[git commit master] dhcp: truncate packets instead of padding them to 574 bytes. closes bug 1849

Denys Vlasenko vda.linux at googlemail.com
Sun May 30 22:45:09 UTC 2010


commit: http://git.busybox.net/busybox/commit/?id=b8b72f02f01017d0f9584666fa572221f2b58613
branch: http://git.busybox.net/busybox/commit/?id=refs/heads/master

function                                             old     new   delta
udhcp_send_raw_packet                                411     456     +45
udhcp_send_kernel_packet                             259     271     +12

Signed-off-by: Denys Vlasenko <vda.linux at googlemail.com>
---
 networking/udhcp/packet.c |   32 ++++++++++++++++++++------------
 1 files changed, 20 insertions(+), 12 deletions(-)

diff --git a/networking/udhcp/packet.c b/networking/udhcp/packet.c
index c01fb7d..4badc9c 100644
--- a/networking/udhcp/packet.c
+++ b/networking/udhcp/packet.c
@@ -173,6 +173,7 @@ int FAST_FUNC udhcp_send_raw_packet(struct dhcp_packet *dhcp_pkt,
 {
 	struct sockaddr_ll dest_sll;
 	struct ip_udp_dhcp_packet packet;
+	unsigned padding;
 	int fd;
 	int result = -1;
 	const char *msg;
@@ -198,34 +199,40 @@ int FAST_FUNC udhcp_send_raw_packet(struct dhcp_packet *dhcp_pkt,
 		goto ret_close;
 	}
 
+	/* We were sending full-sized DHCP packets (zero padded),
+	 * but some badly configured servers were seen dropping them.
+	 * Apparently they drop all DHCP packets >576 *ethernet* octets big,
+	 * whereas they may only drop packets >576 *IP* octets big
+	 * (which for typical Ethernet II means 590 octets: 6+6+2 + 576).
+	 *
+	 * In order to work with those buggy servers,
+	 * we truncate packets after end option byte.
+	 */
+	padding = DHCP_OPTIONS_BUFSIZE - 1 - udhcp_end_option(packet.data.options);
+
 	packet.ip.protocol = IPPROTO_UDP;
 	packet.ip.saddr = source_nip;
 	packet.ip.daddr = dest_nip;
 	packet.udp.source = htons(source_port);
 	packet.udp.dest = htons(dest_port);
 	/* size, excluding IP header: */
-	packet.udp.len = htons(UPD_DHCP_SIZE);
+	packet.udp.len = htons(UPD_DHCP_SIZE - padding);
 	/* for UDP checksumming, ip.len is set to UDP packet len */
 	packet.ip.tot_len = packet.udp.len;
-	packet.udp.check = udhcp_checksum(&packet, IP_UPD_DHCP_SIZE);
+	packet.udp.check = udhcp_checksum(&packet, IP_UPD_DHCP_SIZE - padding);
 	/* but for sending, it is set to IP packet len */
-	packet.ip.tot_len = htons(IP_UPD_DHCP_SIZE);
+	packet.ip.tot_len = htons(IP_UPD_DHCP_SIZE - padding);
 	packet.ip.ihl = sizeof(packet.ip) >> 2;
 	packet.ip.version = IPVERSION;
 	packet.ip.ttl = IPDEFTTL;
 	packet.ip.check = udhcp_checksum(&packet.ip, sizeof(packet.ip));
 
-	/* Currently we send full-sized DHCP packets (zero padded).
-	 * If you need to change this: last byte of the packet is
-	 * packet.data.options[udhcp_end_option(packet.data.options)]
-	 */
 	udhcp_dump_packet(dhcp_pkt);
-	result = sendto(fd, &packet, IP_UPD_DHCP_SIZE, 0,
+	result = sendto(fd, &packet, IP_UPD_DHCP_SIZE - padding, /*flags:*/ 0,
 			(struct sockaddr *) &dest_sll, sizeof(dest_sll));
 	msg = "sendto";
  ret_close:
 	close(fd);
-	/* FIXME: and if result >= 0 but != IP_UPD_DHCP_SIZE? */
 	if (result < 0) {
  ret_msg:
 		bb_perror_msg(msg, "PACKET");
@@ -239,6 +246,7 @@ int FAST_FUNC udhcp_send_kernel_packet(struct dhcp_packet *dhcp_pkt,
 		uint32_t dest_nip, int dest_port)
 {
 	struct sockaddr_in client;
+	unsigned padding;
 	int fd;
 	int result = -1;
 	const char *msg;
@@ -268,13 +276,13 @@ int FAST_FUNC udhcp_send_kernel_packet(struct dhcp_packet *dhcp_pkt,
 		goto ret_close;
 	}
 
-	/* Currently we send full-sized DHCP packets (see above) */
 	udhcp_dump_packet(dhcp_pkt);
-	result = safe_write(fd, dhcp_pkt, DHCP_SIZE);
+
+	padding = DHCP_OPTIONS_BUFSIZE - 1 - udhcp_end_option(dhcp_pkt->options);
+	result = safe_write(fd, dhcp_pkt, DHCP_SIZE - padding);
 	msg = "write";
  ret_close:
 	close(fd);
-	/* FIXME: and if result >= 0 but != DHCP_SIZE? */
 	if (result < 0) {
  ret_msg:
 		bb_perror_msg(msg, "UDP");
-- 
1.6.3.3



More information about the busybox-cvs mailing list