markers: fix: teardown should be tagged with "destructor" attribute

[ust.git] / libust / marker.c

/*
 * Copyright (C) 2007 Mathieu Desnoyers
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301 USA
 */

#include <stdlib.h>
#include <errno.h>
#define _LGPL_SOURCE
#include <urcu-bp.h>
#include <urcu/rculist.h>
#include <urcu/hlist.h>

#include <ust/core.h>
#include <ust/marker.h>
#include <ust/tracepoint.h>

#include "usterr.h"
#include "channels.h"
#include "tracercore.h"
#include "tracer.h"

__thread long ust_reg_stack[500];
volatile __thread long *ust_reg_stack_ptr = (long *) 0;

extern struct marker * const __start___markers_ptrs[] __attribute__((visibility("hidden")));
extern struct marker * const __stop___markers_ptrs[] __attribute__((visibility("hidden")));

/* Set to 1 to enable marker debug output */
static const int marker_debug;

/*
 * markers_mutex nests inside module_mutex. Markers mutex protects the builtin
 * and module markers and the hash table.
 */
static DEFINE_MUTEX(markers_mutex);

static CDS_LIST_HEAD(libs);


void lock_markers(void)
{
        pthread_mutex_lock(&markers_mutex);
}

void unlock_markers(void)
{
        pthread_mutex_unlock(&markers_mutex);
}

/*
 * Marker hash table, containing the active markers.
 * Protected by module_mutex.
 */
#define MARKER_HASH_BITS 6
#define MARKER_TABLE_SIZE (1 << MARKER_HASH_BITS)
static struct cds_hlist_head marker_table[MARKER_TABLE_SIZE];

/*
 * Note about RCU :
 * It is used to make sure every handler has finished using its private data
 * between two consecutive operation (add or remove) on a given marker.  It is
 * also used to delay the free of multiple probes array until a quiescent state
 * is reached.
 * marker entries modifications are protected by the markers_mutex.
 */
struct marker_entry {
        struct cds_hlist_node hlist;
        char *format;
        char *name;
                        /* Probe wrapper */
        void (*call)(const struct marker *mdata, void *call_private, struct registers *regs, ...);
        struct marker_probe_closure single;
        struct marker_probe_closure *multi;
        int refcount;   /* Number of times armed. 0 if disarmed. */
        struct rcu_head rcu;
        void *oldptr;
        int rcu_pending;
        u16 channel_id;
        u16 event_id;
        unsigned char ptype:1;
        unsigned char format_allocated:1;
        char channel[0];        /* Contains channel'\0'name'\0'format'\0' */
};

#ifdef CONFIG_MARKERS_USERSPACE
static void marker_update_processes(void);
#else
static void marker_update_processes(void)
{
}
#endif

/**
 * __mark_empty_function - Empty probe callback
 * @mdata: marker data
 * @probe_private: probe private data
 * @call_private: call site private data
 * @fmt: format string
 * @...: variable argument list
 *
 * Empty callback provided as a probe to the markers. By providing this to a
 * disabled marker, we make sure the  execution flow is always valid even
 * though the function pointer change and the marker enabling are two distinct
 * operations that modifies the execution flow of preemptible code.
 */
notrace void __mark_empty_function(const struct marker *mdata,
        void *probe_private, struct registers *regs, void *call_private, const char *fmt, va_list *args)
{
}
//ust// EXPORT_SYMBOL_GPL(__mark_empty_function);

/*
 * marker_probe_cb Callback that prepares the variable argument list for probes.
 * @mdata: pointer of type struct marker
 * @call_private: caller site private data
 * @...:  Variable argument list.
 *
 * Since we do not use "typical" pointer based RCU in the 1 argument case, we
 * need to put a full cmm_smp_rmb() in this branch. This is why we do not use
 * rcu_dereference() for the pointer read.
 */
notrace void marker_probe_cb(const struct marker *mdata,
                void *call_private, struct registers *regs, ...)
{
        va_list args;
        char ptype;

        /*
         * rcu_read_lock_sched does two things : disabling preemption to make
         * sure the teardown of the callbacks can be done correctly when they
         * are in modules and they insure RCU read coherency.
         */
//ust// rcu_read_lock_sched_notrace();
        ptype = mdata->ptype;
        if (likely(!ptype)) {
                marker_probe_func *func;
                /* Must read the ptype before ptr. They are not data dependant,
                 * so we put an explicit cmm_smp_rmb() here. */
                cmm_smp_rmb();
                func = mdata->single.func;
                /* Must read the ptr before private data. They are not data
                 * dependant, so we put an explicit cmm_smp_rmb() here. */
                cmm_smp_rmb();
                va_start(args, regs);
                func(mdata, mdata->single.probe_private, regs, call_private,
                        mdata->format, &args);
                va_end(args);
        } else {
                struct marker_probe_closure *multi;
                int i;
                /*
                 * Read mdata->ptype before mdata->multi.
                 */
                cmm_smp_rmb();
                multi = mdata->multi;
                /*
                 * multi points to an array, therefore accessing the array
                 * depends on reading multi. However, even in this case,
                 * we must insure that the pointer is read _before_ the array
                 * data. Same as rcu_dereference, but we need a full cmm_smp_rmb()
                 * in the fast path, so put the explicit cmm_barrier here.
                 */
                cmm_smp_read_barrier_depends();
                for (i = 0; multi[i].func; i++) {
                        va_start(args, regs);
                        multi[i].func(mdata, multi[i].probe_private,
                                regs, call_private, mdata->format, &args);
                        va_end(args);
                }
        }
//ust// rcu_read_unlock_sched_notrace();
}
//ust// EXPORT_SYMBOL_GPL(marker_probe_cb);

/*
 * marker_probe_cb Callback that does not prepare the variable argument list.
 * @mdata: pointer of type struct marker
 * @call_private: caller site private data
 * @...:  Variable argument list.
 *
 * Should be connected to markers "MARK_NOARGS".
 */
static notrace void marker_probe_cb_noarg(const struct marker *mdata,
                void *call_private, struct registers *regs, ...)
{
        va_list args;   /* not initialized */
        char ptype;

//ust// rcu_read_lock_sched_notrace();
        ptype = mdata->ptype;
        if (likely(!ptype)) {
                marker_probe_func *func;
                /* Must read the ptype before ptr. They are not data dependant,
                 * so we put an explicit cmm_smp_rmb() here. */
                cmm_smp_rmb();
                func = mdata->single.func;
                /* Must read the ptr before private data. They are not data
                 * dependant, so we put an explicit cmm_smp_rmb() here. */
                cmm_smp_rmb();
                func(mdata, mdata->single.probe_private, regs, call_private,
                        mdata->format, &args);
        } else {
                struct marker_probe_closure *multi;
                int i;
                /*
                 * Read mdata->ptype before mdata->multi.
                 */
                cmm_smp_rmb();
                multi = mdata->multi;
                /*
                 * multi points to an array, therefore accessing the array
                 * depends on reading multi. However, even in this case,
                 * we must insure that the pointer is read _before_ the array
                 * data. Same as rcu_dereference, but we need a full cmm_smp_rmb()
                 * in the fast path, so put the explicit cmm_barrier here.
                 */
                cmm_smp_read_barrier_depends();
                for (i = 0; multi[i].func; i++)
                        multi[i].func(mdata, multi[i].probe_private, regs,
                                call_private, mdata->format, &args);
        }
//ust// rcu_read_unlock_sched_notrace();
}

static void free_old_closure(struct rcu_head *head)
{
        struct marker_entry *entry = _ust_container_of(head,
                struct marker_entry, rcu);
        free(entry->oldptr);
        /* Make sure we free the data before setting the pending flag to 0 */
        cmm_smp_wmb();
        entry->rcu_pending = 0;
}

static void debug_print_probes(struct marker_entry *entry)
{
        int i;

        if (!marker_debug)
                return;

        if (!entry->ptype) {
                DBG("Single probe : %p %p",
                        entry->single.func,
                        entry->single.probe_private);
        } else {
                for (i = 0; entry->multi[i].func; i++)
                        DBG("Multi probe %d : %p %p", i,
                                entry->multi[i].func,
                                entry->multi[i].probe_private);
        }
}

static struct marker_probe_closure *
marker_entry_add_probe(struct marker_entry *entry,
                marker_probe_func *probe, void *probe_private)
{
        int nr_probes = 0;
        struct marker_probe_closure *old, *new;

        WARN_ON(!probe);

        debug_print_probes(entry);
        old = entry->multi;
        if (!entry->ptype) {
                if (entry->single.func == probe &&
                                entry->single.probe_private == probe_private)
                        return ERR_PTR(-EBUSY);
                if (entry->single.func == __mark_empty_function) {
                        /* 0 -> 1 probes */
                        entry->single.func = probe;
                        entry->single.probe_private = probe_private;
                        entry->refcount = 1;
                        entry->ptype = 0;
                        debug_print_probes(entry);
                        return NULL;
                } else {
                        /* 1 -> 2 probes */
                        nr_probes = 1;
                        old = NULL;
                }
        } else {
                /* (N -> N+1), (N != 0, 1) probes */
                for (nr_probes = 0; old[nr_probes].func; nr_probes++)
                        if (old[nr_probes].func == probe
                                        && old[nr_probes].probe_private
                                                == probe_private)
                                return ERR_PTR(-EBUSY);
        }
        /* + 2 : one for new probe, one for NULL func */
        new = zmalloc((nr_probes + 2) * sizeof(struct marker_probe_closure));
        if (new == NULL)
                return ERR_PTR(-ENOMEM);
        if (!old)
                new[0] = entry->single;
        else
                memcpy(new, old,
                        nr_probes * sizeof(struct marker_probe_closure));
        new[nr_probes].func = probe;
        new[nr_probes].probe_private = probe_private;
        entry->refcount = nr_probes + 1;
        entry->multi = new;
        entry->ptype = 1;
        debug_print_probes(entry);
        return old;
}

static struct marker_probe_closure *
marker_entry_remove_probe(struct marker_entry *entry,
                marker_probe_func *probe, void *probe_private)
{
        int nr_probes = 0, nr_del = 0, i;
        struct marker_probe_closure *old, *new;

        old = entry->multi;

        debug_print_probes(entry);
        if (!entry->ptype) {
                /* 0 -> N is an error */
                WARN_ON(entry->single.func == __mark_empty_function);
                /* 1 -> 0 probes */
                WARN_ON(probe && entry->single.func != probe);
                WARN_ON(entry->single.probe_private != probe_private);
                entry->single.func = __mark_empty_function;
                entry->refcount = 0;
                entry->ptype = 0;
                debug_print_probes(entry);
                return NULL;
        } else {
                /* (N -> M), (N > 1, M >= 0) probes */
                for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
                        if ((!probe || old[nr_probes].func == probe)
                                        && old[nr_probes].probe_private
                                                == probe_private)
                                nr_del++;
                }
        }

        if (nr_probes - nr_del == 0) {
                /* N -> 0, (N > 1) */
                entry->single.func = __mark_empty_function;
                entry->refcount = 0;
                entry->ptype = 0;
        } else if (nr_probes - nr_del == 1) {
                /* N -> 1, (N > 1) */
                for (i = 0; old[i].func; i++)
                        if ((probe && old[i].func != probe) ||
                                        old[i].probe_private != probe_private)
                                entry->single = old[i];
                entry->refcount = 1;
                entry->ptype = 0;
        } else {
                int j = 0;
                /* N -> M, (N > 1, M > 1) */
                /* + 1 for NULL */
                new = zmalloc((nr_probes - nr_del + 1) * sizeof(struct marker_probe_closure));
                if (new == NULL)
                        return ERR_PTR(-ENOMEM);
                for (i = 0; old[i].func; i++)
                        if ((probe && old[i].func != probe) ||
                                        old[i].probe_private != probe_private)
                                new[j++] = old[i];
                entry->refcount = nr_probes - nr_del;
                entry->ptype = 1;
                entry->multi = new;
        }
        debug_print_probes(entry);
        return old;
}

/*
 * Get marker if the marker is present in the marker hash table.
 * Must be called with markers_mutex held.
 * Returns NULL if not present.
 */
static struct marker_entry *get_marker(const char *channel, const char *name)
{
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;
        struct marker_entry *e;
        size_t channel_len = strlen(channel) + 1;
        size_t name_len = strlen(name) + 1;
        u32 hash;

        hash = jhash(channel, channel_len-1, 0) ^ jhash(name, name_len-1, 0);
        head = &marker_table[hash & ((1 << MARKER_HASH_BITS)-1)];
        cds_hlist_for_each_entry(e, node, head, hlist) {
                if (!strcmp(channel, e->channel) && !strcmp(name, e->name))
                        return e;
        }
        return NULL;
}

/*
 * Add the marker to the marker hash table. Must be called with markers_mutex
 * held.
 */
static struct marker_entry *add_marker(const char *channel, const char *name,
                const char *format)
{
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;
        struct marker_entry *e;
        size_t channel_len = strlen(channel) + 1;
        size_t name_len = strlen(name) + 1;
        size_t format_len = 0;
        u32 hash;

        hash = jhash(channel, channel_len-1, 0) ^ jhash(name, name_len-1, 0);
        if (format)
                format_len = strlen(format) + 1;
        head = &marker_table[hash & ((1 << MARKER_HASH_BITS)-1)];
        cds_hlist_for_each_entry(e, node, head, hlist) {
                if (!strcmp(channel, e->channel) && !strcmp(name, e->name)) {
                        DBG("Marker %s.%s busy", channel, name);
                        return ERR_PTR(-EBUSY); /* Already there */
                }
        }
        /*
         * Using zmalloc here to allocate a variable length element. Could
         * cause some memory fragmentation if overused.
         */
        e = zmalloc(sizeof(struct marker_entry)
                    + channel_len + name_len + format_len);
        if (!e)
                return ERR_PTR(-ENOMEM);
        memcpy(e->channel, channel, channel_len);
        e->name = &e->channel[channel_len];
        memcpy(e->name, name, name_len);
        if (format) {
                e->format = &e->name[name_len];
                memcpy(e->format, format, format_len);
                if (strcmp(e->format, MARK_NOARGS) == 0)
                        e->call = marker_probe_cb_noarg;
                else
                        e->call = marker_probe_cb;
                trace_mark(metadata, core_marker_format,
                           "channel %s name %s format %s",
                           e->channel, e->name, e->format);
        } else {
                e->format = NULL;
                e->call = marker_probe_cb;
        }
        e->single.func = __mark_empty_function;
        e->single.probe_private = NULL;
        e->multi = NULL;
        e->ptype = 0;
        e->format_allocated = 0;
        e->refcount = 0;
        e->rcu_pending = 0;
        cds_hlist_add_head(&e->hlist, head);
        return e;
}

/*
 * Remove the marker from the marker hash table. Must be called with mutex_lock
 * held.
 */
static int remove_marker(const char *channel, const char *name)
{
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;
        struct marker_entry *e;
        int found = 0;
        size_t channel_len = strlen(channel) + 1;
        size_t name_len = strlen(name) + 1;
        u32 hash;
        int ret;

        hash = jhash(channel, channel_len-1, 0) ^ jhash(name, name_len-1, 0);
        head = &marker_table[hash & ((1 << MARKER_HASH_BITS)-1)];
        cds_hlist_for_each_entry(e, node, head, hlist) {
                if (!strcmp(channel, e->channel) && !strcmp(name, e->name)) {
                        found = 1;
                        break;
                }
        }
        if (!found)
                return -ENOENT;
        if (e->single.func != __mark_empty_function)
                return -EBUSY;
        cds_hlist_del(&e->hlist);
        if (e->format_allocated)
                free(e->format);
        ret = ltt_channels_unregister(e->channel);
        WARN_ON(ret);
        /* Make sure the call_rcu has been executed */
//ust// if (e->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        free(e);
        return 0;
}

/*
 * Set the mark_entry format to the format found in the element.
 */
static int marker_set_format(struct marker_entry *entry, const char *format)
{
        entry->format = strdup(format);
        if (!entry->format)
                return -ENOMEM;
        entry->format_allocated = 1;

        trace_mark(metadata, core_marker_format,
                   "channel %s name %s format %s",
                   entry->channel, entry->name, entry->format);
        return 0;
}

/*
 * Sets the probe callback corresponding to one marker.
 */
static int set_marker(struct marker_entry *entry, struct marker *elem,
                int active)
{
        int ret = 0;
        WARN_ON(strcmp(entry->name, elem->name) != 0);

        if (entry->format) {
                if (strcmp(entry->format, elem->format) != 0) {
                        ERR("Format mismatch for probe %s (%s), marker (%s)",
                                entry->name,
                                entry->format,
                                elem->format);
                        return -EPERM;
                }
        } else {
                ret = marker_set_format(entry, elem->format);
                if (ret)
                        return ret;
        }

        /*
         * probe_cb setup (statically known) is done here. It is
         * asynchronous with the rest of execution, therefore we only
         * pass from a "safe" callback (with argument) to an "unsafe"
         * callback (does not set arguments).
         */
        elem->call = entry->call;
        elem->channel_id = entry->channel_id;
        elem->event_id = entry->event_id;
        /*
         * Sanity check :
         * We only update the single probe private data when the ptr is
         * set to a _non_ single probe! (0 -> 1 and N -> 1, N != 1)
         */
        WARN_ON(elem->single.func != __mark_empty_function
                && elem->single.probe_private != entry->single.probe_private
                && !elem->ptype);
        elem->single.probe_private = entry->single.probe_private;
        /*
         * Make sure the private data is valid when we update the
         * single probe ptr.
         */
        cmm_smp_wmb();
        elem->single.func = entry->single.func;
        /*
         * We also make sure that the new probe callbacks array is consistent
         * before setting a pointer to it.
         */
        rcu_assign_pointer(elem->multi, entry->multi);
        /*
         * Update the function or multi probe array pointer before setting the
         * ptype.
         */
        cmm_smp_wmb();
        elem->ptype = entry->ptype;

        if (elem->tp_name && (active ^ _imv_read(elem->state))) {
                WARN_ON(!elem->tp_cb);
                /*
                 * It is ok to directly call the probe registration because type
                 * checking has been done in the __trace_mark_tp() macro.
                 */

                if (active) {
                        /*
                         * try_module_get should always succeed because we hold
                         * markers_mutex to get the tp_cb address.
                         */
//ust//                 ret = try_module_get(__module_text_address(
//ust//                         (unsigned long)elem->tp_cb));
//ust//                 BUG_ON(!ret);
                        ret = tracepoint_probe_register_noupdate(
                                elem->tp_name,
                                elem->tp_cb, NULL);
                } else {
                        ret = tracepoint_probe_unregister_noupdate(
                                elem->tp_name,
                                elem->tp_cb, NULL);
                        /*
                         * tracepoint_probe_update_all() must be called
                         * before the module containing tp_cb is unloaded.
                         */
//ust//                 module_put(__module_text_address(
//ust//                         (unsigned long)elem->tp_cb));
                }
        }
        elem->state__imv = active;

        return ret;
}

/*
 * Disable a marker and its probe callback.
 * Note: only waiting an RCU period after setting elem->call to the empty
 * function insures that the original callback is not used anymore. This insured
 * by rcu_read_lock_sched around the call site.
 */
static void disable_marker(struct marker *elem)
{
        int ret;

        /* leave "call" as is. It is known statically. */
        if (elem->tp_name && _imv_read(elem->state)) {
                WARN_ON(!elem->tp_cb);
                /*
                 * It is ok to directly call the probe registration because type
                 * checking has been done in the __trace_mark_tp() macro.
                 */
                ret = tracepoint_probe_unregister_noupdate(elem->tp_name,
                        elem->tp_cb, NULL);
                WARN_ON(ret);
                /*
                 * tracepoint_probe_update_all() must be called
                 * before the module containing tp_cb is unloaded.
                 */
//ust//         module_put(__module_text_address((unsigned long)elem->tp_cb));
        }
        elem->state__imv = 0;
        elem->single.func = __mark_empty_function;
        /* Update the function before setting the ptype */
        cmm_smp_wmb();
        elem->ptype = 0;        /* single probe */
        /*
         * Leave the private data and channel_id/event_id there, because removal
         * is racy and should be done only after an RCU period. These are never
         * used until the next initialization anyway.
         */
}

/*
 * is_marker_enabled - Check if a marker is enabled
 * @channel: channel name
 * @name: marker name
 *
 * Returns 1 if the marker is enabled, 0 if disabled.
 */
int is_marker_enabled(const char *channel, const char *name)
{
        struct marker_entry *entry;

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        pthread_mutex_unlock(&markers_mutex);

        return entry && !!entry->refcount;
}

/**
 * marker_update_probe_range - Update a probe range
 * @begin: beginning of the range
 * @end: end of the range
 *
 * Updates the probe callback corresponding to a range of markers.
 */
void marker_update_probe_range(struct marker * const *begin,
        struct marker * const *end)
{
        struct marker * const *iter;
        struct marker_entry *mark_entry;

        pthread_mutex_lock(&markers_mutex);
        for (iter = begin; iter < end; iter++) {
                mark_entry = get_marker((*iter)->channel, (*iter)->name);
                if (mark_entry) {
                        set_marker(mark_entry, *iter, !!mark_entry->refcount);
                        /*
                         * ignore error, continue
                         */

                        /* This is added for UST. We emit a core_marker_id event
                         * for markers that are already registered to a probe
                         * upon library load. Otherwise, no core_marker_id will
                         * be generated for these markers. Is this the right thing
                         * to do?
                         */
                        trace_mark(metadata, core_marker_id,
                                   "channel %s name %s event_id %hu "
                                   "int #1u%zu long #1u%zu pointer #1u%zu "
                                   "size_t #1u%zu alignment #1u%u",
                                   (*iter)->channel, (*iter)->name, mark_entry->event_id,
                                   sizeof(int), sizeof(long), sizeof(void *),
                                   sizeof(size_t), ltt_get_alignment());
                } else {
                        disable_marker(*iter);
                }
        }
        pthread_mutex_unlock(&markers_mutex);
}

static void lib_update_markers(void)
{
        struct lib *lib;

        /* FIXME: we should probably take a mutex here on libs */
//ust// pthread_mutex_lock(&module_mutex);
        cds_list_for_each_entry(lib, &libs, list)
                marker_update_probe_range(lib->markers_start,
                                lib->markers_start + lib->markers_count);
//ust// pthread_mutex_unlock(&module_mutex);
}

/*
 * Update probes, removing the faulty probes.
 *
 * Internal callback only changed before the first probe is connected to it.
 * Single probe private data can only be changed on 0 -> 1 and 2 -> 1
 * transitions.  All other transitions will leave the old private data valid.
 * This makes the non-atomicity of the callback/private data updates valid.
 *
 * "special case" updates :
 * 0 -> 1 callback
 * 1 -> 0 callback
 * 1 -> 2 callbacks
 * 2 -> 1 callbacks
 * Other updates all behave the same, just like the 2 -> 3 or 3 -> 2 updates.
 * Site effect : marker_set_format may delete the marker entry (creating a
 * replacement).
 */
static void marker_update_probes(void)
{
        lib_update_markers();
        tracepoint_probe_update_all();
        /* Update immediate values */
        core_imv_update();
//ust// module_imv_update(); /* FIXME: need to port for libs? */
        marker_update_processes();
}

/**
 * marker_probe_register -  Connect a probe to a marker
 * @channel: marker channel
 * @name: marker name
 * @format: format string
 * @probe: probe handler
 * @probe_private: probe private data
 *
 * private data must be a valid allocated memory address, or NULL.
 * Returns 0 if ok, error value on error.
 * The probe address must at least be aligned on the architecture pointer size.
 */
int marker_probe_register(const char *channel, const char *name,
                          const char *format, marker_probe_func *probe,
                          void *probe_private)
{
        struct marker_entry *entry;
        int ret = 0, ret_err;
        struct marker_probe_closure *old;
        int first_probe = 0;

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        if (!entry) {
                first_probe = 1;
                entry = add_marker(channel, name, format);
                if (IS_ERR(entry))
                        ret = PTR_ERR(entry);
                if (ret)
                        goto end;
                ret = ltt_channels_register(channel);
                if (ret)
                        goto error_remove_marker;
                ret = ltt_channels_get_index_from_name(channel);
                if (ret < 0)
                        goto error_unregister_channel;
                entry->channel_id = ret;
                ret = ltt_channels_get_event_id(channel, name);
                if (ret < 0)
                        goto error_unregister_channel;
                entry->event_id = ret;
                ret = 0;
                trace_mark(metadata, core_marker_id,
                           "channel %s name %s event_id %hu "
                           "int #1u%zu long #1u%zu pointer #1u%zu "
                           "size_t #1u%zu alignment #1u%u",
                           channel, name, entry->event_id,
                           sizeof(int), sizeof(long), sizeof(void *),
                           sizeof(size_t), ltt_get_alignment());
        } else if (format) {
                if (!entry->format)
                        ret = marker_set_format(entry, format);
                else if (strcmp(entry->format, format))
                        ret = -EPERM;
                if (ret)
                        goto end;
        }

        /*
         * If we detect that a call_rcu is pending for this marker,
         * make sure it's executed now.
         */
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        old = marker_entry_add_probe(entry, probe, probe_private);
        if (IS_ERR(old)) {
                ret = PTR_ERR(old);
                if (first_probe)
                        goto error_unregister_channel;
                else
                        goto end;
        }
        pthread_mutex_unlock(&markers_mutex);

        /* Activate marker if necessary */
        marker_update_probes();

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        if (!entry)
                goto end;
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        entry->oldptr = old;
        entry->rcu_pending = 1;
        /* write rcu_pending before calling the RCU callback */
        cmm_smp_wmb();
//ust// call_rcu_sched(&entry->rcu, free_old_closure);
        synchronize_rcu(); free_old_closure(&entry->rcu);
        goto end;

error_unregister_channel:
        ret_err = ltt_channels_unregister(channel);
        WARN_ON(ret_err);
error_remove_marker:
        ret_err = remove_marker(channel, name);
        WARN_ON(ret_err);
end:
        pthread_mutex_unlock(&markers_mutex);
        return ret;
}
//ust// EXPORT_SYMBOL_GPL(marker_probe_register);

/**
 * marker_probe_unregister -  Disconnect a probe from a marker
 * @channel: marker channel
 * @name: marker name
 * @probe: probe function pointer
 * @probe_private: probe private data
 *
 * Returns the private data given to marker_probe_register, or an ERR_PTR().
 * We do not need to call a synchronize_sched to make sure the probes have
 * finished running before doing a module unload, because the module unload
 * itself uses stop_machine(), which insures that every preempt disabled section
 * have finished.
 */
int marker_probe_unregister(const char *channel, const char *name,
                            marker_probe_func *probe, void *probe_private)
{
        struct marker_entry *entry;
        struct marker_probe_closure *old;
        int ret = -ENOENT;

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        if (!entry)
                goto end;
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        old = marker_entry_remove_probe(entry, probe, probe_private);
        pthread_mutex_unlock(&markers_mutex);

        marker_update_probes();

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        if (!entry)
                goto end;
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        entry->oldptr = old;
        entry->rcu_pending = 1;
        /* write rcu_pending before calling the RCU callback */
        cmm_smp_wmb();
//ust// call_rcu_sched(&entry->rcu, free_old_closure);
        synchronize_rcu(); free_old_closure(&entry->rcu);
        remove_marker(channel, name);   /* Ignore busy error message */
        ret = 0;
end:
        pthread_mutex_unlock(&markers_mutex);
        return ret;
}
//ust// EXPORT_SYMBOL_GPL(marker_probe_unregister);

static struct marker_entry *
get_marker_from_private_data(marker_probe_func *probe, void *probe_private)
{
        struct marker_entry *entry;
        unsigned int i;
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;

        for (i = 0; i < MARKER_TABLE_SIZE; i++) {
                head = &marker_table[i];
                cds_hlist_for_each_entry(entry, node, head, hlist) {
                        if (!entry->ptype) {
                                if (entry->single.func == probe
                                                && entry->single.probe_private
                                                == probe_private)
                                        return entry;
                        } else {
                                struct marker_probe_closure *closure;
                                closure = entry->multi;
                                for (i = 0; closure[i].func; i++) {
                                        if (closure[i].func == probe &&
                                                        closure[i].probe_private
                                                        == probe_private)
                                                return entry;
                                }
                        }
                }
        }
        return NULL;
}

/**
 * marker_probe_unregister_private_data -  Disconnect a probe from a marker
 * @probe: probe function
 * @probe_private: probe private data
 *
 * Unregister a probe by providing the registered private data.
 * Only removes the first marker found in hash table.
 * Return 0 on success or error value.
 * We do not need to call a synchronize_sched to make sure the probes have
 * finished running before doing a module unload, because the module unload
 * itself uses stop_machine(), which insures that every preempt disabled section
 * have finished.
 */
int marker_probe_unregister_private_data(marker_probe_func *probe,
                void *probe_private)
{
        struct marker_entry *entry;
        int ret = 0;
        struct marker_probe_closure *old;
        char *channel = NULL, *name = NULL;

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker_from_private_data(probe, probe_private);
        if (!entry) {
                ret = -ENOENT;
                goto end;
        }
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        old = marker_entry_remove_probe(entry, NULL, probe_private);
        channel = strdup(entry->channel);
        name = strdup(entry->name);
        pthread_mutex_unlock(&markers_mutex);

        marker_update_probes();

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        if (!entry)
                goto end;
//ust// if (entry->rcu_pending)
//ust//         rcu_cmm_barrier_sched();
        entry->oldptr = old;
        entry->rcu_pending = 1;
        /* write rcu_pending before calling the RCU callback */
        cmm_smp_wmb();
//ust// call_rcu_sched(&entry->rcu, free_old_closure);
        synchronize_rcu(); free_old_closure(&entry->rcu);
        /* Ignore busy error message */
        remove_marker(channel, name);
end:
        pthread_mutex_unlock(&markers_mutex);
        free(channel);
        free(name);
        return ret;
}
//ust// EXPORT_SYMBOL_GPL(marker_probe_unregister_private_data);

/**
 * marker_get_private_data - Get a marker's probe private data
 * @channel: marker channel
 * @name: marker name
 * @probe: probe to match
 * @num: get the nth matching probe's private data
 *
 * Returns the nth private data pointer (starting from 0) matching, or an
 * ERR_PTR.
 * Returns the private data pointer, or an ERR_PTR.
 * The private data pointer should _only_ be dereferenced if the caller is the
 * owner of the data, or its content could vanish. This is mostly used to
 * confirm that a caller is the owner of a registered probe.
 */
void *marker_get_private_data(const char *channel, const char *name,
                              marker_probe_func *probe, int num)
{
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;
        struct marker_entry *e;
        size_t channel_len = strlen(channel) + 1;
        size_t name_len = strlen(name) + 1;
        int i;
        u32 hash;

        hash = jhash(channel, channel_len-1, 0) ^ jhash(name, name_len-1, 0);
        head = &marker_table[hash & ((1 << MARKER_HASH_BITS)-1)];
        cds_hlist_for_each_entry(e, node, head, hlist) {
                if (!strcmp(channel, e->channel) && !strcmp(name, e->name)) {
                        if (!e->ptype) {
                                if (num == 0 && e->single.func == probe)
                                        return e->single.probe_private;
                        } else {
                                struct marker_probe_closure *closure;
                                int match = 0;
                                closure = e->multi;
                                for (i = 0; closure[i].func; i++) {
                                        if (closure[i].func != probe)
                                                continue;
                                        if (match++ == num)
                                                return closure[i].probe_private;
                                }
                        }
                        break;
                }
        }
        return ERR_PTR(-ENOENT);
}
//ust// EXPORT_SYMBOL_GPL(marker_get_private_data);

/**
 * markers_compact_event_ids - Compact markers event IDs and reassign channels
 *
 * Called when no channel users are active by the channel infrastructure.
 * Called with lock_markers() and channel mutex held.
 */
//ust// void markers_compact_event_ids(void)
//ust// {
//ust//         struct marker_entry *entry;
//ust//         unsigned int i;
//ust//         struct hlist_head *head;
//ust//         struct hlist_node *node;
//ust//         int ret;
//ust// 
//ust//         for (i = 0; i < MARKER_TABLE_SIZE; i++) {
//ust//                 head = &marker_table[i];
//ust//                 hlist_for_each_entry(entry, node, head, hlist) {
//ust//                         ret = ltt_channels_get_index_from_name(entry->channel);
//ust//                         WARN_ON(ret < 0);
//ust//                         entry->channel_id = ret;
//ust//                         ret = _ltt_channels_get_event_id(entry->channel,
//ust//                                                          entry->name);
//ust//                         WARN_ON(ret < 0);
//ust//                         entry->event_id = ret;
//ust//                 }
//ust//         }
//ust// }

//ust//#ifdef CONFIG_MODULES

/*
 * Returns 0 if current not found.
 * Returns 1 if current found.
 */
int lib_get_iter_markers(struct marker_iter *iter)
{
        struct lib *iter_lib;
        int found = 0;

//ust// pthread_mutex_lock(&module_mutex);
        cds_list_for_each_entry(iter_lib, &libs, list) {
                if (iter_lib < iter->lib)
                        continue;
                else if (iter_lib > iter->lib)
                        iter->marker = NULL;
                found = marker_get_iter_range(&iter->marker,
                        iter_lib->markers_start,
                        iter_lib->markers_start + iter_lib->markers_count);
                if (found) {
                        iter->lib = iter_lib;
                        break;
                }
        }
//ust// pthread_mutex_unlock(&module_mutex);
        return found;
}

/**
 * marker_get_iter_range - Get a next marker iterator given a range.
 * @marker: current markers (in), next marker (out)
 * @begin: beginning of the range
 * @end: end of the range
 *
 * Returns whether a next marker has been found (1) or not (0).
 * Will return the first marker in the range if the input marker is NULL.
 */
int marker_get_iter_range(struct marker * const **marker,
        struct marker * const *begin,
        struct marker * const *end)
{
        if (!*marker && begin != end) {
                *marker = begin;
                return 1;
        }
        if (*marker >= begin && *marker < end)
                return 1;
        return 0;
}
//ust// EXPORT_SYMBOL_GPL(marker_get_iter_range);

static void marker_get_iter(struct marker_iter *iter)
{
        int found = 0;

        found = lib_get_iter_markers(iter);
        if (!found)
                marker_iter_reset(iter);
}

void marker_iter_start(struct marker_iter *iter)
{
        marker_get_iter(iter);
}
//ust// EXPORT_SYMBOL_GPL(marker_iter_start);

void marker_iter_next(struct marker_iter *iter)
{
        iter->marker++;
        /*
         * iter->marker may be invalid because we blindly incremented it.
         * Make sure it is valid by marshalling on the markers, getting the
         * markers from following modules if necessary.
         */
        marker_get_iter(iter);
}
//ust// EXPORT_SYMBOL_GPL(marker_iter_next);

void marker_iter_stop(struct marker_iter *iter)
{
}
//ust// EXPORT_SYMBOL_GPL(marker_iter_stop);

void marker_iter_reset(struct marker_iter *iter)
{
        iter->lib = NULL;
        iter->marker = NULL;
}
//ust// EXPORT_SYMBOL_GPL(marker_iter_reset);

#ifdef CONFIG_MARKERS_USERSPACE
/*
 * must be called with current->user_markers_mutex held
 */
static void free_user_marker(char __user *state, struct cds_hlist_head *head)
{
        struct user_marker *umark;
        struct cds_hlist_node *pos, *n;

        cds_hlist_for_each_entry_safe(umark, pos, n, head, hlist) {
                if (umark->state == state) {
                        cds_hlist_del(&umark->hlist);
                        free(umark);
                }
        }
}

/*
 * Update current process.
 * Note that we have to wait a whole scheduler period before we are sure that
 * every running userspace threads have their markers updated.
 * (synchronize_sched() can be used to insure this).
 */
//ust// void marker_update_process(void)
//ust// {
//ust//         struct user_marker *umark;
//ust//         struct hlist_node *pos;
//ust//         struct marker_entry *entry;
//ust// 
//ust//         pthread_mutex_lock(&markers_mutex);
//ust//         pthread_mutex_lock(&current->group_leader->user_markers_mutex);
//ust//         if (strcmp(current->comm, "testprog") == 0)
//ust//                 DBG("do update pending for testprog");
//ust//         hlist_for_each_entry(umark, pos,
//ust//                         &current->group_leader->user_markers, hlist) {
//ust//                 DBG("Updating marker %s in %s", umark->name, current->comm);
//ust//                 entry = get_marker("userspace", umark->name);
//ust//                 if (entry) {
//ust//                         if (entry->format &&
//ust//                                 strcmp(entry->format, umark->format) != 0) {
//ust//                                 WARN("error, wrong format in process %s",
//ust//                                         current->comm);
//ust//                                 break;
//ust//                         }
//ust//                         if (put_user(!!entry->refcount, umark->state)) {
//ust//                                 WARN("Marker in %s caused a fault",
//ust//                                         current->comm);
//ust//                                 break;
//ust//                         }
//ust//                 } else {
//ust//                         if (put_user(0, umark->state)) {
//ust//                                 WARN("Marker in %s caused a fault", current->comm);
//ust//                                 break;
//ust//                         }
//ust//                 }
//ust//         }
//ust//         clear_thread_flag(TIF_MARKER_PENDING);
//ust//         pthread_mutex_unlock(&current->group_leader->user_markers_mutex);
//ust//         pthread_mutex_unlock(&markers_mutex);
//ust// }

/*
 * Called at process exit and upon do_execve().
 * We assume that when the leader exits, no more references can be done to the
 * leader structure by the other threads.
 */
void exit_user_markers(struct task_struct *p)
{
        struct user_marker *umark;
        struct cds_hlist_node *pos, *n;

        if (thread_group_leader(p)) {
                pthread_mutex_lock(&markers_mutex);
                pthread_mutex_lock(&p->user_markers_mutex);
                cds_hlist_for_each_entry_safe(umark, pos, n, &p->user_markers,
                        hlist)
                    free(umark);
                INIT_HLIST_HEAD(&p->user_markers);
                p->user_markers_sequence++;
                pthread_mutex_unlock(&p->user_markers_mutex);
                pthread_mutex_unlock(&markers_mutex);
        }
}

int is_marker_enabled(const char *channel, const char *name)
{
        struct marker_entry *entry;

        pthread_mutex_lock(&markers_mutex);
        entry = get_marker(channel, name);
        pthread_mutex_unlock(&markers_mutex);

        return entry && !!entry->refcount;
}
//ust// #endif

int marker_module_notify(struct notifier_block *self,
                         unsigned long val, void *data)
{
        struct module *mod = data;

        switch (val) {
        case MODULE_STATE_COMING:
                marker_update_probe_range(mod->markers,
                        mod->markers + mod->num_markers);
                break;
        case MODULE_STATE_GOING:
                marker_update_probe_range(mod->markers,
                        mod->markers + mod->num_markers);
                break;
        }
        return 0;
}

struct notifier_block marker_module_nb = {
        .notifier_call = marker_module_notify,
        .priority = 0,
};

//ust// static int init_markers(void)
//ust// {
//ust//         return register_module_notifier(&marker_module_nb);
//ust// }
//ust// __initcall(init_markers);
/* TODO: call marker_module_nb() when a library is linked at runtime (dlopen)? */

#endif /* CONFIG_MODULES */

void ltt_dump_marker_state(struct ust_trace *trace)
{
        struct marker_entry *entry;
        struct ltt_probe_private_data call_data;
        struct cds_hlist_head *head;
        struct cds_hlist_node *node;
        unsigned int i;

        pthread_mutex_lock(&markers_mutex);
        call_data.trace = trace;
        call_data.serializer = NULL;

        for (i = 0; i < MARKER_TABLE_SIZE; i++) {
                head = &marker_table[i];
                cds_hlist_for_each_entry(entry, node, head, hlist) {
                        __trace_mark(0, metadata, core_marker_id,
                                &call_data,
                                "channel %s name %s event_id %hu "
                                "int #1u%zu long #1u%zu pointer #1u%zu "
                                "size_t #1u%zu alignment #1u%u",
                                entry->channel,
                                entry->name,
                                entry->event_id,
                                sizeof(int), sizeof(long),
                                sizeof(void *), sizeof(size_t),
                                ltt_get_alignment());
                        if (entry->format)
                                __trace_mark(0, metadata,
                                        core_marker_format,
                                        &call_data,
                                        "channel %s name %s format %s",
                                        entry->channel,
                                        entry->name,
                                        entry->format);
                }
        }
        pthread_mutex_unlock(&markers_mutex);
}
//ust// EXPORT_SYMBOL_GPL(ltt_dump_marker_state);

static void (*new_marker_cb)(struct marker *) = NULL;

void marker_set_new_marker_cb(void (*cb)(struct marker *))
{
        new_marker_cb = cb;
}

static void new_markers(struct marker * const *start, struct marker * const *end)
{
        if (new_marker_cb) {
                struct marker * const *m;
                for(m=start; m < end; m++) {
                        new_marker_cb(*m);
                }
        }
}

int marker_register_lib(struct marker * const *markers_start, int markers_count)
{
        struct lib *pl, *iter;

        pl = (struct lib *) zmalloc(sizeof(struct lib));

        pl->markers_start = markers_start;
        pl->markers_count = markers_count;

        /* FIXME: maybe protect this with its own mutex? */
        lock_markers();

        /*
         * We sort the libs by struct lib pointer address.
         */
        cds_list_for_each_entry_reverse(iter, &libs, list) {
                BUG_ON(iter == pl);    /* Should never be in the list twice */
                if (iter < pl) {
                        /* We belong to the location right after iter. */
                        cds_list_add(&pl->list, &iter->list);
                        goto lib_added;
                }
        }
        /* We should be added at the head of the list */
        cds_list_add(&pl->list, &libs);
lib_added:
        unlock_markers();

        new_markers(markers_start, markers_start + markers_count);

        /* FIXME: update just the loaded lib */
        lib_update_markers();

        DBG("just registered a markers section from %p and having %d markers", markers_start, markers_count);
        
        return 0;
}

int marker_unregister_lib(struct marker * const *markers_start)
{
        struct lib *lib;

        /*FIXME: implement; but before implementing, marker_register_lib must
          have appropriate locking. */

        lock_markers();

        /* FIXME: we should probably take a mutex here on libs */
//ust// pthread_mutex_lock(&module_mutex);
        cds_list_for_each_entry(lib, &libs, list) {
                if(lib->markers_start == markers_start) {
                        struct lib *lib2free = lib;
                        cds_list_del(&lib->list);
                        free(lib2free);
                        break;
                }
        }

        unlock_markers();

        return 0;
}

static int initialized = 0;

void __attribute__((constructor)) init_markers(void)
{
        if (!initialized) {
                marker_register_lib(__start___markers_ptrs,
                        __stop___markers_ptrs
                        - __start___markers_ptrs);
                initialized = 1;
        }
}

void __attribute__((destructor)) destroy_markers(void)
{
        marker_unregister_lib(__start___markers_ptrs);
}