		if (entry->est)
 			priv->estaddr = priv->ioaddr + entry->regs.est_off;
 
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
index 11ab1d6b916a..41581f516ea9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
@@ -265,7 +265,7 @@ static int stmmac_getcrosststamp(struct ptp_clock_info *ptp,
 }
 
 /* structure describing a PTP hardware clock */
-static struct ptp_clock_info stmmac_ptp_clock_ops = {
+const struct ptp_clock_info stmmac_ptp_clock_ops = {
 	.owner = THIS_MODULE,
 	.name = "stmmac ptp",
 	.max_adj = 62500000,
@@ -303,8 +303,6 @@ void stmmac_ptp_register(struct stmmac_priv *priv)
 	if (priv->plat->has_gmac4 && priv->plat->clk_ptp_rate)
 		priv->plat->cdc_error_adj = (2 * NSEC_PER_SEC) / priv->plat->clk_ptp_rate;
 
-	priv->ptp_clock_ops = stmmac_ptp_clock_ops;
-
 	priv->ptp_clock_ops.n_per_out = priv->dma_cap.pps_out_num;
 	priv->ptp_clock_ops.n_ext_ts = priv->dma_cap.aux_snapshot_n;
 
-- 
2.47.0

[PATCH net-next v3 3/9] net: stmmac: Only update the auto-discovered PTP clock features

[Maxime Chevallier]

Some DWMAC variants such as dwmac1000 don't support discovering the
number of output pps and auxiliary snapshots. Allow these parameters to
be defined in default ptp_clock_info, and let them be updated only when
the feature discovery yielded a result.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
index 41581f516ea9..8ea2b4226234 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
@@ -303,8 +303,14 @@ void stmmac_ptp_register(struct stmmac_priv *priv)
 	if (priv->plat->has_gmac4 && priv->plat->clk_ptp_rate)
 		priv->plat->cdc_error_adj = (2 * NSEC_PER_SEC) / priv->plat->clk_ptp_rate;
 
-	priv->ptp_clock_ops.n_per_out = priv->dma_cap.pps_out_num;
-	priv->ptp_clock_ops.n_ext_ts = priv->dma_cap.aux_snapshot_n;
+	/* Update the ptp clock parameters based on feature discovery, when
+	 * available
+	 */
+	if (priv->dma_cap.pps_out_num)
+		priv->ptp_clock_ops.n_per_out = priv->dma_cap.pps_out_num;
+
+	if (priv->dma_cap.aux_snapshot_n)
+		priv->ptp_clock_ops.n_ext_ts = priv->dma_cap.aux_snapshot_n;
 
 	if (priv->plat->ptp_max_adj)
 		priv->ptp_clock_ops.max_adj = priv->plat->ptp_max_adj;
-- 
2.47.0

[PATCH net-next v3 4/9] net: stmmac: Introduce dwmac1000 ptp_clock_info and operations

[Maxime Chevallier]

The PTP configuration for GMAC3_X differs from the other implementations
in several ways :

 - There's only one external snapshot trigger
 - The snapshot configuration is done through the PTP_TCR register,
   whereas the other dwmac variants have a dedicated ACR (auxiliary
   control reg) for that purpose
 - The layout for the PTP_TCR register also differs, as bits 24/25 are
   used for the snapshot configuration. These bits are reserved on other
   variants.

On GMAC3_X, we also can't discover the number of snapshot triggers
automatically.

The GMAC3_X has one PPS output, however it's configuration isn't
supported yet so report 0 n_per_out for now.

Introduce a dedicated set of ptp_clock_info ops and configuration
parameters to reflect these differences specific to GMAC3_X.

This was tested on dwmac_socfpga.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/common.h  |  1 +
 .../net/ethernet/stmicro/stmmac/dwmac1000.h   |  5 +++
 .../ethernet/stmicro/stmmac/dwmac1000_core.c  | 45 +++++++++++++++++++
 drivers/net/ethernet/stmicro/stmmac/hwif.c    |  4 +-
 .../net/ethernet/stmicro/stmmac/stmmac_ptp.c  | 18 ++++++++
 .../net/ethernet/stmicro/stmmac/stmmac_ptp.h  |  6 +++
 6 files changed, 77 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index 4a0a1708c391..6f68a6b298c9 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -552,6 +552,7 @@ extern const struct stmmac_hwtimestamp stmmac_ptp;
 extern const struct stmmac_mode_ops dwmac4_ring_mode_ops;
 
 extern const struct ptp_clock_info stmmac_ptp_clock_ops;
+extern const struct ptp_clock_info dwmac1000_ptp_clock_ops;
 
 struct mac_link {
 	u32 caps;
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h b/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
index 4296ddda8aaa..01eafeb1272f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
@@ -329,5 +329,10 @@ enum rtc_control {
 #define GMAC_MMC_RX_CSUM_OFFLOAD   0x208
 #define GMAC_EXTHASH_BASE  0x500
 
+/* PTP and timestamping registers */
+
+#define GMAC_PTP_TCR_ATSFC	BIT(24)
+#define GMAC_PTP_TCR_ATSEN0	BIT(25)
+
 extern const struct stmmac_dma_ops dwmac1000_dma_ops;
 #endif /* __DWMAC1000_H__ */
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
index d413d76a8936..b6930009ea06 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
@@ -18,6 +18,7 @@
 #include <linux/io.h>
 #include "stmmac.h"
 #include "stmmac_pcs.h"
+#include "stmmac_ptp.h"
 #include "dwmac1000.h"
 
 static void dwmac1000_core_init(struct mac_device_info *hw,
@@ -551,3 +552,47 @@ int dwmac1000_setup(struct stmmac_priv *priv)
 
 	return 0;
 }
+
+/* DWMAC 1000 ptp_clock_info ops */
+
+int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
+			 struct ptp_clock_request *rq, int on)
+{
+	struct stmmac_priv *priv =
+	    container_of(ptp, struct stmmac_priv, ptp_clock_ops);
+	void __iomem *ptpaddr = priv->ptpaddr;
+	int ret = -EOPNOTSUPP;
+	u32 tcr_val;
+
+	switch (rq->type) {
+	case PTP_CLK_REQ_EXTTS:
+		mutex_lock(&priv->aux_ts_lock);
+		tcr_val = readl(ptpaddr + PTP_TCR);
+
+		if (on) {
+			tcr_val |= GMAC_PTP_TCR_ATSEN0;
+			tcr_val |= GMAC_PTP_TCR_ATSFC;
+			priv->plat->flags |= STMMAC_FLAG_EXT_SNAPSHOT_EN;
+		} else {
+			tcr_val &= ~GMAC_PTP_TCR_ATSEN0;
+			priv->plat->flags &= ~STMMAC_FLAG_EXT_SNAPSHOT_EN;
+		}
+
+		netdev_dbg(priv->dev, "Auxiliary Snapshot %s.\n",
+			   on ? "enabled" : "disabled");
+		writel(tcr_val, ptpaddr + PTP_TCR);
+
+		mutex_unlock(&priv->aux_ts_lock);
+
+		/* wait for auxts fifo clear to finish */
+		ret = readl_poll_timeout(ptpaddr + PTP_TCR, tcr_val,
+					 !(tcr_val & GMAC_PTP_TCR_ATSFC),
+					 10, 10000);
+		break;
+
+	default:
+		break;
+	}
+
+	return ret;
+}
diff --git a/drivers/net/ethernet/stmicro/stmmac/hwif.c b/drivers/net/ethernet/stmicro/stmmac/hwif.c
index 47458cbcbc94..1f508843fb5a 100644
--- a/drivers/net/ethernet/stmicro/stmmac/hwif.c
+++ b/drivers/net/ethernet/stmicro/stmmac/hwif.c
@@ -135,7 +135,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac100_dma_ops,
 		.mac = &dwmac100_ops,
 		.hwtimestamp = &stmmac_ptp,
-		.ptp = &stmmac_ptp_clock_ops,
+		.ptp = &dwmac1000_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
 		.mmc = &dwmac_mmc_ops,
@@ -154,7 +154,7 @@ static const struct stmmac_hwif_entry {
 		.dma = &dwmac1000_dma_ops,
 		.mac = &dwmac1000_ops,
 		.hwtimestamp = &stmmac_ptp,
-		.ptp = &stmmac_ptp_clock_ops,
+		.ptp = &dwmac1000_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
 		.mmc = &dwmac_mmc_ops,
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
index 8ea2b4226234..430905f591b2 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
@@ -282,6 +282,24 @@ const struct ptp_clock_info stmmac_ptp_clock_ops = {
 	.getcrosststamp = stmmac_getcrosststamp,
 };
 
+/* structure describing a PTP hardware clock */
+const struct ptp_clock_info dwmac1000_ptp_clock_ops = {
+	.owner = THIS_MODULE,
+	.name = "stmmac ptp",
+	.max_adj = 62500000,
+	.n_alarm = 0,
+	.n_ext_ts = 1,
+	.n_per_out = 0,
+	.n_pins = 0,
+	.pps = 0,
+	.adjfine = stmmac_adjust_freq,
+	.adjtime = stmmac_adjust_time,
+	.gettime64 = stmmac_get_time,
+	.settime64 = stmmac_set_time,
+	.enable = dwmac1000_ptp_enable,
+	.getcrosststamp = stmmac_getcrosststamp,
+};
+
 /**
  * stmmac_ptp_register
  * @priv: driver private structure
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
index fce3fba2ffd2..fa4611855311 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
@@ -94,4 +94,10 @@ enum aux_snapshot {
 	AUX_SNAPSHOT3 = 0x80,
 };
 
+struct ptp_clock_info;
+struct ptp_clock_request;
+
+int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
+			 struct ptp_clock_request *rq, int on);
+
 #endif	/* __STMMAC_PTP_H__ */
-- 
2.47.0

Re: [PATCH net-next v3 4/9] net: stmmac: Introduce dwmac1000 ptp_clock_info and operations

[Jakub Kicinski]

On Wed,  6 Nov 2024 10:03:25 +0100 Maxime Chevallier wrote:
> +		mutex_unlock(&priv->aux_ts_lock);
> +
> +		/* wait for auxts fifo clear to finish */
> +		ret = readl_poll_timeout(ptpaddr + PTP_TCR, tcr_val,
> +					 !(tcr_val & GMAC_PTP_TCR_ATSFC),
> +					 10, 10000);

Is there a good reason to wait for the flush to complete outside of 
the mutex? 

Re: [PATCH net-next v3 4/9] net: stmmac: Introduce dwmac1000 ptp_clock_info and operations

[Paolo Abeni]

On 11/12/24 01:12, Jakub Kicinski wrote:
> On Wed,  6 Nov 2024 10:03:25 +0100 Maxime Chevallier wrote:
>> +		mutex_unlock(&priv->aux_ts_lock);
>> +
>> +		/* wait for auxts fifo clear to finish */
>> +		ret = readl_poll_timeout(ptpaddr + PTP_TCR, tcr_val,
>> +					 !(tcr_val & GMAC_PTP_TCR_ATSFC),
>> +					 10, 10000);
> 
> Is there a good reason to wait for the flush to complete outside of 
> the mutex? 

Indeed looking at other `ptpaddr` access use-case, it looks like the
mutex protects both read and write accesses.

@Maxime: is the above intentional? looks race-prone

Thanks,

Paolo

Re: [PATCH net-next v3 4/9] net: stmmac: Introduce dwmac1000 ptp_clock_info and operations

[Maxime Chevallier]

Hello Jakub, Paolo,

On Tue, 12 Nov 2024 10:28:21 +0100
Paolo Abeni <pabeni@redhat.com> wrote:

> On 11/12/24 01:12, Jakub Kicinski wrote:
> > On Wed,  6 Nov 2024 10:03:25 +0100 Maxime Chevallier wrote:  
> >> +		mutex_unlock(&priv->aux_ts_lock);
> >> +
> >> +		/* wait for auxts fifo clear to finish */
> >> +		ret = readl_poll_timeout(ptpaddr + PTP_TCR, tcr_val,
> >> +					 !(tcr_val & GMAC_PTP_TCR_ATSFC),
> >> +					 10, 10000);  
> > 
> > Is there a good reason to wait for the flush to complete outside of 
> > the mutex?   
> 
> Indeed looking at other `ptpaddr` access use-case, it looks like the
> mutex protects both read and write accesses.
> 
> @Maxime: is the above intentional? looks race-prone

You're right, this is racy... It wasn't intentionnal, it's actually the
same logic as dwmac4 uses so looks like dwmac4 is also incorrect in
that regard.

I'll send a v4 with that change, and a fix for dwmac4 along the way
then.

Thanks for spotting this,

Maxime

[PATCH net-next v3 5/9] net: stmmac: Introduce dwmac1000 timestamping operations

[Maxime Chevallier]

In GMAC3_X, the timestamping configuration differs from GMAC4 in the
layout of the registers accessed to grab the number of snapshots in FIFO
as well as the register offset to grab the aux snapshot timestamp.

Introduce dedicated ops to configure timestamping on dwmac100 and
dwmac1000. The latency correction doesn't seem to exist on GMAC3, so its
corresponding operation isn't populated.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/common.h  |  1 +
 .../net/ethernet/stmicro/stmmac/dwmac1000.h   |  7 ++++
 .../ethernet/stmicro/stmmac/dwmac1000_core.c  | 40 +++++++++++++++++++
 drivers/net/ethernet/stmicro/stmmac/hwif.c    |  4 +-
 .../ethernet/stmicro/stmmac/stmmac_hwtstamp.c | 11 +++++
 .../net/ethernet/stmicro/stmmac/stmmac_ptp.h  |  4 ++
 6 files changed, 65 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/common.h b/drivers/net/ethernet/stmicro/stmmac/common.h
index 6f68a6b298c9..1367fa5c9b8e 100644
--- a/drivers/net/ethernet/stmicro/stmmac/common.h
+++ b/drivers/net/ethernet/stmicro/stmmac/common.h
@@ -549,6 +549,7 @@ extern const struct stmmac_desc_ops ndesc_ops;
 struct mac_device_info;
 
 extern const struct stmmac_hwtimestamp stmmac_ptp;
+extern const struct stmmac_hwtimestamp dwmac1000_ptp;
 extern const struct stmmac_mode_ops dwmac4_ring_mode_ops;
 
 extern const struct ptp_clock_info stmmac_ptp_clock_ops;
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h b/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
index 01eafeb1272f..600fea8f712f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000.h
@@ -331,8 +331,15 @@ enum rtc_control {
 
 /* PTP and timestamping registers */
 
+#define GMAC3_X_ATSNS       GENMASK(19, 16)
+#define GMAC3_X_ATSNS_SHIFT 16
+
 #define GMAC_PTP_TCR_ATSFC	BIT(24)
 #define GMAC_PTP_TCR_ATSEN0	BIT(25)
 
+#define GMAC3_X_TIMESTAMP_STATUS	0x28
+#define GMAC_PTP_ATNR	0x30
+#define GMAC_PTP_ATSR	0x34
+
 extern const struct stmmac_dma_ops dwmac1000_dma_ops;
 #endif /* __DWMAC1000_H__ */
diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
index b6930009ea06..dbbd834f9fc8 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
@@ -553,6 +553,46 @@ int dwmac1000_setup(struct stmmac_priv *priv)
 	return 0;
 }
 
+/* DWMAC 1000 HW Timestaming ops */
+
+void dwmac1000_get_ptptime(void __iomem *ptpaddr, u64 *ptp_time)
+{
+	u64 ns;
+
+	ns = readl(ptpaddr + GMAC_PTP_ATNR);
+	ns += readl(ptpaddr + GMAC_PTP_ATSR) * NSEC_PER_SEC;
+
+	*ptp_time = ns;
+}
+
+void dwmac1000_timestamp_interrupt(struct stmmac_priv *priv)
+{
+	struct ptp_clock_event event;
+	u32 ts_status, num_snapshot;
+	unsigned long flags;
+	u64 ptp_time;
+	int i;
+
+	/* Clears the timestamp interrupt */
+	ts_status = readl(priv->ptpaddr + GMAC3_X_TIMESTAMP_STATUS);
+
+	if (!(priv->plat->flags & STMMAC_FLAG_EXT_SNAPSHOT_EN))
+		return;
+
+	num_snapshot = (ts_status & GMAC3_X_ATSNS) >> GMAC3_X_ATSNS_SHIFT;
+
+	for (i = 0; i < num_snapshot; i++) {
+		read_lock_irqsave(&priv->ptp_lock, flags);
+		stmmac_get_ptptime(priv, priv->ptpaddr, &ptp_time);
+		read_unlock_irqrestore(&priv->ptp_lock, flags);
+
+		event.type = PTP_CLOCK_EXTTS;
+		event.index = 0;
+		event.timestamp = ptp_time;
+		ptp_clock_event(priv->ptp_clock, &event);
+	}
+}
+
 /* DWMAC 1000 ptp_clock_info ops */
 
 int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
diff --git a/drivers/net/ethernet/stmicro/stmmac/hwif.c b/drivers/net/ethernet/stmicro/stmmac/hwif.c
index 1f508843fb5a..a72d336a8350 100644
--- a/drivers/net/ethernet/stmicro/stmmac/hwif.c
+++ b/drivers/net/ethernet/stmicro/stmmac/hwif.c
@@ -134,7 +134,7 @@ static const struct stmmac_hwif_entry {
 		.desc = NULL,
 		.dma = &dwmac100_dma_ops,
 		.mac = &dwmac100_ops,
-		.hwtimestamp = &stmmac_ptp,
+		.hwtimestamp = &dwmac1000_ptp,
 		.ptp = &dwmac1000_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
@@ -153,7 +153,7 @@ static const struct stmmac_hwif_entry {
 		.desc = NULL,
 		.dma = &dwmac1000_dma_ops,
 		.mac = &dwmac1000_ops,
-		.hwtimestamp = &stmmac_ptp,
+		.hwtimestamp = &dwmac1000_ptp,
 		.ptp = &dwmac1000_ptp_clock_ops,
 		.mode = NULL,
 		.tc = NULL,
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
index 5ef52ef2698f..a94829ef8cfb 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
@@ -269,3 +269,14 @@ const struct stmmac_hwtimestamp stmmac_ptp = {
 	.timestamp_interrupt = timestamp_interrupt,
 	.hwtstamp_correct_latency = hwtstamp_correct_latency,
 };
+
+const struct stmmac_hwtimestamp dwmac1000_ptp = {
+	.config_hw_tstamping = config_hw_tstamping,
+	.init_systime = init_systime,
+	.config_sub_second_increment = config_sub_second_increment,
+	.config_addend = config_addend,
+	.adjust_systime = adjust_systime,
+	.get_systime = get_systime,
+	.get_ptptime = dwmac1000_get_ptptime,
+	.timestamp_interrupt = dwmac1000_timestamp_interrupt,
+};
diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
index fa4611855311..4cc70480ce0f 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.h
@@ -96,8 +96,12 @@ enum aux_snapshot {
 
 struct ptp_clock_info;
 struct ptp_clock_request;
+struct stmmac_priv;
 
 int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
 			 struct ptp_clock_request *rq, int on);
 
+void dwmac1000_get_ptptime(void __iomem *ptpaddr, u64 *ptp_time);
+void dwmac1000_timestamp_interrupt(struct stmmac_priv *priv);
+
 #endif	/* __STMMAC_PTP_H__ */
-- 
2.47.0

[PATCH net-next v3 6/9] net: stmmac: Enable timestamping interrupt on dwmac1000

[Maxime Chevallier]

The default configuration for the interrupts on dwmac1000 have the
timestamping interrupt masked. Now that the timestamping has been
adapted to dwmac1000, enable the timestamping interrupt on these
platforms.

On dwmac1000, the external snapshot interrupt is configured through a
dedicated bit, that is set as reserved on other dwmac variants. The
timestaming interrupt is acknowledged by reading the
GMAC3_X_TIMESTAMP_STATUS register.

Make sure that this interrupt is enabled when snapshot is enabled, and
masked when disabled.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 .../net/ethernet/stmicro/stmmac/dwmac1000_core.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
index dbbd834f9fc8..37374f5a15c4 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac1000_core.c
@@ -595,6 +595,20 @@ void dwmac1000_timestamp_interrupt(struct stmmac_priv *priv)
 
 /* DWMAC 1000 ptp_clock_info ops */
 
+static void dwmac1000_timestamp_interrupt_cfg(struct stmmac_priv *priv, bool en)
+{
+	void __iomem *ioaddr = priv->ioaddr;
+
+	u32 intr_mask = readl(ioaddr + GMAC_INT_MASK);
+
+	if (en)
+		intr_mask &= ~GMAC_INT_DISABLE_TIMESTAMP;
+	else
+		intr_mask |= GMAC_INT_DISABLE_TIMESTAMP;
+
+	writel(intr_mask, ioaddr + GMAC_INT_MASK);
+}
+
 int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
 			 struct ptp_clock_request *rq, int on)
 {
@@ -628,6 +642,8 @@ int dwmac1000_ptp_enable(struct ptp_clock_info *ptp,
 		ret = readl_poll_timeout(ptpaddr + PTP_TCR, tcr_val,
 					 !(tcr_val & GMAC_PTP_TCR_ATSFC),
 					 10, 10000);
+
+		dwmac1000_timestamp_interrupt_cfg(priv, on);
 		break;
 
 	default:
-- 
2.47.0

[PATCH net-next v3 7/9] net: stmmac: Don't include dwmac4 definitions in stmmac_ptp

[Maxime Chevallier]

The stmmac_ptp code doesn't need the dwmac4 register definitions, remove
the inclusion.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
index 430905f591b2..429b2d357813 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_ptp.c
@@ -9,7 +9,6 @@
 *******************************************************************************/
 #include "stmmac.h"
 #include "stmmac_ptp.h"
-#include "dwmac4.h"
 
 /**
  * stmmac_adjust_freq
-- 
2.47.0

[PATCH net-next v3 8/9] net: stmmac: Configure only the relevant bits for timestamping setup

[Maxime Chevallier]

The PTP_TCR (Timestamp Control Register) is used to configure several
features related to packet timestamping.

On one hand, it configures the 1588 packet processing, to indicate what
types of frames should be timestamped (all, only 1588v1 or 1588v2, using
L2 or L4 timestamping, on IPv4 or IPv6, etc.). This is congfigured
usually through the ioctl / ndo dedicated for such setup. This
configuration is done by setting some fields in that register, that seem
to behave the same way on all dwmac variants, including DWMAC1000.

On the other hand, and only on DWMAC1000 apparently, some fields in that
register are used to configure external snapshots (bits 24/25).
On DWMAC4 and others, these fields are reserved and external
snapshots are configured through a dedicated register that simply
doesn't seem to exist on DWMAC1000.

This configuration is done in the dwmac1000-specific ptp_clock_info ops
(cf dwmac1000_ptp_enable()).

So to avoid the timestamping configuration interfering with the external
snapshots, this commit makes sure that the config_hw_tstamping only
configures the relevant bits in PTP_TCR, so that the DWMAC1000
timestamping can correctly rely on these otherwise reserved fields.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 .../net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
index a94829ef8cfb..0f59aa982604 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_hwtstamp.c
@@ -18,9 +18,22 @@
 #include "dwmac4.h"
 #include "stmmac.h"
 
+#define STMMAC_HWTS_CFG_MASK	(PTP_TCR_TSENA | PTP_TCR_TSCFUPDT | \
+				 PTP_TCR_TSINIT | PTP_TCR_TSUPDT | \
+				 PTP_TCR_TSCTRLSSR | PTP_TCR_SNAPTYPSEL_1 | \
+				 PTP_TCR_TSIPV4ENA | PTP_TCR_TSIPV6ENA | \
+				 PTP_TCR_TSEVNTENA | PTP_TCR_TSMSTRENA | \
+				 PTP_TCR_TSVER2ENA | PTP_TCR_TSIPENA | \
+				 PTP_TCR_TSTRIG | PTP_TCR_TSENALL)
+
 static void config_hw_tstamping(void __iomem *ioaddr, u32 data)
 {
-	writel(data, ioaddr + PTP_TCR);
+	u32 regval = readl(ioaddr + PTP_TCR);
+
+	regval &= ~STMMAC_HWTS_CFG_MASK;
+	regval |= data;
+
+	writel(regval, ioaddr + PTP_TCR);
 }
 
 static void config_sub_second_increment(void __iomem *ioaddr,
-- 
2.47.0

[PATCH net-next v3 9/9] net: stmmac: dwmac_socfpga: This platform has GMAC

[Maxime Chevallier]

Indicate that dwmac_socfpga has a gmac. This will make sure that
gmac-specific interrupt processing is done, including timestamp
interrupt handling. Without this, the external snapshot interrupt is
never ack'd and we have an interrupt storm on external snapshot event.

Signed-off-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
---
 drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c
index 0745117d5872..248b30d7b864 100644
--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-socfpga.c
@@ -485,6 +485,7 @@ static int socfpga_dwmac_probe(struct platform_device *pdev)
 	plat_dat->pcs_init = socfpga_dwmac_pcs_init;
 	plat_dat->pcs_exit = socfpga_dwmac_pcs_exit;
 	plat_dat->select_pcs = socfpga_dwmac_select_pcs;
+	plat_dat->has_gmac = true;
 
 	ret = stmmac_dvr_probe(&pdev->dev, plat_dat, &stmmac_res);
 	if (ret)
-- 
2.47.0

Re: [PATCH net-next v3 0/9] Support external snapshots on dwmac1000

[Daniel Machon]

> Hi,
> 
> This V3 is just a rebase a V2 on top of net-next to address some minor
> conflicts. No changes were made.
> 
> This series is another take on the pervious work [1] done by
> Alexis Lothoré, that fixes the support for external snapshots
> timestamping in GMAC3-based devices.
> 
> Details on why this is needed are mentionned on the cover [2] from V1.
> 
> This V2 addresses multiple issues found in V1 :
> 
>  - The PTP_TCR register is configured from multiple places, as reported
>    by Alexis, so we need to make sure that the extts configuration
>    doesn't interfere with the hwtstamp configuration.
> 
>  - The interrupt management in V1 was incomplete, as the interrupt
>    wasn't correctly acked.
> 
>  - This series also makes so that we only enable the extts interrupt
>    when necessary.
> 
> [1]: https://lore.kernel.org/netdev/20230616100409.164583-1-alexis.lothore@bootlin.com/
> [2]: https://lore.kernel.org/netdev/20241029115419.1160201-1-maxime.chevallier@bootlin.com/
> 
> Thanks Alexis for laying the groundwork for this,
> 
> Best regards,
> 
> Maxime
> 
> Link to V1: https://lore.kernel.org/netdev/20241029115419.1160201-1-maxime.chevallier@bootlin.com/
> Link to V2: https://lore.kernel.org/netdev/20241104170251.2202270-1-maxime.chevallier@bootlin.com/

Hi Maxime,

Dont know much about this particular driver, but the patches looked good
to me!

For the series:

Reviewed-by: Daniel Machon <daniel.machon@microchip.com>