Hi Marcel,
On ma, 2014-05-05 at 19:11 -0700, Marcel Holtmann wrote:
> Hi Jukka,
>
> > Create a CoC dynamically instead of one fixed channel for communication
> > to peer devices.
> >
> > Signed-off-by: Jukka Rissanen <jukka.rissanen@xxxxxxxxxxxxxxx>
> > ---
> > include/net/bluetooth/l2cap.h | 3 +-
> > net/bluetooth/6lowpan.c | 470 ++++++++++++++++++++++++++----------------
> > net/bluetooth/6lowpan.h | 19 +-
> > net/bluetooth/l2cap_core.c | 19 +-
> > 4 files changed, 308 insertions(+), 203 deletions(-)
> >
> > diff --git a/include/net/bluetooth/l2cap.h b/include/net/bluetooth/l2cap.h
> > index 4abdcb2..ddf0b35 100644
> > --- a/include/net/bluetooth/l2cap.h
> > +++ b/include/net/bluetooth/l2cap.h
> > @@ -137,7 +137,6 @@ struct l2cap_conninfo {
> > #define L2CAP_FC_L2CAP 0x02
> > #define L2CAP_FC_CONNLESS 0x04
> > #define L2CAP_FC_A2MP 0x08
> > -#define L2CAP_FC_6LOWPAN 0x3e /* reserved and temporary value */
> >
> > /* L2CAP Control Field bit masks */
> > #define L2CAP_CTRL_SAR 0xC000
> > @@ -244,6 +243,7 @@ struct l2cap_conn_rsp {
> > #define L2CAP_PSM_SDP 0x0001
> > #define L2CAP_PSM_RFCOMM 0x0003
> > #define L2CAP_PSM_3DSP 0x0021
> > +#define L2CAP_PSM_6LOWPAN 0x003e
>
> since we have no officially assigned value, please mark this one as temporary as well. It might be actually a good idea to allow providing this value via debugfs.
Sure, via debugfs it is then.
>
> >
> > /* channel identifier */
> > #define L2CAP_CID_SIGNALING 0x0001
> > @@ -626,6 +626,7 @@ struct l2cap_conn {
> >
> > struct sk_buff_head pending_rx;
> > struct work_struct pending_rx_work;
> > + struct work_struct pending_6lowpan_work;
>
> Please start thinking on how we can reach the level that bluetooth_6lowpan.ko can be its own module. So all 6loWPAN specific pieces should be externally to l2cap_conn.
The 6lowpan work_struct was needed so that we do not call 6lowpan stuff
with locking active, so this should probably still be here even if we
have a separate 6lowpan module?
>
> >
> > __u8 disc_reason;
> >
> > diff --git a/net/bluetooth/6lowpan.c b/net/bluetooth/6lowpan.c
> > index adb3ea0..f885836 100644
> > --- a/net/bluetooth/6lowpan.c
> > +++ b/net/bluetooth/6lowpan.c
> > @@ -1,5 +1,5 @@
> > /*
> > - Copyright (c) 2013 Intel Corp.
> > + Copyright (c) 2014 Intel Corp.
>
> Make it 2013-2014. the 2013 copyright is not magically going away ;)
>
> >
> > This program is free software; you can redistribute it and/or modify
> > it under the terms of the GNU General Public License version 2 and
> > @@ -29,12 +29,14 @@
> >
> > #include "../ieee802154/6lowpan.h" /* for the compression support */
> >
> > +static struct l2cap_chan *chan_create(struct l2cap_conn *conn);
> > +
> > #define IFACE_NAME_TEMPLATE "bt%d"
> > #define EUI64_ADDR_LEN 8
> >
> > struct skb_cb {
> > struct in6_addr addr;
> > - struct l2cap_conn *conn;
> > + struct l2cap_chan *chan;
> > };
> > #define lowpan_cb(skb) ((struct skb_cb *)((skb)->cb))
> >
> > @@ -50,7 +52,7 @@ static DEFINE_RWLOCK(devices_lock);
> >
> > struct lowpan_peer {
> > struct list_head list;
> > - struct l2cap_conn *conn;
> > + struct l2cap_chan *chan;
> >
> > /* peer addresses in various formats */
> > unsigned char eui64_addr[EUI64_ADDR_LEN];
> > @@ -102,12 +104,26 @@ static inline struct lowpan_peer *peer_lookup_ba(struct lowpan_dev *dev,
> >
> > list_for_each_entry_safe(peer, tmp, &dev->peers, list) {
> > BT_DBG("addr %pMR type %d",
> > - &peer->conn->hcon->dst, peer->conn->hcon->dst_type);
> > + &peer->chan->conn->hcon->dst,
> > + peer->chan->conn->hcon->dst_type);
> >
> > - if (bacmp(&peer->conn->hcon->dst, ba))
> > + if (bacmp(&peer->chan->conn->hcon->dst, ba))
> > continue;
> >
> > - if (type == peer->conn->hcon->dst_type)
> > + if (type == peer->chan->conn->hcon->dst_type)
> > + return peer;
> > + }
>
> I wonder if we need simple accessor helpers for BD_ADDR and BD_ADDR type. The dereference chain looks scary.
Sure
>
> > +
> > + return NULL;
> > +}
> > +
> > +static inline struct lowpan_peer *peer_lookup_chan(struct lowpan_dev *dev,
> > + struct l2cap_chan *chan)
> > +{
> > + struct lowpan_peer *peer, *tmp;
> > +
> > + list_for_each_entry_safe(peer, tmp, &dev->peers, list) {
> > + if (peer->chan == chan)
> > return peer;
> > }
> >
> > @@ -120,7 +136,7 @@ static inline struct lowpan_peer *peer_lookup_conn(struct lowpan_dev *dev,
> > struct lowpan_peer *peer, *tmp;
> >
> > list_for_each_entry_safe(peer, tmp, &dev->peers, list) {
> > - if (peer->conn == conn)
> > + if (peer->chan->conn == conn)
> > return peer;
> > }
> >
> > @@ -185,7 +201,7 @@ static int give_skb_to_upper(struct sk_buff *skb, struct net_device *dev)
> > }
> >
> > static int process_data(struct sk_buff *skb, struct net_device *netdev,
> > - struct l2cap_conn *conn)
> > + struct l2cap_chan *chan)
> > {
> > const u8 *saddr, *daddr;
> > u8 iphc0, iphc1;
> > @@ -196,7 +212,7 @@ static int process_data(struct sk_buff *skb, struct net_device *netdev,
> > dev = lowpan_dev(netdev);
> >
> > read_lock_irqsave(&devices_lock, flags);
> > - peer = peer_lookup_conn(dev, conn);
> > + peer = peer_lookup_chan(dev, chan);
> > read_unlock_irqrestore(&devices_lock, flags);
> > if (!peer)
> > goto drop;
> > @@ -225,7 +241,7 @@ drop:
> > }
> >
> > static int recv_pkt(struct sk_buff *skb, struct net_device *dev,
> > - struct l2cap_conn *conn)
> > + struct l2cap_chan *chan)
> > {
> > struct sk_buff *local_skb;
> > int ret;
> > @@ -269,7 +285,7 @@ static int recv_pkt(struct sk_buff *skb, struct net_device *dev,
> > if (!local_skb)
> > goto drop;
> >
> > - ret = process_data(local_skb, dev, conn);
> > + ret = process_data(local_skb, dev, chan);
> > if (ret != NET_RX_SUCCESS)
> > goto drop;
> >
> > @@ -291,135 +307,27 @@ drop:
> > }
> >
> > /* Packet from BT LE device */
> > -int bt_6lowpan_recv(struct l2cap_conn *conn, struct sk_buff *skb)
> > +static int bt_6lowpan_chan_recv_cb(struct l2cap_chan *chan, struct sk_buff *skb)
> > {
> > struct lowpan_dev *dev;
> > struct lowpan_peer *peer;
> > int err;
> >
> > - peer = lookup_peer(conn);
> > + peer = lookup_peer(chan->conn);
> > if (!peer)
> > return -ENOENT;
> >
> > - dev = lookup_dev(conn);
> > + dev = lookup_dev(chan->conn);
> > if (!dev || !dev->netdev)
> > return -ENOENT;
> >
> > - err = recv_pkt(skb, dev->netdev, conn);
> > + err = recv_pkt(skb, dev->netdev, chan);
> > +
> > BT_DBG("recv pkt %d", err);
> >
> > return err;
> > }
> >
> > -static inline int skbuff_copy(void *msg, int len, int count, int mtu,
> > - struct sk_buff *skb, struct net_device *dev)
> > -{
> > - struct sk_buff **frag;
> > - int sent = 0;
> > -
> > - memcpy(skb_put(skb, count), msg, count);
> > -
> > - sent += count;
> > - msg += count;
> > - len -= count;
> > -
> > - dev->stats.tx_bytes += count;
> > - dev->stats.tx_packets++;
> > -
> > - raw_dump_table(__func__, "Sending", skb->data, skb->len);
> > -
> > - /* Continuation fragments (no L2CAP header) */
> > - frag = &skb_shinfo(skb)->frag_list;
> > - while (len > 0) {
> > - struct sk_buff *tmp;
> > -
> > - count = min_t(unsigned int, mtu, len);
> > -
> > - tmp = bt_skb_alloc(count, GFP_ATOMIC);
> > - if (!tmp)
> > - return -ENOMEM;
> > -
> > - *frag = tmp;
> > -
> > - memcpy(skb_put(*frag, count), msg, count);
> > -
> > - raw_dump_table(__func__, "Sending fragment",
> > - (*frag)->data, count);
> > -
> > - (*frag)->priority = skb->priority;
> > -
> > - sent += count;
> > - msg += count;
> > - len -= count;
> > -
> > - skb->len += (*frag)->len;
> > - skb->data_len += (*frag)->len;
> > -
> > - frag = &(*frag)->next;
> > -
> > - dev->stats.tx_bytes += count;
> > - dev->stats.tx_packets++;
> > - }
> > -
> > - return sent;
> > -}
> > -
> > -static struct sk_buff *create_pdu(struct l2cap_conn *conn, void *msg,
> > - size_t len, u32 priority,
> > - struct net_device *dev)
> > -{
> > - struct sk_buff *skb;
> > - int err, count;
> > - struct l2cap_hdr *lh;
> > -
> > - /* FIXME: This mtu check should be not needed and atm is only used for
> > - * testing purposes
> > - */
> > - if (conn->mtu > (L2CAP_LE_MIN_MTU + L2CAP_HDR_SIZE))
> > - conn->mtu = L2CAP_LE_MIN_MTU + L2CAP_HDR_SIZE;
> > -
> > - count = min_t(unsigned int, (conn->mtu - L2CAP_HDR_SIZE), len);
> > -
> > - BT_DBG("conn %p len %zu mtu %d count %d", conn, len, conn->mtu, count);
> > -
> > - skb = bt_skb_alloc(count + L2CAP_HDR_SIZE, GFP_ATOMIC);
> > - if (!skb)
> > - return ERR_PTR(-ENOMEM);
> > -
> > - skb->priority = priority;
> > -
> > - lh = (struct l2cap_hdr *)skb_put(skb, L2CAP_HDR_SIZE);
> > - lh->cid = cpu_to_le16(L2CAP_FC_6LOWPAN);
> > - lh->len = cpu_to_le16(len);
> > -
> > - err = skbuff_copy(msg, len, count, conn->mtu, skb, dev);
> > - if (unlikely(err < 0)) {
> > - kfree_skb(skb);
> > - BT_DBG("skbuff copy %d failed", err);
> > - return ERR_PTR(err);
> > - }
> > -
> > - return skb;
> > -}
> > -
> > -static int conn_send(struct l2cap_conn *conn,
> > - void *msg, size_t len, u32 priority,
> > - struct net_device *dev)
> > -{
> > - struct sk_buff *skb;
> > -
> > - skb = create_pdu(conn, msg, len, priority, dev);
> > - if (IS_ERR(skb))
> > - return -EINVAL;
> > -
> > - BT_DBG("conn %p skb %p len %d priority %u", conn, skb, skb->len,
> > - skb->priority);
> > -
> > - hci_send_acl(conn->hchan, skb, ACL_START);
> > -
> > - return 0;
> > -}
> > -
> > static void get_dest_bdaddr(struct in6_addr *ip6_daddr,
> > bdaddr_t *addr, u8 *addr_type)
> > {
> > @@ -466,7 +374,7 @@ static int header_create(struct sk_buff *skb, struct net_device *netdev,
> > if (ipv6_addr_is_multicast(&hdr->daddr)) {
> > memcpy(&lowpan_cb(skb)->addr, &hdr->daddr,
> > sizeof(struct in6_addr));
> > - lowpan_cb(skb)->conn = NULL;
> > + lowpan_cb(skb)->chan = NULL;
> > } else {
> > unsigned long flags;
> >
> > @@ -490,7 +398,7 @@ static int header_create(struct sk_buff *skb, struct net_device *netdev,
> >
> > memcpy(&lowpan_cb(skb)->addr, &hdr->daddr,
> > sizeof(struct in6_addr));
> > - lowpan_cb(skb)->conn = peer->conn;
> > + lowpan_cb(skb)->chan = peer->chan;
> > }
> >
> > saddr = dev->netdev->dev_addr;
> > @@ -499,14 +407,32 @@ static int header_create(struct sk_buff *skb, struct net_device *netdev,
> > }
> >
> > /* Packet to BT LE device */
> > -static int send_pkt(struct l2cap_conn *conn, const void *saddr,
> > +static int send_pkt(struct l2cap_chan *chan, const void *saddr,
> > const void *daddr, struct sk_buff *skb,
> > struct net_device *netdev)
> > {
> > - raw_dump_table(__func__, "raw skb data dump before fragmentation",
> > - skb->data, skb->len);
> > + struct kvec iv;
> > + struct msghdr msg;
> > + int err;
> > +
> > + iv.iov_base = skb->data;
> > + iv.iov_len = skb->len;
> > +
> > + memset(&msg, 0, sizeof(msg));
> > +
> > + msg.msg_iov = (struct iovec *) &iv;
> > + msg.msg_iovlen = 1;
> > +
> > + chan->data = skb->sk;
>
> This assignment here is scary. Why?
The idea was to remember the senders socket so that we can do the voodoo
magic in suspend cb. I seem to have missed the logic behind the suspend
stuff :(
>
> > +
> > + err = l2cap_chan_send(chan, &msg, skb->len, 0);
>
> so this here pretty much proofs my point. I think l2cap_chan_send needs to change to take a single buffer only. Single vector IOV are pretty much useless. We just push it into a single vector for no apparent reason. Lets just hand over the SKB itself to L2CAP core. Then we even have proper reference counting there.
>
> > + if (err > 0) {
> > + netdev->stats.tx_bytes += err;
> > + netdev->stats.tx_packets++;
> > + err = 0;
> > + }
> >
> > - return conn_send(conn, skb->data, skb->len, 0, netdev);
> > + return err;
> > }
> >
> > static void send_mcast_pkt(struct sk_buff *skb, struct net_device *netdev)
> > @@ -529,7 +455,7 @@ static void send_mcast_pkt(struct sk_buff *skb, struct net_device *netdev)
> > list_for_each_entry_safe(pentry, ptmp, &dev->peers, list) {
> > local_skb = skb_clone(skb, GFP_ATOMIC);
> >
> > - send_pkt(pentry->conn, netdev->dev_addr,
> > + send_pkt(pentry->chan, netdev->dev_addr,
> > pentry->eui64_addr, local_skb, netdev);
> >
> > kfree_skb(local_skb);
> > @@ -567,8 +493,8 @@ static netdev_tx_t bt_xmit(struct sk_buff *skb, struct net_device *netdev)
> > BT_DBG("xmit from %s to %pMR (%pI6c) peer %p", netdev->name,
> > &addr, &lowpan_cb(skb)->addr, peer);
> >
> > - if (peer && peer->conn)
> > - err = send_pkt(peer->conn, netdev->dev_addr,
> > + if (peer && peer->chan)
> > + err = send_pkt(peer->chan, netdev->dev_addr,
> > eui64_addr, skb, netdev);
> > }
> > dev_kfree_skb(skb);
> > @@ -664,23 +590,40 @@ static bool is_bt_6lowpan(struct hci_conn *hcon)
> > return test_bit(HCI_CONN_6LOWPAN, &hcon->flags);
> > }
> >
> > -static int add_peer_conn(struct l2cap_conn *conn, struct lowpan_dev *dev)
> > +static struct l2cap_chan *chan_open(struct l2cap_chan *pchan)
> > +{
> > + struct l2cap_chan *chan;
> > +
> > + chan = chan_create(pchan->conn);
> > + if (!chan)
> > + return NULL;
> > +
> > + chan->remote_mps = chan->omtu;
> > + chan->mps = chan->omtu;
> > +
> > + chan->state = BT_CONNECTED;
> > +
> > + return chan;
> > +}
> > +
> > +static struct l2cap_chan *add_peer_chan(struct l2cap_chan *chan,
> > + struct lowpan_dev *dev)
> > {
> > struct lowpan_peer *peer;
> > unsigned long flags;
> >
> > peer = kzalloc(sizeof(*peer), GFP_ATOMIC);
> > if (!peer)
> > - return -ENOMEM;
> > + return NULL;
> >
> > - peer->conn = conn;
> > + peer->chan = chan;
> > memset(&peer->peer_addr, 0, sizeof(struct in6_addr));
> >
> > /* RFC 2464 ch. 5 */
> > peer->peer_addr.s6_addr[0] = 0xFE;
> > peer->peer_addr.s6_addr[1] = 0x80;
> > - set_addr((u8 *)&peer->peer_addr.s6_addr + 8, conn->hcon->dst.b,
> > - conn->hcon->dst_type);
> > + set_addr((u8 *)&peer->peer_addr.s6_addr + 8, chan->conn->hcon->dst.b,
> > + chan->conn->hcon->dst_type);
> >
> > memcpy(&peer->eui64_addr, (u8 *)&peer->peer_addr.s6_addr + 8,
> > EUI64_ADDR_LEN);
> > @@ -701,33 +644,17 @@ static int add_peer_conn(struct l2cap_conn *conn, struct lowpan_dev *dev)
> > INIT_DELAYED_WORK(&dev->notify_peers, do_notify_peers);
> > schedule_delayed_work(&dev->notify_peers, msecs_to_jiffies(100));
> >
> > - return 0;
> > + return peer->chan;
> > }
> >
> > -/* This gets called when BT LE 6LoWPAN device is connected. We then
> > - * create network device that acts as a proxy between BT LE device
> > - * and kernel network stack.
> > - */
> > -int bt_6lowpan_add_conn(struct l2cap_conn *conn)
> > +static int setup_netdev(struct l2cap_conn *conn, struct lowpan_dev **dev)
> > {
> > - struct lowpan_peer *peer = NULL;
> > - struct lowpan_dev *dev;
> > struct net_device *netdev;
> > int err = 0;
> > unsigned long flags;
> >
> > - if (!is_bt_6lowpan(conn->hcon))
> > - return 0;
> > -
> > - peer = lookup_peer(conn);
> > - if (peer)
> > - return -EEXIST;
> > -
> > - dev = lookup_dev(conn);
> > - if (dev)
> > - return add_peer_conn(conn, dev);
> > -
> > - netdev = alloc_netdev(sizeof(*dev), IFACE_NAME_TEMPLATE, netdev_setup);
> > + netdev = alloc_netdev(sizeof(struct lowpan_dev), IFACE_NAME_TEMPLATE,
> > + netdev_setup);
> > if (!netdev)
> > return -ENOMEM;
> >
> > @@ -748,24 +675,49 @@ int bt_6lowpan_add_conn(struct l2cap_conn *conn)
> > netdev->ifindex, &conn->hcon->dst, &conn->hcon->src);
> > set_bit(__LINK_STATE_PRESENT, &netdev->state);
> >
> > - dev = netdev_priv(netdev);
> > - dev->netdev = netdev;
> > - dev->hdev = conn->hcon->hdev;
> > - INIT_LIST_HEAD(&dev->peers);
> > + *dev = netdev_priv(netdev);
> > + (*dev)->netdev = netdev;
> > + (*dev)->hdev = conn->hcon->hdev;
> > + INIT_LIST_HEAD(&(*dev)->peers);
> >
> > write_lock_irqsave(&devices_lock, flags);
> > - INIT_LIST_HEAD(&dev->list);
> > - list_add(&dev->list, &bt_6lowpan_devices);
> > + INIT_LIST_HEAD(&(*dev)->list);
> > + list_add(&(*dev)->list, &bt_6lowpan_devices);
> > write_unlock_irqrestore(&devices_lock, flags);
> >
> > - ifup(netdev);
> > -
> > - return add_peer_conn(conn, dev);
> > + return 0;
> >
> > out:
> > return err;
> > }
> >
> > +static inline void bt_6lowpan_chan_ready_cb(struct l2cap_chan *chan)
> > +{
> > + struct lowpan_dev *dev;
> > +
> > + dev = lookup_dev(chan->conn);
> > +
> > + BT_DBG("chan %p conn %p dev %p", chan, chan->conn, dev);
> > +
> > + if (!dev) {
> > + if (setup_netdev(chan->conn, &dev) < 0) {
> > + l2cap_chan_del(chan, -ENOENT);
> > + return;
> > + }
> > + }
> > +
> > + add_peer_chan(chan, dev);
> > + ifup(dev->netdev);
> > +}
> > +
> > +static inline
> > +struct l2cap_chan *bt_6lowpan_chan_new_connection_cb(struct l2cap_chan *chan)
> > +{
> > + BT_DBG("chan %p", chan);
> > +
> > + return chan_open(chan);
> > +}
> > +
> > static void delete_netdev(struct work_struct *work)
> > {
> > struct lowpan_dev *entry = container_of(work, struct lowpan_dev,
> > @@ -776,26 +728,38 @@ static void delete_netdev(struct work_struct *work)
> > /* The entry pointer is deleted in device_event() */
> > }
> >
> > -int bt_6lowpan_del_conn(struct l2cap_conn *conn)
> > +static void bt_6lowpan_chan_close_cb(struct l2cap_chan *chan)
> > {
> > struct lowpan_dev *entry, *tmp;
> > struct lowpan_dev *dev = NULL;
> > struct lowpan_peer *peer;
> > int err = -ENOENT;
> > unsigned long flags;
> > - bool last = false;
> > + bool last = false, removed = true;
> > +
> > + BT_DBG("chan %p conn %p", chan, chan->conn);
> > +
> > + if (chan->conn && chan->conn->hcon) {
> > + if (!is_bt_6lowpan(chan->conn->hcon))
> > + return;
> >
> > - if (!conn || !is_bt_6lowpan(conn->hcon))
> > - return 0;
> > + /*
> > + * If conn is set, then the netdev is also there and we should
> > + * remove it.
> > + */
> > + removed = false;
> > + }
> >
> > write_lock_irqsave(&devices_lock, flags);
> >
> > list_for_each_entry_safe(entry, tmp, &bt_6lowpan_devices, list) {
> > dev = lowpan_dev(entry->netdev);
> > - peer = peer_lookup_conn(dev, conn);
> > + peer = peer_lookup_chan(dev, chan);
> > if (peer) {
> > last = peer_del(dev, peer);
> > err = 0;
> > + BT_DBG("dev %p removing %speer %p", dev,
> > + last ? "last " : "1 ", peer);
> > break;
> > }
> > }
> > @@ -805,16 +769,166 @@ int bt_6lowpan_del_conn(struct l2cap_conn *conn)
> >
> > cancel_delayed_work_sync(&dev->notify_peers);
> >
> > - /* bt_6lowpan_del_conn() is called with hci dev lock held which
> > - * means that we must delete the netdevice in worker thread.
> > - */
> > - INIT_WORK(&entry->delete_netdev, delete_netdev);
> > - schedule_work(&entry->delete_netdev);
> > + if (!removed) {
> > + INIT_WORK(&entry->delete_netdev, delete_netdev);
> > + schedule_work(&entry->delete_netdev);
> > + }
> > } else {
> > write_unlock_irqrestore(&devices_lock, flags);
> > }
> >
> > - return err;
> > + return;
> > +}
> > +
> > +static void bt_6lowpan_chan_state_change_cb(struct l2cap_chan *chan, int state,
> > + int err)
> > +{
> > + BT_DBG("chan %p conn %p", chan, chan->conn);
> > +}
> > +
> > +static struct sk_buff *bt_6lowpan_chan_alloc_skb_cb(struct l2cap_chan *chan,
> > + unsigned long len, int nb)
> > +{
> > + return bt_skb_alloc(len, GFP_ATOMIC);
> > +}
> > +
> > +static void bt_6lowpan_chan_suspend_cb(struct l2cap_chan *chan)
> > +{
> > + struct sock *sk = chan->data;
> > +
> > + BT_DBG("chan %p conn %p sk %p", chan, chan->conn, sk);
> > +
> > + if (!sk)
> > + return;
> > +
> > + set_bit(BT_SK_SUSPEND, &bt_sk(sk)->flags);
> > + sk->sk_state_change(sk);
> > +}
> > +
> > +static void bt_6lowpan_chan_resume_cb(struct l2cap_chan *chan)
> > +{
> > + struct sock *sk = chan->data;
> > +
> > + BT_DBG("chan %p conn %p sk %p", chan, chan->conn, sk);
> > +
> > + if (!sk)
> > + return;
> > +
> > + clear_bit(BT_SK_SUSPEND, &bt_sk(sk)->flags);
> > + sk->sk_state_change(sk);
> > +}
>
> Why are you messing with socket states here for suspend/resume. These should be internal states to 6loWPAN since it is essentially there to stall IPv6 packets. Involving the socket makes no sense here. You do now have a socket in the first place. Where is this socket coming from?
Took the similar code from l2cap_sock.c so I seem to have missed the
logic in this voodoo code. I will rework this.
>
> > +
> > +static long bt_l2cap_chan_get_sndtimeo(struct l2cap_chan *chan)
> > +{
> > + return msecs_to_jiffies(1000);
> > +}
>
> Why is this needed. You do not have a socket? Do we have this specific timeout handling with LE L2CAP connection oriented channels? We should default to proper defaults or errors if this is not needed.
Hmm, I clearly remember needing this callback, otherwise the connection
was dropped immediately. I will recheck this one.
>
> > +
> > +static struct l2cap_ops bt_6lowpan_chan_ops = {
> > + .name = "L2CAP 6LoWPAN channel",
> > + .recv = bt_6lowpan_chan_recv_cb,
> > + .close = bt_6lowpan_chan_close_cb,
> > + .state_change = bt_6lowpan_chan_state_change_cb,
> > + .alloc_skb = bt_6lowpan_chan_alloc_skb_cb,
> > + .new_connection = bt_6lowpan_chan_new_connection_cb,
> > + .suspend = bt_6lowpan_chan_suspend_cb,
> > + .resume = bt_6lowpan_chan_resume_cb,
> > + .ready = bt_6lowpan_chan_ready_cb,
> > + .get_sndtimeo = bt_l2cap_chan_get_sndtimeo,
> > +
> > + /* Not implemented for 6LoWPAN */
> > + .defer = l2cap_chan_no_defer,
> > + .teardown = l2cap_chan_no_teardown,
> > + .set_shutdown = l2cap_chan_no_set_shutdown,
>
> please keep them in order as L2CAP socket does it. No need to mention that they are not implemented. The _no_ version of the empty helpers do that just fine.
>
> > +};
> > +
> > +static struct l2cap_chan *chan_create(struct l2cap_conn *conn)
> > +{
> > + struct l2cap_chan *chan;
> > +
> > + chan = l2cap_chan_create();
> > + if (!chan)
> > + return NULL;
> > +
> > + l2cap_chan_set_defaults(chan);
> > + skb_queue_head_init(&chan->tx_q);
>
> I now wonder why the channel creation need to repeat the init of tx_q.
Copied the initialization code from other part of the bt subsystem, so
there might be some cruft left in this part.
>
> > +
> > + chan->chan_type = L2CAP_CHAN_CONN_ORIENTED;
> > + chan->flush_to = L2CAP_DEFAULT_FLUSH_TO;
> > + chan->remote_max_tx = chan->max_tx;
> > + chan->remote_tx_win = chan->tx_win;
> > + chan->retrans_timeout = L2CAP_DEFAULT_RETRANS_TO;
> > + chan->monitor_timeout = L2CAP_DEFAULT_MONITOR_TO;
>
> Do we have to repeat the default here? Isn’t l2cap_chan_set_defaults already doing that?
I will recheck what is needed.
>
> > + chan->mode = L2CAP_MODE_LE_FLOWCTL;
> > + chan->conf_state = 0;
>
> This one as well. Defaults should be in place. If not we need to fix the other code first to make sure we are clear here. Users of l2cap_ops should be as simple as possible.
>
> > + chan->omtu = chan->imtu = 65535;
>
> Personally I do not like the a = b = c assignments. Just use single assignments here.
>
> > + chan->ops = &bt_6lowpan_chan_ops;
> > +
> > + return chan;
> > +}
> > +
> > +static struct l2cap_chan *bt_6lowpan_channel_create(struct l2cap_conn *conn)
> > +{
> > + struct lowpan_peer *peer;
> > +
> > + if (!is_bt_6lowpan(conn->hcon))
> > + return NULL;
> > +
> > + BT_DBG("conn %p", conn);
> > +
> > + peer = lookup_peer(conn);
> > + if (peer) {
> > + BT_DBG("6LoWPAN connection and channel already exists");
> > + return NULL;
> > + }
> > +
> > + return chan_create(conn);
> > +}
> > +
> > +static inline __u8 bdaddr_type(struct hci_conn *hcon, __u8 type)
> > +{
> > + if (hcon->type == LE_LINK) {
> > + if (type == ADDR_LE_DEV_PUBLIC)
> > + return BDADDR_LE_PUBLIC;
> > + else
> > + return BDADDR_LE_RANDOM;
> > + }
> > +
> > + return BDADDR_BREDR;
> > +}
> > +
> > +void bt_6lowpan_check(struct l2cap_conn *conn)
> > +{
> > + struct hci_conn *hcon = conn->hcon;
> > + struct l2cap_chan *pchan;
> > + u8 dst_type;
> > + int err;
> > +
> > + dst_type = bdaddr_type(hcon, hcon->dst_type);
> > +
> > + BT_DBG("conn %p dst type %d", conn, dst_type);
> > +
> > + if (hci_blacklist_lookup(hcon->hdev, &hcon->dst, dst_type))
> > + return;
> > +
> > + pchan = bt_6lowpan_channel_create(conn);
> > + if (!pchan)
> > + return;
> > +
> > + if (test_bit(HCI_CONN_6LOWPAN_ROLE_NODE, &conn->hcon->flags)) {
> > + pchan->state = BT_LISTEN;
> > + pchan->src_type = bdaddr_type(hcon, hcon->src_type);
> > +
> > + err = l2cap_add_psm(pchan, &hcon->src,
> > + cpu_to_le16(L2CAP_PSM_6LOWPAN));
> > + if (err) {
> > + BT_ERR("psm cannot be added err %d", err);
> > + return;
> > + }
> > + } else {
> > + err = l2cap_chan_connect(pchan, L2CAP_PSM_6LOWPAN, 0,
> > + &hcon->dst, dst_type);
> > + BT_DBG("chan %p err %d", pchan, err);
> > + }
> > }
>
> So I am thinking that just creating the 6loWPAN connection whenever we connect a LE link is not a good idea. Better have a debugfs trigger to manually create it. I am thinking the following:
>
> /sys/kernel/debug/bluetooth/hci0/6lowpan
>
> pam <val> # sets PSM, 0 for disable (default)
> connect <addr> <type> # connect 6loWPAN channel
> disconnect <addr> <type> # disconnect 6loWPAN channel
>
> When reading that file it will show you the active 6loWPAN connections. This also means that all the hci_dev and hci_conn flags can go away and it is kept internally inside bluetooth_6lowpan.ko module.
Sure.
>
> That will be pretty close to an API that we want to expose via mgmt at some point.
>
> > static int device_event(struct notifier_block *unused,
> > diff --git a/net/bluetooth/6lowpan.h b/net/bluetooth/6lowpan.h
> > index 5d281f1..1ba7af1 100644
> > --- a/net/bluetooth/6lowpan.h
> > +++ b/net/bluetooth/6lowpan.h
> > @@ -1,5 +1,5 @@
> > /*
> > - Copyright (c) 2013 Intel Corp.
> > + Copyright (c) 2014 Intel Corp.
>
> Same as above, 2013-2014.
>
> >
> > This program is free software; you can redistribute it and/or modify
> > it under the terms of the GNU General Public License version 2 and
> > @@ -19,29 +19,16 @@
> > #include <net/bluetooth/l2cap.h>
> >
> > #if IS_ENABLED(CONFIG_BT_6LOWPAN)
> > -int bt_6lowpan_recv(struct l2cap_conn *conn, struct sk_buff *skb);
> > -int bt_6lowpan_add_conn(struct l2cap_conn *conn);
> > -int bt_6lowpan_del_conn(struct l2cap_conn *conn);
> > int bt_6lowpan_init(void);
> > void bt_6lowpan_cleanup(void);
> > +void bt_6lowpan_check(struct l2cap_conn *conn);
> > #else
> > -static int bt_6lowpan_recv(struct l2cap_conn *conn, struct sk_buff *skb)
> > -{
> > - return -EOPNOTSUPP;
> > -}
> > -static int bt_6lowpan_add_conn(struct l2cap_conn *conn)
> > -{
> > - return -EOPNOTSUPP;
> > -}
> > -int bt_6lowpan_del_conn(struct l2cap_conn *conn)
> > -{
> > - return -EOPNOTSUPP;
> > -}
> > static int bt_6lowpan_init(void)
> > {
> > return -EOPNOTSUPP;
> > }
> > static void bt_6lowpan_cleanup(void) { }
> > +void bt_6lowpan_check(struct l2cap_conn *conn) { }
> > #endif
>
> With bluetooth_6lowpan.ko this should go away completely.
>
> > #endif /* __6LOWPAN_H */
> > diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
> > index 528b38a..ef46498 100644
> > --- a/net/bluetooth/l2cap_core.c
> > +++ b/net/bluetooth/l2cap_core.c
> > @@ -1454,8 +1454,6 @@ static void l2cap_le_conn_ready(struct l2cap_conn *conn)
> >
> > BT_DBG("");
> >
> > - bt_6lowpan_add_conn(conn);
> > -
> > /* Check if we have socket listening on cid */
> > pchan = l2cap_global_chan_by_scid(BT_LISTEN, L2CAP_CID_ATT,
> > &hcon->src, &hcon->dst);
> > @@ -1530,6 +1528,7 @@ static void l2cap_conn_ready(struct l2cap_conn *conn)
> >
> > mutex_unlock(&conn->chan_lock);
> >
> > + queue_work(hcon->hdev->workqueue, &conn->pending_6lowpan_work);
> > queue_work(hcon->hdev->workqueue, &conn->pending_rx_work);
> > }
> >
> > @@ -6945,10 +6944,6 @@ static void l2cap_recv_frame(struct l2cap_conn *conn, struct sk_buff *skb)
> > l2cap_conn_del(conn->hcon, EACCES);
> > break;
> >
> > - case L2CAP_FC_6LOWPAN:
> > - bt_6lowpan_recv(conn, skb);
> > - break;
> > -
> > default:
> > l2cap_data_channel(conn, cid, skb);
> > break;
> > @@ -6967,6 +6962,15 @@ static void process_pending_rx(struct work_struct *work)
> > l2cap_recv_frame(conn, skb);
> > }
> >
> > +static void process_pending_6lowpan(struct work_struct *work)
> > +{
> > + struct l2cap_conn *conn = container_of(work, struct l2cap_conn,
> > + pending_6lowpan_work);
> > + BT_DBG("");
> > +
> > + bt_6lowpan_check(conn);
> > +}
> > +
> > static struct l2cap_conn *l2cap_conn_add(struct hci_conn *hcon)
> > {
> > struct l2cap_conn *conn = hcon->l2cap_data;
> > @@ -7024,6 +7028,7 @@ static struct l2cap_conn *l2cap_conn_add(struct hci_conn *hcon)
> >
> > skb_queue_head_init(&conn->pending_rx);
> > INIT_WORK(&conn->pending_rx_work, process_pending_rx);
> > + INIT_WORK(&conn->pending_6lowpan_work, process_pending_6lowpan);
> >
> > conn->disc_reason = HCI_ERROR_REMOTE_USER_TERM;
> >
> > @@ -7257,8 +7262,6 @@ void l2cap_disconn_cfm(struct hci_conn *hcon, u8 reason)
> > {
> > BT_DBG("hcon %p reason %d", hcon, reason);
> >
> > - bt_6lowpan_del_conn(hcon->l2cap_data);
> > -
> > l2cap_conn_del(hcon, bt_to_errno(reason));
> > }
>
> Regards
>
> Marcel
>
Cheers,
Jukka
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
