Rewrite last GPL bits in relay.c and relay.h
authorPierre-Marc Fournier <pierre-marc.fournier@polymtl.ca>
Tue, 20 Oct 2009 18:08:52 +0000 (14:08 -0400)
committerPierre-Marc Fournier <pierre-marc.fournier@polymtl.ca>
Mon, 16 Nov 2009 03:08:44 +0000 (22:08 -0500)
Also update other files accordingly.

17 files changed:
libust/Makefile.am
libust/buffer.h [deleted file]
libust/buffers.c [new file with mode: 0644]
libust/buffers.h [new file with mode: 0644]
libust/channels.c
libust/channels.h
libust/relay.c [deleted file]
libust/relay.h [deleted file]
libust/serialize.c
libust/tracectl.c
libust/tracer.c
libust/tracer.h
libustcmd/ustcmd.c
tests/hello/Makefile.am
tests/hello/hello.c
ustd/lowlevel.c
ustd/ustd.c

index 6d99d0d5d01b8b7f3ccd4743efcf48e78cc2bf2a..087dc0a978a3475ae56ef799243a1d09ddbea4db 100644 (file)
@@ -1,7 +1,7 @@
 AM_CPPFLAGS = -I$(top_srcdir)/include -I$(top_srcdir)/libustcomm
 
 lib_LTLIBRARIES = libust.la
-libust_la_SOURCES = buffer.h marker.c tracepoint.c channels.c channels.h marker-control.c marker-control.h relay.c relay.h tracer.c tracer.h tracercore.c tracercore.h serialize.c tracectl.c $(top_builddir)/libustcomm/ustcomm.c
+libust_la_SOURCES = marker.c tracepoint.c channels.c channels.h marker-control.c marker-control.h buffers.c buffers.h tracer.c tracer.h tracercore.c tracercore.h serialize.c tracectl.c $(top_builddir)/libustcomm/ustcomm.c
 libust_la_LDFLAGS = -no-undefined -version-info 0:0:0
 libust_la_LIBADD = -lpthread
 libust_la_CFLAGS = -DUST_COMPONENT="libust"
diff --git a/libust/buffer.h b/libust/buffer.h
deleted file mode 100644 (file)
index 086f55c..0000000
+++ /dev/null
@@ -1,39 +0,0 @@
-/* Copyright (C) 2005-2009  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
- */
-
-#ifndef UST_BUFFER_H
-#define UST_BUFFER_H
-
-/* Buffer offset macros */
-
-/*
- * BUFFER_TRUNC zeroes the subbuffer offset and the subbuffer number parts of
- * the offset, which leaves only the buffer number.
- */
-#define BUFFER_TRUNC(offset, chan) \
-       ((offset) & (~((chan)->alloc_size-1)))
-#define BUFFER_OFFSET(offset, chan) ((offset) & ((chan)->alloc_size - 1))
-#define SUBBUF_OFFSET(offset, chan) ((offset) & ((chan)->subbuf_size - 1))
-#define SUBBUF_ALIGN(offset, chan) \
-       (((offset) + (chan)->subbuf_size) & (~((chan)->subbuf_size - 1)))
-#define SUBBUF_TRUNC(offset, chan) \
-       ((offset) & (~((chan)->subbuf_size - 1)))
-#define SUBBUF_INDEX(offset, chan) \
-       (BUFFER_OFFSET((offset), chan) >> (chan)->subbuf_size_order)
-
-
-#endif /* UST_BUFFER_H */
diff --git a/libust/buffers.c b/libust/buffers.c
new file mode 100644 (file)
index 0000000..f3331d0
--- /dev/null
@@ -0,0 +1,1422 @@
+/*
+ * buffers.c
+ * LTTng userspace tracer buffering system
+ *
+ * Copyright (C) 2009 - Pierre-Marc Fournier (pierre-marc dot fournier at polymtl dot ca)
+ * Copyright (C) 2008 - Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
+ *
+ * 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 <sys/mman.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <fcntl.h>
+#include <ust/kernelcompat.h>
+#include <kcompat/kref.h>
+#include "buffers.h"
+#include "channels.h"
+#include "tracer.h"
+#include "tracercore.h"
+#include "usterr.h"
+
+static DEFINE_MUTEX(ust_buffers_channels_mutex);
+static LIST_HEAD(ust_buffers_channels);
+
+static int ust_buffers_init_buffer(struct ltt_trace_struct *trace,
+               struct ust_channel *ltt_chan,
+               struct ust_buffer *buf,
+               unsigned int n_subbufs);
+
+static int ust_buffers_alloc_buf(struct ust_buffer *buf, size_t *size)
+{
+       void *ptr;
+       int result;
+
+       *size = PAGE_ALIGN(*size);
+
+       result = buf->shmid = shmget(getpid(), *size, IPC_CREAT | IPC_EXCL | 0700);
+       if(result == -1 && errno == EINVAL) {
+               ERR("shmget() returned EINVAL; maybe /proc/sys/kernel/shmmax should be increased.");
+               return -1;
+       }
+       else if(result == -1) {
+               PERROR("shmget");
+               return -1;
+       }
+
+       ptr = shmat(buf->shmid, NULL, 0);
+       if(ptr == (void *) -1) {
+               perror("shmat");
+               goto destroy_shmem;
+       }
+
+       /* Already mark the shared memory for destruction. This will occur only
+         * when all users have detached.
+        */
+       result = shmctl(buf->shmid, IPC_RMID, NULL);
+       if(result == -1) {
+               perror("shmctl");
+               return -1;
+       }
+
+       buf->buf_data = ptr;
+       buf->buf_size = *size;
+
+       return 0;
+
+       destroy_shmem:
+       result = shmctl(buf->shmid, IPC_RMID, NULL);
+       if(result == -1) {
+               perror("shmctl");
+       }
+
+       return -1;
+}
+
+static struct ust_buffer *ust_buffers_create_buf(struct ust_channel *channel)
+{
+       int result;
+
+       result = ust_buffers_alloc_buf(channel->buf, &channel->alloc_size);
+       if(result)
+               goto free_buf;
+
+       ((struct ust_buffer *)channel->buf)->chan = channel;
+       kref_get(&channel->kref);
+       return channel->buf;
+
+free_buf:
+       return NULL;
+}
+
+static void ust_buffers_destroy_channel(struct kref *kref)
+{
+       struct ust_channel *chan = container_of(kref, struct ust_channel, kref);
+       free(chan);
+}
+
+static void ust_buffers_destroy_buf(struct ust_buffer *buf)
+{
+       struct ust_channel *chan = buf->chan;
+       int result;
+
+       result = munmap(buf->buf_data, buf->buf_size);
+       if(result == -1) {
+               PERROR("munmap");
+       }
+
+       free(buf);
+       kref_put(&chan->kref, ust_buffers_destroy_channel);
+}
+
+/* called from kref_put */
+static void ust_buffers_remove_buf(struct kref *kref)
+{
+       struct ust_buffer *buf = container_of(kref, struct ust_buffer, kref);
+       ust_buffers_destroy_buf(buf);
+}
+
+static struct ust_buffer *ust_buffers_open_buf(struct ust_channel *chan)
+{
+       struct ust_buffer *buf = NULL;
+       int err;
+
+       buf = ust_buffers_create_buf(chan);
+       if (!buf)
+               return NULL;
+
+       kref_init(&buf->kref);
+
+       err = ust_buffers_init_buffer(chan->trace, chan, buf, chan->subbuf_cnt);
+
+       if (err)
+               return ERR_PTR(err);
+
+       return buf;
+
+       /* FIXME: decrementally destroy on error? */
+}
+
+/**
+ *     ust_buffers_close_buf - close a channel buffer
+ *     @buf: buffer
+ */
+static void ust_buffers_close_buf(struct ust_buffer *buf)
+{
+       kref_put(&buf->kref, ust_buffers_remove_buf);
+}
+
+int ust_buffers_channel_open(struct ust_channel *chan, size_t subbuf_size, size_t subbuf_cnt)
+{
+       if(subbuf_size == 0 || subbuf_cnt == 0)
+               return -1;
+
+       chan->version = UST_CHANNEL_VERSION;
+       chan->subbuf_cnt = subbuf_cnt;
+       chan->subbuf_size = subbuf_size;
+       chan->subbuf_size_order = get_count_order(subbuf_size);
+       chan->alloc_size = FIX_SIZE(subbuf_size * subbuf_cnt);
+       kref_init(&chan->kref);
+
+       mutex_lock(&ust_buffers_channels_mutex);
+       chan->buf = ust_buffers_open_buf(chan);
+       if (!chan->buf)
+               goto error;
+       list_add(&chan->list, &ust_buffers_channels);
+       mutex_unlock(&ust_buffers_channels_mutex);
+
+       return 0;
+
+       error:
+       kref_put(&chan->kref, ust_buffers_destroy_channel);
+       mutex_unlock(&ust_buffers_channels_mutex);
+       return -1;
+}
+
+void ust_buffers_channel_close(struct ust_channel *chan)
+{
+       if (!chan)
+               return;
+
+       mutex_lock(&ust_buffers_channels_mutex);
+       if (chan->buf)
+               ust_buffers_close_buf(chan->buf);
+
+       list_del(&chan->list);
+       kref_put(&chan->kref, ust_buffers_destroy_channel);
+       mutex_unlock(&ust_buffers_channels_mutex);
+}
+
+/* _ust_buffers_write()
+ *
+ * @buf: destination buffer
+ * @offset: offset in destination
+ * @src: source buffer
+ * @len: length of source
+ * @cpy: already copied
+ */
+
+void _ust_buffers_write(struct ust_buffer *buf, size_t offset,
+       const void *src, size_t len, ssize_t cpy)
+{
+       do {
+               len -= cpy;
+               src += cpy;
+               offset += cpy;
+               WARN_ON(offset >= buf->buf_size);
+
+               cpy = min_t(size_t, len, buf->buf_size - offset);
+               ust_buffers_do_copy(buf->buf_data + offset, src, cpy);
+       } while (unlikely(len != cpy));
+}
+
+/**
+ * ltt_buffers_offset_address - get address of a location within the buffer
+ * @buf : buffer
+ * @offset : offset within the buffer.
+ *
+ * Return the address where a given offset is located.
+ * Should be used to get the current subbuffer header pointer. Given we know
+ * it's never on a page boundary, it's safe to write directly to this address,
+ * as long as the write is never bigger than a page size.
+ */
+void *ltt_buffers_offset_address(struct ust_buffer *buf, size_t offset)
+{
+       return ((char *)buf->buf_data)+offset;
+}
+
+/*
+ * -------
+ */
+
+/*
+ * 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.
+ */
+
+/* FIXME: does this test work properly? */
+#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 ust_buffer *ltt_buf,
+                                       u64 tsc)
+{
+       ltt_buf->last_tsc = (unsigned long)tsc;
+}
+
+static inline int last_tsc_overflow(struct ust_buffer *ltt_buf,
+                                       u64 tsc)
+{
+       if (unlikely((tsc - ltt_buf->last_tsc) >> LTT_TSC_BITS))
+               return 1;
+       else
+               return 0;
+}
+#endif
+
+/*
+ * 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 void ust_buffers_destroy_buffer(struct ust_channel *ltt_chan);
+
+static void ltt_force_switch(struct ust_buffer *buf,
+               enum force_switch_mode mode);
+
+/*
+ * Trace callbacks
+ */
+static void ltt_buffer_begin_callback(struct ust_buffer *buf,
+                       u64 tsc, unsigned int subbuf_idx)
+{
+       struct ust_channel *channel = buf->chan;
+       struct ltt_subbuffer_header *header =
+               (struct ltt_subbuffer_header *)
+                       ltt_buffers_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 ust_buffer *buf,
+               u64 tsc, unsigned int offset, unsigned int subbuf_idx)
+{
+       struct ltt_subbuffer_header *header =
+               (struct ltt_subbuffer_header *)
+                       ltt_buffers_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(&buf->events_lost);
+       header->subbuf_corrupt = local_read(&buf->corrupted_subbuffers);
+
+}
+
+void (*wake_consumer)(void *, int) = NULL;
+
+void relay_set_wake_consumer(void (*wake)(void *, int))
+{
+       wake_consumer = wake;
+}
+
+void relay_wake_consumer(void *arg, int finished)
+{
+       if(wake_consumer)
+               wake_consumer(arg, finished);
+}
+
+static notrace void ltt_deliver(struct ust_buffer *buf, unsigned int subbuf_idx,
+               long commit_count)
+{
+       int result;
+
+//ust// #ifdef CONFIG_LTT_VMCORE
+       local_set(&buf->commit_seq[subbuf_idx], commit_count);
+//ust// #endif
+
+       /* wakeup consumer */
+       result = write(buf->data_ready_fd_write, "1", 1);
+       if(result == -1) {
+               PERROR("write (in ltt_relay_buffer_flush)");
+               ERR("this should never happen!");
+       }
+//ust//        atomic_set(&ltt_buf->wakeup_readers, 1);
+}
+
+/*
+ * This function should not be called from NMI interrupt context
+ */
+static notrace void ltt_buf_unfull(struct ust_buffer *buf,
+               unsigned int subbuf_idx,
+               long offset)
+{
+//ust//        struct ltt_channel_struct *ltt_channel =
+//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
+//ust//        struct ltt_channel_buf_struct *ltt_buf = ltt_channel->buf;
+//ust//
+//ust//        ltt_relay_wake_writers(ltt_buf);
+}
+
+int ust_buffers_do_get_subbuf(struct ust_buffer *buf, long *pconsumed_old)
+{
+       struct ust_channel *channel = buf->chan;
+       long consumed_old, consumed_idx, commit_count, write_offset;
+       consumed_old = atomic_long_read(&buf->consumed);
+       consumed_idx = SUBBUF_INDEX(consumed_old, buf->chan);
+       commit_count = local_read(&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(&buf->offset);
+       /*
+        * Check that the subbuffer we are trying to consume has been
+        * already fully committed.
+        */
+       if (((commit_count - buf->chan->subbuf_size)
+            & channel->commit_count_mask)
+           - (BUFFER_TRUNC(consumed_old, buf->chan)
+              >> 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;
+}
+
+int ust_buffers_do_put_subbuf(struct ust_buffer *buf, u32 uconsumed_old)
+{
+       long consumed_new, consumed_old;
+
+       consumed_old = atomic_long_read(&buf->consumed);
+       consumed_old = consumed_old & (~0xFFFFFFFFL);
+       consumed_old = consumed_old | uconsumed_old;
+       consumed_new = SUBBUF_ALIGN(consumed_old, buf->chan);
+
+//ust//        spin_lock(&ltt_buf->full_lock);
+       if (atomic_long_cmpxchg(&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. */
+//ust//                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);
+//ust//                spin_unlock(&ltt_buf->full_lock);
+       }
+       return 0;
+}
+
+static void ltt_relay_print_subbuffer_errors(
+               struct ust_channel *channel,
+               long cons_off)
+{
+       struct ust_buffer *ltt_buf = channel->buf;
+       long cons_idx, commit_count, write_offset;
+
+       cons_idx = SUBBUF_INDEX(cons_off, channel);
+       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);
+       WARN( "LTT : unread channel %s offset is %ld "
+               "and cons_off : %ld\n",
+               channel->channel_name, write_offset, cons_off);
+       /* Check each sub-buffer for non filled commit count */
+       if (((commit_count - channel->subbuf_size) & channel->commit_count_mask)
+           - (BUFFER_TRUNC(cons_off, channel) >> channel->n_subbufs_order) != 0) {
+               ERR("LTT : %s : subbuffer %lu has non filled "
+                       "commit count %lu.\n",
+                       channel->channel_name, cons_idx, commit_count);
+       }
+       ERR("LTT : %s : commit count : %lu, subbuf size %zd\n",
+                       channel->channel_name, commit_count,
+                       channel->subbuf_size);
+}
+
+static void ltt_relay_print_errors(struct ltt_trace_struct *trace,
+               struct ust_channel *channel)
+{
+       struct ust_buffer *ltt_buf = channel->buf;
+       long cons_off;
+
+       for (cons_off = atomic_long_read(&ltt_buf->consumed);
+                       (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
+                                     channel)
+                        - cons_off) > 0;
+                       cons_off = SUBBUF_ALIGN(cons_off, channel))
+               ltt_relay_print_subbuffer_errors(channel, cons_off);
+}
+
+static void ltt_relay_print_buffer_errors(struct ust_channel *channel)
+{
+       struct ltt_trace_struct *trace = channel->trace;
+       struct ust_buffer *ltt_buf = channel->buf;
+
+       if (local_read(&ltt_buf->events_lost))
+               printk(KERN_ALERT
+                       "LTT : %s : %ld events lost "
+                       "in %s channel.\n",
+                       channel->channel_name,
+                       local_read(&ltt_buf->events_lost),
+                       channel->channel_name);
+       if (local_read(&ltt_buf->corrupted_subbuffers))
+               printk(KERN_ALERT
+                       "LTT : %s : %ld corrupted subbuffers "
+                       "in %s channel.\n",
+                       channel->channel_name,
+                       local_read(&ltt_buf->corrupted_subbuffers),
+                       channel->channel_name);
+
+       ltt_relay_print_errors(trace, channel);
+}
+
+static void ltt_relay_release_channel(struct kref *kref)
+{
+       struct ust_channel *ltt_chan = container_of(kref,
+                       struct ust_channel, kref);
+       free(ltt_chan->buf);
+}
+
+/*
+ * Create ltt buffer.
+ */
+//ust// static int ltt_relay_create_buffer(struct ltt_trace_struct *trace,
+//ust//                struct ltt_channel_struct *ltt_chan, struct rchan_buf *buf,
+//ust//                unsigned int cpu, unsigned int n_subbufs)
+//ust// {
+//ust//        struct ltt_channel_buf_struct *ltt_buf =
+//ust//                percpu_ptr(ltt_chan->buf, cpu);
+//ust//        unsigned int j;
+//ust// 
+//ust//        ltt_buf->commit_count =
+//ust//                kzalloc_node(sizeof(ltt_buf->commit_count) * n_subbufs,
+//ust//                        GFP_KERNEL, cpu_to_node(cpu));
+//ust//        if (!ltt_buf->commit_count)
+//ust//                return -ENOMEM;
+//ust//        kref_get(&trace->kref);
+//ust//        kref_get(&trace->ltt_transport_kref);
+//ust//        kref_get(&ltt_chan->kref);
+//ust//        local_set(&ltt_buf->offset, ltt_subbuffer_header_size());
+//ust//        atomic_long_set(&ltt_buf->consumed, 0);
+//ust//        atomic_long_set(&ltt_buf->active_readers, 0);
+//ust//        for (j = 0; j < n_subbufs; j++)
+//ust//                local_set(&ltt_buf->commit_count[j], 0);
+//ust//        init_waitqueue_head(&ltt_buf->write_wait);
+//ust//        atomic_set(&ltt_buf->wakeup_readers, 0);
+//ust//        spin_lock_init(&ltt_buf->full_lock);
+//ust// 
+//ust//        ltt_buffer_begin_callback(buf, trace->start_tsc, 0);
+//ust//        /* atomic_add made on local variable on data that belongs to
+//ust//         * various CPUs : ok because tracing not started (for this cpu). */
+//ust//        local_add(ltt_subbuffer_header_size(), &ltt_buf->commit_count[0]);
+//ust// 
+//ust//        local_set(&ltt_buf->events_lost, 0);
+//ust//        local_set(&ltt_buf->corrupted_subbuffers, 0);
+//ust// 
+//ust//        return 0;
+//ust// }
+
+static int ust_buffers_init_buffer(struct ltt_trace_struct *trace,
+               struct ust_channel *ltt_chan, struct ust_buffer *buf,
+               unsigned int n_subbufs)
+{
+       unsigned int j;
+       int fds[2];
+       int result;
+
+       buf->commit_count =
+               zmalloc(sizeof(buf->commit_count) * n_subbufs);
+       if (!buf->commit_count)
+               return -ENOMEM;
+       kref_get(&trace->kref);
+       kref_get(&trace->ltt_transport_kref);
+       kref_get(&ltt_chan->kref);
+       local_set(&buf->offset, ltt_subbuffer_header_size());
+       atomic_long_set(&buf->consumed, 0);
+       atomic_long_set(&buf->active_readers, 0);
+       for (j = 0; j < n_subbufs; j++)
+               local_set(&buf->commit_count[j], 0);
+//ust//        init_waitqueue_head(&buf->write_wait);
+//ust//        atomic_set(&buf->wakeup_readers, 0);
+//ust//        spin_lock_init(&buf->full_lock);
+
+       ltt_buffer_begin_callback(buf, trace->start_tsc, 0);
+
+       local_add(ltt_subbuffer_header_size(), &buf->commit_count[0]);
+
+       local_set(&buf->events_lost, 0);
+       local_set(&buf->corrupted_subbuffers, 0);
+
+       result = pipe(fds);
+       if(result == -1) {
+               PERROR("pipe");
+               return -1;
+       }
+       buf->data_ready_fd_read = fds[0];
+       buf->data_ready_fd_write = fds[1];
+
+       /* FIXME: do we actually need this? */
+       result = fcntl(fds[0], F_SETFL, O_NONBLOCK);
+       if(result == -1) {
+               PERROR("fcntl");
+       }
+
+//ust//        buf->commit_seq = malloc(sizeof(buf->commit_seq) * n_subbufs);
+//ust//        if(!ltt_buf->commit_seq) {
+//ust//                return -1;
+//ust//        }
+
+       /* FIXME: decrementally destroy on error */
+
+       return 0;
+}
+
+/* FIXME: use this function */
+static void ust_buffers_destroy_buffer(struct ust_channel *ltt_chan)
+{
+       struct ltt_trace_struct *trace = ltt_chan->trace;
+       struct ust_buffer *ltt_buf = ltt_chan->buf;
+
+       kref_put(&ltt_chan->trace->ltt_transport_kref,
+               ltt_release_transport);
+       ltt_relay_print_buffer_errors(ltt_chan);
+//ust//        free(ltt_buf->commit_seq);
+       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);
+//ust//        wake_up_interruptible(&trace->kref_wq);
+}
+
+static void ltt_chan_alloc_ltt_buf(struct ust_channel *chan)
+{
+       void *ptr;
+       int result;
+
+       /* Get one page */
+       /* FIXME: increase size if we have a seq_commit array that overflows the page */
+       size_t size = PAGE_ALIGN(1);
+
+       result = chan->buf_shmid = shmget(getpid(), size, IPC_CREAT | IPC_EXCL | 0700);
+       if(chan->buf_shmid == -1) {
+               PERROR("shmget");
+               return;
+       }
+
+       ptr = shmat(chan->buf_shmid, NULL, 0);
+       if(ptr == (void *) -1) {
+               perror("shmat");
+               goto destroy_shmem;
+       }
+
+       /* Already mark the shared memory for destruction. This will occur only
+         * when all users have detached.
+        */
+       result = shmctl(chan->buf_shmid, IPC_RMID, NULL);
+       if(result == -1) {
+               perror("shmctl");
+               return;
+       }
+
+       chan->buf = ptr;
+
+       return;
+
+       destroy_shmem:
+       result = shmctl(chan->buf_shmid, IPC_RMID, NULL);
+       if(result == -1) {
+               perror("shmctl");
+       }
+
+       return;
+}
+
+/*
+ * Create channel.
+ */
+static int ust_buffers_create_channel(const char *trace_name, struct ltt_trace_struct *trace,
+       const char *channel_name, struct ust_channel *ltt_chan,
+       unsigned int subbuf_size, unsigned int n_subbufs, int overwrite)
+{
+       int err = 0;
+       int result;
+
+       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);
+//ust//        ltt_chan->buf = percpu_alloc_mask(sizeof(struct ltt_channel_buf_struct), GFP_KERNEL, cpu_possible_map);
+
+       ltt_chan_alloc_ltt_buf(ltt_chan);
+
+//ust//        ltt_chan->buf = malloc(sizeof(struct ltt_channel_buf_struct));
+       if (!ltt_chan->buf)
+               goto alloc_error;
+       /* FIXME: handle error of this call */
+       result = ust_buffers_channel_open(ltt_chan, subbuf_size, n_subbufs);
+       if (result == -1) {
+               printk(KERN_ERR "LTT : Can't open channel for trace %s\n",
+                               trace_name);
+               goto relay_open_error;
+       }
+
+       err = 0;
+       goto end;
+
+relay_open_error:
+//ust//        percpu_free(ltt_chan->buf);
+alloc_error:
+       err = EPERM;
+end:
+       return err;
+}
+
+/*
+ * 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 ust_buffer *buf)
+{
+       int result;
+
+//ust//        buf->finalized = 1;
+       ltt_force_switch(buf, FORCE_FLUSH);
+
+       result = write(buf->data_ready_fd_write, "1", 1);
+       if(result == -1) {
+               PERROR("write (in ltt_relay_buffer_flush)");
+               ERR("this should never happen!");
+       }
+}
+
+static void ltt_relay_async_wakeup_chan(struct ust_channel *ltt_channel)
+{
+//ust//        unsigned int i;
+//ust//        struct rchan *rchan = ltt_channel->trans_channel_data;
+//ust//
+//ust//        for_each_possible_cpu(i) {
+//ust//                struct ltt_channel_buf_struct *ltt_buf =
+//ust//                        percpu_ptr(ltt_channel->buf, i);
+//ust//
+//ust//                if (atomic_read(&ltt_buf->wakeup_readers) == 1) {
+//ust//                        atomic_set(&ltt_buf->wakeup_readers, 0);
+//ust//                        wake_up_interruptible(&rchan->buf[i]->read_wait);
+//ust//                }
+//ust//        }
+}
+
+static void ltt_relay_finish_buffer(struct ust_channel *channel)
+{
+//     int result;
+
+       if (channel->buf) {
+               struct ust_buffer *buf = channel->buf;
+               ltt_relay_buffer_flush(buf);
+//ust//                ltt_relay_wake_writers(ltt_buf);
+               /* closing the pipe tells the consumer the buffer is finished */
+               
+               //result = write(ltt_buf->data_ready_fd_write, "D", 1);
+               //if(result == -1) {
+               //      PERROR("write (in ltt_relay_finish_buffer)");
+               //      ERR("this should never happen!");
+               //}
+               close(buf->data_ready_fd_write);
+       }
+}
+
+
+static void ltt_relay_finish_channel(struct ust_channel *channel)
+{
+//ust//        unsigned int i;
+
+//ust//        for_each_possible_cpu(i)
+               ltt_relay_finish_buffer(channel);
+}
+
+static void ltt_relay_remove_channel(struct ust_channel *channel)
+{
+       ust_buffers_channel_close(channel);
+       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 ust_channel *channel, struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets, size_t data_size,
+               u64 *tsc, unsigned int *rflags, int largest_align)
+{
+       offsets->begin = local_read(&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(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 = ust_get_header_size(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)
+                        >> channel->n_subbufs_order)
+                       - (local_read(&buf->commit_count[subbuf_index])
+                               & channel->commit_count_mask);
+               if (offsets->reserve_commit_diff == 0) {
+                       long consumed;
+
+                       consumed = atomic_long_read(&buf->consumed);
+
+                       /* Next buffer not corrupted. */
+                       if (!channel->overwrite &&
+                               (SUBBUF_TRUNC(offsets->begin, buf->chan)
+                                - SUBBUF_TRUNC(consumed, buf->chan))
+                               >= channel->alloc_size) {
+
+                               long consumed_idx = SUBBUF_INDEX(consumed, buf->chan);
+                               long commit_count = local_read(&buf->commit_count[consumed_idx]);
+                               if(((commit_count - buf->chan->subbuf_size) & channel->commit_count_mask) - (BUFFER_TRUNC(consumed, buf->chan) >> channel->n_subbufs_order) != 0) {
+                                       WARN("Event dropped. Caused by non-committed event.");
+                               }
+                               else {
+                                       WARN("Event dropped. Caused by non-consumed buffer.");
+                               }
+                               /*
+                                * We do not overwrite non consumed buffers
+                                * and we are full : event is lost.
+                                */
+                               local_inc(&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 = ust_get_header_size(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(&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 ust_channel *channel,
+               struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets,
+               u64 *tsc)
+{
+       long subbuf_index;
+
+       offsets->begin = local_read(&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)
+                >> channel->n_subbufs_order)
+               - (local_read(&buf->commit_count[subbuf_index])
+                       & channel->commit_count_mask);
+       if (offsets->reserve_commit_diff == 0) {
+               /* Next buffer not corrupted. */
+               if (mode == FORCE_ACTIVE
+                   && !channel->overwrite
+                   && offsets->begin - atomic_long_read(&buf->consumed)
+                      >= channel->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 ust_channel *channel,
+               struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets)
+{
+       long consumed_old, consumed_new;
+
+       do {
+               consumed_old = atomic_long_read(&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))
+                  >= channel->alloc_size)
+                       consumed_new = SUBBUF_ALIGN(consumed_old, buf->chan);
+               else {
+                       consumed_new = consumed_old;
+                       break;
+               }
+       } while (atomic_long_cmpxchg(&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,
+                                 &buf->commit_count[
+                                       SUBBUF_INDEX(offsets->begin,
+                                                    buf->chan)]);
+                       if (!channel->overwrite
+                           || offsets->reserve_commit_diff
+                              != channel->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(&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 ust_channel *channel,
+               struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
+{
+       long oldidx = SUBBUF_INDEX(offsets->old - 1, channel);
+
+       channel->buffer_end(buf, *tsc, offsets->old, oldidx);
+       /* Must write buffer end before incrementing commit count */
+       smp_wmb();
+       offsets->commit_count =
+               local_add_return(channel->subbuf_size
+                                - (SUBBUF_OFFSET(offsets->old - 1, channel)
+                                + 1),
+                                &buf->commit_count[oldidx]);
+       if ((BUFFER_TRUNC(offsets->old - 1, channel)
+                       >> channel->n_subbufs_order)
+                       - ((offsets->commit_count - channel->subbuf_size)
+                               & channel->commit_count_mask) == 0)
+               ltt_deliver(buf, oldidx, offsets->commit_count);
+}
+
+/*
+ * 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 ust_channel *channel,
+               struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
+{
+       long beginidx = SUBBUF_INDEX(offsets->begin, channel);
+
+       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(),
+                       &buf->commit_count[beginidx]);
+       /* Check if the written buffer has to be delivered */
+       if ((BUFFER_TRUNC(offsets->begin, channel)
+                       >> channel->n_subbufs_order)
+                       - ((offsets->commit_count - channel->subbuf_size)
+                               & channel->commit_count_mask) == 0)
+               ltt_deliver(buf, beginidx, offsets->commit_count);
+}
+
+
+/*
+ * 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 ust_channel *channel,
+               struct ust_buffer *buf,
+               struct ltt_reserve_switch_offsets *offsets, u64 *tsc)
+{
+       long endidx = SUBBUF_INDEX(offsets->end - 1, channel);
+
+       channel->buffer_end(buf, *tsc, offsets->end, endidx);
+       /* Must write buffer begin before incrementing commit count */
+       smp_wmb();
+       offsets->commit_count =
+               local_add_return(channel->subbuf_size
+                                - (SUBBUF_OFFSET(offsets->end - 1, channel)
+                                + 1),
+                                &buf->commit_count[endidx]);
+       if ((BUFFER_TRUNC(offsets->end - 1, channel)
+                       >> channel->n_subbufs_order)
+                       - ((offsets->commit_count - channel->subbuf_size)
+                               & channel->commit_count_mask) == 0)
+               ltt_deliver(buf, endidx, offsets->commit_count);
+}
+
+/**
+ * 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 ust_channel *channel, void **transport_data,
+               size_t data_size, size_t *slot_size, long *buf_offset, u64 *tsc,
+               unsigned int *rflags, int largest_align)
+{
+       struct ust_buffer *buf = *transport_data = channel->buf;
+       struct ltt_reserve_switch_offsets offsets;
+
+       offsets.reserve_commit_diff = 0;
+       offsets.size = 0;
+
+       /*
+        * Perform retryable operations.
+        */
+       if (ltt_nesting > 4) {
+               local_inc(&buf->events_lost);
+               return -EPERM;
+       }
+       do {
+               if (ltt_relay_try_reserve(channel, buf, &offsets, data_size, tsc, rflags,
+                               largest_align))
+                       return -ENOSPC;
+       } while (local_cmpxchg(&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(buf, *tsc);
+
+       /*
+        * Push the reader if necessary
+        */
+       ltt_reserve_push_reader(channel, buf, &offsets);
+
+       /*
+        * Switch old subbuffer if needed.
+        */
+       if (offsets.end_switch_old)
+               ltt_reserve_switch_old_subbuf(channel, buf, &offsets, tsc);
+
+       /*
+        * Populate new subbuffer.
+        */
+       if (offsets.begin_switch)
+               ltt_reserve_switch_new_subbuf(channel, buf, &offsets, tsc);
+
+       if (offsets.end_switch_current)
+               ltt_reserve_end_switch_current(channel, 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 ust_buffer *buf,
+               enum force_switch_mode mode)
+{
+       struct ust_channel *channel = buf->chan;
+       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, channel, buf, &offsets, &tsc))
+                       return;
+       } while (local_cmpxchg(&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(buf, tsc);
+
+       /*
+        * Push the reader if necessary
+        */
+       if (mode == FORCE_ACTIVE)
+               ltt_reserve_push_reader(channel, buf, &offsets);
+
+       /*
+        * Switch old subbuffer if needed.
+        */
+       if (offsets.end_switch_old)
+               ltt_reserve_switch_old_subbuf(channel, buf, &offsets, &tsc);
+
+       /*
+        * Populate new subbuffer.
+        */
+       if (mode == FORCE_ACTIVE)
+               ltt_reserve_switch_new_subbuf(channel, buf, &offsets, &tsc);
+}
+
+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)
+{
+       struct ust_channel *channel;
+       struct ust_buffer *buf;
+
+       channel = &trace->channels[chan_index];
+       buf = channel->buf;
+
+       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", ltt_nesting);
+       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(&buf->offset);
+       dbg->read = atomic_long_read(&buf->consumed);
+
+       printk(KERN_ERR "LTT cur size %lu.\n",
+               dbg->write + LTT_RESERVE_CRITICAL + data_size
+               - SUBBUF_TRUNC(dbg->read, channel));
+       printk(KERN_ERR "cur write : %lu, read : %lu\n",
+                       dbg->write, dbg->read);
+}
+
+static struct ltt_transport ust_relay_transport = {
+       .name = "ustrelay",
+       .ops = {
+               .create_channel = ust_buffers_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_errors = ltt_relay_print_user_errors,
+       },
+};
+
+/*
+ * 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.
+ */
+static /* inline */ void ltt_write_commit_counter(struct ust_buffer *buf,
+               struct ust_buffer *ltt_buf,
+               long idx, long buf_offset, long commit_count, size_t data_size)
+{
+       long offset;
+       long commit_seq_old;
+
+       offset = buf_offset + data_size;
+
+       /*
+        * SUBBUF_OFFSET includes commit_count_mask. We can simply
+        * compare the offsets within the subbuffer without caring about
+        * buffer full/empty mismatch because offset is never zero here
+        * (subbuffer header and event headers have non-zero length).
+        */
+       if (unlikely(SUBBUF_OFFSET(offset - commit_count, buf->chan)))
+               return;
+
+       commit_seq_old = local_read(&ltt_buf->commit_seq[idx]);
+       while (commit_seq_old < commit_count)
+               commit_seq_old = local_cmpxchg(&ltt_buf->commit_seq[idx],
+                                        commit_seq_old, commit_count);
+}
+
+/*
+ * 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.
+ * @data_size : size of the event data.
+ * @slot_size : size of the reserved slot.
+ */
+/* FIXME: make this function static inline in the .h! */
+/*static*/ /* inline */ notrace void ltt_commit_slot(
+               struct ust_channel *channel,
+               void **transport_data, long buf_offset,
+               size_t data_size, size_t slot_size)
+{
+       struct ust_buffer *buf = *transport_data;
+       long offset_end = buf_offset;
+       long endidx = SUBBUF_INDEX(offset_end - 1, channel);
+       long commit_count;
+
+       /* Must write slot data before incrementing commit count */
+       smp_wmb();
+       commit_count = local_add_return(slot_size,
+               &buf->commit_count[endidx]);
+       /* Check if all commits have been done */
+       if ((BUFFER_TRUNC(offset_end - 1, channel)
+                       >> channel->n_subbufs_order)
+                       - ((commit_count - channel->subbuf_size)
+                          & channel->commit_count_mask) == 0)
+               ltt_deliver(buf, endidx, commit_count);
+       /*
+        * Update lost_size for each commit. It's needed only for extracting
+        * ltt buffers from vmcore, after crash.
+        */
+       ltt_write_commit_counter(buf, buf, endidx,
+                                buf_offset, commit_count, data_size);
+}
+
+
+static char initialized = 0;
+
+void __attribute__((constructor)) init_ustrelay_transport(void)
+{
+       if(!initialized) {
+               ltt_transport_register(&ust_relay_transport);
+               initialized = 1;
+       }
+}
+
+static void __attribute__((destructor)) ltt_relay_exit(void)
+{
+       ltt_transport_unregister(&ust_relay_transport);
+}
diff --git a/libust/buffers.h b/libust/buffers.h
new file mode 100644 (file)
index 0000000..247c7b0
--- /dev/null
@@ -0,0 +1,146 @@
+/*
+ * buffers.h
+ *
+ * Copyright (C) 2009 - Pierre-Marc Fournier (pierre-marc dot fournier at polymtl dot ca)
+ * Copyright (C) 2008 - Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
+ *
+ */
+
+#ifndef _UST_BUFFERS_H
+#define _UST_BUFFERS_H
+
+#include <kcompat/kref.h>
+#include <assert.h>
+#include "channels.h"
+#include "buffers.h"
+
+/* Return the size of the minimum number of pages that can contain x. */
+#define FIX_SIZE(x) ((((x) - 1) & PAGE_MASK) + PAGE_SIZE)
+
+/*
+ * BUFFER_TRUNC zeroes the subbuffer offset and the subbuffer number parts of
+ * the offset, which leaves only the buffer number.
+ */
+#define BUFFER_TRUNC(offset, chan) \
+       ((offset) & (~((chan)->alloc_size-1)))
+#define BUFFER_OFFSET(offset, chan) ((offset) & ((chan)->alloc_size - 1))
+#define SUBBUF_OFFSET(offset, chan) ((offset) & ((chan)->subbuf_size - 1))
+#define SUBBUF_ALIGN(offset, chan) \
+       (((offset) + (chan)->subbuf_size) & (~((chan)->subbuf_size - 1)))
+#define SUBBUF_TRUNC(offset, chan) \
+       ((offset) & (~((chan)->subbuf_size - 1)))
+#define SUBBUF_INDEX(offset, chan) \
+       (BUFFER_OFFSET((offset), chan) >> (chan)->subbuf_size_order)
+
+/*
+ * Tracks changes to rchan/rchan_buf structs
+ */
+#define UST_CHANNEL_VERSION            8
+
+struct ust_buffer {
+       /* 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;
+       /* one byte is written to this pipe when data is available, in order
+           to wake the consumer */
+       /* portability: Single byte writes must be as quick as possible. The kernel-side
+          buffer must be large enough so the writer doesn't block. From the pipe(7)
+           man page: Since linux 2.6.11, the pipe capacity is 65536 bytes. */
+       int data_ready_fd_write;
+       /* the reading end of the pipe */
+       int data_ready_fd_read;
+
+       struct ust_channel *chan;
+       struct kref kref;
+       void *buf_data;
+       size_t buf_size;
+       int shmid;
+
+       /* commit count per subbuffer; must be at end of struct */
+       local_t commit_seq[0] ____cacheline_aligned;
+} ____cacheline_aligned;
+
+extern void _ust_buffers_write(struct ust_buffer *buf, size_t offset,
+       const void *src, size_t len, ssize_t cpy);
+
+/*
+ * Return the address where a given offset is located.
+ * Should be used to get the current subbuffer header pointer. Given we know
+ * it's never on a page boundary, it's safe to write directly to this address,
+ * as long as the write is never bigger than a page size.
+ */
+extern void *ust_buffers_offset_address(struct ust_buffer *buf,
+       size_t offset);
+
+/* FIXME: lttng has a version for systems with inefficient unaligned access */
+static inline void ust_buffers_do_copy(void *dest, const void *src, size_t len)
+{
+       union {
+               const void *src;
+               const u8 *src8;
+               const u16 *src16;
+               const u32 *src32;
+               const u64 *src64;
+       } u = { .src = src };
+
+       switch (len) {
+       case 0: break;
+       case 1: *(u8 *)dest = *u.src8;
+               break;
+       case 2: *(u16 *)dest = *u.src16;
+               break;
+       case 4: *(u32 *)dest = *u.src32;
+               break;
+       case 8: *(u64 *)dest = *u.src64;
+               break;
+       default:
+               memcpy(dest, src, len);
+       }
+}
+
+/* FIXME: there is both a static inline and a '_' non static inline version ?? */
+static inline int ust_buffers_write(struct ust_buffer *buf, size_t offset,
+       const void *src, size_t len)
+{
+       size_t cpy;
+       size_t buf_offset = BUFFER_OFFSET(offset, buf->chan);
+
+       assert(buf_offset < buf->chan->subbuf_size*buf->chan->subbuf_cnt);
+
+       cpy = min_t(size_t, len, buf->buf_size - buf_offset);
+       ust_buffers_do_copy(buf->buf_data + buf_offset, src, cpy);
+       
+       if (unlikely(len != cpy))
+               _ust_buffers_write(buf, buf_offset, src, len, cpy);
+       return len;
+}
+
+int ust_buffers_channel_open(struct ust_channel *chan, size_t subbuf_size, size_t n_subbufs);
+extern void ust_buffers_channel_close(struct ust_channel *chan);
+
+extern int ust_buffers_do_get_subbuf(struct ust_buffer *buf, long *pconsumed_old);
+
+extern int ust_buffers_do_put_subbuf(struct ust_buffer *buf, u32 uconsumed_old);
+
+extern void init_ustrelay_transport(void);
+
+/*static*/ /* inline */ notrace void ltt_commit_slot(
+               struct ust_channel *channel,
+               void **transport_data, long buf_offset,
+               size_t data_size, size_t slot_size);
+
+#endif /* _UST_BUFFERS_H */
index 27ceefc6d6a1cd7921ebb09f91c6d585615348e3..c90733cbbc863061684adb6495bfec232736d8e7 100644 (file)
@@ -247,11 +247,11 @@ int ltt_channels_get_index_from_name(const char *name)
  * Called with trace lock held. Does not perform the trace buffer allocation,
  * because we must let the user overwrite specific channel sizes.
  */
-struct ltt_channel_struct *ltt_channels_trace_alloc(unsigned int *nr_channels,
+struct ust_channel *ltt_channels_trace_alloc(unsigned int *nr_channels,
                                                    int overwrite,
                                                    int active)
 {
-       struct ltt_channel_struct *channel = NULL;
+       struct ust_channel *channel = NULL;
        struct ltt_channel_setting *iter;
 
        mutex_lock(&ltt_channel_mutex);
@@ -264,7 +264,7 @@ struct ltt_channel_struct *ltt_channels_trace_alloc(unsigned int *nr_channels,
        else
                kref_get(&index_kref);
        *nr_channels = free_index;
-       channel = kzalloc(sizeof(struct ltt_channel_struct) * free_index,
+       channel = kzalloc(sizeof(struct ust_channel) * free_index,
                          GFP_KERNEL);
        if (!channel) {
                WARN("ltt_channel_struct: channel null after alloc");
@@ -292,7 +292,7 @@ end:
  * Called with trace lock held. The actual channel buffers must be freed before
  * this function is called.
  */
-void ltt_channels_trace_free(struct ltt_channel_struct *channels)
+void ltt_channels_trace_free(struct ust_channel *channels)
 {
        lock_markers();
        mutex_lock(&ltt_channel_mutex);
index e460e12b916f0faa9ad58634ae180fa791db2a2e..0b179e5cc670fe1b75f765a66d0b5f253d41e709 100644 (file)
@@ -22,8 +22,6 @@
  */
 
 #include <linux/limits.h>
-//ust// #include <linux/kref.h>
-//ust// #include <linux/list.h>
 #include <errno.h>
 
 #include <ust/kernelcompat.h>
 #define EVENTS_PER_CHANNEL     65536
 
 struct ltt_trace_struct;
-struct rchan_buf;
 
-struct ltt_channel_struct {
+struct ust_buffer;
+
+struct ust_channel {
        /* First 32 bytes cache-hot cacheline */
        struct ltt_trace_struct *trace;
        void *buf;
-       void *trans_channel_data;
        int overwrite:1;
        int active:1;
        unsigned int n_subbufs_order;
@@ -54,20 +52,25 @@ struct ltt_channel_struct {
         * buffer_begin - called on buffer-switch to a new sub-buffer
         * @buf: the channel buffer containing the new sub-buffer
         */
-       void (*buffer_begin) (struct rchan_buf *buf,
+       void (*buffer_begin) (struct ust_buffer *buf,
                        u64 tsc, unsigned int subbuf_idx);
        /*
         * buffer_end - called on buffer-switch to a new sub-buffer
         * @buf: the channel buffer containing the previous sub-buffer
         */
-       void (*buffer_end) (struct rchan_buf *buf,
+       void (*buffer_end) (struct ust_buffer *buf,
                        u64 tsc, unsigned int offset, unsigned int subbuf_idx);
        struct kref kref;       /* Channel transport reference count */
-       unsigned int subbuf_size;
+       size_t subbuf_size;
+       int subbuf_size_order;
        unsigned int subbuf_cnt;
        const char *channel_name;
 
        int buf_shmid;
+
+       u32 version;
+       size_t alloc_size;
+       struct list_head list;
 } ____cacheline_aligned;
 
 struct ltt_channel_setting {
@@ -87,10 +90,10 @@ extern int ltt_channels_set_default(const char *name,
                             unsigned int subbuf_cnt);
 extern const char *ltt_channels_get_name_from_index(unsigned int index);
 extern int ltt_channels_get_index_from_name(const char *name);
-extern struct ltt_channel_struct *ltt_channels_trace_alloc(unsigned int *nr_channels,
+extern struct ust_channel *ltt_channels_trace_alloc(unsigned int *nr_channels,
                                                    int overwrite,
                                                    int active);
-extern void ltt_channels_trace_free(struct ltt_channel_struct *channels);
+extern void ltt_channels_trace_free(struct ust_channel *channels);
 extern int _ltt_channels_get_event_id(const char *channel, const char *name);
 extern int ltt_channels_get_event_id(const char *channel, const char *name);
 
diff --git a/libust/relay.c b/libust/relay.c
deleted file mode 100644 (file)
index 408ce90..0000000
+++ /dev/null
@@ -1,2516 +0,0 @@
-/*
- * Public API and common code for kernel->userspace relay file support.
- *
- * Copyright (C) 2002-2005 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999-2005 - Karim Yaghmour (karim@opersys.com)
- * Copyright (C) 2008 - Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
- *
- * Moved to kernel/relay.c by Paul Mundt, 2006.
- * November 2006 - CPU hotplug support by Mathieu Desnoyers
- *     (mathieu.desnoyers@polymtl.ca)
- *
- * This file is released under the GPL.
- */
-//ust// #include <linux/errno.h>
-//ust// #include <linux/stddef.h>
-//ust// #include <linux/slab.h>
-//ust// #include <linux/module.h>
-//ust// #include <linux/string.h>
-//ust// #include <linux/ltt-relay.h>
-//ust// #include <linux/vmalloc.h>
-//ust// #include <linux/mm.h>
-//ust// #include <linux/cpu.h>
-//ust// #include <linux/splice.h>
-//ust// #include <linux/bitops.h>
-#include <ust/kernelcompat.h>
-#include <sys/mman.h>
-#include <sys/ipc.h>
-#include <sys/shm.h>
-#include <fcntl.h>
-//#include "list.h"
-#include "relay.h"
-#include "channels.h"
-#include <kcompat/kref.h>
-#include "tracer.h"
-#include "tracercore.h"
-#include "usterr.h"
-
-/* list of open channels, for cpu hotplug */
-static DEFINE_MUTEX(relay_channels_mutex);
-static LIST_HEAD(relay_channels);
-
-
-static struct dentry *ltt_create_buf_file_callback(struct rchan_buf *buf);
-
-/**
- *     relay_alloc_buf - allocate a channel buffer
- *     @buf: the buffer struct
- *     @size: total size of the buffer
- */
-//ust// static int relay_alloc_buf(struct rchan_buf *buf, size_t *size)
-//ust//{
-//ust//        unsigned int i, n_pages;
-//ust//        struct buf_page *buf_page, *n;
-//ust//
-//ust//        *size = PAGE_ALIGN(*size);
-//ust//        n_pages = *size >> PAGE_SHIFT;
-//ust//
-//ust//        INIT_LIST_HEAD(&buf->pages);
-//ust//
-//ust//        for (i = 0; i < n_pages; i++) {
-//ust//                buf_page = kmalloc_node(sizeof(*buf_page), GFP_KERNEL,
-//ust//                        cpu_to_node(buf->cpu));
-//ust//                if (unlikely(!buf_page))
-//ust//                        goto depopulate;
-//ust//                buf_page->page = alloc_pages_node(cpu_to_node(buf->cpu),
-//ust//                        GFP_KERNEL | __GFP_ZERO, 0);
-//ust//                if (unlikely(!buf_page->page)) {
-//ust//                        kfree(buf_page);
-//ust//                        goto depopulate;
-//ust//                }
-//ust//                list_add_tail(&buf_page->list, &buf->pages);
-//ust//                buf_page->offset = (size_t)i << PAGE_SHIFT;
-//ust//                buf_page->buf = buf;
-//ust//                set_page_private(buf_page->page, (unsigned long)buf_page);
-//ust//                if (i == 0) {
-//ust//                        buf->wpage = buf_page;
-//ust//                        buf->hpage[0] = buf_page;
-//ust//                        buf->hpage[1] = buf_page;
-//ust//                        buf->rpage = buf_page;
-//ust//                }
-//ust//        }
-//ust//        buf->page_count = n_pages;
-//ust//        return 0;
-//ust//
-//ust//depopulate:
-//ust//        list_for_each_entry_safe(buf_page, n, &buf->pages, list) {
-//ust//                list_del_init(&buf_page->list);
-//ust//                __free_page(buf_page->page);
-//ust//                kfree(buf_page);
-//ust//        }
-//ust//        return -ENOMEM;
-//ust//}
-
-static int relay_alloc_buf(struct rchan_buf *buf, size_t *size)
-{
-//ust//        unsigned int n_pages;
-//ust//        struct buf_page *buf_page, *n;
-
-       void *ptr;
-       int result;
-
-       *size = PAGE_ALIGN(*size);
-
-       result = buf->shmid = shmget(getpid(), *size, IPC_CREAT | IPC_EXCL | 0700);
-       if(result == -1 && errno == EINVAL) {
-               ERR("shmget() returned EINVAL; maybe /proc/sys/kernel/shmmax should be increased.");
-               return -1;
-       }
-       else if(result == -1) {
-               PERROR("shmget");
-               return -1;
-       }
-
-       ptr = shmat(buf->shmid, NULL, 0);
-       if(ptr == (void *) -1) {
-               perror("shmat");
-               goto destroy_shmem;
-       }
-
-       /* Already mark the shared memory for destruction. This will occur only
-         * when all users have detached.
-        */
-       result = shmctl(buf->shmid, IPC_RMID, NULL);
-       if(result == -1) {
-               perror("shmctl");
-               return -1;
-       }
-
-       buf->buf_data = ptr;
-       buf->buf_size = *size;
-
-       return 0;
-
-       destroy_shmem:
-       result = shmctl(buf->shmid, IPC_RMID, NULL);
-       if(result == -1) {
-               perror("shmctl");
-       }
-
-       return -1;
-}
-
-/**
- *     relay_create_buf - allocate and initialize a channel buffer
- *     @chan: the relay channel
- *     @cpu: cpu the buffer belongs to
- *
- *     Returns channel buffer if successful, %NULL otherwise.
- */
-static struct rchan_buf *relay_create_buf(struct rchan *chan)
-{
-       int ret;
-       struct rchan_buf *buf = kzalloc(sizeof(struct rchan_buf), GFP_KERNEL);
-       if (!buf)
-               return NULL;
-
-//     buf->cpu = cpu;
-       ret = relay_alloc_buf(buf, &chan->alloc_size);
-       if (ret)
-               goto free_buf;
-
-       buf->chan = chan;
-       kref_get(&buf->chan->kref);
-       return buf;
-
-free_buf:
-       kfree(buf);
-       return NULL;
-}
-
-/**
- *     relay_destroy_channel - free the channel struct
- *     @kref: target kernel reference that contains the relay channel
- *
- *     Should only be called from kref_put().
- */
-static void relay_destroy_channel(struct kref *kref)
-{
-       struct rchan *chan = container_of(kref, struct rchan, kref);
-       kfree(chan);
-}
-
-/**
- *     relay_destroy_buf - destroy an rchan_buf struct and associated buffer
- *     @buf: the buffer struct
- */
-static void relay_destroy_buf(struct rchan_buf *buf)
-{
-       struct rchan *chan = buf->chan;
-//ust//        struct buf_page *buf_page;
-       int result;
-
-       result = munmap(buf->buf_data, buf->buf_size);
-       if(result == -1) {
-               PERROR("munmap");
-       }
-
-//ust//        chan->buf[buf->cpu] = NULL;
-       kfree(buf);
-       kref_put(&chan->kref, relay_destroy_channel);
-}
-
-/**
- *     relay_remove_buf - remove a channel buffer
- *     @kref: target kernel reference that contains the relay buffer
- *
- *     Removes the file from the fileystem, which also frees the
- *     rchan_buf_struct and the channel buffer.  Should only be called from
- *     kref_put().
- */
-static void relay_remove_buf(struct kref *kref)
-{
-       struct rchan_buf *buf = container_of(kref, struct rchan_buf, kref);
-//ust//        buf->chan->cb->remove_buf_file(buf);
-       relay_destroy_buf(buf);
-}
-
-/*
- * High-level relay kernel API and associated functions.
- */
-
-/*
- * rchan_callback implementations defining default channel behavior.  Used
- * in place of corresponding NULL values in client callback struct.
- */
-
-/*
- * create_buf_file_create() default callback.  Does nothing.
- */
-//ust// static struct dentry *create_buf_file_default_callback(const char *filename,
-//ust//                                                       struct dentry *parent,
-//ust//                                                       int mode,
-//ust//                                                       struct rchan_buf *buf)
-//ust// {
-//ust//        return NULL;
-//ust// }
-
-//ust// /*
-//ust//  * remove_buf_file() default callback.  Does nothing.
-//ust//  */
-//ust// static int remove_buf_file_default_callback(struct dentry *dentry)
-//ust// {
-//ust//        return -EINVAL;
-//ust// }
-
-/**
- *     wakeup_readers - wake up readers waiting on a channel
- *     @data: contains the channel buffer
- *
- *     This is the timer function used to defer reader waking.
- */
-//ust// static void wakeup_readers(unsigned long data)
-//ust// {
-//ust//        struct rchan_buf *buf = (struct rchan_buf *)data;
-//ust//        wake_up_interruptible(&buf->read_wait);
-//ust// }
-
-/**
- *     __relay_reset - reset a channel buffer
- *     @buf: the channel buffer
- *     @init: 1 if this is a first-time initialization
- *
- *     See relay_reset() for description of effect.
- */
-static void __relay_reset(struct rchan_buf *buf, unsigned int init)
-{
-       if (init) {
-//ust//                init_waitqueue_head(&buf->read_wait);
-               kref_init(&buf->kref);
-//ust//                setup_timer(&buf->timer, wakeup_readers, (unsigned long)buf);
-       } //ust// else
-//ust//                del_timer_sync(&buf->timer);
-
-       buf->finalized = 0;
-}
-
-/*
- *     relay_open_buf - create a new relay channel buffer
- *
- *     used by relay_open() and CPU hotplug.
- */
-static struct rchan_buf *relay_open_buf(struct rchan *chan)
-{
-       struct rchan_buf *buf = NULL;
-//ust//        struct dentry *dentry;
-//ust//        char *tmpname;
-
-//ust//        tmpname = kzalloc(NAME_MAX + 1, GFP_KERNEL);
-//ust//        if (!tmpname)
-//ust//                goto end;
-//ust//        snprintf(tmpname, NAME_MAX, "%s%d", chan->base_filename, cpu);
-
-       buf = relay_create_buf(chan);
-       if (!buf)
-               goto free_name;
-
-       __relay_reset(buf, 1);
-
-       /* Create file in fs */
-//ust//        dentry = chan->cb->create_buf_file(tmpname, chan->parent, S_IRUSR,
-//ust//                                           buf);
-
-       ltt_create_buf_file_callback(buf); // ust //
-
-//ust//        if (!dentry)
-//ust//                goto free_buf;
-//ust//
-//ust//        buf->dentry = dentry;
-
-       goto free_name;
-
-//ust//free_buf:
-       relay_destroy_buf(buf);
-       buf = NULL;
-free_name:
-//ust//        kfree(tmpname);
-//ust//end:
-       return buf;
-}
-
-/**
- *     relay_close_buf - close a channel buffer
- *     @buf: channel buffer
- *
- *     Marks the buffer finalized and restores the default callbacks.
- *     The channel buffer and channel buffer data structure are then freed
- *     automatically when the last reference is given up.
- */
-static void relay_close_buf(struct rchan_buf *buf)
-{
-//ust//        del_timer_sync(&buf->timer);
-       kref_put(&buf->kref, relay_remove_buf);
-}
-
-//ust// static void setup_callbacks(struct rchan *chan,
-//ust//                                   struct rchan_callbacks *cb)
-//ust// {
-//ust//        if (!cb) {
-//ust//                chan->cb = &default_channel_callbacks;
-//ust//                return;
-//ust//        }
-//ust// 
-//ust//        if (!cb->create_buf_file)
-//ust//                cb->create_buf_file = create_buf_file_default_callback;
-//ust//        if (!cb->remove_buf_file)
-//ust//                cb->remove_buf_file = remove_buf_file_default_callback;
-//ust//        chan->cb = cb;
-//ust// }
-
-/**
- *     relay_hotcpu_callback - CPU hotplug callback
- *     @nb: notifier block
- *     @action: hotplug action to take
- *     @hcpu: CPU number
- *
- *     Returns the success/failure of the operation. (%NOTIFY_OK, %NOTIFY_BAD)
- */
-//ust// static int __cpuinit relay_hotcpu_callback(struct notifier_block *nb,
-//ust//                                unsigned long action,
-//ust//                                void *hcpu)
-//ust// {
-//ust//        unsigned int hotcpu = (unsigned long)hcpu;
-//ust//        struct rchan *chan;
-//ust// 
-//ust//        switch (action) {
-//ust//        case CPU_UP_PREPARE:
-//ust//        case CPU_UP_PREPARE_FROZEN:
-//ust//                mutex_lock(&relay_channels_mutex);
-//ust//                list_for_each_entry(chan, &relay_channels, list) {
-//ust//                        if (chan->buf[hotcpu])
-//ust//                                continue;
-//ust//                        chan->buf[hotcpu] = relay_open_buf(chan, hotcpu);
-//ust//                        if (!chan->buf[hotcpu]) {
-//ust//                                printk(KERN_ERR
-//ust//                                        "relay_hotcpu_callback: cpu %d buffer "
-//ust//                                        "creation failed\n", hotcpu);
-//ust//                                mutex_unlock(&relay_channels_mutex);
-//ust//                                return NOTIFY_BAD;
-//ust//                        }
-//ust//                }
-//ust//                mutex_unlock(&relay_channels_mutex);
-//ust//                break;
-//ust//        case CPU_DEAD:
-//ust//        case CPU_DEAD_FROZEN:
-//ust//                /* No need to flush the cpu : will be flushed upon
-//ust//                 * final relay_flush() call. */
-//ust//                break;
-//ust//        }
-//ust//        return NOTIFY_OK;
-//ust// }
-
-/**
- *     ltt_relay_open - create a new relay channel
- *     @base_filename: base name of files to create
- *     @parent: dentry of parent directory, %NULL for root directory
- *     @subbuf_size: size of sub-buffers
- *     @n_subbufs: number of sub-buffers
- *     @cb: client callback functions
- *     @private_data: user-defined data
- *
- *     Returns channel pointer if successful, %NULL otherwise.
- *
- *     Creates a channel buffer for each cpu using the sizes and
- *     attributes specified.  The created channel buffer files
- *     will be named base_filename0...base_filenameN-1.  File
- *     permissions will be %S_IRUSR.
- */
-struct rchan *ltt_relay_open(const char *base_filename,
-                        struct dentry *parent,
-                        size_t subbuf_size,
-                        size_t n_subbufs,
-                        void *private_data)
-{
-//ust//        unsigned int i;
-       struct rchan *chan;
-//ust//        if (!base_filename)
-//ust//                return NULL;
-
-       if (!(subbuf_size && n_subbufs))
-               return NULL;
-
-       chan = kzalloc(sizeof(struct rchan), GFP_KERNEL);
-       if (!chan)
-               return NULL;
-
-       chan->version = LTT_RELAY_CHANNEL_VERSION;
-       chan->n_subbufs = n_subbufs;
-       chan->subbuf_size = subbuf_size;
-       chan->subbuf_size_order = get_count_order(subbuf_size);
-       chan->alloc_size = FIX_SIZE(subbuf_size * n_subbufs);
-       chan->parent = parent;
-       chan->private_data = private_data;
-//ust//        strlcpy(chan->base_filename, base_filename, NAME_MAX);
-//ust//        setup_callbacks(chan, cb);
-       kref_init(&chan->kref);
-
-       mutex_lock(&relay_channels_mutex);
-//ust//        for_each_online_cpu(i) {
-               chan->buf = relay_open_buf(chan);
-               if (!chan->buf)
-                       goto error;
-//ust//        }
-       list_add(&chan->list, &relay_channels);
-       mutex_unlock(&relay_channels_mutex);
-
-       return chan;
-
-//ust//free_bufs:
-//ust//        for_each_possible_cpu(i) {
-//ust//                if (!chan->buf[i])
-//ust//                        break;
-//ust//                relay_close_buf(chan->buf[i]);
-//ust//        }
-
-       error:
-       kref_put(&chan->kref, relay_destroy_channel);
-       mutex_unlock(&relay_channels_mutex);
-       return NULL;
-}
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_open);
-
-/**
- *     ltt_relay_close - close the channel
- *     @chan: the channel
- *
- *     Closes all channel buffers and frees the channel.
- */
-void ltt_relay_close(struct rchan *chan)
-{
-//ust//        unsigned int i;
-
-       if (!chan)
-               return;
-
-       mutex_lock(&relay_channels_mutex);
-//ust//        for_each_possible_cpu(i)
-               if (chan->buf)
-                       relay_close_buf(chan->buf);
-
-       list_del(&chan->list);
-       kref_put(&chan->kref, relay_destroy_channel);
-       mutex_unlock(&relay_channels_mutex);
-}
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_close);
-
-/*
- * Start iteration at the previous element. Skip the real list head.
- */
-//ust// struct buf_page *ltt_relay_find_prev_page(struct rchan_buf *buf,
-//ust//        struct buf_page *page, size_t offset, ssize_t diff_offset)
-//ust// {
-//ust//        struct buf_page *iter;
-//ust//        size_t orig_iter_off;
-//ust//        unsigned int i = 0;
-//ust// 
-//ust//        orig_iter_off = page->offset;
-//ust//        list_for_each_entry_reverse(iter, &page->list, list) {
-//ust//                /*
-//ust//                 * Skip the real list head.
-//ust//                 */
-//ust//                if (&iter->list == &buf->pages)
-//ust//                        continue;
-//ust//                i++;
-//ust//                if (offset >= iter->offset
-//ust//                        && offset < iter->offset + PAGE_SIZE) {
-//ust// #ifdef CONFIG_LTT_RELAY_CHECK_RANDOM_ACCESS
-//ust//                        if (i > 1) {
-//ust//                                printk(KERN_WARNING
-//ust//                                        "Backward random access detected in "
-//ust//                                        "ltt_relay. Iterations %u, "
-//ust//                                        "offset %zu, orig iter->off %zu, "
-//ust//                                        "iter->off %zu diff_offset %zd.\n", i,
-//ust//                                        offset, orig_iter_off, iter->offset,
-//ust//                                        diff_offset);
-//ust//                                WARN_ON(1);
-//ust//                        }
-//ust// #endif
-//ust//                        return iter;
-//ust//                }
-//ust//        }
-//ust//        WARN_ON(1);
-//ust//        return NULL;
-//ust// }
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_find_prev_page);
-
-/*
- * Start iteration at the next element. Skip the real list head.
- */
-//ust// struct buf_page *ltt_relay_find_next_page(struct rchan_buf *buf,
-//ust//        struct buf_page *page, size_t offset, ssize_t diff_offset)
-//ust// {
-//ust//        struct buf_page *iter;
-//ust//        unsigned int i = 0;
-//ust//        size_t orig_iter_off;
-//ust// 
-//ust//        orig_iter_off = page->offset;
-//ust//        list_for_each_entry(iter, &page->list, list) {
-//ust//                /*
-//ust//                 * Skip the real list head.
-//ust//                 */
-//ust//                if (&iter->list == &buf->pages)
-//ust//                        continue;
-//ust//                i++;
-//ust//                if (offset >= iter->offset
-//ust//                        && offset < iter->offset + PAGE_SIZE) {
-//ust// #ifdef CONFIG_LTT_RELAY_CHECK_RANDOM_ACCESS
-//ust//                        if (i > 1) {
-//ust//                                printk(KERN_WARNING
-//ust//                                        "Forward random access detected in "
-//ust//                                        "ltt_relay. Iterations %u, "
-//ust//                                        "offset %zu, orig iter->off %zu, "
-//ust//                                        "iter->off %zu diff_offset %zd.\n", i,
-//ust//                                        offset, orig_iter_off, iter->offset,
-//ust//                                        diff_offset);
-//ust//                                WARN_ON(1);
-//ust//                        }
-//ust// #endif
-//ust//                        return iter;
-//ust//                }
-//ust//        }
-//ust//        WARN_ON(1);
-//ust//        return NULL;
-//ust// }
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_find_next_page);
-
-/**
- * ltt_relay_write - write data to a ltt_relay buffer.
- * @buf : buffer
- * @offset : offset within the buffer
- * @src : source address
- * @len : length to write
- * @page : cached buffer page
- * @pagecpy : page size copied so far
- */
-void _ltt_relay_write(struct rchan_buf *buf, size_t offset,
-       const void *src, size_t len, ssize_t cpy)
-{
-       do {
-               len -= cpy;
-               src += cpy;
-               offset += cpy;
-               /*
-                * Underlying layer should never ask for writes across
-                * subbuffers.
-                */
-               WARN_ON(offset >= buf->buf_size);
-
-               cpy = min_t(size_t, len, buf->buf_size - offset);
-               ltt_relay_do_copy(buf->buf_data + offset, src, cpy);
-       } while (unlikely(len != cpy));
-}
-//ust// EXPORT_SYMBOL_GPL(_ltt_relay_write);
-
-/**
- * ltt_relay_read - read data from ltt_relay_buffer.
- * @buf : buffer
- * @offset : offset within the buffer
- * @dest : destination address
- * @len : length to write
- */
-//ust// int ltt_relay_read(struct rchan_buf *buf, size_t offset,
-//ust//        void *dest, size_t len)
-//ust// {
-//ust//        struct buf_page *page;
-//ust//        ssize_t pagecpy, orig_len;
-//ust// 
-//ust//        orig_len = len;
-//ust//        offset &= buf->chan->alloc_size - 1;
-//ust//        page = buf->rpage;
-//ust//        if (unlikely(!len))
-//ust//                return 0;
-//ust//        for (;;) {
-//ust//                page = ltt_relay_cache_page(buf, &buf->rpage, page, offset);
-//ust//                pagecpy = min_t(size_t, len, PAGE_SIZE - (offset & ~PAGE_MASK));
-//ust//                memcpy(dest, page_address(page->page) + (offset & ~PAGE_MASK),
-//ust//                        pagecpy);
-//ust//                len -= pagecpy;
-//ust//                if (likely(!len))
-//ust//                        break;
-//ust//                dest += pagecpy;
-//ust//                offset += pagecpy;
-//ust//                /*
-//ust//                 * Underlying layer should never ask for reads across
-//ust//                 * subbuffers.
-//ust//                 */
-//ust//                WARN_ON(offset >= buf->chan->alloc_size);
-//ust//        }
-//ust//        return orig_len;
-//ust// }
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_read);
-
-/**
- * ltt_relay_read_get_page - Get a whole page to read from
- * @buf : buffer
- * @offset : offset within the buffer
- */
-//ust// struct buf_page *ltt_relay_read_get_page(struct rchan_buf *buf, size_t offset)
-//ust// {
-//ust//        struct buf_page *page;
-
-//ust//        offset &= buf->chan->alloc_size - 1;
-//ust//        page = buf->rpage;
-//ust//        page = ltt_relay_cache_page(buf, &buf->rpage, page, offset);
-//ust//        return page;
-//ust// }
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_read_get_page);
-
-/**
- * ltt_relay_offset_address - get address of a location within the buffer
- * @buf : buffer
- * @offset : offset within the buffer.
- *
- * Return the address where a given offset is located.
- * Should be used to get the current subbuffer header pointer. Given we know
- * it's never on a page boundary, it's safe to write directly to this address,
- * as long as the write is never bigger than a page size.
- */
-void *ltt_relay_offset_address(struct rchan_buf *buf, size_t offset)
-{
-//ust//        struct buf_page *page;
-//ust//        unsigned int odd;
-//ust// 
-//ust//        offset &= buf->chan->alloc_size - 1;
-//ust//        odd = !!(offset & buf->chan->subbuf_size);
-//ust//        page = buf->hpage[odd];
-//ust//        if (offset < page->offset || offset >= page->offset + PAGE_SIZE)
-//ust//                buf->hpage[odd] = page = buf->wpage;
-//ust//        page = ltt_relay_cache_page(buf, &buf->hpage[odd], page, offset);
-//ust//        return page_address(page->page) + (offset & ~PAGE_MASK);
-       return ((char *)buf->buf_data)+offset;
-       return NULL;
-}
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_offset_address);
-
-/**
- *     relay_file_open - open file op for relay files
- *     @inode: the inode
- *     @filp: the file
- *
- *     Increments the channel buffer refcount.
- */
-//ust// static int relay_file_open(struct inode *inode, struct file *filp)
-//ust// {
-//ust//        struct rchan_buf *buf = inode->i_private;
-//ust//        kref_get(&buf->kref);
-//ust//        filp->private_data = buf;
-//ust// 
-//ust//        return nonseekable_open(inode, filp);
-//ust// }
-
-/**
- *     relay_file_release - release file op for relay files
- *     @inode: the inode
- *     @filp: the file
- *
- *     Decrements the channel refcount, as the filesystem is
- *     no longer using it.
- */
-//ust// static int relay_file_release(struct inode *inode, struct file *filp)
-//ust// {
-//ust//        struct rchan_buf *buf = filp->private_data;
-//ust//        kref_put(&buf->kref, relay_remove_buf);
-//ust// 
-//ust//        return 0;
-//ust// }
-
-//ust// const struct file_operations ltt_relay_file_operations = {
-//ust//        .open           = relay_file_open,
-//ust//        .release        = relay_file_release,
-//ust// };
-//ust// EXPORT_SYMBOL_GPL(ltt_relay_file_operations);
-
-//ust// static __init int relay_init(void)
-//ust// {
-//ust//        hotcpu_notifier(relay_hotcpu_callback, 5);
-//ust//        return 0;
-//ust// }
-
-//ust// module_init(relay_init);
-/*
- * 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.
- *   }
- * }
- */
-
-//ust// #include <linux/time.h>
-//ust// #include <linux/ltt-tracer.h>
-//ust// #include <linux/ltt-relay.h>
-//ust// #include <linux/module.h>
-//ust// #include <linux/string.h>
-//ust// #include <linux/slab.h>
-//ust// #include <linux/init.h>
-//ust// #include <linux/rcupdate.h>
-//ust// #include <linux/sched.h>
-//ust// #include <linux/bitops.h>
-//ust// #include <linux/fs.h>
-//ust// #include <linux/smp_lock.h>
-//ust// #include <linux/debugfs.h>
-//ust// #include <linux/stat.h>
-//ust// #include <linux/cpu.h>
-//ust// #include <linux/pipe_fs_i.h>
-//ust// #include <linux/splice.h>
-//ust// #include <asm/atomic.h>
-//ust// #include <asm/local.h>
-
-#if 0
-#define printk_dbg(fmt, args...) printk(fmt, args)
-#else
-#define printk_dbg(fmt, args...)
-#endif
-
-/*
- * 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
-
-//ust// 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 n_subbufs);
-
-static void ltt_relay_destroy_buffer(struct ltt_channel_struct *ltt_chan);
-
-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 = channel->buf;
-       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);
-
-}
-
-void (*wake_consumer)(void *, int) = NULL;
-
-void relay_set_wake_consumer(void (*wake)(void *, int))
-{
-       wake_consumer = wake;
-}
-
-void relay_wake_consumer(void *arg, int finished)
-{
-       if(wake_consumer)
-               wake_consumer(arg, finished);
-}
-
-static notrace void ltt_deliver(struct rchan_buf *buf, unsigned int subbuf_idx,
-               long commit_count)
-{
-       struct ltt_channel_struct *channel =
-               (struct ltt_channel_struct *)buf->chan->private_data;
-       struct ltt_channel_buf_struct *ltt_buf = channel->buf;
-       int result;
-
-//ust// #ifdef CONFIG_LTT_VMCORE
-       local_set(&ltt_buf->commit_seq[subbuf_idx], commit_count);
-//ust// #endif
-
-       /* wakeup consumer */
-       result = write(ltt_buf->data_ready_fd_write, "1", 1);
-       if(result == -1) {
-               PERROR("write (in ltt_relay_buffer_flush)");
-               ERR("this should never happen!");
-       }
-//ust//        atomic_set(&ltt_buf->wakeup_readers, 1);
-}
-
-static struct dentry *ltt_create_buf_file_callback(struct rchan_buf *buf)
-{
-       struct ltt_channel_struct *ltt_chan;
-       int err;
-//ust//        struct dentry *dentry;
-
-       ltt_chan = buf->chan->private_data;
-       err = ltt_relay_create_buffer(ltt_chan->trace, ltt_chan, buf, buf->chan->n_subbufs);
-       if (err)
-               return ERR_PTR(err);
-
-//ust//        dentry = debugfs_create_file(filename, mode, parent, buf,
-//ust//                        &ltt_file_operations);
-//ust//        if (!dentry)
-//ust//                goto error;
-//ust//        return dentry;
-       return NULL; //ust//
-//ust//error:
-       ltt_relay_destroy_buffer(ltt_chan);
-       return NULL;
-}
-
-//ust// static int ltt_remove_buf_file_callback(struct rchan_buf *buf)
-//ust// {
-//ust// //ust//        struct rchan_buf *buf = dentry->d_inode->i_private;
-//ust//        struct ltt_channel_struct *ltt_chan = buf->chan->private_data;
-//ust// 
-//ust// //ust//        debugfs_remove(dentry);
-//ust//        ltt_relay_destroy_buffer(ltt_chan);
-//ust// 
-//ust//        return 0;
-//ust// }
-
-/*
- * Wake writers :
- *
- * This must be done after the trace is removed from the RCU list so that there
- * are no stalled writers.
- */
-//ust// static void ltt_relay_wake_writers(struct ltt_channel_buf_struct *ltt_buf)
-//ust// {
-//ust// 
-//ust//        if (waitqueue_active(&ltt_buf->write_wait))
-//ust//                wake_up_interruptible(&ltt_buf->write_wait);
-//ust// }
-
-/*
- * 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)
-{
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf = ltt_channel->buf;
-//ust//
-//ust//        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.
- */
-//ust// static int ltt_open(struct inode *inode, struct file *file)
-//ust// {
-//ust//        struct rchan_buf *buf = inode->i_private;
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_channel->buf, buf->cpu);
-//ust// 
-//ust//        if (!atomic_long_add_unless(&ltt_buf->active_readers, 1, 1))
-//ust//                return -EBUSY;
-//ust//        return ltt_relay_file_operations.open(inode, file);
-//ust// }
-
-/**
- *     ltt_release - release file op for ltt files
- *     @inode: opened inode
- *     @file: opened file
- *
- *     Release implementation.
- */
-//ust// static int ltt_release(struct inode *inode, struct file *file)
-//ust// {
-//ust//        struct rchan_buf *buf = inode->i_private;
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_channel->buf, buf->cpu);
-//ust//        int ret;
-//ust// 
-//ust//        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
-//ust//        atomic_long_dec(&ltt_buf->active_readers);
-//ust//        ret = ltt_relay_file_operations.release(inode, file);
-//ust//        WARN_ON(ret);
-//ust//        return ret;
-//ust// }
-
-/**
- *     ltt_poll - file op for ltt files
- *     @filp: the file
- *     @wait: poll table
- *
- *     Poll implementation.
- */
-//ust// static unsigned int ltt_poll(struct file *filp, poll_table *wait)
-//ust// {
-//ust//        unsigned int mask = 0;
-//ust//        struct inode *inode = filp->f_dentry->d_inode;
-//ust//        struct rchan_buf *buf = inode->i_private;
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_channel->buf, buf->cpu);
-//ust// 
-//ust//        if (filp->f_mode & FMODE_READ) {
-//ust//                poll_wait_set_exclusive(wait);
-//ust//                poll_wait(filp, &buf->read_wait, wait);
-//ust// 
-//ust//                WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
-//ust//                if (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
-//ust//                                                        buf->chan)
-//ust//                  - SUBBUF_TRUNC(atomic_long_read(&ltt_buf->consumed),
-//ust//                                                        buf->chan)
-//ust//                  == 0) {
-//ust//                        if (buf->finalized)
-//ust//                                return POLLHUP;
-//ust//                        else
-//ust//                                return 0;
-//ust//                } else {
-//ust//                        struct rchan *rchan =
-//ust//                                ltt_channel->trans_channel_data;
-//ust//                        if (SUBBUF_TRUNC(local_read(&ltt_buf->offset),
-//ust//                                        buf->chan)
-//ust//                          - SUBBUF_TRUNC(atomic_long_read(
-//ust//                                                &ltt_buf->consumed),
-//ust//                                        buf->chan)
-//ust//                          >= rchan->alloc_size)
-//ust//                                return POLLPRI | POLLRDBAND;
-//ust//                        else
-//ust//                                return POLLIN | POLLRDNORM;
-//ust//                }
-//ust//        }
-//ust//        return mask;
-//ust// }
-
-int ltt_do_get_subbuf(struct rchan_buf *buf, struct ltt_channel_buf_struct *ltt_buf, long *pconsumed_old)
-{
-       struct ltt_channel_struct *ltt_channel = (struct ltt_channel_struct *)buf->chan->private_data;
-       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;
-}
-
-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);
-
-//ust//        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. */
-//ust//                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);
-//ust//                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.
- */
-//ust// static int ltt_ioctl(struct inode *inode, struct file *filp,
-//ust//                unsigned int cmd, unsigned long arg)
-//ust// {
-//ust//        struct rchan_buf *buf = inode->i_private;
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_channel->buf, buf->cpu);
-//ust//        u32 __user *argp = (u32 __user *)arg;
-//ust// 
-//ust//        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
-//ust//        switch (cmd) {
-//ust//        case RELAY_GET_SUBBUF:
-//ust//        {
-//ust//                int ret;
-//ust//                ret = ltt_do_get_subbuf(buf, ltt_buf, &consumed_old);
-//ust//                if(ret < 0)
-//ust//                        return ret;
-//ust//                return put_user((u32)consumed_old, argp);
-//ust//        }
-//ust//        case RELAY_PUT_SUBBUF:
-//ust//        {
-//ust//                int ret;
-//ust//                u32 uconsumed_old;
-//ust//                ret = get_user(uconsumed_old, argp);
-//ust//                if (ret)
-//ust//                        return ret; /* will return -EFAULT */
-//ust//                return ltt_do_put_subbuf(buf, ltt_buf, uconsumed_old);
-//ust//        }
-//ust//        case RELAY_GET_N_SUBBUFS:
-//ust//                return put_user((u32)buf->chan->n_subbufs, argp);
-//ust//                break;
-//ust//        case RELAY_GET_SUBBUF_SIZE:
-//ust//                return put_user((u32)buf->chan->subbuf_size, argp);
-//ust//                break;
-//ust//        default:
-//ust//                return -ENOIOCTLCMD;
-//ust//        }
-//ust//        return 0;
-//ust// }
-
-//ust// #ifdef CONFIG_COMPAT
-//ust// static long ltt_compat_ioctl(struct file *file, unsigned int cmd,
-//ust//                unsigned long arg)
-//ust// {
-//ust//        long ret = -ENOIOCTLCMD;
-//ust// 
-//ust//        lock_kernel();
-//ust//        ret = ltt_ioctl(file->f_dentry->d_inode, file, cmd, arg);
-//ust//        unlock_kernel();
-//ust// 
-//ust//        return ret;
-//ust// }
-//ust// #endif
-
-//ust// static void ltt_relay_pipe_buf_release(struct pipe_inode_info *pipe,
-//ust//                                   struct pipe_buffer *pbuf)
-//ust// {
-//ust// }
-//ust// 
-//ust// static struct pipe_buf_operations ltt_relay_pipe_buf_ops = {
-//ust//        .can_merge = 0,
-//ust//        .map = generic_pipe_buf_map,
-//ust//        .unmap = generic_pipe_buf_unmap,
-//ust//        .confirm = generic_pipe_buf_confirm,
-//ust//        .release = ltt_relay_pipe_buf_release,
-//ust//        .steal = generic_pipe_buf_steal,
-//ust//        .get = generic_pipe_buf_get,
-//ust// };
-
-//ust// static void ltt_relay_page_release(struct splice_pipe_desc *spd, unsigned int i)
-//ust// {
-//ust// }
-
-/*
- *     subbuf_splice_actor - splice up to one subbuf's worth of data
- */
-//ust// static int subbuf_splice_actor(struct file *in,
-//ust//                               loff_t *ppos,
-//ust//                               struct pipe_inode_info *pipe,
-//ust//                               size_t len,
-//ust//                               unsigned int flags)
-//ust// {
-//ust//        struct rchan_buf *buf = in->private_data;
-//ust//        struct ltt_channel_struct *ltt_channel =
-//ust//                (struct ltt_channel_struct *)buf->chan->private_data;
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_channel->buf, buf->cpu);
-//ust//        unsigned int poff, subbuf_pages, nr_pages;
-//ust//        struct page *pages[PIPE_BUFFERS];
-//ust//        struct partial_page partial[PIPE_BUFFERS];
-//ust//        struct splice_pipe_desc spd = {
-//ust//                .pages = pages,
-//ust//                .nr_pages = 0,
-//ust//                .partial = partial,
-//ust//                .flags = flags,
-//ust//                .ops = &ltt_relay_pipe_buf_ops,
-//ust//                .spd_release = ltt_relay_page_release,
-//ust//        };
-//ust//        long consumed_old, consumed_idx, roffset;
-//ust//        unsigned long bytes_avail;
-//ust// 
-//ust//        /*
-//ust//         * Check that a GET_SUBBUF ioctl has been done before.
-//ust//         */
-//ust//        WARN_ON(atomic_long_read(&ltt_buf->active_readers) != 1);
-//ust//        consumed_old = atomic_long_read(&ltt_buf->consumed);
-//ust//        consumed_old += *ppos;
-//ust//        consumed_idx = SUBBUF_INDEX(consumed_old, buf->chan);
-//ust// 
-//ust//        /*
-//ust//         * Adjust read len, if longer than what is available
-//ust//         */
-//ust//        bytes_avail = SUBBUF_TRUNC(local_read(&ltt_buf->offset), buf->chan)
-//ust//                    - consumed_old;
-//ust//        WARN_ON(bytes_avail > buf->chan->alloc_size);
-//ust//        len = min_t(size_t, len, bytes_avail);
-//ust//        subbuf_pages = bytes_avail >> PAGE_SHIFT;
-//ust//        nr_pages = min_t(unsigned int, subbuf_pages, PIPE_BUFFERS);
-//ust//        roffset = consumed_old & PAGE_MASK;
-//ust//        poff = consumed_old & ~PAGE_MASK;
-//ust//        printk_dbg(KERN_DEBUG "SPLICE actor len %zu pos %zd write_pos %ld\n",
-//ust//                len, (ssize_t)*ppos, local_read(&ltt_buf->offset));
-//ust// 
-//ust//        for (; spd.nr_pages < nr_pages; spd.nr_pages++) {
-//ust//                unsigned int this_len;
-//ust//                struct buf_page *page;
-//ust// 
-//ust//                if (!len)
-//ust//                        break;
-//ust//                printk_dbg(KERN_DEBUG "SPLICE actor loop len %zu roffset %ld\n",
-//ust//                        len, roffset);
-//ust// 
-//ust//                this_len = PAGE_SIZE - poff;
-//ust//                page = ltt_relay_read_get_page(buf, roffset);
-//ust//                spd.pages[spd.nr_pages] = page->page;
-//ust//                spd.partial[spd.nr_pages].offset = poff;
-//ust//                spd.partial[spd.nr_pages].len = this_len;
-//ust// 
-//ust//                poff = 0;
-//ust//                roffset += PAGE_SIZE;
-//ust//                len -= this_len;
-//ust//        }
-//ust// 
-//ust//        if (!spd.nr_pages)
-//ust//                return 0;
-//ust// 
-//ust//        return splice_to_pipe(pipe, &spd);
-//ust// }
-
-//ust// static ssize_t ltt_relay_file_splice_read(struct file *in,
-//ust//                                      loff_t *ppos,
-//ust//                                      struct pipe_inode_info *pipe,
-//ust//                                      size_t len,
-//ust//                                      unsigned int flags)
-//ust// {
-//ust//        ssize_t spliced;
-//ust//        int ret;
-//ust// 
-//ust//        ret = 0;
-//ust//        spliced = 0;
-//ust// 
-//ust//        printk_dbg(KERN_DEBUG "SPLICE read len %zu pos %zd\n",
-//ust//                len, (ssize_t)*ppos);
-//ust//        while (len && !spliced) {
-//ust//                ret = subbuf_splice_actor(in, ppos, pipe, len, flags);
-//ust//                printk_dbg(KERN_DEBUG "SPLICE read loop ret %d\n", ret);
-//ust//                if (ret < 0)
-//ust//                        break;
-//ust//                else if (!ret) {
-//ust//                        if (flags & SPLICE_F_NONBLOCK)
-//ust//                                ret = -EAGAIN;
-//ust//                        break;
-//ust//                }
-//ust// 
-//ust//                *ppos += ret;
-//ust//                if (ret > len)
-//ust//                        len = 0;
-//ust//                else
-//ust//                        len -= ret;
-//ust//                spliced += ret;
-//ust//        }
-//ust// 
-//ust//        if (spliced)
-//ust//                return spliced;
-//ust// 
-//ust//        return ret;
-//ust// }
-
-static void ltt_relay_print_subbuffer_errors(
-               struct ltt_channel_struct *ltt_chan,
-               long cons_off)
-{
-       struct rchan *rchan = ltt_chan->trans_channel_data;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_chan->buf;
-       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\n",
-               ltt_chan->channel_name, write_offset, cons_off);
-       /* 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)
-{
-       struct rchan *rchan = ltt_chan->trans_channel_data;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_chan->buf;
-       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);
-}
-
-static void ltt_relay_print_buffer_errors(struct ltt_channel_struct *ltt_chan)
-{
-       struct ltt_trace_struct *trace = ltt_chan->trace;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_chan->buf;
-
-       if (local_read(&ltt_buf->events_lost))
-               printk(KERN_ALERT
-                       "LTT : %s : %ld events lost "
-                       "in %s channel.\n",
-                       ltt_chan->channel_name,
-                       local_read(&ltt_buf->events_lost),
-                       ltt_chan->channel_name);
-       if (local_read(&ltt_buf->corrupted_subbuffers))
-               printk(KERN_ALERT
-                       "LTT : %s : %ld corrupted subbuffers "
-                       "in %s channel.\n",
-                       ltt_chan->channel_name,
-                       local_read(&ltt_buf->corrupted_subbuffers),
-                       ltt_chan->channel_name);
-
-       ltt_relay_print_errors(trace, ltt_chan);
-}
-
-static void ltt_relay_remove_dirs(struct ltt_trace_struct *trace)
-{
-//ust//        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);
-       free(ltt_chan->buf);
-}
-
-/*
- * Create ltt buffer.
- */
-//ust// static int ltt_relay_create_buffer(struct ltt_trace_struct *trace,
-//ust//                struct ltt_channel_struct *ltt_chan, struct rchan_buf *buf,
-//ust//                unsigned int cpu, unsigned int n_subbufs)
-//ust// {
-//ust//        struct ltt_channel_buf_struct *ltt_buf =
-//ust//                percpu_ptr(ltt_chan->buf, cpu);
-//ust//        unsigned int j;
-//ust// 
-//ust//        ltt_buf->commit_count =
-//ust//                kzalloc_node(sizeof(ltt_buf->commit_count) * n_subbufs,
-//ust//                        GFP_KERNEL, cpu_to_node(cpu));
-//ust//        if (!ltt_buf->commit_count)
-//ust//                return -ENOMEM;
-//ust//        kref_get(&trace->kref);
-//ust//        kref_get(&trace->ltt_transport_kref);
-//ust//        kref_get(&ltt_chan->kref);
-//ust//        local_set(&ltt_buf->offset, ltt_subbuffer_header_size());
-//ust//        atomic_long_set(&ltt_buf->consumed, 0);
-//ust//        atomic_long_set(&ltt_buf->active_readers, 0);
-//ust//        for (j = 0; j < n_subbufs; j++)
-//ust//                local_set(&ltt_buf->commit_count[j], 0);
-//ust//        init_waitqueue_head(&ltt_buf->write_wait);
-//ust//        atomic_set(&ltt_buf->wakeup_readers, 0);
-//ust//        spin_lock_init(&ltt_buf->full_lock);
-//ust// 
-//ust//        ltt_buffer_begin_callback(buf, trace->start_tsc, 0);
-//ust//        /* atomic_add made on local variable on data that belongs to
-//ust//         * various CPUs : ok because tracing not started (for this cpu). */
-//ust//        local_add(ltt_subbuffer_header_size(), &ltt_buf->commit_count[0]);
-//ust// 
-//ust//        local_set(&ltt_buf->events_lost, 0);
-//ust//        local_set(&ltt_buf->corrupted_subbuffers, 0);
-//ust// 
-//ust//        return 0;
-//ust// }
-
-static int ltt_relay_create_buffer(struct ltt_trace_struct *trace,
-               struct ltt_channel_struct *ltt_chan, struct rchan_buf *buf,
-               unsigned int n_subbufs)
-{
-       struct ltt_channel_buf_struct *ltt_buf = ltt_chan->buf;
-       unsigned int j;
-       int fds[2];
-       int result;
-
-       ltt_buf->commit_count =
-               zmalloc(sizeof(ltt_buf->commit_count) * n_subbufs);
-       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);
-//ust//        init_waitqueue_head(&ltt_buf->write_wait);
-//ust//        atomic_set(&ltt_buf->wakeup_readers, 0);
-//ust//        spin_lock_init(&ltt_buf->full_lock);
-
-       ltt_buffer_begin_callback(buf, trace->start_tsc, 0);
-
-       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);
-
-       result = pipe(fds);
-       if(result == -1) {
-               PERROR("pipe");
-               return -1;
-       }
-       ltt_buf->data_ready_fd_read = fds[0];
-       ltt_buf->data_ready_fd_write = fds[1];
-
-       /* FIXME: do we actually need this? */
-       result = fcntl(fds[0], F_SETFL, O_NONBLOCK);
-       if(result == -1) {
-               PERROR("fcntl");
-       }
-
-//ust//        ltt_buf->commit_seq = malloc(sizeof(ltt_buf->commit_seq) * n_subbufs);
-//ust//        if(!ltt_buf->commit_seq) {
-//ust//                return -1;
-//ust//        }
-
-       /* FIXME: decrementally destroy on error */
-
-       return 0;
-}
-
-static void ltt_relay_destroy_buffer(struct ltt_channel_struct *ltt_chan)
-{
-       struct ltt_trace_struct *trace = ltt_chan->trace;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_chan->buf;
-
-       kref_put(&ltt_chan->trace->ltt_transport_kref,
-               ltt_release_transport);
-       ltt_relay_print_buffer_errors(ltt_chan);
-//ust//        free(ltt_buf->commit_seq);
-       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);
-//ust//        wake_up_interruptible(&trace->kref_wq);
-}
-
-static void ltt_chan_alloc_ltt_buf(struct ltt_channel_struct *ltt_chan)
-{
-       void *ptr;
-       int result;
-
-       /* Get one page */
-       /* FIXME: increase size if we have a seq_commit array that overflows the page */
-       size_t size = PAGE_ALIGN(1);
-
-       result = ltt_chan->buf_shmid = shmget(getpid(), size, IPC_CREAT | IPC_EXCL | 0700);
-       if(ltt_chan->buf_shmid == -1) {
-               PERROR("shmget");
-               return;
-       }
-
-       ptr = shmat(ltt_chan->buf_shmid, NULL, 0);
-       if(ptr == (void *) -1) {
-               perror("shmat");
-               goto destroy_shmem;
-       }
-
-       /* Already mark the shared memory for destruction. This will occur only
-         * when all users have detached.
-        */
-       result = shmctl(ltt_chan->buf_shmid, IPC_RMID, NULL);
-       if(result == -1) {
-               perror("shmctl");
-               return;
-       }
-
-       ltt_chan->buf = ptr;
-
-       return;
-
-       destroy_shmem:
-       result = shmctl(ltt_chan->buf_shmid, IPC_RMID, NULL);
-       if(result == -1) {
-               perror("shmctl");
-       }
-
-       return;
-}
-
-/*
- * 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);
-//ust//        ltt_chan->buf = percpu_alloc_mask(sizeof(struct ltt_channel_buf_struct), GFP_KERNEL, cpu_possible_map);
-
-       ltt_chan_alloc_ltt_buf(ltt_chan);
-
-//ust//        ltt_chan->buf = malloc(sizeof(struct ltt_channel_buf_struct));
-       if (!ltt_chan->buf)
-               goto alloc_error;
-       ltt_chan->trans_channel_data = ltt_relay_open(tmpname,
-                       dir,
-                       subbuf_size,
-                       n_subbufs,
-                       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:
-//ust//        percpu_free(ltt_chan->buf);
-alloc_error:
-       err = EPERM;
-end:
-       kfree(tmpname);
-       return err;
-}
-
-static int ltt_relay_create_dirs(struct ltt_trace_struct *new_trace)
-{
-//ust//        new_trace->dentry.trace_root = debugfs_create_dir(new_trace->trace_name,
-//ust//                        get_ltt_root());
-//ust//        if (new_trace->dentry.trace_root == NULL) {
-//ust//                printk(KERN_ERR "LTT : Trace directory name %s already taken\n",
-//ust//                                new_trace->trace_name);
-//ust//                return EEXIST;
-//ust//        }
-
-//ust//        new_trace->callbacks.create_buf_file = ltt_create_buf_file_callback;
-//ust//        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)
-{
-       struct ltt_channel_struct *channel =
-               (struct ltt_channel_struct *)buf->chan->private_data;
-       struct ltt_channel_buf_struct *ltt_buf = channel->buf;
-       int result;
-
-       buf->finalized = 1;
-       ltt_force_switch(buf, FORCE_FLUSH);
-
-       result = write(ltt_buf->data_ready_fd_write, "1", 1);
-       if(result == -1) {
-               PERROR("write (in ltt_relay_buffer_flush)");
-               ERR("this should never happen!");
-       }
-}
-
-static void ltt_relay_async_wakeup_chan(struct ltt_channel_struct *ltt_channel)
-{
-//ust//        unsigned int i;
-//ust//        struct rchan *rchan = ltt_channel->trans_channel_data;
-//ust//
-//ust//        for_each_possible_cpu(i) {
-//ust//                struct ltt_channel_buf_struct *ltt_buf =
-//ust//                        percpu_ptr(ltt_channel->buf, i);
-//ust//
-//ust//                if (atomic_read(&ltt_buf->wakeup_readers) == 1) {
-//ust//                        atomic_set(&ltt_buf->wakeup_readers, 0);
-//ust//                        wake_up_interruptible(&rchan->buf[i]->read_wait);
-//ust//                }
-//ust//        }
-}
-
-static void ltt_relay_finish_buffer(struct ltt_channel_struct *ltt_channel)
-{
-       struct rchan *rchan = ltt_channel->trans_channel_data;
-//     int result;
-
-       if (rchan->buf) {
-               struct ltt_channel_buf_struct *ltt_buf = ltt_channel->buf;
-               ltt_relay_buffer_flush(rchan->buf);
-//ust//                ltt_relay_wake_writers(ltt_buf);
-               /* closing the pipe tells the consumer the buffer is finished */
-               
-               //result = write(ltt_buf->data_ready_fd_write, "D", 1);
-               //if(result == -1) {
-               //      PERROR("write (in ltt_relay_finish_buffer)");
-               //      ERR("this should never happen!");
-               //}
-               close(ltt_buf->data_ready_fd_write);
-       }
-}
-
-
-static void ltt_relay_finish_channel(struct ltt_channel_struct *ltt_channel)
-{
-//ust//        unsigned int i;
-
-//ust//        for_each_possible_cpu(i)
-               ltt_relay_finish_buffer(ltt_channel);
-}
-
-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) {
-                       long consumed;
-
-                       consumed = atomic_long_read(&ltt_buf->consumed);
-
-                       /* Next buffer not corrupted. */
-                       if (!ltt_channel->overwrite &&
-                               (SUBBUF_TRUNC(offsets->begin, buf->chan)
-                                - SUBBUF_TRUNC(consumed, buf->chan))
-                               >= rchan->alloc_size) {
-
-                               long consumed_idx = SUBBUF_INDEX(consumed, buf->chan);
-                               long commit_count = local_read(&ltt_buf->commit_count[consumed_idx]);
-                               if(((commit_count - buf->chan->subbuf_size) & ltt_channel->commit_count_mask) - (BUFFER_TRUNC(consumed, buf->chan) >> ltt_channel->n_subbufs_order) != 0) {
-                                       WARN("Event dropped. Caused by non-committed event.");
-                               }
-                               else {
-                                       WARN("Event dropped. Caused by non-consumed buffer.");
-                               }
-                               /*
-                                * 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, offsets->commit_count);
-}
-
-/*
- * 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, offsets->commit_count);
-}
-
-
-/*
- * 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, offsets->commit_count);
-}
-
-/**
- * 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)
-{
-       struct rchan *rchan = ltt_channel->trans_channel_data;
-       struct rchan_buf *buf = *transport_data = rchan->buf;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_channel->buf;
-       struct ltt_reserve_switch_offsets offsets;
-
-       offsets.reserve_commit_diff = 0;
-       offsets.size = 0;
-
-       /*
-        * Perform retryable operations.
-        */
-       if (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 = ltt_channel->buf;
-       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);
-}
-
-/*
- * 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)
-{
-//ust//        struct rchan *rchan;
-//ust//        struct ltt_channel_buf_struct *ltt_buf;
-//ust//        struct ltt_channel_struct *channel;
-//ust//        struct rchan_buf *relay_buf;
-//ust//        int cpu;
-//ust//        DECLARE_WAITQUEUE(wait, current);
-//ust// 
-//ust//        channel = &trace->channels[chan_index];
-//ust//        rchan = channel->trans_channel_data;
-//ust//        cpu = smp_processor_id();
-//ust//        relay_buf = rchan->buf[cpu];
-//ust//        ltt_buf = percpu_ptr(channel->buf, cpu);
-//ust// 
-//ust//        /*
-//ust//         * Check if data is too big for the channel : do not
-//ust//         * block for it.
-//ust//         */
-//ust//        if (LTT_RESERVE_CRITICAL + data_size > relay_buf->chan->subbuf_size)
-//ust//                return 0;
-//ust// 
-//ust//        /*
-//ust//         * If free space too low, we block. We restart from the
-//ust//         * beginning after we resume (cpu id may have changed
-//ust//         * while preemption is active).
-//ust//         */
-//ust//        spin_lock(&ltt_buf->full_lock);
-//ust//        if (!channel->overwrite) {
-//ust//                dbg->write = local_read(&ltt_buf->offset);
-//ust//                dbg->read = atomic_long_read(&ltt_buf->consumed);
-//ust//                dbg->avail_size = dbg->write + LTT_RESERVE_CRITICAL + data_size
-//ust//                                  - SUBBUF_TRUNC(dbg->read,
-//ust//                                                 relay_buf->chan);
-//ust//                if (dbg->avail_size > rchan->alloc_size) {
-//ust//                        __set_current_state(TASK_INTERRUPTIBLE);
-//ust//                        add_wait_queue(&ltt_buf->write_wait, &wait);
-//ust//                        spin_unlock(&ltt_buf->full_lock);
-//ust//                        preempt_enable();
-//ust//                        schedule();
-//ust//                        __set_current_state(TASK_RUNNING);
-//ust//                        remove_wait_queue(&ltt_buf->write_wait, &wait);
-//ust//                        if (signal_pending(current))
-//ust//                                return -ERESTARTSYS;
-//ust//                        preempt_disable();
-//ust//                        return 1;
-//ust//                }
-//ust//        }
-//ust//        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)
-{
-       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;
-       ltt_buf = channel->buf;
-
-       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", ltt_nesting);
-       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);
-}
-
-//ust// static struct ltt_transport ltt_relay_transport = {
-//ust//        .name = "relay",
-//ust//        .owner = THIS_MODULE,
-//ust//        .ops = {
-//ust//                .create_dirs = ltt_relay_create_dirs,
-//ust//                .remove_dirs = ltt_relay_remove_dirs,
-//ust//                .create_channel = ltt_relay_create_channel,
-//ust//                .finish_channel = ltt_relay_finish_channel,
-//ust//                .remove_channel = ltt_relay_remove_channel,
-//ust//                .wakeup_channel = ltt_relay_async_wakeup_chan,
-//ust//                .commit_slot = ltt_relay_commit_slot,
-//ust//                .reserve_slot = ltt_relay_reserve_slot,
-//ust//                .user_blocking = ltt_relay_user_blocking,
-//ust//                .user_errors = ltt_relay_print_user_errors,
-//ust//        },
-//ust// };
-
-static struct ltt_transport ust_relay_transport = {
-       .name = "ustrelay",
-//ust//        .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,
-       },
-};
-
-//ust// static int __init ltt_relay_init(void)
-//ust// {
-//ust//        printk(KERN_INFO "LTT : ltt-relay init\n");
-//ust//
-//ust//        ltt_file_operations = ltt_relay_file_operations;
-//ust//        ltt_file_operations.owner = THIS_MODULE;
-//ust//        ltt_file_operations.open = ltt_open;
-//ust//        ltt_file_operations.release = ltt_release;
-//ust//        ltt_file_operations.poll = ltt_poll;
-//ust//        ltt_file_operations.splice_read = ltt_relay_file_splice_read,
-//ust//        ltt_file_operations.ioctl = ltt_ioctl;
-//ust//#ifdef CONFIG_COMPAT
-//ust//        ltt_file_operations.compat_ioctl = ltt_compat_ioctl;
-//ust//#endif
-//ust// 
-//ust//        ltt_transport_register(&ltt_relay_transport);
-//ust// 
-//ust//        return 0;
-//ust// }
-
-/*
- * 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.
- */
-//ust// #ifdef CONFIG_LTT_VMCORE
-static /* inline */ void ltt_write_commit_counter(struct rchan_buf *buf,
-               struct ltt_channel_buf_struct *ltt_buf,
-               long idx, long buf_offset, long commit_count, size_t data_size)
-{
-       long offset;
-       long commit_seq_old;
-
-       offset = buf_offset + data_size;
-
-       /*
-        * SUBBUF_OFFSET includes commit_count_mask. We can simply
-        * compare the offsets within the subbuffer without caring about
-        * buffer full/empty mismatch because offset is never zero here
-        * (subbuffer header and event headers have non-zero length).
-        */
-       if (unlikely(SUBBUF_OFFSET(offset - commit_count, buf->chan)))
-               return;
-
-       commit_seq_old = local_read(&ltt_buf->commit_seq[idx]);
-       while (commit_seq_old < commit_count)
-               commit_seq_old = local_cmpxchg(&ltt_buf->commit_seq[idx],
-                                        commit_seq_old, commit_count);
-}
-//ust// #else
-//ust// static inline void ltt_write_commit_counter(struct rchan_buf *buf,
-//ust//                long buf_offset, size_t slot_size)
-//ust// {
-//ust// }
-//ust// #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.
- * @data_size : size of the event data.
- * @slot_size : size of the reserved slot.
- */
-/* FIXME: make this function static inline in the .h! */
-/*static*/ /* inline */ notrace void ltt_commit_slot(
-               struct ltt_channel_struct *ltt_channel,
-               void **transport_data, long buf_offset,
-               size_t data_size, size_t slot_size)
-{
-       struct rchan_buf *buf = *transport_data;
-       struct ltt_channel_buf_struct *ltt_buf = ltt_channel->buf;
-       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, commit_count);
-       /*
-        * Update lost_size for each commit. It's needed only for extracting
-        * ltt buffers from vmcore, after crash.
-        */
-       ltt_write_commit_counter(buf, ltt_buf, endidx,
-                                buf_offset, commit_count, data_size);
-}
-
-
-static char initialized = 0;
-
-void __attribute__((constructor)) init_ustrelay_transport(void)
-{
-       if(!initialized) {
-               ltt_transport_register(&ust_relay_transport);
-               initialized = 1;
-       }
-}
-
-static void __attribute__((destructor)) ltt_relay_exit(void)
-{
-//ust//        printk(KERN_INFO "LTT : ltt-relay exit\n");
-
-       ltt_transport_unregister(&ust_relay_transport);
-}
-
-//ust// module_init(ltt_relay_init);
-//ust// module_exit(ltt_relay_exit);
-//ust// 
-//ust// MODULE_LICENSE("GPL");
-//ust// MODULE_AUTHOR("Mathieu Desnoyers");
-//ust// MODULE_DESCRIPTION("Linux Trace Toolkit Next Generation Lockless Relay");
diff --git a/libust/relay.h b/libust/relay.h
deleted file mode 100644 (file)
index 0e18279..0000000
+++ /dev/null
@@ -1,389 +0,0 @@
-/*
- * linux/include/linux/ltt-relay.h
- *
- * Copyright (C) 2002, 2003 - Tom Zanussi (zanussi@us.ibm.com), IBM Corp
- * Copyright (C) 1999, 2000, 2001, 2002 - Karim Yaghmour (karim@opersys.com)
- * Copyright (C) 2008 - Mathieu Desnoyers (mathieu.desnoyers@polymtl.ca)
- *
- * CONFIG_RELAY definitions and declarations
- */
-
-#ifndef _LINUX_LTT_RELAY_H
-#define _LINUX_LTT_RELAY_H
-
-//ust// #include <linux/types.h>
-//ust// #include <linux/sched.h>
-//ust// #include <linux/timer.h>
-//ust// #include <linux/wait.h>
-//ust// #include <linux/list.h>
-//ust// #include <linux/fs.h>
-//ust// #include <linux/poll.h>
-//ust// #include <linux/kref.h>
-//ust// #include <linux/mm.h>
-//ust// #include <linux/ltt-core.h>
-#include <assert.h>
-#include <kcompat/kref.h>
-//#include "list.h"
-#include "channels.h"
-#include "buffer.h"
-
-/* Needs a _much_ better name... */
-#define FIX_SIZE(x) ((((x) - 1) & PAGE_MASK) + PAGE_SIZE)
-
-/*
- * Tracks changes to rchan/rchan_buf structs
- */
-#define LTT_RELAY_CHANNEL_VERSION              8
-
-struct rchan_buf;
-
-struct buf_page {
-       struct page *page;
-       struct rchan_buf *buf;  /* buffer the page belongs to */
-       size_t offset;          /* page offset in the buffer */
-       struct list_head list;  /* buffer linked list */
-};
-
-/*
- * Per-cpu relay channel buffer
- */
-struct rchan_buf { 
-       struct rchan *chan;             /* associated channel */ 
-//ust//        wait_queue_head_t read_wait;    /* reader wait queue */ 
-//ust//        struct timer_list timer;        /* reader wake-up timer */ 
-//ust//        struct dentry *dentry;          /* channel file dentry */ 
-       struct kref kref;               /* channel buffer refcount */ 
-//ust//        struct list_head pages;         /* list of buffer pages */ 
-       void *buf_data; //ust//
-       size_t buf_size;
-//ust//        struct buf_page *wpage;         /* current write page (cache) */ 
-//ust//        struct buf_page *hpage[2];      /* current subbuf header page (cache) */ 
-//ust//        struct buf_page *rpage;         /* current subbuf read page (cache) */ 
-//ust//        unsigned int page_count;        /* number of current buffer pages */ 
-       unsigned int finalized;         /* buffer has been finalized */ 
-//ust//        unsigned int cpu;               /* this buf's cpu */ 
-       int shmid;                      /* the shmid of the buffer data pages */
-} ____cacheline_aligned; 
-
-/*
- * Relay channel data structure
- */
-struct rchan {
-       u32 version;                    /* the version of this struct */
-       size_t subbuf_size;             /* sub-buffer size */
-       size_t n_subbufs;               /* number of sub-buffers per buffer */
-       size_t alloc_size;              /* total buffer size allocated */
-       struct rchan_callbacks *cb;     /* client callbacks */
-       struct kref kref;               /* channel refcount */
-       void *private_data;             /* for user-defined data */
-//ust//        struct rchan_buf *buf[NR_CPUS]; /* per-cpu channel buffers */
-       struct rchan_buf *buf;
-       struct list_head list;          /* for channel list */
-       struct dentry *parent;          /* parent dentry passed to open */
-       int subbuf_size_order;          /* order of sub-buffer size */
-//ust//        char base_filename[NAME_MAX];   /* saved base filename */
-};
-
-/*
- * Relay channel client callbacks
- */
-struct rchan_callbacks {
-       /*
-        * subbuf_start - called on buffer-switch to a new sub-buffer
-        * @buf: the channel buffer containing the new sub-buffer
-        * @subbuf: the start of the new sub-buffer
-        * @prev_subbuf: the start of the previous sub-buffer
-        * @prev_padding: unused space at the end of previous sub-buffer
-        *
-        * The client should return 1 to continue logging, 0 to stop
-        * logging.
-        *
-        * NOTE: subbuf_start will also be invoked when the buffer is
-        *       created, so that the first sub-buffer can be initialized
-        *       if necessary.  In this case, prev_subbuf will be NULL.
-        *
-        * NOTE: the client can reserve bytes at the beginning of the new
-        *       sub-buffer by calling subbuf_start_reserve() in this callback.
-        */
-       int (*subbuf_start) (struct rchan_buf *buf,
-                            void *subbuf,
-                            void *prev_subbuf,
-                            size_t prev_padding);
-
-       /*
-        * create_buf_file - create file to represent a relay channel buffer
-        * @filename: the name of the file to create
-        * @parent: the parent of the file to create
-        * @mode: the mode of the file to create
-        * @buf: the channel buffer
-        *
-        * Called during relay_open(), once for each per-cpu buffer,
-        * to allow the client to create a file to be used to
-        * represent the corresponding channel buffer.  If the file is
-        * created outside of relay, the parent must also exist in
-        * that filesystem.
-        *
-        * The callback should return the dentry of the file created
-        * to represent the relay buffer.
-        *
-        * Setting the is_global outparam to a non-zero value will
-        * cause relay_open() to create a single global buffer rather
-        * than the default set of per-cpu buffers.
-        *
-        * See Documentation/filesystems/relayfs.txt for more info.
-        */
-       struct dentry *(*create_buf_file)(const char *filename,
-                                         struct dentry *parent,
-                                         int mode,
-                                         struct rchan_buf *buf);
-
-       /*
-        * remove_buf_file - remove file representing a relay channel buffer
-        * @dentry: the dentry of the file to remove
-        *
-        * Called during relay_close(), once for each per-cpu buffer,
-        * to allow the client to remove a file used to represent a
-        * channel buffer.
-        *
-        * The callback should return 0 if successful, negative if not.
-        */
-//ust//        int (*remove_buf_file)(struct rchan_buf *buf);
-};
-
-extern struct buf_page *ltt_relay_find_prev_page(struct rchan_buf *buf,
-       struct buf_page *page, size_t offset, ssize_t diff_offset);
-
-extern struct buf_page *ltt_relay_find_next_page(struct rchan_buf *buf,
-       struct buf_page *page, size_t offset, ssize_t diff_offset);
-
-extern void _ltt_relay_write(struct rchan_buf *buf, size_t offset,
-       const void *src, size_t len, ssize_t cpy);
-
-extern int ltt_relay_read(struct rchan_buf *buf, size_t offset,
-       void *dest, size_t len);
-
-extern struct buf_page *ltt_relay_read_get_page(struct rchan_buf *buf,
-       size_t offset);
-
-/*
- * Return the address where a given offset is located.
- * Should be used to get the current subbuffer header pointer. Given we know
- * it's never on a page boundary, it's safe to write directly to this address,
- * as long as the write is never bigger than a page size.
- */
-extern void *ltt_relay_offset_address(struct rchan_buf *buf,
-       size_t offset);
-
-/*
- * Find the page containing "offset". Cache it if it is after the currently
- * cached page.
- */
-static inline struct buf_page *ltt_relay_cache_page(struct rchan_buf *buf,
-               struct buf_page **page_cache,
-               struct buf_page *page, size_t offset)
-{
-       ssize_t diff_offset;
-       ssize_t half_buf_size = buf->chan->alloc_size >> 1;
-
-       /*
-        * Make sure this is the page we want to write into. The current
-        * page is changed concurrently by other writers. [wrh]page are
-        * used as a cache remembering the last page written
-        * to/read/looked up for header address. No synchronization;
-        * could have to find the previous page is a nested write
-        * occured. Finding the right page is done by comparing the
-        * dest_offset with the buf_page offsets.
-        * When at the exact opposite of the buffer, bias towards forward search
-        * because it will be cached.
-        */
-
-       diff_offset = (ssize_t)offset - (ssize_t)page->offset;
-       if (diff_offset <= -(ssize_t)half_buf_size)
-               diff_offset += buf->chan->alloc_size;
-       else if (diff_offset > half_buf_size)
-               diff_offset -= buf->chan->alloc_size;
-
-       if (unlikely(diff_offset >= (ssize_t)PAGE_SIZE)) {
-               page = ltt_relay_find_next_page(buf, page, offset, diff_offset);
-               *page_cache = page;
-       } else if (unlikely(diff_offset < 0)) {
-               page = ltt_relay_find_prev_page(buf, page, offset, diff_offset);
-       }
-       return page;
-}
-
-//ust// #ifdef CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS
-static inline void ltt_relay_do_copy(void *dest, const void *src, size_t len)
-{
-       union {
-               const void *src;
-               const u8 *src8;
-               const u16 *src16;
-               const u32 *src32;
-               const u64 *src64;
-       } u = { .src = src };
-
-       switch (len) {
-       case 0: break;
-       case 1: *(u8 *)dest = *u.src8;
-               break;
-       case 2: *(u16 *)dest = *u.src16;
-               break;
-       case 4: *(u32 *)dest = *u.src32;
-               break;
-//ust// #if (BITS_PER_LONG == 64)
-       case 8: *(u64 *)dest = *u.src64;
-               break;
-//ust// #endif
-       default:
-               memcpy(dest, src, len);
-       }
-}
-//ust// #else
-//ust// /*
-//ust//  * Returns whether the dest and src addresses are aligned on
-//ust//  * min(sizeof(void *), len). Call this with statically known len for efficiency.
-//ust//  */
-//ust// static inline int addr_aligned(const void *dest, const void *src, size_t len)
-//ust// {
-//ust//        if (ltt_align((size_t)dest, len))
-//ust//                return 0;
-//ust//        if (ltt_align((size_t)src, len))
-//ust//                return 0;
-//ust//        return 1;
-//ust// }
-//ust// 
-//ust// static inline void ltt_relay_do_copy(void *dest, const void *src, size_t len)
-//ust// {
-//ust//        switch (len) {
-//ust//        case 0: break;
-//ust//        case 1: *(u8 *)dest = *(const u8 *)src;
-//ust//                break;
-//ust//        case 2: if (unlikely(!addr_aligned(dest, src, 2)))
-//ust//                        goto memcpy_fallback;
-//ust//                *(u16 *)dest = *(const u16 *)src;
-//ust//                break;
-//ust//        case 4: if (unlikely(!addr_aligned(dest, src, 4)))
-//ust//                        goto memcpy_fallback;
-//ust//                *(u32 *)dest = *(const u32 *)src;
-//ust//                break;
-//ust// #if (BITS_PER_LONG == 64)
-//ust//        case 8: if (unlikely(!addr_aligned(dest, src, 8)))
-//ust//                        goto memcpy_fallback;
-//ust//                *(u64 *)dest = *(const u64 *)src;
-//ust//                break;
-//ust// #endif
-//ust//        default:
-//ust//                goto memcpy_fallback;
-//ust//        }
-//ust//        return;
-//ust// memcpy_fallback:
-//ust//        memcpy(dest, src, len);
-//ust// }
-//ust// #endif
-
-static inline int ltt_relay_write(struct rchan_buf *buf, size_t offset,
-       const void *src, size_t len)
-{
-//ust//        struct buf_page *page;
-//ust//        ssize_t pagecpy;
-//ust//
-//ust//        offset &= buf->chan->alloc_size - 1;
-//ust//        page = buf->wpage;
-//ust//
-//ust//        page = ltt_relay_cache_page(buf, &buf->wpage, page, offset);
-//ust//        pagecpy = min_t(size_t, len, PAGE_SIZE - (offset & ~PAGE_MASK));
-//ust//        ltt_relay_do_copy(page_address(page->page)
-//ust//                + (offset & ~PAGE_MASK), src, pagecpy);
-//ust//
-//ust//        if (unlikely(len != pagecpy))
-//ust//                _ltt_relay_write(buf, offset, src, len, page, pagecpy);
-//ust//        return len;
-
-       size_t cpy;
-       size_t buf_offset = BUFFER_OFFSET(offset, buf->chan);
-
-       assert(buf_offset < buf->chan->subbuf_size*buf->chan->n_subbufs);
-
-       cpy = min_t(size_t, len, buf->buf_size - buf_offset);
-       ltt_relay_do_copy(buf->buf_data + buf_offset, src, cpy);
-       
-       if (unlikely(len != cpy))
-               _ltt_relay_write(buf, buf_offset, src, len, cpy);
-       return len;
-}
-
-/*
- * CONFIG_LTT_RELAY kernel API, ltt/ltt-relay-alloc.c
- */
-
-extern struct rchan *ltt_relay_open(const char *base_filename,
-                        struct dentry *parent,
-                        size_t subbuf_size,
-                        size_t n_subbufs,
-                        void *private_data);
-extern void ltt_relay_close(struct rchan *chan);
-
-/*
- * exported ltt_relay file operations, ltt/ltt-relay-alloc.c
- */
-extern const struct file_operations ltt_relay_file_operations;
-
-
-/* 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 */
-//ust// #ifdef CONFIG_LTT_VMCORE
-//ust//        local_t *commit_seq;            /* Consecutive commits */
-//ust// #endif
-       atomic_long_t active_readers;   /*
-                                        * Active readers count
-                                        * standard atomic access (shared)
-                                        */
-       local_t events_lost;
-       local_t corrupted_subbuffers;
-//ust//        spinlock_t full_lock;           /*
-//ust//                                         * buffer full condition spinlock, only
-//ust//                                         * for userspace tracing blocking mode
-//ust//                                         * synchronization with reader.
-//ust//                                         */
-//ust//        wait_queue_head_t write_wait;   /*
-//ust//                                         * Wait queue for blocking user space
-//ust//                                         * writers
-//ust//                                         */
-//ust//        atomic_t wakeup_readers;        /* Boolean : wakeup readers waiting ? */
-       /* one byte is written to this pipe when data is available, in order
-           to wake the consumer */
-       /* portability: Single byte writes must be as quick as possible. The kernel-side
-          buffer must be large enough so the writer doesn't block. From the pipe(7)
-           man page: Since linux 2.6.11, the pipe capacity is 65536 bytes. */
-       int data_ready_fd_write;
-       /* the reading end of the pipe */
-       int data_ready_fd_read;
-
-       /* commit count per subbuffer; must be at end of struct */
-       local_t commit_seq[0] ____cacheline_aligned;
-} ____cacheline_aligned;
-
-extern int ltt_do_get_subbuf(struct rchan_buf *buf, struct ltt_channel_buf_struct *ltt_buf, long *pconsumed_old);
-
-extern int ltt_do_put_subbuf(struct rchan_buf *buf, struct ltt_channel_buf_struct *ltt_buf, u32 uconsumed_old);
-
-extern void init_ustrelay_transport(void);
-
-/*static*/ /* inline */ notrace void ltt_commit_slot(
-               struct ltt_channel_struct *ltt_channel,
-               void **transport_data, long buf_offset,
-               size_t data_size, size_t slot_size);
-
-#endif /* _LINUX_LTT_RELAY_H */
index b1859ed9a8095f63bda857a72dd14ac1d492a52c..d603815b50621e74332f96301d1b6b9a38266200 100644 (file)
  */
 
 #include <stdarg.h>
-//ust// #include <linux/ctype.h>
-//ust// #include <linux/string.h>
-//ust// #include <linux/module.h>
-//ust// #include <linux/ltt-tracer.h>
 #include <string.h>
 #include <stdint.h>
 #include <stdio.h>
@@ -38,7 +34,7 @@
 #include <urcu-bp.h>
 #include <urcu/rculist.h>
 
-#include "relay.h"
+#include "buffers.h"
 #include "tracer.h"
 //#include "list.h"
 #include "usterr.h"
@@ -337,7 +333,7 @@ parse_end:
        return fmt;
 }
 
-static inline size_t serialize_trace_data(struct rchan_buf *buf,
+static inline size_t serialize_trace_data(struct ust_buffer *buf,
                size_t buf_offset,
                char trace_size, enum ltt_type trace_type,
                char c_size, enum ltt_type c_type,
@@ -402,7 +398,7 @@ static inline size_t serialize_trace_data(struct rchan_buf *buf,
                        tmp.v_string.s = "<NULL>";
                tmp.v_string.len = strlen(tmp.v_string.s)+1;
                if (buf)
-                       ltt_relay_write(buf, buf_offset, tmp.v_string.s,
+                       ust_buffers_write(buf, buf_offset, tmp.v_string.s,
                                tmp.v_string.len);
                buf_offset += tmp.v_string.len;
                goto copydone;
@@ -430,31 +426,31 @@ static inline size_t serialize_trace_data(struct rchan_buf *buf,
                        switch (trace_size) {
                        case 1:
                                if (c_size == 8)
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint8_t[]){ (uint8_t)tmp.v_uint64 },
                                        sizeof(uint8_t));
                                else
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint8_t[]){ (uint8_t)tmp.v_ulong },
                                        sizeof(uint8_t));
                                break;
                        case 2:
                                if (c_size == 8)
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint16_t[]){ (uint16_t)tmp.v_uint64 },
                                        sizeof(uint16_t));
                                else
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint16_t[]){ (uint16_t)tmp.v_ulong },
                                        sizeof(uint16_t));
                                break;
                        case 4:
                                if (c_size == 8)
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint32_t[]){ (uint32_t)tmp.v_uint64 },
                                        sizeof(uint32_t));
                                else
-                                       ltt_relay_write(buf, buf_offset,
+                                       ust_buffers_write(buf, buf_offset,
                                        (uint32_t[]){ (uint32_t)tmp.v_ulong },
                                        sizeof(uint32_t));
                                break;
@@ -463,7 +459,7 @@ static inline size_t serialize_trace_data(struct rchan_buf *buf,
                                 * c_size cannot be other than 8 here because
                                 * trace_size > 4.
                                 */
-                               ltt_relay_write(buf, buf_offset,
+                               ust_buffers_write(buf, buf_offset,
                                (uint64_t[]){ (uint64_t)tmp.v_uint64 },
                                sizeof(uint64_t));
                                break;
@@ -480,12 +476,12 @@ static inline size_t serialize_trace_data(struct rchan_buf *buf,
                if (buf) {
                        switch (trace_type) {
                        case LTT_TYPE_SIGNED_INT:
-                               ltt_relay_write(buf, buf_offset,
+                               ust_buffers_write(buf, buf_offset,
                                        (int64_t[]){ (int64_t)tmp.v_ulong },
                                        sizeof(int64_t));
                                break;
                        case LTT_TYPE_UNSIGNED_INT:
-                               ltt_relay_write(buf, buf_offset,
+                               ust_buffers_write(buf, buf_offset,
                                        (uint64_t[]){ (uint64_t)tmp.v_ulong },
                                        sizeof(uint64_t));
                                break;
@@ -587,7 +583,7 @@ int serialize_to_text(char *outbuf, int bufsize, const char *fmt, va_list ap)
        return result;
 }
 
-notrace size_t ltt_serialize_data(struct rchan_buf *buf, size_t buf_offset,
+notrace size_t ltt_serialize_data(struct ust_buffer *buf, size_t buf_offset,
                        struct ltt_serialize_closure *closure,
                        void *serialize_private, int *largest_align,
                        const char *fmt, va_list *args)
@@ -657,7 +653,7 @@ static notrace size_t ltt_get_data_size(struct ltt_serialize_closure *closure,
 }
 
 static notrace
-void ltt_write_event_data(struct rchan_buf *buf, size_t buf_offset,
+void ltt_write_event_data(struct ust_buffer *buf, size_t buf_offset,
                                struct ltt_serialize_closure *closure,
                                void *serialize_private, int largest_align,
                                const char *fmt, va_list *args)
@@ -678,9 +674,9 @@ notrace void ltt_vtrace(const struct marker *mdata, void *probe_data,
        uint16_t eID;
        size_t data_size, slot_size;
        unsigned int chan_index;
-       struct ltt_channel_struct *channel;
+       struct ust_channel *channel;
        struct ltt_trace_struct *trace, *dest_trace = NULL;
-       struct rchan_buf *buf;
+       struct ust_buffer *buf;
        void *transport_data;
        u64 tsc;
        long buf_offset;
@@ -765,10 +761,10 @@ notrace void ltt_vtrace(const struct marker *mdata, void *probe_data,
                va_copy(args_copy, *args);
                /* FIXME : could probably encapsulate transport better. */
 //ust//                buf = ((struct rchan *)channel->trans_channel_data)->buf[cpu];
-               buf = ((struct rchan *)channel->trans_channel_data)->buf;
+               buf = channel->buf;
                /* Out-of-order write : header and data */
                buf_offset = ltt_write_event_header(trace,
-                                       channel, buf, buf_offset,
+                                       buf, buf_offset,
                                        eID, data_size, tsc, rflags);
                ltt_write_event_data(buf, buf_offset, &closure,
                                        serialize_private,
index e28b3159d73c1408b58b072850f2bb89a45d59eb..80d09ae093045fa88b6c1fcac8923ced5ef2402b 100644 (file)
@@ -33,7 +33,7 @@
 #include "tracer.h"
 #include "usterr.h"
 #include "ustcomm.h"
-#include "relay.h" /* FIXME: remove */
+#include "buffers.h" /* FIXME: remove */
 #include "marker-control.h"
 
 //#define USE_CLONE
@@ -87,9 +87,8 @@ struct blocked_consumer {
        struct ustcomm_server server;
        struct ustcomm_source src;
 
-       /* args to ltt_do_get_subbuf */
-       struct rchan_buf *rbuf;
-       struct ltt_channel_buf_struct *lttbuf;
+       /* args to ust_buffers_do_get_subbuf */
+       struct ust_buffer *buf;
 
        struct list_head list;
 };
@@ -250,13 +249,13 @@ void process_blocked_consumers(void)
                                continue;
                        }
 
-                       result = ltt_do_get_subbuf(bc->rbuf, bc->lttbuf, &consumed_old);
+                       result = ust_buffers_do_get_subbuf(bc->buf, &consumed_old);
                        if(result == -EAGAIN) {
                                WARN("missed buffer?");
                                continue;
                        }
                        else if(result < 0) {
-                               DBG("ltt_do_get_subbuf: error: %s", strerror(-result));
+                               DBG("ust_buffers_do_get_subbuf: error: %s", strerror(-result));
                        }
                        asprintf(&reply, "%s %ld", "OK", consumed_old);
                        result = ustcomm_send_reply(&bc->server, reply, &bc->src);
@@ -423,16 +422,15 @@ void *listener_main(void *p)
                        }
 
                        for(i=0; i<trace->nr_channels; i++) {
-                               struct rchan *rchan = trace->channels[i].trans_channel_data;
-                               struct rchan_buf *rbuf = rchan->buf;
-                               struct ltt_channel_struct *ltt_channel = (struct ltt_channel_struct *)rchan->private_data;
+                               struct ust_channel *channel = &trace->channels[i];
+                               struct ust_buffer *buf = channel->buf;
 
                                if(!strcmp(trace->channels[i].channel_name, channel_name)) {
                                        char *reply;
 
-                                       DBG("the shmid for the requested channel is %d", rbuf->shmid);
-                                       DBG("the shmid for its buffer structure is %d", ltt_channel->buf_shmid);
-                                       asprintf(&reply, "%d %d", rbuf->shmid, ltt_channel->buf_shmid);
+                                       DBG("the shmid for the requested channel is %d", buf->shmid);
+                                       DBG("the shmid for its buffer structure is %d", channel->buf_shmid);
+                                       asprintf(&reply, "%d %d", buf->shmid, channel->buf_shmid);
 
                                        result = ustcomm_send_reply(&ustcomm_app.server, reply, &src);
                                        if(result) {
@@ -472,13 +470,13 @@ void *listener_main(void *p)
                        }
 
                        for(i=0; i<trace->nr_channels; i++) {
-                               struct rchan *rchan = trace->channels[i].trans_channel_data;
+                               struct ust_channel *channel = &trace->channels[i];
 
                                if(!strcmp(trace->channels[i].channel_name, channel_name)) {
                                        char *reply;
 
-                                       DBG("the n_subbufs for the requested channel is %zd", rchan->n_subbufs);
-                                       asprintf(&reply, "%zd", rchan->n_subbufs);
+                                       DBG("the n_subbufs for the requested channel is %d", channel->subbuf_cnt);
+                                       asprintf(&reply, "%d", channel->subbuf_cnt);
 
                                        result = ustcomm_send_reply(&ustcomm_app.server, reply, &src);
                                        if(result) {
@@ -516,13 +514,13 @@ void *listener_main(void *p)
                        }
 
                        for(i=0; i<trace->nr_channels; i++) {
-                               struct rchan *rchan = trace->channels[i].trans_channel_data;
+                               struct ust_channel *channel = &trace->channels[i];
 
                                if(!strcmp(trace->channels[i].channel_name, channel_name)) {
                                        char *reply;
 
-                                       DBG("the subbuf_size for the requested channel is %zd", rchan->subbuf_size);
-                                       asprintf(&reply, "%zd", rchan->subbuf_size);
+                                       DBG("the subbuf_size for the requested channel is %zd", channel->subbuf_size);
+                                       asprintf(&reply, "%zd", channel->subbuf_size);
 
                                        result = ustcomm_send_reply(&ustcomm_app.server, reply, &src);
                                        if(result) {
@@ -567,11 +565,10 @@ void *listener_main(void *p)
                        }
 
                        for(i=0; i<trace->nr_channels; i++) {
-                               struct rchan *rchan = trace->channels[i].trans_channel_data;
+                               struct ust_channel *channel = &trace->channels[i];
 
                                if(!strcmp(trace->channels[i].channel_name, channel_name)) {
-                                       struct rchan_buf *rbuf = rchan->buf;
-                                       struct ltt_channel_buf_struct *lttbuf = trace->channels[i].buf;
+                                       struct ust_buffer *buf = channel->buf;
                                        struct blocked_consumer *bc;
 
                                        bc = (struct blocked_consumer *) malloc(sizeof(struct blocked_consumer));
@@ -580,9 +577,8 @@ void *listener_main(void *p)
                                                goto next_cmd;
                                        }
                                        bc->fd_consumer = src.fd;
-                                       bc->fd_producer = lttbuf->data_ready_fd_read;
-                                       bc->rbuf = rbuf;
-                                       bc->lttbuf = lttbuf;
+                                       bc->fd_producer = buf->data_ready_fd_read;
+                                       bc->buf = buf;
                                        bc->src = src;
                                        bc->server = ustcomm_app.server;
 
@@ -630,21 +626,20 @@ void *listener_main(void *p)
                        }
 
                        for(i=0; i<trace->nr_channels; i++) {
-                               struct rchan *rchan = trace->channels[i].trans_channel_data;
+                               struct ust_channel *channel = &trace->channels[i];
 
                                if(!strcmp(trace->channels[i].channel_name, channel_name)) {
-                                       struct rchan_buf *rbuf = rchan->buf;
-                                       struct ltt_channel_buf_struct *lttbuf = trace->channels[i].buf;
+                                       struct ust_buffer *buf = channel->buf;
                                        char *reply;
                                        long consumed_old=0;
 
-                                       result = ltt_do_put_subbuf(rbuf, lttbuf, consumed_old);
+                                       result = ust_buffers_do_put_subbuf(buf, consumed_old);
                                        if(result < 0) {
-                                               WARN("ltt_do_put_subbuf: error (subbuf=%s)", channel_name);
+                                               WARN("ust_buffers_do_put_subbuf: error (subbuf=%s)", channel_name);
                                                asprintf(&reply, "%s", "ERROR");
                                        }
                                        else {
-                                               DBG("ltt_do_put_subbuf: success (subbuf=%s)", channel_name);
+                                               DBG("ust_buffers_do_put_subbuf: success (subbuf=%s)", channel_name);
                                                asprintf(&reply, "%s", "OK");
                                        }
 
index a3ace72c0811de72d9cf08824dfab2137f80f229..e3538e0d5706e86368e03136fdaf98255525b86d 100644 (file)
@@ -52,6 +52,7 @@
 //ust// #include <linux/delay.h>
 //ust// #include <linux/vmalloc.h>
 //ust// #include <asm/atomic.h>
+#include <urcu-bp.h>
 #include <urcu/rculist.h>
 
 #include <ust/kernelcompat.h>
@@ -295,7 +296,7 @@ void notrace ltt_write_trace_header(struct ltt_trace_struct *trace,
 static void trace_async_wakeup(struct ltt_trace_struct *trace)
 {
        int i;
-       struct ltt_channel_struct *chan;
+       struct ust_channel *chan;
 
        /* Must check each channel for pending read wakeup */
        for (i = 0; i < trace->nr_channels; i++) {
@@ -737,7 +738,6 @@ int ltt_trace_alloc(const char *trace_name)
                subbuf_cnt = trace->channels[chan].subbuf_cnt;
                prepare_chan_size_num(&subbuf_size, &subbuf_cnt);
                err = trace->ops->create_channel(trace_name, trace,
-                               trace->dentry.trace_root,
                                channel_name,
                                &trace->channels[chan],
                                subbuf_size,
@@ -846,7 +846,7 @@ traces_error:
 static void __ltt_trace_destroy(struct ltt_trace_struct        *trace)
 {
        int i;
-       struct ltt_channel_struct *chan;
+       struct ust_channel *chan;
 
        for (i = 0; i < trace->nr_channels; i++) {
                chan = &trace->channels[i];
index 502cdcce436b763e4ba06c7a6cfe00c55e6a54ed..58aae583c6aa67d34684f5c8f896ce5116c214ab 100644 (file)
@@ -28,8 +28,7 @@
 #include <stdarg.h>
 //#include "list.h"
 #include <ust/kernelcompat.h>
-#include "buffer.h"
-#include "relay.h"
+#include "buffers.h"
 #include "channels.h"
 #include "tracercore.h"
 #include <ust/marker.h>
@@ -57,7 +56,7 @@ struct ltt_serialize_closure;
 struct ltt_probe_private_data;
 
 /* Serialization callback '%k' */
-typedef size_t (*ltt_serialize_cb)(struct rchan_buf *buf, size_t buf_offset,
+typedef size_t (*ltt_serialize_cb)(struct ust_buffer *buf, size_t buf_offset,
                        struct ltt_serialize_closure *closure,
                        void *serialize_private, int *largest_align,
                        const char *fmt, va_list *args);
@@ -68,7 +67,7 @@ struct ltt_serialize_closure {
        unsigned int cb_idx;
 };
 
-extern size_t ltt_serialize_data(struct rchan_buf *buf, size_t buf_offset,
+extern size_t ltt_serialize_data(struct ust_buffer *buf, size_t buf_offset,
                        struct ltt_serialize_closure *closure,
                        void *serialize_private,
                        int *largest_align, const char *fmt, va_list *args);
@@ -143,7 +142,7 @@ struct user_dbg_data {
 struct ltt_trace_ops {
        /* First 32 bytes cache-hot cacheline */
        int (*reserve_slot) (struct ltt_trace_struct *trace,
-                               struct ltt_channel_struct *channel,
+                               struct ust_channel *channel,
                                void **transport_data, size_t data_size,
                                size_t *slot_size, long *buf_offset, u64 *tsc,
                                unsigned int *rflags,
@@ -151,24 +150,24 @@ struct ltt_trace_ops {
 //ust//        void (*commit_slot) (struct ltt_channel_struct *channel,
 //ust//                                void **transport_data, long buf_offset,
 //ust//                                size_t slot_size);
-       void (*wakeup_channel) (struct ltt_channel_struct *ltt_channel);
+       void (*wakeup_channel) (struct ust_channel *channel);
        int (*user_blocking) (struct ltt_trace_struct *trace,
                                unsigned int index, size_t data_size,
                                struct user_dbg_data *dbg);
-       /* End of first 32 bytes cacheline */
-       int (*create_dirs) (struct ltt_trace_struct *new_trace);
-       void (*remove_dirs) (struct ltt_trace_struct *new_trace);
-       int (*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);
-       void (*finish_channel) (struct ltt_channel_struct *channel);
-       void (*remove_channel) (struct ltt_channel_struct *channel);
-       void (*user_errors) (struct ltt_trace_struct *trace,
-                               unsigned int index, size_t data_size,
-                               struct user_dbg_data *dbg);
+       /* End of first 32 bytes cacheline */
+       int (*create_dirs) (struct ltt_trace_struct *new_trace);
+       void (*remove_dirs) (struct ltt_trace_struct *new_trace);
+       int (*create_channel) (const char *trace_name,
+                               struct ltt_trace_struct *trace,
+                               const char *channel_name,
+                               struct ust_channel *channel,
+                               unsigned int subbuf_size,
+                               unsigned int n_subbufs, int overwrite);
+       void (*finish_channel) (struct ust_channel *channel);
+       void (*remove_channel) (struct ust_channel *channel);
+       void (*user_errors) (struct ltt_trace_struct *trace,
+                               unsigned int index, size_t data_size,
+                               struct user_dbg_data *dbg);
 } ____cacheline_aligned;
 
 struct ltt_transport {
@@ -190,7 +189,7 @@ struct ltt_trace_struct {
        struct ltt_trace_ops *ops;
        int active;
        /* Second 32 bytes cache-hot cacheline */
-       struct ltt_channel_struct *channels;
+       struct ust_channel *channels;
        unsigned int nr_channels;
        u32 freq_scale;
        u64 start_freq;
@@ -299,7 +298,7 @@ static inline size_t ltt_subbuffer_header_size(void)
 }
 
 /*
- * ltt_get_header_size
+ * ust_get_header_size
  *
  * Calculate alignment offset to 32-bits. This is the alignment offset of the
  * event header.
@@ -317,8 +316,8 @@ static inline size_t ltt_subbuffer_header_size(void)
  * The payload must itself determine its own alignment from the biggest type it
  * contains.
  * */
-static inline unsigned char ltt_get_header_size(
-               struct ltt_channel_struct *channel,
+static inline unsigned char ust_get_header_size(
+               struct ust_channel *channel,
                size_t offset,
                size_t data_size,
                size_t *before_hdr_pad,
@@ -370,8 +369,7 @@ static inline unsigned char ltt_get_header_size(
  * returns : offset where the event data must be written.
  */
 static inline size_t ltt_write_event_header(struct ltt_trace_struct *trace,
-               struct ltt_channel_struct *channel,
-               struct rchan_buf *buf, long buf_offset,
+               struct ust_buffer *buf, long buf_offset,
                u16 eID, size_t event_size,
                u64 tsc, unsigned int rflags)
 {
@@ -393,44 +391,44 @@ static inline size_t ltt_write_event_header(struct ltt_trace_struct *trace,
                break;
        }
        header.id_time |= (u32)tsc & LTT_TSC_MASK;
-       ltt_relay_write(buf, buf_offset, &header, sizeof(header));
+       ust_buffers_write(buf, buf_offset, &header, sizeof(header));
        buf_offset += sizeof(header);
 
        switch (rflags) {
        case LTT_RFLAG_ID_SIZE_TSC:
                small_size = min_t(size_t, event_size, 0xFFFFU);
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u16[]){ (u16)eID }, sizeof(u16));
                buf_offset += sizeof(u16);
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u16[]){ (u16)small_size }, sizeof(u16));
                buf_offset += sizeof(u16);
                if (small_size == 0xFFFFU) {
-                       ltt_relay_write(buf, buf_offset,
+                       ust_buffers_write(buf, buf_offset,
                                (u32[]){ (u32)event_size }, sizeof(u32));
                        buf_offset += sizeof(u32);
                }
                buf_offset += ltt_align(buf_offset, sizeof(u64));
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u64[]){ (u64)tsc }, sizeof(u64));
                buf_offset += sizeof(u64);
                break;
        case LTT_RFLAG_ID_SIZE:
                small_size = min_t(size_t, event_size, 0xFFFFU);
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u16[]){ (u16)eID }, sizeof(u16));
                buf_offset += sizeof(u16);
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u16[]){ (u16)small_size }, sizeof(u16));
                buf_offset += sizeof(u16);
                if (small_size == 0xFFFFU) {
-                       ltt_relay_write(buf, buf_offset,
+                       ust_buffers_write(buf, buf_offset,
                                (u32[]){ (u32)event_size }, sizeof(u32));
                        buf_offset += sizeof(u32);
                }
                break;
        case LTT_RFLAG_ID:
-               ltt_relay_write(buf, buf_offset,
+               ust_buffers_write(buf, buf_offset,
                        (u16[]){ (u16)eID }, sizeof(u16));
                buf_offset += sizeof(u16);
                break;
@@ -465,7 +463,7 @@ static inline size_t ltt_write_event_header(struct ltt_trace_struct *trace,
  */
 static inline int ltt_reserve_slot(
                struct ltt_trace_struct *trace,
-               struct ltt_channel_struct *channel,
+               struct ust_channel *channel,
                void **transport_data,
                size_t data_size,
                size_t *slot_size,
@@ -598,7 +596,7 @@ extern struct dentry *get_filter_root(void);
 
 extern void ltt_write_trace_header(struct ltt_trace_struct *trace,
                struct ltt_subbuffer_header *header);
-extern void ltt_buffer_destroy(struct ltt_channel_struct *ltt_chan);
+extern void ltt_buffer_destroy(struct ust_channel *ltt_chan);
 
 extern void ltt_core_register(int (*function)(u8, void *));
 
index 49b76f541af9f5c8bd883093d117f4c7e8b4d09d..d573a7fe2ef84ae845c698053322808252d2778d 100644 (file)
@@ -111,7 +111,7 @@ int ustcmd_destroy_trace(pid_t pid)
 {
        int result;
 
-       result = ustcmd_send_cmd("destroy", pid, NULL);
+       result = ustcmd_send_cmd("trace_destroy", pid, NULL);
        if (result) {
                return USTCMD_ERR_GEN;
        }
index aa92ee8cc3b66db244640e6e7f59ab55d516d5f2..27ff78df09065c368b68120e9d7dfb0b8e5e5131 100644 (file)
@@ -1,4 +1,4 @@
-AM_CPPFLAGS = -I$(top_srcdir)/include
+AM_CPPFLAGS = -I$(top_srcdir)/include -I$(top_srcdir)/libust
 
 noinst_PROGRAMS = hello
 hello_SOURCES = hello.c tp.c tp.h
index 4463a89f2fcf0e128302a16109f3db5ef6c2fd72..561d92e464d7f00871a97f42265a190a4c0ee141 100644 (file)
@@ -10,7 +10,7 @@
 #include <ust/marker.h>
 #include "usterr.h"
 #include "tp.h"
-
+#include "tracer.h"
 
 void inthandler(int sig)
 {
index 6a91accbb5bac43c7a55f0233fd447c473338461..48bd580ee71d59d93a8be604561b148081e7119f 100644 (file)
 
 void finish_consuming_dead_subbuffer(struct buffer_info *buf)
 {
-       struct ltt_channel_buf_struct *ltt_buf = buf->bufstruct_mem;
+       struct ust_buffer *ustbuf = buf->bufstruct_mem;
 
-       long write_offset = local_read(&ltt_buf->offset);
-       long consumed_offset = atomic_long_read(&ltt_buf->consumed);
+       long write_offset = local_read(&ustbuf->offset);
+       long consumed_offset = atomic_long_read(&ustbuf->consumed);
 
        long i_subbuf;
 
@@ -59,7 +59,7 @@ void finish_consuming_dead_subbuffer(struct buffer_info *buf)
                void *tmp;
                /* commit_seq is the offset in the buffer of the end of the last sequential commit.
                 * Bytes beyond this limit cannot be recovered. This is a free-running counter. */
-               long commit_seq = local_read(&ltt_buf->commit_seq[i_subbuf]);
+               long commit_seq = local_read(&ustbuf->commit_seq[i_subbuf]);
 
                unsigned long valid_length = buf->subbuf_size;
                long n_subbufs_order = get_count_order(buf->n_subbufs);
index 5a4569519b78138d3ed7c29fc985a58b7b8c1b33..26f87a2ba3f991f0e7a5acba10aa8a6d3b539b17 100644 (file)
@@ -698,7 +698,6 @@ int start_ustd_daemon()
        }
        else {
                char buf;
-               FILE *pidfp;
 
                result = read(fd[0], &buf, 1);
                if(result == -1) {
This page took 0.08295 seconds and 4 git commands to generate.