/*
- * Copyright (c) 2009, 2010, 2012 Nicira, Inc.
+ * Copyright (c) 2009, 2010, 2012, 2013, 2014 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
#include "pcap-file.h"
#include <errno.h>
#include <inttypes.h>
+#include <stdlib.h>
#include <string.h>
+#include <sys/stat.h>
+#include "byte-order.h"
#include "compiler.h"
+#include "flow.h"
+#include "hmap.h"
#include "ofpbuf.h"
+#include "packets.h"
+#include "timeval.h"
+#include "unaligned.h"
#include "vlog.h"
VLOG_DEFINE_THIS_MODULE(pcap);
BUILD_ASSERT_DECL(sizeof(struct pcaprec_hdr) == 16);
FILE *
-pcap_open(const char *file_name, const char *mode)
+ovs_pcap_open(const char *file_name, const char *mode)
{
+ struct stat s;
FILE *file;
+ int error;
- ovs_assert(!strcmp(mode, "rb") || !strcmp(mode, "wb"));
+ ovs_assert(!strcmp(mode, "rb") ||
+ !strcmp(mode, "wb") ||
+ !strcmp(mode, "ab"));
file = fopen(file_name, mode);
if (file == NULL) {
- VLOG_WARN("%s: failed to open pcap file for %s",
- file_name, mode[0] == 'r' ? "reading" : "writing");
+ VLOG_WARN("%s: failed to open pcap file for %s (%s)", file_name,
+ (mode[0] == 'r' ? "reading"
+ : mode[0] == 'w' ? "writing"
+ : "appending"),
+ ovs_strerror(errno));
return NULL;
}
- if (mode[0] == 'r') {
- if (!pcap_read_header(file)) {
+ switch (mode[0]) {
+ case 'r':
+ error = ovs_pcap_read_header(file);
+ if (error) {
+ errno = error;
fclose(file);
return NULL;
}
- } else {
- pcap_write_header(file);
+ break;
+
+ case 'w':
+ ovs_pcap_write_header(file);
+ break;
+
+ case 'a':
+ if (!fstat(fileno(file), &s) && !s.st_size) {
+ ovs_pcap_write_header(file);
+ }
+ break;
+
+ default:
+ OVS_NOT_REACHED();
}
return file;
}
int
-pcap_read_header(FILE *file)
+ovs_pcap_read_header(FILE *file)
{
struct pcap_hdr ph;
if (fread(&ph, sizeof ph, 1, file) != 1) {
}
void
-pcap_write_header(FILE *file)
+ovs_pcap_write_header(FILE *file)
{
/* The pcap reader is responsible for figuring out endianness based on the
* magic number, so the lack of htonX calls here is intentional. */
}
int
-pcap_read(FILE *file, struct ofpbuf **bufp)
+ovs_pcap_read(FILE *file, struct ofpbuf **bufp, long long int *when)
{
struct pcaprec_hdr prh;
struct ofpbuf *buf;
void *data;
size_t len;
+ bool swap;
*bufp = NULL;
/* Read header. */
if (fread(&prh, sizeof prh, 1, file) != 1) {
- int error = ferror(file) ? errno : EOF;
- VLOG_WARN("failed to read pcap record header: %s",
- ovs_retval_to_string(error));
- return error;
+ if (ferror(file)) {
+ int error = errno;
+ VLOG_WARN("failed to read pcap record header: %s",
+ ovs_retval_to_string(error));
+ return error;
+ } else {
+ return EOF;
+ }
}
/* Calculate length. */
len = prh.incl_len;
- if (len > 0xffff) {
- uint32_t swapped_len = (((len & 0xff000000) >> 24) |
- ((len & 0x00ff0000) >> 8) |
- ((len & 0x0000ff00) << 8) |
- ((len & 0x000000ff) << 24));
- if (swapped_len > 0xffff) {
- VLOG_WARN("bad packet length %zu or %"PRIu32" "
+ swap = len > 0xffff;
+ if (swap) {
+ len = uint32_byteswap(len);
+ if (len > 0xffff) {
+ VLOG_WARN("bad packet length %"PRIuSIZE" or %"PRIu32
"reading pcap file",
- len, swapped_len);
+ len, uint32_byteswap(len));
return EPROTO;
}
- len = swapped_len;
+ }
+
+ /* Calculate time. */
+ if (when) {
+ uint32_t ts_sec = swap ? uint32_byteswap(prh.ts_sec) : prh.ts_sec;
+ uint32_t ts_usec = swap ? uint32_byteswap(prh.ts_usec) : prh.ts_usec;
+ *when = ts_sec * 1000LL + ts_usec / 1000;
}
/* Read packet. */
}
void
-pcap_write(FILE *file, struct ofpbuf *buf)
+ovs_pcap_write(FILE *file, struct ofpbuf *buf)
{
struct pcaprec_hdr prh;
- prh.ts_sec = 0;
- prh.ts_usec = 0;
- prh.incl_len = buf->size;
- prh.orig_len = buf->size;
+ struct timeval tv;
+
+ xgettimeofday(&tv);
+ prh.ts_sec = tv.tv_sec;
+ prh.ts_usec = tv.tv_usec;
+ prh.incl_len = ofpbuf_size(buf);
+ prh.orig_len = ofpbuf_size(buf);
ignore(fwrite(&prh, sizeof prh, 1, file));
- ignore(fwrite(buf->data, buf->size, 1, file));
+ ignore(fwrite(ofpbuf_data(buf), ofpbuf_size(buf), 1, file));
+}
+\f
+struct tcp_key {
+ ovs_be32 nw_src, nw_dst;
+ ovs_be16 tp_src, tp_dst;
+};
+
+struct tcp_stream {
+ struct hmap_node hmap_node;
+ struct tcp_key key;
+ uint32_t seq_no;
+ struct ofpbuf payload;
+};
+
+struct tcp_reader {
+ struct hmap streams;
+};
+
+static void
+tcp_stream_destroy(struct tcp_reader *r, struct tcp_stream *stream)
+{
+ hmap_remove(&r->streams, &stream->hmap_node);
+ ofpbuf_uninit(&stream->payload);
+ free(stream);
+}
+
+/* Returns a new data structure for extracting TCP stream data from an
+ * Ethernet packet capture */
+struct tcp_reader *
+tcp_reader_open(void)
+{
+ struct tcp_reader *r;
+
+ r = xmalloc(sizeof *r);
+ hmap_init(&r->streams);
+ return r;
+}
+
+/* Closes and frees 'r'. */
+void
+tcp_reader_close(struct tcp_reader *r)
+{
+ struct tcp_stream *stream, *next_stream;
+
+ HMAP_FOR_EACH_SAFE (stream, next_stream, hmap_node, &r->streams) {
+ tcp_stream_destroy(r, stream);
+ }
+ hmap_destroy(&r->streams);
+ free(r);
+}
+
+static struct tcp_stream *
+tcp_stream_lookup(struct tcp_reader *r,
+ const struct tcp_key *key, uint32_t hash)
+{
+ struct tcp_stream *stream;
+
+ HMAP_FOR_EACH_WITH_HASH (stream, hmap_node, hash, &r->streams) {
+ if (!memcmp(&stream->key, key, sizeof *key)) {
+ return stream;
+ }
+ }
+ return NULL;
+}
+
+static struct tcp_stream *
+tcp_stream_new(struct tcp_reader *r, const struct tcp_key *key, uint32_t hash)
+{
+ struct tcp_stream *stream;
+
+ stream = xmalloc(sizeof *stream);
+ hmap_insert(&r->streams, &stream->hmap_node, hash);
+ memcpy(&stream->key, key, sizeof *key);
+ stream->seq_no = 0;
+ ofpbuf_init(&stream->payload, 2048);
+ return stream;
+}
+
+/* Processes 'packet' through TCP reader 'r'. The caller must have already
+ * extracted the packet's headers into 'flow', using flow_extract().
+ *
+ * If 'packet' is a TCP packet, then the reader attempts to reconstruct the
+ * data stream. If successful, it returns an ofpbuf that represents the data
+ * stream so far. The caller may examine the data in the ofpbuf and pull off
+ * any data that it has fully processed. The remaining data that the caller
+ * does not pull off will be presented again in future calls if more data
+ * arrives in the stream.
+ *
+ * Returns null if 'packet' doesn't add new data to a TCP stream. */
+struct ofpbuf *
+tcp_reader_run(struct tcp_reader *r, const struct flow *flow,
+ const struct ofpbuf *packet)
+{
+ struct tcp_stream *stream;
+ struct tcp_header *tcp;
+ struct ofpbuf *payload;
+ unsigned int l7_length;
+ struct tcp_key key;
+ uint32_t hash;
+ uint32_t seq;
+ uint8_t flags;
+ const char *l7 = ofpbuf_get_tcp_payload(packet);
+
+ if (flow->dl_type != htons(ETH_TYPE_IP)
+ || flow->nw_proto != IPPROTO_TCP
+ || !l7) {
+ return NULL;
+ }
+ tcp = ofpbuf_l4(packet);
+ flags = TCP_FLAGS(tcp->tcp_ctl);
+ l7_length = (char *) ofpbuf_tail(packet) - l7;
+ seq = ntohl(get_16aligned_be32(&tcp->tcp_seq));
+
+ /* Construct key. */
+ memset(&key, 0, sizeof key);
+ key.nw_src = flow->nw_src;
+ key.nw_dst = flow->nw_dst;
+ key.tp_src = flow->tp_src;
+ key.tp_dst = flow->tp_dst;
+ hash = hash_bytes(&key, sizeof key, 0);
+
+ /* Find existing stream or start a new one for a SYN or if there's data. */
+ stream = tcp_stream_lookup(r, &key, hash);
+ if (!stream) {
+ if (flags & TCP_SYN || l7_length) {
+ stream = tcp_stream_new(r, &key, hash);
+ stream->seq_no = flags & TCP_SYN ? seq + 1 : seq;
+ } else {
+ return NULL;
+ }
+ }
+
+ payload = &stream->payload;
+ if (flags & TCP_SYN || !stream->seq_no) {
+ ofpbuf_clear(payload);
+ stream->seq_no = seq + 1;
+ return NULL;
+ } else if (flags & (TCP_FIN | TCP_RST)) {
+ tcp_stream_destroy(r, stream);
+ return NULL;
+ } else if (seq == stream->seq_no) {
+ /* Shift all of the existing payload to the very beginning of the
+ * allocated space, so that we reuse allocated space instead of
+ * continually expanding it. */
+ ofpbuf_shift(payload, (char *) ofpbuf_base(payload) - (char *) ofpbuf_data(payload));
+
+ ofpbuf_put(payload, l7, l7_length);
+ stream->seq_no += l7_length;
+ return payload;
+ } else {
+ return NULL;
+ }
}