[lttng-ust.git] / libtracing / relay.c

/*
 * ltt/ltt-relay.c
 *
 * (C) Copyright 2005-2008 - Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
 *
 * LTTng lockless buffer space management (reader/writer).
 *
 * Author:
 *      Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
 *
 * Inspired from LTT :
 *  Karim Yaghmour (karim@opersys.com)
 *  Tom Zanussi (zanussi@us.ibm.com)
 *  Bob Wisniewski (bob@watson.ibm.com)
 * And from K42 :
 *  Bob Wisniewski (bob@watson.ibm.com)
 *
 * Changelog:
 *  08/10/08, Cleanup.
 *  19/10/05, Complete lockless mechanism.
 *  27/05/05, Modular redesign and rewrite.
 *
 * Userspace reader semantic :
 * while (poll fd != POLLHUP) {
 *   - ioctl RELAY_GET_SUBBUF_SIZE
 *   while (1) {
 *     - ioctl GET_SUBBUF
 *     - splice 1 subbuffer worth of data to a pipe
 *     - splice the data from pipe to disk/network
 *     - ioctl PUT_SUBBUF, check error value
 *       if err val < 0, previous subbuffer was corrupted.
 *   }
 * }
 */

#include <linux/time.h>
#include <linux/ltt-tracer.h>
#include <linux/ltt-relay.h>
#include <linux/module.h>
#include <linux/string.h>
#include <linux/slab.h>
#include <linux/init.h>
#include <linux/rcupdate.h>
#include <linux/sched.h>
#include <linux/bitops.h>
#include <linux/fs.h>
#include <linux/smp_lock.h>
#include <linux/debugfs.h>
#include <linux/stat.h>
#include <linux/cpu.h>
#include <linux/pipe_fs_i.h>
#include <linux/splice.h>
#include <asm/atomic.h>
#include <asm/local.h>

#if 0
#define printk_dbg(fmt, args...) printk(fmt, args)
#else
#define printk_dbg(fmt, args...)
#endif

/* LTTng lockless logging buffer info */
struct ltt_channel_buf_struct {
        /* First 32 bytes cache-hot cacheline */
        local_t offset;                 /* Current offset in the buffer */
        local_t *commit_count;          /* Commit count per sub-buffer */
        atomic_long_t consumed;         /*
                                         * Current offset in the buffer
                                         * standard atomic access (shared)
                                         */
        unsigned long last_tsc;         /*
                                         * Last timestamp written in the buffer.
                                         */
        /* End of first 32 bytes cacheline */
        atomic_long_t active_readers;   /*
                                         * Active readers count
                                         * standard atomic access (shared)
                                         */
        local_t events_lost;
        local_t corrupted_subbuffers;
        spinlock_t full_lock;           /*
                                         * buffer full condition spinlock, only
                                         * for userspace tracing blocking mode
                                         * synchronization with reader.
                                         */
        wait_queue_head_t write_wait;   /*
                                         * Wait queue for blocking user space
                                         * writers
                                         */
        atomic_t wakeup_readers;        /* Boolean : wakeup readers waiting ? */
} ____cacheline_aligned;

/*
 * Last TSC comparison functions. Check if the current TSC overflows
 * LTT_TSC_BITS bits from the last TSC read. Reads and writes last_tsc
 * atomically.
 */

#if (BITS_PER_LONG == 32)
static inline void save_last_tsc(struct ltt_channel_buf_struct *ltt_buf,
                                        u64 tsc)
{
        ltt_buf->last_tsc = (unsigned long)(tsc >> LTT_TSC_BITS);
}

static inline int last_tsc_overflow(struct ltt_channel_buf_struct *ltt_buf,
                                        u64 tsc)
{
        unsigned long tsc_shifted = (unsigned long)(tsc >> LTT_TSC_BITS);

        if (unlikely((tsc_shifted - ltt_buf->last_tsc)))
                return 1;
        else
                return 0;
}
#else
static inline void save_last_tsc(struct ltt_channel_buf_struct *ltt_buf,
                                        u64 tsc)
{
        ltt_buf->last_tsc = (unsigned long)tsc;
}

static inline int last_tsc_overflow(struct ltt_channel_buf_struct *ltt_buf,
                                        u64 tsc)
{
        if (unlikely((tsc - ltt_buf->last_tsc) >> LTT_TSC_BITS))
                return 1;
        else
                return 0;
}
#endif

static struct file_operations ltt_file_operations;

/*
 * A switch is done during tracing or as a final flush after tracing (so it
 * won't write in the new sub-buffer).
 */
enum force_switch_mode { FORCE_ACTIVE, FORCE_FLUSH };

static int ltt_relay_create_buffer(struct ltt_trace_struct *trace,
                struct ltt_channel_struct *ltt_chan,
                struct rchan_buf *buf,
                unsigned int cpu,
                unsigned int n_subbufs);

static void ltt_relay_destroy_buffer(struct ltt_channel_struct *ltt_chan,
                unsigned int cpu);

static void ltt_force_switch(struct rchan_buf *buf,
                enum force_switch_mode mode);

/*
 * Trace callbacks
 */
static void ltt_buffer_begin_callback(struct rchan_buf *buf,
                        u64 tsc, unsigned int subbuf_idx)
{
        struct ltt_channel_struct *channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_subbuffer_header *header =
                (struct ltt_subbuffer_header *)
                        ltt_relay_offset_address(buf,
                                subbuf_idx * buf->chan->subbuf_size);

        header->cycle_count_begin = tsc;
        header->lost_size = 0xFFFFFFFF; /* for debugging */
        header->buf_size = buf->chan->subbuf_size;
        ltt_write_trace_header(channel->trace, header);
}

/*
 * offset is assumed to never be 0 here : never deliver a completely empty
 * subbuffer. The lost size is between 0 and subbuf_size-1.
 */
static notrace void ltt_buffer_end_callback(struct rchan_buf *buf,
                u64 tsc, unsigned int offset, unsigned int subbuf_idx)
{
        struct ltt_channel_struct *channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(channel->buf, buf->cpu);
        struct ltt_subbuffer_header *header =
                (struct ltt_subbuffer_header *)
                        ltt_relay_offset_address(buf,
                                subbuf_idx * buf->chan->subbuf_size);

        header->lost_size = SUBBUF_OFFSET((buf->chan->subbuf_size - offset),
                                buf->chan);
        header->cycle_count_end = tsc;
        header->events_lost = local_read(&ltt_buf->events_lost);
        header->subbuf_corrupt = local_read(&ltt_buf->corrupted_subbuffers);
}

static notrace void ltt_deliver(struct rchan_buf *buf, unsigned int subbuf_idx,
                void *subbuf)
{
        struct ltt_channel_struct *channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(channel->buf, buf->cpu);

        atomic_set(&ltt_buf->wakeup_readers, 1);
}

static struct dentry *ltt_create_buf_file_callback(const char *filename,
                struct dentry *parent, int mode,
                struct rchan_buf *buf)
{
        struct ltt_channel_struct *ltt_chan;
        int err;
        struct dentry *dentry;

        ltt_chan = buf->chan->private_data;
        err = ltt_relay_create_buffer(ltt_chan->trace, ltt_chan,
                                        buf, buf->cpu,
                                        buf->chan->n_subbufs);
        if (err)
                return ERR_PTR(err);

        dentry = debugfs_create_file(filename, mode, parent, buf,
                        &ltt_file_operations);
        if (!dentry)
                goto error;
        return dentry;
error:
        ltt_relay_destroy_buffer(ltt_chan, buf->cpu);
        return NULL;
}

static int ltt_remove_buf_file_callback(struct dentry *dentry)
{
        struct rchan_buf *buf = dentry->d_inode->i_private;
        struct ltt_channel_struct *ltt_chan = buf->chan->private_data;

        debugfs_remove(dentry);
        ltt_relay_destroy_buffer(ltt_chan, buf->cpu);

        return 0;
}

/*
 * Wake writers :
 *
 * This must be done after the trace is removed from the RCU list so that there
 * are no stalled writers.
 */
static void ltt_relay_wake_writers(struct ltt_channel_buf_struct *ltt_buf)
{

        if (waitqueue_active(&ltt_buf->write_wait))
                wake_up_interruptible(&ltt_buf->write_wait);
}

/*
 * This function should not be called from NMI interrupt context
 */
static notrace void ltt_buf_unfull(struct rchan_buf *buf,
                unsigned int subbuf_idx,
                long offset)
{
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);

        ltt_relay_wake_writers(ltt_buf);
}

/**
 *      ltt_open - open file op for ltt files
 *      @inode: opened inode
 *      @file: opened file
 *
 *      Open implementation. Makes sure only one open instance of a buffer is
 *      done at a given moment.
 */
static int ltt_open(struct inode *inode, struct file *file)
{
        struct rchan_buf *buf = inode->i_private;
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);

        if (!atomic_long_add_unless(&ltt_buf->active_readers, 1, 1))
                return -EBUSY;
        return ltt_relay_file_operations.open(inode, file);
}

/**
 *      ltt_release - release file op for ltt files
 *      @inode: opened inode
 *      @file: opened file
 *
 *      Release implementation.
 */
static int ltt_release(struct inode *inode, struct file *file)
{
        struct rchan_buf *buf = inode->i_private;
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);
        int ret;

        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
        atomic_long_dec(&ltt_buf->active_readers);
        ret = ltt_relay_file_operations.release(inode, file);
        WARN_ON(ret);
        return ret;
}

/**
 *      ltt_poll - file op for ltt files
 *      @filp: the file
 *      @wait: poll table
 *
 *      Poll implementation.
 */
static unsigned int ltt_poll(struct file *filp, poll_table *wait)
{
        unsigned int mask = 0;
        struct inode *inode = filp->f_dentry->d_inode;
        struct rchan_buf *buf = inode->i_private;
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);

        if (filp->f_mode & FMODE_READ) {
                poll_wait_set_exclusive(wait);
                poll_wait(filp, &buf->read_wait, wait);

                WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
                if (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
                                                        buf->chan)
                  - SUBBUF_TRUNC(atomic_long_read(&ltt_buf->consumed),
                                                        buf->chan)
                  == 0) {
                        if (buf->finalized)
                                return POLLHUP;
                        else
                                return 0;
                } else {
                        struct rchan *rchan =
                                ltt_channel->trans_channel_data;
                        if (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
                                        buf->chan)
                          - SUBBUF_TRUNC(atomic_long_read(
                                                &ltt_buf->consumed),
                                        buf->chan)
                          >= rchan->alloc_size)
                                return POLLPRI | POLLRDBAND;
                        else
                                return POLLIN | POLLRDNORM;
                }
        }
        return mask;
}

static int ltt_do_get_subbuf(struct rchan_buf *buf, struct ltt_channel_buf_struct *ltt_buf, long *pconsumed_old)
{
        long consumed_old, consumed_idx, commit_count, write_offset;
        consumed_old = atomic_long_read(&ltt_buf->consumed);
        consumed_idx = SUBBUF_INDEX(consumed_old, buf->chan);
        commit_count = local_read(&ltt_buf->commit_count[consumed_idx]);
        /*
         * Make sure we read the commit count before reading the buffer
         * data and the write offset. Correct consumed offset ordering
         * wrt commit count is insured by the use of cmpxchg to update
         * the consumed offset.
         */
        smp_rmb();
        write_offset = local_read(&ltt_buf->offset);
        /*
         * Check that the subbuffer we are trying to consume has been
         * already fully committed.
         */
        if (((commit_count - buf->chan->subbuf_size)
             & ltt_channel->commit_count_mask)
            - (BUFFER_TRUNC(consumed_old, buf->chan)
               >> ltt_channel->n_subbufs_order)
            != 0) {
                return -EAGAIN;
        }
        /*
         * Check that we are not about to read the same subbuffer in
         * which the writer head is.
         */
        if ((SUBBUF_TRUNC(write_offset, buf->chan)
           - SUBBUF_TRUNC(consumed_old, buf->chan))
           == 0) {
                return -EAGAIN;
        }

        *pconsumed_old = consumed_old;
        return 0;
}

static int ltt_do_put_subbuf(struct rchan_buf *buf, struct ltt_channel_buf_struct *ltt_buf, u32 uconsumed_old)
{
        long consumed_new, consumed_old;

        consumed_old = atomic_long_read(&ltt_buf->consumed);
        consumed_old = consumed_old & (~0xFFFFFFFFL);
        consumed_old = consumed_old | uconsumed_old;
        consumed_new = SUBBUF_ALIGN(consumed_old, buf->chan);

        spin_lock(&ltt_buf->full_lock);
        if (atomic_long_cmpxchg(&ltt_buf->consumed, consumed_old,
                                consumed_new)
            != consumed_old) {
                /* We have been pushed by the writer : the last
                 * buffer read _is_ corrupted! It can also
                 * happen if this is a buffer we never got. */
                spin_unlock(&ltt_buf->full_lock);
                return -EIO;
        } else {
                /* tell the client that buffer is now unfull */
                int index;
                long data;
                index = SUBBUF_INDEX(consumed_old, buf->chan);
                data = BUFFER_OFFSET(consumed_old, buf->chan);
                ltt_buf_unfull(buf, index, data);
                spin_unlock(&ltt_buf->full_lock);
        }
        return 0;
}

/**
 *      ltt_ioctl - control on the debugfs file
 *
 *      @inode: the inode
 *      @filp: the file
 *      @cmd: the command
 *      @arg: command arg
 *
 *      This ioctl implements three commands necessary for a minimal
 *      producer/consumer implementation :
 *      RELAY_GET_SUBBUF
 *              Get the next sub buffer that can be read. It never blocks.
 *      RELAY_PUT_SUBBUF
 *              Release the currently read sub-buffer. Parameter is the last
 *              put subbuffer (returned by GET_SUBBUF).
 *      RELAY_GET_N_BUBBUFS
 *              returns the number of sub buffers in the per cpu channel.
 *      RELAY_GET_SUBBUF_SIZE
 *              returns the size of the sub buffers.
 */
static int ltt_ioctl(struct inode *inode, struct file *filp,
                unsigned int cmd, unsigned long arg)
{
        struct rchan_buf *buf = inode->i_private;
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);
        u32 __user *argp = (u32 __user *)arg;

        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
        switch (cmd) {
        case RELAY_GET_SUBBUF:
        {
                int ret;
                ret = ltt_do_get_subbuf(buf, ltt_buf, &consumed_old);
                if(ret < 0)
                        return ret;
                return put_user((u32)consumed_old, argp);
        }
        case RELAY_PUT_SUBBUF:
        {
                int ret;
                u32 uconsumed_old;
                ret = get_user(uconsumed_old, argp);
                if (ret)
                        return ret; /* will return -EFAULT */
                return ltt_do_put_subbuf(buf, ltt_buf, uconsumed_old);
        }
        case RELAY_GET_N_SUBBUFS:
                return put_user((u32)buf->chan->n_subbufs, argp);
                break;
        case RELAY_GET_SUBBUF_SIZE:
                return put_user((u32)buf->chan->subbuf_size, argp);
                break;
        default:
                return -ENOIOCTLCMD;
        }
        return 0;
}

#ifdef CONFIG_COMPAT
static long ltt_compat_ioctl(struct file *file, unsigned int cmd,
                unsigned long arg)
{
        long ret = -ENOIOCTLCMD;

        lock_kernel();
        ret = ltt_ioctl(file->f_dentry->d_inode, file, cmd, arg);
        unlock_kernel();

        return ret;
}
#endif

static void ltt_relay_pipe_buf_release(struct pipe_inode_info *pipe,
                                   struct pipe_buffer *pbuf)
{
}

static struct pipe_buf_operations ltt_relay_pipe_buf_ops = {
        .can_merge = 0,
        .map = generic_pipe_buf_map,
        .unmap = generic_pipe_buf_unmap,
        .confirm = generic_pipe_buf_confirm,
        .release = ltt_relay_pipe_buf_release,
        .steal = generic_pipe_buf_steal,
        .get = generic_pipe_buf_get,
};

static void ltt_relay_page_release(struct splice_pipe_desc *spd, unsigned int i)
{
}

/*
 *      subbuf_splice_actor - splice up to one subbuf's worth of data
 */
static int subbuf_splice_actor(struct file *in,
                               loff_t *ppos,
                               struct pipe_inode_info *pipe,
                               size_t len,
                               unsigned int flags)
{
        struct rchan_buf *buf = in->private_data;
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_channel->buf, buf->cpu);
        unsigned int poff, subbuf_pages, nr_pages;
        struct page *pages[PIPE_BUFFERS];
        struct partial_page partial[PIPE_BUFFERS];
        struct splice_pipe_desc spd = {
                .pages = pages,
                .nr_pages = 0,
                .partial = partial,
                .flags = flags,
                .ops = &ltt_relay_pipe_buf_ops,
                .spd_release = ltt_relay_page_release,
        };
        long consumed_old, consumed_idx, roffset;
        unsigned long bytes_avail;

        /*
         * Check that a GET_SUBBUF ioctl has been done before.
         */
        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
        consumed_old = atomic_long_read(&ltt_buf->consumed);
        consumed_old += *ppos;
        consumed_idx = SUBBUF_INDEX(consumed_old, buf->chan);

        /*
         * Adjust read len, if longer than what is available
         */
        bytes_avail = SUBBUF_TRUNC(local_read(&ltt_buf->offset), buf->chan)
                    - consumed_old;
        WARN_ON(bytes_avail > buf->chan->alloc_size);
        len = min_t(size_t, len, bytes_avail);
        subbuf_pages = bytes_avail >> PAGE_SHIFT;
        nr_pages = min_t(unsigned int, subbuf_pages, PIPE_BUFFERS);
        roffset = consumed_old & PAGE_MASK;
        poff = consumed_old & ~PAGE_MASK;
        printk_dbg(KERN_DEBUG "SPLICE actor len %zu pos %zd write_pos %ld\n",
                len, (ssize_t)*ppos, local_read(&ltt_buf->offset));

        for (; spd.nr_pages < nr_pages; spd.nr_pages++) {
                unsigned int this_len;
                struct buf_page *page;

                if (!len)
                        break;
                printk_dbg(KERN_DEBUG "SPLICE actor loop len %zu roffset %ld\n",
                        len, roffset);

                this_len = PAGE_SIZE - poff;
                page = ltt_relay_read_get_page(buf, roffset);
                spd.pages[spd.nr_pages] = page->page;
                spd.partial[spd.nr_pages].offset = poff;
                spd.partial[spd.nr_pages].len = this_len;

                poff = 0;
                roffset += PAGE_SIZE;
                len -= this_len;
        }

        if (!spd.nr_pages)
                return 0;

        return splice_to_pipe(pipe, &spd);
}

static ssize_t ltt_relay_file_splice_read(struct file *in,
                                      loff_t *ppos,
                                      struct pipe_inode_info *pipe,
                                      size_t len,
                                      unsigned int flags)
{
        ssize_t spliced;
        int ret;

        ret = 0;
        spliced = 0;

        printk_dbg(KERN_DEBUG "SPLICE read len %zu pos %zd\n",
                len, (ssize_t)*ppos);
        while (len && !spliced) {
                ret = subbuf_splice_actor(in, ppos, pipe, len, flags);
                printk_dbg(KERN_DEBUG "SPLICE read loop ret %d\n", ret);
                if (ret < 0)
                        break;
                else if (!ret) {
                        if (flags & SPLICE_F_NONBLOCK)
                                ret = -EAGAIN;
                        break;
                }

                *ppos += ret;
                if (ret > len)
                        len = 0;
                else
                        len -= ret;
                spliced += ret;
        }

        if (spliced)
                return spliced;

        return ret;
}

static void ltt_relay_print_subbuffer_errors(
                struct ltt_channel_struct *ltt_chan,
                long cons_off, unsigned int cpu)
{
        struct rchan *rchan = ltt_chan->trans_channel_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_chan->buf, cpu);
        long cons_idx, commit_count, write_offset;

        cons_idx = SUBBUF_INDEX(cons_off, rchan);
        commit_count = local_read(&ltt_buf->commit_count[cons_idx]);
        /*
         * No need to order commit_count and write_offset reads because we
         * execute after trace is stopped when there are no readers left.
         */
        write_offset = local_read(&ltt_buf->offset);
        printk(KERN_WARNING
                "LTT : unread channel %s offset is %ld "
                "and cons_off : %ld (cpu %u)\n",
                ltt_chan->channel_name, write_offset, cons_off, cpu);
        /* Check each sub-buffer for non filled commit count */
        if (((commit_count - rchan->subbuf_size) & ltt_chan->commit_count_mask)
            - (BUFFER_TRUNC(cons_off, rchan) >> ltt_chan->n_subbufs_order)
            != 0)
                printk(KERN_ALERT
                        "LTT : %s : subbuffer %lu has non filled "
                        "commit count %lu.\n",
                        ltt_chan->channel_name, cons_idx, commit_count);
        printk(KERN_ALERT "LTT : %s : commit count : %lu, subbuf size %zd\n",
                        ltt_chan->channel_name, commit_count,
                        rchan->subbuf_size);
}

static void ltt_relay_print_errors(struct ltt_trace_struct *trace,
                struct ltt_channel_struct *ltt_chan, int cpu)
{
        struct rchan *rchan = ltt_chan->trans_channel_data;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_chan->buf, cpu);
        long cons_off;

        for (cons_off = atomic_long_read(&ltt_buf->consumed);
                        (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
                                      rchan)
                         - cons_off) > 0;
                        cons_off = SUBBUF_ALIGN(cons_off, rchan))
                ltt_relay_print_subbuffer_errors(ltt_chan, cons_off, cpu);
}

static void ltt_relay_print_buffer_errors(struct ltt_channel_struct *ltt_chan,
                unsigned int cpu)
{
        struct ltt_trace_struct *trace = ltt_chan->trace;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_chan->buf, cpu);

        if (local_read(&ltt_buf->events_lost))
                printk(KERN_ALERT
                        "LTT : %s : %ld events lost "
                        "in %s channel (cpu %u).\n",
                        ltt_chan->channel_name,
                        local_read(&ltt_buf->events_lost),
                        ltt_chan->channel_name, cpu);
        if (local_read(&ltt_buf->corrupted_subbuffers))
                printk(KERN_ALERT
                        "LTT : %s : %ld corrupted subbuffers "
                        "in %s channel (cpu %u).\n",
                        ltt_chan->channel_name,
                        local_read(&ltt_buf->corrupted_subbuffers),
                        ltt_chan->channel_name, cpu);

        ltt_relay_print_errors(trace, ltt_chan, cpu);
}

static void ltt_relay_remove_dirs(struct ltt_trace_struct *trace)
{
        debugfs_remove(trace->dentry.trace_root);
}

static void ltt_relay_release_channel(struct kref *kref)
{
        struct ltt_channel_struct *ltt_chan = container_of(kref,
                        struct ltt_channel_struct, kref);
        percpu_free(ltt_chan->buf);
}

/*
 * Create ltt buffer.
 */
static int ltt_relay_create_buffer(struct ltt_trace_struct *trace,
                struct ltt_channel_struct *ltt_chan, struct rchan_buf *buf,
                unsigned int cpu, unsigned int n_subbufs)
{
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_chan->buf, cpu);
        unsigned int j;

        ltt_buf->commit_count =
                kzalloc_node(sizeof(ltt_buf->commit_count) * n_subbufs,
                        GFP_KERNEL, cpu_to_node(cpu));
        if (!ltt_buf->commit_count)
                return -ENOMEM;
        kref_get(&trace->kref);
        kref_get(&trace->ltt_transport_kref);
        kref_get(&ltt_chan->kref);
        local_set(&ltt_buf->offset, ltt_subbuffer_header_size());
        atomic_long_set(&ltt_buf->consumed, 0);
        atomic_long_set(&ltt_buf->active_readers, 0);
        for (j = 0; j < n_subbufs; j++)
                local_set(&ltt_buf->commit_count[j], 0);
        init_waitqueue_head(&ltt_buf->write_wait);
        atomic_set(&ltt_buf->wakeup_readers, 0);
        spin_lock_init(&ltt_buf->full_lock);

        ltt_buffer_begin_callback(buf, trace->start_tsc, 0);
        /* atomic_add made on local variable on data that belongs to
         * various CPUs : ok because tracing not started (for this cpu). */
        local_add(ltt_subbuffer_header_size(), &ltt_buf->commit_count[0]);

        local_set(&ltt_buf->events_lost, 0);
        local_set(&ltt_buf->corrupted_subbuffers, 0);

        return 0;
}

static void ltt_relay_destroy_buffer(struct ltt_channel_struct *ltt_chan,
                unsigned int cpu)
{
        struct ltt_trace_struct *trace = ltt_chan->trace;
        struct ltt_channel_buf_struct *ltt_buf =
                percpu_ptr(ltt_chan->buf, cpu);

        kref_put(&ltt_chan->trace->ltt_transport_kref,
                ltt_release_transport);
        ltt_relay_print_buffer_errors(ltt_chan, cpu);
        kfree(ltt_buf->commit_count);
        ltt_buf->commit_count = NULL;
        kref_put(&ltt_chan->kref, ltt_relay_release_channel);
        kref_put(&trace->kref, ltt_release_trace);
        wake_up_interruptible(&trace->kref_wq);
}

/*
 * Create channel.
 */
static int ltt_relay_create_channel(const char *trace_name,
                struct ltt_trace_struct *trace, struct dentry *dir,
                const char *channel_name, struct ltt_channel_struct *ltt_chan,
                unsigned int subbuf_size, unsigned int n_subbufs,
                int overwrite)
{
        char *tmpname;
        unsigned int tmpname_len;
        int err = 0;

        tmpname = kmalloc(PATH_MAX, GFP_KERNEL);
        if (!tmpname)
                return EPERM;
        if (overwrite) {
                strncpy(tmpname, LTT_FLIGHT_PREFIX, PATH_MAX-1);
                strncat(tmpname, channel_name,
                        PATH_MAX-1-sizeof(LTT_FLIGHT_PREFIX));
        } else {
                strncpy(tmpname, channel_name, PATH_MAX-1);
        }
        strncat(tmpname, "_", PATH_MAX-1-strlen(tmpname));

        kref_init(&ltt_chan->kref);

        ltt_chan->trace = trace;
        ltt_chan->buffer_begin = ltt_buffer_begin_callback;
        ltt_chan->buffer_end = ltt_buffer_end_callback;
        ltt_chan->overwrite = overwrite;
        ltt_chan->n_subbufs_order = get_count_order(n_subbufs);
        ltt_chan->commit_count_mask = (~0UL >> ltt_chan->n_subbufs_order);
        ltt_chan->buf = percpu_alloc_mask(sizeof(struct ltt_channel_buf_struct),
                                          GFP_KERNEL, cpu_possible_map);
        if (!ltt_chan->buf)
                goto ltt_percpu_alloc_error;
        ltt_chan->trans_channel_data = ltt_relay_open(tmpname,
                        dir,
                        subbuf_size,
                        n_subbufs,
                        &trace->callbacks,
                        ltt_chan);
        tmpname_len = strlen(tmpname);
        if (tmpname_len > 0) {
                /* Remove final _ for pretty printing */
                tmpname[tmpname_len-1] = '\0';
        }
        if (ltt_chan->trans_channel_data == NULL) {
                printk(KERN_ERR "LTT : Can't open %s channel for trace %s\n",
                                tmpname, trace_name);
                goto relay_open_error;
        }

        err = 0;
        goto end;

relay_open_error:
        percpu_free(ltt_chan->buf);
ltt_percpu_alloc_error:
        err = EPERM;
end:
        kfree(tmpname);
        return err;
}

static int ltt_relay_create_dirs(struct ltt_trace_struct *new_trace)
{
        new_trace->dentry.trace_root = debugfs_create_dir(new_trace->trace_name,
                        get_ltt_root());
        if (new_trace->dentry.trace_root == NULL) {
                printk(KERN_ERR "LTT : Trace directory name %s already taken\n",
                                new_trace->trace_name);
                return EEXIST;
        }

        new_trace->callbacks.create_buf_file = ltt_create_buf_file_callback;
        new_trace->callbacks.remove_buf_file = ltt_remove_buf_file_callback;

        return 0;
}

/*
 * LTTng channel flush function.
 *
 * Must be called when no tracing is active in the channel, because of
 * accesses across CPUs.
 */
static notrace void ltt_relay_buffer_flush(struct rchan_buf *buf)
{
        buf->finalized = 1;
        ltt_force_switch(buf, FORCE_FLUSH);
}

static void ltt_relay_async_wakeup_chan(struct ltt_channel_struct *ltt_channel)
{
        unsigned int i;
        struct rchan *rchan = ltt_channel->trans_channel_data;

        for_each_possible_cpu(i) {
                struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, i);

                if (atomic_read(&ltt_buf->wakeup_readers) == 1) {
                        atomic_set(&ltt_buf->wakeup_readers, 0);
                        wake_up_interruptible(&rchan->buf[i]->read_wait);
                }
        }
}

static void ltt_relay_finish_buffer(struct ltt_channel_struct *ltt_channel,
                unsigned int cpu)
{
        struct rchan *rchan = ltt_channel->trans_channel_data;

        if (rchan->buf[cpu]) {
                struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, cpu);
                ltt_relay_buffer_flush(rchan->buf[cpu]);
                ltt_relay_wake_writers(ltt_buf);
        }
}


static void ltt_relay_finish_channel(struct ltt_channel_struct *ltt_channel)
{
        unsigned int i;

        for_each_possible_cpu(i)
                ltt_relay_finish_buffer(ltt_channel, i);
}

static void ltt_relay_remove_channel(struct ltt_channel_struct *channel)
{
        struct rchan *rchan = channel->trans_channel_data;

        ltt_relay_close(rchan);
        kref_put(&channel->kref, ltt_relay_release_channel);
}

struct ltt_reserve_switch_offsets {
        long begin, end, old;
        long begin_switch, end_switch_current, end_switch_old;
        long commit_count, reserve_commit_diff;
        size_t before_hdr_pad, size;
};

/*
 * Returns :
 * 0 if ok
 * !0 if execution must be aborted.
 */
static inline int ltt_relay_try_reserve(
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf, struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets, size_t data_size,
                u64 *tsc, unsigned int *rflags, int largest_align)
{
        offsets->begin = local_read(&ltt_buf->offset);
        offsets->old = offsets->begin;
        offsets->begin_switch = 0;
        offsets->end_switch_current = 0;
        offsets->end_switch_old = 0;

        *tsc = trace_clock_read64();
        if (last_tsc_overflow(ltt_buf, *tsc))
                *rflags = LTT_RFLAG_ID_SIZE_TSC;

        if (SUBBUF_OFFSET(offsets->begin, buf->chan) == 0) {
                offsets->begin_switch = 1;              /* For offsets->begin */
        } else {
                offsets->size = ltt_get_header_size(ltt_channel,
                                        offsets->begin, data_size,
                                        &offsets->before_hdr_pad, *rflags);
                offsets->size += ltt_align(offsets->begin + offsets->size,
                                           largest_align)
                                 + data_size;
                if ((SUBBUF_OFFSET(offsets->begin, buf->chan) + offsets->size)
                                > buf->chan->subbuf_size) {
                        offsets->end_switch_old = 1;    /* For offsets->old */
                        offsets->begin_switch = 1;      /* For offsets->begin */
                }
        }
        if (offsets->begin_switch) {
                long subbuf_index;

                if (offsets->end_switch_old)
                        offsets->begin = SUBBUF_ALIGN(offsets->begin,
                                                      buf->chan);
                offsets->begin = offsets->begin + ltt_subbuffer_header_size();
                /* Test new buffer integrity */
                subbuf_index = SUBBUF_INDEX(offsets->begin, buf->chan);
                offsets->reserve_commit_diff =
                        (BUFFER_TRUNC(offsets->begin, buf->chan)
                         >> ltt_channel->n_subbufs_order)
                        - (local_read(&ltt_buf->commit_count[subbuf_index])
                                & ltt_channel->commit_count_mask);
                if (offsets->reserve_commit_diff == 0) {
                        /* Next buffer not corrupted. */
                        if (!ltt_channel->overwrite &&
                                (SUBBUF_TRUNC(offsets->begin, buf->chan)
                                 - SUBBUF_TRUNC(atomic_long_read(
                                                        &ltt_buf->consumed),
                                                buf->chan))
                                >= rchan->alloc_size) {
                                /*
                                 * We do not overwrite non consumed buffers
                                 * and we are full : event is lost.
                                 */
                                local_inc(&ltt_buf->events_lost);
                                return -1;
                        } else {
                                /*
                                 * next buffer not corrupted, we are either in
                                 * overwrite mode or the buffer is not full.
                                 * It's safe to write in this new subbuffer.
                                 */
                        }
                } else {
                        /*
                         * Next subbuffer corrupted. Force pushing reader even
                         * in normal mode. It's safe to write in this new
                         * subbuffer.
                         */
                }
                offsets->size = ltt_get_header_size(ltt_channel,
                                        offsets->begin, data_size,
                                        &offsets->before_hdr_pad, *rflags);
                offsets->size += ltt_align(offsets->begin + offsets->size,
                                           largest_align)
                                 + data_size;
                if ((SUBBUF_OFFSET(offsets->begin, buf->chan) + offsets->size)
                                > buf->chan->subbuf_size) {
                        /*
                         * Event too big for subbuffers, report error, don't
                         * complete the sub-buffer switch.
                         */
                        local_inc(&ltt_buf->events_lost);
                        return -1;
                } else {
                        /*
                         * We just made a successful buffer switch and the event
                         * fits in the new subbuffer. Let's write.
                         */
                }
        } else {
                /*
                 * Event fits in the current buffer and we are not on a switch
                 * boundary. It's safe to write.
                 */
        }
        offsets->end = offsets->begin + offsets->size;

        if ((SUBBUF_OFFSET(offsets->end, buf->chan)) == 0) {
                /*
                 * The offset_end will fall at the very beginning of the next
                 * subbuffer.
                 */
                offsets->end_switch_current = 1;        /* For offsets->begin */
        }
        return 0;
}

/*
 * Returns :
 * 0 if ok
 * !0 if execution must be aborted.
 */
static inline int ltt_relay_try_switch(
                enum force_switch_mode mode,
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf, struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets,
                u64 *tsc)
{
        long subbuf_index;

        offsets->begin = local_read(&ltt_buf->offset);
        offsets->old = offsets->begin;
        offsets->begin_switch = 0;
        offsets->end_switch_old = 0;

        *tsc = trace_clock_read64();

        if (SUBBUF_OFFSET(offsets->begin, buf->chan) != 0) {
                offsets->begin = SUBBUF_ALIGN(offsets->begin, buf->chan);
                offsets->end_switch_old = 1;
        } else {
                /* we do not have to switch : buffer is empty */
                return -1;
        }
        if (mode == FORCE_ACTIVE)
                offsets->begin += ltt_subbuffer_header_size();
        /*
         * Always begin_switch in FORCE_ACTIVE mode.
         * Test new buffer integrity
         */
        subbuf_index = SUBBUF_INDEX(offsets->begin, buf->chan);
        offsets->reserve_commit_diff =
                (BUFFER_TRUNC(offsets->begin, buf->chan)
                 >> ltt_channel->n_subbufs_order)
                - (local_read(&ltt_buf->commit_count[subbuf_index])
                        & ltt_channel->commit_count_mask);
        if (offsets->reserve_commit_diff == 0) {
                /* Next buffer not corrupted. */
                if (mode == FORCE_ACTIVE
                    && !ltt_channel->overwrite
                    && offsets->begin - atomic_long_read(&ltt_buf->consumed)
                       >= rchan->alloc_size) {
                        /*
                         * We do not overwrite non consumed buffers and we are
                         * full : ignore switch while tracing is active.
                         */
                        return -1;
                }
        } else {
                /*
                 * Next subbuffer corrupted. Force pushing reader even in normal
                 * mode
                 */
        }
        offsets->end = offsets->begin;
        return 0;
}

static inline void ltt_reserve_push_reader(
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf,
                struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets)
{
        long consumed_old, consumed_new;

        do {
                consumed_old = atomic_long_read(&ltt_buf->consumed);
                /*
                 * If buffer is in overwrite mode, push the reader consumed
                 * count if the write position has reached it and we are not
                 * at the first iteration (don't push the reader farther than
                 * the writer). This operation can be done concurrently by many
                 * writers in the same buffer, the writer being at the farthest
                 * write position sub-buffer index in the buffer being the one
                 * which will win this loop.
                 * If the buffer is not in overwrite mode, pushing the reader
                 * only happens if a sub-buffer is corrupted.
                 */
                if ((SUBBUF_TRUNC(offsets->end-1, buf->chan)
                   - SUBBUF_TRUNC(consumed_old, buf->chan))
                   >= rchan->alloc_size)
                        consumed_new = SUBBUF_ALIGN(consumed_old, buf->chan);
                else {
                        consumed_new = consumed_old;
                        break;
                }
        } while (atomic_long_cmpxchg(&ltt_buf->consumed, consumed_old,
                        consumed_new) != consumed_old);

        if (consumed_old != consumed_new) {
                /*
                 * Reader pushed : we are the winner of the push, we can
                 * therefore reequilibrate reserve and commit. Atomic increment
                 * of the commit count permits other writers to play around
                 * with this variable before us. We keep track of
                 * corrupted_subbuffers even in overwrite mode :
                 * we never want to write over a non completely committed
                 * sub-buffer : possible causes : the buffer size is too low
                 * compared to the unordered data input, or there is a writer
                 * that died between the reserve and the commit.
                 */
                if (offsets->reserve_commit_diff) {
                        /*
                         * We have to alter the sub-buffer commit count.
                         * We do not deliver the previous subbuffer, given it
                         * was either corrupted or not consumed (overwrite
                         * mode).
                         */
                        local_add(offsets->reserve_commit_diff,
                                  &ltt_buf->commit_count[
                                        SUBBUF_INDEX(offsets->begin,
                                                     buf->chan)]);
                        if (!ltt_channel->overwrite
                            || offsets->reserve_commit_diff
                               != rchan->subbuf_size) {
                                /*
                                 * The reserve commit diff was not subbuf_size :
                                 * it means the subbuffer was partly written to
                                 * and is therefore corrupted. If it is multiple
                                 * of subbuffer size and we are in flight
                                 * recorder mode, we are skipping over a whole
                                 * subbuffer.
                                 */
                                local_inc(&ltt_buf->corrupted_subbuffers);
                        }
                }
        }
}


/*
 * ltt_reserve_switch_old_subbuf: switch old subbuffer
 *
 * Concurrency safe because we are the last and only thread to alter this
 * sub-buffer. As long as it is not delivered and read, no other thread can
 * alter the offset, alter the reserve_count or call the
 * client_buffer_end_callback on this sub-buffer.
 *
 * The only remaining threads could be the ones with pending commits. They will
 * have to do the deliver themselves.  Not concurrency safe in overwrite mode.
 * We detect corrupted subbuffers with commit and reserve counts. We keep a
 * corrupted sub-buffers count and push the readers across these sub-buffers.
 *
 * Not concurrency safe if a writer is stalled in a subbuffer and another writer
 * switches in, finding out it's corrupted.  The result will be than the old
 * (uncommited) subbuffer will be declared corrupted, and that the new subbuffer
 * will be declared corrupted too because of the commit count adjustment.
 *
 * Note : offset_old should never be 0 here.
 */
static inline void ltt_reserve_switch_old_subbuf(
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf, struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
{
        long oldidx = SUBBUF_INDEX(offsets->old - 1, rchan);

        ltt_channel->buffer_end(buf, *tsc, offsets->old, oldidx);
        /* Must write buffer end before incrementing commit count */
        smp_wmb();
        offsets->commit_count =
                local_add_return(rchan->subbuf_size
                                 - (SUBBUF_OFFSET(offsets->old - 1, rchan)
                                 + 1),
                                 &ltt_buf->commit_count[oldidx]);
        if ((BUFFER_TRUNC(offsets->old - 1, rchan)
                        >> ltt_channel->n_subbufs_order)
                        - ((offsets->commit_count - rchan->subbuf_size)
                                & ltt_channel->commit_count_mask) == 0)
                ltt_deliver(buf, oldidx, NULL);
}

/*
 * ltt_reserve_switch_new_subbuf: Populate new subbuffer.
 *
 * This code can be executed unordered : writers may already have written to the
 * sub-buffer before this code gets executed, caution.  The commit makes sure
 * that this code is executed before the deliver of this sub-buffer.
 */
static inline void ltt_reserve_switch_new_subbuf(
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf, struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
{
        long beginidx = SUBBUF_INDEX(offsets->begin, rchan);

        ltt_channel->buffer_begin(buf, *tsc, beginidx);
        /* Must write buffer end before incrementing commit count */
        smp_wmb();
        offsets->commit_count = local_add_return(ltt_subbuffer_header_size(),
                        &ltt_buf->commit_count[beginidx]);
        /* Check if the written buffer has to be delivered */
        if ((BUFFER_TRUNC(offsets->begin, rchan)
                        >> ltt_channel->n_subbufs_order)
                        - ((offsets->commit_count - rchan->subbuf_size)
                                & ltt_channel->commit_count_mask) == 0)
                ltt_deliver(buf, beginidx, NULL);
}


/*
 * ltt_reserve_end_switch_current: finish switching current subbuffer
 *
 * Concurrency safe because we are the last and only thread to alter this
 * sub-buffer. As long as it is not delivered and read, no other thread can
 * alter the offset, alter the reserve_count or call the
 * client_buffer_end_callback on this sub-buffer.
 *
 * The only remaining threads could be the ones with pending commits. They will
 * have to do the deliver themselves.  Not concurrency safe in overwrite mode.
 * We detect corrupted subbuffers with commit and reserve counts. We keep a
 * corrupted sub-buffers count and push the readers across these sub-buffers.
 *
 * Not concurrency safe if a writer is stalled in a subbuffer and another writer
 * switches in, finding out it's corrupted.  The result will be than the old
 * (uncommited) subbuffer will be declared corrupted, and that the new subbuffer
 * will be declared corrupted too because of the commit count adjustment.
 */
static inline void ltt_reserve_end_switch_current(
                struct ltt_channel_struct *ltt_channel,
                struct ltt_channel_buf_struct *ltt_buf, struct rchan *rchan,
                struct rchan_buf *buf,
                struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
{
        long endidx = SUBBUF_INDEX(offsets->end - 1, rchan);

        ltt_channel->buffer_end(buf, *tsc, offsets->end, endidx);
        /* Must write buffer begin before incrementing commit count */
        smp_wmb();
        offsets->commit_count =
                local_add_return(rchan->subbuf_size
                                 - (SUBBUF_OFFSET(offsets->end - 1, rchan)
                                 + 1),
                                 &ltt_buf->commit_count[endidx]);
        if ((BUFFER_TRUNC(offsets->end - 1, rchan)
                        >> ltt_channel->n_subbufs_order)
                        - ((offsets->commit_count - rchan->subbuf_size)
                                & ltt_channel->commit_count_mask) == 0)
                ltt_deliver(buf, endidx, NULL);
}

/**
 * ltt_relay_reserve_slot - Atomic slot reservation in a LTTng buffer.
 * @trace: the trace structure to log to.
 * @ltt_channel: channel structure
 * @transport_data: data structure specific to ltt relay
 * @data_size: size of the variable length data to log.
 * @slot_size: pointer to total size of the slot (out)
 * @buf_offset : pointer to reserved buffer offset (out)
 * @tsc: pointer to the tsc at the slot reservation (out)
 * @cpu: cpuid
 *
 * Return : -ENOSPC if not enough space, else returns 0.
 * It will take care of sub-buffer switching.
 */
static notrace int ltt_relay_reserve_slot(struct ltt_trace_struct *trace,
                struct ltt_channel_struct *ltt_channel, void **transport_data,
                size_t data_size, size_t *slot_size, long *buf_offset, u64 *tsc,
                unsigned int *rflags, int largest_align, int cpu)
{
        struct rchan *rchan = ltt_channel->trans_channel_data;
        struct rchan_buf *buf = *transport_data =
                        rchan->buf[cpu];
        struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, buf->cpu);
        struct ltt_reserve_switch_offsets offsets;

        offsets.reserve_commit_diff = 0;
        offsets.size = 0;

        /*
         * Perform retryable operations.
         */
        if (__get_cpu_var(ltt_nesting) > 4) {
                local_inc(&ltt_buf->events_lost);
                return -EPERM;
        }
        do {
                if (ltt_relay_try_reserve(ltt_channel, ltt_buf,
                                rchan, buf, &offsets, data_size, tsc, rflags,
                                largest_align))
                        return -ENOSPC;
        } while (local_cmpxchg(&ltt_buf->offset, offsets.old,
                        offsets.end) != offsets.old);

        /*
         * Atomically update last_tsc. This update races against concurrent
         * atomic updates, but the race will always cause supplementary full TSC
         * events, never the opposite (missing a full TSC event when it would be
         * needed).
         */
        save_last_tsc(ltt_buf, *tsc);

        /*
         * Push the reader if necessary
         */
        ltt_reserve_push_reader(ltt_channel, ltt_buf, rchan, buf, &offsets);

        /*
         * Switch old subbuffer if needed.
         */
        if (offsets.end_switch_old)
                ltt_reserve_switch_old_subbuf(ltt_channel, ltt_buf, rchan, buf,
                        &offsets, tsc);

        /*
         * Populate new subbuffer.
         */
        if (offsets.begin_switch)
                ltt_reserve_switch_new_subbuf(ltt_channel, ltt_buf, rchan,
                        buf, &offsets, tsc);

        if (offsets.end_switch_current)
                ltt_reserve_end_switch_current(ltt_channel, ltt_buf, rchan,
                        buf, &offsets, tsc);

        *slot_size = offsets.size;
        *buf_offset = offsets.begin + offsets.before_hdr_pad;
        return 0;
}

/*
 * Force a sub-buffer switch for a per-cpu buffer. This operation is
 * completely reentrant : can be called while tracing is active with
 * absolutely no lock held.
 *
 * Note, however, that as a local_cmpxchg is used for some atomic
 * operations, this function must be called from the CPU which owns the buffer
 * for a ACTIVE flush.
 */
static notrace void ltt_force_switch(struct rchan_buf *buf,
                enum force_switch_mode mode)
{
        struct ltt_channel_struct *ltt_channel =
                        (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, buf->cpu);
        struct rchan *rchan = ltt_channel->trans_channel_data;
        struct ltt_reserve_switch_offsets offsets;
        u64 tsc;

        offsets.reserve_commit_diff = 0;
        offsets.size = 0;

        /*
         * Perform retryable operations.
         */
        do {
                if (ltt_relay_try_switch(mode, ltt_channel, ltt_buf,
                                rchan, buf, &offsets, &tsc))
                        return;
        } while (local_cmpxchg(&ltt_buf->offset, offsets.old,
                        offsets.end) != offsets.old);

        /*
         * Atomically update last_tsc. This update races against concurrent
         * atomic updates, but the race will always cause supplementary full TSC
         * events, never the opposite (missing a full TSC event when it would be
         * needed).
         */
        save_last_tsc(ltt_buf, tsc);

        /*
         * Push the reader if necessary
         */
        if (mode == FORCE_ACTIVE)
                ltt_reserve_push_reader(ltt_channel, ltt_buf, rchan,
                                        buf, &offsets);

        /*
         * Switch old subbuffer if needed.
         */
        if (offsets.end_switch_old)
                ltt_reserve_switch_old_subbuf(ltt_channel, ltt_buf, rchan, buf,
                        &offsets, &tsc);

        /*
         * Populate new subbuffer.
         */
        if (mode == FORCE_ACTIVE)
                ltt_reserve_switch_new_subbuf(ltt_channel,
                        ltt_buf, rchan, buf, &offsets, &tsc);
}

/*
 * for flight recording. must be called after relay_commit.
 * This function decrements de subbuffer's lost_size each time the commit count
 * reaches back the reserve offset (module subbuffer size). It is useful for
 * crash dump.
 * We use slot_size - 1 to make sure we deal correctly with the case where we
 * fill the subbuffer completely (so the subbuf index stays in the previous
 * subbuffer).
 */
#ifdef CONFIG_LTT_VMCORE
static inline void ltt_write_commit_counter(struct rchan_buf *buf,
                long buf_offset, size_t slot_size)
{
        struct ltt_channel_struct *ltt_channel =
                (struct ltt_channel_struct *)buf->chan->private_data;
        struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, buf->cpu);
        struct ltt_subbuffer_header *header;
        long offset, subbuf_idx, commit_count;
        uint32_t lost_old, lost_new;

        subbuf_idx = SUBBUF_INDEX(buf_offset - 1, buf->chan);
        offset = buf_offset + slot_size;
        header = (struct ltt_subbuffer_header *)
                        ltt_relay_offset_address(buf,
                                subbuf_idx * buf->chan->subbuf_size);
        for (;;) {
                lost_old = header->lost_size;
                commit_count =
                        local_read(&ltt_buf->commit_count[subbuf_idx]);
                /* SUBBUF_OFFSET includes commit_count_mask */
                if (!SUBBUF_OFFSET(offset - commit_count, buf->chan)) {
                        lost_new = (uint32_t)buf->chan->subbuf_size
                                   - SUBBUF_OFFSET(commit_count, buf->chan);
                        lost_old = cmpxchg_local(&header->lost_size, lost_old,
                                                lost_new);
                        if (lost_old <= lost_new)
                                break;
                } else {
                        break;
                }
        }
}
#else
static inline void ltt_write_commit_counter(struct rchan_buf *buf,
                long buf_offset, size_t slot_size)
{
}
#endif

/*
 * Atomic unordered slot commit. Increments the commit count in the
 * specified sub-buffer, and delivers it if necessary.
 *
 * Parameters:
 *
 * @ltt_channel : channel structure
 * @transport_data: transport-specific data
 * @buf_offset : offset following the event header.
 * @slot_size : size of the reserved slot.
 */
static notrace void ltt_relay_commit_slot(
                struct ltt_channel_struct *ltt_channel,
                void **transport_data, long buf_offset, size_t slot_size)
{
        struct rchan_buf *buf = *transport_data;
        struct ltt_channel_buf_struct *ltt_buf =
                        percpu_ptr(ltt_channel->buf, buf->cpu);
        struct rchan *rchan = buf->chan;
        long offset_end = buf_offset;
        long endidx = SUBBUF_INDEX(offset_end - 1, rchan);
        long commit_count;

        /* Must write slot data before incrementing commit count */
        smp_wmb();
        commit_count = local_add_return(slot_size,
                &ltt_buf->commit_count[endidx]);
        /* Check if all commits have been done */
        if ((BUFFER_TRUNC(offset_end - 1, rchan)
                        >> ltt_channel->n_subbufs_order)
                        - ((commit_count - rchan->subbuf_size)
                           & ltt_channel->commit_count_mask) == 0)
                ltt_deliver(buf, endidx, NULL);
        /*
         * Update lost_size for each commit. It's needed only for extracting
         * ltt buffers from vmcore, after crash.
         */
        ltt_write_commit_counter(buf, buf_offset, slot_size);
}

/*
 * This is called with preemption disabled when user space has requested
 * blocking mode.  If one of the active traces has free space below a
 * specific threshold value, we reenable preemption and block.
 */
static int ltt_relay_user_blocking(struct ltt_trace_struct *trace,
                unsigned int chan_index, size_t data_size,
                struct user_dbg_data *dbg)
{
        struct rchan *rchan;
        struct ltt_channel_buf_struct *ltt_buf;
        struct ltt_channel_struct *channel;
        struct rchan_buf *relay_buf;
        int cpu;
        DECLARE_WAITQUEUE(wait, current);

        channel = &trace->channels[chan_index];
        rchan = channel->trans_channel_data;
        cpu = smp_processor_id();
        relay_buf = rchan->buf[cpu];
        ltt_buf = percpu_ptr(channel->buf, cpu);

        /*
         * Check if data is too big for the channel : do not
         * block for it.
         */
        if (LTT_RESERVE_CRITICAL + data_size > relay_buf->chan->subbuf_size)
                return 0;

        /*
         * If free space too low, we block. We restart from the
         * beginning after we resume (cpu id may have changed
         * while preemption is active).
         */
        spin_lock(&ltt_buf->full_lock);
        if (!channel->overwrite) {
                dbg->write = local_read(&ltt_buf->offset);
                dbg->read = atomic_long_read(&ltt_buf->consumed);
                dbg->avail_size = dbg->write + LTT_RESERVE_CRITICAL + data_size
                                  - SUBBUF_TRUNC(dbg->read,
                                                 relay_buf->chan);
                if (dbg->avail_size > rchan->alloc_size) {
                        __set_current_state(TASK_INTERRUPTIBLE);
                        add_wait_queue(&ltt_buf->write_wait, &wait);
                        spin_unlock(&ltt_buf->full_lock);
                        preempt_enable();
                        schedule();
                        __set_current_state(TASK_RUNNING);
                        remove_wait_queue(&ltt_buf->write_wait, &wait);
                        if (signal_pending(current))
                                return -ERESTARTSYS;
                        preempt_disable();
                        return 1;
                }
        }
        spin_unlock(&ltt_buf->full_lock);
        return 0;
}

static void ltt_relay_print_user_errors(struct ltt_trace_struct *trace,
                unsigned int chan_index, size_t data_size,
                struct user_dbg_data *dbg, int cpu)
{
        struct rchan *rchan;
        struct ltt_channel_buf_struct *ltt_buf;
        struct ltt_channel_struct *channel;
        struct rchan_buf *relay_buf;

        channel = &trace->channels[chan_index];
        rchan = channel->trans_channel_data;
        relay_buf = rchan->buf[cpu];
        ltt_buf = percpu_ptr(channel->buf, cpu);

        printk(KERN_ERR "Error in LTT usertrace : "
        "buffer full : event lost in blocking "
        "mode. Increase LTT_RESERVE_CRITICAL.\n");
        printk(KERN_ERR "LTT nesting level is %u.\n",
                per_cpu(ltt_nesting, cpu));
        printk(KERN_ERR "LTT avail size %lu.\n",
                dbg->avail_size);
        printk(KERN_ERR "avai write : %lu, read : %lu\n",
                        dbg->write, dbg->read);

        dbg->write = local_read(&ltt_buf->offset);
        dbg->read = atomic_long_read(&ltt_buf->consumed);

        printk(KERN_ERR "LTT cur size %lu.\n",
                dbg->write + LTT_RESERVE_CRITICAL + data_size
                - SUBBUF_TRUNC(dbg->read, relay_buf->chan));
        printk(KERN_ERR "cur write : %lu, read : %lu\n",
                        dbg->write, dbg->read);
}

static struct ltt_transport ltt_relay_transport = {
        .name = "relay",
        .owner = THIS_MODULE,
        .ops = {
                .create_dirs = ltt_relay_create_dirs,
                .remove_dirs = ltt_relay_remove_dirs,
                .create_channel = ltt_relay_create_channel,
                .finish_channel = ltt_relay_finish_channel,
                .remove_channel = ltt_relay_remove_channel,
                .wakeup_channel = ltt_relay_async_wakeup_chan,
                .commit_slot = ltt_relay_commit_slot,
                .reserve_slot = ltt_relay_reserve_slot,
                .user_blocking = ltt_relay_user_blocking,
                .user_errors = ltt_relay_print_user_errors,
        },
};

static int __init ltt_relay_init(void)
{
        printk(KERN_INFO "LTT : ltt-relay init\n");

        ltt_file_operations = ltt_relay_file_operations;
        ltt_file_operations.owner = THIS_MODULE;
        ltt_file_operations.open = ltt_open;
        ltt_file_operations.release = ltt_release;
        ltt_file_operations.poll = ltt_poll;
        ltt_file_operations.splice_read = ltt_relay_file_splice_read,
        ltt_file_operations.ioctl = ltt_ioctl;
#ifdef CONFIG_COMPAT
        ltt_file_operations.compat_ioctl = ltt_compat_ioctl;
#endif

        ltt_transport_register(&ltt_relay_transport);

        return 0;
}

static void __exit ltt_relay_exit(void)
{
        printk(KERN_INFO "LTT : ltt-relay exit\n");

        ltt_transport_unregister(&ltt_relay_transport);
}

module_init(ltt_relay_init);
module_exit(ltt_relay_exit);

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Mathieu Desnoyers");
MODULE_DESCRIPTION("Linux Trace Toolkit Next Generation Lockless Relay");