* [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints
@ 2016-07-19 20:43 Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list Sven Eckelmann
` (3 more replies)
0 siblings, 4 replies; 7+ messages in thread
From: Sven Eckelmann @ 2016-07-19 20:43 UTC (permalink / raw)
To: b.a.t.m.a.n
[-- Attachment #1: Type: text/plain, Size: 1806 bytes --]
Hi,
The current endpoint for batadv_icmp* is implemented in the kernel module and
can be accessed via debugfs. But the debugfs cannot be accessed from non-
default netns or when debugfs is disabled. Thus it has be possible to use it
via the netlink infrastructure to make it compatible with future
setups.
The use of the socket file is completely removed and instead raw sockets with
BPF filters are used to send/receive batadv_icmp_packet* directly. All
information about interfaces and available originators are received via
rtnetlink and the batman-adv netlink.
The originators debugfs file is used when the batman-adv netlink commands are
not available. The routing of batadv_icmp_packets is still done inside the
kernel module.
The patchset is based on the netlink and rtnl patchset for batctl.
Kind regards,
Sven
Sven Eckelmann (4):
batctl: Replace list implementation with linux-like-list
batctl: Use monotonic time source for icmp timing
batctl: Add helper to generate instant random bytes
batctl: Implement non-routing batadv_icmp in userspace
Makefile | 4 +-
bisect_iv.c | 40 ++-
bisect_iv.h | 10 +-
functions.c | 82 +++++-
functions.h | 2 +
icmp_helper.c | 633 ++++++++++++++++++++++++++++++++++++++++++++
icmp_helper.h | 58 ++++
list-batman.c | 123 ---------
list-batman.h | 120 ---------
list.h | 834 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++
netlink.c | 178 ++++++++++++-
netlink.h | 3 +
ping.c | 42 +--
tcpdump.c | 7 +-
tcpdump.h | 2 +-
traceroute.c | 42 +--
16 files changed, 1832 insertions(+), 348 deletions(-)
create mode 100644 icmp_helper.c
create mode 100644 icmp_helper.h
delete mode 100644 list-batman.c
delete mode 100644 list-batman.h
create mode 100644 list.h
[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread
* [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list
2016-07-19 20:43 [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints Sven Eckelmann
@ 2016-07-19 20:44 ` Sven Eckelmann
2016-10-18 12:24 ` [B.A.T.M.A.N.] [1/4] " Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 2/4] batctl: Use monotonic time source for icmp timing Sven Eckelmann
` (2 subsequent siblings)
3 siblings, 1 reply; 7+ messages in thread
From: Sven Eckelmann @ 2016-07-19 20:44 UTC (permalink / raw)
To: b.a.t.m.a.n
The list-batman implementation of linked lists has an interface which tries
to resample the linux double linked list implementation but in reality is
incompatible. This resulted in code which moves most of the pointer
arithmetic back to the actual code. This is rather unfortunate because this
abstraction should reduce the pointer arithmetic and not increase it.
Also the incompatibilities between the list implementation used in
batman-adv and the one used by batctl made coding features in both
components error prone because each component required a slightly different
implementation for the same list manipulation.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
Makefile | 1 -
bisect_iv.c | 40 ++-
bisect_iv.h | 10 +-
list-batman.c | 123 ---------
list-batman.h | 120 ---------
list.h | 834 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
tcpdump.c | 7 +-
tcpdump.h | 2 +-
8 files changed, 860 insertions(+), 277 deletions(-)
delete mode 100644 list-batman.c
delete mode 100644 list-batman.h
create mode 100644 list.h
diff --git a/Makefile b/Makefile
index 5dc8fe8..e980587 100755
--- a/Makefile
+++ b/Makefile
@@ -32,7 +32,6 @@ OBJ += genl.o
OBJ += hash.o
OBJ += interface.o
OBJ += ioctl.o
-OBJ += list-batman.o
OBJ += main.o
OBJ += netlink.o
OBJ += ping.o
diff --git a/bisect_iv.c b/bisect_iv.c
index 0f2c0cb..1cb79eb 100644
--- a/bisect_iv.c
+++ b/bisect_iv.c
@@ -92,8 +92,8 @@ static struct bat_node *node_get(char *name)
strncpy(bat_node->name, name, NAME_LEN);
bat_node->name[NAME_LEN - 1] = '\0';
- INIT_LIST_HEAD_FIRST(bat_node->orig_event_list);
- INIT_LIST_HEAD_FIRST(bat_node->rt_table_list);
+ INIT_LIST_HEAD(&bat_node->orig_event_list);
+ INIT_LIST_HEAD(&bat_node->rt_table_list);
memset(bat_node->loop_magic, 0, sizeof(bat_node->loop_magic));
memset(bat_node->loop_magic2, 0, sizeof(bat_node->loop_magic2));
hash_add(node_hash, bat_node);
@@ -112,9 +112,8 @@ static struct orig_event *orig_event_new(struct bat_node *bat_node, struct bat_n
return NULL;
}
- INIT_LIST_HEAD(&orig_event->list);
- INIT_LIST_HEAD_FIRST(orig_event->event_list);
- INIT_LIST_HEAD_FIRST(orig_event->rt_hist_list);
+ INIT_LIST_HEAD(&orig_event->event_list);
+ INIT_LIST_HEAD(&orig_event->rt_hist_list);
orig_event->orig_node = orig_node;
list_add_tail(&orig_event->list, &bat_node->orig_event_list);
@@ -166,21 +165,21 @@ static void node_free(void *data)
list_for_each_entry_safe(orig_event, orig_event_tmp, &bat_node->orig_event_list, list) {
list_for_each_entry_safe(seqno_event, seqno_event_tmp, &orig_event->event_list, list) {
- list_del((struct list_head *)&orig_event->event_list, &seqno_event->list, &orig_event->event_list);
+ list_del(&seqno_event->list);
free(seqno_event);
}
list_for_each_entry_safe(rt_hist, rt_hist_tmp, &orig_event->rt_hist_list, list) {
- list_del((struct list_head *)&orig_event->rt_hist_list, &rt_hist->list, &orig_event->rt_hist_list);
+ list_del(&rt_hist->list);
free(rt_hist);
}
- list_del((struct list_head *)&bat_node->orig_event_list, &orig_event->list, &bat_node->orig_event_list);
+ list_del(&orig_event->list);
free(orig_event);
}
list_for_each_entry_safe(rt_table, rt_table_tmp, &bat_node->rt_table_list, list) {
- list_del((struct list_head *)&bat_node->rt_table_list, &rt_table->list, &bat_node->rt_table_list);
+ list_del(&rt_table->list);
free(rt_table->entries);
free(rt_table);
@@ -248,10 +247,8 @@ static int routing_table_new(char *orig, char *next_hop, char *old_next_hop, cha
goto table_free;
}
- INIT_LIST_HEAD(&rt_table->list);
rt_table->num_entries = 1;
- INIT_LIST_HEAD(&rt_hist->list);
rt_hist->prev_rt_hist = NULL;
rt_hist->next_hop = next_hop_node;
rt_hist->flags = rt_flag;
@@ -310,7 +307,6 @@ static int routing_table_new(char *orig, char *next_hop, char *old_next_hop, cha
goto rt_hist_free;
}
- INIT_LIST_HEAD(&seqno_event->list);
seqno_event->orig = node_get(orig);
seqno_event->neigh = NULL;
seqno_event->prev_sender = NULL;
@@ -449,7 +445,6 @@ static int seqno_event_new(char *iface_addr, char *orig, char *prev_sender, char
goto err;
}
- INIT_LIST_HEAD(&seqno_event->list);
seqno_event->orig = orig_node;
seqno_event->neigh = neigh_node;
seqno_event->prev_sender = prev_sender_node;
@@ -986,7 +981,7 @@ static void seqno_trace_print_neigh(struct seqno_trace_neigh *seqno_trace_neigh,
}
}
-static void seqno_trace_print(struct list_head_first *trace_list, char *trace_orig,
+static void seqno_trace_print(struct list_head *trace_list, char *trace_orig,
long long seqno_min, long long seqno_max, char *filter_orig, int read_opt)
{
struct seqno_trace *seqno_trace;
@@ -1180,7 +1175,6 @@ static struct seqno_trace *seqno_trace_new(struct seqno_event *seqno_event)
return NULL;
}
- INIT_LIST_HEAD(&seqno_trace->list);
seqno_trace->seqno = seqno_event->seqno;
seqno_trace->print = 0;
@@ -1199,7 +1193,7 @@ static void seqno_trace_free(struct seqno_trace *seqno_trace)
free(seqno_trace);
}
-static int seqno_trace_add(struct list_head_first *trace_list, struct bat_node *bat_node,
+static int seqno_trace_add(struct list_head *trace_list, struct bat_node *bat_node,
struct seqno_event *seqno_event, char print_trace)
{
struct seqno_trace *seqno_trace = NULL, *seqno_trace_tmp = NULL, *seqno_trace_prev = NULL;
@@ -1223,12 +1217,12 @@ static int seqno_trace_add(struct list_head_first *trace_list, struct bat_node *
goto err;
if ((list_empty(trace_list)) ||
- (seqno_event->seqno > ((struct seqno_trace *)trace_list->prev)->seqno))
+ (seqno_event->seqno > list_last_entry(trace_list, struct seqno_trace, list)->seqno))
list_add_tail(&seqno_trace->list, trace_list);
- else if (seqno_event->seqno < ((struct seqno_trace *)trace_list->next)->seqno)
- list_add_before((struct list_head *)trace_list, trace_list->next, &seqno_trace->list);
+ else if (seqno_event->seqno < list_first_entry(trace_list, struct seqno_trace, list)->seqno)
+ list_add(&seqno_trace->list, trace_list);
else
- list_add_before(&seqno_trace_prev->list, &seqno_trace_tmp->list, &seqno_trace->list);
+ list_add_behind(&seqno_trace->list, &seqno_trace_prev->list);
}
if (print_trace)
@@ -1260,14 +1254,14 @@ static void trace_seqnos(char *trace_orig, long long seqno_min, long long seqno_
struct orig_event *orig_event;
struct seqno_event *seqno_event;
struct hash_it_t *hashit = NULL;
- struct list_head_first trace_list;
+ struct list_head trace_list;
struct seqno_trace *seqno_trace, *seqno_trace_tmp;
char check_orig[NAME_LEN], print_trace;
int res;
/* if no option was given filter_orig is empty */
memset(check_orig, 0, NAME_LEN);
- INIT_LIST_HEAD_FIRST(trace_list);
+ INIT_LIST_HEAD(&trace_list);
while (NULL != (hashit = hash_iterate(node_hash, hashit))) {
bat_node = hashit->bucket->data;
@@ -1313,7 +1307,7 @@ static void trace_seqnos(char *trace_orig, long long seqno_min, long long seqno_
out:
list_for_each_entry_safe(seqno_trace, seqno_trace_tmp, &trace_list, list) {
- list_del((struct list_head *)&trace_list, &seqno_trace->list, &trace_list);
+ list_del(&seqno_trace->list);
seqno_trace_free(seqno_trace);
}
diff --git a/bisect_iv.h b/bisect_iv.h
index ace76c7..43aa80b 100644
--- a/bisect_iv.h
+++ b/bisect_iv.h
@@ -22,7 +22,7 @@
#ifndef _BATCTL_BISECT_IV_H
#define _BATCTL_BISECT_IV_H
-#include "list-batman.h"
+#include "list.h"
#define NAME_LEN 18
#define MAX_LINE 256
@@ -38,8 +38,8 @@ int bisect_iv(int argc, char **argv);
struct bat_node {
char name[NAME_LEN];
- struct list_head_first orig_event_list;
- struct list_head_first rt_table_list;
+ struct list_head orig_event_list;
+ struct list_head rt_table_list;
char loop_magic[LOOP_MAGIC_LEN];
char loop_magic2[LOOP_MAGIC_LEN];
};
@@ -47,8 +47,8 @@ struct bat_node {
struct orig_event {
struct list_head list;
struct bat_node *orig_node;
- struct list_head_first event_list;
- struct list_head_first rt_hist_list;
+ struct list_head event_list;
+ struct list_head rt_hist_list;
};
struct rt_table {
diff --git a/list-batman.c b/list-batman.c
deleted file mode 100644
index affedb6..0000000
--- a/list-batman.c
+++ /dev/null
@@ -1,123 +0,0 @@
-/*
- * Copyright (C) 2006-2016 B.A.T.M.A.N. contributors:
- *
- * Marek Lindner
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of version 2 of the GNU General Public
- * License as published by the Free Software Foundation.
- *
- * 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., 51 Franklin Street, Fifth Floor, Boston, MA
- * 02110-1301, USA
- *
- */
-
-
-
-#include "list-batman.h"
-
-
-
-/*
- * Insert a new entry between two known consecutive entries.
- *
- * This is only for internal list manipulation where we know
- * the next entries already!
- */
-static void __list_add( struct list_head *new, struct list_head *prev, struct list_head *next ) {
-
- new->next = next;
- prev->next = new;
-
-}
-
-/**
- * list_add - add a new entry
- * @new: new entry to be added
- * @head: list head to add it after
- *
- * Insert a new entry after the specified head.
- * This is good for implementing stacks.
- */
-void list_add( struct list_head *new, struct list_head_first *head ) {
-
- __list_add( new, (struct list_head *)head, head->next );
-
- if ( head->prev == (struct list_head *)head )
- head->prev = new;
-
-}
-
-/**
- * list_add_tail - add a new entry
- * @new: new entry to be added
- * @head: list head to add it before
- *
- * Insert a new entry before the specified head.
- * This is useful for implementing queues.
- */
-void list_add_tail( struct list_head *new, struct list_head_first *head ) {
-
- __list_add( new, head->prev, (struct list_head *)head );
-
- head->prev = new;
-
-}
-
-void list_add_before( struct list_head *prev_node, struct list_head *next_node, struct list_head *new_node ) {
-
- prev_node->next = new_node;
- new_node->next = next_node;
-
-}
-
-
-
-/*
- * Delete a list entry by making the next entries
- * point to each other.
- *
- * This is only for internal list manipulation where we know
- * the next entries already!
- */
-static void __list_del( struct list_head *prev, struct list_head *next ) {
-
- prev->next = next;
-
-}
-
-/**
- * list_del - deletes entry from list.
- * @entry: the element to delete from the list.
- * Note: list_empty on entry does not return true after this, the entry is in an undefined state.
- */
-void list_del( struct list_head *prev_entry, struct list_head *entry, struct list_head_first *head ) {
-
- if ( head->prev == entry )
- head->prev = prev_entry;
-
- __list_del( prev_entry, entry->next );
-
- entry->next = (void *) 0;
-
-}
-
-
-
-/**
- * list_empty - tests whether a list is empty
- * @head: the list to test.
- */
-int list_empty( struct list_head_first *head ) {
-
- return head->next == (struct list_head *)head;
-
-}
-
diff --git a/list-batman.h b/list-batman.h
deleted file mode 100644
index 53b8413..0000000
--- a/list-batman.h
+++ /dev/null
@@ -1,120 +0,0 @@
-/*
- * Copyright (C) 2006-2016 B.A.T.M.A.N. contributors:
- *
- * Marek Lindner
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of version 2 of the GNU General Public
- * License as published by the Free Software Foundation.
- *
- * 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., 51 Franklin Street, Fifth Floor, Boston, MA
- * 02110-1301, USA
- *
- */
-
-#ifndef _LINUX_LIST_H
-#define _LINUX_LIST_H
-
-/*
- * XXX: Resolve conflict between this file and <sys/queue.h> on BSD systems.
- */
-#ifdef LIST_HEAD
-#undef LIST_HEAD
-#endif
-
-/*
- * Simple linked list implementation.
- *
- * Some of the internal functions ("__xxx") are useful when
- * manipulating whole lists rather than single entries, as
- * sometimes we already know the next entries and we can
- * generate better code by using them directly rather than
- * using the generic single-entry routines.
- */
-
-struct list_head {
- struct list_head *next;
-};
-
-struct list_head_first {
- struct list_head *next, *prev;
-};
-
-
-void list_add( struct list_head *new, struct list_head_first *head );
-void list_add_tail( struct list_head *new, struct list_head_first *head );
-void list_add_before( struct list_head *prev_node, struct list_head *next_node, struct list_head *new_node );
-void list_del( struct list_head *prev_entry, struct list_head *entry, struct list_head_first *head );
-int list_empty( struct list_head_first *head );
-
-
-
-#define INIT_LIST_HEAD(ptr) do { \
- (ptr)->next = (ptr); \
-} while (0)
-
-#define INIT_LIST_HEAD_FIRST(ptr) \
- ptr.next = (struct list_head *)&ptr; ptr.prev = (struct list_head *)&ptr; \
-
-
-/**
- * list_entry - get the struct for this entry
- * @ptr: the &struct list_head pointer.
- * @type: the type of the struct this is embedded in.
- * @member: the name of the list_struct within the struct.
- */
-#define list_entry(ptr, type, member) \
- ((type *)((char *)(ptr)-(unsigned long)(&((type *)0)->member)))
-
-/**
- * list_for_each - iterate over a list
- * @pos: the &struct list_head to use as a loop counter.
- * @head: the head for your list.
- */
-#define list_for_each(pos, head) \
- for (pos = (head)->next; pos != (struct list_head *)(head); \
- pos = pos->next)
-
-/**
- * list_for_each_safe - iterate over a list safe against removal of list entry
- * @pos: the &struct list_head to use as a loop counter.
- * @n: another &struct list_head to use as temporary storage
- * @head: the head for your list.
- */
-#define list_for_each_safe(pos, n, head) \
- for (pos = (head)->next, n = pos->next; pos != (struct list_head *)(head); \
- pos = n, n = pos->next)
-
-/**
- * list_for_each_entry - iterate over list of given type
- * @pos: the type * to use as a loop counter.
- * @head: the head for your list.
- * @member: the name of the list_struct within the struct.
- */
-#define list_for_each_entry(pos, head, member) \
- for (pos = list_entry((head)->next, typeof(*pos), member); \
- &pos->member != (struct list_head *)(head); \
- pos = list_entry(pos->member.next, typeof(*pos), member))
-
-/**
- * list_for_each_entry_safe - iterate over list of given type safe against removal of list entry
- * @pos: the type * to use as a loop counter.
- * @n: another type * to use as temporary storage
- * @head: the head for your list.
- * @member: the name of the list_struct within the struct.
- */
-#define list_for_each_entry_safe(pos, n, head, member) \
- for (pos = list_entry((head)->next, typeof(*pos), member), \
- n = list_entry(pos->member.next, typeof(*pos), member); \
- &pos->member != (struct list_head *)(head); \
- pos = n, n = list_entry(n->member.next, typeof(*n), member))
-
-
-#endif
diff --git a/list.h b/list.h
new file mode 100644
index 0000000..41feaa0
--- /dev/null
+++ b/list.h
@@ -0,0 +1,834 @@
+/* Minimal Linux-like double-linked list helper functions
+ *
+ * Copyright (c) 2012-2016, Sven Eckelmann <sven@narfation.org>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef __LINUX_LIKE_LIST_H__
+#define __LINUX_LIKE_LIST_H__
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#include <stddef.h>
+
+#if defined(__GNUC__)
+#define LIST_TYPEOF_USE 1
+#endif
+
+#if defined(_MSC_VER)
+#define __inline__ __inline
+#endif
+
+/**
+ * container_of() - Calculate address of object that contains address ptr
+ * @ptr: pointer to member variable
+ * @type: type of the structure containing ptr
+ * @member: name of the member variable in struct @type
+ *
+ * Return: @type pointer of object containing ptr
+ */
+#ifndef container_of
+#ifdef LIST_TYPEOF_USE
+#define container_of(ptr, type, member) __extension__ ({ \
+ const __typeof__(((type *)0)->member) *__pmember = (ptr); \
+ (type *)((char *)__pmember - offsetof(type, member)); })
+#else
+#define container_of(ptr, type, member) \
+ ((type *)((char *)(ptr) - offsetof(type, member)))
+#endif
+#endif
+
+/**
+ * struct list_head - Head and node of a double-linked list
+ * @prev: pointer to the previous node in the list
+ * @next: pointer to the next node in the list
+ *
+ * The simple double-linked list consists of a head and nodes attached to
+ * this head. Both node and head share the same struct type. The list_*
+ * functions and macros can be used to access and modify this data structure.
+ *
+ * The @prev pointer of the list head points to the last list node of the
+ * list and @next points to the first list node of the list. For an empty list,
+ * both member variables point to the head.
+ *
+ * The list nodes are usually embedded in a container structure which holds the
+ * actual data. Such an container object is called entry. The helper list_entry
+ * can be used to calculate the object address from the address of the node.
+ */
+struct list_head {
+ struct list_head *prev;
+ struct list_head *next;
+};
+
+/**
+ * LIST_HEAD - Declare list head and initialize it
+ * @head: name of the new object
+ */
+#define LIST_HEAD(head) \
+ struct list_head head = { &(head), &(head) }
+
+/**
+ * INIT_LIST_HEAD() - Initialize empty list head
+ * @head: pointer to list head
+ *
+ * This can also be used to initialize a unlinked list node.
+ *
+ * A node is usually linked inside a list, will be added to a list in
+ * the near future or the entry containing the node will be free'd soon.
+ *
+ * But an unlinked node may be given to a function which uses list_del(_init)
+ * before it ends up in a previously mentioned state. The list_del(_init) on an
+ * initialized node is well defined and safe. But the result of a
+ * list_del(_init) on an uninitialized node is undefined (unrelated memory is
+ * modified, crashes, ...).
+ */
+static __inline__ void INIT_LIST_HEAD(struct list_head *head)
+{
+ head->next = head;
+ head->prev = head;
+}
+
+/**
+ * list_add() - Add a list node to the beginning of the list
+ * @node: pointer to the new node
+ * @head: pointer to the head of the list
+ */
+static __inline__ void list_add(struct list_head *node,
+ struct list_head *head)
+{
+ struct list_head *next = head->next;
+
+ next->prev = node;
+ node->next = next;
+ node->prev = head;
+ head->next = node;
+}
+
+/**
+ * list_add_tail() - Add a list node to the end of the list
+ * @node: pointer to the new node
+ * @head: pointer to the head of the list
+ */
+static __inline__ void list_add_tail(struct list_head *node,
+ struct list_head *head)
+{
+ struct list_head *prev = head->prev;
+
+ prev->next = node;
+ node->next = head;
+ node->prev = prev;
+ head->prev = node;
+}
+
+/**
+ * list_add_before() - Add a list node before another node to the list
+ * @new_node: pointer to the new node
+ * @node: pointer to the reference node in the list
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define list_add_before(new_node, node) \
+ list_add_tail(new_node, node)
+
+/**
+ * list_add_behind() - Add a list node behind another node to the list
+ * @new_node: pointer to the new node
+ * @node: pointer to the reference node in the list
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define list_add_behind(new_node, node) \
+ list_add(new_node, node)
+
+/**
+ * list_del() - Remove a list node from the list
+ * @node: pointer to the node
+ *
+ * The node is only removed from the list. Neither the memory of the removed
+ * node nor the memory of the entry containing the node is free'd. The node
+ * has to be handled like an uninitialized node. Accessing the next or prev
+ * pointer of the node is not safe.
+ *
+ * Unlinked, initialized nodes are also uninitialized after list_del.
+ *
+ * LIST_POISONING can be enabled during build-time to provoke an invalid memory
+ * access when the memory behind the next/prev pointer is used after a list_del.
+ * This only works on systems which prohibit access to the predefined memory
+ * addresses.
+ */
+static __inline__ void list_del(struct list_head *node)
+{
+ struct list_head *next = node->next;
+ struct list_head *prev = node->prev;
+
+ next->prev = prev;
+ prev->next = next;
+
+#ifdef LIST_POISONING
+ node->prev = (struct list_head *)(0x00100100);
+ node->next = (struct list_head *)(0x00200200);
+#endif
+}
+
+/**
+ * list_del_init() - Remove a list node from the list and reinitialize it
+ * @node: pointer to the node
+ *
+ * The removed node will not end up in an uninitialized state like when using
+ * list_del. Instead the node is initialized again to the unlinked state.
+ */
+static __inline__ void list_del_init(struct list_head *node)
+{
+ list_del(node);
+ INIT_LIST_HEAD(node);
+}
+
+/**
+ * list_empty() - Check if list head has no nodes attached
+ * @head: pointer to the head of the list
+ *
+ * Return: 0 - list is not empty !0 - list is empty
+ */
+static __inline__ int list_empty(const struct list_head *head)
+{
+ return (head->next == head);
+}
+
+/**
+ * list_is_singular() - Check if list head has exactly one node attached
+ * @head: pointer to the head of the list
+ *
+ * Return: 0 - list is not singular !0 -list has exactly one entry
+ */
+static __inline__ int list_is_singular(const struct list_head *head)
+{
+ return (!list_empty(head) && head->prev == head->next);
+}
+
+/**
+ * list_splice() - Add list nodes from a list to beginning of another list
+ * @list: pointer to the head of the list with the node entries
+ * @head: pointer to the head of the list
+ *
+ * All nodes from @list are added to to the beginning of the list of @head.
+ * It is similar to list_add but for multiple nodes. The @list head is not
+ * modified and has to be initialized to be used as a valid list head/node
+ * again.
+ */
+static __inline__ void list_splice(struct list_head *list,
+ struct list_head *head)
+{
+ struct list_head *head_first = head->next;
+ struct list_head *list_first = list->next;
+ struct list_head *list_last = list->prev;
+
+ if (list_empty(list))
+ return;
+
+ head->next = list_first;
+ list_first->prev = head;
+
+ list_last->next = head_first;
+ head_first->prev = list_last;
+}
+
+/**
+ * list_splice_tail() - Add list nodes from a list to end of another list
+ * @list: pointer to the head of the list with the node entries
+ * @head: pointer to the head of the list
+ *
+ * All nodes from @list are added to to the end of the list of @head.
+ * It is similar to list_add_tail but for multiple nodes. The @list head is not
+ * modified and has to be initialized to be used as a valid list head/node
+ * again.
+ */
+static __inline__ void list_splice_tail(struct list_head *list,
+ struct list_head *head)
+{
+ struct list_head *head_last = head->prev;
+ struct list_head *list_first = list->next;
+ struct list_head *list_last = list->prev;
+
+ if (list_empty(list))
+ return;
+
+ head->prev = list_last;
+ list_last->next = head;
+
+ list_first->prev = head_last;
+ head_last->next = list_first;
+}
+
+/**
+ * list_splice_init() - Move list nodes from a list to beginning of another list
+ * @list: pointer to the head of the list with the node entries
+ * @head: pointer to the head of the list
+ *
+ * All nodes from @list are added to to the beginning of the list of @head.
+ * It is similar to list_add but for multiple nodes.
+ *
+ * The @list head will not end up in an uninitialized state like when using
+ * list_splice. Instead the @list is initialized again to the an empty
+ * list/unlinked state.
+ */
+static __inline__ void list_splice_init(struct list_head *list,
+ struct list_head *head)
+{
+ list_splice(list, head);
+ INIT_LIST_HEAD(list);
+}
+
+/**
+ * list_splice_tail_init() - Move list nodes from a list to end of another list
+ * @list: pointer to the head of the list with the node entries
+ * @head: pointer to the head of the list
+ *
+ * All nodes from @list are added to to the end of the list of @head.
+ * It is similar to list_add_tail but for multiple nodes.
+ *
+ * The @list head will not end up in an uninitialized state like when using
+ * list_splice. Instead the @list is initialized again to the an empty
+ * list/unlinked state.
+ */
+static __inline__ void list_splice_tail_init(struct list_head *list,
+ struct list_head *head)
+{
+ list_splice_tail(list, head);
+ INIT_LIST_HEAD(list);
+}
+
+/**
+ * list_cut_position() - Move beginning of a list to another list
+ * @head_to: pointer to the head of the list which receives nodes
+ * @head_from: pointer to the head of the list
+ * @node: pointer to the node in which defines the cutting point
+ *
+ * All entries from the beginning of the list @head_from to (including) the
+ * @node is moved to @head_from.
+ *
+ * @head_to is replaced when @head_from is not empty. @node must be a real
+ * list node from @head_from or the behavior is undefined.
+ */
+static __inline__ void list_cut_position(struct list_head *head_to,
+ struct list_head *head_from,
+ struct list_head *node)
+{
+ struct list_head *head_from_first = head_from->next;
+
+ if (list_empty(head_from))
+ return;
+
+ if (head_from == node) {
+ INIT_LIST_HEAD(head_to);
+ return;
+ }
+
+ head_from->next = node->next;
+ head_from->next->prev = head_from;
+
+ head_to->prev = node;
+ node->next = head_to;
+ head_to->next = head_from_first;
+ head_to->next->prev = head_to;
+}
+
+/**
+ * list_move() - Move a list node to the beginning of the list
+ * @node: pointer to the node
+ * @head: pointer to the head of the list
+ *
+ * The @node is removed from its old position/node and add to the beginning of
+ * @head
+ */
+static __inline__ void list_move(struct list_head *node, struct list_head *head)
+{
+ list_del(node);
+ list_add(node, head);
+}
+
+/**
+ * list_move_tail() - Move a list node to the end of the list
+ * @node: pointer to the node
+ * @head: pointer to the head of the list
+ *
+ * The @node is removed from its old position/node and add to the end of @head
+ */
+static __inline__ void list_move_tail(struct list_head *node,
+ struct list_head *head)
+{
+ list_del(node);
+ list_add_tail(node, head);
+}
+
+/**
+ * list_entry() - Calculate address of entry that contains list node
+ * @node: pointer to list node
+ * @type: type of the entry containing the list node
+ * @member: name of the list_head member variable in struct @type
+ *
+ * Return: @type pointer of entry containing node
+ */
+#define list_entry(node, type, member) container_of(node, type, member)
+
+/**
+ * list_first_entry() - get first entry of the list
+ * @head: pointer to the head of the list
+ * @type: type of the entry containing the list node
+ * @member: name of the list_head member variable in struct @type
+ *
+ * Return: @type pointer of first entry in list
+ */
+#define list_first_entry(head, type, member) \
+ list_entry((head)->next, type, member)
+
+/**
+ * list_last_entry() - get last entry of the list
+ * @head: pointer to the head of the list
+ * @type: type of the entry containing the list node
+ * @member: name of the list_head member variable in struct @type
+ *
+ * Return: @type pointer of last entry in list
+ */
+#define list_last_entry(head, type, member) \
+ list_entry((head)->prev, type, member)
+
+/**
+ * list_for_each - iterate over list nodes
+ * @node: list_head pointer used as iterator
+ * @head: pointer to the head of the list
+ *
+ * The nodes and the head of the list must must be kept unmodified while
+ * iterating through it. Any modifications to the the list will cause undefined
+ * behavior.
+ */
+#define list_for_each(node, head) \
+ for (node = (head)->next; \
+ node != (head); \
+ node = node->next)
+
+/**
+ * list_for_each_entry_t - iterate over list entries
+ * @entry: @type pointer used as iterator
+ * @head: pointer to the head of the list
+ * @type: type of the entries containing the list nodes
+ * @member: name of the list_head member variable in struct @type
+ *
+ * The nodes and the head of the list must must be kept unmodified while
+ * iterating through it. Any modifications to the the list will cause undefined
+ * behavior.
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define list_for_each_entry_t(entry, head, type, member) \
+ for (entry = list_entry((head)->next, type, member); \
+ &entry->member != (head); \
+ entry = list_entry(entry->member.next, type, member))
+
+/**
+ * list_for_each_entry - iterate over list entries
+ * @entry: pointer used as iterator
+ * @head: pointer to the head of the list
+ * @member: name of the list_head member variable in struct type of @entry
+ *
+ * The nodes and the head of the list must must be kept unmodified while
+ * iterating through it. Any modifications to the the list will cause undefined
+ * behavior.
+ */
+#ifdef LIST_TYPEOF_USE
+#define list_for_each_entry(entry, head, member) \
+ list_for_each_entry_t(entry, head, __typeof__(*entry), member)
+#endif
+
+/**
+ * list_for_each_safe - iterate over list nodes and allow deletes
+ * @node: list_head pointer used as iterator
+ * @safe: list_head pointer used to store info for next entry in list
+ * @head: pointer to the head of the list
+ *
+ * The current node (iterator) is allowed to be removed from the list. Any
+ * other modifications to the the list will cause undefined behavior.
+ */
+#define list_for_each_safe(node, safe, head) \
+ for (node = (head)->next, safe = node->next; \
+ node != (head); \
+ node = safe, safe = node->next)
+
+/**
+ * list_for_each_entry_safe_t - iterate over list entries and allow deletes
+ * @entry: @type pointer used as iterator
+ * @safe: @type pointer used to store info for next entry in list
+ * @head: pointer to the head of the list
+ * @type: type of the entries containing the list nodes
+ * @member: name of the list_head member variable in struct @type
+ *
+ * The current node (iterator) is allowed to be removed from the list. Any
+ * other modifications to the the list will cause undefined behavior.
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define list_for_each_entry_safe_t(entry, safe, head, type, member) \
+ for (entry = list_entry((head)->next, type, member), \
+ safe = list_entry(entry->member.next, type, member); \
+ &entry->member != (head); \
+ entry = safe, \
+ safe = list_entry(safe->member.next, type, member))
+
+/**
+ * list_for_each_entry_safe - iterate over list entries and allow deletes
+ * @entry: pointer used as iterator
+ * @safe: @type pointer used to store info for next entry in list
+ * @head: pointer to the head of the list
+ * @member: name of the list_head member variable in struct type of @entry
+ *
+ * The current node (iterator) is allowed to be removed from the list. Any
+ * other modifications to the the list will cause undefined behavior.
+ */
+#ifdef LIST_TYPEOF_USE
+#define list_for_each_entry_safe(entry, safe, head, member) \
+ list_for_each_entry_safe_t(entry, safe, head, __typeof__(*entry), \
+ member)
+#endif
+
+/**
+ * struct hlist_node - Node of a double-linked list with single pointer head
+ * @next: pointer to the next node in the list
+ * @pprev: pointer to @next of the previous node in the hlist
+ *
+ * The double-linked list with single pointer head consists of a head and nodes
+ * attached to this head. The hlist_* functions and macros can be used to access
+ * and modify this data structure.
+ *
+ * The @pprev pointer is used to find the previous node (or head) in the list
+ * when doing hlist_del operations
+ *
+ * The hlist nodes are usually embedded in a container structure which holds the
+ * actual data. Such an container object is called entry. The helper hlist_entry
+ * can be used to calculate the object address from the address of the node.
+ */
+struct hlist_node {
+ struct hlist_node *next;
+ struct hlist_node **pprev;
+};
+
+/**
+ * struct hlist_head - Head of a double-linked list with single pointer head
+ * @first: pointer to the first node in the hlist
+ *
+ * The hlist doesn't have a pointer to the last node. This makes it harder to
+ * access or modify the tail of the list. But the single pointer to the first
+ * entry makes it well suited for implementation of hash tables because it
+ * cuts the size cost of the head pointers by half compared to the list_head.
+ */
+struct hlist_head {
+ struct hlist_node *first;
+};
+
+/**
+ * HLIST_HEAD - Declare hlist head and initialize it
+ * @head: name of the new object
+ */
+#define HLIST_HEAD(head) \
+ struct hlist_head head = { NULL }
+
+/**
+ * INIT_HLIST_HEAD() - Initialize empty hlist head
+ * @head: pointer to hlist head
+ */
+static __inline__ void INIT_HLIST_HEAD(struct hlist_head *head)
+{
+ head->first = NULL;
+}
+
+/**
+ * INIT_HLIST_NODE() - Initialize unhashed hlist node
+ * @node: pointer to hlist node
+ *
+ * A hlist_node is usually linked inside a hlist, will be added to a hlist in
+ * the near future or the entry containing the node will be free'd soon.
+ *
+ * But an unlinked node may be given to a function which uses hlist_del(_init)
+ * before it ends up in a previously mentioned state. The hlist_del(_init) on an
+ * initialized node is well defined and safe. But the result of a
+ * hlist_del(_init) on an uninitialized node is undefined (unrelated memory is
+ * modified, crashes, ...).
+ */
+static __inline__ void INIT_HLIST_NODE(struct hlist_node *node)
+{
+ node->next = NULL;
+ node->pprev = NULL;
+}
+
+/**
+ * hlist_add_head() - Add a hlist node to the beginning of the hlist
+ * @node: pointer to the new node
+ * @head: pointer to the head of the hlist
+ */
+static __inline__ void hlist_add_head(struct hlist_node *node,
+ struct hlist_head *head)
+{
+ struct hlist_node *first = head->first;
+
+ head->first = node;
+ node->next = first;
+ node->pprev = &head->first;
+ if (first)
+ first->pprev = &node->next;
+}
+
+/**
+ * hlist_add_before() - Add a hlist node before another node to the hlist
+ * @new_node: pointer to the new node
+ * @node: pointer to the reference node in the hlist
+ */
+static __inline__ void hlist_add_before(struct hlist_node *new_node,
+ struct hlist_node *node)
+{
+ struct hlist_node **pprev = node->pprev;
+
+ *pprev = new_node;
+ new_node->next = node;
+ new_node->pprev = pprev;
+ node->pprev = &new_node->next;
+}
+
+/**
+ * hlist_add_behind() - Add a hlist node behind another node to the hlist
+ * @new_node: pointer to the new node
+ * @node: pointer to the reference node in the hlist
+ */
+static __inline__ void hlist_add_behind(struct hlist_node *new_node,
+ struct hlist_node *node)
+{
+ struct hlist_node *next = node->next;
+
+ node->next = new_node;
+ new_node->pprev = &node->next;
+ new_node->next = next;
+ if (next)
+ next->pprev = &new_node->next;
+}
+
+/**
+ * hlist_del() - Remove a hlist node from the hlist
+ * @node: pointer to the node
+ *
+ * The node is only removed from the hlist. Neither the memory of the removed
+ * node nor the memory of the entry containing the node is free'd. The node
+ * has to be handled like an uninitialized node. Accessing the next or pprev
+ * pointer of the node is not safe.
+ *
+ * Unlinked, initialized nodes are also uninitialized after hlist_del.
+ *
+ * LIST_POISONING can be enabled during build-time to provoke an invalid memory
+ * access when the memory behind the next/prev pointer is used after an
+ * hlist_del. This only works on systems which prohibit access to the predefined
+ * memory addresses.
+ */
+static __inline__ void hlist_del(struct hlist_node *node)
+{
+ struct hlist_node *next = node->next;
+ struct hlist_node **pprev = node->pprev;
+
+ if (pprev)
+ *pprev = next;
+
+ if (next)
+ next->pprev = pprev;
+
+#ifdef LIST_POISONING
+ node->pprev = (struct hlist_node **)(0x00100100);
+ node->next = (struct hlist_node *)(0x00200200);
+#endif
+}
+
+/**
+ * hlist_del_init() - Remove a hlist node from the hlist and reinitialize it
+ * @node: pointer to the node
+ *
+ * The removed node will not end up in an uninitialized state like when using
+ * hlist_del. Instead the node is initialized again to the unlinked state.
+ */
+static __inline__ void hlist_del_init(struct hlist_node *node)
+{
+ hlist_del(node);
+ INIT_HLIST_NODE(node);
+}
+
+/**
+ * hlist_empty() - Check if hlist head has no nodes attached
+ * @head: pointer to the head of the hlist
+ *
+ * Return: 0 - hlist is not empty !0 - hlist is empty
+ */
+static __inline__ int hlist_empty(const struct hlist_head *head)
+{
+ return !head->first;
+}
+
+/**
+ * hlist_move_list() - Move hlist nodes from a hlist head new hlist head
+ * @list: pointer to the head of the hlist with the node entries
+ * @head: pointer to the head of the hlist
+ *
+ * All nodes from @list are added to to the beginning of the list of @head.
+ * @head can be uninitialized or an empty, initialized hlist. All entries of
+ * a non-empty hlist @head would be lost after this operation.
+ *
+ * The @list head will not end up in an uninitialized state. Instead the @list
+ * is initialized again to an empty hlist.
+ */
+static __inline__ void hlist_move_list(struct hlist_head *list,
+ struct hlist_head *head)
+{
+ head->first = list->first;
+ if (head->first)
+ head->first->pprev = &head->first;
+ INIT_HLIST_HEAD(list);
+}
+
+/**
+ * hlist_entry() - Calculate address of entry that contains hlist node
+ * @node: pointer to hlist node
+ * @type: type of the entry containing the hlist node
+ * @member: name of the hlist_node member variable in struct @type
+ *
+ * Return: @type pointer of entry containing node
+ */
+#define hlist_entry(node, type, member) container_of(node, type, member)
+
+/**
+ * hlist_entry_safe() - Calculate address of entry that contains hlist node
+ * @node: pointer to hlist node or (struct hlist_node *)NULL
+ * @type: type of the entry containing the hlist node
+ * @member: name of the hlist_node member variable in struct @type
+ *
+ * Return: @type pointer of entry containing node or NULL
+ */
+#ifdef LIST_TYPEOF_USE
+#define hlist_entry_safe(node, type, member) __extension__ ({ \
+ __typeof__(node) __node = (node); \
+ !__node ? NULL : hlist_entry(__node, type, member); })
+#else
+#define hlist_entry_safe(node, type, member) \
+ (node) ? hlist_entry(node, type, member) : NULL
+#endif
+
+/**
+ * hlist_for_each - iterate over hlist nodes
+ * @node: hlist_node pointer used as iterator
+ * @head: pointer to the head of the hlist
+ *
+ * The nodes and the head of the hlist must must be kept unmodified while
+ * iterating through it. Any modifications to the the hlist will cause undefined
+ * behavior.
+ */
+#define hlist_for_each(node, head) \
+ for (node = (head)->first; \
+ node; \
+ node = node->next)
+
+/**
+ * hlist_for_each_entry_t - iterate over hlist entries
+ * @entry: @type pointer used as iterator
+ * @head: pointer to the head of the hlist
+ * @type: type of the entries containing the hlist nodes
+ * @member: name of the hlist_node member variable in struct @type
+ *
+ * The nodes and the head of the hlist must must be kept unmodified while
+ * iterating through it. Any modifications to the the hlist will cause undefined
+ * behavior.
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define hlist_for_each_entry_t(entry, head, type, member) \
+ for (entry = hlist_entry_safe((head)->first, type, member); \
+ entry; \
+ entry = hlist_entry_safe(entry->member.next, type, member))
+
+/**
+ * hlist_for_each_entry - iterate over hlist entries
+ * @entry: pointer used as iterator
+ * @head: pointer to the head of the hlist
+ * @member: name of the hlist_node member variable in struct type of @entry
+ *
+ * The nodes and the head of the hlist must must be kept unmodified while
+ * iterating through it. Any modifications to the the hlist will cause undefined
+ * behavior.
+ */
+#ifdef LIST_TYPEOF_USE
+#define hlist_for_each_entry(entry, head, member) \
+ hlist_for_each_entry_t(entry, head, __typeof__(*entry), member)
+#endif
+
+/**
+ * hlist_for_each_safe - iterate over hlist nodes and allow deletes
+ * @node: hlist_node pointer used as iterator
+ * @safe: hlist_node pointer used to store info for next entry in hlist
+ * @head: pointer to the head of the hlist
+ *
+ * The current node (iterator) is allowed to be removed from the hlist. Any
+ * other modifications to the the hlist will cause undefined behavior.
+ */
+#define hlist_for_each_safe(node, safe, head) \
+ for (node = (head)->first; \
+ node && ((safe = node->next) || 1); \
+ node = safe)
+
+/**
+ * hlist_for_each_entry_safe_t - iterate over hlist entries and allow deletes
+ * @entry: @type pointer used as iterator
+ * @safe: hlist_node pointer used to store info for next entry in hlist
+ * @head: pointer to the head of the hlist
+ * @type: type of the entries containing the hlist nodes
+ * @member: name of the hlist_node member variable in struct @type
+ *
+ * The current node (iterator) is allowed to be removed from the hlist. Any
+ * other modifications to the the hlist will cause undefined behavior.
+ *
+ * WARNING this functionality is not available in the Linux list implementation
+ */
+#define hlist_for_each_entry_safe_t(entry, safe, head, type, member) \
+ for (entry = hlist_entry_safe((head)->first, type, member); \
+ entry && ((safe = entry->member.next) || 1); \
+ entry = hlist_entry_safe(safe, type, member))
+
+/**
+ * hlist_for_each_entry_safe - iterate over hlist entries and allow deletes
+ * @entry: pointer used as iterator
+ * @safe: hlist_node pointer used to store info for next entry in hlist
+ * @head: pointer to the head of the hlist
+ * @member: name of the hlist_node member variable in struct type of @entry
+ *
+ * The current node (iterator) is allowed to be removed from the hlist. Any
+ * other modifications to the the hlist will cause undefined behavior.
+ */
+#ifdef LIST_TYPEOF_USE
+#define hlist_for_each_entry_safe(entry, safe, head, member) \
+ hlist_for_each_entry_safe_t(entry, safe, head, __typeof__(*entry),\
+ member)
+#endif
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* __LINUX_LIKE_LIST_H__ */
diff --git a/tcpdump.c b/tcpdump.c
index be0c4f0..2125b66 100644
--- a/tcpdump.c
+++ b/tcpdump.c
@@ -1118,7 +1118,7 @@ int tcpdump(int argc, char **argv)
struct ifreq req;
struct timeval tv;
struct dump_if *dump_if, *dump_if_tmp;
- struct list_head_first dump_if_list;
+ struct list_head dump_if_list;
fd_set wait_sockets, tmp_wait_sockets;
ssize_t read_len;
int ret = EXIT_FAILURE, res, optchar, found_args = 1, max_sock = 0, tmp;
@@ -1168,7 +1168,7 @@ int tcpdump(int argc, char **argv)
bat_hosts_init(read_opt);
/* init interfaces list */
- INIT_LIST_HEAD_FIRST(dump_if_list);
+ INIT_LIST_HEAD(&dump_if_list);
FD_ZERO(&wait_sockets);
while (argc > found_args) {
@@ -1176,7 +1176,6 @@ int tcpdump(int argc, char **argv)
dump_if = malloc(sizeof(struct dump_if));
memset(dump_if, 0, sizeof(struct dump_if));
dump_if->raw_sock = -1;
- INIT_LIST_HEAD(&dump_if->list);
dump_if->dev = argv[found_args];
@@ -1306,7 +1305,7 @@ out:
if (dump_if->raw_sock >= 0)
close(dump_if->raw_sock);
- list_del((struct list_head *)&dump_if_list, &dump_if->list, &dump_if_list);
+ list_del(&dump_if->list);
free(dump_if);
}
diff --git a/tcpdump.h b/tcpdump.h
index 229ee70..38d07c5 100644
--- a/tcpdump.h
+++ b/tcpdump.h
@@ -27,7 +27,7 @@
#include <net/if_arp.h>
#include <sys/types.h>
#include "main.h"
-#include "list-batman.h"
+#include "list.h"
#ifndef ARPHRD_IEEE80211_PRISM
#define ARPHRD_IEEE80211_PRISM 802
--
2.8.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [B.A.T.M.A.N.] [PATCH 2/4] batctl: Use monotonic time source for icmp timing
2016-07-19 20:43 [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list Sven Eckelmann
@ 2016-07-19 20:44 ` Sven Eckelmann
2016-10-18 12:25 ` [B.A.T.M.A.N.] [2/4] " Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 3/4] batctl: Add helper to generate instant random bytes Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 4/4] batctl: Implement non-routing batadv_icmp in userspace Sven Eckelmann
3 siblings, 1 reply; 7+ messages in thread
From: Sven Eckelmann @ 2016-07-19 20:44 UTC (permalink / raw)
To: b.a.t.m.a.n
gettimeofday is not monotonic. It would give wrong results in some
situation like leap seconds, when switching between daylight saving time
and non-DST, NTP changes the time or when the administrator of the system
changes it manually. The function is also obsolete since POSIX.1-2008.
clock_gettime is recommended to get a monotonic timesource and as general
replacement of gettimeofday.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
Makefile | 2 +-
functions.c | 17 +++++++++--------
2 files changed, 10 insertions(+), 9 deletions(-)
diff --git a/Makefile b/Makefile
index e980587..57d9a4e 100755
--- a/Makefile
+++ b/Makefile
@@ -46,7 +46,7 @@ MANPAGE = man/batctl.8
# batctl flags and options
CFLAGS += -Wall -W -std=gnu99 -fno-strict-aliasing -MD -MP
CPPFLAGS += -D_GNU_SOURCE
-LDLIBS += -lm
+LDLIBS += -lm -lrt
# disable verbose output
ifneq ($(findstring $(MAKEFLAGS),s),s)
diff --git a/functions.c b/functions.c
index f9feca4..bc4f9f8 100644
--- a/functions.c
+++ b/functions.c
@@ -48,6 +48,7 @@
#include <netlink/handlers.h>
#include <netlink/msg.h>
#include <netlink/attr.h>
+#include <time.h>
#include "main.h"
#include "functions.h"
@@ -59,7 +60,7 @@
#define PATH_BUFF_LEN 200
-static struct timeval start_time;
+static struct timespec start_time;
static char *host_name;
char *line_ptr = NULL;
@@ -80,23 +81,23 @@ const char *fs_compile_out_param[] = {
void start_timer(void)
{
- gettimeofday(&start_time, NULL);
+ clock_gettime(CLOCK_MONOTONIC, &start_time);
}
double end_timer(void)
{
- struct timeval end_time, diff;
+ struct timespec end_time, diff;
- gettimeofday(&end_time, NULL);
+ clock_gettime(CLOCK_MONOTONIC, &end_time);
diff.tv_sec = end_time.tv_sec - start_time.tv_sec;
- diff.tv_usec = end_time.tv_usec - start_time.tv_usec;
+ diff.tv_nsec = end_time.tv_nsec - start_time.tv_nsec;
- if (diff.tv_usec < 0) {
+ if (diff.tv_nsec < 0) {
diff.tv_sec--;
- diff.tv_usec += 1000000;
+ diff.tv_nsec += 1000000000;
}
- return (((double)diff.tv_sec * 1000) + ((double)diff.tv_usec / 1000));
+ return (((double)diff.tv_sec * 1000) + ((double)diff.tv_nsec / 1000000));
}
char *ether_ntoa_long(const struct ether_addr *addr)
--
2.8.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [B.A.T.M.A.N.] [PATCH 3/4] batctl: Add helper to generate instant random bytes
2016-07-19 20:43 [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 2/4] batctl: Use monotonic time source for icmp timing Sven Eckelmann
@ 2016-07-19 20:44 ` Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 4/4] batctl: Implement non-routing batadv_icmp in userspace Sven Eckelmann
3 siblings, 0 replies; 7+ messages in thread
From: Sven Eckelmann @ 2016-07-19 20:44 UTC (permalink / raw)
To: b.a.t.m.a.n
Linux provides different ways to get instant random bytes. These are not
all supported on all systems and thus a fallback may have to be used.
Abstract all this in a single function which can be used from different
parts of the code.
The current implementations are
* get random data from urandom pool via SYS_getrandom syscall
* get random data from reading /dev/urandom
* fallback to per-program prng initialized via the current time (seconds +
nanoseconds)
All are tried in this order in hope to get a high quality random number
source before falling back to some really low quality one.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
functions.c | 65 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
functions.h | 2 ++
2 files changed, 67 insertions(+)
diff --git a/functions.c b/functions.c
index bc4f9f8..d9f93c2 100644
--- a/functions.c
+++ b/functions.c
@@ -41,6 +41,7 @@
#include <linux/if_link.h>
#include <linux/rtnetlink.h>
#include <linux/neighbour.h>
+#include <sys/syscall.h>
#include <errno.h>
#include <net/if.h>
#include <netlink/socket.h>
@@ -1071,3 +1072,67 @@ int check_mesh_iface_ownership(char *mesh_iface, char *hard_iface)
return EXIT_SUCCESS;
}
+
+static int get_random_bytes_syscall(void *buf __unused, size_t buflen __unused)
+{
+#ifdef SYS_getrandom
+ return syscall(SYS_getrandom, buf, buflen, 0);
+#else
+ return -EOPNOTSUPP;
+#endif
+}
+
+static int get_random_bytes_urandom(void *buf, size_t buflen)
+{
+ int fd;
+ ssize_t r;
+
+ fd = open("/dev/urandom", O_RDONLY);
+ if (fd < 0)
+ return -EOPNOTSUPP;
+
+ r = read(fd, buf, buflen);
+ close(fd);
+ if (r < 0)
+ return -EOPNOTSUPP;
+
+ if ((size_t)r != buflen)
+ return -EOPNOTSUPP;
+
+ return 0;
+}
+
+static int get_random_bytes_fallback(void *buf, size_t buflen)
+{
+ struct timespec now;
+ static int initialized = 0;
+ size_t i;
+ uint8_t *bufc = buf;
+
+ /* this is not a good source for randomness */
+ if (!initialized) {
+ clock_gettime(CLOCK_MONOTONIC, &now);
+ srand(now.tv_sec ^ now.tv_nsec);
+ initialized = 1;
+ }
+
+ for (i = 0; i < buflen; i++)
+ bufc[i] = rand() & 0xff;
+
+ return 0;
+}
+
+void get_random_bytes(void *buf, size_t buflen)
+{
+ int ret;
+
+ ret = get_random_bytes_syscall(buf, buflen);
+ if (ret != -EOPNOTSUPP)
+ return;
+
+ ret = get_random_bytes_urandom(buf, buflen);
+ if (ret != -EOPNOTSUPP)
+ return;
+
+ get_random_bytes_fallback(buf, buflen);
+}
diff --git a/functions.h b/functions.h
index e413d6b..95cd6cf 100644
--- a/functions.h
+++ b/functions.h
@@ -53,6 +53,8 @@ int netlink_simple_request(struct nl_msg *msg);
int check_mesh_iface(char *mesh_iface);
int check_mesh_iface_ownership(char *mesh_iface, char *hard_iface);
+void get_random_bytes(void *buf, size_t buflen);
+
int print_routing_algos(void);
extern char *line_ptr;
--
2.8.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [B.A.T.M.A.N.] [PATCH 4/4] batctl: Implement non-routing batadv_icmp in userspace
2016-07-19 20:43 [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints Sven Eckelmann
` (2 preceding siblings ...)
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 3/4] batctl: Add helper to generate instant random bytes Sven Eckelmann
@ 2016-07-19 20:44 ` Sven Eckelmann
3 siblings, 0 replies; 7+ messages in thread
From: Sven Eckelmann @ 2016-07-19 20:44 UTC (permalink / raw)
To: b.a.t.m.a.n
The current endpoint for batadv_icmp* is implemented in the kernel module
and can be accessed via debugfs. But the debugfs cannot be accessed from
non-default netns or when debugfs is disabled. Thus it has be possible to
use it via the netlink infrastructure to make it compatible with future
setups.
The use of the socket file is completely removed and instead raw sockets
with BPF filters are used to send/receive batadv_icmp_packet* directly. All
information about interfaces and available originators are received via
rtnetlink and the batman-adv netlink.
The originators debugfs file is used when the batman-adv netlink commands
are not available. The routing of batadv_icmp_packets is still done inside
the kernel module.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
---
Makefile | 1 +
icmp_helper.c | 633 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
icmp_helper.h | 58 ++++++
netlink.c | 178 ++++++++++++++++-
netlink.h | 3 +
ping.c | 42 ++--
traceroute.c | 42 ++--
7 files changed, 895 insertions(+), 62 deletions(-)
create mode 100644 icmp_helper.c
create mode 100644 icmp_helper.h
diff --git a/Makefile b/Makefile
index 57d9a4e..cfc9465 100755
--- a/Makefile
+++ b/Makefile
@@ -30,6 +30,7 @@ OBJ += debug.o
OBJ += functions.o
OBJ += genl.o
OBJ += hash.o
+OBJ += icmp_helper.o
OBJ += interface.o
OBJ += ioctl.o
OBJ += main.o
diff --git a/icmp_helper.c b/icmp_helper.c
new file mode 100644
index 0000000..87c6e15
--- /dev/null
+++ b/icmp_helper.c
@@ -0,0 +1,633 @@
+/*
+ * Copyright (C) 2007-2016 B.A.T.M.A.N. contributors:
+ *
+ * Marek Lindner <mareklindner@neomailbox.ch>, Simon Wunderlich
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of version 2 of the GNU General Public
+ * License as published by the Free Software Foundation.
+ *
+ * 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., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
+ *
+ */
+
+#include "icmp_helper.h"
+#include "main.h"
+
+#include <errno.h>
+#include <linux/filter.h>
+#include <linux/if_packet.h>
+#include <net/ethernet.h>
+#include <netinet/ether.h>
+#include <stdbool.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+#include <sys/select.h>
+#include <sys/socket.h>
+#include <sys/uio.h>
+#include <unistd.h>
+
+#include "debug.h"
+#include "debugfs.h"
+#include "functions.h"
+#include "list.h"
+#include "netlink.h"
+#include "packet.h"
+
+#ifndef ETH_P_BATMAN
+#define ETH_P_BATMAN 0x4305
+#endif /* ETH_P_BATMAN */
+
+static LIST_HEAD(interface_list);
+static size_t direct_reply_len;
+static uint8_t uid;
+static uint8_t primary_mac[ETH_ALEN];
+static uint8_t icmp_buffer[BATADV_ICMP_MAX_PACKET_SIZE];
+
+#define BATADV_ICMP_MIN_PACKET_SIZE sizeof(struct batadv_icmp_packet)
+
+#define BADADV_ICMP_ETH_OFFSET(member) \
+ (ETH_HLEN + offsetof(struct batadv_icmp_packet, member))
+
+static struct icmp_interface *icmp_interface_find(const char *ifname)
+{
+ struct icmp_interface *found = NULL;
+ struct icmp_interface *iface;
+
+ list_for_each_entry(iface, &interface_list, list) {
+ if (strcmp(iface->name, ifname) == 0) {
+ found = iface;
+ break;
+ }
+ }
+
+ return found;
+}
+
+static bool icmp_interfaces_is_my_mac(uint8_t dst[ETH_ALEN])
+{
+ struct icmp_interface *iface;
+
+ list_for_each_entry(iface, &interface_list, list) {
+ if (memcmp(iface->mac, dst, ETH_ALEN) == 0)
+ return true;
+ }
+
+ return false;
+}
+
+void icmp_interface_destroy(struct icmp_interface *iface)
+{
+ close(iface->sock);
+ list_del(&iface->list);
+ free(iface);
+}
+
+static int icmp_interface_filter(int sock, int uid)
+{
+ struct sock_fprog filter;
+ struct sock_filter accept_icmp[] = {
+ /* load ethernet proto */
+ BPF_STMT(BPF_LD + BPF_H + BPF_ABS,
+ offsetof(struct ether_header, ether_type)),
+ /* jump to ret 0 when it is != 0x4305 */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ ETH_P_BATMAN,
+ 0, 14),
+ /* load pktlen */
+ BPF_STMT(BPF_LD + BPF_W + BPF_LEN,
+ 0),
+ /* jump to ret 0 when it is < 34 */
+ BPF_JUMP(BPF_JMP + BPF_JGE + BPF_K,
+ ETH_HLEN + BATADV_ICMP_MIN_PACKET_SIZE,
+ 0, 12),
+ /* jump to ret 0 when it is > 130 */
+ BPF_JUMP(BPF_JMP + BPF_JGT + BPF_K,
+ ETH_HLEN + BATADV_ICMP_MAX_PACKET_SIZE,
+ 11, 0),
+ /* load batman-adv type */
+ BPF_STMT(BPF_LD + BPF_B + BPF_ABS,
+ BADADV_ICMP_ETH_OFFSET(packet_type)),
+ /* jump to ret 0 when it is != BATADV_ICMP */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ BATADV_ICMP,
+ 0, 9),
+ /* load batman-adv version */
+ BPF_STMT(BPF_LD + BPF_B + BPF_ABS,
+ BADADV_ICMP_ETH_OFFSET(version)),
+ /* jump to ret 0 when it is != 15 */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ BATADV_COMPAT_VERSION,
+ 0, 7),
+ /* load batman-adv icmp msg_type */
+ BPF_STMT(BPF_LD + BPF_B + BPF_ABS,
+ BADADV_ICMP_ETH_OFFSET(msg_type)),
+ /* accept BATADV_ECHO_REPLY or go to next check */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ BATADV_ECHO_REPLY,
+ 2, 0),
+ /* accept BATADV_DESTINATION_UNREACHABLE or go to next check */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ BATADV_DESTINATION_UNREACHABLE,
+ 1, 0),
+ /* accept BATADV_TTL_EXCEEDED or go to ret 0 */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ BATADV_TTL_EXCEEDED,
+ 0, 3),
+ /* load batman-adv icmp uid */
+ BPF_STMT(BPF_LD + BPF_B + BPF_ABS,
+ BADADV_ICMP_ETH_OFFSET(uid)),
+ /* jump to ret 0 when it is not our uid */
+ BPF_JUMP(BPF_JMP + BPF_JEQ + BPF_K,
+ uid,
+ 0, 1),
+ /* accept 130 bytes */
+ BPF_STMT(BPF_RET + BPF_K,
+ ETH_HLEN + BATADV_ICMP_MAX_PACKET_SIZE),
+ /* ret 0 -> reject packet */
+ BPF_STMT(BPF_RET + BPF_K,
+ 0),
+ };
+
+ memset(&filter, 0, sizeof(filter));
+ filter.len = sizeof(accept_icmp) / sizeof(*accept_icmp);
+ filter.filter = accept_icmp;
+
+ if (setsockopt(sock, SOL_SOCKET, SO_ATTACH_FILTER, &filter,
+ sizeof(filter)))
+ return -1;
+
+ return 0;
+}
+
+static int icmp_interface_add(const char *ifname, const uint8_t mac[ETH_ALEN])
+{
+ struct icmp_interface *iface;
+ struct sockaddr_ll sll;
+ struct ifreq req;
+ int ret;
+
+ iface = malloc(sizeof(*iface));
+ if (!iface)
+ return -ENOMEM;
+
+ iface->mark = 1;
+ memcpy(iface->mac, mac, ETH_ALEN);
+
+ strncpy(iface->name, ifname, IFNAMSIZ);
+ iface->name[sizeof(iface->name) - 1] = '\0';
+
+ iface->sock = socket(PF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
+ if (iface->sock < 0) {
+ fprintf(stderr, "Error - can't create raw socket: %s\n", strerror(errno));
+ ret = -errno;
+ goto free_iface;
+ }
+
+ memset(&req, 0, sizeof(struct ifreq));
+ strncpy(req.ifr_name, ifname, IFNAMSIZ);
+ req.ifr_name[sizeof(req.ifr_name) - 1] = '\0';
+
+ ret = ioctl(iface->sock, SIOCGIFINDEX, &req);
+ if (ret < 0) {
+ fprintf(stderr, "Error - can't create raw socket (SIOCGIFINDEX): %s\n", strerror(errno));
+ ret = -errno;
+ goto close_sock;
+ }
+
+ memset(&sll, 0, sizeof(sll));
+ sll.sll_family = AF_PACKET;
+ sll.sll_protocol = htons(ETH_P_ALL);
+ sll.sll_pkttype = PACKET_HOST;
+ sll.sll_ifindex = req.ifr_ifindex;
+
+ ret = bind(iface->sock, (struct sockaddr *)&sll, sizeof(struct sockaddr_ll));
+ if (ret < 0) {
+ fprintf(stderr, "Error - can't bind raw socket: %s\n", strerror(errno));
+ ret = -errno;
+ goto close_sock;
+ }
+
+ ret = icmp_interface_filter(iface->sock, uid);
+ if (ret < 0) {
+ fprintf(stderr, "Error - can't add filter to raw socket: %s\n", strerror(-ret));
+ goto close_sock;
+ }
+
+ list_add(&iface->list, &interface_list);
+
+ return 0;
+
+close_sock:
+ close(iface->sock);
+free_iface:
+ free(iface);
+
+ return ret;
+}
+
+int icmp_interfaces_init(void)
+{
+ get_random_bytes(&uid, 1);
+
+ return 0;
+}
+
+static struct nla_policy link_policy[IFLA_MAX + 1] = {
+ [IFLA_IFNAME] = { .type = NLA_STRING, .maxlen = IFNAMSIZ },
+ [IFLA_MASTER] = { .type = NLA_U32 },
+ [IFLA_ADDRESS] = { .type = NLA_UNSPEC,
+ .minlen = ETH_ALEN,
+ .maxlen = ETH_ALEN,
+ },
+};
+
+struct icmp_interface_update_arg {
+ int ifindex;
+};
+
+static int icmp_interface_update_parse(struct nl_msg *msg, void *arg)
+{
+ struct icmp_interface_update_arg *update_arg = arg;
+ struct nlattr *attrs[IFLA_MAX + 1];
+ struct icmp_interface *iface;
+ struct ifinfomsg *ifm;
+ char *ifname;
+ int ret;
+ int master;
+ uint8_t *mac;
+
+ ifm = nlmsg_data(nlmsg_hdr(msg));
+ ret = nlmsg_parse(nlmsg_hdr(msg), sizeof(*ifm), attrs, IFLA_MAX,
+ link_policy);
+ if (ret < 0)
+ goto err;
+
+ if (!attrs[IFLA_IFNAME])
+ goto err;
+
+ if (!attrs[IFLA_MASTER])
+ goto err;
+
+ if (!attrs[IFLA_ADDRESS])
+ goto err;
+
+ ifname = nla_get_string(attrs[IFLA_IFNAME]);
+ master = nla_get_u32(attrs[IFLA_MASTER]);
+ mac = nla_data(attrs[IFLA_ADDRESS]);
+
+ /* required on older kernels which don't prefilter the results */
+ if (master != update_arg->ifindex)
+ goto err;
+
+ /* update or add interface */
+ iface = icmp_interface_find(ifname);
+ if (!iface) {
+ icmp_interface_add(ifname, mac);
+ goto err;
+ }
+
+ /* update */
+ iface->mark = 1;
+ memcpy(iface->mac, mac, ETH_ALEN);
+
+err:
+ return NL_OK;
+}
+
+static int get_nexthop_debugfs(const char *mesh_iface,
+ struct ether_addr *mac,
+ uint8_t nexthop[ETH_ALEN],
+ char ifname[IF_NAMESIZE])
+{
+ char *tptr;
+ char *temp1, *temp2;
+ char *dest, *neigh, *iface;
+ int lnum, tnum;
+ struct ether_addr *mac_tmp;
+ char path[1024];
+ FILE *f = NULL;
+ size_t len = 0;
+ char *line = NULL;
+ char *primary_info;
+
+ debugfs_make_path(DEBUG_BATIF_PATH_FMT "/" "originators", mesh_iface, path, sizeof(path));
+
+ f = fopen(path, "r");
+ if (!f)
+ return -EOPNOTSUPP;
+
+ lnum = 0;
+ while (getline(&line, &len, f) != -1) {
+ lnum++;
+
+ /* find primary mac */
+ if (lnum == 1) {
+ primary_info = strstr(line, "MainIF/MAC: ");
+ if (!primary_info)
+ continue;
+
+ primary_info += 12;
+ temp1 = strstr(primary_info, "/");
+ if (!temp1)
+ continue;
+
+ temp1++;
+ temp2 = strstr(line, " ");
+ if (!temp2)
+ continue;
+
+ temp2[0] = '\0';
+ mac_tmp = ether_aton(temp1);
+ if (!mac_tmp)
+ continue;
+
+ memcpy(primary_mac, mac_tmp, ETH_ALEN);
+ }
+
+ if (lnum < 3)
+ continue;
+
+ /* find correct neighbor */
+ for (tptr = line, tnum = 0;; tptr = NULL, tnum++) {
+ tptr = strtok_r(tptr, "\t []()", &temp2);
+ if (!tptr)
+ break;
+ switch (tnum) {
+ case 0: dest = tptr; break;
+ case 2: break;
+ case 3: neigh = tptr; break;
+ case 4: iface = tptr; break;
+ default: break;
+ }
+ }
+ if (tnum <= 4)
+ continue;
+
+ mac_tmp = ether_aton(dest);
+ if (!mac_tmp || memcmp(mac_tmp, mac, ETH_ALEN) != 0)
+ continue;
+
+ mac_tmp = ether_aton(neigh);
+ if (!mac_tmp)
+ continue;
+
+ memcpy(nexthop, mac_tmp, ETH_ALEN);
+ strncpy(ifname, iface, IF_NAMESIZE);
+ ifname[IF_NAMESIZE - 1] = '\0';
+ break;
+ }
+ free(line);
+ fclose(f);
+
+ return 0;
+}
+
+static void icmp_interface_unmark(void)
+{
+ struct icmp_interface *iface;
+
+ list_for_each_entry(iface, &interface_list, list)
+ iface->mark = 0;
+}
+
+
+static void icmp_interface_sweep(void)
+{
+ struct icmp_interface *iface, *safe;
+
+ list_for_each_entry_safe(iface, safe, &interface_list, list) {
+ if (iface->mark)
+ continue;
+
+ icmp_interface_destroy(iface);
+ }
+}
+
+static int icmp_interface_update(const char *mesh_iface)
+{
+ struct icmp_interface_update_arg update_arg;
+
+ update_arg.ifindex = if_nametoindex(mesh_iface);
+ if (!update_arg.ifindex)
+ return -errno;
+
+ /* unmark current interface - will be marked again by query */
+ icmp_interface_unmark();
+
+ query_rtnl_link(update_arg.ifindex, icmp_interface_update_parse,
+ &update_arg);
+
+ /* remove old interfaces */
+ icmp_interface_sweep();
+
+ get_primarymac_netlink(mesh_iface, primary_mac);
+
+ return 0;
+}
+
+static int icmp_interface_send(struct batadv_icmp_header *icmp_packet,
+ size_t packet_len, struct icmp_interface *iface,
+ uint8_t nexthop[ETH_ALEN])
+{
+ struct ether_header header;
+ struct iovec vector[2];
+
+ header.ether_type = htons(ETH_P_BATMAN);
+ memcpy(header.ether_shost, iface->mac, ETH_ALEN);
+ memcpy(header.ether_dhost, nexthop, ETH_ALEN);
+
+ vector[0].iov_base = &header;
+ vector[0].iov_len = sizeof(struct ether_header);
+ vector[1].iov_base = icmp_packet;
+ vector[1].iov_len = packet_len;
+
+ return (int)writev(iface->sock, vector, 2);
+}
+
+int icmp_interface_write(const char *mesh_iface,
+ struct batadv_icmp_header *icmp_packet, size_t len)
+{
+ struct batadv_icmp_packet_rr *icmp_packet_rr;
+ struct icmp_interface *iface;
+ uint8_t nexthop[ETH_ALEN];
+ char ifname[IF_NAMESIZE];
+ struct ether_addr mac;
+ size_t packet_len;
+ int ret;
+
+ if (len < sizeof(*icmp_packet))
+ return -EINVAL;
+
+ if (len >= BATADV_ICMP_MAX_PACKET_SIZE)
+ packet_len = BATADV_ICMP_MAX_PACKET_SIZE;
+ else
+ packet_len = len;
+
+ if (icmp_packet->packet_type != BATADV_ICMP)
+ return -EINVAL;
+
+ if (icmp_packet->msg_type != BATADV_ECHO_REQUEST)
+ return -EINVAL;
+
+ icmp_interface_update(mesh_iface);
+
+ if (list_empty(&interface_list))
+ return -EFAULT;
+
+ /* find best neighbor */
+ memcpy(&mac, icmp_packet->dst, ETH_ALEN);
+
+ ret = get_nexthop_netlink(mesh_iface, &mac, nexthop, ifname);
+ if (ret == -EOPNOTSUPP)
+ ret = get_nexthop_debugfs(mesh_iface, &mac, nexthop, ifname);
+ if (ret < 0)
+ goto dst_unreachable;
+
+ iface = icmp_interface_find(ifname);
+ if (!iface)
+ goto dst_unreachable;
+
+ direct_reply_len = 0;
+
+ icmp_packet->uid = uid;
+ memcpy(icmp_packet->orig, primary_mac, ETH_ALEN);
+
+ /* start RR packet */
+ icmp_packet_rr = (struct batadv_icmp_packet_rr *)icmp_packet;
+ if (packet_len == sizeof(*icmp_packet_rr))
+ memcpy(icmp_packet_rr->rr[0], iface->mac, ETH_ALEN);
+
+ return icmp_interface_send(icmp_packet, packet_len, iface, nexthop);
+
+dst_unreachable:
+ memcpy(icmp_buffer, icmp_packet, packet_len);
+ icmp_packet = (struct batadv_icmp_header *)icmp_buffer;
+ icmp_packet->msg_type = BATADV_DESTINATION_UNREACHABLE;
+ direct_reply_len = packet_len;
+ return 0;
+}
+
+static int icmp_interface_preselect(fd_set *read_sockets)
+{
+ struct icmp_interface *iface;
+ int max = 0;
+
+ FD_ZERO(read_sockets);
+
+ list_for_each_entry(iface, &interface_list, list) {
+ FD_SET(iface->sock, read_sockets);
+ if (max <= iface->sock)
+ max = iface->sock + 1;
+ }
+
+ return max;
+}
+
+static ssize_t icmp_interface_get_read_sock(fd_set *read_sockets,
+ struct icmp_interface **piface)
+{
+ struct icmp_interface *iface;
+ int sock = -1;
+
+ list_for_each_entry(iface, &interface_list, list) {
+ if (!FD_ISSET(iface->sock, read_sockets))
+ continue;
+
+ sock = iface->sock;
+ *piface = iface;
+ break;
+ }
+
+ return sock;
+}
+
+ssize_t icmp_interface_read(struct batadv_icmp_header *icmp_packet, size_t len,
+ struct timeval *tv)
+{
+ struct batadv_icmp_packet_rr *icmp_packet_rr;
+ struct icmp_interface *iface;
+ struct ether_header header;
+ struct iovec vector[2];
+ fd_set read_sockets;
+ size_t packet_len;
+ int max_sock;
+ ssize_t read_len;
+ int read_sock;
+ int res;
+
+ if (len < sizeof(*icmp_packet))
+ return -EINVAL;
+
+ if (len >= BATADV_ICMP_MAX_PACKET_SIZE)
+ packet_len = BATADV_ICMP_MAX_PACKET_SIZE;
+ else
+ packet_len = len;
+
+ if (direct_reply_len > 0) {
+ memcpy(icmp_packet, icmp_buffer, packet_len);
+ direct_reply_len = 0;
+ return (ssize_t)packet_len;
+ }
+
+retry:
+ max_sock = icmp_interface_preselect(&read_sockets);
+
+ res = select(max_sock, &read_sockets, NULL, NULL, tv);
+ /* timeout, or < 0 error */
+ if (res <= 0)
+ return res;
+
+ read_sock = icmp_interface_get_read_sock(&read_sockets, &iface);
+ if (read_sock < 0)
+ return read_sock;
+
+ vector[0].iov_base = &header;
+ vector[0].iov_len = sizeof(struct ether_header);
+ vector[1].iov_base = icmp_packet;
+ vector[1].iov_len = packet_len;
+
+ read_len = readv(read_sock, vector, 2);
+ if (read_len < 0)
+ return read_len;
+
+ if (read_len < ETH_HLEN)
+ goto retry;
+
+ read_len -= ETH_HLEN;
+
+ if (read_len < (ssize_t)sizeof(*icmp_packet))
+ goto retry;
+
+ if (!icmp_interfaces_is_my_mac(icmp_packet->dst))
+ goto retry;
+
+ /* end RR packet */
+ icmp_packet_rr = (struct batadv_icmp_packet_rr *)icmp_packet;
+ if (read_len == sizeof(*icmp_packet_rr) &&
+ icmp_packet_rr->rr_cur < BATADV_RR_LEN) {
+ memcpy(icmp_packet_rr->rr[icmp_packet_rr->rr_cur], iface->mac,
+ ETH_ALEN);
+ icmp_packet_rr->rr_cur++;
+ }
+
+ return read_len;
+}
+
+void icmp_interfaces_clean(void)
+{
+ struct icmp_interface *iface, *safe;
+
+ list_for_each_entry_safe(iface, safe, &interface_list, list)
+ icmp_interface_destroy(iface);
+}
diff --git a/icmp_helper.h b/icmp_helper.h
new file mode 100644
index 0000000..37971da
--- /dev/null
+++ b/icmp_helper.h
@@ -0,0 +1,58 @@
+/*
+ * Copyright (C) 2007-2016 B.A.T.M.A.N. contributors:
+ *
+ * Andreas Langer <an.langer@gmx.de>, Marek Lindner <mareklindner@neomailbox.ch>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of version 2 of the GNU General Public
+ * License as published by the Free Software Foundation.
+ *
+ * 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., 51 Franklin Street, Fifth Floor, Boston, MA
+ * 02110-1301, USA
+ *
+ */
+
+#ifndef _BATCTL_ICMP_HELPER_H
+#define _BATCTL_ICMP_HELPER_H
+
+#include "main.h"
+
+#include <linux/if_link.h>
+#include <net/ethernet.h>
+#include <net/if.h>
+#include <netlink/netlink.h>
+#include <netlink/msg.h>
+#include <netlink/attr.h>
+#include <stddef.h>
+#include <stdint.h>
+
+#include "list.h"
+#include "packet.h"
+
+struct timeval;
+
+struct icmp_interface {
+ char name[IFNAMSIZ];
+ uint8_t mac[ETH_ALEN];
+
+ int sock;
+
+ int mark;
+ struct list_head list;
+};
+
+int icmp_interfaces_init(void);
+int icmp_interface_write(const char *mesh_iface,
+ struct batadv_icmp_header *icmp_packet, size_t len);
+void icmp_interfaces_clean(void);
+ssize_t icmp_interface_read(struct batadv_icmp_header *icmp_packet, size_t len,
+ struct timeval *tv);
+
+#endif
diff --git a/netlink.c b/netlink.c
index e8e94b2..0d07f70 100644
--- a/netlink.c
+++ b/netlink.c
@@ -1309,7 +1309,7 @@ static int nlquery_stop_cb(struct nl_msg *msg, void *arg)
}
static int netlink_query_common(const char *mesh_iface, uint8_t nl_cmd,
- nl_recvmsg_msg_cb_t callback,
+ nl_recvmsg_msg_cb_t callback, int flags,
struct nlquery_opts *query_opts)
{
struct nl_sock *sock;
@@ -1359,7 +1359,7 @@ static int netlink_query_common(const char *mesh_iface, uint8_t nl_cmd,
goto err_free_cb;
}
- genlmsg_put(msg, NL_AUTO_PID, NL_AUTO_SEQ, family, 0, NLM_F_DUMP,
+ genlmsg_put(msg, NL_AUTO_PID, NL_AUTO_SEQ, family, 0, flags,
nl_cmd, 1);
nla_put_u32(msg, BATADV_ATTR_MESH_IFINDEX, ifindex);
@@ -1448,7 +1448,8 @@ int translate_mac_netlink(const char *mesh_iface, const struct ether_addr *mac,
ret = netlink_query_common(mesh_iface,
BATADV_CMD_GET_TRANSTABLE_GLOBAL,
- translate_mac_netlink_cb, &opts.query_opts);
+ translate_mac_netlink_cb, NLM_F_DUMP,
+ &opts.query_opts);
if (ret < 0)
return ret;
@@ -1459,3 +1460,174 @@ int translate_mac_netlink(const char *mesh_iface, const struct ether_addr *mac,
return 0;
}
+
+static const int get_nexthop_netlink_mandatory[] = {
+ BATADV_ATTR_ORIG_ADDRESS,
+ BATADV_ATTR_NEIGH_ADDRESS,
+ BATADV_ATTR_HARD_IFINDEX,
+};
+
+struct get_nexthop_netlink_opts {
+ struct ether_addr mac;
+ uint8_t *nexthop;
+ char *ifname;
+ bool found;
+ struct nlquery_opts query_opts;
+};
+
+static int get_nexthop_netlink_cb(struct nl_msg *msg, void *arg)
+{
+ struct nlattr *attrs[BATADV_ATTR_MAX+1];
+ struct nlmsghdr *nlh = nlmsg_hdr(msg);
+ struct nlquery_opts *query_opts = arg;
+ struct get_nexthop_netlink_opts *opts;
+ struct genlmsghdr *ghdr;
+ const uint8_t *orig;
+ const uint8_t *neigh;
+ uint32_t index;
+ const char *ifname;
+
+ opts = container_of(query_opts, struct get_nexthop_netlink_opts,
+ query_opts);
+
+ if (!genlmsg_valid_hdr(nlh, 0))
+ return NL_OK;
+
+ ghdr = nlmsg_data(nlh);
+
+ if (ghdr->cmd != BATADV_CMD_GET_ORIGINATORS)
+ return NL_OK;
+
+ if (nla_parse(attrs, BATADV_ATTR_MAX, genlmsg_attrdata(ghdr, 0),
+ genlmsg_len(ghdr), batadv_netlink_policy)) {
+ return NL_OK;
+ }
+
+ if (missing_mandatory_attrs(attrs, get_nexthop_netlink_mandatory,
+ ARRAY_SIZE(get_nexthop_netlink_mandatory)))
+ return NL_OK;
+
+ orig = nla_data(attrs[BATADV_ATTR_ORIG_ADDRESS]);
+ neigh = nla_data(attrs[BATADV_ATTR_NEIGH_ADDRESS]);
+ index = nla_get_u32(attrs[BATADV_ATTR_HARD_IFINDEX]);
+
+ if (!attrs[BATADV_ATTR_FLAG_BEST])
+ return NL_OK;
+
+ if (memcmp(&opts->mac, orig, ETH_ALEN) != 0)
+ return NL_OK;
+
+ /* save result */
+ memcpy(opts->nexthop, neigh, ETH_ALEN);
+ ifname = if_indextoname(index, opts->ifname);
+ if (!ifname)
+ return NL_OK;
+
+ opts->found = true;
+ opts->query_opts.err = 0;
+
+ return NL_STOP;
+}
+
+int get_nexthop_netlink(const char *mesh_iface, const struct ether_addr *mac,
+ uint8_t *nexthop, char *ifname)
+{
+ struct get_nexthop_netlink_opts opts = {
+ .nexthop = 0,
+ .found = false,
+ .query_opts = {
+ .err = 0,
+ },
+ };
+ int ret;
+
+ memcpy(&opts.mac, mac, ETH_ALEN);
+ opts.nexthop = nexthop;
+ opts.ifname = ifname;
+
+ ret = netlink_query_common(mesh_iface, BATADV_CMD_GET_ORIGINATORS,
+ get_nexthop_netlink_cb, NLM_F_DUMP,
+ &opts.query_opts);
+ if (ret < 0)
+ return ret;
+
+ if (!opts.found)
+ return -ENOENT;
+
+ return 0;
+}
+
+static const int get_primarymac_netlink_mandatory[] = {
+ BATADV_ATTR_HARD_ADDRESS,
+};
+
+struct get_primarymac_netlink_opts {
+ uint8_t *primarymac;
+ bool found;
+ struct nlquery_opts query_opts;
+};
+
+static int get_primarymac_netlink_cb(struct nl_msg *msg, void *arg)
+{
+ struct nlattr *attrs[BATADV_ATTR_MAX+1];
+ struct nlmsghdr *nlh = nlmsg_hdr(msg);
+ struct nlquery_opts *query_opts = arg;
+ struct get_primarymac_netlink_opts *opts;
+ struct genlmsghdr *ghdr;
+ const uint8_t *primary_mac;
+
+ opts = container_of(query_opts, struct get_primarymac_netlink_opts,
+ query_opts);
+
+ if (!genlmsg_valid_hdr(nlh, 0))
+ return NL_OK;
+
+ ghdr = nlmsg_data(nlh);
+
+ if (ghdr->cmd != BATADV_CMD_GET_MESH_INFO)
+ return NL_OK;
+
+ if (nla_parse(attrs, BATADV_ATTR_MAX, genlmsg_attrdata(ghdr, 0),
+ genlmsg_len(ghdr), batadv_netlink_policy)) {
+ return NL_OK;
+ }
+
+ if (missing_mandatory_attrs(attrs, get_primarymac_netlink_mandatory,
+ ARRAY_SIZE(get_primarymac_netlink_mandatory)))
+ return NL_OK;
+
+ primary_mac = nla_data(attrs[BATADV_ATTR_HARD_ADDRESS]);
+
+ /* save result */
+ memcpy(opts->primarymac, primary_mac, ETH_ALEN);
+
+ opts->found = true;
+ opts->query_opts.err = 0;
+
+ return NL_STOP;
+}
+
+int get_primarymac_netlink(const char *mesh_iface, uint8_t *primarymac)
+{
+ struct get_primarymac_netlink_opts opts = {
+ .primarymac = 0,
+ .found = false,
+ .query_opts = {
+ .err = 0,
+ },
+ };
+ int ret;
+
+ opts.primarymac = primarymac;
+
+ ret = netlink_query_common(mesh_iface, BATADV_CMD_GET_MESH_INFO,
+ get_primarymac_netlink_cb, 0,
+ &opts.query_opts);
+ if (ret < 0)
+ return ret;
+
+ if (!opts.found)
+ return -ENOENT;
+
+ return 0;
+}
diff --git a/netlink.h b/netlink.h
index b2084fb..3e7713d 100644
--- a/netlink.h
+++ b/netlink.h
@@ -47,6 +47,9 @@ int netlink_print_bla_backbone(char *mesh_iface, char *orig_iface, int read_opt,
int translate_mac_netlink(const char *mesh_iface, const struct ether_addr *mac,
struct ether_addr *mac_out);
+int get_nexthop_netlink(const char *mesh_iface, const struct ether_addr *mac,
+ uint8_t *nexthop, char *ifname);
+int get_primarymac_netlink(const char *mesh_iface, uint8_t *primarymac);
extern struct nla_policy batadv_netlink_policy[];
diff --git a/ping.c b/ping.c
index 05d6e97..1a0e2fe 100644
--- a/ping.c
+++ b/ping.c
@@ -42,6 +42,7 @@
#include "packet.h"
#include "bat-hosts.h"
#include "debugfs.h"
+#include "icmp_helper.h"
char is_aborted = 0;
@@ -78,8 +79,7 @@ int ping(char *mesh_iface, int argc, char **argv)
struct ether_addr *dst_mac = NULL, *rr_mac = NULL;
struct bat_host *bat_host, *rr_host;
ssize_t read_len;
- fd_set read_socket;
- int ret = EXIT_FAILURE, ping_fd = -1, res, optchar, found_args = 1;
+ int ret = EXIT_FAILURE, res, optchar, found_args = 1;
int loop_count = -1, loop_interval = 0, timeout = 1, rr = 0, i;
unsigned int seq_counter = 0, packets_out = 0, packets_in = 0, packets_loss;
char *dst_string, *mac_string, *rr_string;
@@ -88,7 +88,6 @@ int ping(char *mesh_iface, int argc, char **argv)
uint8_t last_rr_cur = 0, last_rr[BATADV_RR_LEN][ETH_ALEN];
size_t packet_len;
char *debugfs_mnt;
- char icmp_socket[MAX_PATH+1];
int disable_translate_mac = 0;
while ((optchar = getopt(argc, argv, "hc:i:t:RT")) != -1) {
@@ -163,17 +162,7 @@ int ping(char *mesh_iface, int argc, char **argv)
goto out;
}
- debugfs_make_path(SOCKET_PATH_FMT, mesh_iface, icmp_socket, sizeof(icmp_socket));
-
- ping_fd = open(icmp_socket, O_RDWR);
-
- if (ping_fd < 0) {
- fprintf(stderr, "Error - can't open a connection to the batman adv kernel module via the socket '%s': %s\n",
- icmp_socket, strerror(errno));
- printf("Check whether the module is loaded and active.\n");
- goto out;
- }
-
+ icmp_interfaces_init();
packet_len = sizeof(struct batadv_icmp_packet);
memset(&icmp_packet_out, 0, sizeof(icmp_packet_out));
@@ -208,36 +197,32 @@ int ping(char *mesh_iface, int argc, char **argv)
icmp_packet_out.seqno = htons(++seq_counter);
- if (write(ping_fd, (char *)&icmp_packet_out, packet_len) < 0) {
- fprintf(stderr, "Error - can't write to batman adv kernel file '%s': %s\n", icmp_socket, strerror(errno));
+ res = icmp_interface_write(mesh_iface,
+ (struct batadv_icmp_header *)&icmp_packet_out,
+ packet_len);
+ if (res < 0) {
+ fprintf(stderr, "Error - can't send icmp packet: %s\n", strerror(res));
goto sleep;
}
read_packet:
start_timer();
- FD_ZERO(&read_socket);
- FD_SET(ping_fd, &read_socket);
-
- res = select(ping_fd + 1, &read_socket, NULL, NULL, &tv);
+ read_len = icmp_interface_read((struct batadv_icmp_header *)&icmp_packet_in,
+ packet_len, &tv);
if (is_aborted)
break;
packets_out++;
- if (res == 0) {
+ if (read_len == 0) {
printf("Reply from host %s timed out\n", dst_string);
goto sleep;
}
- if (res < 0)
- goto sleep;
-
- read_len = read(ping_fd, (char *)&icmp_packet_in, packet_len);
-
if (read_len < 0) {
- fprintf(stderr, "Error - can't read from batman adv kernel file '%s': %s\n", icmp_socket, strerror(errno));
+ fprintf(stderr, "Error - can't receive icmp packets: %s\n", strerror(read_len));
goto sleep;
}
@@ -353,8 +338,7 @@ sleep:
ret = EXIT_NOSUCCESS;
out:
+ icmp_interfaces_clean();
bat_hosts_free();
- if (ping_fd >= 0)
- close(ping_fd);
return ret;
}
diff --git a/traceroute.c b/traceroute.c
index 2cd29d9..89e5ad9 100644
--- a/traceroute.c
+++ b/traceroute.c
@@ -39,6 +39,7 @@
#include "packet.h"
#include "bat-hosts.h"
#include "debugfs.h"
+#include "icmp_helper.h"
#define TTL_MAX 50
@@ -60,14 +61,12 @@ int traceroute(char *mesh_iface, int argc, char **argv)
struct bat_host *bat_host;
struct ether_addr *dst_mac = NULL;
struct timeval tv;
- fd_set read_socket;
ssize_t read_len;
char *dst_string, *mac_string, *return_mac, dst_reached = 0;
- int ret = EXIT_FAILURE, res, trace_fd = -1, i;
+ int ret = EXIT_FAILURE, res, i;
int found_args = 1, optchar, seq_counter = 0, read_opt = USE_BAT_HOSTS;
double time_delta[NUM_PACKETS];
char *debugfs_mnt;
- char icmp_socket[MAX_PATH+1];
int disable_translate_mac = 0;
while ((optchar = getopt(argc, argv, "hnT")) != -1) {
@@ -122,16 +121,7 @@ int traceroute(char *mesh_iface, int argc, char **argv)
goto out;
}
- debugfs_make_path(SOCKET_PATH_FMT, mesh_iface, icmp_socket, sizeof(icmp_socket));
-
- trace_fd = open(icmp_socket, O_RDWR);
-
- if (trace_fd < 0) {
- fprintf(stderr, "Error - can't open a connection to the batman adv kernel module via the socket '%s': %s\n",
- icmp_socket, strerror(errno));
- fprintf(stderr, "Check whether the module is loaded and active.\n");
- goto out;
- }
+ icmp_interfaces_init();
memset(&icmp_packet_out, 0, sizeof(icmp_packet_out));
memcpy(&icmp_packet_out.dst, dst_mac, ETH_ALEN);
@@ -154,8 +144,11 @@ int traceroute(char *mesh_iface, int argc, char **argv)
icmp_packet_out.seqno = htons(++seq_counter);
time_delta[i] = 0.0;
- if (write(trace_fd, (char *)&icmp_packet_out, sizeof(icmp_packet_out)) < 0) {
- fprintf(stderr, "Error - can't write to batman adv kernel file '%s': %s\n", icmp_socket, strerror(errno));
+ res = icmp_interface_write(mesh_iface,
+ (struct batadv_icmp_header *)&icmp_packet_out,
+ sizeof(icmp_packet_out));
+ if (res < 0) {
+ fprintf(stderr, "Error - can't send icmp packet: %s\n", strerror(res));
continue;
}
@@ -165,21 +158,11 @@ read_packet:
tv.tv_sec = 2;
tv.tv_usec = 0;
- FD_ZERO(&read_socket);
- FD_SET(trace_fd, &read_socket);
-
- res = select(trace_fd + 1, &read_socket, NULL, NULL, &tv);
-
- if (res <= 0)
+ read_len = icmp_interface_read((struct batadv_icmp_header *)&icmp_packet_in,
+ sizeof(icmp_packet_in), &tv);
+ if (read_len <= 0)
continue;
- read_len = read(trace_fd, (char *)&icmp_packet_in, sizeof(icmp_packet_in));
-
- if (read_len < 0) {
- fprintf(stderr, "Error - can't read from batman adv kernel file '%s': %s\n", icmp_socket, strerror(errno));
- continue;
- }
-
if ((size_t)read_len < sizeof(icmp_packet_in)) {
printf("Warning - dropping received packet as it is smaller than expected (%zu): %zd\n",
sizeof(icmp_packet_in), read_len);
@@ -241,8 +224,7 @@ read_packet:
ret = EXIT_SUCCESS;
out:
+ icmp_interfaces_clean();
bat_hosts_free();
- if (trace_fd >= 0)
- close(trace_fd);
return ret;
}
--
2.8.1
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [B.A.T.M.A.N.] [1/4] batctl: Replace list implementation with linux-like-list
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list Sven Eckelmann
@ 2016-10-18 12:24 ` Sven Eckelmann
0 siblings, 0 replies; 7+ messages in thread
From: Sven Eckelmann @ 2016-10-18 12:24 UTC (permalink / raw)
To: b.a.t.m.a.n
[-- Attachment #1: Type: text/plain, Size: 1371 bytes --]
On Dienstag, 19. Juli 2016 22:44:08 CEST Sven Eckelmann wrote:
> The list-batman implementation of linked lists has an interface which tries
> to resample the linux double linked list implementation but in reality is
> incompatible. This resulted in code which moves most of the pointer
> arithmetic back to the actual code. This is rather unfortunate because this
> abstraction should reduce the pointer arithmetic and not increase it.
>
> Also the incompatibilities between the list implementation used in
> batman-adv and the one used by batctl made coding features in both
> components error prone because each component required a slightly different
> implementation for the same list manipulation.
>
> Signed-off-by: Sven Eckelmann <sven@narfation.org>
> ---
> Makefile | 1 -
> bisect_iv.c | 40 ++-
> bisect_iv.h | 10 +-
> list-batman.c | 123 ---------
> list-batman.h | 120 ---------
> list.h | 834
> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ tcpdump.c |
> 7 +-
> tcpdump.h | 2 +-
> 8 files changed, 860 insertions(+), 277 deletions(-)
> delete mode 100644 list-batman.c
> delete mode 100644 list-batman.h
> create mode 100644 list.h
Applied in 315b3b0fac147b9d8dc8f83a6409fde91eca683a [1].
Kind regards,
Sven
[1] https://git.open-mesh.org/batctl.git/commit/315b3b0fac147b9d8dc8f83a6409fde91eca683a
[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 801 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [B.A.T.M.A.N.] [2/4] batctl: Use monotonic time source for icmp timing
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 2/4] batctl: Use monotonic time source for icmp timing Sven Eckelmann
@ 2016-10-18 12:25 ` Sven Eckelmann
0 siblings, 0 replies; 7+ messages in thread
From: Sven Eckelmann @ 2016-10-18 12:25 UTC (permalink / raw)
To: b.a.t.m.a.n
[-- Attachment #1: Type: text/plain, Size: 811 bytes --]
On Dienstag, 19. Juli 2016 22:44:09 CEST Sven Eckelmann wrote:
> gettimeofday is not monotonic. It would give wrong results in some
> situation like leap seconds, when switching between daylight saving time
> and non-DST, NTP changes the time or when the administrator of the system
> changes it manually. The function is also obsolete since POSIX.1-2008.
>
> clock_gettime is recommended to get a monotonic timesource and as general
> replacement of gettimeofday.
>
> Signed-off-by: Sven Eckelmann <sven@narfation.org>
> ---
> Makefile | 2 +-
> functions.c | 17 +++++++++--------
> 2 files changed, 10 insertions(+), 9 deletions(-)
Applied in ef7cfaa81f72292f67afe50303ac657f41280853 [1].
Kind regards,
Sven
[1] https://git.open-mesh.org/batctl.git/commit/ef7cfaa81f72292f67afe50303ac657f41280853
[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 801 bytes --]
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2016-10-18 12:25 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-07-19 20:43 [B.A.T.M.A.N.] [PATCH 0/4] batctl: Userspace batadv_icmp endpoints Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 1/4] batctl: Replace list implementation with linux-like-list Sven Eckelmann
2016-10-18 12:24 ` [B.A.T.M.A.N.] [1/4] " Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 2/4] batctl: Use monotonic time source for icmp timing Sven Eckelmann
2016-10-18 12:25 ` [B.A.T.M.A.N.] [2/4] " Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 3/4] batctl: Add helper to generate instant random bytes Sven Eckelmann
2016-07-19 20:44 ` [B.A.T.M.A.N.] [PATCH 4/4] batctl: Implement non-routing batadv_icmp in userspace Sven Eckelmann
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).