* Driver support for Realtek RTL8125 2.5GB Ethernet
From: Jian-Hong Pan @ 2019-07-24 8:19 UTC (permalink / raw)
To: Linux Netdev List, Yan-Hsuan Chuang; +Cc: Linux Upstreaming Team
Hi all,
We have got a consumer desktop equipped with Realtek RTL8125 2.5GB
Ethernet [1] recently. But, there is no related driver in mainline
kernel yet. So, we can only use the vendor driver [2] and customize
it [3] right now.
Is anyone working on an upstream driver for this hardware?
[1] https://www.realtek.com/en/press-room/news-releases/item/realtek-launches-world-s-first-single-chip-2-5g-ethernet-controller-for-multiple-applications-including-gaming-solution
[2] https://www.realtek.com/en/component/zoo/category/network-interface-controllers-10-100-1000m-gigabit-ethernet-pci-express-software
[3] https://github.com/endlessm/linux/commit/da1e43f58850d272eb72f571524ed71fd237d32b
Jian-Hong Pan
^ permalink raw reply
* Re: [RFC PATCH 1/2] gianfar: convert to phylink
From: Russell King - ARM Linux admin @ 2019-07-24 8:19 UTC (permalink / raw)
To: Arseny Solokha; +Cc: Claudiu Manoil, Ioana Ciornei, Andrew Lunn, netdev
In-Reply-To: <20190723151702.14430-2-asolokha@kb.kras.ru>
On Tue, Jul 23, 2019 at 10:17:01PM +0700, Arseny Solokha wrote:
> -static noinline void gfar_update_link_state(struct gfar_private *priv)
> +static void gfar_mac_config(struct phylink_config *config, unsigned int mode,
> + const struct phylink_link_state *state)
> {
> + struct gfar_private *priv = netdev_priv(to_net_dev(config->dev));
> struct gfar __iomem *regs = priv->gfargrp[0].regs;
> - struct net_device *ndev = priv->ndev;
> - struct phy_device *phydev = ndev->phydev;
> - struct gfar_priv_rx_q *rx_queue = NULL;
> - int i;
> + u32 maccfg1, new_maccfg1;
> + u32 maccfg2, new_maccfg2;
> + u32 ecntrl, new_ecntrl;
> + u32 tx_flow, new_tx_flow;
>
> if (unlikely(test_bit(GFAR_RESETTING, &priv->state)))
> return;
>
> - if (phydev->link) {
> - u32 tempval1 = gfar_read(®s->maccfg1);
> - u32 tempval = gfar_read(®s->maccfg2);
> - u32 ecntrl = gfar_read(®s->ecntrl);
> - u32 tx_flow_oldval = (tempval1 & MACCFG1_TX_FLOW);
> + if (unlikely(phylink_autoneg_inband(mode)))
> + return;
Given that SFPs can be either SGMII or 1000BASE-X (which require
different configuration) and that the intention here is to support
SFPs, I don't see how this works with the above.
How is the difference between SGMII and 1000BASE-X handled?
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
^ permalink raw reply
* [PATCH net-next v3 8/8] net: mscc: PTP Hardware Clock (PHC) support
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
This patch adds support for PTP Hardware Clock (PHC) to the Ocelot
switch for both PTP 1-step and 2-step modes.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
drivers/net/ethernet/mscc/ocelot.c | 394 ++++++++++++++++++++++-
drivers/net/ethernet/mscc/ocelot.h | 39 +++
drivers/net/ethernet/mscc/ocelot_board.c | 111 ++++++-
3 files changed, 536 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/mscc/ocelot.c b/drivers/net/ethernet/mscc/ocelot.c
index b71e4ecbe469..4f287202bf03 100644
--- a/drivers/net/ethernet/mscc/ocelot.c
+++ b/drivers/net/ethernet/mscc/ocelot.c
@@ -14,6 +14,7 @@
#include <linux/module.h>
#include <linux/netdevice.h>
#include <linux/phy.h>
+#include <linux/ptp_clock_kernel.h>
#include <linux/skbuff.h>
#include <linux/iopoll.h>
#include <net/arp.h>
@@ -538,7 +539,7 @@ static int ocelot_port_stop(struct net_device *dev)
*/
static int ocelot_gen_ifh(u32 *ifh, struct frame_info *info)
{
- ifh[0] = IFH_INJ_BYPASS;
+ ifh[0] = IFH_INJ_BYPASS | ((0x1ff & info->rew_op) << 21);
ifh[1] = (0xf00 & info->port) >> 8;
ifh[2] = (0xff & info->port) << 24;
ifh[3] = (info->tag_type << 16) | info->vid;
@@ -548,6 +549,7 @@ static int ocelot_gen_ifh(u32 *ifh, struct frame_info *info)
static int ocelot_port_xmit(struct sk_buff *skb, struct net_device *dev)
{
+ struct skb_shared_info *shinfo = skb_shinfo(skb);
struct ocelot_port *port = netdev_priv(dev);
struct ocelot *ocelot = port->ocelot;
u32 val, ifh[IFH_LEN];
@@ -566,6 +568,14 @@ static int ocelot_port_xmit(struct sk_buff *skb, struct net_device *dev)
info.port = BIT(port->chip_port);
info.tag_type = IFH_TAG_TYPE_C;
info.vid = skb_vlan_tag_get(skb);
+
+ /* Check if timestamping is needed */
+ if (ocelot->ptp && shinfo->tx_flags & SKBTX_HW_TSTAMP) {
+ info.rew_op = port->ptp_cmd;
+ if (port->ptp_cmd == IFH_REW_OP_TWO_STEP_PTP)
+ info.rew_op |= (port->ts_id % 4) << 3;
+ }
+
ocelot_gen_ifh(ifh, &info);
for (i = 0; i < IFH_LEN; i++)
@@ -596,11 +606,51 @@ static int ocelot_port_xmit(struct sk_buff *skb, struct net_device *dev)
dev->stats.tx_packets++;
dev->stats.tx_bytes += skb->len;
- dev_kfree_skb_any(skb);
+
+ if (ocelot->ptp && shinfo->tx_flags & SKBTX_HW_TSTAMP &&
+ port->ptp_cmd == IFH_REW_OP_TWO_STEP_PTP) {
+ struct ocelot_skb *oskb =
+ kzalloc(sizeof(struct ocelot_skb), GFP_ATOMIC);
+
+ oskb->skb = skb;
+ oskb->id = port->ts_id % 4;
+ port->ts_id++;
+
+ list_add_tail(&oskb->head, &port->skbs);
+ } else {
+ dev_kfree_skb_any(skb);
+ }
return NETDEV_TX_OK;
}
+void ocelot_get_hwtimestamp(struct ocelot *ocelot, struct timespec64 *ts)
+{
+ unsigned long flags;
+ u32 val;
+
+ spin_lock_irqsave(&ocelot->ptp_clock_lock, flags);
+
+ /* Read current PTP time to get seconds */
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_SAVE);
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+ ts->tv_sec = ocelot_read_rix(ocelot, PTP_PIN_TOD_SEC_LSB, TOD_ACC_PIN);
+
+ /* Read packet HW timestamp from FIFO */
+ val = ocelot_read(ocelot, SYS_PTP_TXSTAMP);
+ ts->tv_nsec = SYS_PTP_TXSTAMP_PTP_TXSTAMP(val);
+
+ /* Sec has incremented since the ts was registered */
+ if ((ts->tv_sec & 0x1) != !!(val & SYS_PTP_TXSTAMP_PTP_TXSTAMP_SEC))
+ ts->tv_sec--;
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+}
+EXPORT_SYMBOL(ocelot_get_hwtimestamp);
+
static int ocelot_mc_unsync(struct net_device *dev, const unsigned char *addr)
{
struct ocelot_port *port = netdev_priv(dev);
@@ -917,6 +967,97 @@ static int ocelot_get_port_parent_id(struct net_device *dev,
return 0;
}
+static int ocelot_hwstamp_get(struct ocelot_port *port, struct ifreq *ifr)
+{
+ struct ocelot *ocelot = port->ocelot;
+
+ return copy_to_user(ifr->ifr_data, &ocelot->hwtstamp_config,
+ sizeof(ocelot->hwtstamp_config)) ? -EFAULT : 0;
+}
+
+static int ocelot_hwstamp_set(struct ocelot_port *port, struct ifreq *ifr)
+{
+ struct ocelot *ocelot = port->ocelot;
+ struct hwtstamp_config cfg;
+
+ if (copy_from_user(&cfg, ifr->ifr_data, sizeof(cfg)))
+ return -EFAULT;
+
+ /* reserved for future extensions */
+ if (cfg.flags)
+ return -EINVAL;
+
+ /* Tx type sanity check */
+ switch (cfg.tx_type) {
+ case HWTSTAMP_TX_ON:
+ port->ptp_cmd = IFH_REW_OP_TWO_STEP_PTP;
+ break;
+ case HWTSTAMP_TX_ONESTEP_SYNC:
+ /* IFH_REW_OP_ONE_STEP_PTP updates the correctional field, we
+ * need to update the origin time.
+ */
+ port->ptp_cmd = IFH_REW_OP_ORIGIN_PTP;
+ break;
+ case HWTSTAMP_TX_OFF:
+ port->ptp_cmd = 0;
+ break;
+ default:
+ return -ERANGE;
+ }
+
+ mutex_lock(&ocelot->ptp_lock);
+
+ switch (cfg.rx_filter) {
+ case HWTSTAMP_FILTER_NONE:
+ break;
+ case HWTSTAMP_FILTER_ALL:
+ case HWTSTAMP_FILTER_SOME:
+ 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_NTP_ALL:
+ case HWTSTAMP_FILTER_PTP_V2_L4_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L4_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L4_DELAY_REQ:
+ case HWTSTAMP_FILTER_PTP_V2_L2_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_L2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_L2_DELAY_REQ:
+ case HWTSTAMP_FILTER_PTP_V2_EVENT:
+ case HWTSTAMP_FILTER_PTP_V2_SYNC:
+ case HWTSTAMP_FILTER_PTP_V2_DELAY_REQ:
+ cfg.rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT;
+ break;
+ default:
+ mutex_unlock(&ocelot->ptp_lock);
+ return -ERANGE;
+ }
+
+ /* Commit back the result & save it */
+ memcpy(&ocelot->hwtstamp_config, &cfg, sizeof(cfg));
+ mutex_unlock(&ocelot->ptp_lock);
+
+ return copy_to_user(ifr->ifr_data, &cfg, sizeof(cfg)) ? -EFAULT : 0;
+}
+
+static int ocelot_ioctl(struct net_device *dev, struct ifreq *ifr, int cmd)
+{
+ struct ocelot_port *port = netdev_priv(dev);
+ struct ocelot *ocelot = port->ocelot;
+
+ /* The function is only used for PTP operations for now */
+ if (!ocelot->ptp)
+ return -EOPNOTSUPP;
+
+ switch (cmd) {
+ case SIOCSHWTSTAMP:
+ return ocelot_hwstamp_set(port, ifr);
+ case SIOCGHWTSTAMP:
+ return ocelot_hwstamp_get(port, ifr);
+ default:
+ return -EOPNOTSUPP;
+ }
+}
+
static const struct net_device_ops ocelot_port_netdev_ops = {
.ndo_open = ocelot_port_open,
.ndo_stop = ocelot_port_stop,
@@ -933,6 +1074,7 @@ static const struct net_device_ops ocelot_port_netdev_ops = {
.ndo_set_features = ocelot_set_features,
.ndo_get_port_parent_id = ocelot_get_port_parent_id,
.ndo_setup_tc = ocelot_setup_tc,
+ .ndo_do_ioctl = ocelot_ioctl,
};
static void ocelot_get_strings(struct net_device *netdev, u32 sset, u8 *data)
@@ -1014,12 +1156,37 @@ static int ocelot_get_sset_count(struct net_device *dev, int sset)
return ocelot->num_stats;
}
+static int ocelot_get_ts_info(struct net_device *dev,
+ struct ethtool_ts_info *info)
+{
+ struct ocelot_port *ocelot_port = netdev_priv(dev);
+ struct ocelot *ocelot = ocelot_port->ocelot;
+
+ if (!ocelot->ptp)
+ return ethtool_op_get_ts_info(dev, info);
+
+ info->phc_index = ocelot->ptp_clock ?
+ ptp_clock_index(ocelot->ptp_clock) : -1;
+ info->so_timestamping |= SOF_TIMESTAMPING_TX_SOFTWARE |
+ SOF_TIMESTAMPING_RX_SOFTWARE |
+ SOF_TIMESTAMPING_SOFTWARE |
+ SOF_TIMESTAMPING_TX_HARDWARE |
+ SOF_TIMESTAMPING_RX_HARDWARE |
+ SOF_TIMESTAMPING_RAW_HARDWARE;
+ info->tx_types = BIT(HWTSTAMP_TX_OFF) | BIT(HWTSTAMP_TX_ON) |
+ BIT(HWTSTAMP_TX_ONESTEP_SYNC);
+ info->rx_filters = BIT(HWTSTAMP_FILTER_NONE) | BIT(HWTSTAMP_FILTER_ALL);
+
+ return 0;
+}
+
static const struct ethtool_ops ocelot_ethtool_ops = {
.get_strings = ocelot_get_strings,
.get_ethtool_stats = ocelot_get_ethtool_stats,
.get_sset_count = ocelot_get_sset_count,
.get_link_ksettings = phy_ethtool_get_link_ksettings,
.set_link_ksettings = phy_ethtool_set_link_ksettings,
+ .get_ts_info = ocelot_get_ts_info,
};
static int ocelot_port_attr_stp_state_set(struct ocelot_port *ocelot_port,
@@ -1629,6 +1796,196 @@ struct notifier_block ocelot_switchdev_blocking_nb __read_mostly = {
};
EXPORT_SYMBOL(ocelot_switchdev_blocking_nb);
+int ocelot_ptp_gettime64(struct ptp_clock_info *ptp, struct timespec64 *ts)
+{
+ struct ocelot *ocelot = container_of(ptp, struct ocelot, ptp_info);
+ unsigned long flags;
+ time64_t s;
+ u32 val;
+ s64 ns;
+
+ spin_lock_irqsave(&ocelot->ptp_clock_lock, flags);
+
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_SAVE);
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ s = ocelot_read_rix(ocelot, PTP_PIN_TOD_SEC_MSB, TOD_ACC_PIN) & 0xffff;
+ s <<= 32;
+ s += ocelot_read_rix(ocelot, PTP_PIN_TOD_SEC_LSB, TOD_ACC_PIN);
+ ns = ocelot_read_rix(ocelot, PTP_PIN_TOD_NSEC, TOD_ACC_PIN);
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+
+ /* Deal with negative values */
+ if (ns >= 0x3ffffff0 && ns <= 0x3fffffff) {
+ s--;
+ ns &= 0xf;
+ ns += 999999984;
+ }
+
+ set_normalized_timespec64(ts, s, ns);
+ return 0;
+}
+EXPORT_SYMBOL(ocelot_ptp_gettime64);
+
+static int ocelot_ptp_settime64(struct ptp_clock_info *ptp,
+ const struct timespec64 *ts)
+{
+ struct ocelot *ocelot = container_of(ptp, struct ocelot, ptp_info);
+ unsigned long flags;
+ u32 val;
+
+ spin_lock_irqsave(&ocelot->ptp_clock_lock, flags);
+
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_IDLE);
+
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ ocelot_write_rix(ocelot, lower_32_bits(ts->tv_sec), PTP_PIN_TOD_SEC_LSB,
+ TOD_ACC_PIN);
+ ocelot_write_rix(ocelot, upper_32_bits(ts->tv_sec), PTP_PIN_TOD_SEC_MSB,
+ TOD_ACC_PIN);
+ ocelot_write_rix(ocelot, ts->tv_nsec, PTP_PIN_TOD_NSEC, TOD_ACC_PIN);
+
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_LOAD);
+
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+ return 0;
+}
+
+static int ocelot_ptp_adjtime(struct ptp_clock_info *ptp, s64 delta)
+{
+ if (delta > -(NSEC_PER_SEC / 2) && delta < (NSEC_PER_SEC / 2)) {
+ struct ocelot *ocelot = container_of(ptp, struct ocelot, ptp_info);
+ unsigned long flags;
+ u32 val;
+
+ spin_lock_irqsave(&ocelot->ptp_clock_lock, flags);
+
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_IDLE);
+
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ ocelot_write_rix(ocelot, 0, PTP_PIN_TOD_SEC_LSB, TOD_ACC_PIN);
+ ocelot_write_rix(ocelot, 0, PTP_PIN_TOD_SEC_MSB, TOD_ACC_PIN);
+ ocelot_write_rix(ocelot, delta, PTP_PIN_TOD_NSEC, TOD_ACC_PIN);
+
+ val = ocelot_read_rix(ocelot, PTP_PIN_CFG, TOD_ACC_PIN);
+ val &= ~(PTP_PIN_CFG_SYNC | PTP_PIN_CFG_ACTION_MASK | PTP_PIN_CFG_DOM);
+ val |= PTP_PIN_CFG_ACTION(PTP_PIN_ACTION_DELTA);
+
+ ocelot_write_rix(ocelot, val, PTP_PIN_CFG, TOD_ACC_PIN);
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+ } else {
+ /* Fall back using ocelot_ptp_settime64 which is not exact. */
+ struct timespec64 ts;
+ u64 now;
+
+ ocelot_ptp_gettime64(ptp, &ts);
+
+ now = ktime_to_ns(timespec64_to_ktime(ts));
+ ts = ns_to_timespec64(now + delta);
+
+ ocelot_ptp_settime64(ptp, &ts);
+ }
+ return 0;
+}
+
+static int ocelot_ptp_adjfine(struct ptp_clock_info *ptp, long scaled_ppm)
+{
+ struct ocelot *ocelot = container_of(ptp, struct ocelot, ptp_info);
+ u32 unit = 0, direction = 0;
+ unsigned long flags;
+ u64 adj = 0;
+
+ if (!scaled_ppm)
+ goto disable_adj;
+
+ if (scaled_ppm < 0) {
+ direction = PTP_CFG_CLK_ADJ_CFG_DIR;
+ scaled_ppm = -scaled_ppm;
+ }
+
+ adj = PSEC_PER_SEC << 16;
+ do_div(adj, scaled_ppm);
+ do_div(adj, 1000);
+
+ /* If the adjustment value is too large, use ns instead */
+ if (adj >= (1L << 30)) {
+ unit = PTP_CFG_CLK_ADJ_FREQ_NS;
+ do_div(adj, 1000);
+ }
+
+ spin_lock_irqsave(&ocelot->ptp_clock_lock, flags);
+
+ /* Still too big */
+ if (adj >= (1L << 30))
+ goto disable_adj;
+
+ ocelot_write(ocelot, unit | adj, PTP_CLK_CFG_ADJ_FREQ);
+ ocelot_write(ocelot, PTP_CFG_CLK_ADJ_CFG_ENA | direction,
+ PTP_CLK_CFG_ADJ_CFG);
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+ return 0;
+
+disable_adj:
+ ocelot_write(ocelot, 0, PTP_CLK_CFG_ADJ_CFG);
+
+ spin_unlock_irqrestore(&ocelot->ptp_clock_lock, flags);
+ return 0;
+}
+
+static struct ptp_clock_info ocelot_ptp_clock_info = {
+ .owner = THIS_MODULE,
+ .name = "ocelot ptp",
+ .max_adj = 0x7fffffff,
+ .n_alarm = 0,
+ .n_ext_ts = 0,
+ .n_per_out = 0,
+ .n_pins = 0,
+ .pps = 0,
+ .gettime64 = ocelot_ptp_gettime64,
+ .settime64 = ocelot_ptp_settime64,
+ .adjtime = ocelot_ptp_adjtime,
+ .adjfine = ocelot_ptp_adjfine,
+};
+
+static int ocelot_init_timestamp(struct ocelot *ocelot)
+{
+ ocelot->ptp_info = ocelot_ptp_clock_info;
+ ocelot->ptp_clock = ptp_clock_register(&ocelot->ptp_info, ocelot->dev);
+ if (IS_ERR(ocelot->ptp_clock))
+ return PTR_ERR(ocelot->ptp_clock);
+ /* Check if PHC support is missing at the configuration level */
+ if (!ocelot->ptp_clock)
+ return 0;
+
+ ocelot_write(ocelot, SYS_PTP_CFG_PTP_STAMP_WID(30), SYS_PTP_CFG);
+ ocelot_write(ocelot, 0xffffffff, ANA_TABLES_PTP_ID_LOW);
+ ocelot_write(ocelot, 0xffffffff, ANA_TABLES_PTP_ID_HIGH);
+
+ ocelot_write(ocelot, PTP_CFG_MISC_PTP_EN, PTP_CFG_MISC);
+
+ /* There is no device reconfiguration, PTP Rx stamping is always
+ * enabled.
+ */
+ ocelot->hwtstamp_config.rx_filter = HWTSTAMP_FILTER_PTP_V2_EVENT;
+
+ return 0;
+}
+
int ocelot_probe_port(struct ocelot *ocelot, u8 port,
void __iomem *regs,
struct phy_device *phy)
@@ -1661,6 +2018,8 @@ int ocelot_probe_port(struct ocelot *ocelot, u8 port,
ocelot_mact_learn(ocelot, PGID_CPU, dev->dev_addr, ocelot_port->pvid,
ENTRYTYPE_LOCKED);
+ INIT_LIST_HEAD(&ocelot_port->skbs);
+
err = register_netdev(dev);
if (err) {
dev_err(ocelot->dev, "register_netdev failed\n");
@@ -1684,7 +2043,7 @@ EXPORT_SYMBOL(ocelot_probe_port);
int ocelot_init(struct ocelot *ocelot)
{
u32 port;
- int i, cpu = ocelot->num_phys_ports;
+ int i, ret, cpu = ocelot->num_phys_ports;
char queue_name[32];
ocelot->lags = devm_kcalloc(ocelot->dev, ocelot->num_phys_ports,
@@ -1699,6 +2058,8 @@ int ocelot_init(struct ocelot *ocelot)
return -ENOMEM;
mutex_init(&ocelot->stats_lock);
+ mutex_init(&ocelot->ptp_lock);
+ spin_lock_init(&ocelot->ptp_clock_lock);
snprintf(queue_name, sizeof(queue_name), "%s-stats",
dev_name(ocelot->dev));
ocelot->stats_queue = create_singlethread_workqueue(queue_name);
@@ -1812,15 +2173,42 @@ int ocelot_init(struct ocelot *ocelot)
INIT_DELAYED_WORK(&ocelot->stats_work, ocelot_check_stats_work);
queue_delayed_work(ocelot->stats_queue, &ocelot->stats_work,
OCELOT_STATS_CHECK_DELAY);
+
+ if (ocelot->ptp) {
+ ret = ocelot_init_timestamp(ocelot);
+ if (ret) {
+ dev_err(ocelot->dev,
+ "Timestamp initialization failed\n");
+ return ret;
+ }
+ }
+
return 0;
}
EXPORT_SYMBOL(ocelot_init);
void ocelot_deinit(struct ocelot *ocelot)
{
+ struct list_head *pos, *tmp;
+ struct ocelot_port *port;
+ struct ocelot_skb *entry;
+ int i;
+
destroy_workqueue(ocelot->stats_queue);
mutex_destroy(&ocelot->stats_lock);
ocelot_ace_deinit();
+
+ for (i = 0; i < ocelot->num_phys_ports; i++) {
+ port = ocelot->ports[i];
+
+ list_for_each_safe(pos, tmp, &port->skbs) {
+ entry = list_entry(pos, struct ocelot_skb, head);
+
+ list_del(pos);
+ dev_kfree_skb_any(entry->skb);
+ kfree(entry);
+ }
+ }
}
EXPORT_SYMBOL(ocelot_deinit);
diff --git a/drivers/net/ethernet/mscc/ocelot.h b/drivers/net/ethernet/mscc/ocelot.h
index 515dee6fa8a6..e40773c01a44 100644
--- a/drivers/net/ethernet/mscc/ocelot.h
+++ b/drivers/net/ethernet/mscc/ocelot.h
@@ -11,9 +11,11 @@
#include <linux/bitops.h>
#include <linux/etherdevice.h>
#include <linux/if_vlan.h>
+#include <linux/net_tstamp.h>
#include <linux/phy.h>
#include <linux/phy/phy.h>
#include <linux/platform_device.h>
+#include <linux/ptp_clock_kernel.h>
#include <linux/regmap.h>
#include "ocelot_ana.h"
@@ -39,6 +41,8 @@
#define OCELOT_STATS_CHECK_DELAY (2 * HZ)
+#define OCELOT_PTP_QUEUE_SZ 128
+
#define IFH_LEN 4
struct frame_info {
@@ -46,6 +50,8 @@ struct frame_info {
u16 port;
u16 vid;
u8 tag_type;
+ u16 rew_op;
+ u32 timestamp; /* rew_val */
};
#define IFH_INJ_BYPASS BIT(31)
@@ -54,6 +60,12 @@ struct frame_info {
#define IFH_TAG_TYPE_C 0
#define IFH_TAG_TYPE_S 1
+#define IFH_REW_OP_NOOP 0x0
+#define IFH_REW_OP_DSCP 0x1
+#define IFH_REW_OP_ONE_STEP_PTP 0x2
+#define IFH_REW_OP_TWO_STEP_PTP 0x3
+#define IFH_REW_OP_ORIGIN_PTP 0x5
+
#define OCELOT_SPEED_2500 0
#define OCELOT_SPEED_1000 1
#define OCELOT_SPEED_100 2
@@ -401,6 +413,13 @@ enum ocelot_regfield {
REGFIELD_MAX
};
+enum ocelot_clk_pins {
+ ALT_PPS_PIN = 1,
+ EXT_CLK_PIN,
+ ALT_LDST_PIN,
+ TOD_ACC_PIN
+};
+
struct ocelot_multicast {
struct list_head list;
unsigned char addr[ETH_ALEN];
@@ -450,6 +469,13 @@ struct ocelot {
u64 *stats;
struct delayed_work stats_work;
struct workqueue_struct *stats_queue;
+
+ u8 ptp:1;
+ struct ptp_clock *ptp_clock;
+ struct ptp_clock_info ptp_info;
+ struct hwtstamp_config hwtstamp_config;
+ struct mutex ptp_lock; /* Protects the PTP interface state */
+ spinlock_t ptp_clock_lock; /* Protects the PTP clock */
};
struct ocelot_port {
@@ -473,6 +499,16 @@ struct ocelot_port {
struct phy *serdes;
struct ocelot_port_tc tc;
+
+ u8 ptp_cmd;
+ struct list_head skbs;
+ u8 ts_id;
+};
+
+struct ocelot_skb {
+ struct list_head head;
+ struct sk_buff *skb;
+ u8 id;
};
u32 __ocelot_read_ix(struct ocelot *ocelot, u32 reg, u32 offset);
@@ -517,4 +553,7 @@ extern struct notifier_block ocelot_netdevice_nb;
extern struct notifier_block ocelot_switchdev_nb;
extern struct notifier_block ocelot_switchdev_blocking_nb;
+int ocelot_ptp_gettime64(struct ptp_clock_info *ptp, struct timespec64 *ts);
+void ocelot_get_hwtimestamp(struct ocelot *ocelot, struct timespec64 *ts);
+
#endif
diff --git a/drivers/net/ethernet/mscc/ocelot_board.c b/drivers/net/ethernet/mscc/ocelot_board.c
index df8d15994a89..0b14e7110e7f 100644
--- a/drivers/net/ethernet/mscc/ocelot_board.c
+++ b/drivers/net/ethernet/mscc/ocelot_board.c
@@ -31,6 +31,8 @@ static int ocelot_parse_ifh(u32 *_ifh, struct frame_info *info)
info->len = OCELOT_BUFFER_CELL_SZ * wlen + llen - 80;
+ info->timestamp = IFH_EXTRACT_BITFIELD64(ifh[0], 21, 32);
+
info->port = IFH_EXTRACT_BITFIELD64(ifh[1], 43, 4);
info->tag_type = IFH_EXTRACT_BITFIELD64(ifh[1], 16, 1);
@@ -98,7 +100,11 @@ static irqreturn_t ocelot_xtr_irq_handler(int irq, void *arg)
int sz, len, buf_len;
u32 ifh[4];
u32 val;
- struct frame_info info;
+ struct frame_info info = {};
+ struct timespec64 ts;
+ struct skb_shared_hwtstamps *shhwtstamps;
+ u64 tod_in_ns;
+ u64 full_ts_in_ns;
for (i = 0; i < IFH_LEN; i++) {
err = ocelot_rx_frame_word(ocelot, grp, true, &ifh[i]);
@@ -145,6 +151,22 @@ static irqreturn_t ocelot_xtr_irq_handler(int irq, void *arg)
break;
}
+ if (ocelot->ptp) {
+ ocelot_ptp_gettime64(&ocelot->ptp_info, &ts);
+
+ tod_in_ns = ktime_set(ts.tv_sec, ts.tv_nsec);
+ if ((tod_in_ns & 0xffffffff) < info.timestamp)
+ full_ts_in_ns = (((tod_in_ns >> 32) - 1) << 32) |
+ info.timestamp;
+ else
+ full_ts_in_ns = (tod_in_ns & GENMASK_ULL(63, 32)) |
+ info.timestamp;
+
+ shhwtstamps = skb_hwtstamps(skb);
+ memset(shhwtstamps, 0, sizeof(struct skb_shared_hwtstamps));
+ shhwtstamps->hwtstamp = full_ts_in_ns;
+ }
+
/* Everything we see on an interface that is in the HW bridge
* has already been forwarded.
*/
@@ -164,6 +186,70 @@ static irqreturn_t ocelot_xtr_irq_handler(int irq, void *arg)
return IRQ_HANDLED;
}
+static irqreturn_t ocelot_ptp_rdy_irq_handler(int irq, void *arg)
+{
+ int budget = OCELOT_PTP_QUEUE_SZ;
+ struct ocelot *ocelot = arg;
+
+ do {
+ struct skb_shared_hwtstamps shhwtstamps;
+ struct list_head *pos, *tmp;
+ struct sk_buff *skb = NULL;
+ struct ocelot_skb *entry;
+ struct ocelot_port *port;
+ struct timespec64 ts;
+ u32 val, id, txport;
+
+ /* Prevent from infinite loop */
+ if (unlikely(!--budget))
+ break;
+
+ val = ocelot_read(ocelot, SYS_PTP_STATUS);
+
+ /* Check if a timestamp can be retrieved */
+ if (!(val & SYS_PTP_STATUS_PTP_MESS_VLD))
+ break;
+
+ WARN_ON(val & SYS_PTP_STATUS_PTP_OVFL);
+
+ /* Retrieve the ts ID and Tx port */
+ id = SYS_PTP_STATUS_PTP_MESS_ID_X(val);
+ txport = SYS_PTP_STATUS_PTP_MESS_TXPORT_X(val);
+
+ /* Retrieve its associated skb */
+ port = ocelot->ports[txport];
+
+ list_for_each_safe(pos, tmp, &port->skbs) {
+ entry = list_entry(pos, struct ocelot_skb, head);
+ if (entry->id != id)
+ continue;
+
+ skb = entry->skb;
+
+ list_del(pos);
+ kfree(entry);
+ }
+
+ /* Next ts */
+ ocelot_write(ocelot, SYS_PTP_NXT_PTP_NXT, SYS_PTP_NXT);
+
+ if (unlikely(!skb))
+ continue;
+
+ /* Get the h/w timestamp */
+ ocelot_get_hwtimestamp(ocelot, &ts);
+
+ /* Set the timestamp into the skb */
+ memset(&shhwtstamps, 0, sizeof(shhwtstamps));
+ shhwtstamps.hwtstamp = ktime_set(ts.tv_sec, ts.tv_nsec);
+ skb_tstamp_tx(skb, &shhwtstamps);
+
+ dev_kfree_skb_any(skb);
+ } while (true);
+
+ return IRQ_HANDLED;
+}
+
static const struct of_device_id mscc_ocelot_match[] = {
{ .compatible = "mscc,vsc7514-switch" },
{ }
@@ -172,8 +258,8 @@ MODULE_DEVICE_TABLE(of, mscc_ocelot_match);
static int mscc_ocelot_probe(struct platform_device *pdev)
{
- int err, irq;
unsigned int i;
+ int err, irq_xtr, irq_ptp_rdy;
struct device_node *np = pdev->dev.of_node;
struct device_node *ports, *portnp;
struct ocelot *ocelot;
@@ -232,16 +318,31 @@ static int mscc_ocelot_probe(struct platform_device *pdev)
if (err)
return err;
- irq = platform_get_irq_byname(pdev, "xtr");
- if (irq < 0)
+ irq_xtr = platform_get_irq_byname(pdev, "xtr");
+ if (irq_xtr < 0)
return -ENODEV;
- err = devm_request_threaded_irq(&pdev->dev, irq, NULL,
+ err = devm_request_threaded_irq(&pdev->dev, irq_xtr, NULL,
ocelot_xtr_irq_handler, IRQF_ONESHOT,
"frame extraction", ocelot);
if (err)
return err;
+
+ irq_ptp_rdy = platform_get_irq_byname(pdev, "ptp_rdy");
+ if (irq_ptp_rdy > 0) {
+ err = devm_request_threaded_irq(&pdev->dev, irq_ptp_rdy, NULL,
+ ocelot_ptp_rdy_irq_handler,
+ IRQF_ONESHOT, "ptp ready",
+ ocelot);
+ if (err)
+ return err;
+
+ /* Check if we can support PTP */
+ if (ocelot->targets[PTP])
+ ocelot->ptp = 1;
+ }
+
regmap_field_write(ocelot->regfields[SYS_RESET_CFG_MEM_INIT], 1);
regmap_field_write(ocelot->regfields[SYS_RESET_CFG_MEM_ENA], 1);
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 7/8] net: mscc: remove the frame_info cpuq member
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
In struct frame_info, the cpuq member is never used. This cosmetic patch
removes it from the structure, and from the parsing of the frame header
as it's only set but never used.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
drivers/net/ethernet/mscc/ocelot.h | 1 -
drivers/net/ethernet/mscc/ocelot_board.c | 1 -
2 files changed, 2 deletions(-)
diff --git a/drivers/net/ethernet/mscc/ocelot.h b/drivers/net/ethernet/mscc/ocelot.h
index e0da8b4eddf2..515dee6fa8a6 100644
--- a/drivers/net/ethernet/mscc/ocelot.h
+++ b/drivers/net/ethernet/mscc/ocelot.h
@@ -45,7 +45,6 @@ struct frame_info {
u32 len;
u16 port;
u16 vid;
- u8 cpuq;
u8 tag_type;
};
diff --git a/drivers/net/ethernet/mscc/ocelot_board.c b/drivers/net/ethernet/mscc/ocelot_board.c
index 5e4f1718dd99..df8d15994a89 100644
--- a/drivers/net/ethernet/mscc/ocelot_board.c
+++ b/drivers/net/ethernet/mscc/ocelot_board.c
@@ -33,7 +33,6 @@ static int ocelot_parse_ifh(u32 *_ifh, struct frame_info *info)
info->port = IFH_EXTRACT_BITFIELD64(ifh[1], 43, 4);
- info->cpuq = IFH_EXTRACT_BITFIELD64(ifh[1], 20, 8);
info->tag_type = IFH_EXTRACT_BITFIELD64(ifh[1], 16, 1);
info->vid = IFH_EXTRACT_BITFIELD64(ifh[1], 0, 12);
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 0/8] net: mscc: PTP Hardware Clock (PHC) support
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
Hello,
This series introduces the PTP Hardware Clock (PHC) support to the Mscc
Ocelot switch driver. In order to make use of this, a new register bank
is added and described in the device tree, as well as a new interrupt.
The use this bank and interrupt was made optional in the driver for dt
compatibility reasons.
Patches 2 and 4 should probably go through the MIPS tree.
Thanks!
Antoine
Since v2:
- Prevented from a possible infinite loop when reading the h/w
timestamps.
- s/GFP_KERNEL/GFP_ATOMIC/ in the Tx path.
- Set rx_filter to HWTSTAMP_FILTER_PTP_V2_EVENT at probe.
- Fixed s/w timestamping dependencies.
- Added Paul Burton's Acked-by on patches 2 and 4.
Since v1:
- Used list_for_each_safe() in ocelot_deinit().
- Fixed a memory leak in ocelot_deinit() by calling
dev_kfree_skb_any().
- Fixed a locking issue in get_hwtimestamp().
- Handled the NULL case of ptp_clock_register().
- Added comments on optional dt properties.
Antoine Tenart (8):
Documentation/bindings: net: ocelot: document the PTP bank
MIPS: dts: mscc: describe the PTP register range
Documentation/bindings: net: ocelot: document the PTP ready IRQ
MIPS: dts: mscc: describe the PTP ready interrupt
net: mscc: describe the PTP register range
net: mscc: improve the frame header parsing readability
net: mscc: remove the frame_info cpuq member
net: mscc: PTP Hardware Clock (PHC) support
.../devicetree/bindings/net/mscc-ocelot.txt | 20 +-
arch/mips/boot/dts/mscc/ocelot.dtsi | 7 +-
drivers/net/ethernet/mscc/ocelot.c | 394 +++++++++++++++++-
drivers/net/ethernet/mscc/ocelot.h | 49 ++-
drivers/net/ethernet/mscc/ocelot_board.c | 144 ++++++-
drivers/net/ethernet/mscc/ocelot_ptp.h | 41 ++
drivers/net/ethernet/mscc/ocelot_regs.c | 11 +
7 files changed, 634 insertions(+), 32 deletions(-)
create mode 100644 drivers/net/ethernet/mscc/ocelot_ptp.h
--
2.21.0
^ permalink raw reply
* [PATCH net-next v3 6/8] net: mscc: improve the frame header parsing readability
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
This cosmetic patch improves the frame header parsing readability by
introducing a new macro to access and mask its fields.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
drivers/net/ethernet/mscc/ocelot_board.c | 24 +++++++++++++-----------
1 file changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/mscc/ocelot_board.c b/drivers/net/ethernet/mscc/ocelot_board.c
index 990027f04d1b..5e4f1718dd99 100644
--- a/drivers/net/ethernet/mscc/ocelot_board.c
+++ b/drivers/net/ethernet/mscc/ocelot_board.c
@@ -16,24 +16,26 @@
#include "ocelot.h"
-static int ocelot_parse_ifh(u32 *ifh, struct frame_info *info)
+#define IFH_EXTRACT_BITFIELD64(x, o, w) (((x) >> (o)) & GENMASK_ULL((w) - 1, 0))
+
+static int ocelot_parse_ifh(u32 *_ifh, struct frame_info *info)
{
- int i;
u8 llen, wlen;
+ u64 ifh[2];
+
+ ifh[0] = be64_to_cpu(((__force __be64 *)_ifh)[0]);
+ ifh[1] = be64_to_cpu(((__force __be64 *)_ifh)[1]);
- /* The IFH is in network order, switch to CPU order */
- for (i = 0; i < IFH_LEN; i++)
- ifh[i] = ntohl((__force __be32)ifh[i]);
+ wlen = IFH_EXTRACT_BITFIELD64(ifh[0], 7, 8);
+ llen = IFH_EXTRACT_BITFIELD64(ifh[0], 15, 6);
- wlen = (ifh[1] >> 7) & 0xff;
- llen = (ifh[1] >> 15) & 0x3f;
info->len = OCELOT_BUFFER_CELL_SZ * wlen + llen - 80;
- info->port = (ifh[2] & GENMASK(14, 11)) >> 11;
+ info->port = IFH_EXTRACT_BITFIELD64(ifh[1], 43, 4);
- info->cpuq = (ifh[3] & GENMASK(27, 20)) >> 20;
- info->tag_type = (ifh[3] & BIT(16)) >> 16;
- info->vid = ifh[3] & GENMASK(11, 0);
+ info->cpuq = IFH_EXTRACT_BITFIELD64(ifh[1], 20, 8);
+ info->tag_type = IFH_EXTRACT_BITFIELD64(ifh[1], 16, 1);
+ info->vid = IFH_EXTRACT_BITFIELD64(ifh[1], 0, 12);
return 0;
}
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 5/8] net: mscc: describe the PTP register range
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
This patch adds support for using the PTP register range, and adds a
description of its registers. This bank is used when configuring PTP.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
drivers/net/ethernet/mscc/ocelot.h | 9 ++++++
drivers/net/ethernet/mscc/ocelot_board.c | 10 +++++-
drivers/net/ethernet/mscc/ocelot_ptp.h | 41 ++++++++++++++++++++++++
drivers/net/ethernet/mscc/ocelot_regs.c | 11 +++++++
4 files changed, 70 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/ethernet/mscc/ocelot_ptp.h
diff --git a/drivers/net/ethernet/mscc/ocelot.h b/drivers/net/ethernet/mscc/ocelot.h
index f7eeb4806897..e0da8b4eddf2 100644
--- a/drivers/net/ethernet/mscc/ocelot.h
+++ b/drivers/net/ethernet/mscc/ocelot.h
@@ -23,6 +23,7 @@
#include "ocelot_sys.h"
#include "ocelot_qs.h"
#include "ocelot_tc.h"
+#include "ocelot_ptp.h"
#define PGID_AGGR 64
#define PGID_SRC 80
@@ -71,6 +72,7 @@ enum ocelot_target {
SYS,
S2,
HSIO,
+ PTP,
TARGET_MAX,
};
@@ -343,6 +345,13 @@ enum ocelot_reg {
S2_CACHE_ACTION_DAT,
S2_CACHE_CNT_DAT,
S2_CACHE_TG_DAT,
+ PTP_PIN_CFG = PTP << TARGET_OFFSET,
+ PTP_PIN_TOD_SEC_MSB,
+ PTP_PIN_TOD_SEC_LSB,
+ PTP_PIN_TOD_NSEC,
+ PTP_CFG_MISC,
+ PTP_CLK_CFG_ADJ_CFG,
+ PTP_CLK_CFG_ADJ_FREQ,
};
enum ocelot_regfield {
diff --git a/drivers/net/ethernet/mscc/ocelot_board.c b/drivers/net/ethernet/mscc/ocelot_board.c
index 2451d4a96490..990027f04d1b 100644
--- a/drivers/net/ethernet/mscc/ocelot_board.c
+++ b/drivers/net/ethernet/mscc/ocelot_board.c
@@ -182,6 +182,7 @@ static int mscc_ocelot_probe(struct platform_device *pdev)
struct {
enum ocelot_target id;
char *name;
+ u8 optional:1;
} res[] = {
{ SYS, "sys" },
{ REW, "rew" },
@@ -189,6 +190,7 @@ static int mscc_ocelot_probe(struct platform_device *pdev)
{ ANA, "ana" },
{ QS, "qs" },
{ S2, "s2" },
+ { PTP, "ptp", 1 },
};
if (!np && !pdev->dev.platform_data)
@@ -205,8 +207,14 @@ static int mscc_ocelot_probe(struct platform_device *pdev)
struct regmap *target;
target = ocelot_io_platform_init(ocelot, pdev, res[i].name);
- if (IS_ERR(target))
+ if (IS_ERR(target)) {
+ if (res[i].optional) {
+ ocelot->targets[res[i].id] = NULL;
+ continue;
+ }
+
return PTR_ERR(target);
+ }
ocelot->targets[res[i].id] = target;
}
diff --git a/drivers/net/ethernet/mscc/ocelot_ptp.h b/drivers/net/ethernet/mscc/ocelot_ptp.h
new file mode 100644
index 000000000000..9ede14a12573
--- /dev/null
+++ b/drivers/net/ethernet/mscc/ocelot_ptp.h
@@ -0,0 +1,41 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR MIT) */
+/*
+ * Microsemi Ocelot Switch driver
+ *
+ * License: Dual MIT/GPL
+ * Copyright (c) 2017 Microsemi Corporation
+ */
+
+#ifndef _MSCC_OCELOT_PTP_H_
+#define _MSCC_OCELOT_PTP_H_
+
+#define PTP_PIN_CFG_RSZ 0x20
+#define PTP_PIN_TOD_SEC_MSB_RSZ PTP_PIN_CFG_RSZ
+#define PTP_PIN_TOD_SEC_LSB_RSZ PTP_PIN_CFG_RSZ
+#define PTP_PIN_TOD_NSEC_RSZ PTP_PIN_CFG_RSZ
+
+#define PTP_PIN_CFG_DOM BIT(0)
+#define PTP_PIN_CFG_SYNC BIT(2)
+#define PTP_PIN_CFG_ACTION(x) ((x) << 3)
+#define PTP_PIN_CFG_ACTION_MASK PTP_PIN_CFG_ACTION(0x7)
+
+enum {
+ PTP_PIN_ACTION_IDLE = 0,
+ PTP_PIN_ACTION_LOAD,
+ PTP_PIN_ACTION_SAVE,
+ PTP_PIN_ACTION_CLOCK,
+ PTP_PIN_ACTION_DELTA,
+ PTP_PIN_ACTION_NOSYNC,
+ PTP_PIN_ACTION_SYNC,
+};
+
+#define PTP_CFG_MISC_PTP_EN BIT(2)
+
+#define PSEC_PER_SEC 1000000000000LL
+
+#define PTP_CFG_CLK_ADJ_CFG_ENA BIT(0)
+#define PTP_CFG_CLK_ADJ_CFG_DIR BIT(1)
+
+#define PTP_CFG_CLK_ADJ_FREQ_NS BIT(30)
+
+#endif
diff --git a/drivers/net/ethernet/mscc/ocelot_regs.c b/drivers/net/ethernet/mscc/ocelot_regs.c
index 6c387f994ec5..e59977d20400 100644
--- a/drivers/net/ethernet/mscc/ocelot_regs.c
+++ b/drivers/net/ethernet/mscc/ocelot_regs.c
@@ -234,6 +234,16 @@ static const u32 ocelot_s2_regmap[] = {
REG(S2_CACHE_TG_DAT, 0x000388),
};
+static const u32 ocelot_ptp_regmap[] = {
+ REG(PTP_PIN_CFG, 0x000000),
+ REG(PTP_PIN_TOD_SEC_MSB, 0x000004),
+ REG(PTP_PIN_TOD_SEC_LSB, 0x000008),
+ REG(PTP_PIN_TOD_NSEC, 0x00000c),
+ REG(PTP_CFG_MISC, 0x0000a0),
+ REG(PTP_CLK_CFG_ADJ_CFG, 0x0000a4),
+ REG(PTP_CLK_CFG_ADJ_FREQ, 0x0000a8),
+};
+
static const u32 *ocelot_regmap[] = {
[ANA] = ocelot_ana_regmap,
[QS] = ocelot_qs_regmap,
@@ -241,6 +251,7 @@ static const u32 *ocelot_regmap[] = {
[REW] = ocelot_rew_regmap,
[SYS] = ocelot_sys_regmap,
[S2] = ocelot_s2_regmap,
+ [PTP] = ocelot_ptp_regmap,
};
static const struct reg_field ocelot_regfields[] = {
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 4/8] MIPS: dts: mscc: describe the PTP ready interrupt
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
This patch adds a description of the PTP ready interrupt, which can be
triggered when a PTP timestamp is available on an hardware FIFO.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Acked-by: Paul Burton <paul.burton@mips.com>
---
arch/mips/boot/dts/mscc/ocelot.dtsi | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
index 1e55a778def5..797d336db54d 100644
--- a/arch/mips/boot/dts/mscc/ocelot.dtsi
+++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
@@ -139,8 +139,8 @@
"port2", "port3", "port4", "port5", "port6",
"port7", "port8", "port9", "port10", "qsys",
"ana", "s2";
- interrupts = <21 22>;
- interrupt-names = "xtr", "inj";
+ interrupts = <18 21 22>;
+ interrupt-names = "ptp_rdy", "xtr", "inj";
ethernet-ports {
#address-cells = <1>;
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 3/8] Documentation/bindings: net: ocelot: document the PTP ready IRQ
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
One additional interrupt needs to be described within the Ocelot device
tree node: the PTP ready one. This patch documents the binding needed to
do so.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
Documentation/devicetree/bindings/net/mscc-ocelot.txt | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/mscc-ocelot.txt b/Documentation/devicetree/bindings/net/mscc-ocelot.txt
index 4d05a3b0f786..3b6290b45ce5 100644
--- a/Documentation/devicetree/bindings/net/mscc-ocelot.txt
+++ b/Documentation/devicetree/bindings/net/mscc-ocelot.txt
@@ -17,9 +17,10 @@ Required properties:
- "ana"
- "portX" with X from 0 to the number of last port index available on that
switch
-- interrupts: Should contain the switch interrupts for frame extraction and
- frame injection
-- interrupt-names: should contain the interrupt names: "xtr", "inj"
+- interrupts: Should contain the switch interrupts for frame extraction,
+ frame injection and PTP ready.
+- interrupt-names: should contain the interrupt names: "xtr", "inj". Can contain
+ "ptp_rdy" which is optional due to backward compatibility.
- ethernet-ports: A container for child nodes representing switch ports.
The ethernet-ports container has the following properties
@@ -63,8 +64,8 @@ Example:
"port2", "port3", "port4", "port5", "port6",
"port7", "port8", "port9", "port10", "qsys",
"ana";
- interrupts = <21 22>;
- interrupt-names = "xtr", "inj";
+ interrupts = <18 21 22>;
+ interrupt-names = "ptp_rdy", "xtr", "inj";
ethernet-ports {
#address-cells = <1>;
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 2/8] MIPS: dts: mscc: describe the PTP register range
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
This patch adds one register range within the mscc,vsc7514-switch node,
to describe the PTP registers.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
Acked-by: Paul Burton <paul.burton@mips.com>
---
arch/mips/boot/dts/mscc/ocelot.dtsi | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
index 33ae74aaa1bb..1e55a778def5 100644
--- a/arch/mips/boot/dts/mscc/ocelot.dtsi
+++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
@@ -120,6 +120,7 @@
reg = <0x1010000 0x10000>,
<0x1030000 0x10000>,
<0x1080000 0x100>,
+ <0x10e0000 0x10000>,
<0x11e0000 0x100>,
<0x11f0000 0x100>,
<0x1200000 0x100>,
@@ -134,7 +135,7 @@
<0x1800000 0x80000>,
<0x1880000 0x10000>,
<0x1060000 0x10000>;
- reg-names = "sys", "rew", "qs", "port0", "port1",
+ reg-names = "sys", "rew", "qs", "ptp", "port0", "port1",
"port2", "port3", "port4", "port5", "port6",
"port7", "port8", "port9", "port10", "qsys",
"ana", "s2";
--
2.21.0
^ permalink raw reply related
* [PATCH net-next v3 1/8] Documentation/bindings: net: ocelot: document the PTP bank
From: Antoine Tenart @ 2019-07-24 8:17 UTC (permalink / raw)
To: davem, richardcochran, alexandre.belloni, UNGLinuxDriver, ralf,
paul.burton, jhogan
Cc: Antoine Tenart, netdev, linux-mips, thomas.petazzoni,
allan.nielsen
In-Reply-To: <20190724081715.29159-1-antoine.tenart@bootlin.com>
One additional register range needs to be described within the Ocelot
device tree node: the PTP. This patch documents the binding needed to do
so.
Signed-off-by: Antoine Tenart <antoine.tenart@bootlin.com>
---
Documentation/devicetree/bindings/net/mscc-ocelot.txt | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/Documentation/devicetree/bindings/net/mscc-ocelot.txt b/Documentation/devicetree/bindings/net/mscc-ocelot.txt
index 9e5c17d426ce..4d05a3b0f786 100644
--- a/Documentation/devicetree/bindings/net/mscc-ocelot.txt
+++ b/Documentation/devicetree/bindings/net/mscc-ocelot.txt
@@ -12,6 +12,7 @@ Required properties:
- "sys"
- "rew"
- "qs"
+ - "ptp" (optional due to backward compatibility)
- "qsys"
- "ana"
- "portX" with X from 0 to the number of last port index available on that
@@ -44,6 +45,7 @@ Example:
reg = <0x1010000 0x10000>,
<0x1030000 0x10000>,
<0x1080000 0x100>,
+ <0x10e0000 0x10000>,
<0x11e0000 0x100>,
<0x11f0000 0x100>,
<0x1200000 0x100>,
@@ -57,9 +59,10 @@ Example:
<0x1280000 0x100>,
<0x1800000 0x80000>,
<0x1880000 0x10000>;
- reg-names = "sys", "rew", "qs", "port0", "port1", "port2",
- "port3", "port4", "port5", "port6", "port7",
- "port8", "port9", "port10", "qsys", "ana";
+ reg-names = "sys", "rew", "qs", "ptp", "port0", "port1",
+ "port2", "port3", "port4", "port5", "port6",
+ "port7", "port8", "port9", "port10", "qsys",
+ "ana";
interrupts = <21 22>;
interrupt-names = "xtr", "inj";
--
2.21.0
^ permalink raw reply related
* [PATCH] iproute2: devlink: use sys/queue.h from libbsd as a fallback
From: Sergei Trofimovich @ 2019-07-24 8:18 UTC (permalink / raw)
To: netdev; +Cc: Sergei Trofimovich, Stephen Hemminger
On sys/queue.h does not exist linux-musl targets and
fails build as:
devlink.c:28:10: fatal error: sys/queue.h: No such file or directory
28 | #include <sys/queue.h>
| ^~~~~~~~~~~~~
The change pulls in 'sys/queue.h' from libbsd in case
system headers don't already provides it.
Tested on linux-musl and linux-glibc.
Bug: https://bugs.gentoo.org/690486
CC: Stephen Hemminger <stephen@networkplumber.org>
CC: netdev@vger.kernel.org
Signed-off-by: Sergei Trofimovich <slyfox@gentoo.org>
---
configure | 30 ++++++++++++++++++++++++++++++
devlink/devlink.c | 9 ++++++++-
2 files changed, 38 insertions(+), 1 deletion(-)
diff --git a/configure b/configure
index 45fcffb6..a1ee946f 100755
--- a/configure
+++ b/configure
@@ -323,6 +323,33 @@ check_cap()
fi
}
+check_sys_queue()
+{
+ cat >$TMPDIR/queue_test.c <<EOF
+#include <sys/queue.h>
+struct nest_qentry {
+ int attr_type;
+ TAILQ_ENTRY(nest_qentry) nest_entries;
+};
+int main(int argc, char **argv) {
+ return 0;
+}
+EOF
+ if $CC -I$INCLUDE -o $TMPDIR/queue_test $TMPDIR/queue_test.c >/dev/null 2>&1; then
+ echo "no"
+ else
+ if ${PKG_CONFIG} libbsd --exists; then
+ echo 'CFLAGS += -DHAVE_LIBBSD_SYS_QUEUE' `${PKG_CONFIG} libbsd --cflags` >>$CONFIG
+ echo 'LDLIBS +=' `${PKG_CONFIG} libbsd --libs` >> $CONFIG
+ echo "no"
+ else
+ echo 'CFLAGS += -DNEED_SYS_QUEUE' >>$CONFIG
+ echo "yes"
+ fi
+ fi
+ rm -f $TMPDIR/queue_test.c $TMPDIR/queue_test
+}
+
quiet_config()
{
cat <<EOF
@@ -398,6 +425,9 @@ check_strlcpy
echo -n "libcap support: "
check_cap
+echo -n "need for sys/queue.h API: "
+check_sys_queue
+
echo >> $CONFIG
echo "%.o: %.c" >> $CONFIG
echo ' $(QUIET_CC)$(CC) $(CFLAGS) $(EXTRA_CFLAGS) $(CPPFLAGS) -c -o $@ $<' >> $CONFIG
diff --git a/devlink/devlink.c b/devlink/devlink.c
index bb023c0c..fd91198c 100644
--- a/devlink/devlink.c
+++ b/devlink/devlink.c
@@ -25,7 +25,14 @@
#include <linux/devlink.h>
#include <libmnl/libmnl.h>
#include <netinet/ether.h>
-#include <sys/queue.h>
+#ifdef HAVE_LIBBSD_SYS_QUEUE
+# include <bsd/sys/queue.h>
+#else
+# include <sys/queue.h>
+#endif
+#ifdef NEED_SYS_QUEUE
+# error "No <sys/queue.h> implementation found."
+#endif
#include "SNAPSHOT.h"
#include "list.h"
--
2.22.0
^ permalink raw reply related
* RE: [PATCH net-next] dpaa2-eth: Don't use netif_receive_skb_list for TCP frames
From: Ioana Ciocoi Radulescu @ 2019-07-24 8:15 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, Ioana Ciornei, Vladimir Oltean,
Eric Dumazet, Edward Cree
In-Reply-To: <20190723.140255.1785812525450069326.davem@davemloft.net>
> -----Original Message-----
> From: David Miller <davem@davemloft.net>
> Sent: Wednesday, July 24, 2019 12:03 AM
> To: Ioana Ciocoi Radulescu <ruxandra.radulescu@nxp.com>
> Cc: netdev@vger.kernel.org; Ioana Ciornei <ioana.ciornei@nxp.com>; Vladimir
> Oltean <vladimir.oltean@nxp.com>
> Subject: Re: [PATCH net-next] dpaa2-eth: Don't use netif_receive_skb_list for
> TCP frames
>
> From: Ioana Radulescu <ruxandra.radulescu@nxp.com>
> Date: Tue, 23 Jul 2019 20:28:43 +0300
>
> > Using Rx skb bulking for all frames may negatively impact the
> > performance in some TCP termination scenarios, as it effectively
> > bypasses GRO.
>
> "may"?
>
> Please provide numbers so that we know exactly whether it actually
> hurts performance one way or another.
We observed the worst degradation running netperf TCP_STREAM on a
setup with two 16cores LX2160A boards connected back-to-back, one
stream per cpu:
With netif_receive_skb_list() on all packets, we get a total bandwidth
of 41.6Gbps, with rx cpu load of 97%; after applying current patch, bw
increases to 45.8Gbps with rx cpu at 64%, which is similar to what we
got before using netif_receive_skb_list() in the first place.
On other platforms/setups the impact is lower, in some cases there is
no difference in throughput, only higher cpu load when skb batching
is used for TCP packets.
Anyway, based on feedback so far I guess the best path forward is to
withdraw this patch and wait for Edward's GRO batching work to be
accepted. I can help test those patches if needed.
Thanks,
Ioana
^ permalink raw reply
* Re: [PATCH net-next v2 0/8] Use dev_get_drvdata where possible
From: Chuhong Yuan @ 2019-07-24 8:13 UTC (permalink / raw)
Cc: Steffen Klassert, David S . Miller, Jay Cliburn, Chris Snook,
Rasesh Mody, Michael Chan, Siva Reddy Kallam, Prashant Sreedharan,
GR-Linux-NIC-Dev, Jeff Kirsher, Guo-Fu Tseng, intel-wired-lan,
netdev, linux-kernel
In-Reply-To: <20190724060512.23899-1-hslester96@gmail.com>
On Wed, Jul 24, 2019 at 2:05 PM Chuhong Yuan <hslester96@gmail.com> wrote:
>
> These patches use dev_get_drvdata instead of
> using to_pci_dev + pci_get_drvdata to make
> code simpler where possible.
>
> Changelog:
>
> v1 -> v2:
> - Change pci_set_drvdata to dev_set_drvdata
> to keep consistency.
>
Hi all,
I checked the cases which mentioned the consistency
of get/set_drvdata usages.
The cases' commit IDs are
488d040e3a3452a0dceef5d3ec4f61942262f57f
b77c98780e682fe780d899b91543769d4cf94585
After checking, I think that the consistency problem
refers to inconsistency between probe and remove.
But the changes of these patches are not related
to probe and remove.
So I think the previously sent and applied v1 patches
which do not change pci_set_drvdata to dev_set_drvdata
are okay.
Therefore there may be no need to use these v2 patches.
Regards,
Chuhong
> Chuhong Yuan (8):
> net: 3com: 3c59x: Use dev_get_drvdata
> net: atheros: Use dev_get_drvdata
> net: broadcom: Use dev_get_drvdata
> e1000e: Use dev_get_drvdata where possible
> fm10k: Use dev_get_drvdata
> i40e: Use dev_get_drvdata
> igb: Use dev_get_drvdata where possible
> net: jme: Use dev_get_drvdata
>
> drivers/net/ethernet/3com/3c59x.c | 8 +++-----
> drivers/net/ethernet/atheros/alx/main.c | 8 +++-----
> drivers/net/ethernet/atheros/atl1c/atl1c_main.c | 10 ++++------
> drivers/net/ethernet/atheros/atlx/atl1.c | 8 +++-----
> drivers/net/ethernet/broadcom/bnx2.c | 8 +++-----
> drivers/net/ethernet/broadcom/bnxt/bnxt.c | 8 +++-----
> drivers/net/ethernet/broadcom/tg3.c | 8 +++-----
> drivers/net/ethernet/intel/e1000e/netdev.c | 9 ++++-----
> drivers/net/ethernet/intel/fm10k/fm10k_pci.c | 6 +++---
> drivers/net/ethernet/intel/i40e/i40e_main.c | 10 ++++------
> drivers/net/ethernet/intel/igb/igb_main.c | 5 ++---
> drivers/net/ethernet/jme.c | 8 +++-----
> 12 files changed, 38 insertions(+), 58 deletions(-)
>
> --
> 2.20.1
>
^ permalink raw reply
* Re: [RFC PATCH net-next 00/12] drop_monitor: Capture dropped packets and metadata
From: Ido Schimmel @ 2019-07-24 8:10 UTC (permalink / raw)
To: Toke Høiland-Jørgensen
Cc: netdev, davem, nhorman, dsahern, roopa, nikolay, jakub.kicinski,
andy, f.fainelli, andrew, vivien.didelot, mlxsw, Ido Schimmel
In-Reply-To: <87ftmw3f9m.fsf@toke.dk>
On Tue, Jul 23, 2019 at 06:08:21PM +0200, Toke Høiland-Jørgensen wrote:
> Also, presumably the queue will have to change from a struct
> sk_buff_head to something that can hold XDP frames and whatever devlink
> puts there as well, right?
Good point!
For HW drops we get an SKB and relevant metadata from devlink about why
the packet was dropped etc. I plan to store a pointer to this metadata
in the SKB control block.
Let me see how the implementation goes. Even if use sk_buff_head for
now, I will make sure that converting to a more generalized data
structure is straightforward.
^ permalink raw reply
* Re: [PATCH 07/12] vhost-scsi: convert put_page() to put_user_page*()
From: Michael S. Tsirkin @ 2019-07-24 8:07 UTC (permalink / raw)
To: john.hubbard
Cc: Andrew Morton, Alexander Viro, Anna Schumaker, David S . Miller,
Dominique Martinet, Eric Van Hensbergen, Jason Gunthorpe,
Jason Wang, Jens Axboe, Latchesar Ionkov, Miklos Szeredi,
Trond Myklebust, Christoph Hellwig, Matthew Wilcox, linux-mm,
LKML, ceph-devel, kvm, linux-block, linux-cifs, linux-fsdevel,
linux-nfs, linux-rdma, netdev, samba-technical, v9fs-developer,
virtualization, Jérôme Glisse, John Hubbard, Jan Kara,
Dan Williams, Johannes Thumshirn, Ming Lei, Dave Chinner,
Boaz Harrosh, Paolo Bonzini, Stefan Hajnoczi
In-Reply-To: <20190724042518.14363-8-jhubbard@nvidia.com>
On Tue, Jul 23, 2019 at 09:25:13PM -0700, john.hubbard@gmail.com wrote:
> From: Jérôme Glisse <jglisse@redhat.com>
>
> For pages that were retained via get_user_pages*(), release those pages
> via the new put_user_page*() routines, instead of via put_page().
>
> This is part a tree-wide conversion, as described in commit fc1d8e7cca2d
> ("mm: introduce put_user_page*(), placeholder versions").
>
> Changes from Jérôme's original patch:
>
> * Changed a WARN_ON to a BUG_ON.
>
> Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
> Signed-off-by: John Hubbard <jhubbard@nvidia.com>
> Cc: virtualization@lists.linux-foundation.org
> Cc: linux-fsdevel@vger.kernel.org
> Cc: linux-block@vger.kernel.org
> Cc: linux-mm@kvack.org
> Cc: Jan Kara <jack@suse.cz>
> Cc: Dan Williams <dan.j.williams@intel.com>
> Cc: Alexander Viro <viro@zeniv.linux.org.uk>
> Cc: Johannes Thumshirn <jthumshirn@suse.de>
> Cc: Christoph Hellwig <hch@lst.de>
> Cc: Jens Axboe <axboe@kernel.dk>
> Cc: Ming Lei <ming.lei@redhat.com>
> Cc: Dave Chinner <david@fromorbit.com>
> Cc: Jason Gunthorpe <jgg@ziepe.ca>
> Cc: Matthew Wilcox <willy@infradead.org>
> Cc: Boaz Harrosh <boaz@plexistor.com>
> Cc: Miklos Szeredi <miklos@szeredi.hu>
> Cc: "Michael S. Tsirkin" <mst@redhat.com>
> Cc: Jason Wang <jasowang@redhat.com>
> Cc: Paolo Bonzini <pbonzini@redhat.com>
> Cc: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
> ---
> drivers/vhost/scsi.c | 13 ++++++++++---
> 1 file changed, 10 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/vhost/scsi.c b/drivers/vhost/scsi.c
> index a9caf1bc3c3e..282565ab5e3f 100644
> --- a/drivers/vhost/scsi.c
> +++ b/drivers/vhost/scsi.c
> @@ -329,11 +329,11 @@ static void vhost_scsi_release_cmd(struct se_cmd *se_cmd)
>
> if (tv_cmd->tvc_sgl_count) {
> for (i = 0; i < tv_cmd->tvc_sgl_count; i++)
> - put_page(sg_page(&tv_cmd->tvc_sgl[i]));
> + put_user_page(sg_page(&tv_cmd->tvc_sgl[i]));
> }
> if (tv_cmd->tvc_prot_sgl_count) {
> for (i = 0; i < tv_cmd->tvc_prot_sgl_count; i++)
> - put_page(sg_page(&tv_cmd->tvc_prot_sgl[i]));
> + put_user_page(sg_page(&tv_cmd->tvc_prot_sgl[i]));
> }
>
> vhost_scsi_put_inflight(tv_cmd->inflight);
> @@ -630,6 +630,13 @@ vhost_scsi_map_to_sgl(struct vhost_scsi_cmd *cmd,
> size_t offset;
> unsigned int npages = 0;
>
> + /*
> + * Here in all cases we should have an IOVEC which use GUP. If that is
> + * not the case then we will wrongly call put_user_page() and the page
> + * refcount will go wrong (this is in vhost_scsi_release_cmd())
> + */
> + WARN_ON(!iov_iter_get_pages_use_gup(iter));
> +
> bytes = iov_iter_get_pages(iter, pages, LONG_MAX,
> VHOST_SCSI_PREALLOC_UPAGES, &offset);
> /* No pages were pinned */
> @@ -681,7 +688,7 @@ vhost_scsi_iov_to_sgl(struct vhost_scsi_cmd *cmd, bool write,
> while (p < sg) {
> struct page *page = sg_page(p++);
> if (page)
> - put_page(page);
> + put_user_page(page);
> }
> return ret;
> }
> --
> 2.22.0
^ permalink raw reply
* Re: Reminder: 3 open syzbot bugs in "net/ax25" subsystem
From: Dmitry Vyukov @ 2019-07-24 8:01 UTC (permalink / raw)
To: Joe; +Cc: linux-hams, netdev, Ralf Baechle, David S. Miller, LKML,
syzkaller-bugs
In-Reply-To: <623ff230-5883-560c-22d0-3e6b0eeaba39@mcn.org>
On Wed, Jul 24, 2019 at 5:42 AM Joe <joego@mcn.org> wrote:
>
> Hi Eric, How do I get off of this thread? When I try to unsubscribe it
> tells me I'm not a member of the group.
Hi Joe,
If you received it via netdev or linux-hams mailing lists, here are
instructions on how to unsubscribe:
http://vger.kernel.org/vger-lists.html#netdev
http://vger.kernel.org/vger-lists.html#linux-hams
> On 7/23/19 7:40 PM, Eric Biggers wrote:
> > [This email was generated by a script. Let me know if you have any suggestions
> > to make it better, or if you want it re-generated with the latest status.]
> >
> > Of the currently open syzbot reports against the upstream kernel, I've manually
> > marked 3 of them as possibly being bugs in the "net/ax25" subsystem. I've
> > listed these reports below, sorted by an algorithm that tries to list first the
> > reports most likely to be still valid, important, and actionable.
> >
> > If you believe a bug is no longer valid, please close the syzbot report by
> > sending a '#syz fix', '#syz dup', or '#syz invalid' command in reply to the
> > original thread, as explained at https://goo.gl/tpsmEJ#status
> >
> > If you believe I misattributed a bug to the "net/ax25" subsystem, please let me
> > know, and if possible forward the report to the correct people or mailing list.
> >
> > Here are the bugs:
> >
> > --------------------------------------------------------------------------------
> > Title: general protection fault in ax25_send_frame
> > Last occurred: 0 days ago
> > Reported: 204 days ago
> > Branches: Mainline and others
> > Dashboard link: https://syzkaller.appspot.com/bug?id=1cdd5b120f129364fc8e9b2b027826cf99fa696e
> > Original thread: https://lkml.kernel.org/lkml/0000000000009ea37c057e58d787@google.com/T/#u
> >
> > Unfortunately, this bug does not have a reproducer.
> >
> > No one replied to the original thread for this bug.
> >
> > If you fix this bug, please add the following tag to the commit:
> > Reported-by: syzbot+e0b81535a27b8be39502@syzkaller.appspotmail.com
> >
> > If you send any email or patch for this bug, please consider replying to the
> > original thread. For the git send-email command to use, or tips on how to reply
> > if the thread isn't in your mailbox, see the "Reply instructions" at
> > https://lkml.kernel.org/r/0000000000009ea37c057e58d787@google.com
> >
> > --------------------------------------------------------------------------------
> > Title: KASAN: stack-out-of-bounds Write in ax25_getname
> > Last occurred: 90 days ago
> > Reported: 206 days ago
> > Branches: Mainline and others
> > Dashboard link: https://syzkaller.appspot.com/bug?id=fb195f91dc044978c1b186f1288b1eff61edcc20
> > Original thread: https://lkml.kernel.org/lkml/000000000000ed4120057e2df0c6@google.com/T/#u
> >
> > This bug has a syzkaller reproducer only.
> >
> > No one replied to the original thread for this bug.
> >
> > If you fix this bug, please add the following tag to the commit:
> > Reported-by: syzbot+6a29097222b4d3b8617c@syzkaller.appspotmail.com
> >
> > If you send any email or patch for this bug, please consider replying to the
> > original thread. For the git send-email command to use, or tips on how to reply
> > if the thread isn't in your mailbox, see the "Reply instructions" at
> > https://lkml.kernel.org/r/000000000000ed4120057e2df0c6@google.com
> >
> > --------------------------------------------------------------------------------
> > Title: inconsistent lock state in ax25_std_heartbeat_expiry
> > Last occurred: 122 days ago
> > Reported: 120 days ago
> > Branches: net
> > Dashboard link: https://syzkaller.appspot.com/bug?id=9086a8eac930890b2730d6441093bd478e32913f
> > Original thread: https://lkml.kernel.org/lkml/0000000000001b07250584efbee3@google.com/T/#u
> >
> > Unfortunately, this bug does not have a reproducer.
> >
> > The original thread for this bug received 2 replies; the last was 119 days ago.
> >
> > If you fix this bug, please add the following tag to the commit:
> > Reported-by: syzbot+e350b81e95a6a214da8a@syzkaller.appspotmail.com
> >
> > If you send any email or patch for this bug, please consider replying to the
> > original thread. For the git send-email command to use, or tips on how to reply
> > if the thread isn't in your mailbox, see the "Reply instructions" at
> > https://lkml.kernel.org/r/0000000000001b07250584efbee3@google.com
> >
>
> --
> You received this message because you are subscribed to the Google Groups "syzkaller-bugs" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller-bugs+unsubscribe@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/syzkaller-bugs/623ff230-5883-560c-22d0-3e6b0eeaba39%40mcn.org.
^ permalink raw reply
* Re: [RFC PATCH net-next 00/12] drop_monitor: Capture dropped packets and metadata
From: Ido Schimmel @ 2019-07-24 7:57 UTC (permalink / raw)
To: David Ahern
Cc: Toke Høiland-Jørgensen, netdev, davem, nhorman, roopa,
nikolay, jakub.kicinski, andy, f.fainelli, andrew, vivien.didelot,
mlxsw, Ido Schimmel
In-Reply-To: <c02f9b6a-f343-89ee-1047-79c1fb4e3436@gmail.com>
On Tue, Jul 23, 2019 at 08:47:57AM -0700, David Ahern wrote:
> On 7/23/19 8:14 AM, Ido Schimmel wrote:
> > On Tue, Jul 23, 2019 at 02:17:49PM +0200, Toke Høiland-Jørgensen wrote:
> >> Ido Schimmel <idosch@idosch.org> writes:
> >>
> >>> On Mon, Jul 22, 2019 at 09:43:15PM +0200, Toke Høiland-Jørgensen wrote:
> >>>> Is there a mechanism for the user to filter the packets before they are
> >>>> sent to userspace? A bpf filter would be the obvious choice I guess...
> >>>
> >>> Hi Toke,
> >>>
> >>> Yes, it's on my TODO list to write an eBPF program that only lets
> >>> "unique" packets to be enqueued on the netlink socket. Where "unique" is
> >>> defined as {5-tuple, PC}. The rest of the copies will be counted in an
> >>> eBPF map, which is just a hash table keyed by {5-tuple, PC}.
> >>
> >> Yeah, that's a good idea. Or even something simpler like tcpdump-style
> >> filters for the packets returned by drop monitor (say if I'm just trying
> >> to figure out what happens to my HTTP requests).
> >
> > Yep, that's a good idea. I guess different users will use different
> > programs. Will look into both options.
>
> Perhaps I am missing something, but the dropmon code only allows a
> single user at the moment (in my attempts to run 2 instances the second
> one failed).
Yes, you're correct. By "different users" I meant users on different
systems with different needs. For example, someone trying to monitor
dropped packets on a laptop versus someone trying to do the same on a
ToR switch.
> If that part stays with the design
This stays the same.
> it afford better options for the design. e.g., attributes that control
> the enqueued packets when the event occurs as opposed to bpf filters
> which run much later when the message is enqueued to the socket.
I'm going to add an attribute that will control the number of packets
we're enqueuing on the per-CPU drop list. I'm not sure, but are you
suggesting to add even more attributes? If so, how do you imagine these
will look like?
>
> >
> >>> I think it would be good to have the program as part of the bcc
> >>> repository [1]. What do you think?
> >>
> >> Sure. We could also add it to the XDP tutorial[2]; it could go into a
> >> section on introspection and debugging (just added a TODO about that[3]).
> >
> > Great!
> >
> >>>> For integrating with XDP the trick would be to find a way to do it that
> >>>> doesn't incur any overhead when it's not enabled. Are you envisioning
> >>>> that this would be enabled separately for the different "modes" (kernel,
> >>>> hardware, XDP, etc)?
> >>>
> >>> Yes. Drop monitor have commands to enable and disable tracing, but they
> >>> don't carry any attributes at the moment. My plan is to add an attribute
> >>> (e.g., 'NET_DM_ATTR_DROP_TYPE') that will specify the type of drops
> >>> you're interested in - SW/HW/XDP. If the attribute is not specified,
> >>> then current behavior is maintained and all the drop types are traced.
> >>> But if you're only interested in SW drops, then overhead for the rest
> >>> should be zero.
> >>
> >> Makes sense (although "should be" is the key here ;)).
>
> static_key is used in other parts of the packet fast path.
>
> Toke/Jesper: Any reason to believe it is too much overhead for this path?
>
> >>
> >> I'm also worried about the drop monitor getting overwhelmed; if you turn
> >> it on for XDP and you're running a filtering program there, you'll
> >> suddenly get *a lot* of drops.
> >>
> >> As I read your patch, the current code can basically queue up an
> >> unbounded number of packets waiting to go out over netlink, can't it?
> >
> > That's a very good point. Each CPU holds a drop list. It probably makes
> > sense to limit it by default (to 1000?) and allow user to change it
> > later, if needed. I can expose a counter that shows how many packets
> > were dropped because of this limit. It can be used as an indication to
> > adjust the queue length (or flip to 'summary' mode).
> >
>
> And then with a single user limit, you can have an attribute that
> controls the backlog.
Yep, already on my list of changes for v1 :)
Thanks, David.
^ permalink raw reply
* Re: [PATCH] net: sctp: fix memory leak in sctp_send_reset_streams
From: Xin Long @ 2019-07-24 7:56 UTC (permalink / raw)
To: Neil Horman
Cc: Hillf Danton, linux-sctp, network dev, syzkaller,
David S . Miller, LKML, syzkaller-bugs, syzbot,
Marcelo Ricardo Leitner, Vlad Yasevich, Eric Dumazet
In-Reply-To: <CADvbK_dUDjK3UAF49uo+DZv+QiuEsaMmZeqDwBJ0suRwu4yXJw@mail.gmail.com>
On Sun, Jun 2, 2019 at 9:36 PM Xin Long <lucien.xin@gmail.com> wrote:
>
> On Sun, Jun 2, 2019 at 6:52 PM Neil Horman <nhorman@tuxdriver.com> wrote:
> >
> > On Sun, Jun 02, 2019 at 11:44:29AM +0800, Hillf Danton wrote:
> > >
> > > syzbot found the following crash on:
> > >
> > > HEAD commit: 036e3431 Merge git://git.kernel.org/pub/scm/linux/kernel/g..
> > > git tree: upstream
> > > console output: https://syzkaller.appspot.com/x/log.txt?x=153cff12a00000
> > > kernel config: https://syzkaller.appspot.com/x/.config?x=8f0f63a62bb5b13c
> > > dashboard link: https://syzkaller.appspot.com/bug?extid=6ad9c3bd0a218a2ab41d
> > > compiler: gcc (GCC) 9.0.0 20181231 (experimental)
> > > syz repro: https://syzkaller.appspot.com/x/repro.syz?x=12561c86a00000
> > > C reproducer: https://syzkaller.appspot.com/x/repro.c?x=15b76fd8a00000
> > >
> > > executing program
> > > executing program
> > > executing program
> > > executing program
> > > executing program
> > > BUG: memory leak
> > > unreferenced object 0xffff888123894820 (size 32):
> > > comm "syz-executor045", pid 7267, jiffies 4294943559 (age 13.660s)
> > > hex dump (first 32 bytes):
> > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > > 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> > > backtrace:
> > > [<00000000c7e71c69>] kmemleak_alloc_recursive
> > > include/linux/kmemleak.h:55 [inline]
> > > [<00000000c7e71c69>] slab_post_alloc_hook mm/slab.h:439 [inline]
> > > [<00000000c7e71c69>] slab_alloc mm/slab.c:3326 [inline]
> > > [<00000000c7e71c69>] __do_kmalloc mm/slab.c:3658 [inline]
> > > [<00000000c7e71c69>] __kmalloc+0x161/0x2c0 mm/slab.c:3669
> > > [<000000003250ed8e>] kmalloc_array include/linux/slab.h:670 [inline]
> > > [<000000003250ed8e>] kcalloc include/linux/slab.h:681 [inline]
> > > [<000000003250ed8e>] sctp_send_reset_streams+0x1ab/0x5a0 net/sctp/stream.c:302
> > > [<00000000cd899c6e>] sctp_setsockopt_reset_streams net/sctp/socket.c:4314 [inline]
> > > [<00000000cd899c6e>] sctp_setsockopt net/sctp/socket.c:4765 [inline]
> > > [<00000000cd899c6e>] sctp_setsockopt+0xc23/0x2bf0 net/sctp/socket.c:4608
> > > [<00000000ff3a21a2>] sock_common_setsockopt+0x38/0x50 net/core/sock.c:3130
> > > [<000000009eb87ae7>] __sys_setsockopt+0x98/0x120 net/socket.c:2078
> > > [<00000000e0ede6ca>] __do_sys_setsockopt net/socket.c:2089 [inline]
> > > [<00000000e0ede6ca>] __se_sys_setsockopt net/socket.c:2086 [inline]
> > > [<00000000e0ede6ca>] __x64_sys_setsockopt+0x26/0x30 net/socket.c:2086
> > > [<00000000c61155f5>] do_syscall_64+0x76/0x1a0 arch/x86/entry/common.c:301
> > > [<00000000e540958c>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
> > >
> > >
> > > It was introduced in commit d570a59c5b5f ("sctp: only allow the out stream
> > > reset when the stream outq is empty"), in orde to check stream outqs before
> > > sending SCTP_STRRESET_IN_PROGRESS back to the peer of the stream. EAGAIN is
> > > returned, however, without the nstr_list slab released, if any outq is found
> > > to be non empty.
> > >
> > > Freeing the slab in question before bailing out fixes it.
> > >
> > > Fixes: d570a59c5b5f ("sctp: only allow the out stream reset when the stream outq is empty")
> > > Reported-by: syzbot <syzbot+6ad9c3bd0a218a2ab41d@syzkaller.appspotmail.com>
> > > Reported-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> > > Tested-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> > > Cc: Xin Long <lucien.xin@gmail.com>
> > > Cc: Neil Horman <nhorman@tuxdriver.com>
> > > Cc: Vlad Yasevich <vyasevich@gmail.com>
> > > Cc: Eric Dumazet <edumazet@google.com>
> > > Signed-off-by: Hillf Danton <hdanton@sina.com>
> > > ---
> > > net/sctp/stream.c | 1 +
> > > 1 file changed, 1 insertion(+)
> > >
> > > diff --git a/net/sctp/stream.c b/net/sctp/stream.c
> > > index 93ed078..d3e2f03 100644
> > > --- a/net/sctp/stream.c
> > > +++ b/net/sctp/stream.c
> > > @@ -310,6 +310,7 @@ int sctp_send_reset_streams(struct sctp_association *asoc,
> > >
> > > if (out && !sctp_stream_outq_is_empty(stream, str_nums, nstr_list)) {
> > > retval = -EAGAIN;
> > > + kfree(nstr_list);
> > > goto out;
> > > }
> > >
> > > --
> > >
> > >
> > Acked-by: Neil Horman <nhorman@tuxdriver.com>
> Reviewed-by: Xin Long <lucien.xin@gmail.com>
This fix is not applied, pls resend it with:
to = network dev <netdev@vger.kernel.org>
cc = davem@davemloft.net
to = linux-sctp@vger.kernel.org
cc = Neil Horman <nhorman@tuxdriver.com>
cc = Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
^ permalink raw reply
* Re: [RFC PATCH 1/2] gianfar: convert to phylink
From: Arseny Solokha @ 2019-07-24 7:36 UTC (permalink / raw)
To: Claudiu Manoil
Cc: Ioana Ciornei, Russell King, Andrew Lunn, netdev@vger.kernel.org
In-Reply-To: <VI1PR04MB48809AFBB9DF01001AA5E2CA96C70@VI1PR04MB4880.eurprd04.prod.outlook.com>
>>-----Original Message-----
>>From: Arseny Solokha <asolokha@kb.kras.ru>
>>Sent: Tuesday, July 23, 2019 6:17 PM
>>To: Claudiu Manoil <claudiu.manoil@nxp.com>; Ioana Ciornei
>><ioana.ciornei@nxp.com>; Russell King <linux@armlinux.org.uk>; Andrew Lunn
>><andrew@lunn.ch>
>>Cc: netdev@vger.kernel.org; Arseny Solokha <asolokha@kb.kras.ru>
>>Subject: [RFC PATCH 1/2] gianfar: convert to phylink
>>
>>Convert gianfar to use the phylink API for better SFP modules support.
>>
>>The driver still uses phylib for serdes configuration over the TBI
>>interface, as there seems to be no functionally equivalent API present
>>in phylink (yet). phylib usage is basically confined in two functions.
>>
>
> Thanks for your patch. Phylink in gianfar... that would be something!
> At first glance a lot of code has changed with this patch or got relocated.
> To make it easier to swallow, I think a few cleanup patches could be
> separated before migrating to phylink. Like for instance getting rid of the
> old* link state variables, which I think are an artifact from early phylib usage.
> Nonetheless good to see this implemented, I'll have a closer look asap.
Hi,
meanwhile I'll have to post v2 of this patch because it has some issues which
initially escaped my attention. For now I'm pasting the diff against v1 here for
reference:
--- a/drivers/net/ethernet/freescale/gianfar.c
+++ b/drivers/net/ethernet/freescale/gianfar.c
@@ -1666,7 +1666,7 @@ static int gfar_suspend(struct device *dev)
gfar_start_wol_filer(priv);
} else {
- phylink_stop(phy->phylink);
+ phylink_stop(priv->phylink);
}
priv->speed = SPEED_UNKNOWN;
@@ -3699,9 +3699,6 @@ static void gfar_mac_config(struct phylink_config *config, unsigned int mode,
if (unlikely(test_bit(GFAR_RESETTING, &priv->state)))
return;
- if (unlikely(phylink_autoneg_inband(mode)))
- return;
-
maccfg1 = gfar_read(®s->maccfg1);
maccfg2 = gfar_read(®s->maccfg2);
ecntrl = gfar_read(®s->ecntrl);
The first hunk here fixes a typo which broke build with PM enabled. The second
one removes an early return from gfar_mac_config() which I believe is really
bogus and also breaks coalesce parameters calculation for SGMII and 1000Base-X
attached PHYs.
I'd like to submit a real v2 after the patches gets actual review, though.
Thanks,
Arseny
^ permalink raw reply
* Re: [PATCH net-next 1/4] sctp: check addr_size with sa_family_t size in __sctp_setsockopt_connectx
From: Xin Long @ 2019-07-24 7:21 UTC (permalink / raw)
To: Neil Horman; +Cc: network dev, linux-sctp, Marcelo Ricardo Leitner, davem
In-Reply-To: <20190723152449.GB8419@localhost.localdomain>
On Tue, Jul 23, 2019 at 11:25 PM Neil Horman <nhorman@tuxdriver.com> wrote:
>
> On Tue, Jul 23, 2019 at 01:37:57AM +0800, Xin Long wrote:
> > Now __sctp_connect() is called by __sctp_setsockopt_connectx() and
> > sctp_inet_connect(), the latter has done addr_size check with size
> > of sa_family_t.
> >
> > In the next patch to clean up __sctp_connect(), we will remove
> > addr_size check with size of sa_family_t from __sctp_connect()
> > for the 1st address.
> >
> > So before doing that, __sctp_setsockopt_connectx() should do
> > this check first, as sctp_inet_connect() does.
> >
> > Signed-off-by: Xin Long <lucien.xin@gmail.com>
> > ---
> > net/sctp/socket.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/net/sctp/socket.c b/net/sctp/socket.c
> > index aa80cda..5f92e4a 100644
> > --- a/net/sctp/socket.c
> > +++ b/net/sctp/socket.c
> > @@ -1311,7 +1311,7 @@ static int __sctp_setsockopt_connectx(struct sock *sk,
> > pr_debug("%s: sk:%p addrs:%p addrs_size:%d\n",
> > __func__, sk, addrs, addrs_size);
> >
> > - if (unlikely(addrs_size <= 0))
> > + if (unlikely(addrs_size < sizeof(sa_family_t)))
> I don't think this is what you want to check for here. sa_family_t is
> an unsigned short, and addrs_size is the number of bytes in the addrs
> array. The addrs array should be at least the size of one struct
> sockaddr (16 bytes iirc), and, if larger, should be a multiple of
> sizeof(struct sockaddr)
sizeof(struct sockaddr) is not the right value to check either.
The proper check will be done later in __sctp_connect():
af = sctp_get_af_specific(daddr->sa.sa_family);
if (!af || af->sockaddr_len > addrs_size)
return -EINVAL;
So the check 'addrs_size < sizeof(sa_family_t)' in this patch is
just to make sure daddr->sa.sa_family is accessible. the same
check is also done in sctp_inet_connect().
>
> Neil
>
> > return -EINVAL;
> >
> > kaddrs = memdup_user(addrs, addrs_size);
> > --
> > 2.1.0
> >
> >
^ permalink raw reply
* Re: [PATCH net-next] dpaa2-eth: Don't use netif_receive_skb_list for TCP frames
From: Eric Dumazet @ 2019-07-24 7:10 UTC (permalink / raw)
To: Ioana Radulescu, netdev, davem; +Cc: ioana.ciornei, vladimir.oltean
In-Reply-To: <1563902923-26178-1-git-send-email-ruxandra.radulescu@nxp.com>
On 7/23/19 7:28 PM, Ioana Radulescu wrote:
> Using Rx skb bulking for all frames may negatively impact the
> performance in some TCP termination scenarios, as it effectively
> bypasses GRO.
>
> - list_add_tail(&skb->list, ch->rx_list);
> + if (frame_is_tcp(fd, fas))
> + napi_gro_receive(&ch->napi, skb);
> + else
> + list_add_tail(&skb->list, ch->rx_list);
>
> return;
>
This is really bad.
This is exactly why I suggested to add the batching capability to GRO,
instead having to change all drivers.
Edward Cree is working on this.
^ permalink raw reply
* Re: [RFC PATCH net-next 10/12] drop_monitor: Add packet alert mode
From: Ido Schimmel @ 2019-07-24 7:10 UTC (permalink / raw)
To: Neil Horman
Cc: netdev, davem, dsahern, roopa, nikolay, jakub.kicinski, toke,
andy, f.fainelli, andrew, vivien.didelot, mlxsw, Ido Schimmel
In-Reply-To: <20190723151431.GA8419@localhost.localdomain>
On Tue, Jul 23, 2019 at 11:14:31AM -0400, Neil Horman wrote:
> On Tue, Jul 23, 2019 at 05:16:25PM +0300, Ido Schimmel wrote:
> > On Tue, Jul 23, 2019 at 08:43:40AM -0400, Neil Horman wrote:
> > > On Mon, Jul 22, 2019 at 09:31:32PM +0300, Ido Schimmel wrote:
> > > > +static void net_dm_packet_work(struct work_struct *work)
> > > > +{
> > > > + struct per_cpu_dm_data *data;
> > > > + struct sk_buff_head list;
> > > > + struct sk_buff *skb;
> > > > + unsigned long flags;
> > > > +
> > > > + data = container_of(work, struct per_cpu_dm_data, dm_alert_work);
> > > > +
> > > > + __skb_queue_head_init(&list);
> > > > +
> > > > + spin_lock_irqsave(&data->drop_queue.lock, flags);
> > > > + skb_queue_splice_tail_init(&data->drop_queue, &list);
> > > > + spin_unlock_irqrestore(&data->drop_queue.lock, flags);
> > > > +
> > > These functions are all executed in a per-cpu context. While theres nothing
> > > wrong with using a spinlock here, I think you can get away with just doing
> > > local_irqsave and local_irq_restore.
> >
> > Hi Neil,
> >
> > Thanks a lot for reviewing. I might be missing something, but please
> > note that this function is executed from a workqueue and therefore the
> > CPU it is running on does not have to be the same CPU to which 'data'
> > belongs to. If so, I'm not sure how I can avoid taking the spinlock, as
> > otherwise two different CPUs can modify the list concurrently.
> >
> Ah, my bad, I was under the impression that the schedule_work call for
> that particular work queue was actually a call to schedule_work_on,
> which would have affined it to a specific cpu. That said, looking at
> it, I think using schedule_work_on was my initial intent, as the work
> queue is registered per cpu. And converting it to schedule_work_on
> would allow you to reduce the spin_lock to a faster local_irqsave
Yes, this can work, but I'm not sure we can justify it. The CPU that is
dropping packets is potentially very busy processing all incoming
packets and with schedule_work_on() we force the same CPU to be used to
allocate and prepare the netlink messages. With schedule_work() the
system can choose an idle CPU and better utilize system resources. Also,
the scope of the lock is very limited and it is only ever contended by
at most two CPUs: The CPU the list belongs to and the CPU executing the
work item.
I will limit the number of skbs we can enqueue, add a counter to see how
many packets we tail drop and benchmark both approaches.
Thanks!
^ permalink raw reply
* Re: [PATCH v12 1/5] can: m_can: Create a m_can platform framework
From: Greg KH @ 2019-07-24 6:47 UTC (permalink / raw)
To: Dan Murphy; +Cc: wg, mkl, davem, linux-can, netdev, linux-kernel
In-Reply-To: <437b6371-8488-a0ff-fa68-d1fb5a81bb8b@ti.com>
On Tue, Jul 23, 2019 at 10:14:14AM -0500, Dan Murphy wrote:
> Hello
>
> On 7/10/19 7:08 AM, Dan Murphy wrote:
> > Hello
> >
> > On 6/17/19 10:09 AM, Dan Murphy wrote:
> > > Marc
> > >
> > > On 6/10/19 11:35 AM, Dan Murphy wrote:
> > > > Bump
> > > >
> > > > On 6/6/19 8:16 AM, Dan Murphy wrote:
> > > > > Marc
> > > > >
> > > > > Bump
> > > > >
> > > > > On 5/31/19 6:51 AM, Dan Murphy wrote:
> > > > > > Marc
> > > > > >
> > > > > > On 5/15/19 3:54 PM, Dan Murphy wrote:
> > > > > > > Marc
> > > > > > >
> > > > > > > On 5/9/19 11:11 AM, Dan Murphy wrote:
> > > > > > > > Create a m_can platform framework that peripheral
> > > > > > > > devices can register to and use common code and register sets.
> > > > > > > > The peripheral devices may provide read/write and configuration
> > > > > > > > support of the IP.
> > > > > > > >
> > > > > > > > Acked-by: Wolfgang Grandegger <wg@grandegger.com>
> > > > > > > > Signed-off-by: Dan Murphy <dmurphy@ti.com>
> > > > > > > > ---
> > > > > > > >
> > > > > > > > v12 - Update the m_can_read/write functions to
> > > > > > > > create a backtrace if the callback
> > > > > > > > pointer is NULL. - https://lore.kernel.org/patchwork/patch/1052302/
> > > > > > > >
> > > > > > > Is this able to be merged now?
> > > > > >
> > > > > > ping
> > >
> > > Wondering if there is anything else we need to do?
> > >
> > > The part has officially shipped and we had hoped to have driver
> > > support in Linux as part of the announcement.
> > >
> > Is this being sent in a PR for 5.3?
> >
> > Dan
> >
> Adding Greg to this thread as I have no idea what is going on with this.
Why me? What am I supposed to do here? I see no patches at all to do
anything with :(
thanks,
greg "not a miracle worker" k-h
^ permalink raw reply
* Re: Reminder: 99 open syzbot bugs in net subsystem
From: Eric Dumazet @ 2019-07-24 6:39 UTC (permalink / raw)
To: netdev, David S. Miller, Florian Westphal, Ilya Maximets,
Eric Dumazet, David Ahern, linux-kernel, syzkaller-bugs
In-Reply-To: <20190724013813.GB643@sol.localdomain>
On 7/24/19 3:38 AM, Eric Biggers wrote:
> [This email was generated by a script. Let me know if you have any suggestions
> to make it better, or if you want it re-generated with the latest status.]
>
> Of the currently open syzbot reports against the upstream kernel, I've manually
> marked 99 of them as possibly being bugs in the net subsystem. This category
> only includes the networking bugs that I couldn't assign to a more specific
> component (bpf, xfrm, bluetooth, tls, tipc, sctp, wireless, etc.). I've listed
> these reports below, sorted by an algorithm that tries to list first the reports
> most likely to be still valid, important, and actionable.
>
> Of these 99 bugs, 17 were seen in mainline in the last week.
>
> Of these 99 bugs, 4 were bisected to commits from the following people:
>
> Florian Westphal <fw@strlen.de>
> Ilya Maximets <i.maximets@samsung.com>
> Eric Dumazet <edumazet@google.com>
> David Ahern <dsahern@gmail.com>
>
> If you believe a bug is no longer valid, please close the syzbot report by
> sending a '#syz fix', '#syz dup', or '#syz invalid' command in reply to the
> original thread, as explained at https://goo.gl/tpsmEJ#status
>
> If you believe I misattributed a bug to the net subsystem, please let me know,
> and if possible forward the report to the correct people or mailing list.
>
Some of the bugs have been fixed already, before syzbot found them.
Why force human to be gentle to bots and actually replying to them ?
I usually simply wait that syzbot is finding the bug does not repro anymore,
but now if you send these emails, we will have even more pressure on us.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox