Initial import of kconsumerd and libkernelctl
authorJulien Desfossez <ju@klipix.org>
Fri, 6 May 2011 17:45:00 +0000 (13:45 -0400)
committerDavid Goulet <david.goulet@polymtl.ca>
Wed, 18 May 2011 14:21:16 +0000 (10:21 -0400)
Import the new kernel consumer along with the kernel tracing control
library.

kernel-ioctl.h is taken from the lttng-modules tree.

Missing:
- Send back error to session daemon on splice error
- kconsumerd clean tear down on error for both threads

Signed-off-by: David Goulet <david.goulet@polymtl.ca>
.gitignore
Makefile.am
configure.ac
kconsumerd/Makefile.am [new file with mode: 0644]
kconsumerd/kconsumerd.c [new file with mode: 0644]
kconsumerd/kconsumerd.h [new file with mode: 0644]
libkernelctl/kernel-ioctl.h
libkernelctl/libkernelctl.c
libkernelctl/libkernelctl.h
liblttsessiondcomm/liblttsessiondcomm.c
liblttsessiondcomm/liblttsessiondcomm.h

index 82001ff50987e4105490f21f55f067ef0a7afc98..6b33473d1b67c3e8c93b583ddc0f2ce2cb130b34 100644 (file)
@@ -25,3 +25,4 @@ libtool
 
 ltt-sessiond/ltt-sessiond
 lttng/lttng
+kconsumerd/kconsumerd
index dce10f8363f39f34f49a14445e1b83d646e976ea..5c27eabe3558080b97c47e4cefd284d2bafb646b 100644 (file)
@@ -5,4 +5,5 @@ SUBDIRS = liblttsessiondcomm \
                  liblttngctl \
                  lttng \
                  ltt-sessiond \
+                 kconsumerd \
                  tests
index 43b2f3f73bee68ff29c6c69b65eaef883db525fd..5eb64a5f9cd47b2c1973587b9c6ca897d4c7c325 100644 (file)
@@ -51,6 +51,7 @@ AC_CONFIG_FILES([
        liblttsessiondcomm/Makefile
        ltt-sessiond/Makefile
        lttng/Makefile
+       kconsumerd/Makefile
        tests/Makefile
        tests/ltt-sessiond/Makefile
 ])
diff --git a/kconsumerd/Makefile.am b/kconsumerd/Makefile.am
new file mode 100644 (file)
index 0000000..f96a552
--- /dev/null
@@ -0,0 +1,9 @@
+AM_CPPFLAGS = -I$(top_srcdir)/include -I$(top_srcdir)/liblttsessiondcomm -I$(top_srcdir)/libkernelctl
+
+bin_PROGRAMS = kconsumerd
+
+kconsumerd_SOURCES = kconsumerd.c
+
+kconsumerd_LDADD = \
+                $(top_builddir)/liblttsessiondcomm/liblttsessiondcomm.la \
+                $(top_builddir)/libkernelctl/libkernelctl.la
diff --git a/kconsumerd/kconsumerd.c b/kconsumerd/kconsumerd.c
new file mode 100644 (file)
index 0000000..7e166b8
--- /dev/null
@@ -0,0 +1,861 @@
+/*
+ * Copyright (C) 2011 - Julien Desfossez <julien.desfossez@polymtl.ca>
+ *                      Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+ */
+
+#define _GNU_SOURCE
+#include <fcntl.h>
+#include <getopt.h>
+#include <grp.h>
+#include <limits.h>
+#include <pthread.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <urcu/list.h>
+#include <poll.h>
+#include <unistd.h>
+
+#include "lttngerr.h"
+#include "libkernelctl.h"
+#include "liblttsessiondcomm.h"
+#include "kconsumerd.h"
+
+/* Init the list of FDs */
+static struct ltt_kconsumerd_fd_list kconsumerd_fd_list = {
+       .head = CDS_LIST_HEAD_INIT(kconsumerd_fd_list.head),
+};
+
+/* Number of element for the list below. */
+static unsigned int fds_count;
+
+/* If the local array of FDs needs update in the poll function */
+static unsigned int update_fd_array = 1;
+
+/* lock the fd array and structures */
+static pthread_mutex_t kconsumerd_lock_fds;
+
+/* the two threads (receive fd and poll) */
+static pthread_t threads[2];
+
+/* communication with splice */
+static int thread_pipe[2];
+
+/* socket to communicate errors with sessiond */
+static int error_socket = -1;
+
+/* Argument variables */
+int opt_quiet;
+int opt_verbose;
+static int opt_daemon;
+static const char *progname;
+static char command_sock_path[PATH_MAX]; /* Global command socket path */
+static char error_sock_path[PATH_MAX]; /* Global error path */
+
+/*
+ *  cleanup
+ *
+ *  Cleanup the daemon's socket on exit
+ */
+static void cleanup()
+{
+       unlink(command_sock_path);
+}
+
+/* send_error
+ *
+ * send return code to ltt-sessiond
+ */
+static int send_error(enum lttcomm_return_code cmd)
+{
+       if (error_socket > 0) {
+               return lttcomm_send_unix_sock(error_socket, &cmd,
+                               sizeof(enum lttcomm_sessiond_command));
+       } else {
+               return 0;
+       }
+}
+
+/*
+ * cleanup_kconsumerd_fd
+ *
+ * Close the FDs and frees a ltt_kconsumerd_fd struct
+ */
+static void cleanup_kconsumerd_fd(struct ltt_kconsumerd_fd *lcf)
+{
+       if (lcf != NULL) {
+               close(lcf->out_fd);
+               close(lcf->consumerd_fd);
+               free(lcf);
+               lcf = NULL;
+       }
+}
+
+/*
+ * add_fd
+ *
+ * Add a fd to the global list protected by a mutex
+ */
+static int add_fd(struct lttcomm_kconsumerd_msg *buf, int consumerd_fd)
+{
+       struct ltt_kconsumerd_fd *tmp_fd;
+       int ret;
+
+       tmp_fd = malloc(sizeof(struct ltt_kconsumerd_fd));
+       tmp_fd->sessiond_fd = buf->fd;
+       tmp_fd->consumerd_fd = consumerd_fd;
+       tmp_fd->state = buf->state;
+       tmp_fd->max_sb_size = buf->max_sb_size;
+       strncpy(tmp_fd->path_name, buf->path_name, PATH_MAX);
+
+       /* Opening the tracefile in write mode */
+       DBG("Opening %s for writing", tmp_fd->path_name);
+       ret = open(tmp_fd->path_name,
+                       O_WRONLY|O_CREAT, S_IRWXU|S_IRWXG|S_IRWXO);
+       if (ret < 0) {
+               ERR("Opening %s", tmp_fd->path_name);
+               perror("open");
+               goto end;
+       }
+       tmp_fd->out_fd = ret;
+       tmp_fd->out_fd_offset = 0;
+
+       DBG("Adding %s (%d, %d, %d)", tmp_fd->path_name,
+                       tmp_fd->sessiond_fd, tmp_fd->consumerd_fd, tmp_fd->out_fd);
+
+       pthread_mutex_lock(&kconsumerd_lock_fds);
+       cds_list_add(&tmp_fd->list, &kconsumerd_fd_list.head);
+       fds_count++;
+       pthread_mutex_unlock(&kconsumerd_lock_fds);
+
+end:
+       return ret;
+}
+
+/*
+ * del_fd
+ *
+ * Remove a fd from the global list protected by a mutex
+ */
+static void del_fd(struct ltt_kconsumerd_fd *lcf)
+{
+       pthread_mutex_lock(&kconsumerd_lock_fds);
+       cds_list_del(&lcf->list);
+       if (fds_count > 0) {
+               fds_count--;
+               DBG("Removed ltt_kconsumerd_fd");
+               cleanup_kconsumerd_fd(lcf);
+       }
+       pthread_mutex_unlock(&kconsumerd_lock_fds);
+}
+
+/*
+ * close_outfds
+ *
+ * Close all fds in the previous fd_list
+ * Must be used with kconsumerd_lock_fds lock held
+ */
+static void close_outfds()
+{
+       struct ltt_kconsumerd_fd *iter;
+       cds_list_for_each_entry(iter, &kconsumerd_fd_list.head, list) {
+               del_fd(iter);
+       }
+}
+
+/*
+ *  sighandler
+ *
+ *  Signal handler for the daemon
+ */
+static void sighandler(int sig)
+{
+       /* unblock the threads */
+       pthread_cancel(threads[0]);
+       pthread_cancel(threads[1]);
+
+       close_outfds();
+       cleanup();
+
+       return;
+}
+
+/*
+ *  set_signal_handler
+ *
+ *  Setup signal handler for :
+ *      SIGINT, SIGTERM, SIGPIPE
+ */
+static int set_signal_handler(void)
+{
+       int ret = 0;
+       struct sigaction sa;
+       sigset_t sigset;
+
+       if ((ret = sigemptyset(&sigset)) < 0) {
+               perror("sigemptyset");
+               return ret;
+       }
+
+       sa.sa_handler = sighandler;
+       sa.sa_mask = sigset;
+       sa.sa_flags = 0;
+       if ((ret = sigaction(SIGTERM, &sa, NULL)) < 0) {
+               perror("sigaction");
+               return ret;
+       }
+
+       if ((ret = sigaction(SIGINT, &sa, NULL)) < 0) {
+               perror("sigaction");
+               return ret;
+       }
+
+       if ((ret = sigaction(SIGPIPE, &sa, NULL)) < 0) {
+               perror("sigaction");
+               return ret;
+       }
+
+       return ret;
+}
+
+/*
+ * on_read_subbuffer
+ *
+ * Splice the data from the ring buffer to the tracefile.
+ * Returns the number of bytes spliced
+ */
+static int on_read_subbuffer(struct ltt_kconsumerd_fd *kconsumerd_fd,
+               unsigned long len)
+{
+       long ret = 0;
+       loff_t offset = 0;
+       off_t orig_offset = kconsumerd_fd->out_fd_offset;
+       int fd = kconsumerd_fd->consumerd_fd;
+       int outfd = kconsumerd_fd->out_fd;
+
+       while (len > 0) {
+               DBG("splice chan to pipe offset %lu (fd : %d)",
+                               (unsigned long)offset, fd);
+               ret = splice(fd, &offset, thread_pipe[1], NULL, len,
+                               SPLICE_F_MOVE | SPLICE_F_MORE);
+               DBG("splice chan to pipe ret %ld", ret);
+               if (ret < 0) {
+                       perror("Error in relay splice");
+                       goto write_end;
+               }
+
+               ret = splice(thread_pipe[0], NULL, outfd, NULL, ret,
+                               SPLICE_F_MOVE | SPLICE_F_MORE);
+               DBG("splice pipe to file %ld", ret);
+               if (ret < 0) {
+                       perror("Error in file splice");
+                       goto write_end;
+               }
+               if (ret >= len) {
+                       len = 0;
+               }
+               /* This won't block, but will start writeout asynchronously */
+               sync_file_range(outfd, kconsumerd_fd->out_fd_offset, ret,
+                               SYNC_FILE_RANGE_WRITE);
+               kconsumerd_fd->out_fd_offset += ret;
+       }
+write_end:
+       /*
+        * This does a blocking write-and-wait on any page that belongs to the
+        * subbuffer prior to the one we just wrote.
+        * Don't care about error values, as these are just hints and ways to
+        * limit the amount of page cache used.
+        */
+       if (orig_offset >= kconsumerd_fd->max_sb_size) {
+               sync_file_range(outfd, orig_offset - kconsumerd_fd->max_sb_size,
+                               kconsumerd_fd->max_sb_size,
+                               SYNC_FILE_RANGE_WAIT_BEFORE
+                               | SYNC_FILE_RANGE_WRITE
+                               | SYNC_FILE_RANGE_WAIT_AFTER);
+               /*
+                * Give hints to the kernel about how we access the file:
+                * POSIX_FADV_DONTNEED : we won't re-access data in a near
+                * future after we write it.
+                * We need to call fadvise again after the file grows because
+                * the kernel does not seem to apply fadvise to non-existing
+                * parts of the file.
+                * Call fadvise _after_ having waited for the page writeback to
+                * complete because the dirty page writeback semantic is not
+                * well defined. So it can be expected to lead to lower
+                * throughput in streaming.
+                */
+               posix_fadvise(outfd, orig_offset - kconsumerd_fd->max_sb_size,
+                               kconsumerd_fd->max_sb_size, POSIX_FADV_DONTNEED);
+       }
+       return ret;
+}
+
+/*
+ * read_subbuffer
+ *
+ * Consume data on a file descriptor and write it on a trace file
+ */
+static int read_subbuffer(struct ltt_kconsumerd_fd *kconsumerd_fd)
+{
+       unsigned long len;
+       int err;
+       long ret = 0;
+       int infd = kconsumerd_fd->consumerd_fd;
+
+       DBG("In read_subbuffer");
+       /* Get the next subbuffer */
+       err = kernctl_get_next_subbuf(infd);
+       if (err != 0) {
+               ret = errno;
+               perror("Reserving sub buffer failed (everything is normal, "
+                               "it is due to concurrency)");
+               goto end;
+       }
+
+       /* read the whole subbuffer */
+       err = kernctl_get_padded_subbuf_size(infd, &len);
+       if (err != 0) {
+               ret = errno;
+               perror("Getting sub-buffer len failed.");
+               goto end;
+       }
+
+       /* splice the subbuffer to the tracefile */
+       ret = on_read_subbuffer(kconsumerd_fd, len);
+       if (ret < 0) {
+               /*
+                * display the error but continue processing to try
+                * to release the subbuffer
+                */
+               ERR("Error splicing to tracefile");
+       }
+
+       err = kernctl_put_next_subbuf(infd);
+       if (err != 0) {
+               ret = errno;
+               if (errno == EFAULT) {
+                       perror("Error in unreserving sub buffer\n");
+               } else if (errno == EIO) {
+                       /* Should never happen with newer LTTng versions */
+                       perror("Reader has been pushed by the writer, last sub-buffer corrupted.");
+               }
+               goto end;
+       }
+
+end:
+       return ret;
+}
+
+/*
+ * change_fd_state
+ *
+ * Update a fd according to what we just received
+ */
+static void change_fd_state(int sessiond_fd,
+               enum lttcomm_kconsumerd_fd_state state)
+{
+       struct ltt_kconsumerd_fd *iter;
+       cds_list_for_each_entry(iter, &kconsumerd_fd_list.head, list) {
+               if (iter->sessiond_fd == sessiond_fd) {
+                       iter->state = state;
+                       break;
+               }
+       }
+}
+
+/*
+ * consumerd_recv_fd
+ *
+ * Receives an array of file descriptors and the associated
+ * structures describing each fd (path name).
+ * Returns the size of received data
+ */
+static int consumerd_recv_fd(int sfd, int size,
+               enum lttcomm_consumerd_command cmd_type)
+{
+       struct msghdr msg;
+       struct iovec iov[1];
+       int ret, i;
+       struct cmsghdr *cmsg;
+       int nb_fd;
+       char tmp[CMSG_SPACE(size)];
+       struct lttcomm_kconsumerd_msg *buf;
+       /* the number of fds we are about to receive */
+       nb_fd = size/sizeof(struct lttcomm_kconsumerd_msg);
+
+       buf = malloc(size);
+
+       memset(&msg, 0, sizeof(msg));
+
+       /* Prepare to receive the structures */
+       iov[0].iov_base = buf;
+       iov[0].iov_len = size;
+       msg.msg_iov = iov;
+       msg.msg_iovlen = 1;
+
+       msg.msg_control = tmp;
+       msg.msg_controllen = sizeof(tmp);
+
+       DBG("Waiting to receive fds");
+       if ((ret = recvmsg(sfd, &msg, 0)) < 0) {
+               perror("recvmsg");
+       }
+       if (ret != size) {
+               ERR("Received only %d, expected %d", ret, size);
+               send_error(KCONSUMERD_ERROR_RECV_FD);
+               goto end;
+       }
+
+       cmsg = CMSG_FIRSTHDR(&msg);
+       if (!cmsg) {
+               ERR("Invalid control message header");
+               ret = -1;
+               send_error(KCONSUMERD_ERROR_RECV_FD);
+               goto end;
+       }
+
+       /* if we received fds */
+       if (cmsg->cmsg_level == SOL_SOCKET && cmsg->cmsg_type == SCM_RIGHTS) {
+               DBG("Receive : expecting %d fds", nb_fd);
+               for (i = 0; i < nb_fd; i++) {
+                       switch (cmd_type) {
+                               case LTTCOMM_ADD_STREAM:
+                                       DBG("add_fd %s (%d)", buf[i].path_name, ((int *)CMSG_DATA(cmsg))[i]);
+                                       ret = add_fd(&buf[i], ((int *)CMSG_DATA(cmsg))[i]);
+                                       if (ret < 0) {
+                                               send_error(KCONSUMERD_OUTFD_ERROR);
+                                               goto end;
+                                       }
+                                       break;
+                               case LTTCOMM_UPDATE_STREAM:
+                                       change_fd_state(buf[i].fd, buf[i].state);
+                                       break;
+                               default:
+                                       break;
+                       }
+               }
+               /* flag to tell the polling thread to update its fd array */
+               update_fd_array = 1;
+               send_error(KCONSUMERD_SUCCESS_RECV_FD);
+       } else {
+               ERR("Didn't received any fd");
+               send_error(KCONSUMERD_ERROR_RECV_FD);
+               ret = -1;
+               goto end;
+       }
+
+end:
+       if (buf != NULL) {
+               free(buf);
+               buf = NULL;
+       }
+       return ret;
+}
+
+/*
+ *  thread_receive_fds
+ *
+ *  This thread listens on the consumerd socket and
+ *  receives the file descriptors from ltt-sessiond
+ */
+static void *thread_receive_fds(void *data)
+{
+       int sock, client_socket, ret;
+       struct lttcomm_kconsumerd_header tmp;
+
+       DBG("Creating command socket %s", command_sock_path);
+       unlink(command_sock_path);
+       client_socket = lttcomm_create_unix_sock(command_sock_path);
+       if (client_socket < 0) {
+               ERR("Cannot create command socket");
+               goto error;
+       }
+
+       ret = lttcomm_listen_unix_sock(client_socket);
+       if (ret < 0) {
+               goto error;
+       }
+
+       DBG("Sending ready command to ltt-sessiond");
+       ret = send_error(KCONSUMERD_COMMAND_SOCK_READY);
+       if (ret < 0) {
+               ERR("Error sending ready command to ltt-sessiond");
+               goto error;
+       }
+
+       while (1) {
+               /* Blocking call, waiting for transmission */
+               sock = lttcomm_accept_unix_sock(client_socket);
+               if (sock <= 0) {
+                       continue;
+               }
+
+               /* We first get the number of fd we are about to receive */
+               ret = lttcomm_recv_unix_sock(sock, &tmp,
+                               sizeof(struct lttcomm_kconsumerd_header));
+               if (ret < 0) {
+                       ERR("Receiving the lttcomm_kconsumerd_header");
+                       continue;
+               }
+               ret = consumerd_recv_fd(sock, tmp.payload_size, tmp.cmd_type);
+               if (ret < 0) {
+                       continue;
+               }
+       }
+
+error:
+       return NULL;
+}
+
+/*
+ * update_poll_array
+ *
+ * Allocate the pollfd structure and the local view of the out fds
+ * to avoid doing a lookup in the linked list and concurrency issues
+ * when writing is needed.
+ * Returns the number of fds in the structures
+ */
+static int update_poll_array(struct pollfd **pollfd,
+               struct ltt_kconsumerd_fd **local_kconsumerd_fd)
+{
+       struct ltt_kconsumerd_fd *iter;
+       int i = 0;
+
+       if (*pollfd != NULL) {
+               free(*pollfd);
+               *pollfd = NULL;
+       }
+
+       if (*local_kconsumerd_fd != NULL) {
+               free(*local_kconsumerd_fd);
+               *local_kconsumerd_fd = NULL;
+       }
+
+       *pollfd = malloc(fds_count * sizeof(struct pollfd));
+       if (*pollfd == NULL) {
+               perror("pollfd malloc");
+               goto error_mem;
+       }
+
+       *local_kconsumerd_fd = malloc(fds_count * sizeof(struct ltt_kconsumerd_fd));
+       if (*local_kconsumerd_fd == NULL) {
+               perror("local_kconsumerd_fd malloc");
+               goto error_mem;
+       }
+
+       DBG("Updating poll fd array");
+       pthread_mutex_lock(&kconsumerd_lock_fds);
+
+       cds_list_for_each_entry(iter, &kconsumerd_fd_list.head, list) {
+               DBG("Inside for each");
+               if (iter->state == ACTIVE_FD) {
+                       DBG("Active FD %d", iter->consumerd_fd);
+                       pollfd[i]->fd = iter->consumerd_fd;
+                       pollfd[i]->events = POLLIN | POLLPRI;
+                       local_kconsumerd_fd[i] = iter;
+                       i++;
+               } else if (iter->state == DELETE_FD) {
+                       del_fd(iter);
+               }
+       }
+       update_fd_array = 0;
+       pthread_mutex_unlock(&kconsumerd_lock_fds);
+       return i;
+
+error_mem:
+       return -ENOMEM;
+}
+
+/*
+ *  thread_poll_fds
+ *
+ *  This thread polls the fds in the ltt_fd_list to consume the data
+ *  and write it to tracefile if necessary.
+ */
+static void *thread_poll_fds(void *data)
+{
+       int num_rdy, num_hup, high_prio, ret, i;
+       struct pollfd *pollfd = NULL;
+       /* local view of the fds */
+       struct ltt_kconsumerd_fd *local_kconsumerd_fd = NULL;
+       /* local view of fds_count */
+       int nb_fd = 0;
+
+       ret = pipe(thread_pipe);
+       if (ret < 0) {
+               perror("Error creating pipe");
+               goto end;
+       }
+
+       while (1) {
+               high_prio = 0;
+               num_hup = 0;
+
+               /*
+                * the ltt_fd_list has been updated, we need to update our
+                * local array as well
+                */
+               if (update_fd_array) {
+                       ret = update_poll_array(&pollfd, &local_kconsumerd_fd);
+                       if (ret < 0) {
+                               ERR("Error in allocating pollfd or local_outfds");
+                               send_error(KCONSUMERD_POLL_ERROR);
+                               goto end;
+                       }
+                       nb_fd = ret;
+               }
+
+               /* poll on the array of fds */
+               DBG("polling on %d fd", nb_fd);
+               num_rdy = poll(pollfd, nb_fd, POLL_TIMEOUT);
+               DBG("poll num_rdy : %d", num_rdy);
+               if (num_rdy == -1) {
+                       perror("Poll error");
+                       send_error(KCONSUMERD_POLL_ERROR);
+                       goto end;
+               }
+
+               /* Take care of high priority channels first. */
+               for (i = 0; i < nb_fd; i++) {
+                       switch(pollfd[i].revents) {
+                               case POLLERR:
+                                       ERR("Error returned in polling fd %d.", pollfd[i].fd);
+                                       num_hup++;
+                                       send_error(KCONSUMERD_POLL_ERROR);
+                                       break;
+                               case POLLHUP:
+                                       ERR("Polling fd %d tells it has hung up.", pollfd[i].fd);
+                                       num_hup++;
+                                       break;
+                               case POLLNVAL:
+                                       ERR("Polling fd %d tells fd is not open.", pollfd[i].fd);
+                                       send_error(KCONSUMERD_POLL_NVAL);
+                                       num_hup++;
+                                       break;
+                               case POLLPRI:
+                                       DBG("Urgent read on fd %d", pollfd[i].fd);
+                                       high_prio = 1;
+                                       ret = read_subbuffer(&local_kconsumerd_fd[i]);
+                                       /* it's ok to have an unavailable sub-buffer (FIXME : is it ?) */
+                                       if (ret == EAGAIN) {
+                                               ret = 0;
+                                       }
+                                       break;
+                       }
+               }
+
+               /* If every buffer FD has hung up, we end the read loop here */
+               if (nb_fd > 0 && num_hup == nb_fd) {
+                       DBG("every buffer FD has hung up\n");
+                       send_error(KCONSUMERD_POLL_HUP);
+                       continue;
+               }
+
+               /* Take care of low priority channels. */
+               if (!high_prio) {
+                       for (i = 0; i < nb_fd; i++) {
+                               switch(pollfd[i].revents) {
+                                       case POLLIN:
+                                               DBG("Normal read on fd %d", pollfd[i].fd);
+                                               ret = read_subbuffer(&local_kconsumerd_fd[i]);
+                                               /* it's ok to have an unavailable subbuffer (FIXME : is it ?) */
+                                               if (ret == EAGAIN) {
+                                                       ret = 0;
+                                               }
+                                               break;
+                               }
+                       }
+               }
+       }
+end:
+       if (pollfd != NULL) {
+               free(pollfd);
+               pollfd = NULL;
+       }
+       if (local_kconsumerd_fd != NULL) {
+               free(local_kconsumerd_fd);
+               local_kconsumerd_fd = NULL;
+       }
+       return NULL;
+}
+
+/*
+ * usage function on stderr
+ */
+static void usage(void)
+{
+       fprintf(stderr, "Usage: %s OPTIONS\n\nOptions:\n", progname);
+       fprintf(stderr, "  -h, --help                         "
+                       "Display this usage.\n");
+       fprintf(stderr, "  -c, --kconsumerd-cmd-sock PATH     "
+                       "Specify path for the command socket\n");
+       fprintf(stderr, "  -e, --kconsumerd-err-sock PATH     "
+                       "Specify path for the error socket\n");
+       fprintf(stderr, "  -d, --daemonize                    "
+                       "Start as a daemon.\n");
+       fprintf(stderr, "  -q, --quiet                        "
+                       "No output at all.\n");
+       fprintf(stderr, "  -v, --verbose                      "
+                       "Verbose mode. Activate DBG() macro.\n");
+       fprintf(stderr, "  -V, --version                      "
+                       "Show version number.\n");
+}
+
+/*
+ * daemon argument parsing
+ */
+static void parse_args(int argc, char **argv)
+{
+       int c;
+
+       static struct option long_options[] = {
+               { "kconsumerd-cmd-sock", 1, 0, 'c' },
+               { "kconsumerd-err-sock", 1, 0, 'e' },
+               { "daemonize", 0, 0, 'd' },
+               { "help", 0, 0, 'h' },
+               { "quiet", 0, 0, 'q' },
+               { "verbose", 0, 0, 'v' },
+               { "version", 0, 0, 'V' },
+               { NULL, 0, 0, 0 }
+       };
+
+       while (1) {
+               int option_index = 0;
+               c = getopt_long(argc, argv, "dhqvV" "c:e:", long_options, &option_index);
+               if (c == -1) {
+                       break;
+               }
+
+               switch (c) {
+                       case 0:
+                               fprintf(stderr, "option %s", long_options[option_index].name);
+                               if (optarg) {
+                                       fprintf(stderr, " with arg %s\n", optarg);
+                               }
+                               break;
+                       case 'c':
+                               snprintf(command_sock_path, PATH_MAX, "%s", optarg);
+                               break;
+                       case 'e':
+                               snprintf(error_sock_path, PATH_MAX, "%s", optarg);
+                               break;
+                       case 'd':
+                               opt_daemon = 1;
+                               break;
+                       case 'h':
+                               usage();
+                               exit(EXIT_FAILURE);
+                       case 'q':
+                               opt_quiet = 1;
+                               break;
+                       case 'v':
+                               opt_verbose = 1;
+                               break;
+                       case 'V':
+                               fprintf(stdout, "%s\n", VERSION);
+                               exit(EXIT_SUCCESS);
+                       default:
+                               usage();
+                               exit(EXIT_FAILURE);
+               }
+       }
+}
+
+
+/*
+ * main
+ */
+int main(int argc, char **argv)
+{
+       int i;
+       int ret = 0;
+       void *status;
+
+       /* Parse arguments */
+       progname = argv[0];
+       parse_args(argc, argv);
+
+       /* Daemonize */
+       if (opt_daemon) {
+               ret = daemon(0, 0);
+               if (ret < 0) {
+                       perror("daemon");
+                       goto error;
+               }
+       }
+
+       if (strlen(command_sock_path) == 0) {
+               snprintf(command_sock_path, PATH_MAX,
+                               KCONSUMERD_CMD_SOCK_PATH);
+       }
+       if (strlen(error_sock_path) == 0) {
+               snprintf(error_sock_path, PATH_MAX,
+                               KCONSUMERD_ERR_SOCK_PATH);
+       }
+
+       if (set_signal_handler() < 0) {
+               goto error;
+       }
+
+       /* Connect to the socket created by ltt-sessiond to report errors */
+       DBG("Connecting to error socket %s", error_sock_path);
+       error_socket = lttcomm_connect_unix_sock(error_sock_path);
+       /* not a fatal error, but all communication with ltt-sessiond will fail */
+       if (error_socket < 0) {
+               WARN("Cannot connect to error socket, is ltt-sessiond started ?");
+       }
+
+       /* Create the thread to manage the receive of fd */
+       ret = pthread_create(&threads[0], NULL, thread_receive_fds, (void *) NULL);
+       if (ret != 0) {
+               perror("pthread_create");
+               goto error;
+       }
+
+       /* Create thread to manage the polling/writing of traces */
+       ret = pthread_create(&threads[1], NULL, thread_poll_fds, (void *) NULL);
+       if (ret != 0) {
+               perror("pthread_create");
+               goto error;
+       }
+
+       for (i = 0; i < 2; i++) {
+               ret = pthread_join(threads[i], &status);
+               if (ret != 0) {
+                       perror("pthread_join");
+                       goto error;
+               }
+       }
+       ret = EXIT_SUCCESS;
+       send_error(KCONSUMERD_EXIT_SUCCESS);
+       goto end;
+
+error:
+       ret = EXIT_FAILURE;
+       send_error(KCONSUMERD_EXIT_FAILURE);
+
+end:
+       cleanup();
+
+       return ret;
+}
diff --git a/kconsumerd/kconsumerd.h b/kconsumerd/kconsumerd.h
new file mode 100644 (file)
index 0000000..f5ec288
--- /dev/null
@@ -0,0 +1,44 @@
+/*
+ * Copyright (C) 2011 - Julien Desfossez <julien.desfossez@polymtl.ca>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+ */
+
+#ifndef _LTT_KCONSUMERD_H
+#define _LTT_KCONSUMERD_H
+
+/* timeout of 1s on poll to ensure the fd list is updated when needed */
+#define POLL_TIMEOUT 1000
+
+struct ltt_kconsumerd_fd_list {
+       struct cds_list_head head;
+};
+
+/*
+ * Internal representation of the FDs,
+ * sessiond_fd is used to identify uniquely a fd
+ */
+struct ltt_kconsumerd_fd {
+       struct cds_list_head list;
+       int sessiond_fd; /* used to identify uniquely a fd with sessiond */
+       int consumerd_fd; /* fd to consume */
+       int out_fd; /* output file to write the data */
+       off_t out_fd_offset; /* write position in the output file descriptor */
+       char path_name[PATH_MAX]; /* tracefile name */
+       enum lttcomm_kconsumerd_fd_state state;
+       unsigned long max_sb_size; /* the subbuffer size for this channel */
+};
+
+#endif /* _LTT_KCONSUMERD_H */
index bed6f593449f2e1d34e80f5cc17fd2af36c076ec..9aa0407b4c2fa36a6aac36847c6a9ecbc90032ca 100644 (file)
 #define RING_BUFFER_GET_MMAP_READ_OFFSET    _IOR(0xF6, 0x0B, unsigned long)
 
 /* Create session LTTNG_SESSION */
-#define KERNEL_IO_CREATE_SESSION       _IO(0xF6, 0x40)
-#define KERNEL_IO_SESSION_START _IO(0xF6, 0x41)
-#define KERNEL_IO_SESSION_STOP  _IO(0xF6, 0x42)
+#define KERNEL_CREATE_SESSION       _IO(0xF6, 0x40)
+#define KERNEL_SESSION_START _IO(0xF6, 0x41)
+#define KERNEL_SESSION_STOP  _IO(0xF6, 0x42)
 
 /* Create channel LTTNG_CHANNEL */
-#define KERNEL_IO_CREATE_CHANNEL       _IOW(0xF6, 0x43, struct lttng_channel)
-#define KERNEL_IO_CREATE_STREAM        _IO(0xF6, 0x44)
-#define KERNEL_IO_CREATE_EVENT     _IOW(0xF6, 0x45, struct lttng_event)
+#define KERNEL_CREATE_CHANNEL       _IOW(0xF6, 0x43, struct lttng_channel)
+#define KERNEL_CREATE_STREAM        _IO(0xF6, 0x44)
+#define KERNEL_CREATE_EVENT     _IOW(0xF6, 0x45, struct lttng_event)
+
+#define KERNEL_OPEN_METADATA   _IOW(0xF6, 0x46, struct lttng_channel)
 
 #endif /* _LTT_KERNEL_IOCTL_H */
index e99e777c3ab0b25221e421db9aaca96cc8508cbc..e15dd8da624b9489adf256015950a4b66abeeb9d 100644 (file)
 
 int kernctl_create_channel(int fd, struct lttng_channel *chops)
 {
-       return ioctl(fd, KERNEL_IO_CREATE_CHANNEL, chops);
+       return ioctl(fd, KERNEL_CREATE_CHANNEL, chops);
 }
 
 int kernctl_create_event(int fd, struct lttng_event *ev)
 {
-       return ioctl(fd, KERNEL_IO_CREATE_EVENT, ev);
+       return ioctl(fd, KERNEL_CREATE_EVENT, ev);
 }
 
 int kernctl_create_session(int fd)
 {
-       return ioctl(fd, KERNEL_IO_CREATE_SESSION);
+       return ioctl(fd, KERNEL_CREATE_SESSION);
 }
 
 int kernctl_create_stream(int fd)
 {
-       return ioctl(fd, KERNEL_IO_CREATE_STREAM);
+       return ioctl(fd, KERNEL_CREATE_STREAM);
 }
 
 /* returns the maximum size for sub-buffers. */
@@ -85,6 +85,12 @@ int kernctl_get_subbuf_size(int fd, unsigned long *len)
        return ioctl(fd, RING_BUFFER_GET_SUBBUF_SIZE, len);
 }
 
+/* open the metadata global channel */
+int kernctl_open_metadata(int fd, struct lttng_channel *chops)
+{
+       return ioctl(fd, KERNEL_OPEN_METADATA, chops);
+}
+
 /* Release exclusive sub-buffer access, move consumer forward. */
 int kernctl_put_next_subbuf(int fd)
 {
@@ -117,11 +123,11 @@ int kernctl_snapshot_get_produced(int fd, unsigned long *pos)
 
 int kernctl_start_session(int fd)
 {
-       return ioctl(fd, KERNEL_IO_SESSION_START);
+       return ioctl(fd, KERNEL_SESSION_START);
 }
 
 int kernctl_stop_session(int fd)
 {
-       return ioctl(fd, KERNEL_IO_SESSION_STOP);
+       return ioctl(fd, KERNEL_SESSION_STOP);
 }
 
index 2c03e1088c31cb32184afa843fa7eec17d5ed3f2..eb0c7800efcad3c10949dcf70247f1149681d537 100644 (file)
@@ -33,6 +33,7 @@ int kernctl_get_next_subbuf(int fd);
 int kernctl_get_padded_subbuf_size(int fd, unsigned long *len);
 int kernctl_get_subbuf(int fd, unsigned long *len);
 int kernctl_get_subbuf_size(int fd, unsigned long *len);
+int kernctl_open_metadata(int fd, struct lttng_channel *chops);
 int kernctl_put_next_subbuf(int fd);
 int kernctl_put_subbuf(int fd);
 int kernctl_snapshot(int fd);
index 2c46f0027d58077f1035342ae3fa14fe55de92d6..785d8e827b4dc456703ab8950b49de187d635aca 100644 (file)
@@ -48,6 +48,15 @@ static const char *lttcomm_readable_code[] = {
        [ LTTCOMM_ERR_INDEX(LTTCOMM_NO_TRACEABLE) ] = "App is not traceable",
        [ LTTCOMM_ERR_INDEX(LTTCOMM_SELECT_SESS) ] = "A session MUST be selected",
        [ LTTCOMM_ERR_INDEX(LTTCOMM_EXIST_SESS) ] = "Session name already exist",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_COMMAND_SOCK_READY) ] = "Kconsumerd command socket ready",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_SUCCESS_RECV_FD) ] = "Kconsumerd success on receiving fds",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_ERROR_RECV_FD) ] = "Kconsumerd error on receiving fds",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_POLL_ERROR) ] = "Kconsumerd error in polling thread",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_POLL_NVAL) ] = "Kconsumerd polling on closed fd",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_POLL_HUP) ] = "Kconsumerd all fd hung up",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_EXIT_SUCCESS) ] = "Kconsumerd exiting normally",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_EXIT_FAILURE) ] = "Kconsumerd exiting on error",
+       [ LTTCOMM_ERR_INDEX(KCONSUMERD_OUTFD_ERROR) ] = "Kconsumerd error opening the tracefile",
 };
 
 /*
index 4a7fdfcc48155e44c9937dd471649e38bbec5e3e..8b505eb980f3ea6782ec0cef60083c2f2549f4d0 100644 (file)
@@ -94,6 +94,15 @@ enum lttcomm_return_code {
        LTTCOMM_SELECT_SESS,    /* Must select a session */
        LTTCOMM_EXIST_SESS,             /* Session name already exist */
        LTTCOMM_NR,                             /* Last element */
+       KCONSUMERD_COMMAND_SOCK_READY,  /* when kconsumerd command socket ready */
+       KCONSUMERD_SUCCESS_RECV_FD,     /* success on receiving fds */
+       KCONSUMERD_ERROR_RECV_FD,       /* error on receiving fds */
+       KCONSUMERD_POLL_ERROR,  /* Error in polling thread in kconsumerd */
+       KCONSUMERD_POLL_NVAL,   /* Poll on closed fd */
+       KCONSUMERD_POLL_HUP,    /* All fds have hungup */
+       KCONSUMERD_EXIT_SUCCESS,        /* kconsumerd exiting normally */
+       KCONSUMERD_EXIT_FAILURE,        /* kconsumerd exiting on error */
+       KCONSUMERD_OUTFD_ERROR, /* error opening the tracefile */
 };
 
 /* commands for kconsumerd */
@@ -178,6 +187,7 @@ struct lttcomm_kconsumerd_msg {
        char path_name[PATH_MAX];
        int fd;
        u32 state;    /* enum lttcomm_kconsumerd_fd_state */
+       unsigned long max_sb_size; /* the subbuffer size for this channel */
 };
 
 extern int lttcomm_create_unix_sock(const char *pathname);
This page took 0.040054 seconds and 4 git commands to generate.