[PATCH 6/8] usb: USB over WiFi HCD driver

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Works with devices using Ozmo USB over WiFi technology.
See README file for further details.
Added sixth batch of files to drivers/staging/ozwpan.

Signed-off-by: Chris Kelly <ckelly@xxxxxxxxxxxxxxx>
---
 drivers/staging/ozwpan/ozproto.c |  957 ++++++++++++++++++++++++++++++++++++++
 drivers/staging/ozwpan/ozproto.h |   69 +++
 2 files changed, 1026 insertions(+), 0 deletions(-)
 create mode 100644 drivers/staging/ozwpan/ozproto.c
 create mode 100644 drivers/staging/ozwpan/ozproto.h

diff --git a/drivers/staging/ozwpan/ozproto.c b/drivers/staging/ozwpan/ozproto.c
new file mode 100644
index 0000000..8359bb5
--- /dev/null
+++ b/drivers/staging/ozwpan/ozproto.c
@@ -0,0 +1,957 @@
+/* -----------------------------------------------------------------------------
+ * Copyright (c) 2011 Ozmo Inc
+ * Released under the GNU General Public License Version 2 (GPLv2).
+ * -----------------------------------------------------------------------------
+ */
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/timer.h>
+#include <linux/sched.h>
+#include <linux/netdevice.h>
+#include <linux/errno.h>
+#include <linux/ieee80211.h>
+#include "ozconfig.h"
+#include "ozprotocol.h"
+#include "ozeltbuf.h"
+#include "ozpd.h"
+#include "ozproto.h"
+#include "ozusbsvc.h"
+#include "oztrace.h"
+#include "ozappif.h"
+#include "ozevent.h"
+#include "ozalloc.h"
+#include "ozunaligned.h"
+#include <linux/uaccess.h>
+#include <net/psnap.h>
+/*------------------------------------------------------------------------------
+ */
+#define OZ_CF_CONN_SUCCESS	1
+#define OZ_CF_CONN_FAILURE	2
+
+#define OZ_DO_STOP		1
+#define OZ_DO_SLEEP		2
+
+/* States of the timer.
+ */
+#define OZ_TIMER_IDLE		0
+#define OZ_TIMER_SET		1
+#define OZ_TIMER_IN_HANDLER	2
+
+#define OZ_MAX_TIMER_POOL_SIZE	16
+
+/*------------------------------------------------------------------------------
+ */
+struct oz_binding {
+	struct packet_type ptype;
+	char name[OZ_MAX_BINDING_LEN];
+	struct oz_binding *next;
+};
+
+struct oz_timer {
+	struct list_head link;
+	struct oz_pd *pd;
+	unsigned long due_time;
+	int type;
+};
+/*------------------------------------------------------------------------------
+ * Static external variables.
+ */
+static DEFINE_SPINLOCK(g_polling_lock);
+static LIST_HEAD(g_pd_list);
+static struct oz_binding *g_binding ;
+static DEFINE_SPINLOCK(g_binding_lock);
+static struct sk_buff_head g_rx_queue;
+static u8 g_session_id;
+static u16 g_apps = 0x1;
+static int g_processing_rx;
+static struct timer_list g_timer;
+static struct oz_timer *g_cur_timer;
+static struct list_head *g_timer_pool;
+static int g_timer_pool_count;
+static int g_timer_state = OZ_TIMER_IDLE;
+static LIST_HEAD(g_timer_list);
+/*------------------------------------------------------------------------------
+ */
+static void oz_protocol_timer_start(void);
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static u8 oz_get_new_session_id(u8 exclude)
+{
+	if (++g_session_id == 0)
+		g_session_id = 1;
+	if (g_session_id == exclude) {
+		if (++g_session_id == 0)
+			g_session_id = 1;
+	}
+	return g_session_id;
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static void oz_send_conn_rsp(struct oz_pd *pd, u8 status)
+{
+	struct sk_buff *skb;
+	struct net_device *dev = pd->net_dev;
+	struct oz_hdr *oz_hdr;
+	struct oz_elt *elt;
+	struct oz_elt_connect_rsp *body;
+	int sz = sizeof(struct oz_hdr) + sizeof(struct oz_elt) +
+			sizeof(struct oz_elt_connect_rsp);
+	skb = alloc_skb(sz + OZ_ALLOCATED_SPACE(dev), GFP_ATOMIC);
+	if (skb == 0)
+		return;
+	skb_reserve(skb, LL_RESERVED_SPACE(dev));
+	skb_reset_network_header(skb);
+	oz_hdr = (struct oz_hdr *)skb_put(skb, sz);
+	elt = (struct oz_elt *)(oz_hdr+1);
+	body = (struct oz_elt_connect_rsp *)(elt+1);
+	skb->dev = dev;
+	skb->protocol = htons(OZ_ETHERTYPE);
+	/* Fill in device header */
+	if (dev_hard_header(skb, dev, OZ_ETHERTYPE, pd->mac_addr,
+			dev->dev_addr, skb->len) < 0) {
+		kfree_skb(skb);
+		return;
+	}
+	oz_hdr->control = (OZ_PROTOCOL_VERSION<<OZ_VERSION_SHIFT);
+	oz_hdr->last_pkt_num = 0;
+	put_unaligned(0, &oz_hdr->pkt_num);
+	oz_event_log(OZ_EVT_CONNECT_RSP, 0, 0, 0, 0);
+	elt->type = OZ_ELT_CONNECT_RSP;
+	elt->length = sizeof(struct oz_elt_connect_rsp);
+	memset(body, 0, sizeof(struct oz_elt_connect_rsp));
+	body->status = status;
+	if (status == 0) {
+		body->mode = pd->mode;
+		body->session_id = pd->session_id;
+		put_unaligned(cpu_to_le16(pd->total_apps), &body->apps);
+	}
+	oz_trace("TX: OZ_ELT_CONNECT_RSP %d", status);
+	dev_queue_xmit(skb);
+	return;
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static void pd_set_keepalive(struct oz_pd *pd, u8 kalive)
+{
+	unsigned long keep_alive = kalive & OZ_KALIVE_VALUE_MASK;
+
+	switch (kalive & OZ_KALIVE_TYPE_MASK) {
+	case OZ_KALIVE_SPECIAL:
+		pd->keep_alive_j =
+			oz_ms_to_jiffies(keep_alive * 1000*60*60*24*20);
+		break;
+	case OZ_KALIVE_SECS:
+		pd->keep_alive_j = oz_ms_to_jiffies(keep_alive*1000);
+		break;
+	case OZ_KALIVE_MINS:
+		pd->keep_alive_j = oz_ms_to_jiffies(keep_alive*1000*60);
+		break;
+	case OZ_KALIVE_HOURS:
+		pd->keep_alive_j = oz_ms_to_jiffies(keep_alive*1000*60*60);
+		break;
+	default:
+		pd->keep_alive_j = 0;
+	}
+	oz_trace("Keepalive = %lu jiffies\n", pd->keep_alive_j);
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static void pd_set_presleep(struct oz_pd *pd, u8 presleep)
+{
+	if (presleep)
+		pd->presleep_j = oz_ms_to_jiffies(presleep*100);
+	else
+		pd->presleep_j = OZ_PRESLEEP_TOUT_J;
+	oz_trace("Presleep time = %lu jiffies\n", pd->presleep_j);
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static struct oz_pd *oz_connect_req(struct oz_pd *cur_pd, struct oz_elt *elt,
+			u8 *pd_addr, struct net_device *net_dev)
+{
+	struct oz_pd *pd;
+	struct oz_elt_connect_req *body =
+			(struct oz_elt_connect_req *)(elt+1);
+	u8 rsp_status = OZ_STATUS_SUCCESS;
+	u8 stop_needed = 0;
+	u16 new_apps = g_apps;
+	struct net_device *old_net_dev = 0;
+	struct oz_pd *free_pd = 0;
+	if (cur_pd) {
+		pd = cur_pd;
+		spin_lock_bh(&g_polling_lock);
+	} else {
+		struct oz_pd *pd2 = 0;
+		struct list_head *e;
+		pd = oz_pd_alloc(pd_addr);
+		if (pd == 0)
+			return 0;
+		pd->last_rx_time_j = jiffies;
+		spin_lock_bh(&g_polling_lock);
+		list_for_each(e, &g_pd_list) {
+			pd2 = container_of(e, struct oz_pd, link);
+			if (memcmp(pd2->mac_addr, pd_addr, ETH_ALEN) == 0) {
+				free_pd = pd;
+				pd = pd2;
+				break;
+			}
+		}
+		if (pd != pd2)
+			list_add_tail(&pd->link, &g_pd_list);
+	}
+	if (pd == 0) {
+		spin_unlock_bh(&g_polling_lock);
+		return 0;
+	}
+	if (pd->net_dev != net_dev) {
+		old_net_dev = pd->net_dev;
+		dev_hold(net_dev);
+		pd->net_dev = net_dev;
+	}
+	oz_trace("Host vendor: %d\n", body->host_vendor);
+	pd->max_tx_size = OZ_MAX_TX_SIZE;
+	pd->mode = body->mode;
+	pd->pd_info = body->pd_info;
+	if (pd->mode & OZ_F_ISOC_NO_ELTS) {
+		pd->mode |= OZ_F_ISOC_ANYTIME;
+		pd->ms_per_isoc = body->ms_per_isoc;
+		if (!pd->ms_per_isoc)
+			pd->ms_per_isoc = 4;
+	}
+	if (body->max_len_div16)
+		pd->max_tx_size = ((u16)body->max_len_div16)<<4;
+	oz_trace("Max frame:%u Ms per isoc:%u\n",
+		pd->max_tx_size, pd->ms_per_isoc);
+	pd->max_stream_buffering = 3*1024;
+	pd->timeout_time_j = jiffies + OZ_CONNECTION_TOUT_J;
+	pd->pulse_period_j = OZ_QUANTUM_J;
+	pd_set_presleep(pd, body->presleep);
+	pd_set_keepalive(pd, body->keep_alive);
+
+	new_apps &= le16_to_cpu(get_unaligned(&body->apps));
+	if ((new_apps & 0x1) && (body->session_id)) {
+		if (pd->session_id) {
+			if (pd->session_id != body->session_id) {
+				rsp_status = OZ_STATUS_SESSION_MISMATCH;
+				goto done;
+			}
+		} else {
+			new_apps &= ~0x1;  /* Resume not permitted */
+			pd->session_id =
+				oz_get_new_session_id(body->session_id);
+		}
+	} else {
+		if (pd->session_id && !body->session_id) {
+			rsp_status = OZ_STATUS_SESSION_TEARDOWN;
+			stop_needed = 1;
+		} else {
+			new_apps &= ~0x1;  /* Resume not permitted */
+			pd->session_id =
+				oz_get_new_session_id(body->session_id);
+		}
+	}
+done:
+	if (rsp_status == OZ_STATUS_SUCCESS) {
+		u16 start_apps = new_apps & ~pd->total_apps & ~0x1;
+		u16 stop_apps = pd->total_apps & ~new_apps & ~0x1;
+		u16 resume_apps = new_apps & pd->paused_apps  & ~0x1;
+		spin_unlock_bh(&g_polling_lock);
+		oz_pd_set_state(pd, OZ_PD_S_CONNECTED);
+		oz_timer_delete(pd, OZ_TIMER_STOP);
+		oz_trace("new_apps=0x%x total_apps=0x%x paused_apps=0x%x\n",
+			new_apps, pd->total_apps, pd->paused_apps);
+		if (start_apps) {
+			if (oz_services_start(pd, start_apps, 0))
+				rsp_status = OZ_STATUS_TOO_MANY_PDS;
+		}
+		if (resume_apps)
+			if (oz_services_start(pd, resume_apps, 1))
+				rsp_status = OZ_STATUS_TOO_MANY_PDS;
+		if (stop_apps)
+			oz_services_stop(pd, stop_apps, 0);
+		oz_pd_request_heartbeat(pd);
+	} else {
+		spin_unlock_bh(&g_polling_lock);
+	}
+	oz_send_conn_rsp(pd, rsp_status);
+	if (rsp_status != OZ_STATUS_SUCCESS) {
+		if (stop_needed)
+			oz_pd_stop(pd);
+		oz_pd_put(pd);
+		pd = 0;
+	}
+	if (old_net_dev)
+		dev_put(old_net_dev);
+	if (free_pd)
+		oz_pd_destroy(free_pd);
+	return pd;
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static void oz_add_farewell(struct oz_pd *pd, u8 ep_num, u8 index,
+			u8 *report, u8 len)
+{
+	struct oz_farewell *f;
+	struct oz_farewell *f2;
+	int found = 0;
+	f = oz_alloc(sizeof(struct oz_farewell) + len - 1, GFP_ATOMIC);
+	if (!f)
+		return;
+	f->ep_num = ep_num;
+	f->index = index;
+	memcpy(f->report, report, len);
+	oz_trace("RX: Adding farewell report\n");
+	spin_lock(&g_polling_lock);
+	list_for_each_entry(f2, &pd->farewell_list, link) {
+		if ((f2->ep_num == ep_num) && (f2->index == index)) {
+			found = 1;
+			list_del(&f2->link);
+			break;
+		}
+	}
+	list_add_tail(&f->link, &pd->farewell_list);
+	spin_unlock(&g_polling_lock);
+	if (found)
+		oz_free(f2);
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq-serialized
+ */
+static void oz_rx_frame(struct sk_buff *skb)
+{
+	u8 *mac_hdr;
+	u8 *src_addr;
+	struct oz_elt *elt;
+	int length;
+	struct oz_pd *pd = 0;
+	struct oz_hdr *oz_hdr = (struct oz_hdr *)skb_network_header(skb);
+	int dup = 0;
+	u32 pkt_num;
+
+	oz_event_log(OZ_EVT_RX_PROCESS, 0,
+		(((u16)oz_hdr->control)<<8)|oz_hdr->last_pkt_num,
+		0, oz_hdr->pkt_num);
+	oz_trace2(OZ_TRACE_RX_FRAMES,
+		"RX frame PN=0x%x LPN=0x%x control=0x%x\n",
+		oz_hdr->pkt_num, oz_hdr->last_pkt_num, oz_hdr->control);
+	mac_hdr = skb_mac_header(skb);
+	src_addr = &mac_hdr[ETH_ALEN] ;
+	length = skb->len;
+
+	/* Check the version field */
+	if (oz_get_prot_ver(oz_hdr->control) != OZ_PROTOCOL_VERSION) {
+		oz_trace("Incorrect protocol version: %d\n",
+			oz_get_prot_ver(oz_hdr->control));
+		goto done;
+	}
+
+	pkt_num = le32_to_cpu(get_unaligned(&oz_hdr->pkt_num));
+
+	pd = oz_pd_find(src_addr);
+	if (pd) {
+		pd->last_rx_time_j = jiffies;
+		oz_timer_add(pd, OZ_TIMER_TOUT,
+			pd->last_rx_time_j + pd->presleep_j, 1);
+		if (pkt_num != pd->last_rx_pkt_num) {
+			pd->last_rx_pkt_num = pkt_num;
+		} else {
+			dup = 1;
+			oz_trace("Duplicate frame\n");
+		}
+	}
+
+	if (pd && !dup && ((pd->mode & OZ_MODE_MASK) == OZ_MODE_TRIGGERED)) {
+		pd->last_sent_frame = &pd->tx_queue;
+		if (oz_hdr->control & OZ_F_ACK) {
+			/* Retire completed frames */
+			oz_retire_tx_frames(pd, oz_hdr->last_pkt_num);
+		}
+		if ((oz_hdr->control & OZ_F_ACK_REQUESTED) &&
+				(pd->state == OZ_PD_S_CONNECTED)) {
+			int backlog = pd->nb_queued_frames;
+			pd->trigger_pkt_num = pkt_num;
+			/* Send queued frames */
+			while (oz_prepare_frame(pd, 0) >= 0)
+				;
+			oz_send_queued_frames(pd, backlog);
+		}
+	}
+
+	length -= sizeof(struct oz_hdr);
+	elt = (struct oz_elt *)((u8 *)oz_hdr + sizeof(struct oz_hdr));
+
+	while (length >= sizeof(struct oz_elt)) {
+		length -= sizeof(struct oz_elt) + elt->length;
+		if (length < 0)
+			break;
+		switch (elt->type) {
+		case OZ_ELT_CONNECT_REQ:
+			oz_event_log(OZ_EVT_CONNECT_REQ, 0, 0, 0, 0);
+			oz_trace("RX: OZ_ELT_CONNECT_REQ\n");
+			pd = oz_connect_req(pd, elt, src_addr, skb->dev);
+			break;
+		case OZ_ELT_DISCONNECT:
+			oz_trace("RX: OZ_ELT_DISCONNECT\n");
+			if (pd)
+				oz_pd_sleep(pd);
+			break;
+		case OZ_ELT_UPDATE_PARAM_REQ: {
+				struct oz_elt_update_param *body =
+					(struct oz_elt_update_param *)(elt + 1);
+				oz_trace("RX: OZ_ELT_UPDATE_PARAM_REQ\n");
+				if (pd && (pd->state & OZ_PD_S_CONNECTED)) {
+					spin_lock(&g_polling_lock);
+					pd_set_keepalive(pd, body->keepalive);
+					pd_set_presleep(pd, body->presleep);
+					spin_unlock(&g_polling_lock);
+				}
+			}
+			break;
+		case OZ_ELT_FAREWELL_REQ: {
+				struct oz_elt_farewell *body =
+					(struct oz_elt_farewell *)(elt + 1);
+				oz_trace("RX: OZ_ELT_FAREWELL_REQ\n");
+				oz_add_farewell(pd, body->ep_num,
+					body->index, body->report,
+					elt->length + 1 - sizeof(*body));
+			}
+			break;
+		case OZ_ELT_APP_DATA:
+			if (pd && (pd->state & OZ_PD_S_CONNECTED)) {
+				struct oz_app_hdr *app_hdr =
+					(struct oz_app_hdr *)(elt+1);
+				if (dup)
+					break;
+				oz_handle_app_elt(pd, app_hdr->app_id, elt);
+			}
+			break;
+		default:
+			oz_trace("RX: Unknown elt %02x\n", elt->type);
+		}
+		elt = oz_next_elt(elt);
+	}
+done:
+	if (pd)
+		oz_pd_put(pd);
+	consume_skb(skb);
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+void oz_protocol_term(void)
+{
+	struct list_head *chain = 0;
+	del_timer_sync(&g_timer);
+	/* Walk the list of bindings and remove each one.
+	 */
+	spin_lock_bh(&g_binding_lock);
+	while (g_binding) {
+		struct oz_binding *b = g_binding;
+		g_binding = b->next;
+		spin_unlock_bh(&g_binding_lock);
+		dev_remove_pack(&b->ptype);
+		if (b->ptype.dev)
+			dev_put(b->ptype.dev);
+		oz_free(b);
+		spin_lock_bh(&g_binding_lock);
+	}
+	spin_unlock_bh(&g_binding_lock);
+	/* Walk the list of PDs and stop each one. This causes the PD to be
+	 * removed from the list so we can just pull each one from the head
+	 * of the list.
+	 */
+	spin_lock_bh(&g_polling_lock);
+	while (!list_empty(&g_pd_list)) {
+		struct oz_pd *pd =
+			list_first_entry(&g_pd_list, struct oz_pd, link);
+		oz_pd_get(pd);
+		spin_unlock_bh(&g_polling_lock);
+		oz_pd_stop(pd);
+		oz_pd_put(pd);
+		spin_lock_bh(&g_polling_lock);
+	}
+	chain = g_timer_pool;
+	g_timer_pool = 0;
+	spin_unlock_bh(&g_polling_lock);
+	while (chain) {
+		struct oz_timer *t = container_of(chain, struct oz_timer, link);
+		chain = chain->next;
+		oz_free(t);
+	}
+	oz_trace("Protocol stopped\n");
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq
+ */
+static void oz_pd_handle_timer(struct oz_pd *pd, int type)
+{
+	switch (type) {
+	case OZ_TIMER_TOUT:
+		oz_pd_sleep(pd);
+		break;
+	case OZ_TIMER_STOP:
+		oz_pd_stop(pd);
+		break;
+	case OZ_TIMER_HEARTBEAT: {
+			u16 apps = 0;
+			spin_lock_bh(&g_polling_lock);
+			pd->heartbeat_requested = 0;
+			if (pd->state & OZ_PD_S_CONNECTED)
+				apps = pd->total_apps;
+			spin_unlock_bh(&g_polling_lock);
+			if (apps)
+				oz_pd_heartbeat(pd, apps);
+		}
+		break;
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq
+ */
+static void oz_protocol_timer(unsigned long arg)
+{
+	struct oz_timer *t;
+	struct oz_timer *t2;
+	struct oz_pd *pd;
+	spin_lock_bh(&g_polling_lock);
+	if (!g_cur_timer) {
+		/* This happens if we remove the current timer but can't stop
+		 * the timer from firing. In this case just get out.
+		 */
+		oz_event_log(OZ_EVT_TIMER, 0, 0, 0, 0);
+		spin_unlock_bh(&g_polling_lock);
+		return;
+	}
+	g_timer_state = OZ_TIMER_IN_HANDLER;
+	t = g_cur_timer;
+	g_cur_timer = 0;
+	list_del(&t->link);
+	spin_unlock_bh(&g_polling_lock);
+	do {
+		pd = t->pd;
+		oz_event_log(OZ_EVT_TIMER, 0, t->type, 0, 0);
+		oz_pd_handle_timer(pd, t->type);
+		spin_lock_bh(&g_polling_lock);
+		if (g_timer_pool_count < OZ_MAX_TIMER_POOL_SIZE) {
+			t->link.next = g_timer_pool;
+			g_timer_pool = &t->link;
+			g_timer_pool_count++;
+			t = 0;
+		}
+		if (!list_empty(&g_timer_list)) {
+			t2 =  container_of(g_timer_list.next,
+				struct oz_timer, link);
+			if (time_before_eq(t2->due_time, jiffies))
+				list_del(&t2->link);
+			else
+				t2 = 0;
+		} else {
+			t2 = 0;
+		}
+		spin_unlock_bh(&g_polling_lock);
+		oz_pd_put(pd);
+		if (t)
+			oz_free(t);
+		t = t2;
+	} while (t);
+	g_timer_state = OZ_TIMER_IDLE;
+	oz_protocol_timer_start();
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq
+ */
+static void oz_protocol_timer_start(void)
+{
+	spin_lock_bh(&g_polling_lock);
+	if (!list_empty(&g_timer_list)) {
+		g_cur_timer =
+			container_of(g_timer_list.next, struct oz_timer, link);
+		if (g_timer_state == OZ_TIMER_SET) {
+			oz_event_log(OZ_EVT_TIMER_CTRL, 3,
+				(u16)g_cur_timer->type, 0,
+				(unsigned)g_cur_timer->due_time);
+			mod_timer(&g_timer, g_cur_timer->due_time);
+		} else {
+			oz_event_log(OZ_EVT_TIMER_CTRL, 4,
+				(u16)g_cur_timer->type, 0,
+				(unsigned)g_cur_timer->due_time);
+			g_timer.expires = g_cur_timer->due_time;
+			g_timer.function = oz_protocol_timer;
+			g_timer.data = 0;
+			add_timer(&g_timer);
+		}
+		g_timer_state = OZ_TIMER_SET;
+	} else {
+		oz_trace("No queued timers\n");
+	}
+	spin_unlock_bh(&g_polling_lock);
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq or process
+ */
+void oz_timer_add(struct oz_pd *pd, int type, unsigned long due_time,
+		int remove)
+{
+	struct list_head *e;
+	struct oz_timer *t = 0;
+	int restart_needed = 0;
+	oz_event_log(OZ_EVT_TIMER_CTRL, 1, (u16)type, 0, (unsigned)due_time);
+	spin_lock(&g_polling_lock);
+	if (remove) {
+		list_for_each(e, &g_timer_list) {
+			t = container_of(e, struct oz_timer, link);
+			if ((t->pd == pd) && (t->type == type)) {
+				if (g_cur_timer == t) {
+					restart_needed = 1;
+					g_cur_timer = 0;
+				}
+				list_del(e);
+				break;
+			}
+			t = 0;
+		}
+	}
+	if (!t) {
+		if (g_timer_pool) {
+			t = container_of(g_timer_pool, struct oz_timer, link);
+			g_timer_pool = g_timer_pool->next;
+			g_timer_pool_count--;
+		} else {
+			t = oz_alloc(sizeof(struct oz_timer), GFP_ATOMIC);
+		}
+		if (t) {
+			t->pd = pd;
+			t->type = type;
+			oz_pd_get(pd);
+		}
+	}
+	if (t) {
+		struct oz_timer *t2;
+		t->due_time = due_time;
+		list_for_each(e, &g_timer_list) {
+			t2 = container_of(e, struct oz_timer, link);
+			if (time_before(due_time, t2->due_time)) {
+				if (t2 == g_cur_timer) {
+					g_cur_timer = 0;
+					restart_needed = 1;
+				}
+				break;
+			}
+		}
+		list_add_tail(&t->link, e);
+	}
+	if (g_timer_state == OZ_TIMER_IDLE)
+		restart_needed = 1;
+	else if (g_timer_state == OZ_TIMER_IN_HANDLER)
+		restart_needed = 0;
+	spin_unlock(&g_polling_lock);
+	if (restart_needed)
+		oz_protocol_timer_start();
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq or process
+ */
+void oz_timer_delete(struct oz_pd *pd, int type)
+{
+	struct list_head *chain = 0;
+	struct oz_timer *t;
+	struct oz_timer *n;
+	int restart_needed = 0;
+	int release = 0;
+	oz_event_log(OZ_EVT_TIMER_CTRL, 2, (u16)type, 0, 0);
+	spin_lock(&g_polling_lock);
+	list_for_each_entry_safe(t, n, &g_timer_list, link) {
+		if ((t->pd == pd) && ((type == 0) || (t->type == type))) {
+			if (g_cur_timer == t) {
+				restart_needed = 1;
+				g_cur_timer = 0;
+				del_timer(&g_timer);
+			}
+			list_del(&t->link);
+			release++;
+			if (g_timer_pool_count < OZ_MAX_TIMER_POOL_SIZE) {
+				t->link.next = g_timer_pool;
+				g_timer_pool = &t->link;
+				g_timer_pool_count++;
+			} else {
+				t->link.next = chain;
+				chain = &t->link;
+			}
+			if (type)
+				break;
+		}
+	}
+	if (g_timer_state == OZ_TIMER_IN_HANDLER)
+		restart_needed = 0;
+	else if (restart_needed)
+		g_timer_state = OZ_TIMER_IDLE;
+	spin_unlock(&g_polling_lock);
+	if (restart_needed)
+		oz_protocol_timer_start();
+	while (release--)
+		oz_pd_put(pd);
+	while (chain) {
+		t = container_of(chain, struct oz_timer, link);
+		chain = chain->next;
+		oz_free(t);
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq or process
+ */
+void oz_pd_request_heartbeat(struct oz_pd *pd)
+{
+	unsigned long now = jiffies;
+	unsigned long t;
+	spin_lock(&g_polling_lock);
+	if (pd->heartbeat_requested) {
+		spin_unlock(&g_polling_lock);
+		return;
+	}
+	if (pd->pulse_period_j)
+		t = ((now / pd->pulse_period_j) + 1) * pd->pulse_period_j;
+	else
+		t = now + 1;
+	pd->heartbeat_requested = 1;
+	spin_unlock(&g_polling_lock);
+	oz_timer_add(pd, OZ_TIMER_HEARTBEAT, t, 0);
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq or process
+ */
+struct oz_pd *oz_pd_find(u8 *mac_addr)
+{
+	struct oz_pd *pd;
+	struct list_head *e;
+	spin_lock_bh(&g_polling_lock);
+	list_for_each(e, &g_pd_list) {
+		pd = container_of(e, struct oz_pd, link);
+		if (memcmp(pd->mac_addr, mac_addr, ETH_ALEN) == 0) {
+			atomic_inc(&pd->ref_count);
+			spin_unlock_bh(&g_polling_lock);
+			return pd;
+		}
+	}
+	spin_unlock_bh(&g_polling_lock);
+	return 0;
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+void oz_app_enable(int app_id, int enable)
+{
+	if (app_id <= OZ_APPID_MAX) {
+		spin_lock_bh(&g_polling_lock);
+		if (enable)
+			g_apps |= (1<<app_id);
+		else
+			g_apps &= ~(1<<app_id);
+		spin_unlock_bh(&g_polling_lock);
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: softirq
+ */
+static int oz_pkt_recv(struct sk_buff *skb, struct net_device *dev,
+		struct packet_type *pt, struct net_device *orig_dev)
+{
+	oz_event_log(OZ_EVT_RX_FRAME, 0, 0, 0, 0);
+	skb = skb_share_check(skb, GFP_ATOMIC);
+	if (skb == 0)
+		return 0;
+	spin_lock_bh(&g_rx_queue.lock);
+	if (g_processing_rx) {
+		/* We already hold the lock so use __ variant.
+		 */
+		__skb_queue_head(&g_rx_queue, skb);
+		spin_unlock_bh(&g_rx_queue.lock);
+	} else {
+		g_processing_rx = 1;
+		do {
+
+			spin_unlock_bh(&g_rx_queue.lock);
+			oz_rx_frame(skb);
+			spin_lock_bh(&g_rx_queue.lock);
+			if (skb_queue_empty(&g_rx_queue)) {
+				g_processing_rx = 0;
+				spin_unlock_bh(&g_rx_queue.lock);
+				break;
+			}
+			/* We already hold the lock so use __ variant.
+			 */
+			skb = __skb_dequeue(&g_rx_queue);
+		} while (1);
+	}
+	return 0;
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+void oz_binding_add(char *net_dev)
+{
+	struct oz_binding *binding;
+	binding = oz_alloc(sizeof(struct oz_binding), GFP_ATOMIC);
+	if (binding) {
+		binding->ptype.type = __constant_htons(OZ_ETHERTYPE);
+		binding->ptype.func = oz_pkt_recv;
+		memcpy(binding->name, net_dev, OZ_MAX_BINDING_LEN);
+		if (net_dev && *net_dev) {
+			oz_trace("Adding binding: %s\n", net_dev);
+			binding->ptype.dev =
+				dev_get_by_name(&init_net, net_dev);
+			if (binding->ptype.dev == 0) {
+				oz_trace("Netdev %s not found\n", net_dev);
+				oz_free(binding);
+				binding = 0;
+			}
+		} else {
+			oz_trace("Binding to all netcards\n");
+			binding->ptype.dev = 0;
+		}
+		if (binding) {
+			dev_add_pack(&binding->ptype);
+			spin_lock_bh(&g_binding_lock);
+			binding->next = g_binding;
+			g_binding = binding;
+			spin_unlock_bh(&g_binding_lock);
+		}
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+static int compare_binding_name(char *s1, char *s2)
+{
+	int i;
+	for (i = 0; i < OZ_MAX_BINDING_LEN; i++) {
+		if (*s1 != *s2)
+			return 0;
+		if (!*s1++)
+			return 1;
+		s2++;
+	}
+	return 1;
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+static void pd_stop_all_for_device(struct net_device *net_dev)
+{
+	struct list_head h;
+	struct oz_pd *pd;
+	struct oz_pd *n;
+	INIT_LIST_HEAD(&h);
+	spin_lock_bh(&g_polling_lock);
+	list_for_each_entry_safe(pd, n, &g_pd_list, link) {
+		if (pd->net_dev == net_dev) {
+			list_move(&pd->link, &h);
+			oz_pd_get(pd);
+		}
+	}
+	spin_unlock_bh(&g_polling_lock);
+	while (!list_empty(&h)) {
+		pd = list_first_entry(&h, struct oz_pd, link);
+		oz_pd_stop(pd);
+		oz_pd_put(pd);
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+void oz_binding_remove(char *net_dev)
+{
+	struct oz_binding *binding = 0;
+	struct oz_binding **link;
+	oz_trace("Removing binding: %s\n", net_dev);
+	spin_lock_bh(&g_binding_lock);
+	binding = g_binding;
+	link = &g_binding;
+	while (binding) {
+		if (compare_binding_name(binding->name, net_dev)) {
+			oz_trace("Binding '%s' found\n", net_dev);
+			*link = binding->next;
+			break;
+		} else {
+			link = &binding;
+			binding = binding->next;
+		}
+	}
+	spin_unlock_bh(&g_binding_lock);
+	if (binding) {
+		dev_remove_pack(&binding->ptype);
+		if (binding->ptype.dev) {
+			dev_put(binding->ptype.dev);
+			pd_stop_all_for_device(binding->ptype.dev);
+		}
+		oz_free(binding);
+	}
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+static char *oz_get_next_device_name(char *s, char *dname, int max_size)
+{
+	while (*s == ',')
+		s++;
+	while (*s && (*s != ',') && max_size > 1) {
+		*dname++ = *s++;
+		max_size--;
+	}
+	*dname = 0;
+	return s;
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+int oz_protocol_init(char *devs)
+{
+	skb_queue_head_init(&g_rx_queue);
+	if (devs && (devs[0] == '*')) {
+		oz_binding_add(0);
+	} else {
+		char d[32];
+		while (*devs) {
+			devs = oz_get_next_device_name(devs, d, sizeof(d));
+			if (d[0])
+				oz_binding_add(d);
+		}
+	}
+	init_timer(&g_timer);
+	return 0;
+}
+/*------------------------------------------------------------------------------
+ * Context: process
+ */
+int oz_get_pd_list(struct oz_mac_addr *addr, int max_count)
+{
+	struct oz_pd *pd;
+	struct list_head *e;
+	int count = 0;
+	spin_lock_bh(&g_polling_lock);
+	list_for_each(e, &g_pd_list) {
+		if (count >= max_count)
+			break;
+		pd = container_of(e, struct oz_pd, link);
+		memcpy(&addr[count++], pd->mac_addr, ETH_ALEN);
+	}
+	spin_unlock_bh(&g_polling_lock);
+	return count;
+}
+/*------------------------------------------------------------------------------
+*/
+void oz_polling_lock_bh(void)
+{
+	spin_lock_bh(&g_polling_lock);
+}
+/*------------------------------------------------------------------------------
+*/
+void oz_polling_unlock_bh(void)
+{
+	spin_unlock_bh(&g_polling_lock);
+}
diff --git a/drivers/staging/ozwpan/ozproto.h b/drivers/staging/ozwpan/ozproto.h
new file mode 100644
index 0000000..89aea28
--- /dev/null
+++ b/drivers/staging/ozwpan/ozproto.h
@@ -0,0 +1,69 @@
+/* -----------------------------------------------------------------------------
+ * Copyright (c) 2011 Ozmo Inc
+ * Released under the GNU General Public License Version 2 (GPLv2).
+ * -----------------------------------------------------------------------------
+ */
+#ifndef _OZPROTO_H
+#define _OZPROTO_H
+
+#include <asm/byteorder.h>
+#include "ozconfig.h"
+#include "ozappif.h"
+
+#define OZ_ALLOCATED_SPACE(__x)	(LL_RESERVED_SPACE(__x)+(__x)->needed_tailroom)
+
+/* Converts millisecs to jiffies.
+ */
+#define oz_ms_to_jiffies(__x)	(((__x)*1000)/HZ)
+
+/* Quantum milliseconds.
+ */
+#define OZ_QUANTUM_MS		8
+/* Quantum jiffies
+ */
+#define OZ_QUANTUM_J		(oz_ms_to_jiffies(OZ_QUANTUM_MS))
+/* Default timeouts.
+ */
+#define OZ_CONNECTION_TOUT_J	(2*HZ)
+#define OZ_PRESLEEP_TOUT_J	(11*HZ)
+
+/* Maximun sizes of tx frames. */
+#define OZ_MAX_TX_SIZE		1514
+
+/* Application handler functions.
+ */
+typedef int (*oz_app_init_fn_t)(void);
+typedef void (*oz_app_term_fn_t)(void);
+typedef int (*oz_app_start_fn_t)(struct oz_pd *pd, int resume);
+typedef void (*oz_app_stop_fn_t)(struct oz_pd *pd, int pause);
+typedef void (*oz_app_rx_fn_t)(struct oz_pd *pd, struct oz_elt *elt);
+typedef int (*oz_app_hearbeat_fn_t)(struct oz_pd *pd);
+typedef void (*oz_app_farewell_fn_t)(struct oz_pd *pd, u8 ep_num,
+			u8 *data, u8 len);
+
+struct oz_app_if {
+	oz_app_init_fn_t	init;
+	oz_app_term_fn_t	term;
+	oz_app_start_fn_t	start;
+	oz_app_stop_fn_t	stop;
+	oz_app_rx_fn_t		rx;
+	oz_app_hearbeat_fn_t	heartbeat;
+	oz_app_farewell_fn_t	farewell;
+	int			app_id;
+};
+
+int oz_protocol_init(char *devs);
+void oz_protocol_term(void);
+int oz_get_pd_list(struct oz_mac_addr *addr, int max_count);
+void oz_app_enable(int app_id, int enable);
+struct oz_pd *oz_pd_find(u8 *mac_addr);
+void oz_binding_add(char *net_dev);
+void oz_binding_remove(char *net_dev);
+void oz_timer_add(struct oz_pd *pd, int type, unsigned long due_time,
+		int remove);
+void oz_timer_delete(struct oz_pd *pd, int type);
+void oz_pd_request_heartbeat(struct oz_pd *pd);
+void oz_polling_lock_bh(void);
+void oz_polling_unlock_bh(void);
+
+#endif /* _OZPROTO_H */
-- 
1.7.7.6


--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html


[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux