ptp_qoriq: fix latency in ptp_qoriq_adjtime() operation

1588 driver loses about 1us in adjtime operation at PTP slave
This is because adjtime operation uses a slow non-atomic tmr_cnt_read()
followed by tmr_cnt_write() operation.

In the above sequence, since the timer counter operation keeps
incrementing, it leads to latency. The tmr_offset register
(which is added to TMR_CNT_H/L register giving the current time)
must be programmed with the delta nanoseconds.
Signed-off-by: Nikhil Gupta <nikhil.gupta@nxp.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20230119204034.7969-1-nikhil.gupta@nxp.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/ptp/ptp_qoriq.c

@@ -48,6 +48,29 @@ static void tmr_cnt_write(struct ptp_qoriq *ptp_qoriq, u64 ns)
 	ptp_qoriq->write(&regs->ctrl_regs->tmr_cnt_h, hi);
 }
 
+static u64 tmr_offset_read(struct ptp_qoriq *ptp_qoriq)
+{
+	struct ptp_qoriq_registers *regs = &ptp_qoriq->regs;
+	u32 lo, hi;
+	u64 ns;
+
+	lo = ptp_qoriq->read(&regs->ctrl_regs->tmroff_l);
+	hi = ptp_qoriq->read(&regs->ctrl_regs->tmroff_h);
+	ns = ((u64) hi) << 32;
+	ns |= lo;
+	return ns;
+}
+
+static void tmr_offset_write(struct ptp_qoriq *ptp_qoriq, u64 delta_ns)
+{
+	struct ptp_qoriq_registers *regs = &ptp_qoriq->regs;
+	u32 lo = delta_ns & 0xffffffff;
+	u32 hi = delta_ns >> 32;
+
+	ptp_qoriq->write(&regs->ctrl_regs->tmroff_l, lo);
+	ptp_qoriq->write(&regs->ctrl_regs->tmroff_h, hi);
+}
+
 /* Caller must hold ptp_qoriq->lock. */
 static void set_alarm(struct ptp_qoriq *ptp_qoriq)
 {
@@ -55,7 +78,9 @@ static void set_alarm(struct ptp_qoriq *ptp_qoriq)
 	u64 ns;
 	u32 lo, hi;
 
-	ns = tmr_cnt_read(ptp_qoriq) + 1500000000ULL;
+	ns = tmr_cnt_read(ptp_qoriq) + tmr_offset_read(ptp_qoriq)
+				     + 1500000000ULL;
+
 	ns = div_u64(ns, 1000000000UL) * 1000000000ULL;
 	ns -= ptp_qoriq->tclk_period;
 	hi = ns >> 32;
@@ -207,15 +232,24 @@ EXPORT_SYMBOL_GPL(ptp_qoriq_adjfine);
 
 int ptp_qoriq_adjtime(struct ptp_clock_info *ptp, s64 delta)
 {
-	s64 now;
-	unsigned long flags;
 	struct ptp_qoriq *ptp_qoriq = container_of(ptp, struct ptp_qoriq, caps);
+	s64 now, curr_delta;
+	unsigned long flags;
 
 	spin_lock_irqsave(&ptp_qoriq->lock, flags);
 
+	/* On LS1021A, eTSEC2 and eTSEC3 do not take into account the TMR_OFF
+	 * adjustment
+	 */
+	if (ptp_qoriq->etsec) {
 		now = tmr_cnt_read(ptp_qoriq);
 		now += delta;
 		tmr_cnt_write(ptp_qoriq, now);
+	} else {
+		curr_delta = tmr_offset_read(ptp_qoriq);
+		curr_delta += delta;
+		tmr_offset_write(ptp_qoriq, curr_delta);
+	}
 	set_fipers(ptp_qoriq);
 
 	spin_unlock_irqrestore(&ptp_qoriq->lock, flags);
@@ -232,7 +266,7 @@ int ptp_qoriq_gettime(struct ptp_clock_info *ptp, struct timespec64 *ts)
 
 	spin_lock_irqsave(&ptp_qoriq->lock, flags);
 
-	ns = tmr_cnt_read(ptp_qoriq);
+	ns = tmr_cnt_read(ptp_qoriq) + tmr_offset_read(ptp_qoriq);
 
 	spin_unlock_irqrestore(&ptp_qoriq->lock, flags);
 
@@ -253,6 +287,7 @@ int ptp_qoriq_settime(struct ptp_clock_info *ptp,
 
 	spin_lock_irqsave(&ptp_qoriq->lock, flags);
 
+	tmr_offset_write(ptp_qoriq, 0);
 	tmr_cnt_write(ptp_qoriq, ns);
 	set_fipers(ptp_qoriq);
 
@@ -488,6 +523,7 @@ int ptp_qoriq_init(struct ptp_qoriq *ptp_qoriq, void __iomem *base,
 
 	/* The eTSEC uses differnt memory map with DPAA/ENETC */
 	if (of_device_is_compatible(node, "fsl,etsec-ptp")) {
+		ptp_qoriq->etsec = true;
 		ptp_qoriq->regs.ctrl_regs = base + ETSEC_CTRL_REGS_OFFSET;
 		ptp_qoriq->regs.alarm_regs = base + ETSEC_ALARM_REGS_OFFSET;
 		ptp_qoriq->regs.fiper_regs = base + ETSEC_FIPER_REGS_OFFSET;

include/linux/fsl/ptp_qoriq.h

@@ -149,6 +149,7 @@ struct ptp_qoriq {
 	struct device *dev;
 	bool extts_fifo_support;
 	bool fiper3_support;
+	bool etsec;
 	int irq;
 	int phc_index;
 	u32 tclk_period;  /* nanoseconds */