Router: client IP address replacement.

This commit introduces the replacement of the client address based on the value
of a specified HTTP header.  This is intended for use when Unit is placed
behind a reverse proxy like nginx or a CDN.

You must specify the source addresses of the trusted proxies.  This can be
accomplished with any valid IP pattern supported by Unit's match block:

["10.0.0.1", "10.4.0.0/16", "!192.168.1.1"]

The feature is configured per listener.

The client address replacement functionality only operates when there is a
source IP match and the specified header is present.  Typically this would be
an 'X-Forwarded-For' header.

{
    "listeners": {
        "127.0.0.1:8080": {
            "client_ip": {
                "header": "X-Forwarded-For",
                "source": [
                    "10.0.0.0/8"
                ]
            },
            "pass": "applications/my_app"
        },
    }
}

If a request occurs and Unit receives a header like below:

"X-Forwarded-For: 84.123.23.23"

By default, Unit trusts the last rightmost IP in the header, so REMOTE_ADDR
will be set to 84.123.23.23 if the connection originated from 10.0.0.0/8.

If Unit runs behind consecutive reverse proxies and receives a header similar
to the following:

"X-Forwarded-For: 84.123.23.23, 10.0.0.254"

You will need to enable "recursive" checking, which walks the header from
last address to first and chooses the first non-trusted address it finds.

{
    "listeners": {
        "127.0.0.1:8080": {
            "client_ip": {
                "header": "X-Forwarded-For",
                "source": [
                    "10.0.0.0/8"
                ]
                "recursive": true,
            },
            "pass": "applications/my_app"
        },
    }
}

If a connection from 10.0.0.0/8 occurs, the chain is walked.  Here, 10.0.0.254
is also a trusted address so the client address will be replaced with
84.123.23.23.

If all IP addresses in the header are trusted, the client address is set to
the first address in the header:

If 10.0.0.0/8 is trusted and "X-Forwarded-For: 10.0.0.3, 10.0.0.2, 10.0.0.1",
the client address will be replaced with 10.0.0.3.

6 files changed, 282 insertions, 11 deletions

diff --git a/src/nxt_conf_validation.c b/src/nxt_conf_validation.c
index 0ec2e811..0106ebc8 100644
--- a/src/nxt_conf_validation.c
+++ b/src/nxt_conf_validation.c
@@ -208,6 +208,7 @@ static nxt_conf_vldt_object_t  nxt_conf_vldt_setting_members[];
 static nxt_conf_vldt_object_t  nxt_conf_vldt_http_members[];
 static nxt_conf_vldt_object_t  nxt_conf_vldt_websocket_members[];
 static nxt_conf_vldt_object_t  nxt_conf_vldt_static_members[];
+static nxt_conf_vldt_object_t  nxt_conf_vldt_client_ip_members[];
 #if (NXT_TLS)
 static nxt_conf_vldt_object_t  nxt_conf_vldt_tls_members[];
 static nxt_conf_vldt_object_t  nxt_conf_vldt_session_members[];
@@ -351,6 +352,11 @@ static nxt_conf_vldt_object_t  nxt_conf_vldt_listener_members[] = {
         .name       = nxt_string("application"),
         .type       = NXT_CONF_VLDT_STRING,
         .validator  = nxt_conf_vldt_app_name,
+    }, {
+        .name       = nxt_string("client_ip"),
+        .type       = NXT_CONF_VLDT_OBJECT,
+        .validator  = nxt_conf_vldt_object,
+        .u.members  = nxt_conf_vldt_client_ip_members
     },
 
 #if (NXT_TLS)
@@ -366,6 +372,25 @@ static nxt_conf_vldt_object_t  nxt_conf_vldt_listener_members[] = {
 };
 
 
+static nxt_conf_vldt_object_t  nxt_conf_vldt_client_ip_members[] = {
+    {
+        .name       = nxt_string("source"),
+        .type       = NXT_CONF_VLDT_STRING | NXT_CONF_VLDT_ARRAY,
+        .validator  = nxt_conf_vldt_match_addrs,
+        .flags      = NXT_CONF_VLDT_REQUIRED
+    }, {
+        .name       = nxt_string("header"),
+        .type       = NXT_CONF_VLDT_STRING,
+        .flags      = NXT_CONF_VLDT_REQUIRED
+    }, {
+        .name       = nxt_string("recursive"),
+        .type       = NXT_CONF_VLDT_BOOLEAN,
+    },
+
+    NXT_CONF_VLDT_END
+};
+
+
 #if (NXT_TLS)
 
 static nxt_conf_vldt_object_t  nxt_conf_vldt_tls_members[] = {
diff --git a/src/nxt_http.h b/src/nxt_http.h
index faba83f2..3bc2fd61 100644
--- a/src/nxt_http.h
+++ b/src/nxt_http.h
@@ -197,8 +197,9 @@ struct nxt_http_request_s {
 };
 
 
-typedef struct nxt_http_route_s       nxt_http_route_t;
-typedef struct nxt_http_route_rule_s  nxt_http_route_rule_t;
+typedef struct nxt_http_route_s            nxt_http_route_t;
+typedef struct nxt_http_route_rule_s       nxt_http_route_rule_t;
+typedef struct nxt_http_route_addr_rule_s  nxt_http_route_addr_rule_t;
 
 
 typedef struct {
@@ -254,6 +255,14 @@ typedef struct {
 } nxt_http_proto_table_t;
 
 
+struct nxt_http_client_ip_s {
+    nxt_http_route_addr_rule_t  *source;
+    nxt_str_t                   *header;
+    uint32_t                    header_hash;
+    uint8_t                     recursive;    /* 1 bit */
+};
+
+
 #define NXT_HTTP_DATE_LEN  nxt_length("Wed, 31 Dec 1986 16:40:00 GMT")
 
 nxt_inline u_char *
@@ -311,6 +320,10 @@ nxt_int_t nxt_http_pass_segments(nxt_mp_t *mp, nxt_str_t *pass,
     nxt_str_t *segments, nxt_uint_t n);
 nxt_http_action_t *nxt_http_pass_application(nxt_task_t *task,
     nxt_router_conf_t *rtcf, nxt_str_t *name);
+nxt_http_route_addr_rule_t *nxt_http_route_addr_rule_create(
+    nxt_task_t *task, nxt_mp_t *mp, nxt_conf_value_t *cv);
+nxt_int_t nxt_http_route_addr_rule(nxt_http_request_t *r,
+    nxt_http_route_addr_rule_t *addr_rule, nxt_sockaddr_t *sockaddr);
 nxt_http_route_rule_t *nxt_http_route_types_rule_create(nxt_task_t *task,
     nxt_mp_t *mp, nxt_conf_value_t *types);
 nxt_int_t nxt_http_route_test_rule(nxt_http_request_t *r,
diff --git a/src/nxt_http_request.c b/src/nxt_http_request.c
index 16563a98..b71b25d9 100644
--- a/src/nxt_http_request.c
+++ b/src/nxt_http_request.c
@@ -10,6 +10,10 @@
 
 static nxt_int_t nxt_http_validate_host(nxt_str_t *host, nxt_mp_t *mp);
 static void nxt_http_request_start(nxt_task_t *task, void *obj, void *data);
+static nxt_int_t nxt_http_request_client_ip(nxt_task_t *task,
+    nxt_http_request_t *r);
+static nxt_sockaddr_t *nxt_http_request_client_ip_sockaddr(
+    nxt_http_request_t *r, u_char *start, size_t len);
 static void nxt_http_request_ready(nxt_task_t *task, void *obj, void *data);
 static void nxt_http_request_proto_info(nxt_task_t *task,
     nxt_http_request_t *r);
@@ -272,16 +276,162 @@ static const nxt_http_request_state_t  nxt_http_request_init_state
 static void
 nxt_http_request_start(nxt_task_t *task, void *obj, void *data)
 {
+    nxt_int_t           ret;
     nxt_http_request_t  *r;
 
     r = obj;
 
     r->state = &nxt_http_request_body_state;
 
+    ret = nxt_http_request_client_ip(task, r);
+    if (nxt_slow_path(ret != NXT_OK)) {
+        nxt_http_request_error(task, r, NXT_HTTP_INTERNAL_SERVER_ERROR);
+    }
+
     nxt_http_request_read_body(task, r);
 }
 
 
+static nxt_int_t
+nxt_http_request_client_ip(nxt_task_t *task, nxt_http_request_t *r)
+{
+    u_char                *start, *p;
+    nxt_int_t             ret, i, len;
+    nxt_str_t             *header;
+    nxt_array_t           *fields_arr;  /* of nxt_http_field_t * */
+    nxt_sockaddr_t        *sa, *prev_sa;
+    nxt_http_field_t      *f, **fields;
+    nxt_http_client_ip_t  *client_ip;
+
+    client_ip = r->conf->socket_conf->client_ip;
+
+    if (client_ip == NULL) {
+        return NXT_OK;
+    }
+
+    ret = nxt_http_route_addr_rule(r, client_ip->source, r->remote);
+    if (ret <= 0) {
+        return NXT_OK;
+    }
+
+    header = client_ip->header;
+
+    fields_arr = nxt_array_create(r->mem_pool, 2, sizeof(nxt_http_field_t *));
+    if (nxt_slow_path(fields_arr == NULL)) {
+        return NXT_ERROR;
+    }
+
+    nxt_list_each(f, r->fields) {
+        if (f->hash == client_ip->header_hash
+            && f->name_length == client_ip->header->length
+            && f->value_length > 0
+            && nxt_memcasecmp(f->name, header->start, header->length) == 0)
+        {
+            fields = nxt_array_add(fields_arr);
+            if (nxt_slow_path(fields == NULL)) {
+                return NXT_ERROR;
+            }
+
+            *fields = f;
+        }
+    } nxt_list_loop;
+
+    prev_sa = r->remote;
+    fields = (nxt_http_field_t **) fields_arr->elts;
+
+    i = fields_arr->nelts;
+
+    while (i-- > 0) {
+        f = fields[i];
+        start = f->value;
+        len = f->value_length;
+
+        do {
+            for (p = start + len - 1; p > start; p--, len--) {
+                if (*p != ' ' && *p != ',') {
+                    break;
+                }
+            }
+
+            for (/* void */; p > start; p--) {
+                if (*p == ' ' || *p == ',') {
+                    p++;
+                    break;
+                }
+            }
+
+            sa = nxt_http_request_client_ip_sockaddr(r, p, len - (p - start));
+            if (nxt_slow_path(sa == NULL)) {
+                if (prev_sa != NULL) {
+                    r->remote = prev_sa;
+                }
+
+                return NXT_OK;
+            }
+
+            if (!client_ip->recursive) {
+                r->remote = sa;
+
+                return NXT_OK;
+            }
+
+            ret = nxt_http_route_addr_rule(r, client_ip->source, sa);
+            if (ret <= 0 || (i == 0 && p == start)) {
+                r->remote = sa;
+
+                return NXT_OK;
+            }
+
+            prev_sa = sa;
+            len = p - 1 - start;
+
+        } while (len > 0);
+    }
+
+    return NXT_OK;
+}
+
+
+static nxt_sockaddr_t *
+nxt_http_request_client_ip_sockaddr(nxt_http_request_t *r, u_char *start,
+    size_t len)
+{
+    nxt_str_t       addr;
+    nxt_sockaddr_t  *sa;
+
+    addr.start = start;
+    addr.length = len;
+
+    sa = nxt_sockaddr_parse_optport(r->mem_pool, &addr);
+    if (nxt_slow_path(sa == NULL)) {
+        return NULL;
+    }
+
+    switch (sa->u.sockaddr.sa_family) {
+        case AF_INET:
+            if (sa->u.sockaddr_in.sin_addr.s_addr == INADDR_ANY) {
+                return NULL;
+            }
+
+            break;
+
+#if (NXT_INET6)
+        case AF_INET6:
+            if (IN6_IS_ADDR_UNSPECIFIED(&sa->u.sockaddr_in6.sin6_addr)) {
+                return NULL;
+            }
+
+            break;
+#endif /* NXT_INET6 */
+
+        default:
+            return NULL;
+    }
+
+    return sa;
+}
+
+
 static const nxt_http_request_state_t  nxt_http_request_body_state
     nxt_aligned(64) =
 {
diff --git a/src/nxt_http_route.c b/src/nxt_http_route.c
index b330796f..cff69f96 100644
--- a/src/nxt_http_route.c
+++ b/src/nxt_http_route.c
@@ -135,12 +135,12 @@ typedef struct {
 } nxt_http_route_table_t;
 
 
-typedef struct {
+struct nxt_http_route_addr_rule_s {
     /* The object must be the first field. */
     nxt_http_route_object_t        object:8;
     uint32_t                       items;
     nxt_http_route_addr_pattern_t  addr_pattern[0];
-} nxt_http_route_addr_rule_t;
+};
 
 
 typedef union {
@@ -194,8 +194,6 @@ static nxt_http_route_ruleset_t *nxt_http_route_ruleset_create(nxt_task_t *task,
 static nxt_http_route_rule_t *nxt_http_route_rule_name_create(nxt_task_t *task,
     nxt_mp_t *mp, nxt_conf_value_t *rule_cv, nxt_str_t *name,
     nxt_bool_t case_sensitive, nxt_http_route_encoding_t encoding);
-static nxt_http_route_addr_rule_t *nxt_http_route_addr_rule_create(
-    nxt_task_t *task, nxt_mp_t *mp, nxt_conf_value_t *cv);
 static nxt_http_route_rule_t *nxt_http_route_rule_create(nxt_task_t *task,
     nxt_mp_t *mp, nxt_conf_value_t *cv, nxt_bool_t case_sensitive,
     nxt_http_route_pattern_case_t pattern_case,
@@ -237,8 +235,6 @@ static nxt_int_t nxt_http_route_table(nxt_http_request_t *r,
     nxt_http_route_table_t *table);
 static nxt_int_t nxt_http_route_ruleset(nxt_http_request_t *r,
     nxt_http_route_ruleset_t *ruleset);
-static nxt_int_t nxt_http_route_addr_rule(nxt_http_request_t *r,
-    nxt_http_route_addr_rule_t *addr_rule, nxt_sockaddr_t *sockaddr);
 static nxt_int_t nxt_http_route_rule(nxt_http_request_t *r,
     nxt_http_route_rule_t *rule);
 static nxt_int_t nxt_http_route_header(nxt_http_request_t *r,
@@ -940,7 +936,7 @@ nxt_http_route_rule_create(nxt_task_t *task, nxt_mp_t *mp,
 }
 
 
-static nxt_http_route_addr_rule_t *
+nxt_http_route_addr_rule_t *
 nxt_http_route_addr_rule_create(nxt_task_t *task, nxt_mp_t *mp,
      nxt_conf_value_t *cv)
 {
@@ -1927,7 +1923,7 @@ nxt_http_route_addr_pattern_match(nxt_http_route_addr_pattern_t *p,
 }
 
 
-static nxt_int_t
+nxt_int_t
 nxt_http_route_addr_rule(nxt_http_request_t *r,
     nxt_http_route_addr_rule_t *addr_rule, nxt_sockaddr_t *sa)
 {
diff --git a/src/nxt_router.c b/src/nxt_router.c
index 8360e75a..1aa919c2 100644
--- a/src/nxt_router.c
+++ b/src/nxt_router.c
@@ -107,6 +107,9 @@ static nxt_int_t nxt_router_conf_create(nxt_task_t *task,
     nxt_router_temp_conf_t *tmcf, u_char *start, u_char *end);
 static nxt_int_t nxt_router_conf_process_static(nxt_task_t *task,
     nxt_router_conf_t *rtcf, nxt_conf_value_t *conf);
+static nxt_int_t nxt_router_conf_process_client_ip(nxt_task_t *task,
+    nxt_router_temp_conf_t *tmcf, nxt_socket_conf_t *skcf,
+    nxt_conf_value_t *conf);
 
 static nxt_app_t *nxt_router_app_find(nxt_queue_t *queue, nxt_str_t *name);
 static nxt_int_t nxt_router_apps_hash_test(nxt_lvlhsh_query_t *lhq, void *data);
@@ -1450,7 +1453,7 @@ nxt_router_conf_create(nxt_task_t *task, nxt_router_temp_conf_t *tmcf,
     nxt_conf_value_t            *conf, *http, *value, *websocket;
     nxt_conf_value_t            *applications, *application;
     nxt_conf_value_t            *listeners, *listener;
-    nxt_conf_value_t            *routes_conf, *static_conf;
+    nxt_conf_value_t            *routes_conf, *static_conf, *client_ip_conf;
     nxt_socket_conf_t           *skcf;
     nxt_http_routes_t           *routes;
     nxt_event_engine_t          *engine;
@@ -1472,6 +1475,7 @@ nxt_router_conf_create(nxt_task_t *task, nxt_router_temp_conf_t *tmcf,
 #endif
     static nxt_str_t  static_path = nxt_string("/settings/http/static");
     static nxt_str_t  websocket_path = nxt_string("/settings/http/websocket");
+    static nxt_str_t  client_ip_path = nxt_string("/client_ip");
 
     conf = nxt_conf_json_parse(tmcf->mem_pool, start, end, NULL);
     if (conf == NULL) {
@@ -1843,6 +1847,13 @@ nxt_router_conf_create(nxt_task_t *task, nxt_router_temp_conf_t *tmcf,
                 t->length = nxt_strlen(t->start);
             }
 
+            client_ip_conf = nxt_conf_get_path(listener, &client_ip_path);
+            ret = nxt_router_conf_process_client_ip(task, tmcf, skcf,
+                                                    client_ip_conf);
+            if (nxt_slow_path(ret != NXT_OK)) {
+                return NXT_ERROR;
+            }
+
 #if (NXT_TLS)
             certificate = nxt_conf_get_path(listener, &certificate_path);
 
@@ -2085,6 +2096,79 @@ nxt_router_conf_process_static(nxt_task_t *task, nxt_router_conf_t *rtcf,
 }
 
 
+static nxt_int_t
+nxt_router_conf_process_client_ip(nxt_task_t *task, nxt_router_temp_conf_t *tmcf,
+    nxt_socket_conf_t *skcf, nxt_conf_value_t *conf)
+{
+    char                        c;
+    size_t                      i;
+    nxt_mp_t                    *mp;
+    uint32_t                    hash;
+    nxt_str_t                   header;
+    nxt_conf_value_t            *source_conf, *header_conf, *recursive_conf;
+    nxt_http_client_ip_t        *client_ip;
+    nxt_http_route_addr_rule_t  *source;
+
+    static nxt_str_t  header_path = nxt_string("/header");
+    static nxt_str_t  source_path = nxt_string("/source");
+    static nxt_str_t  recursive_path = nxt_string("/recursive");
+
+    if (conf == NULL) {
+        skcf->client_ip = NULL;
+
+        return NXT_OK;
+    }
+
+    mp = tmcf->router_conf->mem_pool;
+
+    source_conf = nxt_conf_get_path(conf, &source_path);
+    header_conf = nxt_conf_get_path(conf, &header_path);
+    recursive_conf = nxt_conf_get_path(conf, &recursive_path);
+
+    if (source_conf == NULL || header_conf == NULL) {
+        return NXT_ERROR;
+    }
+
+    client_ip = nxt_mp_zget(mp, sizeof(nxt_http_client_ip_t));
+    if (nxt_slow_path(client_ip == NULL)) {
+        return NXT_ERROR;
+    }
+
+    source = nxt_http_route_addr_rule_create(task, mp, source_conf);
+    if (nxt_slow_path(source == NULL)) {
+        return NXT_ERROR;
+    }
+
+    client_ip->source = source;
+
+    nxt_conf_get_string(header_conf, &header);
+
+    if (recursive_conf != NULL) {
+        client_ip->recursive = nxt_conf_get_boolean(recursive_conf);
+    }
+
+    client_ip->header = nxt_str_dup(mp, NULL, &header);
+    if (nxt_slow_path(client_ip->header == NULL)) {
+        return NXT_ERROR;
+    }
+
+    hash = NXT_HTTP_FIELD_HASH_INIT;
+
+    for (i = 0; i < client_ip->header->length; i++) {
+        c = client_ip->header->start[i];
+        hash = nxt_http_field_hash_char(hash, nxt_lowcase(c));
+    }
+
+    hash = nxt_http_field_hash_end(hash) & 0xFFFF;
+
+    client_ip->header_hash = hash;
+
+    skcf->client_ip = client_ip;
+
+    return NXT_OK;
+}
+
+
 static nxt_app_t *
 nxt_router_app_find(nxt_queue_t *queue, nxt_str_t *name)
 {
diff --git a/src/nxt_router.h b/src/nxt_router.h
index 6611cf45..fc068b53 100644
--- a/src/nxt_router.h
+++ b/src/nxt_router.h
@@ -18,6 +18,7 @@ typedef struct nxt_http_request_s  nxt_http_request_t;
 
 typedef struct nxt_http_action_s        nxt_http_action_t;
 typedef struct nxt_http_routes_s        nxt_http_routes_t;
+typedef struct nxt_http_client_ip_s     nxt_http_client_ip_t;
 typedef struct nxt_upstream_s           nxt_upstream_t;
 typedef struct nxt_upstreams_s          nxt_upstreams_t;
 typedef struct nxt_router_access_log_s  nxt_router_access_log_t;
@@ -195,6 +196,8 @@ typedef struct {
 
     uint8_t                discard_unsafe_fields;  /* 1 bit */
 
+    nxt_http_client_ip_t   *client_ip;
+
 #if (NXT_TLS)
     nxt_tls_conf_t         *tls;
 #endif