ice_ptp.c

// SPDX-License-Identifier: GPL-2.0
/* Copyright (C) 2021, Intel Corporation. */

#include "ice.h"
#include "ice_lib.h"

/**
 * ice_set_tx_tstamp - Enable or disable Tx timestamping
 * @pf: The PF pointer to search in
 * @on: bool value for whether timestamps are enabled or disabled
 */
static void ice_set_tx_tstamp(struct ice_pf *pf, bool on)
{
	struct ice_vsi *vsi;
	u32 val;
	u16 i;

	vsi = ice_get_main_vsi(pf);
	if (!vsi)
		return;

	/* Set the timestamp enable flag for all the Tx rings */
	ice_for_each_rxq(vsi, i) {
		if (!vsi->tx_rings[i])
			continue;
		vsi->tx_rings[i]->ptp_tx = on;
	}

	/* Configure the Tx timestamp interrupt */
	val = rd32(&pf->hw, PFINT_OICR_ENA);
	if (on)
		val |= PFINT_OICR_TSYN_TX_M;
	else
		val &= ~PFINT_OICR_TSYN_TX_M;
	wr32(&pf->hw, PFINT_OICR_ENA, val);
}

/**
 * ice_set_rx_tstamp - Enable or disable Rx timestamping
 * @pf: The PF pointer to search in
 * @on: bool value for whether timestamps are enabled or disabled
 */
static void ice_set_rx_tstamp(struct ice_pf *pf, bool on)
{
	struct ice_vsi *vsi;
	u16 i;

	vsi = ice_get_main_vsi(pf);
	if (!vsi)
		return;

	/* Set the timestamp flag for all the Rx rings */
	ice_for_each_rxq(vsi, i) {
		if (!vsi->rx_rings[i])
			continue;
		vsi->rx_rings[i]->ptp_rx = on;
	}
}

/**
 * ice_ptp_cfg_timestamp - Configure timestamp for init/deinit
 * @pf: Board private structure
 * @ena: bool value to enable or disable time stamp
 *
 * This function will configure timestamping during PTP initialization
 * and deinitialization
 */
static void ice_ptp_cfg_timestamp(struct ice_pf *pf, bool ena)
{
	ice_set_tx_tstamp(pf, ena);
	ice_set_rx_tstamp(pf, ena);

	if (ena) {
		pf->ptp.tstamp_config.rx_filter = HWTSTAMP_FILTER_ALL;
		pf->ptp.tstamp_config.tx_type = HWTSTAMP_TX_ON;
	} else {
		pf->ptp.tstamp_config.rx_filter = HWTSTAMP_FILTER_NONE;
		pf->ptp.tstamp_config.tx_type = HWTSTAMP_TX_OFF;
	}
}

/**
 * ice_get_ptp_clock_index - Get the PTP clock index
 * @pf: the PF pointer
 *
 * Determine the clock index of the PTP clock associated with this device. If
 * this is the PF controlling the clock, just use the local access to the
 * clock device pointer.
 *
 * Otherwise, read from the driver shared parameters to determine the clock
 * index value.
 *
 * Returns: the index of the PTP clock associated with this device, or -1 if
 * there is no associated clock.
 */
int ice_get_ptp_clock_index(struct ice_pf *pf)
{
	struct device *dev = ice_pf_to_dev(pf);
	enum ice_aqc_driver_params param_idx;
	struct ice_hw *hw = &pf->hw;
	u8 tmr_idx;
	u32 value;
	int err;

	/* Use the ptp_clock structure if we're the main PF */
	if (pf->ptp.clock)
		return ptp_clock_index(pf->ptp.clock);

	tmr_idx = hw->func_caps.ts_func_info.tmr_index_assoc;
	if (!tmr_idx)
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR0;
	else
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR1;

	err = ice_aq_get_driver_param(hw, param_idx, &value, NULL);
	if (err) {
		dev_err(dev, "Failed to read PTP clock index parameter, err %d aq_err %s\n",
			err, ice_aq_str(hw->adminq.sq_last_status));
		return -1;
	}

	/* The PTP clock index is an integer, and will be between 0 and
	 * INT_MAX. The highest bit of the driver shared parameter is used to
	 * indicate whether or not the currently stored clock index is valid.
	 */
	if (!(value & PTP_SHARED_CLK_IDX_VALID))
		return -1;

	return value & ~PTP_SHARED_CLK_IDX_VALID;
}

/**
 * ice_set_ptp_clock_index - Set the PTP clock index
 * @pf: the PF pointer
 *
 * Set the PTP clock index for this device into the shared driver parameters,
 * so that other PFs associated with this device can read it.
 *
 * If the PF is unable to store the clock index, it will log an error, but
 * will continue operating PTP.
 */
static void ice_set_ptp_clock_index(struct ice_pf *pf)
{
	struct device *dev = ice_pf_to_dev(pf);
	enum ice_aqc_driver_params param_idx;
	struct ice_hw *hw = &pf->hw;
	u8 tmr_idx;
	u32 value;
	int err;

	if (!pf->ptp.clock)
		return;

	tmr_idx = hw->func_caps.ts_func_info.tmr_index_assoc;
	if (!tmr_idx)
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR0;
	else
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR1;

	value = (u32)ptp_clock_index(pf->ptp.clock);
	if (value > INT_MAX) {
		dev_err(dev, "PTP Clock index is too large to store\n");
		return;
	}
	value |= PTP_SHARED_CLK_IDX_VALID;

	err = ice_aq_set_driver_param(hw, param_idx, value, NULL);
	if (err) {
		dev_err(dev, "Failed to set PTP clock index parameter, err %d aq_err %s\n",
			err, ice_aq_str(hw->adminq.sq_last_status));
	}
}

/**
 * ice_clear_ptp_clock_index - Clear the PTP clock index
 * @pf: the PF pointer
 *
 * Clear the PTP clock index for this device. Must be called when
 * unregistering the PTP clock, in order to ensure other PFs stop reporting
 * a clock object that no longer exists.
 */
static void ice_clear_ptp_clock_index(struct ice_pf *pf)
{
	struct device *dev = ice_pf_to_dev(pf);
	enum ice_aqc_driver_params param_idx;
	struct ice_hw *hw = &pf->hw;
	u8 tmr_idx;
	int err;

	/* Do not clear the index if we don't own the timer */
	if (!hw->func_caps.ts_func_info.src_tmr_owned)
		return;

	tmr_idx = hw->func_caps.ts_func_info.tmr_index_assoc;
	if (!tmr_idx)
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR0;
	else
		param_idx = ICE_AQC_DRIVER_PARAM_CLK_IDX_TMR1;

	err = ice_aq_set_driver_param(hw, param_idx, 0, NULL);
	if (err) {
		dev_dbg(dev, "Failed to clear PTP clock index parameter, err %d aq_err %s\n",
			err, ice_aq_str(hw->adminq.sq_last_status));
	}
}

/**
 * ice_ptp_read_src_clk_reg - Read the source clock register
 * @pf: Board private structure
 * @sts: Optional parameter for holding a pair of system timestamps from
 *       the system clock. Will be ignored if NULL is given.
 */
static u64
ice_ptp_read_src_clk_reg(struct ice_pf *pf, struct ptp_system_timestamp *sts)
{
	struct ice_hw *hw = &pf->hw;
	u32 hi, lo, lo2;
	u8 tmr_idx;

	tmr_idx = ice_get_ptp_src_clock_index(hw);
	/* Read the system timestamp pre PHC read */
	ptp_read_system_prets(sts);

	lo = rd32(hw, GLTSYN_TIME_L(tmr_idx));

	/* Read the system timestamp post PHC read */
	ptp_read_system_postts(sts);

	hi = rd32(hw, GLTSYN_TIME_H(tmr_idx));
	lo2 = rd32(hw, GLTSYN_TIME_L(tmr_idx));

	if (lo2 < lo) {
		/* if TIME_L rolled over read TIME_L again and update
		 * system timestamps
		 */
		ptp_read_system_prets(sts);
		lo = rd32(hw, GLTSYN_TIME_L(tmr_idx));
		ptp_read_system_postts(sts);
		hi = rd32(hw, GLTSYN_TIME_H(tmr_idx));
	}

	return ((u64)hi << 32) | lo;
}

/**
 * ice_ptp_update_cached_phctime - Update the cached PHC time values
 * @pf: Board specific private structure
 *
 * This function updates the system time values which are cached in the PF
 * structure and the Rx rings.
 *
 * This function must be called periodically to ensure that the cached value
 * is never more than 2 seconds old. It must also be called whenever the PHC
 * time has been changed.
 */
static void ice_ptp_update_cached_phctime(struct ice_pf *pf)
{
	u64 systime;
	int i;

	/* Read the current PHC time */
	systime = ice_ptp_read_src_clk_reg(pf, NULL);

	/* Update the cached PHC time stored in the PF structure */
	WRITE_ONCE(pf->ptp.cached_phc_time, systime);

	ice_for_each_vsi(pf, i) {
		struct ice_vsi *vsi = pf->vsi[i];
		int j;

		if (!vsi)
			continue;

		if (vsi->type != ICE_VSI_PF)
			continue;

		ice_for_each_rxq(vsi, j) {
			if (!vsi->rx_rings[j])
				continue;
			WRITE_ONCE(vsi->rx_rings[j]->cached_phctime, systime);
		}
	}
}

/**
 * ice_ptp_extend_32b_ts - Convert a 32b nanoseconds timestamp to 64b
 * @cached_phc_time: recently cached copy of PHC time
 * @in_tstamp: Ingress/egress 32b nanoseconds timestamp value
 *
 * Hardware captures timestamps which contain only 32 bits of nominal
 * nanoseconds, as opposed to the 64bit timestamps that the stack expects.
 * Note that the captured timestamp values may be 40 bits, but the lower
 * 8 bits are sub-nanoseconds and generally discarded.
 *
 * Extend the 32bit nanosecond timestamp using the following algorithm and
 * assumptions:
 *
 * 1) have a recently cached copy of the PHC time
 * 2) assume that the in_tstamp was captured 2^31 nanoseconds (~2.1
 *    seconds) before or after the PHC time was captured.
 * 3) calculate the delta between the cached time and the timestamp
 * 4) if the delta is smaller than 2^31 nanoseconds, then the timestamp was
 *    captured after the PHC time. In this case, the full timestamp is just
 *    the cached PHC time plus the delta.
 * 5) otherwise, if the delta is larger than 2^31 nanoseconds, then the
 *    timestamp was captured *before* the PHC time, i.e. because the PHC
 *    cache was updated after the timestamp was captured by hardware. In this
 *    case, the full timestamp is the cached time minus the inverse delta.
 *
 * This algorithm works even if the PHC time was updated after a Tx timestamp
 * was requested, but before the Tx timestamp event was reported from
 * hardware.
 *
 * This calculation primarily relies on keeping the cached PHC time up to
 * date. If the timestamp was captured more than 2^31 nanoseconds after the
 * PHC time, it is possible that the lower 32bits of PHC time have
 * overflowed more than once, and we might generate an incorrect timestamp.
 *
 * This is prevented by (a) periodically updating the cached PHC time once
 * a second, and (b) discarding any Tx timestamp packet if it has waited for
 * a timestamp for more than one second.
 */
static u64 ice_ptp_extend_32b_ts(u64 cached_phc_time, u32 in_tstamp)
{
	u32 delta, phc_time_lo;
	u64 ns;

	/* Extract the lower 32 bits of the PHC time */
	phc_time_lo = (u32)cached_phc_time;

	/* Calculate the delta between the lower 32bits of the cached PHC
	 * time and the in_tstamp value
	 */
	delta = (in_tstamp - phc_time_lo);

	/* Do not assume that the in_tstamp is always more recent than the
	 * cached PHC time. If the delta is large, it indicates that the
	 * in_tstamp was taken in the past, and should be converted
	 * forward.
	 */
	if (delta > (U32_MAX / 2)) {
		/* reverse the delta calculation here */
		delta = (phc_time_lo - in_tstamp);
		ns = cached_phc_time - delta;
	} else {
		ns = cached_phc_time + delta;
	}

	return ns;
}

/**
 * ice_ptp_extend_40b_ts - Convert a 40b timestamp to 64b nanoseconds
 * @pf: Board private structure
 * @in_tstamp: Ingress/egress 40b timestamp value
 *
 * The Tx and Rx timestamps are 40 bits wide, including 32 bits of nominal
 * nanoseconds, 7 bits of sub-nanoseconds, and a valid bit.
 *
 *  *--------------------------------------------------------------*
 *  | 32 bits of nanoseconds | 7 high bits of sub ns underflow | v |
 *  *--------------------------------------------------------------*
 *
 * The low bit is an indicator of whether the timestamp is valid. The next
 * 7 bits are a capture of the upper 7 bits of the sub-nanosecond underflow,
 * and the remaining 32 bits are the lower 32 bits of the PHC timer.
 *
 * It is assumed that the caller verifies the timestamp is valid prior to
 * calling this function.
 *
 * Extract the 32bit nominal nanoseconds and extend them. Use the cached PHC
 * time stored in the device private PTP structure as the basis for timestamp
 * extension.
 *
 * See ice_ptp_extend_32b_ts for a detailed explanation of the extension
 * algorithm.
 */
static u64 ice_ptp_extend_40b_ts(struct ice_pf *pf, u64 in_tstamp)
{
	const u64 mask = GENMASK_ULL(31, 0);

	return ice_ptp_extend_32b_ts(pf->ptp.cached_phc_time,
				     (in_tstamp >> 8) & mask);
}

/**
 * ice_ptp_read_time - Read the time from the device
 * @pf: Board private structure
 * @ts: timespec structure to hold the current time value
 * @sts: Optional parameter for holding a pair of system timestamps from
 *       the system clock. Will be ignored if NULL is given.
 *
 * This function reads the source clock registers and stores them in a timespec.
 * However, since the registers are 64 bits of nanoseconds, we must convert the
 * result to a timespec before we can return.
 */
static void
ice_ptp_read_time(struct ice_pf *pf, struct timespec64 *ts,
		  struct ptp_system_timestamp *sts)
{
	u64 time_ns = ice_ptp_read_src_clk_reg(pf, sts);

	*ts = ns_to_timespec64(time_ns);
}

/**
 * ice_ptp_write_init - Set PHC time to provided value
 * @pf: Board private structure
 * @ts: timespec structure that holds the new time value
 *
 * Set the PHC time to the specified time provided in the timespec.
 */
static int ice_ptp_write_init(struct ice_pf *pf, struct timespec64 *ts)
{
	u64 ns = timespec64_to_ns(ts);
	struct ice_hw *hw = &pf->hw;

	return ice_ptp_init_time(hw, ns);
}

/**
 * ice_ptp_write_adj - Adjust PHC clock time atomically
 * @pf: Board private structure
 * @adj: Adjustment in nanoseconds
 *
 * Perform an atomic adjustment of the PHC time by the specified number of
 * nanoseconds.
 */
static int ice_ptp_write_adj(struct ice_pf *pf, s32 adj)
{
	struct ice_hw *hw = &pf->hw;

	return ice_ptp_adj_clock(hw, adj);
}

/**
 * ice_ptp_adjfine - Adjust clock increment rate
 * @info: the driver's PTP info structure
 * @scaled_ppm: Parts per million with 16-bit fractional field
 *
 * Adjust the frequency of the clock by the indicated scaled ppm from the
 * base frequency.
 */
static int ice_ptp_adjfine(struct ptp_clock_info *info, long scaled_ppm)
{
	struct ice_pf *pf = ptp_info_to_pf(info);
	u64 freq, divisor = 1000000ULL;
	struct ice_hw *hw = &pf->hw;
	s64 incval, diff;
	int neg_adj = 0;
	int err;

	incval = ICE_PTP_NOMINAL_INCVAL_E810;

	if (scaled_ppm < 0) {
		neg_adj = 1;
		scaled_ppm = -scaled_ppm;
	}

	while ((u64)scaled_ppm > div_u64(U64_MAX, incval)) {
		/* handle overflow by scaling down the scaled_ppm and
		 * the divisor, losing some precision
		 */
		scaled_ppm >>= 2;
		divisor >>= 2;
	}

	freq = (incval * (u64)scaled_ppm) >> 16;
	diff = div_u64(freq, divisor);

	if (neg_adj)
		incval -= diff;
	else
		incval += diff;

	err = ice_ptp_write_incval_locked(hw, incval);
	if (err) {
		dev_err(ice_pf_to_dev(pf), "PTP failed to set incval, err %d\n",
			err);
		return -EIO;
	}

	return 0;
}

/**
 * ice_ptp_gettimex64 - Get the time of the clock
 * @info: the driver's PTP info structure
 * @ts: timespec64 structure to hold the current time value
 * @sts: Optional parameter for holding a pair of system timestamps from
 *       the system clock. Will be ignored if NULL is given.
 *
 * Read the device clock and return the correct value on ns, after converting it
 * into a timespec struct.
 */
static int
ice_ptp_gettimex64(struct ptp_clock_info *info, struct timespec64 *ts,
		   struct ptp_system_timestamp *sts)
{
	struct ice_pf *pf = ptp_info_to_pf(info);
	struct ice_hw *hw = &pf->hw;

	if (!ice_ptp_lock(hw)) {
		dev_err(ice_pf_to_dev(pf), "PTP failed to get time\n");
		return -EBUSY;
	}

	ice_ptp_read_time(pf, ts, sts);
	ice_ptp_unlock(hw);

	return 0;
}

/**
 * ice_ptp_settime64 - Set the time of the clock
 * @info: the driver's PTP info structure
 * @ts: timespec64 structure that holds the new time value
 *
 * Set the device clock to the user input value. The conversion from timespec
 * to ns happens in the write function.
 */
static int
ice_ptp_settime64(struct ptp_clock_info *info, const struct timespec64 *ts)
{
	struct ice_pf *pf = ptp_info_to_pf(info);
	struct timespec64 ts64 = *ts;
	struct ice_hw *hw = &pf->hw;
	int err;

	if (!ice_ptp_lock(hw)) {
		err = -EBUSY;
		goto exit;
	}

	err = ice_ptp_write_init(pf, &ts64);
	ice_ptp_unlock(hw);

	if (!err)
		ice_ptp_update_cached_phctime(pf);

exit:
	if (err) {
		dev_err(ice_pf_to_dev(pf), "PTP failed to set time %d\n", err);
		return err;
	}

	return 0;
}

/**
 * ice_ptp_adjtime_nonatomic - Do a non-atomic clock adjustment
 * @info: the driver's PTP info structure
 * @delta: Offset in nanoseconds to adjust the time by
 */
static int ice_ptp_adjtime_nonatomic(struct ptp_clock_info *info, s64 delta)
{
	struct timespec64 now, then;

	then = ns_to_timespec64(delta);
	ice_ptp_gettimex64(info, &now, NULL);
	now = timespec64_add(now, then);

	return ice_ptp_settime64(info, (const struct timespec64 *)&now);
}

/**
 * ice_ptp_adjtime - Adjust the time of the clock by the indicated delta
 * @info: the driver's PTP info structure
 * @delta: Offset in nanoseconds to adjust the time by
 */
static int ice_ptp_adjtime(struct ptp_clock_info *info, s64 delta)
{
	struct ice_pf *pf = ptp_info_to_pf(info);
	struct ice_hw *hw = &pf->hw;
	struct device *dev;
	int err;

	dev = ice_pf_to_dev(pf);

	/* Hardware only supports atomic adjustments using signed 32-bit
	 * integers. For any adjustment outside this range, perform
	 * a non-atomic get->adjust->set flow.
	 */
	if (delta > S32_MAX || delta < S32_MIN) {
		dev_dbg(dev, "delta = %lld, adjtime non-atomic\n", delta);
		return ice_ptp_adjtime_nonatomic(info, delta);
	}

	if (!ice_ptp_lock(hw)) {
		dev_err(dev, "PTP failed to acquire semaphore in adjtime\n");
		return -EBUSY;
	}

	err = ice_ptp_write_adj(pf, delta);

	ice_ptp_unlock(hw);

	if (err) {
		dev_err(dev, "PTP failed to adjust time, err %d\n", err);
		return err;
	}

	ice_ptp_update_cached_phctime(pf);

	return 0;
}

/**
 * ice_ptp_get_ts_config - ioctl interface to read the timestamping config
 * @pf: Board private structure
 * @ifr: ioctl data
 *
 * Copy the timestamping config to user buffer
 */
int ice_ptp_get_ts_config(struct ice_pf *pf, struct ifreq *ifr)
{
	struct hwtstamp_config *config;

	if (!test_bit(ICE_FLAG_PTP, pf->flags))
		return -EIO;

	config = &pf->ptp.tstamp_config;

	return copy_to_user(ifr->ifr_data, config, sizeof(*config)) ?
		-EFAULT : 0;
}

/**
 * ice_ptp_set_timestamp_mode - Setup driver for requested timestamp mode
 * @pf: Board private structure
 * @config: hwtstamp settings requested or saved
 */
static int
ice_ptp_set_timestamp_mode(struct ice_pf *pf, struct hwtstamp_config *config)
{
	/* Reserved for future extensions. */
	if (config->flags)
		return -EINVAL;

	switch (config->tx_type) {
	case HWTSTAMP_TX_OFF:
		ice_set_tx_tstamp(pf, false);
		break;
	case HWTSTAMP_TX_ON:
		ice_set_tx_tstamp(pf, true);
		break;
	default:
		return -ERANGE;
	}

	switch (config->rx_filter) {
	case HWTSTAMP_FILTER_NONE:
		ice_set_rx_tstamp(pf, false);
		break;
	case HWTSTAMP_FILTER_PTP_V1_L4_EVENT:
	case HWTSTAMP_FILTER_PTP_V1_L4_SYNC:
	case HWTSTAMP_FILTER_PTP_V1_L4_DELAY_REQ:
	case HWTSTAMP_FILTER_PTP_V2_EVENT:
	case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
	case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
	case HWTSTAMP_FILTER_PTP_V2_SYNC:
	case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
	case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
	case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
	case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
	case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
	case HWTSTAMP_FILTER_NTP_ALL:
	case HWTSTAMP_FILTER_ALL:
		config->rx_filter = HWTSTAMP_FILTER_ALL;
		ice_set_rx_tstamp(pf, true);
		break;
	default:
		return -ERANGE;
	}

	return 0;
}

/**
 * ice_ptp_set_ts_config - ioctl interface to control the timestamping
 * @pf: Board private structure
 * @ifr: ioctl data
 *
 * Get the user config and store it
 */
int ice_ptp_set_ts_config(struct ice_pf *pf, struct ifreq *ifr)
{
	struct hwtstamp_config config;
	int err;

	if (!test_bit(ICE_FLAG_PTP, pf->flags))
		return -EAGAIN;

	if (copy_from_user(&config, ifr->ifr_data, sizeof(config)))
		return -EFAULT;

	err = ice_ptp_set_timestamp_mode(pf, &config);
	if (err)
		return err;

	/* Save these settings for future reference */
	pf->ptp.tstamp_config = config;

	return copy_to_user(ifr->ifr_data, &config, sizeof(config)) ?
		-EFAULT : 0;
}

/**
 * ice_ptp_rx_hwtstamp - Check for an Rx timestamp
 * @rx_ring: Ring to get the VSI info
 * @rx_desc: Receive descriptor
 * @skb: Particular skb to send timestamp with
 *
 * The driver receives a notification in the receive descriptor with timestamp.
 * The timestamp is in ns, so we must convert the result first.
 */
void
ice_ptp_rx_hwtstamp(struct ice_ring *rx_ring,
		    union ice_32b_rx_flex_desc *rx_desc, struct sk_buff *skb)
{
	u32 ts_high;
	u64 ts_ns;

	/* Populate timesync data into skb */
	if (rx_desc->wb.time_stamp_low & ICE_PTP_TS_VALID) {
		struct skb_shared_hwtstamps *hwtstamps;

		/* Use ice_ptp_extend_32b_ts directly, using the ring-specific
		 * cached PHC value, rather than accessing the PF. This also
		 * allows us to simply pass the upper 32bits of nanoseconds
		 * directly. Calling ice_ptp_extend_40b_ts is unnecessary as
		 * it would just discard these bits itself.
		 */
		ts_high = le32_to_cpu(rx_desc->wb.flex_ts.ts_high);
		ts_ns = ice_ptp_extend_32b_ts(rx_ring->cached_phctime, ts_high);

		hwtstamps = skb_hwtstamps(skb);
		memset(hwtstamps, 0, sizeof(*hwtstamps));
		hwtstamps->hwtstamp = ns_to_ktime(ts_ns);
	}
}

/**
 * ice_ptp_set_caps - Set PTP capabilities
 * @pf: Board private structure
 */
static void ice_ptp_set_caps(struct ice_pf *pf)
{
	struct ptp_clock_info *info = &pf->ptp.info;
	struct device *dev = ice_pf_to_dev(pf);

	snprintf(info->name, sizeof(info->name) - 1, "%s-%s-clk",
		 dev_driver_string(dev), dev_name(dev));
	info->owner = THIS_MODULE;
	info->max_adj = 999999999;
	info->adjtime = ice_ptp_adjtime;
	info->adjfine = ice_ptp_adjfine;
	info->gettimex64 = ice_ptp_gettimex64;
	info->settime64 = ice_ptp_settime64;
}

/**
 * ice_ptp_create_clock - Create PTP clock device for userspace
 * @pf: Board private structure
 *
 * This function creates a new PTP clock device. It only creates one if we
 * don't already have one. Will return error if it can't create one, but success
 * if we already have a device. Should be used by ice_ptp_init to create clock
 * initially, and prevent global resets from creating new clock devices.
 */
static long ice_ptp_create_clock(struct ice_pf *pf)
{
	struct ptp_clock_info *info;
	struct ptp_clock *clock;
	struct device *dev;

	/* No need to create a clock device if we already have one */
	if (pf->ptp.clock)
		return 0;

	ice_ptp_set_caps(pf);

	info = &pf->ptp.info;
	dev = ice_pf_to_dev(pf);

	/* Attempt to register the clock before enabling the hardware. */
	clock = ptp_clock_register(info, dev);
	if (IS_ERR(clock))
		return PTR_ERR(clock);

	pf->ptp.clock = clock;

	return 0;
}

/**
 * ice_ptp_tx_tstamp_work - Process Tx timestamps for a port
 * @work: pointer to the kthread_work struct
 *
 * Process timestamps captured by the PHY associated with this port. To do
 * this, loop over each index with a waiting skb.
 *
 * If a given index has a valid timestamp, perform the following steps:
 *
 * 1) copy the timestamp out of the PHY register
 * 4) clear the timestamp valid bit in the PHY register
 * 5) unlock the index by clearing the associated in_use bit.
 * 2) extend the 40b timestamp value to get a 64bit timestamp
 * 3) send that timestamp to the stack
 *
 * After looping, if we still have waiting SKBs, then re-queue the work. This
 * may cause us effectively poll even when not strictly necessary. We do this
 * because it's possible a new timestamp was requested around the same time as
 * the interrupt. In some cases hardware might not interrupt us again when the
 * timestamp is captured.
 *
 * Note that we only take the tracking lock when clearing the bit and when
 * checking if we need to re-queue this task. The only place where bits can be
 * set is the hard xmit routine where an SKB has a request flag set. The only
 * places where we clear bits are this work function, or the periodic cleanup
 * thread. If the cleanup thread clears a bit we're processing we catch it
 * when we lock to clear the bit and then grab the SKB pointer. If a Tx thread
 * starts a new timestamp, we might not begin processing it right away but we
 * will notice it at the end when we re-queue the work item. If a Tx thread
 * starts a new timestamp just after this function exits without re-queuing,
 * the interrupt when the timestamp finishes should trigger. Avoiding holding
 * the lock for the entire function is important in order to ensure that Tx
 * threads do not get blocked while waiting for the lock.
 */
static void ice_ptp_tx_tstamp_work(struct kthread_work *work)
{
	struct ice_ptp_port *ptp_port;
	struct ice_ptp_tx *tx;
	struct ice_pf *pf;
	struct ice_hw *hw;
	u8 idx;

	tx = container_of(work, struct ice_ptp_tx, work);
	if (!tx->init)
		return;

	ptp_port = container_of(tx, struct ice_ptp_port, tx);
	pf = ptp_port_to_pf(ptp_port);
	hw = &pf->hw;

	for_each_set_bit(idx, tx->in_use, tx->len) {
		struct skb_shared_hwtstamps shhwtstamps = {};
		u8 phy_idx = idx + tx->quad_offset;
		u64 raw_tstamp, tstamp;
		struct sk_buff *skb;
		int err;

		err = ice_read_phy_tstamp(hw, tx->quad, phy_idx,
					  &raw_tstamp);
		if (err)
			continue;

		/* Check if the timestamp is valid */
		if (!(raw_tstamp & ICE_PTP_TS_VALID))
			continue;

		/* clear the timestamp register, so that it won't show valid
		 * again when re-used.
		 */
		ice_clear_phy_tstamp(hw, tx->quad, phy_idx);

		/* The timestamp is valid, so we'll go ahead and clear this
		 * index and then send the timestamp up to the stack.
		 */
		spin_lock(&tx->lock);
		clear_bit(idx, tx->in_use);
		skb = tx->tstamps[idx].skb;
		tx->tstamps[idx].skb = NULL;
		spin_unlock(&tx->lock);

		/* it's (unlikely but) possible we raced with the cleanup
		 * thread for discarding old timestamp requests.
		 */
		if (!skb)
			continue;

		/* Extend the timestamp using cached PHC time */
		tstamp = ice_ptp_extend_40b_ts(pf, raw_tstamp);
		shhwtstamps.hwtstamp = ns_to_ktime(tstamp);

		skb_tstamp_tx(skb, &shhwtstamps);
		dev_kfree_skb_any(skb);
	}

	/* Check if we still have work to do. If so, re-queue this task to
	 * poll for remaining timestamps.
	 */
	spin_lock(&tx->lock);
	if (!bitmap_empty(tx->in_use, tx->len))
		kthread_queue_work(pf->ptp.kworker, &tx->work);
	spin_unlock(&tx->lock);
}

/**
 * ice_ptp_request_ts - Request an available Tx timestamp index
 * @tx: the PTP Tx timestamp tracker to request from
 * @skb: the SKB to associate with this timestamp request
 */
s8 ice_ptp_request_ts(struct ice_ptp_tx *tx, struct sk_buff *skb)
{
	u8 idx;

	/* Check if this tracker is initialized */
	if (!tx->init)
		return -1;

	spin_lock(&tx->lock);
	/* Find and set the first available index */
	idx = find_first_zero_bit(tx->in_use, tx->len);
	if (idx < tx->len) {
		/* We got a valid index that no other thread could have set. Store
		 * a reference to the skb and the start time to allow discarding old
		 * requests.
		 */
		set_bit(idx, tx->in_use);
		tx->tstamps[idx].start = jiffies;
		tx->tstamps[idx].skb = skb_get(skb);
		skb_shinfo(skb)->tx_flags |= SKBTX_IN_PROGRESS;
	}

	spin_unlock(&tx->lock);

	/* return the appropriate PHY timestamp register index, -1 if no
	 * indexes were available.
	 */
	if (idx >= tx->len)
		return -1;
	else
		return idx + tx->quad_offset;
}

/**
 * ice_ptp_process_ts - Spawn kthread work to handle timestamps
 * @pf: Board private structure
 *
 * Queue work required to process the PTP Tx timestamps outside of interrupt
 * context.
 */
void ice_ptp_process_ts(struct ice_pf *pf)
{
	if (pf->ptp.port.tx.init)
		kthread_queue_work(pf->ptp.kworker, &pf->ptp.port.tx.work);
}

/**
 * ice_ptp_alloc_tx_tracker - Initialize tracking for Tx timestamps
 * @tx: Tx tracking structure to initialize
 *
 * Assumes that the length has already been initialized. Do not call directly,
 * use the ice_ptp_init_tx_e822 or ice_ptp_init_tx_e810 instead.
 */
static int
ice_ptp_alloc_tx_tracker(struct ice_ptp_tx *tx)
{
	tx->tstamps = kcalloc(tx->len, sizeof(*tx->tstamps), GFP_KERNEL);
	if (!tx->tstamps)
		return -ENOMEM;

	tx->in_use = bitmap_zalloc(tx->len, GFP_KERNEL);
	if (!tx->in_use) {
		kfree(tx->tstamps);
		tx->tstamps = NULL;
		return -ENOMEM;
	}

	spin_lock_init(&tx->lock);
	kthread_init_work(&tx->work, ice_ptp_tx_tstamp_work);

	tx->init = 1;

	return 0;
}

/**
 * ice_ptp_flush_tx_tracker - Flush any remaining timestamps from the tracker
 * @pf: Board private structure
 * @tx: the tracker to flush
 */
static void
ice_ptp_flush_tx_tracker(struct ice_pf *pf, struct ice_ptp_tx *tx)
{
	u8 idx;

	for (idx = 0; idx < tx->len; idx++) {
		u8 phy_idx = idx + tx->quad_offset;

		/* Clear any potential residual timestamp in the PHY block */
		if (!pf->hw.reset_ongoing)
			ice_clear_phy_tstamp(&pf->hw, tx->quad, phy_idx);

		if (tx->tstamps[idx].skb) {
			dev_kfree_skb_any(tx->tstamps[idx].skb);
			tx->tstamps[idx].skb = NULL;
		}
	}
}

/**
 * ice_ptp_release_tx_tracker - Release allocated memory for Tx tracker
 * @pf: Board private structure
 * @tx: Tx tracking structure to release
 *
 * Free memory associated with the Tx timestamp tracker.
 */
static void
ice_ptp_release_tx_tracker(struct ice_pf *pf, struct ice_ptp_tx *tx)
{
	tx->init = 0;

	kthread_cancel_work_sync(&tx->work);

	ice_ptp_flush_tx_tracker(pf, tx);

	kfree(tx->tstamps);
	tx->tstamps = NULL;

	kfree(tx->in_use);
	tx->in_use = NULL;

	tx->len = 0;
}

/**
 * ice_ptp_init_tx_e810 - Initialize tracking for Tx timestamps
 * @pf: Board private structure
 * @tx: the Tx tracking structure to initialize
 *
 * Initialize the Tx timestamp tracker for this PF. For E810 devices, each
 * port has its own block of timestamps, independent of the other ports.
 */
static int
ice_ptp_init_tx_e810(struct ice_pf *pf, struct ice_ptp_tx *tx)
{
	tx->quad = pf->hw.port_info->lport;
	tx->quad_offset = 0;
	tx->len = INDEX_PER_QUAD;

	return ice_ptp_alloc_tx_tracker(tx);
}

/**
 * ice_ptp_tx_tstamp_cleanup - Cleanup old timestamp requests that got dropped
 * @tx: PTP Tx tracker to clean up
 *
 * Loop through the Tx timestamp requests and see if any of them have been
 * waiting for a long time. Discard any SKBs that have been waiting for more
 * than 2 seconds. This is long enough to be reasonably sure that the
 * timestamp will never be captured. This might happen if the packet gets
 * discarded before it reaches the PHY timestamping block.
 */
static void ice_ptp_tx_tstamp_cleanup(struct ice_ptp_tx *tx)
{
	u8 idx;

	if (!tx->init)
		return;

	for_each_set_bit(idx, tx->in_use, tx->len) {
		struct sk_buff *skb;

		/* Check if this SKB has been waiting for too long */
		if (time_is_after_jiffies(tx->tstamps[idx].start + 2 * HZ))
			continue;

		spin_lock(&tx->lock);
		skb = tx->tstamps[idx].skb;
		tx->tstamps[idx].skb = NULL;
		clear_bit(idx, tx->in_use);
		spin_unlock(&tx->lock);

		/* Free the SKB after we've cleared the bit */
		dev_kfree_skb_any(skb);
	}
}

static void ice_ptp_periodic_work(struct kthread_work *work)
{
	struct ice_ptp *ptp = container_of(work, struct ice_ptp, work.work);
	struct ice_pf *pf = container_of(ptp, struct ice_pf, ptp);

	if (!test_bit(ICE_FLAG_PTP, pf->flags))
		return;

	ice_ptp_update_cached_phctime(pf);

	ice_ptp_tx_tstamp_cleanup(&pf->ptp.port.tx);

	/* Run twice a second */
	kthread_queue_delayed_work(ptp->kworker, &ptp->work,
				   msecs_to_jiffies(500));
}

/**
 * ice_ptp_init_owner - Initialize PTP_1588_CLOCK device
 * @pf: Board private structure
 *
 * Setup and initialize a PTP clock device that represents the device hardware
 * clock. Save the clock index for other functions connected to the same
 * hardware resource.
 */
static int ice_ptp_init_owner(struct ice_pf *pf)
{
	struct device *dev = ice_pf_to_dev(pf);
	struct ice_hw *hw = &pf->hw;
	struct timespec64 ts;
	u8 src_idx;
	int err;

	wr32(hw, GLTSYN_SYNC_DLAY, 0);

	/* Clear some HW residue and enable source clock */
	src_idx = hw->func_caps.ts_func_info.tmr_index_owned;

	/* Enable source clocks */
	wr32(hw, GLTSYN_ENA(src_idx), GLTSYN_ENA_TSYN_ENA_M);

	/* Enable PHY time sync */
	err = ice_ptp_init_phy_e810(hw);
	if (err)
		goto err_exit;

	/* Clear event status indications for auxiliary pins */
	(void)rd32(hw, GLTSYN_STAT(src_idx));

	/* Acquire the global hardware lock */
	if (!ice_ptp_lock(hw)) {
		err = -EBUSY;
		goto err_exit;
	}

	/* Write the increment time value to PHY and LAN */
	err = ice_ptp_write_incval(hw, ICE_PTP_NOMINAL_INCVAL_E810);
	if (err) {
		ice_ptp_unlock(hw);
		goto err_exit;
	}

	ts = ktime_to_timespec64(ktime_get_real());
	/* Write the initial Time value to PHY and LAN */
	err = ice_ptp_write_init(pf, &ts);
	if (err) {
		ice_ptp_unlock(hw);
		goto err_exit;
	}

	/* Release the global hardware lock */
	ice_ptp_unlock(hw);

	/* Ensure we have a clock device */
	err = ice_ptp_create_clock(pf);
	if (err)
		goto err_clk;

	/* Store the PTP clock index for other PFs */
	ice_set_ptp_clock_index(pf);

	return 0;

err_clk:
	pf->ptp.clock = NULL;
err_exit:
	dev_err(dev, "PTP failed to register clock, err %d\n", err);

	return err;
}

/**
 * ice_ptp_init - Initialize the PTP support after device probe or reset
 * @pf: Board private structure
 *
 * This function sets device up for PTP support. The first time it is run, it
 * will create a clock device. It does not create a clock device if one
 * already exists. It also reconfigures the device after a reset.
 */
void ice_ptp_init(struct ice_pf *pf)
{
	struct device *dev = ice_pf_to_dev(pf);
	struct kthread_worker *kworker;
	struct ice_hw *hw = &pf->hw;
	int err;

	/* PTP is currently only supported on E810 devices */
	if (!ice_is_e810(hw))
		return;

	/* Check if this PF owns the source timer */
	if (hw->func_caps.ts_func_info.src_tmr_owned) {
		err = ice_ptp_init_owner(pf);
		if (err)
			return;
	}

	/* Disable timestamping for both Tx and Rx */
	ice_ptp_cfg_timestamp(pf, false);

	/* Initialize the PTP port Tx timestamp tracker */
	ice_ptp_init_tx_e810(pf, &pf->ptp.port.tx);

	/* Initialize work functions */
	kthread_init_delayed_work(&pf->ptp.work, ice_ptp_periodic_work);

	/* Allocate a kworker for handling work required for the ports
	 * connected to the PTP hardware clock.
	 */
	kworker = kthread_create_worker(0, "ice-ptp-%s", dev_name(dev));
	if (IS_ERR(kworker)) {
		err = PTR_ERR(kworker);
		goto err_kworker;
	}
	pf->ptp.kworker = kworker;

	set_bit(ICE_FLAG_PTP, pf->flags);

	/* Start periodic work going */
	kthread_queue_delayed_work(pf->ptp.kworker, &pf->ptp.work, 0);

	dev_info(dev, "PTP init successful\n");
	return;

err_kworker:
	/* If we registered a PTP clock, release it */
	if (pf->ptp.clock) {
		ptp_clock_unregister(pf->ptp.clock);
		pf->ptp.clock = NULL;
	}
	dev_err(dev, "PTP failed %d\n", err);
}

/**
 * ice_ptp_release - Disable the driver/HW support and unregister the clock
 * @pf: Board private structure
 *
 * This function handles the cleanup work required from the initialization by
 * clearing out the important information and unregistering the clock
 */
void ice_ptp_release(struct ice_pf *pf)
{
	/* Disable timestamping for both Tx and Rx */
	ice_ptp_cfg_timestamp(pf, false);

	ice_ptp_release_tx_tracker(pf, &pf->ptp.port.tx);

	clear_bit(ICE_FLAG_PTP, pf->flags);

	kthread_cancel_delayed_work_sync(&pf->ptp.work);

	if (pf->ptp.kworker) {
		kthread_destroy_worker(pf->ptp.kworker);
		pf->ptp.kworker = NULL;
	}

	if (!pf->ptp.clock)
		return;

	ice_clear_ptp_clock_index(pf);
	ptp_clock_unregister(pf->ptp.clock);
	pf->ptp.clock = NULL;

	dev_info(ice_pf_to_dev(pf), "Removed PTP clock\n");
}