1
0
mirror of https://passt.top/passt synced 2024-12-22 13:45:32 +00:00
passt/dhcp.c

400 lines
9.3 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-or-later
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 06:25:09 +00:00
/* PASST - Plug A Simple Socket Transport
passt: Add PASTA mode, major rework PASTA (Pack A Subtle Tap Abstraction) provides quasi-native host connectivity to an otherwise disconnected, unprivileged network and user namespace, similarly to slirp4netns. Given that the implementation is largely overlapping with PASST, no separate binary is built: 'pasta' (and 'passt4netns' for clarity) both link to 'passt', and the mode of operation is selected depending on how the binary is invoked. Usage example: $ unshare -rUn # echo $$ 1871759 $ ./pasta 1871759 # From another terminal # udhcpc -i pasta0 2>/dev/null # ping -c1 pasta.pizza PING pasta.pizza (64.190.62.111) 56(84) bytes of data. 64 bytes from 64.190.62.111 (64.190.62.111): icmp_seq=1 ttl=255 time=34.6 ms --- pasta.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 34.575/34.575/34.575/0.000 ms # ping -c1 spaghetti.pizza PING spaghetti.pizza(2606:4700:3034::6815:147a (2606:4700:3034::6815:147a)) 56 data bytes 64 bytes from 2606:4700:3034::6815:147a (2606:4700:3034::6815:147a): icmp_seq=1 ttl=255 time=29.0 ms --- spaghetti.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 28.967/28.967/28.967/0.000 ms This entails a major rework, especially with regard to the storage of tracked connections and to the semantics of epoll(7) references. Indexing TCP and UDP bindings merely by socket proved to be inflexible and unsuitable to handle different connection flows: pasta also provides Layer-2 to Layer-2 socket mapping between init and a separate namespace for local connections, using a pair of splice() system calls for TCP, and a recvmmsg()/sendmmsg() pair for UDP local bindings. For instance, building on the previous example: # ip link set dev lo up # iperf3 -s $ iperf3 -c ::1 -Z -w 32M -l 1024k -P2 | tail -n4 [SUM] 0.00-10.00 sec 52.3 GBytes 44.9 Gbits/sec 283 sender [SUM] 0.00-10.43 sec 52.3 GBytes 43.1 Gbits/sec receiver iperf Done. epoll(7) references now include a generic part in order to demultiplex data to the relevant protocol handler, using 24 bits for the socket number, and an opaque portion reserved for usage by the single protocol handlers, in order to track sockets back to corresponding connections and bindings. A number of fixes pertaining to TCP state machine and congestion window handling are also included here. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-07-17 06:34:53 +00:00
* for qemu/UNIX domain socket mode
*
* PASTA - Pack A Subtle Tap Abstraction
* for network namespace/tap device mode
*
* dhcp.c - Minimalistic DHCP server for PASST
*
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 06:25:09 +00:00
* Copyright (c) 2020-2021 Red Hat GmbH
* Author: Stefano Brivio <sbrivio@redhat.com>
*/
#include <arpa/inet.h>
#include <net/if.h>
#include <netinet/if_ether.h>
#include <netinet/ip.h>
#include <netinet/udp.h>
#include <stddef.h>
#include <stdio.h>
#include <stdint.h>
#include <unistd.h>
#include <string.h>
#include <limits.h>
passt: Add PASTA mode, major rework PASTA (Pack A Subtle Tap Abstraction) provides quasi-native host connectivity to an otherwise disconnected, unprivileged network and user namespace, similarly to slirp4netns. Given that the implementation is largely overlapping with PASST, no separate binary is built: 'pasta' (and 'passt4netns' for clarity) both link to 'passt', and the mode of operation is selected depending on how the binary is invoked. Usage example: $ unshare -rUn # echo $$ 1871759 $ ./pasta 1871759 # From another terminal # udhcpc -i pasta0 2>/dev/null # ping -c1 pasta.pizza PING pasta.pizza (64.190.62.111) 56(84) bytes of data. 64 bytes from 64.190.62.111 (64.190.62.111): icmp_seq=1 ttl=255 time=34.6 ms --- pasta.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 34.575/34.575/34.575/0.000 ms # ping -c1 spaghetti.pizza PING spaghetti.pizza(2606:4700:3034::6815:147a (2606:4700:3034::6815:147a)) 56 data bytes 64 bytes from 2606:4700:3034::6815:147a (2606:4700:3034::6815:147a): icmp_seq=1 ttl=255 time=29.0 ms --- spaghetti.pizza ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 28.967/28.967/28.967/0.000 ms This entails a major rework, especially with regard to the storage of tracked connections and to the semantics of epoll(7) references. Indexing TCP and UDP bindings merely by socket proved to be inflexible and unsuitable to handle different connection flows: pasta also provides Layer-2 to Layer-2 socket mapping between init and a separate namespace for local connections, using a pair of splice() system calls for TCP, and a recvmmsg()/sendmmsg() pair for UDP local bindings. For instance, building on the previous example: # ip link set dev lo up # iperf3 -s $ iperf3 -c ::1 -Z -w 32M -l 1024k -P2 | tail -n4 [SUM] 0.00-10.00 sec 52.3 GBytes 44.9 Gbits/sec 283 sender [SUM] 0.00-10.43 sec 52.3 GBytes 43.1 Gbits/sec receiver iperf Done. epoll(7) references now include a generic part in order to demultiplex data to the relevant protocol handler, using 24 bits for the socket number, and an opaque portion reserved for usage by the single protocol handlers, in order to track sockets back to corresponding connections and bindings. A number of fixes pertaining to TCP state machine and congestion window handling are also included here. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-07-17 06:34:53 +00:00
#include "util.h"
#include "ip.h"
#include "checksum.h"
#include "packet.h"
#include "passt.h"
passt: New design and implementation with native Layer 4 sockets This is a reimplementation, partially building on the earlier draft, that uses L4 sockets (SOCK_DGRAM, SOCK_STREAM) instead of SOCK_RAW, providing L4-L2 translation functionality without requiring any security capability. Conceptually, this follows the design presented at: https://gitlab.com/abologna/kubevirt-and-kvm/-/blob/master/Networking.md The most significant novelty here comes from TCP and UDP translation layers. In particular, the TCP state and translation logic follows the intent of being minimalistic, without reimplementing a full TCP stack in either direction, and synchronising as much as possible the TCP dynamic and flows between guest and host kernel. Another important introduction concerns addressing, port translation and forwarding. The Layer 4 implementations now attempt to bind on all unbound ports, in order to forward connections in a transparent way. While at it: - the qemu 'tap' back-end can't be used as-is by qrap anymore, because of explicit checks now introduced in qemu to ensure that the corresponding file descriptor is actually a tap device. For this reason, qrap now operates on a 'socket' back-end type, accounting for and building the additional header reporting frame length - provide a demo script that sets up namespaces, addresses and routes, and starts the daemon. A virtual machine started in the network namespace, wrapped by qrap, will now directly interface with passt and communicate using Layer 4 sockets provided by the host kernel. Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-16 06:25:09 +00:00
#include "tap.h"
#include "log.h"
#include "dhcp.h"
/**
* struct opt - DHCP option
* @sent: Convenience flag, set while filling replies
* @slen: Length of option defined for server, -1 if not going to be sent
* @s: Option payload from server
* @clen: Length of option received from client, -1 if not received
* @c: Option payload from client
*/
struct opt {
int sent;
int slen;
uint8_t s[255];
int clen;
uint8_t c[255];
};
static struct opt opts[255];
#define DHCPDISCOVER 1
#define DHCPOFFER 2
#define DHCPREQUEST 3
#define DHCPDECLINE 4
#define DHCPACK 5
#define DHCPNAK 6
#define DHCPRELEASE 7
#define DHCPINFORM 8
#define DHCPFORCERENEW 9
#define OPT_MIN 60 /* RFC 951 */
/**
* dhcp_init() - Initialise DHCP options
*/
void dhcp_init(void)
{
int i;
for (i = 0; i < ARRAY_SIZE(opts); i++)
opts[i].slen = -1;
opts[1] = (struct opt) { 0, 4, { 0 }, 0, { 0 }, }; /* Mask */
opts[3] = (struct opt) { 0, 4, { 0 }, 0, { 0 }, }; /* Router */
opts[51] = (struct opt) { 0, 4, { 0xff,
0xff,
0xff,
0xff }, 0, { 0 }, }; /* Lease time */
opts[53] = (struct opt) { 0, 1, { 0 }, 0, { 0 }, }; /* Type */
opts[54] = (struct opt) { 0, 4, { 0 }, 0, { 0 }, }; /* Server ID */
}
/**
* struct msg - BOOTP/DHCP message
* @op: BOOTP message type
* @htype: Hardware address type
* @hlen: Hardware address length
* @hops: DHCP relay hops
* @xid: Transaction ID randomly chosen by client
* @secs: Seconds elapsed since beginning of acquisition or renewal
* @flags: DHCP message flags
* @ciaddr: Client IP address in BOUND, RENEW, REBINDING
* @yiaddr: IP address being offered or assigned
* @siaddr: Next server to use in bootstrap
* @giaddr: Relay agent IP address
* @chaddr: Client hardware address
* @sname: Server host name
* @file: Boot file name
* @magic: Magic cookie prefix before options
* @o: Options
*/
struct msg {
uint8_t op;
#define BOOTREQUEST 1
#define BOOTREPLY 2
uint8_t htype;
uint8_t hlen;
uint8_t hops;
uint32_t xid;
uint16_t secs;
uint16_t flags;
uint32_t ciaddr;
struct in_addr yiaddr;
uint32_t siaddr;
uint32_t giaddr;
uint8_t chaddr[16];
uint8_t sname[64];
uint8_t file[128];
uint32_t magic;
uint8_t o[308];
} __attribute__((__packed__));
/**
* fill_one() - Fill a single option in message
* @m: Message to fill
* @o: Option number
* @offset: Current offset within options field, updated on insertion
*/
static void fill_one(struct msg *m, int o, int *offset)
{
m->o[*offset] = o;
m->o[*offset + 1] = opts[o].slen;
memcpy(&m->o[*offset + 2], opts[o].s, opts[o].slen);
opts[o].sent = 1;
*offset += 2 + opts[o].slen;
}
/**
* fill() - Fill options in message
* @m: Message to fill
*
* Return: current size of options field
*/
static int fill(struct msg *m)
{
int i, o, offset = 0;
m->op = BOOTREPLY;
m->secs = 0;
for (o = 0; o < 255; o++)
opts[o].sent = 0;
/* Some clients (wattcp32, mTCP, maybe some others) expect
* option 53 at the beginning of the list.
* Put it there explicitly, unless requested via option 55.
*/
if (opts[55].clen > 0 && !memchr(opts[55].c, 53, opts[55].clen))
fill_one(m, 53, &offset);
for (i = 0; i < opts[55].clen; i++) {
o = opts[55].c[i];
if (opts[o].slen != -1)
fill_one(m, o, &offset);
}
for (o = 0; o < 255; o++) {
if (opts[o].slen != -1 && !opts[o].sent)
fill_one(m, o, &offset);
}
m->o[offset++] = 255;
m->o[offset++] = 0;
if (offset < OPT_MIN) {
memset(&m->o[offset], 0, OPT_MIN - offset);
offset = OPT_MIN;
}
return offset;
}
/**
* opt_dns_search_dup_ptr() - Look for possible domain name compression pointer
* @buf: Current option buffer with existing labels
* @cmp: Portion of domain name being added
* @len: Length of current option buffer
*
* Return: offset to corresponding compression pointer if any, -1 if not found
*/
static int opt_dns_search_dup_ptr(unsigned char *buf, const char *cmp,
size_t len)
{
unsigned int i;
for (i = 0; i < len; i++) {
if (buf[i] == 0 &&
len - i - 1 >= strlen(cmp) &&
!memcmp(buf + i + 1, cmp, strlen(cmp)))
return i;
if ((buf[i] & 0xc0) == 0xc0 &&
len - i - 2 >= strlen(cmp) &&
!memcmp(buf + i + 2, cmp, strlen(cmp)))
return i + 1;
}
return -1;
}
/**
* opt_set_dns_search() - Fill data and set length for Domain Search option
* @c: Execution context
* @max_len: Maximum total length of option buffer
*/
static void opt_set_dns_search(const struct ctx *c, size_t max_len)
{
char buf[NS_MAXDNAME];
int i;
opts[119].slen = 0;
for (i = 0; i < 255; i++)
max_len -= opts[i].slen;
for (i = 0; *c->dns_search[i].n; i++) {
unsigned int n;
int count = -1;
const char *p;
buf[0] = 0;
for (p = c->dns_search[i].n, n = 1; *p; p++) {
if (*p == '.') {
/* RFC 1035 4.1.4 Message compression */
count = opt_dns_search_dup_ptr(opts[119].s,
p + 1,
opts[119].slen);
if (count >= 0) {
buf[n++] = '\xc0';
buf[n++] = count;
break;
}
buf[n++] = '.';
} else {
buf[n++] = *p;
}
}
/* The compression pointer is also an end of label */
if (count < 0)
buf[n++] = 0;
if (n >= max_len)
break;
memcpy(opts[119].s + opts[119].slen, buf, n);
opts[119].slen += n;
max_len -= n;
}
for (i = 0; i < opts[119].slen; i++) {
if (!opts[119].s[i] || opts[119].s[i] == '.') {
opts[119].s[i] = strcspn((char *)opts[119].s + i + 1,
".\xc0");
}
}
if (!opts[119].slen)
opts[119].slen = -1;
}
/**
* dhcp() - Check if this is a DHCP message, reply as needed
* @c: Execution context
* @p: Packet pool, single packet with Ethernet buffer
*
* Return: 0 if it's not a DHCP message, 1 if handled, -1 on failure
*/
int dhcp(const struct ctx *c, const struct pool *p)
{
size_t mlen, dlen, offset = 0, opt_len, opt_off = 0;
char macstr[ETH_ADDRSTRLEN];
const struct ethhdr *eh;
const struct iphdr *iph;
const struct udphdr *uh;
struct in_addr mask;
passt: Assorted fixes from "fresh eyes" review A bunch of fixes not worth single commits at this stage, notably: - make buffer, length parameter ordering consistent in ARP, DHCP, NDP handlers - strict checking of buffer, message and option length in DHCP handler (a malicious client could have easily crashed it) - set up forwarding for IPv4 and IPv6, and masquerading with nft for IPv4, from demo script - get rid of separate slow and fast timers, we don't save any overhead that way - stricter checking of buffer lengths as passed to tap handlers - proper dequeuing from qemu socket back-end: I accidentally trashed messages that were bundled up together in a single tap read operation -- the length header tells us what's the size of the next frame, but there's no apparent limit to the number of messages we get with one single receive - rework some bits of the TCP state machine, now passive and active connection closes appear to be robust -- introduce a new FIN_WAIT_1_SOCK_FIN state indicating a FIN_WAIT_1 with a FIN flag from socket - streamline TCP option parsing routine - track TCP state changes to stderr (this is temporary, proper debugging and syslogging support pending) - observe that multiplying a number by four might very well change its value, and this happens to be the case for the data offset from the TCP header as we check if it's the same as the total length to find out if it's a duplicated ACK segment - recent estimates suggest that the duration of a millisecond is closer to a million nanoseconds than a thousand of them, this trend is now reflected into the timespec_diff_ms() convenience routine Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-21 10:33:38 +00:00
unsigned int i;
struct msg *m;
eh = packet_get(p, 0, offset, sizeof(*eh), NULL);
offset += sizeof(*eh);
passt: Assorted fixes from "fresh eyes" review A bunch of fixes not worth single commits at this stage, notably: - make buffer, length parameter ordering consistent in ARP, DHCP, NDP handlers - strict checking of buffer, message and option length in DHCP handler (a malicious client could have easily crashed it) - set up forwarding for IPv4 and IPv6, and masquerading with nft for IPv4, from demo script - get rid of separate slow and fast timers, we don't save any overhead that way - stricter checking of buffer lengths as passed to tap handlers - proper dequeuing from qemu socket back-end: I accidentally trashed messages that were bundled up together in a single tap read operation -- the length header tells us what's the size of the next frame, but there's no apparent limit to the number of messages we get with one single receive - rework some bits of the TCP state machine, now passive and active connection closes appear to be robust -- introduce a new FIN_WAIT_1_SOCK_FIN state indicating a FIN_WAIT_1 with a FIN flag from socket - streamline TCP option parsing routine - track TCP state changes to stderr (this is temporary, proper debugging and syslogging support pending) - observe that multiplying a number by four might very well change its value, and this happens to be the case for the data offset from the TCP header as we check if it's the same as the total length to find out if it's a duplicated ACK segment - recent estimates suggest that the duration of a millisecond is closer to a million nanoseconds than a thousand of them, this trend is now reflected into the timespec_diff_ms() convenience routine Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-21 10:33:38 +00:00
iph = packet_get(p, 0, offset, sizeof(*iph), NULL);
if (!eh || !iph)
return -1;
passt: Assorted fixes from "fresh eyes" review A bunch of fixes not worth single commits at this stage, notably: - make buffer, length parameter ordering consistent in ARP, DHCP, NDP handlers - strict checking of buffer, message and option length in DHCP handler (a malicious client could have easily crashed it) - set up forwarding for IPv4 and IPv6, and masquerading with nft for IPv4, from demo script - get rid of separate slow and fast timers, we don't save any overhead that way - stricter checking of buffer lengths as passed to tap handlers - proper dequeuing from qemu socket back-end: I accidentally trashed messages that were bundled up together in a single tap read operation -- the length header tells us what's the size of the next frame, but there's no apparent limit to the number of messages we get with one single receive - rework some bits of the TCP state machine, now passive and active connection closes appear to be robust -- introduce a new FIN_WAIT_1_SOCK_FIN state indicating a FIN_WAIT_1 with a FIN flag from socket - streamline TCP option parsing routine - track TCP state changes to stderr (this is temporary, proper debugging and syslogging support pending) - observe that multiplying a number by four might very well change its value, and this happens to be the case for the data offset from the TCP header as we check if it's the same as the total length to find out if it's a duplicated ACK segment - recent estimates suggest that the duration of a millisecond is closer to a million nanoseconds than a thousand of them, this trend is now reflected into the timespec_diff_ms() convenience routine Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-21 10:33:38 +00:00
offset += iph->ihl * 4UL;
uh = packet_get(p, 0, offset, sizeof(*uh), &mlen);
offset += sizeof(*uh);
if (!uh)
return -1;
if (uh->dest != htons(67))
return 0;
if (c->no_dhcp)
return 1;
m = packet_get(p, 0, offset, offsetof(struct msg, o), &opt_len);
if (!m ||
mlen != ntohs(uh->len) - sizeof(*uh) ||
mlen < offsetof(struct msg, o) ||
m->op != BOOTREQUEST)
return -1;
offset += offsetof(struct msg, o);
for (i = 0; i < ARRAY_SIZE(opts); i++)
opts[i].clen = -1;
while (opt_off + 2 < opt_len) {
const uint8_t *olen, *val;
uint8_t *type;
type = packet_get(p, 0, offset + opt_off, 1, NULL);
olen = packet_get(p, 0, offset + opt_off + 1, 1, NULL);
if (!type || !olen)
return -1;
val = packet_get(p, 0, offset + opt_off + 2, *olen, NULL);
if (!val)
passt: Assorted fixes from "fresh eyes" review A bunch of fixes not worth single commits at this stage, notably: - make buffer, length parameter ordering consistent in ARP, DHCP, NDP handlers - strict checking of buffer, message and option length in DHCP handler (a malicious client could have easily crashed it) - set up forwarding for IPv4 and IPv6, and masquerading with nft for IPv4, from demo script - get rid of separate slow and fast timers, we don't save any overhead that way - stricter checking of buffer lengths as passed to tap handlers - proper dequeuing from qemu socket back-end: I accidentally trashed messages that were bundled up together in a single tap read operation -- the length header tells us what's the size of the next frame, but there's no apparent limit to the number of messages we get with one single receive - rework some bits of the TCP state machine, now passive and active connection closes appear to be robust -- introduce a new FIN_WAIT_1_SOCK_FIN state indicating a FIN_WAIT_1 with a FIN flag from socket - streamline TCP option parsing routine - track TCP state changes to stderr (this is temporary, proper debugging and syslogging support pending) - observe that multiplying a number by four might very well change its value, and this happens to be the case for the data offset from the TCP header as we check if it's the same as the total length to find out if it's a duplicated ACK segment - recent estimates suggest that the duration of a millisecond is closer to a million nanoseconds than a thousand of them, this trend is now reflected into the timespec_diff_ms() convenience routine Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-21 10:33:38 +00:00
return -1;
memcpy(&opts[*type].c, val, *olen);
opts[*type].clen = *olen;
opt_off += *olen + 2;
passt: Assorted fixes from "fresh eyes" review A bunch of fixes not worth single commits at this stage, notably: - make buffer, length parameter ordering consistent in ARP, DHCP, NDP handlers - strict checking of buffer, message and option length in DHCP handler (a malicious client could have easily crashed it) - set up forwarding for IPv4 and IPv6, and masquerading with nft for IPv4, from demo script - get rid of separate slow and fast timers, we don't save any overhead that way - stricter checking of buffer lengths as passed to tap handlers - proper dequeuing from qemu socket back-end: I accidentally trashed messages that were bundled up together in a single tap read operation -- the length header tells us what's the size of the next frame, but there's no apparent limit to the number of messages we get with one single receive - rework some bits of the TCP state machine, now passive and active connection closes appear to be robust -- introduce a new FIN_WAIT_1_SOCK_FIN state indicating a FIN_WAIT_1 with a FIN flag from socket - streamline TCP option parsing routine - track TCP state changes to stderr (this is temporary, proper debugging and syslogging support pending) - observe that multiplying a number by four might very well change its value, and this happens to be the case for the data offset from the TCP header as we check if it's the same as the total length to find out if it's a duplicated ACK segment - recent estimates suggest that the duration of a millisecond is closer to a million nanoseconds than a thousand of them, this trend is now reflected into the timespec_diff_ms() convenience routine Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2021-02-21 10:33:38 +00:00
}
if (opts[53].clen > 0 && opts[53].c[0] == DHCPDISCOVER) {
info("DHCP: offer to discover");
opts[53].s[0] = DHCPOFFER;
} else if (opts[53].clen <= 0 || opts[53].c[0] == DHCPREQUEST) {
info("%s: ack to request", /* DHCP needs a valid message type */
(opts[53].clen <= 0) ? "BOOTP" : "DHCP");
opts[53].s[0] = DHCPACK;
} else {
return -1;
}
info(" from %s", eth_ntop(m->chaddr, macstr, sizeof(macstr)));
m->yiaddr = c->ip4.addr;
mask.s_addr = htonl(0xffffffff << (32 - c->ip4.prefix_len));
fwd: Split notion of "our tap address" from gateway for IPv4 ip4.gw conflates 3 conceptually different things, which (for now) have the same value: 1. The router/gateway address as seen by the guest 2. An address to NAT to the host with --no-map-gw isn't specified 3. An address to use as source when nothing else makes sense Case 3 occurs in two situations: a) for our DHCP responses - since they come from passt internally there's no naturally meaningful address for them to come from b) for forwarded connections coming from an address that isn't guest accessible (localhost or the guest's own address). (b) occurs even with --no-map-gw, and the expected behaviour of forwarding local connections requires it. For IPv6 role (3) is now taken by ip6.our_tap_ll (which usually has the same value as ip6.gw). For future flexibility we may want to make this "address of last resort" different from the gateway address, so split them logically for IPv4 as well. Specifically, add a new ip4.our_tap_addr field for the address with this role, and initialise it to ip4.gw for now. Unlike IPv6 where we can always get a link-local address, we might not be able to get a (non 0.0.0.0) address here (e.g. if the host is disconnected or only has a point to point link with no gateway address). In that case we have to disable forwarding of inbound connections with guest-inaccessible source addresses. Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2024-08-21 04:20:13 +00:00
memcpy(opts[1].s, &mask, sizeof(mask));
memcpy(opts[3].s, &c->ip4.guest_gw, sizeof(c->ip4.guest_gw));
fwd: Split notion of "our tap address" from gateway for IPv4 ip4.gw conflates 3 conceptually different things, which (for now) have the same value: 1. The router/gateway address as seen by the guest 2. An address to NAT to the host with --no-map-gw isn't specified 3. An address to use as source when nothing else makes sense Case 3 occurs in two situations: a) for our DHCP responses - since they come from passt internally there's no naturally meaningful address for them to come from b) for forwarded connections coming from an address that isn't guest accessible (localhost or the guest's own address). (b) occurs even with --no-map-gw, and the expected behaviour of forwarding local connections requires it. For IPv6 role (3) is now taken by ip6.our_tap_ll (which usually has the same value as ip6.gw). For future flexibility we may want to make this "address of last resort" different from the gateway address, so split them logically for IPv4 as well. Specifically, add a new ip4.our_tap_addr field for the address with this role, and initialise it to ip4.gw for now. Unlike IPv6 where we can always get a link-local address, we might not be able to get a (non 0.0.0.0) address here (e.g. if the host is disconnected or only has a point to point link with no gateway address). In that case we have to disable forwarding of inbound connections with guest-inaccessible source addresses. Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2024-08-21 04:20:13 +00:00
memcpy(opts[54].s, &c->ip4.our_tap_addr, sizeof(c->ip4.our_tap_addr));
/* If the gateway is not on the assigned subnet, send an option 121
* (Classless Static Routing) adding a dummy route to it.
*/
if ((c->ip4.addr.s_addr & mask.s_addr)
!= (c->ip4.guest_gw.s_addr & mask.s_addr)) {
/* a.b.c.d/32:0.0.0.0, 0:a.b.c.d */
opts[121].slen = 14;
opts[121].s[0] = 32;
memcpy(opts[121].s + 1,
&c->ip4.guest_gw, sizeof(c->ip4.guest_gw));
memcpy(opts[121].s + 10,
&c->ip4.guest_gw, sizeof(c->ip4.guest_gw));
}
if (c->mtu != -1) {
opts[26].slen = 2;
opts[26].s[0] = c->mtu / 256;
opts[26].s[1] = c->mtu % 256;
}
for (i = 0, opts[6].slen = 0;
!c->no_dhcp_dns && !IN4_IS_ADDR_UNSPECIFIED(&c->ip4.dns[i]); i++) {
((struct in_addr *)opts[6].s)[i] = c->ip4.dns[i];
opts[6].slen += sizeof(uint32_t);
}
if (!opts[6].slen)
opts[6].slen = -1;
if (!c->no_dhcp_dns_search)
opt_set_dns_search(c, sizeof(m->o));
dlen = offsetof(struct msg, o) + fill(m);
fwd: Split notion of "our tap address" from gateway for IPv4 ip4.gw conflates 3 conceptually different things, which (for now) have the same value: 1. The router/gateway address as seen by the guest 2. An address to NAT to the host with --no-map-gw isn't specified 3. An address to use as source when nothing else makes sense Case 3 occurs in two situations: a) for our DHCP responses - since they come from passt internally there's no naturally meaningful address for them to come from b) for forwarded connections coming from an address that isn't guest accessible (localhost or the guest's own address). (b) occurs even with --no-map-gw, and the expected behaviour of forwarding local connections requires it. For IPv6 role (3) is now taken by ip6.our_tap_ll (which usually has the same value as ip6.gw). For future flexibility we may want to make this "address of last resort" different from the gateway address, so split them logically for IPv4 as well. Specifically, add a new ip4.our_tap_addr field for the address with this role, and initialise it to ip4.gw for now. Unlike IPv6 where we can always get a link-local address, we might not be able to get a (non 0.0.0.0) address here (e.g. if the host is disconnected or only has a point to point link with no gateway address). In that case we have to disable forwarding of inbound connections with guest-inaccessible source addresses. Signed-off-by: David Gibson <david@gibson.dropbear.id.au> Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
2024-08-21 04:20:13 +00:00
tap_udp4_send(c, c->ip4.our_tap_addr, 67, c->ip4.addr, 68, m, dlen);
return 1;
}