]> git.karo-electronics.de Git - karo-tx-linux.git/commitdiff
[media] cec: add HDMI CEC framework (api)
authorHans Verkuil <hans.verkuil@cisco.com>
Sat, 25 Jun 2016 12:44:43 +0000 (09:44 -0300)
committerMauro Carvalho Chehab <mchehab@s-opensource.com>
Tue, 28 Jun 2016 12:59:24 +0000 (09:59 -0300)
The added HDMI CEC framework provides a generic kernel interface for
HDMI CEC devices.

Note that the CEC framework is added to staging/media and that the
cec.h and cec-funcs.h headers are not exported yet. While the kABI
is mature, I would prefer to allow the uABI some more time before
it is mainlined in case it needs more tweaks.

This adds the cec-api.c source that deals with the public CEC API
and the Kconfig/Makefile plumbing.

The MAINTAINERS file is also updated.

Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
[k.debski@samsung.com: code cleanup and fixes]
Signed-off-by: Kamil Debski <kamil@wypas.org>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
MAINTAINERS
drivers/staging/media/Kconfig
drivers/staging/media/Makefile
drivers/staging/media/cec/Kconfig [new file with mode: 0644]
drivers/staging/media/cec/Makefile [new file with mode: 0644]
drivers/staging/media/cec/cec-api.c [new file with mode: 0644]

index 02299fd2db18dbfd1cfef644929368e9a2ecd673..848d37dafd90ba6a98f9408db58f0f21c6152396 100644 (file)
@@ -2847,6 +2847,22 @@ F:       drivers/net/ieee802154/cc2520.c
 F:     include/linux/spi/cc2520.h
 F:     Documentation/devicetree/bindings/net/ieee802154/cc2520.txt
 
+CEC DRIVER
+M:     Hans Verkuil <hans.verkuil@cisco.com>
+L:     linux-media@vger.kernel.org
+T:     git git://linuxtv.org/media_tree.git
+W:     http://linuxtv.org
+S:     Supported
+F:     Documentation/cec.txt
+F:     Documentation/DocBook/media/v4l/cec*
+F:     drivers/staging/media/cec/
+F:     drivers/media/cec-edid.c
+F:     drivers/media/rc/keymaps/rc-cec.c
+F:     include/media/cec.h
+F:     include/media/cec-edid.h
+F:     include/linux/cec.h
+F:     include/linux/cec-funcs.h
+
 CELL BROADBAND ENGINE ARCHITECTURE
 M:     Arnd Bergmann <arnd@arndb.de>
 L:     linuxppc-dev@lists.ozlabs.org
index ee9186843371f7ce49def99c7ff6207db34eb59d..7ce679eb87e1161f0523ec0ac047a3a7a37bc37e 100644 (file)
@@ -21,6 +21,8 @@ if STAGING_MEDIA
 # Please keep them in alphabetic order
 source "drivers/staging/media/bcm2048/Kconfig"
 
+source "drivers/staging/media/cec/Kconfig"
+
 source "drivers/staging/media/cxd2099/Kconfig"
 
 source "drivers/staging/media/davinci_vpfe/Kconfig"
index 8c05d0ace0fb5a88d8f7cb95656768e40de66353..2d213dd54bc37fcd43312f1d0425df856d8fe7ad 100644 (file)
@@ -1,4 +1,5 @@
 obj-$(CONFIG_I2C_BCM2048)      += bcm2048/
+obj-$(CONFIG_MEDIA_CEC)                += cec/
 obj-$(CONFIG_DVB_CXD2099)      += cxd2099/
 obj-$(CONFIG_LIRC_STAGING)     += lirc/
 obj-$(CONFIG_VIDEO_DM365_VPFE) += davinci_vpfe/
diff --git a/drivers/staging/media/cec/Kconfig b/drivers/staging/media/cec/Kconfig
new file mode 100644 (file)
index 0000000..8a7acee
--- /dev/null
@@ -0,0 +1,14 @@
+config MEDIA_CEC
+       tristate "CEC API (EXPERIMENTAL)"
+       select MEDIA_CEC_EDID
+       ---help---
+         Enable the CEC API.
+
+         To compile this driver as a module, choose M here: the
+         module will be called cec.
+
+config MEDIA_CEC_DEBUG
+       bool "CEC debugfs interface (EXPERIMENTAL)"
+       depends on MEDIA_CEC && DEBUG_FS
+       ---help---
+         Turns on the DebugFS interface for CEC devices.
diff --git a/drivers/staging/media/cec/Makefile b/drivers/staging/media/cec/Makefile
new file mode 100644 (file)
index 0000000..426ef73
--- /dev/null
@@ -0,0 +1,3 @@
+cec-objs := cec-core.o cec-adap.o cec-api.o
+
+obj-$(CONFIG_MEDIA_CEC) += cec.o
diff --git a/drivers/staging/media/cec/cec-api.c b/drivers/staging/media/cec/cec-api.c
new file mode 100644 (file)
index 0000000..d7cba7a
--- /dev/null
@@ -0,0 +1,578 @@
+/*
+ * cec-api.c - HDMI Consumer Electronics Control framework - API
+ *
+ * Copyright 2016 Cisco Systems, Inc. and/or its affiliates. All rights reserved.
+ *
+ * This program is free software; you may redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; version 2 of the License.
+ *
+ * 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. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/errno.h>
+#include <linux/init.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/kmod.h>
+#include <linux/ktime.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/uaccess.h>
+#include <linux/version.h>
+
+#include "cec-priv.h"
+
+static inline struct cec_devnode *cec_devnode_data(struct file *filp)
+{
+       struct cec_fh *fh = filp->private_data;
+
+       return &fh->adap->devnode;
+}
+
+/* CEC file operations */
+
+static unsigned int cec_poll(struct file *filp,
+                            struct poll_table_struct *poll)
+{
+       struct cec_devnode *devnode = cec_devnode_data(filp);
+       struct cec_fh *fh = filp->private_data;
+       struct cec_adapter *adap = fh->adap;
+       unsigned int res = 0;
+
+       if (!devnode->registered)
+               return POLLERR | POLLHUP;
+       mutex_lock(&adap->lock);
+       if (adap->is_configured)
+               res |= POLLOUT | POLLWRNORM;
+       if (fh->queued_msgs)
+               res |= POLLIN | POLLRDNORM;
+       if (fh->pending_events)
+               res |= POLLPRI;
+       poll_wait(filp, &fh->wait, poll);
+       mutex_unlock(&adap->lock);
+       return res;
+}
+
+static bool cec_is_busy(const struct cec_adapter *adap,
+                       const struct cec_fh *fh)
+{
+       bool valid_initiator = adap->cec_initiator && adap->cec_initiator == fh;
+       bool valid_follower = adap->cec_follower && adap->cec_follower == fh;
+
+       /*
+        * Exclusive initiators and followers can always access the CEC adapter
+        */
+       if (valid_initiator || valid_follower)
+               return false;
+       /*
+        * All others can only access the CEC adapter if there is no
+        * exclusive initiator and they are in INITIATOR mode.
+        */
+       return adap->cec_initiator ||
+              fh->mode_initiator == CEC_MODE_NO_INITIATOR;
+}
+
+static long cec_adap_g_caps(struct cec_adapter *adap,
+                           struct cec_caps __user *parg)
+{
+       struct cec_caps caps = {};
+
+       strlcpy(caps.driver, adap->devnode.parent->driver->name,
+               sizeof(caps.driver));
+       strlcpy(caps.name, adap->name, sizeof(caps.name));
+       caps.available_log_addrs = adap->available_log_addrs;
+       caps.capabilities = adap->capabilities;
+       caps.version = LINUX_VERSION_CODE;
+       if (copy_to_user(parg, &caps, sizeof(caps)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_adap_g_phys_addr(struct cec_adapter *adap,
+                                __u16 __user *parg)
+{
+       u16 phys_addr;
+
+       mutex_lock(&adap->lock);
+       phys_addr = adap->phys_addr;
+       mutex_unlock(&adap->lock);
+       if (copy_to_user(parg, &phys_addr, sizeof(phys_addr)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_adap_s_phys_addr(struct cec_adapter *adap, struct cec_fh *fh,
+                                bool block, __u16 __user *parg)
+{
+       u16 phys_addr;
+       long err;
+
+       if (!(adap->capabilities & CEC_CAP_PHYS_ADDR))
+               return -ENOTTY;
+       if (copy_from_user(&phys_addr, parg, sizeof(phys_addr)))
+               return -EFAULT;
+
+       err = cec_phys_addr_validate(phys_addr, NULL, NULL);
+       if (err)
+               return err;
+       mutex_lock(&adap->lock);
+       if (cec_is_busy(adap, fh))
+               err = -EBUSY;
+       else
+               __cec_s_phys_addr(adap, phys_addr, block);
+       mutex_unlock(&adap->lock);
+       return err;
+}
+
+static long cec_adap_g_log_addrs(struct cec_adapter *adap,
+                                struct cec_log_addrs __user *parg)
+{
+       struct cec_log_addrs log_addrs;
+
+       mutex_lock(&adap->lock);
+       log_addrs = adap->log_addrs;
+       if (!adap->is_configured)
+               memset(log_addrs.log_addr, CEC_LOG_ADDR_INVALID,
+                      sizeof(log_addrs.log_addr));
+       mutex_unlock(&adap->lock);
+
+       if (copy_to_user(parg, &log_addrs, sizeof(log_addrs)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_adap_s_log_addrs(struct cec_adapter *adap, struct cec_fh *fh,
+                                bool block, struct cec_log_addrs __user *parg)
+{
+       struct cec_log_addrs log_addrs;
+       long err = -EBUSY;
+
+       if (!(adap->capabilities & CEC_CAP_LOG_ADDRS))
+               return -ENOTTY;
+       if (copy_from_user(&log_addrs, parg, sizeof(log_addrs)))
+               return -EFAULT;
+       log_addrs.flags = 0;
+       mutex_lock(&adap->lock);
+       if (!adap->is_configuring &&
+           (!log_addrs.num_log_addrs || !adap->is_configured) &&
+           !cec_is_busy(adap, fh)) {
+               err = __cec_s_log_addrs(adap, &log_addrs, block);
+               if (!err)
+                       log_addrs = adap->log_addrs;
+       }
+       mutex_unlock(&adap->lock);
+       if (err)
+               return err;
+       if (copy_to_user(parg, &log_addrs, sizeof(log_addrs)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_transmit(struct cec_adapter *adap, struct cec_fh *fh,
+                        bool block, struct cec_msg __user *parg)
+{
+       struct cec_msg msg = {};
+       long err = 0;
+
+       if (!(adap->capabilities & CEC_CAP_TRANSMIT))
+               return -ENOTTY;
+       if (copy_from_user(&msg, parg, sizeof(msg)))
+               return -EFAULT;
+       mutex_lock(&adap->lock);
+       if (!adap->is_configured) {
+               err = -ENONET;
+       } else if (cec_is_busy(adap, fh)) {
+               err = -EBUSY;
+       } else {
+               if (!block || !msg.reply)
+                       fh = NULL;
+               err = cec_transmit_msg_fh(adap, &msg, fh, block);
+       }
+       mutex_unlock(&adap->lock);
+       if (err)
+               return err;
+       if (copy_to_user(parg, &msg, sizeof(msg)))
+               return -EFAULT;
+       return 0;
+}
+
+/* Called by CEC_RECEIVE: wait for a message to arrive */
+static int cec_receive_msg(struct cec_fh *fh, struct cec_msg *msg, bool block)
+{
+       int res;
+
+       do {
+               mutex_lock(&fh->lock);
+               /* Are there received messages queued up? */
+               if (fh->queued_msgs) {
+                       /* Yes, return the first one */
+                       struct cec_msg_entry *entry =
+                               list_first_entry(&fh->msgs,
+                                                struct cec_msg_entry, list);
+
+                       list_del(&entry->list);
+                       *msg = entry->msg;
+                       kfree(entry);
+                       fh->queued_msgs--;
+                       mutex_unlock(&fh->lock);
+                       return 0;
+               }
+
+               /* No, return EAGAIN in non-blocking mode or wait */
+               mutex_unlock(&fh->lock);
+
+               /* Return when in non-blocking mode */
+               if (!block)
+                       return -EAGAIN;
+
+               if (msg->timeout) {
+                       /* The user specified a timeout */
+                       res = wait_event_interruptible_timeout(fh->wait,
+                                                              fh->queued_msgs,
+                               msecs_to_jiffies(msg->timeout));
+                       if (res == 0)
+                               res = -ETIMEDOUT;
+                       else if (res > 0)
+                               res = 0;
+               } else {
+                       /* Wait indefinitely */
+                       res = wait_event_interruptible(fh->wait,
+                                                      fh->queued_msgs);
+               }
+               /* Exit on error, otherwise loop to get the new message */
+       } while (!res);
+       return res;
+}
+
+static long cec_receive(struct cec_adapter *adap, struct cec_fh *fh,
+                       bool block, struct cec_msg __user *parg)
+{
+       struct cec_msg msg = {};
+       long err = 0;
+
+       if (copy_from_user(&msg, parg, sizeof(msg)))
+               return -EFAULT;
+       mutex_lock(&adap->lock);
+       if (!adap->is_configured)
+               err = -ENONET;
+       mutex_unlock(&adap->lock);
+       if (err)
+               return err;
+
+       err = cec_receive_msg(fh, &msg, block);
+       if (err)
+               return err;
+       if (copy_to_user(parg, &msg, sizeof(msg)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_dqevent(struct cec_adapter *adap, struct cec_fh *fh,
+                       bool block, struct cec_event __user *parg)
+{
+       struct cec_event *ev = NULL;
+       u64 ts = ~0ULL;
+       unsigned int i;
+       long err = 0;
+
+       mutex_lock(&fh->lock);
+       while (!fh->pending_events && block) {
+               mutex_unlock(&fh->lock);
+               err = wait_event_interruptible(fh->wait, fh->pending_events);
+               if (err)
+                       return err;
+               mutex_lock(&fh->lock);
+       }
+
+       /* Find the oldest event */
+       for (i = 0; i < CEC_NUM_EVENTS; i++) {
+               if (fh->pending_events & (1 << (i + 1)) &&
+                   fh->events[i].ts <= ts) {
+                       ev = &fh->events[i];
+                       ts = ev->ts;
+               }
+       }
+       if (!ev) {
+               err = -EAGAIN;
+               goto unlock;
+       }
+
+       if (copy_to_user(parg, ev, sizeof(*ev))) {
+               err = -EFAULT;
+               goto unlock;
+       }
+
+       fh->pending_events &= ~(1 << ev->event);
+
+unlock:
+       mutex_unlock(&fh->lock);
+       return err;
+}
+
+static long cec_g_mode(struct cec_adapter *adap, struct cec_fh *fh,
+                      u32 __user *parg)
+{
+       u32 mode = fh->mode_initiator | fh->mode_follower;
+
+       if (copy_to_user(parg, &mode, sizeof(mode)))
+               return -EFAULT;
+       return 0;
+}
+
+static long cec_s_mode(struct cec_adapter *adap, struct cec_fh *fh,
+                      u32 __user *parg)
+{
+       u32 mode;
+       u8 mode_initiator;
+       u8 mode_follower;
+       long err = 0;
+
+       if (copy_from_user(&mode, parg, sizeof(mode)))
+               return -EFAULT;
+       if (mode & ~(CEC_MODE_INITIATOR_MSK | CEC_MODE_FOLLOWER_MSK))
+               return -EINVAL;
+
+       mode_initiator = mode & CEC_MODE_INITIATOR_MSK;
+       mode_follower = mode & CEC_MODE_FOLLOWER_MSK;
+
+       if (mode_initiator > CEC_MODE_EXCL_INITIATOR ||
+           mode_follower > CEC_MODE_MONITOR_ALL)
+               return -EINVAL;
+
+       if (mode_follower == CEC_MODE_MONITOR_ALL &&
+           !(adap->capabilities & CEC_CAP_MONITOR_ALL))
+               return -EINVAL;
+
+       /* Follower modes should always be able to send CEC messages */
+       if ((mode_initiator == CEC_MODE_NO_INITIATOR ||
+            !(adap->capabilities & CEC_CAP_TRANSMIT)) &&
+           mode_follower >= CEC_MODE_FOLLOWER &&
+           mode_follower <= CEC_MODE_EXCL_FOLLOWER_PASSTHRU)
+               return -EINVAL;
+
+       /* Monitor modes require CEC_MODE_NO_INITIATOR */
+       if (mode_initiator && mode_follower >= CEC_MODE_MONITOR)
+               return -EINVAL;
+
+       /* Monitor modes require CAP_NET_ADMIN */
+       if (mode_follower >= CEC_MODE_MONITOR && !capable(CAP_NET_ADMIN))
+               return -EPERM;
+
+       mutex_lock(&adap->lock);
+       /*
+        * You can't become exclusive follower if someone else already
+        * has that job.
+        */
+       if ((mode_follower == CEC_MODE_EXCL_FOLLOWER ||
+            mode_follower == CEC_MODE_EXCL_FOLLOWER_PASSTHRU) &&
+           adap->cec_follower && adap->cec_follower != fh)
+               err = -EBUSY;
+       /*
+        * You can't become exclusive initiator if someone else already
+        * has that job.
+        */
+       if (mode_initiator == CEC_MODE_EXCL_INITIATOR &&
+           adap->cec_initiator && adap->cec_initiator != fh)
+               err = -EBUSY;
+
+       if (!err) {
+               bool old_mon_all = fh->mode_follower == CEC_MODE_MONITOR_ALL;
+               bool new_mon_all = mode_follower == CEC_MODE_MONITOR_ALL;
+
+               if (old_mon_all != new_mon_all) {
+                       if (new_mon_all)
+                               err = cec_monitor_all_cnt_inc(adap);
+                       else
+                               cec_monitor_all_cnt_dec(adap);
+               }
+       }
+
+       if (err) {
+               mutex_unlock(&adap->lock);
+               return err;
+       }
+
+       if (fh->mode_follower == CEC_MODE_FOLLOWER)
+               adap->follower_cnt--;
+       if (mode_follower == CEC_MODE_FOLLOWER)
+               adap->follower_cnt++;
+       if (mode_follower == CEC_MODE_EXCL_FOLLOWER ||
+           mode_follower == CEC_MODE_EXCL_FOLLOWER_PASSTHRU) {
+               adap->passthrough =
+                       mode_follower == CEC_MODE_EXCL_FOLLOWER_PASSTHRU;
+               adap->cec_follower = fh;
+       } else if (adap->cec_follower == fh) {
+               adap->passthrough = false;
+               adap->cec_follower = NULL;
+       }
+       if (mode_initiator == CEC_MODE_EXCL_INITIATOR)
+               adap->cec_initiator = fh;
+       else if (adap->cec_initiator == fh)
+               adap->cec_initiator = NULL;
+       fh->mode_initiator = mode_initiator;
+       fh->mode_follower = mode_follower;
+       mutex_unlock(&adap->lock);
+       return 0;
+}
+
+static long cec_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
+{
+       struct cec_devnode *devnode = cec_devnode_data(filp);
+       struct cec_fh *fh = filp->private_data;
+       struct cec_adapter *adap = fh->adap;
+       bool block = !(filp->f_flags & O_NONBLOCK);
+       void __user *parg = (void __user *)arg;
+
+       if (!devnode->registered)
+               return -EIO;
+
+       switch (cmd) {
+       case CEC_ADAP_G_CAPS:
+               return cec_adap_g_caps(adap, parg);
+
+       case CEC_ADAP_G_PHYS_ADDR:
+               return cec_adap_g_phys_addr(adap, parg);
+
+       case CEC_ADAP_S_PHYS_ADDR:
+               return cec_adap_s_phys_addr(adap, fh, block, parg);
+
+       case CEC_ADAP_G_LOG_ADDRS:
+               return cec_adap_g_log_addrs(adap, parg);
+
+       case CEC_ADAP_S_LOG_ADDRS:
+               return cec_adap_s_log_addrs(adap, fh, block, parg);
+
+       case CEC_TRANSMIT:
+               return cec_transmit(adap, fh, block, parg);
+
+       case CEC_RECEIVE:
+               return cec_receive(adap, fh, block, parg);
+
+       case CEC_DQEVENT:
+               return cec_dqevent(adap, fh, block, parg);
+
+       case CEC_G_MODE:
+               return cec_g_mode(adap, fh, parg);
+
+       case CEC_S_MODE:
+               return cec_s_mode(adap, fh, parg);
+
+       default:
+               return -ENOTTY;
+       }
+}
+
+static int cec_open(struct inode *inode, struct file *filp)
+{
+       struct cec_devnode *devnode =
+               container_of(inode->i_cdev, struct cec_devnode, cdev);
+       struct cec_adapter *adap = to_cec_adapter(devnode);
+       struct cec_fh *fh = kzalloc(sizeof(*fh), GFP_KERNEL);
+       /*
+        * Initial events that are automatically sent when the cec device is
+        * opened.
+        */
+       struct cec_event ev_state = {
+               .event = CEC_EVENT_STATE_CHANGE,
+               .flags = CEC_EVENT_FL_INITIAL_STATE,
+       };
+       int err;
+
+       if (!fh)
+               return -ENOMEM;
+
+       INIT_LIST_HEAD(&fh->msgs);
+       INIT_LIST_HEAD(&fh->xfer_list);
+       mutex_init(&fh->lock);
+       init_waitqueue_head(&fh->wait);
+
+       fh->mode_initiator = CEC_MODE_INITIATOR;
+       fh->adap = adap;
+
+       err = cec_get_device(devnode);
+       if (err) {
+               kfree(fh);
+               return err;
+       }
+
+       filp->private_data = fh;
+
+       mutex_lock(&devnode->fhs_lock);
+       /* Queue up initial state events */
+       ev_state.state_change.phys_addr = adap->phys_addr;
+       ev_state.state_change.log_addr_mask = adap->log_addrs.log_addr_mask;
+       cec_queue_event_fh(fh, &ev_state, 0);
+
+       list_add(&fh->list, &devnode->fhs);
+       mutex_unlock(&devnode->fhs_lock);
+
+       return 0;
+}
+
+/* Override for the release function */
+static int cec_release(struct inode *inode, struct file *filp)
+{
+       struct cec_devnode *devnode = cec_devnode_data(filp);
+       struct cec_adapter *adap = to_cec_adapter(devnode);
+       struct cec_fh *fh = filp->private_data;
+
+       mutex_lock(&adap->lock);
+       if (adap->cec_initiator == fh)
+               adap->cec_initiator = NULL;
+       if (adap->cec_follower == fh) {
+               adap->cec_follower = NULL;
+               adap->passthrough = false;
+       }
+       if (fh->mode_follower == CEC_MODE_FOLLOWER)
+               adap->follower_cnt--;
+       if (fh->mode_follower == CEC_MODE_MONITOR_ALL)
+               cec_monitor_all_cnt_dec(adap);
+       mutex_unlock(&adap->lock);
+
+       mutex_lock(&devnode->fhs_lock);
+       list_del(&fh->list);
+       mutex_unlock(&devnode->fhs_lock);
+
+       /* Unhook pending transmits from this filehandle. */
+       mutex_lock(&adap->lock);
+       while (!list_empty(&fh->xfer_list)) {
+               struct cec_data *data =
+                       list_first_entry(&fh->xfer_list, struct cec_data, xfer_list);
+
+               data->blocking = false;
+               data->fh = NULL;
+               list_del(&data->xfer_list);
+       }
+       mutex_unlock(&adap->lock);
+       while (!list_empty(&fh->msgs)) {
+               struct cec_msg_entry *entry =
+                       list_first_entry(&fh->msgs, struct cec_msg_entry, list);
+
+               list_del(&entry->list);
+               kfree(entry);
+       }
+       kfree(fh);
+
+       cec_put_device(devnode);
+       filp->private_data = NULL;
+       return 0;
+}
+
+const struct file_operations cec_devnode_fops = {
+       .owner = THIS_MODULE,
+       .open = cec_open,
+       .unlocked_ioctl = cec_ioctl,
+       .release = cec_release,
+       .poll = cec_poll,
+       .llseek = no_llseek,
+};