linux/net/bluetooth/hidp/core.c

1426 lines
37 KiB
C
Raw Normal View History

/*
HIDP implementation for Linux Bluetooth stack (BlueZ).
Copyright (C) 2003-2004 Marcel Holtmann <marcel@holtmann.org>
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
Copyright (C) 2013 David Herrmann <dh.herrmann@gmail.com>
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License version 2 as
published by the Free Software Foundation;
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT OF THIRD PARTY RIGHTS.
IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) AND AUTHOR(S) BE LIABLE FOR ANY
CLAIM, OR ANY SPECIAL INDIRECT OR CONSEQUENTIAL DAMAGES, OR ANY DAMAGES
WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
ALL LIABILITY, INCLUDING LIABILITY FOR INFRINGEMENT OF ANY PATENTS,
COPYRIGHTS, TRADEMARKS OR OTHER RIGHTS, RELATING TO USE OF THIS
SOFTWARE IS DISCLAIMED.
*/
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
#include <linux/kref.h>
#include <linux/module.h>
#include <linux/file.h>
#include <linux/kthread.h>
#include <linux/hidraw.h>
#include <net/bluetooth/bluetooth.h>
#include <net/bluetooth/hci_core.h>
#include <net/bluetooth/l2cap.h>
#include "hidp.h"
#define VERSION "1.2"
static DECLARE_RWSEM(hidp_session_sem);
static LIST_HEAD(hidp_session_list);
static unsigned char hidp_keycode[256] = {
0, 0, 0, 0, 30, 48, 46, 32, 18, 33, 34, 35, 23, 36,
37, 38, 50, 49, 24, 25, 16, 19, 31, 20, 22, 47, 17, 45,
21, 44, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 28, 1,
14, 15, 57, 12, 13, 26, 27, 43, 43, 39, 40, 41, 51, 52,
53, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 87, 88,
99, 70, 119, 110, 102, 104, 111, 107, 109, 106, 105, 108, 103, 69,
98, 55, 74, 78, 96, 79, 80, 81, 75, 76, 77, 71, 72, 73,
82, 83, 86, 127, 116, 117, 183, 184, 185, 186, 187, 188, 189, 190,
191, 192, 193, 194, 134, 138, 130, 132, 128, 129, 131, 137, 133, 135,
136, 113, 115, 114, 0, 0, 0, 121, 0, 89, 93, 124, 92, 94,
95, 0, 0, 0, 122, 123, 90, 91, 85, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
29, 42, 56, 125, 97, 54, 100, 126, 164, 166, 165, 163, 161, 115,
114, 113, 150, 158, 159, 128, 136, 177, 178, 176, 142, 152, 173, 140
};
static unsigned char hidp_mkeyspat[] = { 0x01, 0x01, 0x01, 0x01, 0x01, 0x01 };
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
static int hidp_session_probe(struct l2cap_conn *conn,
struct l2cap_user *user);
static void hidp_session_remove(struct l2cap_conn *conn,
struct l2cap_user *user);
static int hidp_session_thread(void *arg);
static void hidp_session_terminate(struct hidp_session *s);
static void hidp_copy_session(struct hidp_session *session, struct hidp_conninfo *ci)
{
memset(ci, 0, sizeof(*ci));
bacpy(&ci->bdaddr, &session->bdaddr);
ci->flags = session->flags;
ci->state = BT_CONNECTED;
if (session->input) {
ci->vendor = session->input->id.vendor;
ci->product = session->input->id.product;
ci->version = session->input->id.version;
if (session->input->name)
strlcpy(ci->name, session->input->name, 128);
else
strlcpy(ci->name, "HID Boot Device", 128);
} else if (session->hid) {
ci->vendor = session->hid->vendor;
ci->product = session->hid->product;
ci->version = session->hid->version;
strlcpy(ci->name, session->hid->name, 128);
}
}
/* assemble skb, queue message on @transmit and wake up the session thread */
static int hidp_send_message(struct hidp_session *session, struct socket *sock,
struct sk_buff_head *transmit, unsigned char hdr,
const unsigned char *data, int size)
{
struct sk_buff *skb;
struct sock *sk = sock->sk;
BT_DBG("session %p data %p size %d", session, data, size);
if (atomic_read(&session->terminate))
return -EIO;
skb = alloc_skb(size + 1, GFP_ATOMIC);
if (!skb) {
BT_ERR("Can't allocate memory for new frame");
return -ENOMEM;
}
*skb_put(skb, 1) = hdr;
if (data && size > 0)
memcpy(skb_put(skb, size), data, size);
skb_queue_tail(transmit, skb);
wake_up_interruptible(sk_sleep(sk));
return 0;
}
static int hidp_send_ctrl_message(struct hidp_session *session,
unsigned char hdr, const unsigned char *data,
int size)
{
return hidp_send_message(session, session->ctrl_sock,
&session->ctrl_transmit, hdr, data, size);
}
static int hidp_send_intr_message(struct hidp_session *session,
unsigned char hdr, const unsigned char *data,
int size)
{
return hidp_send_message(session, session->intr_sock,
&session->intr_transmit, hdr, data, size);
}
static int hidp_input_event(struct input_dev *dev, unsigned int type,
unsigned int code, int value)
{
struct hidp_session *session = input_get_drvdata(dev);
unsigned char newleds;
unsigned char hdr, data[2];
BT_DBG("session %p type %d code %d value %d",
session, type, code, value);
if (type != EV_LED)
return -1;
newleds = (!!test_bit(LED_KANA, dev->led) << 3) |
(!!test_bit(LED_COMPOSE, dev->led) << 3) |
(!!test_bit(LED_SCROLLL, dev->led) << 2) |
(!!test_bit(LED_CAPSL, dev->led) << 1) |
(!!test_bit(LED_NUML, dev->led));
if (session->leds == newleds)
return 0;
session->leds = newleds;
hdr = HIDP_TRANS_DATA | HIDP_DATA_RTYPE_OUPUT;
data[0] = 0x01;
data[1] = newleds;
return hidp_send_intr_message(session, hdr, data, 2);
}
static void hidp_input_report(struct hidp_session *session, struct sk_buff *skb)
{
struct input_dev *dev = session->input;
unsigned char *keys = session->keys;
unsigned char *udata = skb->data + 1;
signed char *sdata = skb->data + 1;
int i, size = skb->len - 1;
switch (skb->data[0]) {
case 0x01: /* Keyboard report */
for (i = 0; i < 8; i++)
input_report_key(dev, hidp_keycode[i + 224], (udata[0] >> i) & 1);
/* If all the key codes have been set to 0x01, it means
* too many keys were pressed at the same time. */
if (!memcmp(udata + 2, hidp_mkeyspat, 6))
break;
for (i = 2; i < 8; i++) {
if (keys[i] > 3 && memscan(udata + 2, keys[i], 6) == udata + 8) {
if (hidp_keycode[keys[i]])
input_report_key(dev, hidp_keycode[keys[i]], 0);
else
BT_ERR("Unknown key (scancode %#x) released.", keys[i]);
}
if (udata[i] > 3 && memscan(keys + 2, udata[i], 6) == keys + 8) {
if (hidp_keycode[udata[i]])
input_report_key(dev, hidp_keycode[udata[i]], 1);
else
BT_ERR("Unknown key (scancode %#x) pressed.", udata[i]);
}
}
memcpy(keys, udata, 8);
break;
case 0x02: /* Mouse report */
input_report_key(dev, BTN_LEFT, sdata[0] & 0x01);
input_report_key(dev, BTN_RIGHT, sdata[0] & 0x02);
input_report_key(dev, BTN_MIDDLE, sdata[0] & 0x04);
input_report_key(dev, BTN_SIDE, sdata[0] & 0x08);
input_report_key(dev, BTN_EXTRA, sdata[0] & 0x10);
input_report_rel(dev, REL_X, sdata[1]);
input_report_rel(dev, REL_Y, sdata[2]);
if (size > 3)
input_report_rel(dev, REL_WHEEL, sdata[3]);
break;
}
input_sync(dev);
}
static int hidp_send_report(struct hidp_session *session, struct hid_report *report)
{
HID: fix data access in implement() implement() is setting bytes in LE data stream. In case the data is not aligned to 64bits, it reads past the allocated buffer. It doesn't really change any value there (it's properly bitmasked), but in case that this read past the boundary hits a page boundary, pagefault happens when accessing 64bits of 'x' in implement(), and kernel oopses. This happens much more often when numbered reports are in use, as the initial 8bit skip in the buffer makes the whole process work on values which are not aligned to 64bits. This problem dates back to attempts in 2005 and 2006 to make implement() and extract() as generic as possible, and even back then the problem was realized by Adam Kroperlin, but falsely assumed to be impossible to cause any harm: http://www.mail-archive.com/linux-usb-devel@lists.sourceforge.net/msg47690.html I have made several attempts at fixing it "on the spot" directly in implement(), but the results were horrible; the special casing for processing last 64bit chunk and switching to different math makes it unreadable mess. I therefore took a path to allocate a few bytes more which will never make it into final report, but are there as a cushion for all the 64bit math operations happening in implement() and extract(). All callers of hid_output_report() are converted at the same time to allocate the buffer by newly introduced hid_alloc_report_buf() helper. Bruno noticed that the whole raw_size test can be dropped as well, as hid_alloc_report_buf() makes sure that the buffer is always of a proper size. Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-07-10 17:56:27 +00:00
unsigned char hdr;
u8 *buf;
int rsize, ret;
HID: fix data access in implement() implement() is setting bytes in LE data stream. In case the data is not aligned to 64bits, it reads past the allocated buffer. It doesn't really change any value there (it's properly bitmasked), but in case that this read past the boundary hits a page boundary, pagefault happens when accessing 64bits of 'x' in implement(), and kernel oopses. This happens much more often when numbered reports are in use, as the initial 8bit skip in the buffer makes the whole process work on values which are not aligned to 64bits. This problem dates back to attempts in 2005 and 2006 to make implement() and extract() as generic as possible, and even back then the problem was realized by Adam Kroperlin, but falsely assumed to be impossible to cause any harm: http://www.mail-archive.com/linux-usb-devel@lists.sourceforge.net/msg47690.html I have made several attempts at fixing it "on the spot" directly in implement(), but the results were horrible; the special casing for processing last 64bit chunk and switching to different math makes it unreadable mess. I therefore took a path to allocate a few bytes more which will never make it into final report, but are there as a cushion for all the 64bit math operations happening in implement() and extract(). All callers of hid_output_report() are converted at the same time to allocate the buffer by newly introduced hid_alloc_report_buf() helper. Bruno noticed that the whole raw_size test can be dropped as well, as hid_alloc_report_buf() makes sure that the buffer is always of a proper size. Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-07-10 17:56:27 +00:00
buf = hid_alloc_report_buf(report, GFP_ATOMIC);
if (!buf)
return -EIO;
hid_output_report(report, buf);
hdr = HIDP_TRANS_DATA | HIDP_DATA_RTYPE_OUPUT;
rsize = ((report->size - 1) >> 3) + 1 + (report->id > 0);
HID: fix data access in implement() implement() is setting bytes in LE data stream. In case the data is not aligned to 64bits, it reads past the allocated buffer. It doesn't really change any value there (it's properly bitmasked), but in case that this read past the boundary hits a page boundary, pagefault happens when accessing 64bits of 'x' in implement(), and kernel oopses. This happens much more often when numbered reports are in use, as the initial 8bit skip in the buffer makes the whole process work on values which are not aligned to 64bits. This problem dates back to attempts in 2005 and 2006 to make implement() and extract() as generic as possible, and even back then the problem was realized by Adam Kroperlin, but falsely assumed to be impossible to cause any harm: http://www.mail-archive.com/linux-usb-devel@lists.sourceforge.net/msg47690.html I have made several attempts at fixing it "on the spot" directly in implement(), but the results were horrible; the special casing for processing last 64bit chunk and switching to different math makes it unreadable mess. I therefore took a path to allocate a few bytes more which will never make it into final report, but are there as a cushion for all the 64bit math operations happening in implement() and extract(). All callers of hid_output_report() are converted at the same time to allocate the buffer by newly introduced hid_alloc_report_buf() helper. Bruno noticed that the whole raw_size test can be dropped as well, as hid_alloc_report_buf() makes sure that the buffer is always of a proper size. Reviewed-by: Benjamin Tissoires <benjamin.tissoires@redhat.com> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-07-10 17:56:27 +00:00
ret = hidp_send_intr_message(session, hdr, buf, rsize);
kfree(buf);
return ret;
}
static int hidp_get_raw_report(struct hid_device *hid,
unsigned char report_number,
unsigned char *data, size_t count,
unsigned char report_type)
{
struct hidp_session *session = hid->driver_data;
struct sk_buff *skb;
size_t len;
int numbered_reports = hid->report_enum[report_type].numbered;
int ret;
if (atomic_read(&session->terminate))
return -EIO;
switch (report_type) {
case HID_FEATURE_REPORT:
report_type = HIDP_TRANS_GET_REPORT | HIDP_DATA_RTYPE_FEATURE;
break;
case HID_INPUT_REPORT:
report_type = HIDP_TRANS_GET_REPORT | HIDP_DATA_RTYPE_INPUT;
break;
case HID_OUTPUT_REPORT:
report_type = HIDP_TRANS_GET_REPORT | HIDP_DATA_RTYPE_OUPUT;
break;
default:
return -EINVAL;
}
if (mutex_lock_interruptible(&session->report_mutex))
return -ERESTARTSYS;
/* Set up our wait, and send the report request to the device. */
session->waiting_report_type = report_type & HIDP_DATA_RTYPE_MASK;
session->waiting_report_number = numbered_reports ? report_number : -1;
set_bit(HIDP_WAITING_FOR_RETURN, &session->flags);
data[0] = report_number;
ret = hidp_send_ctrl_message(session, report_type, data, 1);
if (ret)
goto err;
/* Wait for the return of the report. The returned report
gets put in session->report_return. */
while (test_bit(HIDP_WAITING_FOR_RETURN, &session->flags) &&
!atomic_read(&session->terminate)) {
int res;
res = wait_event_interruptible_timeout(session->report_queue,
!test_bit(HIDP_WAITING_FOR_RETURN, &session->flags)
|| atomic_read(&session->terminate),
5*HZ);
if (res == 0) {
/* timeout */
ret = -EIO;
goto err;
}
if (res < 0) {
/* signal */
ret = -ERESTARTSYS;
goto err;
}
}
skb = session->report_return;
if (skb) {
len = skb->len < count ? skb->len : count;
memcpy(data, skb->data, len);
kfree_skb(skb);
session->report_return = NULL;
} else {
/* Device returned a HANDSHAKE, indicating protocol error. */
len = -EIO;
}
clear_bit(HIDP_WAITING_FOR_RETURN, &session->flags);
mutex_unlock(&session->report_mutex);
return len;
err:
clear_bit(HIDP_WAITING_FOR_RETURN, &session->flags);
mutex_unlock(&session->report_mutex);
return ret;
}
static int hidp_output_raw_report(struct hid_device *hid, unsigned char *data, size_t count,
unsigned char report_type)
{
struct hidp_session *session = hid->driver_data;
int ret;
if (report_type == HID_OUTPUT_REPORT) {
report_type = HIDP_TRANS_DATA | HIDP_DATA_RTYPE_OUPUT;
return hidp_send_intr_message(session, report_type,
data, count);
} else if (report_type != HID_FEATURE_REPORT) {
return -EINVAL;
}
if (mutex_lock_interruptible(&session->report_mutex))
return -ERESTARTSYS;
/* Set up our wait, and send the report request to the device. */
set_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags);
report_type = HIDP_TRANS_SET_REPORT | HIDP_DATA_RTYPE_FEATURE;
ret = hidp_send_ctrl_message(session, report_type, data, count);
if (ret)
goto err;
/* Wait for the ACK from the device. */
while (test_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags) &&
!atomic_read(&session->terminate)) {
int res;
res = wait_event_interruptible_timeout(session->report_queue,
!test_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags)
|| atomic_read(&session->terminate),
10*HZ);
if (res == 0) {
/* timeout */
ret = -EIO;
goto err;
}
if (res < 0) {
/* signal */
ret = -ERESTARTSYS;
goto err;
}
}
if (!session->output_report_success) {
ret = -EIO;
goto err;
}
ret = count;
err:
clear_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags);
mutex_unlock(&session->report_mutex);
return ret;
}
static void hidp_idle_timeout(unsigned long arg)
{
struct hidp_session *session = (struct hidp_session *) arg;
hidp_session_terminate(session);
}
static void hidp_set_timer(struct hidp_session *session)
{
if (session->idle_to > 0)
mod_timer(&session->timer, jiffies + HZ * session->idle_to);
}
static void hidp_del_timer(struct hidp_session *session)
{
if (session->idle_to > 0)
del_timer(&session->timer);
}
static void hidp_process_handshake(struct hidp_session *session,
unsigned char param)
{
BT_DBG("session %p param 0x%02x", session, param);
session->output_report_success = 0; /* default condition */
switch (param) {
case HIDP_HSHK_SUCCESSFUL:
/* FIXME: Call into SET_ GET_ handlers here */
session->output_report_success = 1;
break;
case HIDP_HSHK_NOT_READY:
case HIDP_HSHK_ERR_INVALID_REPORT_ID:
case HIDP_HSHK_ERR_UNSUPPORTED_REQUEST:
case HIDP_HSHK_ERR_INVALID_PARAMETER:
if (test_and_clear_bit(HIDP_WAITING_FOR_RETURN, &session->flags))
wake_up_interruptible(&session->report_queue);
/* FIXME: Call into SET_ GET_ handlers here */
break;
case HIDP_HSHK_ERR_UNKNOWN:
break;
case HIDP_HSHK_ERR_FATAL:
/* Device requests a reboot, as this is the only way this error
* can be recovered. */
hidp_send_ctrl_message(session,
HIDP_TRANS_HID_CONTROL | HIDP_CTRL_SOFT_RESET, NULL, 0);
break;
default:
hidp_send_ctrl_message(session,
HIDP_TRANS_HANDSHAKE | HIDP_HSHK_ERR_INVALID_PARAMETER, NULL, 0);
break;
}
/* Wake up the waiting thread. */
if (test_and_clear_bit(HIDP_WAITING_FOR_SEND_ACK, &session->flags))
wake_up_interruptible(&session->report_queue);
}
static void hidp_process_hid_control(struct hidp_session *session,
unsigned char param)
{
BT_DBG("session %p param 0x%02x", session, param);
if (param == HIDP_CTRL_VIRTUAL_CABLE_UNPLUG) {
/* Flush the transmit queues */
skb_queue_purge(&session->ctrl_transmit);
skb_queue_purge(&session->intr_transmit);
hidp_session_terminate(session);
}
}
/* Returns true if the passed-in skb should be freed by the caller. */
static int hidp_process_data(struct hidp_session *session, struct sk_buff *skb,
unsigned char param)
{
int done_with_skb = 1;
BT_DBG("session %p skb %p len %d param 0x%02x", session, skb, skb->len, param);
switch (param) {
case HIDP_DATA_RTYPE_INPUT:
hidp_set_timer(session);
if (session->input)
hidp_input_report(session, skb);
if (session->hid)
hid_input_report(session->hid, HID_INPUT_REPORT, skb->data, skb->len, 0);
break;
case HIDP_DATA_RTYPE_OTHER:
case HIDP_DATA_RTYPE_OUPUT:
case HIDP_DATA_RTYPE_FEATURE:
break;
default:
hidp_send_ctrl_message(session,
HIDP_TRANS_HANDSHAKE | HIDP_HSHK_ERR_INVALID_PARAMETER, NULL, 0);
}
if (test_bit(HIDP_WAITING_FOR_RETURN, &session->flags) &&
param == session->waiting_report_type) {
if (session->waiting_report_number < 0 ||
session->waiting_report_number == skb->data[0]) {
/* hidp_get_raw_report() is waiting on this report. */
session->report_return = skb;
done_with_skb = 0;
clear_bit(HIDP_WAITING_FOR_RETURN, &session->flags);
wake_up_interruptible(&session->report_queue);
}
}
return done_with_skb;
}
static void hidp_recv_ctrl_frame(struct hidp_session *session,
struct sk_buff *skb)
{
unsigned char hdr, type, param;
int free_skb = 1;
BT_DBG("session %p skb %p len %d", session, skb, skb->len);
hdr = skb->data[0];
skb_pull(skb, 1);
type = hdr & HIDP_HEADER_TRANS_MASK;
param = hdr & HIDP_HEADER_PARAM_MASK;
switch (type) {
case HIDP_TRANS_HANDSHAKE:
hidp_process_handshake(session, param);
break;
case HIDP_TRANS_HID_CONTROL:
hidp_process_hid_control(session, param);
break;
case HIDP_TRANS_DATA:
free_skb = hidp_process_data(session, skb, param);
break;
default:
hidp_send_ctrl_message(session,
HIDP_TRANS_HANDSHAKE | HIDP_HSHK_ERR_UNSUPPORTED_REQUEST, NULL, 0);
break;
}
if (free_skb)
kfree_skb(skb);
}
static void hidp_recv_intr_frame(struct hidp_session *session,
struct sk_buff *skb)
{
unsigned char hdr;
BT_DBG("session %p skb %p len %d", session, skb, skb->len);
hdr = skb->data[0];
skb_pull(skb, 1);
if (hdr == (HIDP_TRANS_DATA | HIDP_DATA_RTYPE_INPUT)) {
hidp_set_timer(session);
if (session->input)
hidp_input_report(session, skb);
if (session->hid) {
hid_input_report(session->hid, HID_INPUT_REPORT, skb->data, skb->len, 1);
BT_DBG("report len %d", skb->len);
}
} else {
BT_DBG("Unsupported protocol header 0x%02x", hdr);
}
kfree_skb(skb);
}
static int hidp_send_frame(struct socket *sock, unsigned char *data, int len)
{
struct kvec iv = { data, len };
struct msghdr msg;
BT_DBG("sock %p data %p len %d", sock, data, len);
if (!len)
return 0;
memset(&msg, 0, sizeof(msg));
return kernel_sendmsg(sock, &msg, &iv, 1, len);
}
/* dequeue message from @transmit and send via @sock */
static void hidp_process_transmit(struct hidp_session *session,
struct sk_buff_head *transmit,
struct socket *sock)
{
struct sk_buff *skb;
int ret;
BT_DBG("session %p", session);
while ((skb = skb_dequeue(transmit))) {
ret = hidp_send_frame(sock, skb->data, skb->len);
if (ret == -EAGAIN) {
skb_queue_head(transmit, skb);
break;
} else if (ret < 0) {
hidp_session_terminate(session);
kfree_skb(skb);
break;
}
hidp_set_timer(session);
kfree_skb(skb);
}
}
static int hidp_setup_input(struct hidp_session *session,
struct hidp_connadd_req *req)
{
struct input_dev *input;
int i;
input = input_allocate_device();
if (!input)
return -ENOMEM;
session->input = input;
input_set_drvdata(input, session);
input->name = "Bluetooth HID Boot Protocol Device";
input->id.bustype = BUS_BLUETOOTH;
input->id.vendor = req->vendor;
input->id.product = req->product;
input->id.version = req->version;
if (req->subclass & 0x40) {
set_bit(EV_KEY, input->evbit);
set_bit(EV_LED, input->evbit);
set_bit(EV_REP, input->evbit);
set_bit(LED_NUML, input->ledbit);
set_bit(LED_CAPSL, input->ledbit);
set_bit(LED_SCROLLL, input->ledbit);
set_bit(LED_COMPOSE, input->ledbit);
set_bit(LED_KANA, input->ledbit);
for (i = 0; i < sizeof(hidp_keycode); i++)
set_bit(hidp_keycode[i], input->keybit);
clear_bit(0, input->keybit);
}
if (req->subclass & 0x80) {
input->evbit[0] = BIT_MASK(EV_KEY) | BIT_MASK(EV_REL);
input->keybit[BIT_WORD(BTN_MOUSE)] = BIT_MASK(BTN_LEFT) |
BIT_MASK(BTN_RIGHT) | BIT_MASK(BTN_MIDDLE);
input->relbit[0] = BIT_MASK(REL_X) | BIT_MASK(REL_Y);
input->keybit[BIT_WORD(BTN_MOUSE)] |= BIT_MASK(BTN_SIDE) |
BIT_MASK(BTN_EXTRA);
input->relbit[0] |= BIT_MASK(REL_WHEEL);
}
input->dev.parent = &session->conn->hcon->dev;
input->event = hidp_input_event;
return 0;
}
static int hidp_open(struct hid_device *hid)
{
return 0;
}
static void hidp_close(struct hid_device *hid)
{
}
static int hidp_parse(struct hid_device *hid)
{
struct hidp_session *session = hid->driver_data;
return hid_parse_report(session->hid, session->rd_data,
session->rd_size);
}
static int hidp_start(struct hid_device *hid)
{
struct hidp_session *session = hid->driver_data;
struct hid_report *report;
if (hid->quirks & HID_QUIRK_NO_INIT_REPORTS)
return 0;
list_for_each_entry(report, &hid->report_enum[HID_INPUT_REPORT].
report_list, list)
hidp_send_report(session, report);
list_for_each_entry(report, &hid->report_enum[HID_FEATURE_REPORT].
report_list, list)
hidp_send_report(session, report);
return 0;
}
static void hidp_stop(struct hid_device *hid)
{
struct hidp_session *session = hid->driver_data;
skb_queue_purge(&session->ctrl_transmit);
skb_queue_purge(&session->intr_transmit);
hid->claimed = 0;
}
static struct hid_ll_driver hidp_hid_driver = {
.parse = hidp_parse,
.start = hidp_start,
.stop = hidp_stop,
.open = hidp_open,
.close = hidp_close,
};
/* This function sets up the hid device. It does not add it
to the HID system. That is done in hidp_add_connection(). */
static int hidp_setup_hid(struct hidp_session *session,
struct hidp_connadd_req *req)
{
struct hid_device *hid;
int err;
session->rd_data = kzalloc(req->rd_size, GFP_KERNEL);
if (!session->rd_data)
return -ENOMEM;
if (copy_from_user(session->rd_data, req->rd_data, req->rd_size)) {
err = -EFAULT;
goto fault;
}
session->rd_size = req->rd_size;
hid = hid_allocate_device();
if (IS_ERR(hid)) {
err = PTR_ERR(hid);
goto fault;
}
session->hid = hid;
hid->driver_data = session;
hid->bus = BUS_BLUETOOTH;
hid->vendor = req->vendor;
hid->product = req->product;
hid->version = req->version;
hid->country = req->country;
strncpy(hid->name, req->name, sizeof(req->name) - 1);
snprintf(hid->phys, sizeof(hid->phys), "%pMR",
&bt_sk(session->ctrl_sock->sk)->src);
snprintf(hid->uniq, sizeof(hid->uniq), "%pMR",
&bt_sk(session->ctrl_sock->sk)->dst);
hid->dev.parent = &session->conn->hcon->dev;
hid->ll_driver = &hidp_hid_driver;
hid->hid_get_raw_report = hidp_get_raw_report;
hid->hid_output_raw_report = hidp_output_raw_report;
/* True if device is blacklisted in drivers/hid/hid-core.c */
if (hid_ignore(hid)) {
hid_destroy_device(session->hid);
session->hid = NULL;
return -ENODEV;
}
return 0;
fault:
kfree(session->rd_data);
session->rd_data = NULL;
return err;
}
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
/* initialize session devices */
static int hidp_session_dev_init(struct hidp_session *session,
struct hidp_connadd_req *req)
{
int ret;
if (req->rd_size > 0) {
ret = hidp_setup_hid(session, req);
if (ret && ret != -ENODEV)
return ret;
}
if (!session->hid) {
ret = hidp_setup_input(session, req);
if (ret < 0)
return ret;
}
return 0;
}
/* destroy session devices */
static void hidp_session_dev_destroy(struct hidp_session *session)
{
if (session->hid)
put_device(&session->hid->dev);
else if (session->input)
input_put_device(session->input);
kfree(session->rd_data);
session->rd_data = NULL;
}
/* add HID/input devices to their underlying bus systems */
static int hidp_session_dev_add(struct hidp_session *session)
{
int ret;
/* Both HID and input systems drop a ref-count when unregistering the
* device but they don't take a ref-count when registering them. Work
* around this by explicitly taking a refcount during registration
* which is dropped automatically by unregistering the devices. */
if (session->hid) {
ret = hid_add_device(session->hid);
if (ret)
return ret;
get_device(&session->hid->dev);
} else if (session->input) {
ret = input_register_device(session->input);
if (ret)
return ret;
input_get_device(session->input);
}
return 0;
}
/* remove HID/input devices from their bus systems */
static void hidp_session_dev_del(struct hidp_session *session)
{
if (session->hid)
hid_destroy_device(session->hid);
else if (session->input)
input_unregister_device(session->input);
}
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
/*
* Asynchronous device registration
* HID device drivers might want to perform I/O during initialization to
* detect device types. Therefore, call device registration in a separate
* worker so the HIDP thread can schedule I/O operations.
* Note that this must be called after the worker thread was initialized
* successfully. This will then add the devices and increase session state
* on success, otherwise it will terminate the session thread.
*/
static void hidp_session_dev_work(struct work_struct *work)
{
struct hidp_session *session = container_of(work,
struct hidp_session,
dev_init);
int ret;
ret = hidp_session_dev_add(session);
if (!ret)
atomic_inc(&session->state);
else
hidp_session_terminate(session);
}
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
/*
* Create new session object
* Allocate session object, initialize static fields, copy input data into the
* object and take a reference to all sub-objects.
* This returns 0 on success and puts a pointer to the new session object in
* \out. Otherwise, an error code is returned.
* The new session object has an initial ref-count of 1.
*/
static int hidp_session_new(struct hidp_session **out, const bdaddr_t *bdaddr,
struct socket *ctrl_sock,
struct socket *intr_sock,
struct hidp_connadd_req *req,
struct l2cap_conn *conn)
{
struct hidp_session *session;
int ret;
struct bt_sock *ctrl, *intr;
ctrl = bt_sk(ctrl_sock->sk);
intr = bt_sk(intr_sock->sk);
session = kzalloc(sizeof(*session), GFP_KERNEL);
if (!session)
return -ENOMEM;
/* object and runtime management */
kref_init(&session->ref);
atomic_set(&session->state, HIDP_SESSION_IDLING);
init_waitqueue_head(&session->state_queue);
session->flags = req->flags & (1 << HIDP_BLUETOOTH_VENDOR_ID);
/* connection management */
bacpy(&session->bdaddr, bdaddr);
session->conn = conn;
session->user.probe = hidp_session_probe;
session->user.remove = hidp_session_remove;
session->ctrl_sock = ctrl_sock;
session->intr_sock = intr_sock;
skb_queue_head_init(&session->ctrl_transmit);
skb_queue_head_init(&session->intr_transmit);
session->ctrl_mtu = min_t(uint, l2cap_pi(ctrl)->chan->omtu,
l2cap_pi(ctrl)->chan->imtu);
session->intr_mtu = min_t(uint, l2cap_pi(intr)->chan->omtu,
l2cap_pi(intr)->chan->imtu);
session->idle_to = req->idle_to;
/* device management */
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
INIT_WORK(&session->dev_init, hidp_session_dev_work);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
setup_timer(&session->timer, hidp_idle_timeout,
(unsigned long)session);
/* session data */
mutex_init(&session->report_mutex);
init_waitqueue_head(&session->report_queue);
ret = hidp_session_dev_init(session, req);
if (ret)
goto err_free;
l2cap_conn_get(session->conn);
get_file(session->intr_sock->file);
get_file(session->ctrl_sock->file);
*out = session;
return 0;
err_free:
kfree(session);
return ret;
}
/* increase ref-count of the given session by one */
static void hidp_session_get(struct hidp_session *session)
{
kref_get(&session->ref);
}
/* release callback */
static void session_free(struct kref *ref)
{
struct hidp_session *session = container_of(ref, struct hidp_session,
ref);
hidp_session_dev_destroy(session);
skb_queue_purge(&session->ctrl_transmit);
skb_queue_purge(&session->intr_transmit);
fput(session->intr_sock->file);
fput(session->ctrl_sock->file);
l2cap_conn_put(session->conn);
kfree(session);
}
/* decrease ref-count of the given session by one */
static void hidp_session_put(struct hidp_session *session)
{
kref_put(&session->ref, session_free);
}
/*
* Search the list of active sessions for a session with target address
* \bdaddr. You must hold at least a read-lock on \hidp_session_sem. As long as
* you do not release this lock, the session objects cannot vanish and you can
* safely take a reference to the session yourself.
*/
static struct hidp_session *__hidp_session_find(const bdaddr_t *bdaddr)
{
struct hidp_session *session;
list_for_each_entry(session, &hidp_session_list, list) {
if (!bacmp(bdaddr, &session->bdaddr))
return session;
}
return NULL;
}
/*
* Same as __hidp_session_find() but no locks must be held. This also takes a
* reference of the returned session (if non-NULL) so you must drop this
* reference if you no longer use the object.
*/
static struct hidp_session *hidp_session_find(const bdaddr_t *bdaddr)
{
struct hidp_session *session;
down_read(&hidp_session_sem);
session = __hidp_session_find(bdaddr);
if (session)
hidp_session_get(session);
up_read(&hidp_session_sem);
return session;
}
/*
* Start session synchronously
* This starts a session thread and waits until initialization
* is done or returns an error if it couldn't be started.
* If this returns 0 the session thread is up and running. You must call
* hipd_session_stop_sync() before deleting any runtime resources.
*/
static int hidp_session_start_sync(struct hidp_session *session)
{
unsigned int vendor, product;
if (session->hid) {
vendor = session->hid->vendor;
product = session->hid->product;
} else if (session->input) {
vendor = session->input->id.vendor;
product = session->input->id.product;
} else {
vendor = 0x0000;
product = 0x0000;
}
session->task = kthread_run(hidp_session_thread, session,
"khidpd_%04x%04x", vendor, product);
if (IS_ERR(session->task))
return PTR_ERR(session->task);
while (atomic_read(&session->state) <= HIDP_SESSION_IDLING)
wait_event(session->state_queue,
atomic_read(&session->state) > HIDP_SESSION_IDLING);
return 0;
}
/*
* Terminate session thread
* Wake up session thread and notify it to stop. This is asynchronous and
* returns immediately. Call this whenever a runtime error occurs and you want
* the session to stop.
* Note: wake_up_process() performs any necessary memory-barriers for us.
*/
static void hidp_session_terminate(struct hidp_session *session)
{
atomic_inc(&session->terminate);
wake_up_process(session->task);
}
/*
* Probe HIDP session
* This is called from the l2cap_conn core when our l2cap_user object is bound
* to the hci-connection. We get the session via the \user object and can now
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
* start the session thread, link it into the global session list and
* schedule HID/input device registration.
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
* The global session-list owns its own reference to the session object so you
* can drop your own reference after registering the l2cap_user object.
*/
static int hidp_session_probe(struct l2cap_conn *conn,
struct l2cap_user *user)
{
struct hidp_session *session = container_of(user,
struct hidp_session,
user);
struct hidp_session *s;
int ret;
down_write(&hidp_session_sem);
/* check that no other session for this device exists */
s = __hidp_session_find(&session->bdaddr);
if (s) {
ret = -EEXIST;
goto out_unlock;
}
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
if (session->input) {
ret = hidp_session_dev_add(session);
if (ret)
goto out_unlock;
}
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
ret = hidp_session_start_sync(session);
if (ret)
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
goto out_del;
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
/* HID device registration is async to allow I/O during probe */
if (session->input)
atomic_inc(&session->state);
else
schedule_work(&session->dev_init);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
hidp_session_get(session);
list_add(&session->list, &hidp_session_list);
ret = 0;
goto out_unlock;
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
out_del:
if (session->input)
hidp_session_dev_del(session);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
out_unlock:
up_write(&hidp_session_sem);
return ret;
}
/*
* Remove HIDP session
* Called from the l2cap_conn core when either we explicitly unregistered
* the l2cap_user object or if the underlying connection is shut down.
* We signal the hidp-session thread to shut down, unregister the HID/input
* devices and unlink the session from the global list.
* This drops the reference to the session that is owned by the global
* session-list.
* Note: We _must_ not synchronosly wait for the session-thread to shut down.
* This is, because the session-thread might be waiting for an HCI lock that is
* held while we are called. Therefore, we only unregister the devices and
* notify the session-thread to terminate. The thread itself owns a reference
* to the session object so it can safely shut down.
*/
static void hidp_session_remove(struct l2cap_conn *conn,
struct l2cap_user *user)
{
struct hidp_session *session = container_of(user,
struct hidp_session,
user);
down_write(&hidp_session_sem);
hidp_session_terminate(session);
HID: Bluetooth: hidp: register HID devices async While l2cap_user callbacks are running, the whole hci_dev is locked. Even if we would add more fine-grained locking to HCI core, it would still be called from the non-reentrant rx work-queue and thus block the event processing. However, if we want to perform synchronous I/O during HID device registration (eg., to perform device-detection), we need the HCI core to be able to dispatch incoming data. Therefore, we now move device-registration to a separate worker. The HCI core can continue running and we add devices asynchronously in another kernel thread. Device removal is synchronized and waits for the worker to exit before calling the usual device removal functions. If l2cap_user->remove is called before the thread registered the devices, we set "terminate" to true and the thread will skip it. If l2cap_user->remove is called after it, we notice this as the device is no longer in HIDP_SESSION_PREPARING state and simply unregister the device as we did before. There is no new deadlock as we now call hidp_session_add_dev() with one lock less held (the HCI lock) and it cannot itself call back into HCI as it was called with the HCI-lock held before. One might wonder whether this can block during device unregistration. But we set "terminate" to true and wake the HIDP thread up _before_ unregistering the HID/input devices. Therefore, all pending HID I/O operations are canceled. All further I/O attempts will fail with ENODEV or EIO. So all latency we can get are few context-switches, but no timeouts or blocking I/O waits! This change also prepares for a long standing HID bug. All HID devices that register power_supply devices need to be able to handle callbacks during registration (a power_supply oddity that cannot easily be fixed). So with this patch available, we can allow HID I/O during registration by calling the recently introduced hid_device_io_start/stop helpers, which currently are a no-op for bluetooth due to this locking. Note that we cannot do the same for input devices. input-core doesn't allow us to call input_event() asynchronously to input_register_device(), which HID-core kindly allows (for good reasons). Fixing input-core to allow this isn't as easy as it sounds and is, beside simplifying HIDP, not really an improvement. Hence, we still register input devices synchronously as we did before. Only HID devices are registered asynchronously. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Jiri Kosina <jkosina@suse.cz> Acked-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Tested-by: Daniel Nicoletti <dantti12@gmail.com> Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2013-05-23 11:10:25 +00:00
cancel_work_sync(&session->dev_init);
if (session->input ||
atomic_read(&session->state) > HIDP_SESSION_PREPARING)
hidp_session_dev_del(session);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
list_del(&session->list);
up_write(&hidp_session_sem);
hidp_session_put(session);
}
/*
* Session Worker
* This performs the actual main-loop of the HIDP worker. We first check
* whether the underlying connection is still alive, then parse all pending
* messages and finally send all outstanding messages.
*/
static void hidp_session_run(struct hidp_session *session)
{
struct sock *ctrl_sk = session->ctrl_sock->sk;
struct sock *intr_sk = session->intr_sock->sk;
struct sk_buff *skb;
for (;;) {
/*
* This thread can be woken up two ways:
* - You call hidp_session_terminate() which sets the
* session->terminate flag and wakes this thread up.
* - Via modifying the socket state of ctrl/intr_sock. This
* thread is woken up by ->sk_state_changed().
*
* Note: set_current_state() performs any necessary
* memory-barriers for us.
*/
set_current_state(TASK_INTERRUPTIBLE);
if (atomic_read(&session->terminate))
break;
if (ctrl_sk->sk_state != BT_CONNECTED ||
intr_sk->sk_state != BT_CONNECTED)
break;
/* parse incoming intr-skbs */
while ((skb = skb_dequeue(&intr_sk->sk_receive_queue))) {
skb_orphan(skb);
if (!skb_linearize(skb))
hidp_recv_intr_frame(session, skb);
else
kfree_skb(skb);
}
/* send pending intr-skbs */
hidp_process_transmit(session, &session->intr_transmit,
session->intr_sock);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
/* parse incoming ctrl-skbs */
while ((skb = skb_dequeue(&ctrl_sk->sk_receive_queue))) {
skb_orphan(skb);
if (!skb_linearize(skb))
hidp_recv_ctrl_frame(session, skb);
else
kfree_skb(skb);
}
/* send pending ctrl-skbs */
hidp_process_transmit(session, &session->ctrl_transmit,
session->ctrl_sock);
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
schedule();
}
atomic_inc(&session->terminate);
set_current_state(TASK_RUNNING);
}
/*
* HIDP session thread
* This thread runs the I/O for a single HIDP session. Startup is synchronous
* which allows us to take references to ourself here instead of doing that in
* the caller.
* When we are ready to run we notify the caller and call hidp_session_run().
*/
static int hidp_session_thread(void *arg)
{
struct hidp_session *session = arg;
wait_queue_t ctrl_wait, intr_wait;
BT_DBG("session %p", session);
/* initialize runtime environment */
hidp_session_get(session);
__module_get(THIS_MODULE);
set_user_nice(current, -15);
hidp_set_timer(session);
init_waitqueue_entry(&ctrl_wait, current);
init_waitqueue_entry(&intr_wait, current);
add_wait_queue(sk_sleep(session->ctrl_sock->sk), &ctrl_wait);
add_wait_queue(sk_sleep(session->intr_sock->sk), &intr_wait);
/* This memory barrier is paired with wq_has_sleeper(). See
* sock_poll_wait() for more information why this is needed. */
smp_mb();
/* notify synchronous startup that we're ready */
atomic_inc(&session->state);
wake_up(&session->state_queue);
/* run session */
hidp_session_run(session);
/* cleanup runtime environment */
remove_wait_queue(sk_sleep(session->intr_sock->sk), &intr_wait);
remove_wait_queue(sk_sleep(session->intr_sock->sk), &ctrl_wait);
wake_up_interruptible(&session->report_queue);
hidp_del_timer(session);
/*
* If we stopped ourself due to any internal signal, we should try to
* unregister our own session here to avoid having it linger until the
* parent l2cap_conn dies or user-space cleans it up.
* This does not deadlock as we don't do any synchronous shutdown.
* Instead, this call has the same semantics as if user-space tried to
* delete the session.
*/
l2cap_unregister_user(session->conn, &session->user);
hidp_session_put(session);
module_put_and_exit(0);
return 0;
}
static int hidp_verify_sockets(struct socket *ctrl_sock,
struct socket *intr_sock)
{
struct bt_sock *ctrl, *intr;
struct hidp_session *session;
if (!l2cap_is_socket(ctrl_sock) || !l2cap_is_socket(intr_sock))
return -EINVAL;
ctrl = bt_sk(ctrl_sock->sk);
intr = bt_sk(intr_sock->sk);
if (bacmp(&ctrl->src, &intr->src) || bacmp(&ctrl->dst, &intr->dst))
return -ENOTUNIQ;
if (ctrl->sk.sk_state != BT_CONNECTED ||
intr->sk.sk_state != BT_CONNECTED)
return -EBADFD;
/* early session check, we check again during session registration */
session = hidp_session_find(&ctrl->dst);
if (session) {
hidp_session_put(session);
return -EEXIST;
}
return 0;
}
int hidp_connection_add(struct hidp_connadd_req *req,
struct socket *ctrl_sock,
struct socket *intr_sock)
{
struct hidp_session *session;
struct l2cap_conn *conn;
struct l2cap_chan *chan = l2cap_pi(ctrl_sock->sk)->chan;
int ret;
ret = hidp_verify_sockets(ctrl_sock, intr_sock);
if (ret)
return ret;
conn = NULL;
l2cap_chan_lock(chan);
if (chan->conn) {
l2cap_conn_get(chan->conn);
conn = chan->conn;
}
l2cap_chan_unlock(chan);
if (!conn)
return -EBADFD;
ret = hidp_session_new(&session, &bt_sk(ctrl_sock->sk)->dst, ctrl_sock,
intr_sock, req, conn);
if (ret)
goto out_conn;
ret = l2cap_register_user(conn, &session->user);
if (ret)
goto out_session;
ret = 0;
out_session:
hidp_session_put(session);
out_conn:
l2cap_conn_put(conn);
return ret;
}
int hidp_connection_del(struct hidp_conndel_req *req)
{
struct hidp_session *session;
session = hidp_session_find(&req->bdaddr);
if (!session)
return -ENOENT;
if (req->flags & (1 << HIDP_VIRTUAL_CABLE_UNPLUG))
hidp_send_ctrl_message(session,
HIDP_TRANS_HID_CONTROL |
HIDP_CTRL_VIRTUAL_CABLE_UNPLUG,
NULL, 0);
else
l2cap_unregister_user(session->conn, &session->user);
hidp_session_put(session);
return 0;
}
int hidp_get_connlist(struct hidp_connlist_req *req)
{
struct hidp_session *session;
int err = 0, n = 0;
BT_DBG("");
down_read(&hidp_session_sem);
list_for_each_entry(session, &hidp_session_list, list) {
struct hidp_conninfo ci;
hidp_copy_session(session, &ci);
if (copy_to_user(req->ci, &ci, sizeof(ci))) {
err = -EFAULT;
break;
}
if (++n >= req->cnum)
break;
req->ci++;
}
req->cnum = n;
up_read(&hidp_session_sem);
return err;
}
int hidp_get_conninfo(struct hidp_conninfo *ci)
{
struct hidp_session *session;
session = hidp_session_find(&ci->bdaddr);
if (session) {
hidp_copy_session(session, ci);
hidp_session_put(session);
}
return session ? 0 : -ENOENT;
}
static int __init hidp_init(void)
{
BT_INFO("HIDP (Human Interface Emulation) ver %s", VERSION);
return hidp_init_sockets();
}
static void __exit hidp_exit(void)
{
hidp_cleanup_sockets();
}
module_init(hidp_init);
module_exit(hidp_exit);
MODULE_AUTHOR("Marcel Holtmann <marcel@holtmann.org>");
Bluetooth: hidp: add new session-management helpers This is a rewrite of the HIDP session management. It implements HIDP as an l2cap_user sub-module so we get proper notification when the underlying connection goes away. The helpers are not yet used but only added in this commit. The old session management is still used and will be removed in a following patch. The old session-management was flawed. Hotplugging is horribly broken and we have no way of getting notified when the underlying connection goes down. The whole idea of removing the HID/input sub-devices from within the session itself is broken and suffers from major dead-locks. We never can guarantee that the session can unregister itself as long as we use synchronous shutdowns. This can only work with asynchronous shutdowns. However, in this case we _must_ be able to unregister the session from the outside as otherwise the l2cap_conn object might be unlinked before we are. The new session-management is based on l2cap_user. There is only one way how to add a session and how to delete a session: "probe" and "remove" callbacks from l2cap_user. This guarantees that the session can be registered and unregistered at _any_ time without any synchronous shutdown. On the other hand, much work has been put into proper session-refcounting. We can unregister/unlink the session only if we can guarantee that it will stay alive. But for asynchronous shutdowns we never know when the last user goes away so we must use proper ref-counting. The old ->conn field has been renamed to ->hconn so we can reuse ->conn in the new session management. No other existing HIDP code is modified. Signed-off-by: David Herrmann <dh.herrmann@gmail.com> Acked-by: Marcel Holtmann <marcel@holtmann.org> Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
2013-04-06 18:28:46 +00:00
MODULE_AUTHOR("David Herrmann <dh.herrmann@gmail.com>");
MODULE_DESCRIPTION("Bluetooth HIDP ver " VERSION);
MODULE_VERSION(VERSION);
MODULE_LICENSE("GPL");
MODULE_ALIAS("bt-proto-6");