libwebsockets/lib/server.c

2770 lines
65 KiB
C
Raw Permalink Normal View History

/*
* libwebsockets - small server side websockets and web server implementation
*
* Copyright (C) 2010-2016 Andy Green <andy@warmcat.com>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation:
* version 2.1 of the License.
*
* This library 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
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston,
* MA 02110-1301 USA
*/
#include "private-libwebsockets.h"
2016-07-23 14:18:25 +08:00
#if defined (LWS_WITH_ESP8266)
#undef memcpy
void *memcpy(void *dest, const void *src, size_t n)
{
return ets_memcpy(dest, src, n);
}
#endif
int
lws_context_init_server(struct lws_context_creation_info *info,
struct lws_vhost *vhost)
{
2016-07-23 14:18:25 +08:00
#if LWS_POSIX
int n, opt = 1, limit = 1;
#endif
lws_sockfd_type sockfd;
struct lws_vhost *vh;
struct lws *wsi;
int m = 0;
/* set up our external listening socket we serve on */
if (info->port == CONTEXT_PORT_NO_LISTEN)
return 0;
vh = vhost->context->vhost_list;
while (vh) {
if (vh->listen_port == info->port) {
if ((!info->iface && !vh->iface) ||
(info->iface && vh->iface &&
!strcmp(info->iface, vh->iface))) {
vhost->listen_port = info->port;
vhost->iface = info->iface;
lwsl_notice(" using listen skt from vhost %s\n",
vh->name);
return 0;
}
}
vh = vh->vhost_next;
}
#if LWS_POSIX
#if defined(__linux__)
limit = vhost->context->count_threads;
#endif
for (m = 0; m < limit; m++) {
2016-03-30 22:47:02 -07:00
#ifdef LWS_USE_UNIX_SOCK
if (LWS_UNIX_SOCK_ENABLED(vhost))
2016-03-30 22:47:02 -07:00
sockfd = socket(AF_UNIX, SOCK_STREAM, 0);
else
#endif
#ifdef LWS_USE_IPV6
if (LWS_IPV6_ENABLED(vhost))
sockfd = socket(AF_INET6, SOCK_STREAM, 0);
else
#endif
sockfd = socket(AF_INET, SOCK_STREAM, 0);
if (sockfd == -1) {
2016-07-23 14:18:25 +08:00
#else
#if defined(LWS_WITH_ESP8266)
sockfd = esp8266_create_tcp_listen_socket(vhost);
if (!lws_sockfd_valid(sockfd)) {
#else
sockfd = mbed3_create_tcp_stream_socket();
if (!lws_sockfd_valid(sockfd)) {
2016-07-23 14:18:25 +08:00
#endif
#endif
lwsl_err("ERROR opening socket\n");
return 1;
}
#if LWS_POSIX
/*
* allow us to restart even if old sockets in TIME_WAIT
*/
if (setsockopt(sockfd, SOL_SOCKET, SO_REUSEADDR,
(const void *)&opt, sizeof(opt)) < 0) {
compatible_close(sockfd);
return 1;
}
#if defined(LWS_USE_IPV6) && defined(IPV6_V6ONLY)
if (LWS_IPV6_ENABLED(vhost)) {
if (vhost->options & LWS_SERVER_OPTION_IPV6_V6ONLY_MODIFY) {
int value = (vhost->options & LWS_SERVER_OPTION_IPV6_V6ONLY_VALUE) ? 1 : 0;
if (setsockopt(sockfd, IPPROTO_IPV6, IPV6_V6ONLY,
(const void*)&value, sizeof(value)) < 0) {
compatible_close(sockfd);
return 1;
}
}
}
#endif
#if defined(__linux__) && defined(SO_REUSEPORT) && LWS_MAX_SMP > 1
if (vhost->context->count_threads > 1)
if (setsockopt(sockfd, SOL_SOCKET, SO_REUSEPORT,
(const void *)&opt, sizeof(opt)) < 0) {
compatible_close(sockfd);
return 1;
}
#endif
#endif
lws_plat_set_socket_options(vhost, sockfd);
#if LWS_POSIX
n = lws_socket_bind(vhost, sockfd, info->port, info->iface);
if (n < 0)
goto bail;
info->port = n;
#endif
vhost->listen_port = info->port;
vhost->iface = info->iface;
wsi = lws_zalloc(sizeof(struct lws));
if (wsi == NULL) {
lwsl_err("Out of mem\n");
goto bail;
}
wsi->context = vhost->context;
wsi->sock = sockfd;
wsi->mode = LWSCM_SERVER_LISTENER;
wsi->protocol = vhost->protocols;
wsi->tsi = m;
wsi->vhost = vhost;
wsi->listener = 1;
vhost->context->pt[m].wsi_listening = wsi;
if (insert_wsi_socket_into_fds(vhost->context, wsi))
goto bail;
vhost->context->count_wsi_allocated++;
vhost->lserv_wsi = wsi;
#if LWS_POSIX
n = listen(wsi->sock, LWS_SOMAXCONN);
if (n < 0) {
lwsl_err("listen failed with error %d\n", LWS_ERRNO);
vhost->lserv_wsi = NULL;
vhost->context->count_wsi_allocated--;
remove_wsi_socket_from_fds(wsi);
vhost->context->pt[m].wsi_listening = NULL;
goto bail;
}
} /* for each thread able to independently listen */
2016-07-23 14:18:25 +08:00
#else
#if defined(LWS_WITH_ESP8266)
esp8266_tcp_stream_bind(wsi->sock, info->port, wsi);
#else
mbed3_tcp_stream_bind(wsi->sock, info->port, wsi);
2016-07-23 14:18:25 +08:00
#endif
#endif
2016-03-30 22:47:02 -07:00
if (!lws_check_opt(info->options, LWS_SERVER_OPTION_EXPLICIT_VHOSTS)) {
#ifdef LWS_USE_UNIX_SOCK
if (LWS_UNIX_SOCK_ENABLED(vhost))
2016-03-30 22:47:02 -07:00
lwsl_notice(" Listening on \"%s\"\n", info->iface);
else
#endif
lwsl_notice(" Listening on port %d\n", info->port);
}
return 0;
bail:
compatible_close(sockfd);
return 1;
}
int
_lws_server_listen_accept_flow_control(struct lws *twsi, int on)
{
struct lws_context_per_thread *pt = &twsi->context->pt[(int)twsi->tsi];
struct lws *wsi = pt->wsi_listening;
int n;
if (!wsi || twsi->context->being_destroyed)
return 0;
lwsl_debug("%s: Thr %d: LISTEN wsi %p: state %d\n",
__func__, twsi->tsi, (void *)wsi, on);
if (on)
n = lws_change_pollfd(wsi, 0, LWS_POLLIN);
else
n = lws_change_pollfd(wsi, LWS_POLLIN, 0);
return n;
}
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
#undef strchr
#define strchr ets_strchr
#endif
struct lws_vhost *
lws_select_vhost(struct lws_context *context, int port, const char *servername)
{
struct lws_vhost *vhost = context->vhost_list;
const char *p;
int n, m, colon;
n = strlen(servername);
colon = n;
p = strchr(servername, ':');
if (p)
colon = p - servername;
/* first try exact matches */
while (vhost) {
if (port == vhost->listen_port &&
!strncmp(vhost->name, servername, colon)) {
lwsl_info("SNI: Found: %s\n", servername);
return vhost;
}
vhost = vhost->vhost_next;
}
/*
* if no exact matches, try matching *.vhost-name
* unintentional matches are possible but resolve to x.com for *.x.com
* which is reasonable. If exact match exists we already chose it and
* never reach here. SSL will still fail it if the cert doesn't allow
* *.x.com.
*/
vhost = context->vhost_list;
while (vhost) {
m = strlen(vhost->name);
if (port == vhost->listen_port &&
m <= (colon - 2) &&
servername[colon - m - 1] == '.' &&
!strncmp(vhost->name, servername + colon - m, m)) {
lwsl_info("SNI: Found %s on wildcard: %s\n",
servername, vhost->name);
return vhost;
}
vhost = vhost->vhost_next;
}
return NULL;
}
LWS_VISIBLE LWS_EXTERN const struct lws_protocols *
lws_vhost_name_to_protocol(struct lws_vhost *vh, const char *name)
{
int n;
for (n = 0; n < vh->count_protocols; n++)
if (!strcmp(name, vh->protocols[n].name))
return &vh->protocols[n];
return NULL;
}
2016-07-03 09:20:11 +08:00
LWS_VISIBLE LWS_EXTERN const char *
lws_get_mimetype(const char *file, const struct lws_http_mount *m)
{
int n = strlen(file);
const struct lws_protocol_vhost_options *pvo = NULL;
if (m)
pvo = m->extra_mimetypes;
if (n < 5)
return NULL;
if (!strcmp(&file[n - 4], ".ico"))
return "image/x-icon";
if (!strcmp(&file[n - 4], ".gif"))
return "image/gif";
if (!strcmp(&file[n - 3], ".js"))
return "text/javascript";
if (!strcmp(&file[n - 4], ".png"))
return "image/png";
if (!strcmp(&file[n - 4], ".jpg"))
return "image/jpeg";
if (!strcmp(&file[n - 3], ".gz"))
return "application/gzip";
if (!strcmp(&file[n - 4], ".JPG"))
return "image/jpeg";
if (!strcmp(&file[n - 5], ".html"))
return "text/html";
if (!strcmp(&file[n - 4], ".css"))
return "text/css";
if (!strcmp(&file[n - 4], ".txt"))
return "text/plain";
2016-09-23 00:04:40 +02:00
if (!strcmp(&file[n - 4], ".svg"))
return "image/svg+xml";
if (!strcmp(&file[n - 4], ".ttf"))
return "application/x-font-ttf";
if (!strcmp(&file[n - 5], ".woff"))
return "application/font-woff";
if (!strcmp(&file[n - 4], ".xml"))
return "application/xml";
while (pvo) {
if (pvo->name[0] == '*') /* ie, match anything */
return pvo->value;
if (!strcmp(&file[n - strlen(pvo->name)], pvo->name))
return pvo->value;
pvo = pvo->next;
}
return NULL;
}
static int
lws_http_serve(struct lws *wsi, char *uri, const char *origin,
const struct lws_http_mount *m)
{
const struct lws_protocol_vhost_options *pvo = m->interpret;
struct lws_process_html_args args;
const char *mimetype;
2016-07-23 14:18:25 +08:00
#if !defined(_WIN32_WCE) && !defined(LWS_WITH_ESP8266)
struct stat st;
2016-07-23 14:18:25 +08:00
int spin = 0;
#endif
char path[256], sym[512];
unsigned char *p = (unsigned char *)sym + 32 + LWS_PRE, *start = p;
unsigned char *end = p + sizeof(sym) - 32 - LWS_PRE;
2016-07-23 14:18:25 +08:00
#if !defined(WIN32) && LWS_POSIX
2016-04-26 22:52:16 +02:00
size_t len;
#endif
2016-07-23 14:18:25 +08:00
int n;
lws_snprintf(path, sizeof(path) - 1, "%s/%s", origin, uri);
2016-07-23 14:18:25 +08:00
#if !defined(_WIN32_WCE) && !defined(LWS_WITH_ESP8266)
do {
spin++;
if (stat(path, &st)) {
lwsl_info("unable to stat %s\n", path);
goto bail;
}
lwsl_debug(" %s mode %d\n", path, S_IFMT & st.st_mode);
2016-07-23 14:18:25 +08:00
#if !defined(WIN32) && LWS_POSIX
if ((S_IFMT & st.st_mode) == S_IFLNK) {
len = readlink(path, sym, sizeof(sym) - 1);
if (len) {
lwsl_err("Failed to read link %s\n", path);
goto bail;
}
sym[len] = '\0';
lwsl_debug("symlink %s -> %s\n", path, sym);
lws_snprintf(path, sizeof(path) - 1, "%s", sym);
}
#endif
if ((S_IFMT & st.st_mode) == S_IFDIR) {
lwsl_debug("default filename append to dir\n");
lws_snprintf(path, sizeof(path) - 1, "%s/%s/index.html",
origin, uri);
}
} while ((S_IFMT & st.st_mode) != S_IFREG && spin < 5);
if (spin == 5)
lwsl_err("symlink loop %s \n", path);
n = sprintf(sym, "%08lX%08lX", (unsigned long)st.st_size,
(unsigned long)st.st_mtime);
if (lws_hdr_total_length(wsi, WSI_TOKEN_HTTP_IF_NONE_MATCH)) {
/*
* he thinks he has some version of it already,
* check if the tag matches
*/
if (!strcmp(sym, lws_hdr_simple_ptr(wsi, WSI_TOKEN_HTTP_IF_NONE_MATCH))) {
lwsl_debug("%s: ETAG match %s %s\n", __func__,
uri, origin);
/* we don't need to send the payload */
if (lws_add_http_header_status(wsi, 304, &p, end))
return -1;
if (lws_add_http_header_by_token(wsi,
WSI_TOKEN_HTTP_ETAG,
(unsigned char *)sym, n, &p, end))
return -1;
if (lws_finalize_http_header(wsi, &p, end))
return -1;
n = lws_write(wsi, start, p - start,
LWS_WRITE_HTTP_HEADERS);
if (n != (p - start)) {
lwsl_err("_write returned %d from %ld\n", n,
(long)(p - start));
return -1;
}
return lws_http_transaction_completed(wsi);
}
}
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_HTTP_ETAG,
(unsigned char *)sym, n, &p, end))
return -1;
#endif
2016-07-03 09:20:11 +08:00
mimetype = lws_get_mimetype(path, m);
if (!mimetype) {
2016-07-23 14:18:25 +08:00
lwsl_err("unknown mimetype for %s\n", path);
goto bail;
}
if (!mimetype[0])
lwsl_debug("sending no mimetype for %s\n", path);
wsi->sending_chunked = 0;
/*
* check if this is in the list of file suffixes to be interpreted by
* a protocol
*/
while (pvo) {
n = strlen(path);
if (n > (int)strlen(pvo->name) &&
!strcmp(&path[n - strlen(pvo->name)], pvo->name)) {
wsi->sending_chunked = 1;
wsi->protocol_interpret_idx = (char)(long)pvo->value;
lwsl_info("want %s interpreted by %s\n", path,
wsi->vhost->protocols[(int)(long)(pvo->value)].name);
wsi->protocol = &wsi->vhost->protocols[(int)(long)(pvo->value)];
if (lws_ensure_user_space(wsi))
return -1;
break;
}
pvo = pvo->next;
}
if (m->protocol) {
const struct lws_protocols *pp = lws_vhost_name_to_protocol(
wsi->vhost, m->protocol);
if (lws_bind_protocol(wsi, pp))
return 1;
args.p = (char *)p;
args.max_len = end - p;
if (pp->callback(wsi, LWS_CALLBACK_ADD_HEADERS,
wsi->user_space, &args, 0))
return -1;
p = (unsigned char *)args.p;
}
n = lws_serve_http_file(wsi, path, mimetype, (char *)start, p - start);
if (n < 0 || ((n > 0) && lws_http_transaction_completed(wsi)))
return -1; /* error or can't reuse connection: close the socket */
return 0;
bail:
return -1;
}
2016-06-26 06:29:20 +08:00
const struct lws_http_mount *
lws_find_mount(struct lws *wsi, const char *uri_ptr, int uri_len)
{
const struct lws_http_mount *hm, *hit = NULL;
int best = 0;
hm = wsi->vhost->mount_list;
while (hm) {
if (uri_len >= hm->mountpoint_len &&
!strncmp(uri_ptr, hm->mountpoint, hm->mountpoint_len) &&
(uri_ptr[hm->mountpoint_len] == '\0' ||
uri_ptr[hm->mountpoint_len] == '/' ||
hm->mountpoint_len == 1)
) {
if (hm->origin_protocol == LWSMPRO_CALLBACK ||
((hm->origin_protocol == LWSMPRO_CGI ||
lws_hdr_total_length(wsi, WSI_TOKEN_GET_URI) ||
hm->protocol) &&
hm->mountpoint_len > best)) {
best = hm->mountpoint_len;
hit = hm;
}
}
hm = hm->mount_next;
}
return hit;
}
int
lws_http_action(struct lws *wsi)
{
struct lws_context_per_thread *pt = &wsi->context->pt[(int)wsi->tsi];
HTTP Version, Keep-alive support, No-copy POST This is a squashed commit from https://github.com/andrew-canaday/libwebsockets, dev/http_keepalive branch (strategies changed a few times, so the commit history is clutteread). This branch is submitted for clarity, but the other can be used as a reference or alternative. * added **enum http_version** to track HTTP/1.0 vs HTTP/1.1 requests * added **enum http_connection_type** to track keep-alive vs close * replaced content_length_seen and body_index with **content_remain** * removed **post_buffer** (see handshake.c modifications) * removed post_buffer free * switch state to WSI_TOKEN_SKIPPING after URI is complete to store version * delete *spill* label (unused) * add vars to track HTTP version and connection type * HTTP version defaults to 1.0 * connection type defaults to 'close' for 1.0, keep-alive for 1.1 * additional checks in **cleanup:** label: * if HTTP version string is present and valid, set enum val appropriately * override connection default with the "Connection:" header, if present * set state to WSI_STATE_HTTP_BODY if content_length > 0 * return 0 on HTTP requests, unless LWS_CALLBACK_HTTP indicates otherwise * add vars to track remaining content_length and body chunk size * re-arrange switch case order to facilitate creation of jump-table * added new labels: * **read_ok**: normal location reach on break from switch; just return 0 * **http_complete**: check for keep-alive + init state, mode, hdr table * **http_new**: jump location for keep-alive when http_complete sees len>0 * after libwebsocket_parse, jump to one of those labels based on state * POST body handling: * don't bother iterating over input byte-by-byte or using memcpy * just pass the relevant portion of the context->service_buffer to callback
2014-07-13 01:07:36 -04:00
enum http_connection_type connection_type;
enum http_version request_version;
char content_length_str[32];
struct lws_process_html_args args;
2016-06-26 06:29:20 +08:00
const struct lws_http_mount *hit = NULL;
unsigned int n, count = 0;
char http_version_str[10];
char http_conn_str[20];
int http_version_len;
2016-06-26 06:29:20 +08:00
char *uri_ptr = NULL, *s;
int uri_len = 0;
int meth = -1;
static const unsigned char methods[] = {
WSI_TOKEN_GET_URI,
WSI_TOKEN_POST_URI,
WSI_TOKEN_OPTIONS_URI,
WSI_TOKEN_PUT_URI,
WSI_TOKEN_PATCH_URI,
WSI_TOKEN_DELETE_URI,
#ifdef LWS_USE_HTTP2
WSI_TOKEN_HTTP_COLON_PATH,
#endif
};
#if defined(_DEBUG) || defined(LWS_WITH_ACCESS_LOG)
static const char * const method_names[] = {
"GET", "POST", "OPTIONS", "PUT", "PATCH", "DELETE",
#ifdef LWS_USE_HTTP2
":path",
#endif
};
#endif
static const char * const oprot[] = {
"http://", "https://"
};
/* it's not websocket.... shall we accept it as http? */
for (n = 0; n < ARRAY_SIZE(methods); n++)
if (lws_hdr_total_length(wsi, methods[n]))
count++;
if (!count) {
lwsl_warn("Missing URI in HTTP request\n");
goto bail_nuke_ah;
}
if (count != 1) {
lwsl_warn("multiple methods?\n");
goto bail_nuke_ah;
}
if (lws_ensure_user_space(wsi))
goto bail_nuke_ah;
for (n = 0; n < ARRAY_SIZE(methods); n++)
if (lws_hdr_total_length(wsi, methods[n])) {
uri_ptr = lws_hdr_simple_ptr(wsi, methods[n]);
uri_len = lws_hdr_total_length(wsi, methods[n]);
lwsl_info("Method: %s request for '%s'\n",
method_names[n], uri_ptr);
meth = n;
break;
}
(void)meth;
/* we insist on absolute paths */
if (uri_ptr[0] != '/') {
lws_return_http_status(wsi, HTTP_STATUS_FORBIDDEN, NULL);
goto bail_nuke_ah;
}
/* HTTP header had a content length? */
wsi->u.http.content_length = 0;
if (lws_hdr_total_length(wsi, WSI_TOKEN_POST_URI) ||
lws_hdr_total_length(wsi, WSI_TOKEN_PATCH_URI) ||
lws_hdr_total_length(wsi, WSI_TOKEN_PUT_URI))
wsi->u.http.content_length = 100 * 1024 * 1024;
if (lws_hdr_total_length(wsi, WSI_TOKEN_HTTP_CONTENT_LENGTH)) {
lws_hdr_copy(wsi, content_length_str,
sizeof(content_length_str) - 1,
WSI_TOKEN_HTTP_CONTENT_LENGTH);
wsi->u.http.content_length = atoi(content_length_str);
}
if (wsi->http2_substream) {
wsi->u.http.request_version = HTTP_VERSION_2;
} else {
/* http_version? Default to 1.0, override with token: */
request_version = HTTP_VERSION_1_0;
/* Works for single digit HTTP versions. : */
http_version_len = lws_hdr_total_length(wsi, WSI_TOKEN_HTTP);
if (http_version_len > 7) {
lws_hdr_copy(wsi, http_version_str,
sizeof(http_version_str) - 1, WSI_TOKEN_HTTP);
if (http_version_str[5] == '1' && http_version_str[7] == '1')
request_version = HTTP_VERSION_1_1;
}
wsi->u.http.request_version = request_version;
/* HTTP/1.1 defaults to "keep-alive", 1.0 to "close" */
if (request_version == HTTP_VERSION_1_1)
connection_type = HTTP_CONNECTION_KEEP_ALIVE;
else
connection_type = HTTP_CONNECTION_CLOSE;
/* Override default if http "Connection:" header: */
if (lws_hdr_total_length(wsi, WSI_TOKEN_CONNECTION)) {
lws_hdr_copy(wsi, http_conn_str, sizeof(http_conn_str) - 1,
WSI_TOKEN_CONNECTION);
http_conn_str[sizeof(http_conn_str) - 1] = '\0';
if (!strcasecmp(http_conn_str, "keep-alive"))
connection_type = HTTP_CONNECTION_KEEP_ALIVE;
else
if (!strcasecmp(http_conn_str, "close"))
connection_type = HTTP_CONNECTION_CLOSE;
}
wsi->u.http.connection_type = connection_type;
}
n = wsi->protocol->callback(wsi, LWS_CALLBACK_FILTER_HTTP_CONNECTION,
wsi->user_space, uri_ptr, uri_len);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
if (n) {
lwsl_info("LWS_CALLBACK_HTTP closing\n");
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
return 1;
}
/*
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
* if there is content supposed to be coming,
* put a timeout on it having arrived
*/
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
lws_set_timeout(wsi, PENDING_TIMEOUT_HTTP_CONTENT,
wsi->context->timeout_secs);
#ifdef LWS_OPENSSL_SUPPORT
if (wsi->redirect_to_https) {
/*
* we accepted http:// only so we could redirect to
* https://, so issue the redirect. Create the redirection
* URI from the host: header and ignore the path part
*/
unsigned char *start = pt->serv_buf + LWS_PRE, *p = start,
*end = p + 512;
if (!lws_hdr_total_length(wsi, WSI_TOKEN_HOST))
goto bail_nuke_ah;
n = sprintf((char *)end, "https://%s/",
lws_hdr_simple_ptr(wsi, WSI_TOKEN_HOST));
n = lws_http_redirect(wsi, HTTP_STATUS_MOVED_PERMANENTLY,
end, n, &p, end);
if ((int)n < 0)
goto bail_nuke_ah;
return lws_http_transaction_completed(wsi);
}
#endif
#ifdef LWS_WITH_ACCESS_LOG
/*
* Produce Apache-compatible log string for wsi, like this:
*
* 2.31.234.19 - - [27/Mar/2016:03:22:44 +0800]
* "GET /aep-screen.png HTTP/1.1"
* 200 152987 "https://libwebsockets.org/index.html"
* "Mozilla/5.0 (Macint... Chrome/49.0.2623.87 Safari/537.36"
*
*/
{
static const char * const hver[] = {
"http/1.0", "http/1.1", "http/2"
};
#ifdef LWS_USE_IPV6
char ads[INET6_ADDRSTRLEN];
#else
char ads[INET_ADDRSTRLEN];
#endif
char da[64];
const char *pa, *me;
struct tm *tmp;
time_t t = time(NULL);
int l = 256;
if (wsi->access_log_pending)
lws_access_log(wsi);
wsi->access_log.header_log = lws_malloc(l);
if (wsi->access_log.header_log) {
tmp = localtime(&t);
if (tmp)
strftime(da, sizeof(da), "%d/%b/%Y:%H:%M:%S %z", tmp);
else
strcpy(da, "01/Jan/1970:00:00:00 +0000");
pa = lws_get_peer_simple(wsi, ads, sizeof(ads));
if (!pa)
pa = "(unknown)";
if (meth >= 0)
me = method_names[meth];
else
me = "unknown";
lws_snprintf(wsi->access_log.header_log, l,
"%s - - [%s] \"%s %s %s\"",
pa, da, me, uri_ptr,
hver[wsi->u.http.request_version]);
l = lws_hdr_total_length(wsi, WSI_TOKEN_HTTP_USER_AGENT);
if (l) {
wsi->access_log.user_agent = lws_malloc(l + 2);
if (wsi->access_log.user_agent)
lws_hdr_copy(wsi, wsi->access_log.user_agent,
l + 1, WSI_TOKEN_HTTP_USER_AGENT);
else
lwsl_err("OOM getting user agent\n");
}
wsi->access_log_pending = 1;
}
}
#endif
/* can we serve it from the mount list? */
2016-06-26 06:29:20 +08:00
hit = lws_find_mount(wsi, uri_ptr, uri_len);
if (!hit) {
/* deferred cleanup and reset to protocols[0] */
2016-06-26 06:29:20 +08:00
lwsl_info("no hit\n");
2016-06-26 06:29:20 +08:00
if (lws_bind_protocol(wsi, &wsi->vhost->protocols[0]))
return 1;
2016-06-26 06:29:20 +08:00
n = wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP,
wsi->user_space, uri_ptr, uri_len);
2016-06-26 06:29:20 +08:00
goto after;
}
2016-06-26 06:29:20 +08:00
s = uri_ptr + hit->mountpoint_len;
2016-06-26 06:29:20 +08:00
/*
* if we have a mountpoint like https://xxx.com/yyy
* there is an implied / at the end for our purposes since
* we can only mount on a "directory".
*
* But if we just go with that, the browser cannot understand
* that he is actually looking down one "directory level", so
* even though we give him /yyy/abc.html he acts like the
* current directory level is /. So relative urls like "x.png"
* wrongly look outside the mountpoint.
*
* Therefore if we didn't come in on a url with an explicit
* / at the end, we must redirect to add it so the browser
* understands he is one "directory level" down.
*/
if ((hit->mountpoint_len > 1 ||
(hit->origin_protocol == LWSMPRO_REDIR_HTTP ||
hit->origin_protocol == LWSMPRO_REDIR_HTTPS)) &&
(*s != '/' ||
(hit->origin_protocol == LWSMPRO_REDIR_HTTP ||
hit->origin_protocol == LWSMPRO_REDIR_HTTPS)) &&
(hit->origin_protocol != LWSMPRO_CGI &&
hit->origin_protocol != LWSMPRO_CALLBACK //&&
//hit->protocol == NULL
)) {
unsigned char *start = pt->serv_buf + LWS_PRE,
*p = start, *end = p + 512;
lwsl_debug("Doing 301 '%s' org %s\n", s, hit->origin);
if (!lws_hdr_total_length(wsi, WSI_TOKEN_HOST))
goto bail_nuke_ah;
2016-06-26 06:29:20 +08:00
/* > at start indicates deal with by redirect */
if (hit->origin_protocol == LWSMPRO_REDIR_HTTP ||
hit->origin_protocol == LWSMPRO_REDIR_HTTPS)
n = lws_snprintf((char *)end, 256, "%s%s",
2016-06-26 06:29:20 +08:00
oprot[hit->origin_protocol & 1],
hit->origin);
else
n = lws_snprintf((char *)end, 256,
2016-06-26 06:29:20 +08:00
"%s%s%s/", oprot[lws_is_ssl(wsi)],
lws_hdr_simple_ptr(wsi, WSI_TOKEN_HOST),
uri_ptr);
2016-06-26 06:29:20 +08:00
n = lws_http_redirect(wsi, HTTP_STATUS_MOVED_PERMANENTLY,
end, n, &p, end);
if ((int)n < 0)
goto bail_nuke_ah;
2016-06-26 06:29:20 +08:00
return lws_http_transaction_completed(wsi);
}
/*
* A particular protocol callback is mounted here?
*
* For the duration of this http transaction, bind us to the
* associated protocol
*/
if (hit->origin_protocol == LWSMPRO_CALLBACK || hit->protocol) {
const struct lws_protocols *pp;
const char *name = hit->origin;
if (hit->protocol)
name = hit->protocol;
pp = lws_vhost_name_to_protocol(wsi->vhost, name);
if (!pp) {
n = -1;
lwsl_err("Unable to find plugin '%s'\n",
hit->origin);
return 1;
}
2016-06-26 06:29:20 +08:00
if (lws_bind_protocol(wsi, pp))
return 1;
args.p = uri_ptr;
args.len = uri_len;
args.max_len = hit->auth_mask;
args.final = 0; /* used to signal callback dealt with it */
n = wsi->protocol->callback(wsi, LWS_CALLBACK_CHECK_ACCESS_RIGHTS,
wsi->user_space, &args, 0);
if (n) {
lws_return_http_status(wsi, HTTP_STATUS_UNAUTHORIZED,
NULL);
goto bail_nuke_ah;
}
if (args.final) /* callback completely handled it well */
return 0;
2016-06-26 06:29:20 +08:00
if (hit->cgienv && wsi->protocol->callback(wsi,
LWS_CALLBACK_HTTP_PMO,
wsi->user_space, (void *)hit->cgienv, 0))
return 1;
2016-06-26 06:29:20 +08:00
if (lws_hdr_total_length(wsi, WSI_TOKEN_POST_URI)) {
n = wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP,
wsi->user_space,
uri_ptr + hit->mountpoint_len,
uri_len - hit->mountpoint_len);
goto after;
}
2016-06-26 06:29:20 +08:00
}
#ifdef LWS_WITH_CGI
2016-06-26 06:29:20 +08:00
/* did we hit something with a cgi:// origin? */
if (hit->origin_protocol == LWSMPRO_CGI) {
const char *cmd[] = {
NULL, /* replace with cgi path */
NULL
};
unsigned char *p, *end, buffer[1024];
2016-06-26 06:29:20 +08:00
lwsl_debug("%s: cgi\n", __func__);
cmd[0] = hit->origin;
2016-06-26 06:29:20 +08:00
n = 5;
if (hit->cgi_timeout)
n = hit->cgi_timeout;
n = lws_cgi(wsi, cmd, hit->mountpoint_len, n,
hit->cgienv);
if (n) {
lwsl_err("%s: cgi failed\n", __func__);
2016-06-26 06:29:20 +08:00
return -1;
}
2016-06-26 06:29:20 +08:00
p = buffer + LWS_PRE;
end = p + sizeof(buffer) - LWS_PRE;
2016-06-26 06:29:20 +08:00
if (lws_add_http_header_status(wsi, 200, &p, end))
return 1;
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_CONNECTION,
(unsigned char *)"close", 5, &p, end))
return 1;
n = lws_write(wsi, buffer + LWS_PRE,
p - (buffer + LWS_PRE),
LWS_WRITE_HTTP_HEADERS);
2016-06-26 06:29:20 +08:00
goto deal_body;
}
#endif
2016-06-26 06:29:20 +08:00
n = strlen(s);
if (s[0] == '\0' || (n == 1 && s[n - 1] == '/'))
s = (char *)hit->def;
if (!s)
s = "index.html";
2016-06-26 06:29:20 +08:00
wsi->cache_secs = hit->cache_max_age;
wsi->cache_reuse = hit->cache_reusable;
wsi->cache_revalidate = hit->cache_revalidate;
wsi->cache_intermediaries = hit->cache_intermediaries;
2016-06-26 06:29:20 +08:00
n = lws_http_serve(wsi, s, hit->origin, hit);
if (n) {
/*
* lws_return_http_status(wsi, HTTP_STATUS_NOT_FOUND, NULL);
*/
if (hit->protocol) {
const struct lws_protocols *pp = lws_vhost_name_to_protocol(
wsi->vhost, hit->protocol);
2016-06-26 06:29:20 +08:00
if (lws_bind_protocol(wsi, pp))
return 1;
2016-06-26 06:29:20 +08:00
n = pp->callback(wsi, LWS_CALLBACK_HTTP,
wsi->user_space,
uri_ptr + hit->mountpoint_len,
uri_len - hit->mountpoint_len);
} else
n = wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP,
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
wsi->user_space, uri_ptr, uri_len);
}
2016-06-26 06:29:20 +08:00
after:
if (n) {
lwsl_info("LWS_CALLBACK_HTTP closing\n");
return 1;
}
#ifdef LWS_WITH_CGI
deal_body:
#endif
/*
* If we're not issuing a file, check for content_length or
* HTTP keep-alive. No keep-alive header allocation for
* ISSUING_FILE, as this uses HTTP/1.0.
*
* In any case, return 0 and let lws_read decide how to
* proceed based on state
*/
if (wsi->state != LWSS_HTTP_ISSUING_FILE)
/* Prepare to read body if we have a content length: */
if (wsi->u.http.content_length > 0)
wsi->state = LWSS_HTTP_BODY;
return 0;
bail_nuke_ah:
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* we're closing, losing some rx is OK */
wsi->u.hdr.ah->rxpos = wsi->u.hdr.ah->rxlen;
2016-07-23 14:18:25 +08:00
// lwsl_notice("%s: drop1\n", __func__);
lws_header_table_detach(wsi, 1);
return 1;
}
int
lws_bind_protocol(struct lws *wsi, const struct lws_protocols *p)
{
// if (wsi->protocol == p)
// return 0;
if (wsi->protocol)
wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP_DROP_PROTOCOL,
wsi->user_space, NULL, 0);
if (!wsi->user_space_externally_allocated)
lws_free_set_NULL(wsi->user_space);
wsi->protocol = p;
if (!p)
return 0;
if (lws_ensure_user_space(wsi))
return 1;
if (wsi->protocol->callback(wsi, LWS_CALLBACK_HTTP_BIND_PROTOCOL,
wsi->user_space, NULL, 0))
return 1;
return 0;
}
int
lws_handshake_server(struct lws *wsi, unsigned char **buf, size_t len)
{
struct lws_context *context = lws_get_context(wsi);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
struct _lws_header_related hdr;
struct allocated_headers *ah;
int protocol_len, n = 0, hit, non_space_char_found = 0;
char protocol_list[128];
char protocol_name[64];
char *p;
if (len >= 10000000) {
lwsl_err("%s: assert: len %ld\n", __func__, (long)len);
assert(0);
}
if (!wsi->u.hdr.ah) {
lwsl_err("%s: assert: NULL ah\n", __func__);
assert(0);
}
while (len--) {
wsi->more_rx_waiting = !!len;
if (wsi->mode != LWSCM_HTTP_SERVING &&
wsi->mode != LWSCM_HTTP_SERVING_ACCEPTED) {
lwsl_err("%s: bad wsi mode %d\n", __func__, wsi->mode);
goto bail_nuke_ah;
}
if (lws_parse(wsi, *(*buf)++)) {
lwsl_info("lws_parse failed\n");
goto bail_nuke_ah;
}
if (wsi->u.hdr.parser_state != WSI_PARSING_COMPLETE)
continue;
lwsl_parser("%s: lws_parse sees parsing complete\n", __func__);
lwsl_debug("%s: wsi->more_rx_waiting=%d\n", __func__,
wsi->more_rx_waiting);
/* select vhost */
if (lws_hdr_total_length(wsi, WSI_TOKEN_HOST)) {
struct lws_vhost *vhost = lws_select_vhost(
context, wsi->vhost->listen_port,
lws_hdr_simple_ptr(wsi, WSI_TOKEN_HOST));
if (vhost)
wsi->vhost = vhost;
} else
lwsl_info("no host\n");
wsi->vhost->trans++;
if (!wsi->conn_stat_done) {
wsi->vhost->conn++;
wsi->conn_stat_done = 1;
}
wsi->mode = LWSCM_PRE_WS_SERVING_ACCEPT;
lws_set_timeout(wsi, NO_PENDING_TIMEOUT, 0);
/* is this websocket protocol or normal http 1.0? */
if (lws_hdr_total_length(wsi, WSI_TOKEN_UPGRADE)) {
if (!strcasecmp(lws_hdr_simple_ptr(wsi, WSI_TOKEN_UPGRADE),
"websocket")) {
wsi->vhost->ws_upgrades++;
lwsl_info("Upgrade to ws\n");
goto upgrade_ws;
}
#ifdef LWS_USE_HTTP2
if (!strcasecmp(lws_hdr_simple_ptr(wsi, WSI_TOKEN_UPGRADE),
"h2c")) {
wsi->vhost->http2_upgrades++;
lwsl_info("Upgrade to h2c\n");
goto upgrade_h2c;
}
#endif
lwsl_info("Unknown upgrade\n");
/* dunno what he wanted to upgrade to */
goto bail_nuke_ah;
}
/* no upgrade ack... he remained as HTTP */
lwsl_info("No upgrade\n");
ah = wsi->u.hdr.ah;
lws_union_transition(wsi, LWSCM_HTTP_SERVING_ACCEPTED);
wsi->state = LWSS_HTTP;
wsi->u.http.fd = LWS_INVALID_FILE;
/* expose it at the same offset as u.hdr */
wsi->u.http.ah = ah;
lwsl_debug("%s: wsi %p: ah %p\n", __func__, (void *)wsi,
(void *)wsi->u.hdr.ah);
n = lws_http_action(wsi);
return n;
#ifdef LWS_USE_HTTP2
upgrade_h2c:
if (!lws_hdr_total_length(wsi, WSI_TOKEN_HTTP2_SETTINGS)) {
lwsl_info("missing http2_settings\n");
goto bail_nuke_ah;
}
lwsl_info("h2c upgrade...\n");
p = lws_hdr_simple_ptr(wsi, WSI_TOKEN_HTTP2_SETTINGS);
/* convert the peer's HTTP-Settings */
n = lws_b64_decode_string(p, protocol_list,
sizeof(protocol_list));
if (n < 0) {
lwsl_parser("HTTP2_SETTINGS too long\n");
return 1;
}
/* adopt the header info */
ah = wsi->u.hdr.ah;
lws_union_transition(wsi, LWSCM_HTTP2_SERVING);
/* http2 union member has http union struct at start */
wsi->u.http.ah = ah;
lws_http2_init(&wsi->u.http2.peer_settings);
lws_http2_init(&wsi->u.http2.my_settings);
/* HTTP2 union */
lws_http2_interpret_settings_payload(&wsi->u.http2.peer_settings,
(unsigned char *)protocol_list, n);
strcpy(protocol_list,
"HTTP/1.1 101 Switching Protocols\x0d\x0a"
"Connection: Upgrade\x0d\x0a"
"Upgrade: h2c\x0d\x0a\x0d\x0a");
n = lws_issue_raw(wsi, (unsigned char *)protocol_list,
strlen(protocol_list));
if (n != strlen(protocol_list)) {
lwsl_debug("http2 switch: ERROR writing to socket\n");
return 1;
}
wsi->state = LWSS_HTTP2_AWAIT_CLIENT_PREFACE;
return 0;
#endif
upgrade_ws:
if (!wsi->protocol)
lwsl_err("NULL protocol at lws_read\n");
/*
* It's websocket
*
* Select the first protocol we support from the list
* the client sent us.
*
* Copy it to remove header fragmentation
*/
if (lws_hdr_copy(wsi, protocol_list, sizeof(protocol_list) - 1,
WSI_TOKEN_PROTOCOL) < 0) {
lwsl_err("protocol list too long");
goto bail_nuke_ah;
}
protocol_len = lws_hdr_total_length(wsi, WSI_TOKEN_PROTOCOL);
protocol_list[protocol_len] = '\0';
p = protocol_list;
hit = 0;
while (*p && !hit) {
n = 0;
non_space_char_found = 0;
while (n < sizeof(protocol_name) - 1 && *p &&
*p != ',') {
// ignore leading spaces
if (!non_space_char_found && *p == ' ') {
n++;
continue;
}
non_space_char_found = 1;
protocol_name[n++] = *p++;
}
protocol_name[n] = '\0';
if (*p)
p++;
lwsl_info("checking %s\n", protocol_name);
n = 0;
while (wsi->vhost->protocols[n].callback) {
if (wsi->vhost->protocols[n].name &&
!strcmp(wsi->vhost->protocols[n].name,
protocol_name)) {
wsi->protocol = &wsi->vhost->protocols[n];
hit = 1;
break;
}
n++;
}
}
/* we didn't find a protocol he wanted? */
if (!hit) {
if (lws_hdr_simple_ptr(wsi, WSI_TOKEN_PROTOCOL)) {
lwsl_info("No protocol from \"%s\" supported\n",
protocol_list);
goto bail_nuke_ah;
}
/*
* some clients only have one protocol and
* do not send the protocol list header...
* allow it and match to the vhost's default
* protocol (which itself defaults to zero)
*/
lwsl_info("defaulting to prot handler %d\n",
wsi->vhost->default_protocol_index);
n = 0;
wsi->protocol = &wsi->vhost->protocols[
(int)wsi->vhost->default_protocol_index];
}
/* allocate wsi->user storage */
if (lws_ensure_user_space(wsi))
goto bail_nuke_ah;
/*
* Give the user code a chance to study the request and
* have the opportunity to deny it
*/
if ((wsi->protocol->callback)(wsi,
LWS_CALLBACK_FILTER_PROTOCOL_CONNECTION,
wsi->user_space,
lws_hdr_simple_ptr(wsi, WSI_TOKEN_PROTOCOL), 0)) {
lwsl_warn("User code denied connection\n");
goto bail_nuke_ah;
}
/*
* Perform the handshake according to the protocol version the
* client announced
*/
switch (wsi->ietf_spec_revision) {
case 13:
lwsl_parser("lws_parse calling handshake_04\n");
if (handshake_0405(context, wsi)) {
lwsl_info("hs0405 has failed the connection\n");
goto bail_nuke_ah;
}
break;
default:
lwsl_info("Unknown client spec version %d\n",
wsi->ietf_spec_revision);
goto bail_nuke_ah;
}
/*
* stitch protocol choice into the vh protocol linked list
* We always insert ourselves at the start of the list
*
* X <-> B
* X <-> pAn <-> pB
*/
//lwsl_err("%s: pre insert vhost start wsi %p, that wsi prev == %p\n",
// __func__,
// wsi->vhost->same_vh_protocol_list[n],
// wsi->same_vh_protocol_prev);
wsi->same_vh_protocol_prev = /* guy who points to us */
&wsi->vhost->same_vh_protocol_list[n];
wsi->same_vh_protocol_next = /* old first guy is our next */
wsi->vhost->same_vh_protocol_list[n];
/* we become the new first guy */
wsi->vhost->same_vh_protocol_list[n] = wsi;
if (wsi->same_vh_protocol_next)
/* old first guy points back to us now */
wsi->same_vh_protocol_next->same_vh_protocol_prev =
&wsi->same_vh_protocol_next;
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* we are upgrading to ws, so http/1.1 and keepalive +
* pipelined header considerations about keeping the ah around
* no longer apply. However it's common for the first ws
* protocol data to have been coalesced with the browser
* upgrade request and to already be in the ah rx buffer.
*/
lwsl_info("%s: %p: inheriting ah in ws mode (rxpos:%d, rxlen:%d)\n",
__func__, wsi, wsi->u.hdr.ah->rxpos,
wsi->u.hdr.ah->rxlen);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
lws_pt_lock(pt);
hdr = wsi->u.hdr;
lws_union_transition(wsi, LWSCM_WS_SERVING);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/*
* first service is WS mode will notice this, use the RX and
* then detach the ah (caution: we are not in u.hdr union
* mode any more then... ah_temp member is at start the same
* though)
*
2016-02-16 15:19:36 +02:00
* Because rxpos/rxlen shows something in the ah, we will get
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
* service guaranteed next time around the event loop
*
* All union members begin with hdr, so we can use it even
* though we transitioned to ws union mode (the ah detach
* code uses it anyway).
*/
wsi->u.hdr = hdr;
lws_pt_unlock(pt);
lws_restart_ws_ping_pong_timer(wsi);
/*
* create the frame buffer for this connection according to the
* size mentioned in the protocol definition. If 0 there, use
* a big default for compatibility
*/
n = wsi->protocol->rx_buffer_size;
if (!n)
n = context->pt_serv_buf_size;
n += LWS_PRE;
wsi->u.ws.rx_ubuf = lws_malloc(n + 4 /* 0x0000ffff zlib */);
if (!wsi->u.ws.rx_ubuf) {
lwsl_err("Out of Mem allocating rx buffer %d\n", n);
return 1;
}
wsi->u.ws.rx_ubuf_alloc = n;
2016-06-26 06:29:20 +08:00
lwsl_debug("Allocating RX buffer %d\n", n);
#if LWS_POSIX
if (setsockopt(wsi->sock, SOL_SOCKET, SO_SNDBUF,
(const char *)&n, sizeof n)) {
lwsl_warn("Failed to set SNDBUF to %d", n);
return 1;
}
#endif
2016-06-26 06:29:20 +08:00
lwsl_parser("accepted v%02d connection\n",
wsi->ietf_spec_revision);
/* notify user code that we're ready to roll */
if (wsi->protocol->callback)
if (wsi->protocol->callback(wsi, LWS_CALLBACK_ESTABLISHED,
wsi->user_space,
#ifdef LWS_OPENSSL_SUPPORT
wsi->ssl,
#else
NULL,
#endif
0))
return 1;
2016-07-23 14:18:25 +08:00
/* !!! drop ah unreservedly after ESTABLISHED */
if (!wsi->more_rx_waiting) {
wsi->u.hdr.ah->rxpos = wsi->u.hdr.ah->rxlen;
//lwsl_notice("%p: dropping ah EST\n", wsi);
lws_header_table_detach(wsi, 1);
}
return 0;
} /* while all chars are handled */
return 0;
bail_nuke_ah:
/* drop the header info */
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* we're closing, losing some rx is OK */
wsi->u.hdr.ah->rxpos = wsi->u.hdr.ah->rxlen;
2016-07-23 14:18:25 +08:00
//lwsl_notice("%s: drop2\n", __func__);
lws_header_table_detach(wsi, 1);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
return 1;
}
static int
lws_get_idlest_tsi(struct lws_context *context)
{
unsigned int lowest = ~0;
int n = 0, hit = -1;
for (; n < context->count_threads; n++) {
if ((unsigned int)context->pt[n].fds_count !=
context->fd_limit_per_thread - 1 &&
(unsigned int)context->pt[n].fds_count < lowest) {
lowest = context->pt[n].fds_count;
hit = n;
}
}
return hit;
}
struct lws *
lws_create_new_server_wsi(struct lws_vhost *vhost)
{
struct lws *new_wsi;
int n = lws_get_idlest_tsi(vhost->context);
if (n < 0) {
lwsl_err("no space for new conn\n");
return NULL;
}
new_wsi = lws_zalloc(sizeof(struct lws));
if (new_wsi == NULL) {
lwsl_err("Out of memory for new connection\n");
return NULL;
}
new_wsi->tsi = n;
lwsl_info("Accepted %p to tsi %d\n", new_wsi, new_wsi->tsi);
new_wsi->vhost = vhost;
new_wsi->context = vhost->context;
new_wsi->pending_timeout = NO_PENDING_TIMEOUT;
new_wsi->rxflow_change_to = LWS_RXFLOW_ALLOW;
/* initialize the instance struct */
new_wsi->state = LWSS_HTTP;
new_wsi->mode = LWSCM_HTTP_SERVING;
new_wsi->hdr_parsing_completed = 0;
#ifdef LWS_OPENSSL_SUPPORT
new_wsi->use_ssl = LWS_SSL_ENABLED(vhost);
#endif
/*
* these can only be set once the protocol is known
* we set an unestablished connection's protocol pointer
* to the start of the supported list, so it can look
* for matching ones during the handshake
*/
new_wsi->protocol = vhost->protocols;
new_wsi->user_space = NULL;
new_wsi->ietf_spec_revision = 0;
new_wsi->sock = LWS_SOCK_INVALID;
vhost->context->count_wsi_allocated++;
/*
* outermost create notification for wsi
* no user_space because no protocol selection
*/
vhost->protocols[0].callback(new_wsi, LWS_CALLBACK_WSI_CREATE,
NULL, NULL, 0);
return new_wsi;
}
LWS_VISIBLE int LWS_WARN_UNUSED_RESULT
lws_http_transaction_completed(struct lws *wsi)
{
int n = NO_PENDING_TIMEOUT;
lws_access_log(wsi);
2016-07-23 14:18:25 +08:00
lwsl_info("%s: wsi %p\n", __func__, wsi);
/* if we can't go back to accept new headers, drop the connection */
if (wsi->u.http.connection_type != HTTP_CONNECTION_KEEP_ALIVE) {
lwsl_info("%s: %p: close connection\n", __func__, wsi);
return 1;
}
if (lws_bind_protocol(wsi, &wsi->vhost->protocols[0]))
return 1;
/* otherwise set ourselves up ready to go again */
wsi->state = LWSS_HTTP;
wsi->mode = LWSCM_HTTP_SERVING;
wsi->u.http.content_length = 0;
wsi->u.http.content_remain = 0;
wsi->hdr_parsing_completed = 0;
#ifdef LWS_WITH_ACCESS_LOG
wsi->access_log.sent = 0;
#endif
if (wsi->vhost->keepalive_timeout)
n = PENDING_TIMEOUT_HTTP_KEEPALIVE_IDLE;
lws_set_timeout(wsi, n, wsi->vhost->keepalive_timeout);
/*
* We already know we are on http1.1 / keepalive and the next thing
* coming will be another header set.
*
* If there is no pending rx and we still have the ah, drop it and
* reacquire a new ah when the new headers start to arrive. (Otherwise
* we needlessly hog an ah indefinitely.)
*
* However if there is pending rx and we know from the keepalive state
* that is already at least the start of another header set, simply
* reset the existing header table and keep it.
*/
if (wsi->u.hdr.ah) {
lwsl_info("%s: wsi->more_rx_waiting=%d\n", __func__,
wsi->more_rx_waiting);
if (!wsi->more_rx_waiting) {
wsi->u.hdr.ah->rxpos = wsi->u.hdr.ah->rxlen;
lws_header_table_detach(wsi, 1);
} else
lws_header_table_reset(wsi, 1);
}
/* If we're (re)starting on headers, need other implied init */
wsi->u.hdr.ues = URIES_IDLE;
lwsl_info("%s: %p: keep-alive await new transaction\n", __func__, wsi);
return 0;
}
2016-07-23 14:18:25 +08:00
struct lws *
lws_adopt_socket_vhost(struct lws_vhost *vh, lws_sockfd_type accept_fd)
{
struct lws_context *context = vh->context;
struct lws *new_wsi = lws_create_new_server_wsi(vh);
if (!new_wsi) {
compatible_close(accept_fd);
return NULL;
}
2016-07-23 14:18:25 +08:00
//lwsl_notice("%s: new wsi %p, sockfd %d, cb %p\n", __func__, new_wsi, accept_fd, context->vhost_list->protocols[0].callback);
new_wsi->sock = accept_fd;
/* the transport is accepted... give him time to negotiate */
lws_set_timeout(new_wsi, PENDING_TIMEOUT_ESTABLISH_WITH_SERVER,
context->timeout_secs);
#if LWS_POSIX == 0
2016-07-23 14:18:25 +08:00
#if defined(LWS_WITH_ESP8266)
esp8266_tcp_stream_accept(accept_fd, new_wsi);
#else
mbed3_tcp_stream_accept(accept_fd, new_wsi);
#endif
2016-07-23 14:18:25 +08:00
#endif
/*
* A new connection was accepted. Give the user a chance to
* set properties of the newly created wsi. There's no protocol
* selected yet so we issue this to protocols[0]
*/
if ((context->vhost_list->protocols[0].callback)(new_wsi,
LWS_CALLBACK_SERVER_NEW_CLIENT_INSTANTIATED, NULL, NULL, 0)) {
/* force us off the timeout list by hand */
lws_set_timeout(new_wsi, NO_PENDING_TIMEOUT, 0);
compatible_close(new_wsi->sock);
lws_free(new_wsi);
return NULL;
}
lws_libev_accept(new_wsi, new_wsi->sock);
lws_libuv_accept(new_wsi, new_wsi->sock);
if (!LWS_SSL_ENABLED(new_wsi->vhost)) {
if (insert_wsi_socket_into_fds(context, new_wsi)) {
lwsl_err("%s: fail inserting socket\n", __func__);
goto fail;
}
} else {
new_wsi->mode = LWSCM_SSL_INIT;
if (lws_server_socket_service_ssl(new_wsi, accept_fd)) {
lwsl_err("%s: fail ssl negotiation\n", __func__);
goto fail;
}
}
if (!lws_header_table_attach(new_wsi, 0))
lwsl_debug("Attached ah immediately\n");
return new_wsi;
fail:
lws_close_free_wsi(new_wsi, LWS_CLOSE_STATUS_NOSTATUS);
return NULL;
}
LWS_VISIBLE struct lws *
lws_adopt_socket(struct lws_context *context, lws_sockfd_type accept_fd)
{
return lws_adopt_socket_vhost(context->vhost_list, accept_fd);
}
LWS_VISIBLE LWS_EXTERN struct lws *
lws_adopt_socket_readbuf(struct lws_context *context, lws_sockfd_type accept_fd,
const char *readbuf, size_t len)
{
struct lws *wsi = lws_adopt_socket(context, accept_fd);
struct lws_context_per_thread *pt;
struct allocated_headers *ah;
struct lws_pollfd *pfd;
if (!wsi)
return NULL;
if (!readbuf)
return wsi;
if (len > sizeof(ah->rx)) {
lwsl_err("%s: rx in too big\n", __func__);
goto bail;
}
2016-07-23 14:18:25 +08:00
/*
* we can't process the initial read data until we can attach an ah.
*
* if one is available, get it and place the data in his ah rxbuf...
* wsi with ah that have pending rxbuf get auto-POLLIN service.
*
* no autoservice because we didn't get a chance to attach the
* readbuf data to wsi or ah yet, and we will do it next if we get
* the ah.
*/
if (wsi->u.hdr.ah || !lws_header_table_attach(wsi, 0)) {
ah = wsi->u.hdr.ah;
memcpy(ah->rx, readbuf, len);
ah->rxpos = 0;
ah->rxlen = len;
lwsl_notice("%s: calling service on readbuf ah\n", __func__);
pt = &context->pt[(int)wsi->tsi];
/* unlike a normal connect, we have the headers already
* (or the first part of them anyway).
* libuv won't come back and service us without a network
* event, so we need to do the header service right here.
*/
pfd = &pt->fds[wsi->position_in_fds_table];
pfd->revents |= LWS_POLLIN;
lwsl_err("%s: calling service\n", __func__);
if (lws_service_fd_tsi(context, pfd, wsi->tsi))
/* service closed us */
return NULL;
return wsi;
}
lwsl_err("%s: deferring handling ah\n", __func__);
/*
* hum if no ah came, we are on the wait list and must defer
* dealing with this until the ah arrives.
*
* later successful lws_header_table_attach() will apply the
* below to the rx buffer (via lws_header_table_reset()).
*/
wsi->u.hdr.preamble_rx = lws_malloc(len);
if (!wsi->u.hdr.preamble_rx) {
lwsl_err("OOM\n");
goto bail;
}
memcpy(wsi->u.hdr.preamble_rx, readbuf, len);
wsi->u.hdr.preamble_rx_len = len;
return wsi;
bail:
lws_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS);
return NULL;
}
LWS_VISIBLE int
lws_server_socket_service(struct lws_context *context, struct lws *wsi,
struct lws_pollfd *pollfd)
{
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
lws_sockfd_type accept_fd = LWS_SOCK_INVALID;
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
struct allocated_headers *ah;
#if LWS_POSIX
struct sockaddr_in cli_addr;
socklen_t clilen;
#endif
int n, len;
2016-07-23 14:18:25 +08:00
// lwsl_notice("%s: mode %d\n", __func__, wsi->mode);
switch (wsi->mode) {
case LWSCM_HTTP_SERVING:
case LWSCM_HTTP_SERVING_ACCEPTED:
case LWSCM_HTTP2_SERVING:
/* handle http headers coming in */
/* pending truncated sends have uber priority */
if (wsi->trunc_len) {
if (!(pollfd->revents & LWS_POLLOUT))
break;
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
if (lws_issue_raw(wsi, wsi->trunc_alloc +
wsi->trunc_offset,
wsi->trunc_len) < 0)
goto fail;
/*
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
* we can't afford to allow input processing to send
* something new, so spin around he event loop until
* he doesn't have any partials
*/
break;
}
/* any incoming data ready? */
if (!(pollfd->revents & pollfd->events & LWS_POLLIN))
goto try_pollout;
/*
* If we previously just did POLLIN when IN and OUT were
* signalled (because POLLIN processing may have used up
* the POLLOUT), don't let that happen twice in a row...
* next time we see the situation favour POLLOUT
*/
2016-07-23 14:18:25 +08:00
#if !defined(LWS_WITH_ESP8266)
if (wsi->favoured_pollin &&
(pollfd->revents & pollfd->events & LWS_POLLOUT)) {
wsi->favoured_pollin = 0;
goto try_pollout;
}
2016-07-23 14:18:25 +08:00
#endif
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* these states imply we MUST have an ah attached */
if (wsi->state == LWSS_HTTP ||
wsi->state == LWSS_HTTP_ISSUING_FILE ||
wsi->state == LWSS_HTTP_HEADERS) {
2016-07-23 14:18:25 +08:00
if (!wsi->u.hdr.ah) {
//lwsl_err("wsi %p: missing ah\n", wsi);
/* no autoservice beacuse we will do it next */
2016-07-23 14:18:25 +08:00
if (lws_header_table_attach(wsi, 0)) {
2017-04-02 13:03:51 +08:00
lwsl_info("wsi %p: failed to acquire ah\n", wsi);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
goto try_pollout;
2016-07-23 14:18:25 +08:00
}
}
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
ah = wsi->u.hdr.ah;
2016-07-23 14:18:25 +08:00
//lwsl_notice("%s: %p: rxpos:%d rxlen:%d\n", __func__, wsi,
// ah->rxpos, ah->rxlen);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* if nothing in ah rx buffer, get some fresh rx */
if (ah->rxpos == ah->rxlen) {
ah->rxlen = lws_ssl_capable_read(wsi, ah->rx,
sizeof(ah->rx));
ah->rxpos = 0;
2016-07-23 14:18:25 +08:00
//lwsl_notice("%s: wsi %p, ah->rxlen = %d\r\n",
// __func__, wsi, ah->rxlen);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
switch (ah->rxlen) {
case 0:
lwsl_info("%s: read 0 len\n", __func__);
/* lwsl_info(" state=%d\n", wsi->state); */
// if (!wsi->hdr_parsing_completed)
// lws_header_table_detach(wsi);
/* fallthru */
case LWS_SSL_CAPABLE_ERROR:
goto fail;
case LWS_SSL_CAPABLE_MORE_SERVICE:
ah->rxlen = ah->rxpos = 0;
goto try_pollout;
}
}
2016-07-23 14:18:25 +08:00
if (!(ah->rxpos != ah->rxlen && ah->rxlen)) {
lwsl_err("%s: assert: rxpos %d, rxlen %d\n",
__func__, ah->rxpos, ah->rxlen);
assert(0);
}
2016-07-23 14:18:25 +08:00
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
/* just ignore incoming if waiting for close */
if (wsi->state != LWSS_FLUSHING_STORED_SEND_BEFORE_CLOSE) {
n = lws_read(wsi, ah->rx + ah->rxpos,
ah->rxlen - ah->rxpos);
if (n < 0) /* we closed wsi */
return 1;
if (wsi->u.hdr.ah) {
if ( wsi->u.hdr.ah->rxlen)
wsi->u.hdr.ah->rxpos += n;
lwsl_debug("%s: wsi %p: ah read rxpos %d, rxlen %d\n", __func__, wsi, wsi->u.hdr.ah->rxpos, wsi->u.hdr.ah->rxlen);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
if (wsi->u.hdr.ah->rxpos == wsi->u.hdr.ah->rxlen &&
(wsi->mode != LWSCM_HTTP_SERVING &&
wsi->mode != LWSCM_HTTP_SERVING_ACCEPTED &&
wsi->mode != LWSCM_HTTP2_SERVING))
lws_header_table_detach(wsi, 1);
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
}
break;
}
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
goto try_pollout;
}
len = lws_ssl_capable_read(wsi, pt->serv_buf,
context->pt_serv_buf_size);
2016-07-23 14:18:25 +08:00
lwsl_notice("%s: wsi %p read %d\r\n", __func__, wsi, len);
switch (len) {
case 0:
lwsl_info("%s: read 0 len\n", __func__);
/* lwsl_info(" state=%d\n", wsi->state); */
ah owns rxbuf This is intended to solve a longstanding problem with the relationship between http/1.1 keep-alive and the service loop. Ah now contain an rx buffer which is used during header processing, and the ah may not be detached from the wsi until the rx buffer is exhausted. Having the rx buffer in the ah means we can delay using the rx until a later service loop. Ah which have pending rx force POLLIN service on the wsi they are attached to automatically, so we can interleave general service / connections with draining each ah rx buffer. The possible http/1.1 situations and their dispositions are: 1) exactly one set of http headers come. After processing, the ah is detached since no pending rx left. If more headers come later, a fresh ah is aqcuired when available and the rx flow control blocks the read until then. 2) more that one whole set of headers come and we remain in http mode (no upgrade). The ah is left attached and returns to the service loop after the first set of headers. We will get forced service due to the ah having pending content (respecting flowcontrol) and process the pending rx in the ah. If we use it all up, we will detach the ah. 3) one set of http headers come with ws traffic appended. We service the headers, do the upgrade, and keep the ah until the remaining ws content is used. When we exhausted the ws traffix in the ah rx buffer, we detach the ah. Since there can be any amount of http/1.1 pipelining on a connection, and each may be expensive to service, it's now enforced there is a return to the service loop after each header set is serviced on a connection. When I added the forced service for ah with pending buffering, I added support for it to the windows plat code. However this is untested. Signed-off-by: Andy Green <andy.green@linaro.org>
2016-02-15 12:37:04 +08:00
// if (!wsi->hdr_parsing_completed)
// lws_header_table_detach(wsi);
/* fallthru */
case LWS_SSL_CAPABLE_ERROR:
goto fail;
case LWS_SSL_CAPABLE_MORE_SERVICE:
goto try_pollout;
}
2016-07-23 14:18:25 +08:00
/* just ignore incoming if waiting for close */
if (wsi->state != LWSS_FLUSHING_STORED_SEND_BEFORE_CLOSE) {
/*
* this may want to send
* (via HTTP callback for example)
*/
n = lws_read(wsi, pt->serv_buf, len);
if (n < 0) /* we closed wsi */
return 1;
/*
* he may have used up the
* writability above, if we will defer POLLOUT
* processing in favour of POLLIN, note it
*/
if (pollfd->revents & LWS_POLLOUT)
wsi->favoured_pollin = 1;
break;
}
try_pollout:
2016-07-23 14:18:25 +08:00
/* this handles POLLOUT for http serving fragments */
if (!(pollfd->revents & LWS_POLLOUT))
break;
/* one shot */
if (lws_change_pollfd(wsi, LWS_POLLOUT, 0)) {
lwsl_notice("%s a\n", __func__);
goto fail;
}
if (!wsi->hdr_parsing_completed)
break;
if (wsi->state != LWSS_HTTP_ISSUING_FILE) {
n = user_callback_handle_rxflow(wsi->protocol->callback,
wsi, LWS_CALLBACK_HTTP_WRITEABLE,
wsi->user_space, NULL, 0);
if (n < 0) {
lwsl_info("writeable_fail\n");
goto fail;
}
break;
}
/* >0 == completion, <0 == error */
n = lws_serve_http_file_fragment(wsi);
if (n < 0 || (n > 0 && lws_http_transaction_completed(wsi))) {
lwsl_info("completed\n");
goto fail;
}
2016-07-23 14:18:25 +08:00
break;
case LWSCM_SERVER_LISTENER:
#if LWS_POSIX
/* pollin means a client has connected to us then */
do {
if (!(pollfd->revents & LWS_POLLIN) || !(pollfd->events & LWS_POLLIN))
break;
/* listen socket got an unencrypted connection... */
clilen = sizeof(cli_addr);
lws_latency_pre(context, wsi);
accept_fd = accept(pollfd->fd, (struct sockaddr *)&cli_addr,
&clilen);
lws_latency(context, wsi, "listener accept", accept_fd,
accept_fd >= 0);
if (accept_fd < 0) {
if (LWS_ERRNO == LWS_EAGAIN ||
LWS_ERRNO == LWS_EWOULDBLOCK) {
// lwsl_err("accept asks to try again\n");
break;
}
lwsl_err("ERROR on accept: %s\n", strerror(LWS_ERRNO));
break;
}
lws_plat_set_socket_options(wsi->vhost, accept_fd);
lwsl_debug("accepted new conn port %u on fd=%d\n",
ntohs(cli_addr.sin_port), accept_fd);
#else
/* not very beautiful... */
accept_fd = (lws_sockfd_type)pollfd;
#endif
/*
* look at who we connected to and give user code a chance
* to reject based on client IP. There's no protocol selected
* yet so we issue this to protocols[0]
*/
if ((wsi->vhost->protocols[0].callback)(wsi,
LWS_CALLBACK_FILTER_NETWORK_CONNECTION,
NULL, (void *)(long)accept_fd, 0)) {
lwsl_debug("Callback denied network connection\n");
compatible_close(accept_fd);
break;
}
if (!lws_adopt_socket_vhost(wsi->vhost, accept_fd))
/* already closed cleanly as necessary */
return 1;
#if LWS_POSIX
} while (pt->fds_count < context->fd_limit_per_thread - 1 &&
lws_poll_listen_fd(&pt->fds[wsi->position_in_fds_table]) > 0);
#endif
return 0;
default:
break;
}
if (!lws_server_socket_service_ssl(wsi, accept_fd))
return 0;
fail:
lws_close_free_wsi(wsi, LWS_CLOSE_STATUS_NOSTATUS);
return 1;
}
2013-02-11 17:52:23 +01:00
LWS_VISIBLE int
lws_serve_http_file(struct lws *wsi, const char *file, const char *content_type,
const char *other_headers, int other_headers_len)
{
static const char * const intermediates[] = { "private", "public" };
struct lws_context *context = lws_get_context(wsi);
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
char cache_control[50], *cc = "no-store";
unsigned char *response = pt->serv_buf + LWS_PRE;
unsigned char *p = response;
unsigned char *end = p + context->pt_serv_buf_size - LWS_PRE;
int ret = 0, cclen = 8;
wsi->u.http.fd = lws_plat_file_open(wsi, file, &wsi->u.http.filelen,
O_RDONLY);
if (wsi->u.http.fd == LWS_INVALID_FILE) {
2016-07-23 14:18:25 +08:00
lwsl_err("Unable to open '%s'\n", file);
return -1;
}
if (lws_add_http_header_status(wsi, 200, &p, end))
return -1;
if (content_type && content_type[0]) {
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_HTTP_CONTENT_TYPE,
(unsigned char *)content_type,
strlen(content_type), &p, end))
return -1;
}
if (!wsi->sending_chunked) {
if (lws_add_http_header_content_length(wsi, wsi->u.http.filelen, &p, end))
return -1;
} else {
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_HTTP_TRANSFER_ENCODING,
(unsigned char *)"chunked",
7, &p, end))
return -1;
}
if (wsi->cache_secs && wsi->cache_reuse) {
if (wsi->cache_revalidate) {
cc = cache_control;
cclen = sprintf(cache_control, "%s max-age: %u",
intermediates[wsi->cache_intermediaries],
wsi->cache_secs);
} else {
cc = "no-cache";
cclen = 8;
}
}
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_HTTP_CACHE_CONTROL,
(unsigned char *)cc, cclen, &p, end))
return -1;
2016-07-23 14:18:25 +08:00
if (wsi->u.http.connection_type == HTTP_CONNECTION_KEEP_ALIVE)
if (lws_add_http_header_by_token(wsi, WSI_TOKEN_CONNECTION,
(unsigned char *)"keep-alive", 10, &p, end))
return -1;
if (other_headers) {
if ((end - p) < other_headers_len)
return -1;
memcpy(p, other_headers, other_headers_len);
p += other_headers_len;
}
if (lws_finalize_http_header(wsi, &p, end))
return -1;
ret = lws_write(wsi, response, p - response, LWS_WRITE_HTTP_HEADERS);
if (ret != (p - response)) {
lwsl_err("_write returned %d from %ld\n", ret, (long)(p - response));
return -1;
}
wsi->u.http.filepos = 0;
wsi->state = LWSS_HTTP_ISSUING_FILE;
return lws_serve_http_file_fragment(wsi);
}
int
lws_interpret_incoming_packet(struct lws *wsi, unsigned char **buf, size_t len)
{
int m;
lwsl_parser("%s: received %d byte packet\n", __func__, (int)len);
#if 0
lwsl_hexdump(*buf, len);
#endif
/* let the rx protocol state machine have as much as it needs */
while (len) {
/*
* we were accepting input but now we stopped doing so
*/
if (!(wsi->rxflow_change_to & LWS_RXFLOW_ALLOW)) {
lws_rxflow_cache(wsi, *buf, 0, len);
lwsl_parser("%s: cached %ld\n", __func__, (long)len);
return 1;
}
if (wsi->u.ws.rx_draining_ext) {
// lwsl_notice("draining with 0\n");
m = lws_rx_sm(wsi, 0);
if (m < 0)
return -1;
continue;
}
/* account for what we're using in rxflow buffer */
if (wsi->rxflow_buffer)
wsi->rxflow_pos++;
/* consume payload bytes efficiently */
if (
wsi->lws_rx_parse_state ==
LWS_RXPS_PAYLOAD_UNTIL_LENGTH_EXHAUSTED) {
m = lws_payload_until_length_exhausted(wsi, buf, &len);
if (wsi->rxflow_buffer)
wsi->rxflow_pos += m;
}
/* process the byte */
m = lws_rx_sm(wsi, *(*buf)++);
if (m < 0)
return -1;
len--;
}
lwsl_parser("%s: exit with %d unused\n", __func__, (int)len);
return 0;
}
LWS_VISIBLE void
lws_server_get_canonical_hostname(struct lws_context *context,
struct lws_context_creation_info *info)
{
if (lws_check_opt(info->options, LWS_SERVER_OPTION_SKIP_SERVER_CANONICAL_NAME))
return;
#if LWS_POSIX
/* find canonical hostname */
gethostname((char *)context->canonical_hostname,
sizeof(context->canonical_hostname) - 1);
lwsl_notice(" canonical_hostname = %s\n", context->canonical_hostname);
#else
(void)context;
#endif
}
#define LWS_MAX_ELEM_NAME 32
enum urldecode_stateful {
US_NAME,
US_IDLE,
US_PC1,
US_PC2,
MT_LOOK_BOUND_IN,
MT_HNAME,
MT_DISP,
MT_TYPE,
MT_IGNORE1,
MT_IGNORE2,
};
static const char * const mp_hdr[] = {
"content-disposition: ",
"content-type: ",
"\x0d\x0a"
};
typedef int (*lws_urldecode_stateful_cb)(void *data,
const char *name, char **buf, int len, int final);
struct lws_urldecode_stateful {
char *out;
void *data;
char name[LWS_MAX_ELEM_NAME];
char temp[LWS_MAX_ELEM_NAME];
char content_type[32];
char content_disp[32];
char content_disp_filename[256];
char mime_boundary[128];
int out_len;
int pos;
int hdr_idx;
int mp;
int sum;
unsigned int multipart_form_data:1;
unsigned int inside_quote:1;
unsigned int subname:1;
unsigned int boundary_real_crlf:1;
enum urldecode_stateful state;
lws_urldecode_stateful_cb output;
};
static struct lws_urldecode_stateful *
lws_urldecode_s_create(struct lws *wsi, char *out, int out_len, void *data,
lws_urldecode_stateful_cb output)
{
struct lws_urldecode_stateful *s = lws_zalloc(sizeof(*s));
char buf[200], *p;
int m = 0;
if (!s)
return NULL;
s->out = out;
s->out_len = out_len;
s->output = output;
s->pos = 0;
s->sum = 0;
s->mp = 0;
s->state = US_NAME;
s->name[0] = '\0';
s->data = data;
if (lws_hdr_copy(wsi, buf, sizeof(buf), WSI_TOKEN_HTTP_CONTENT_TYPE) > 0) {
/* multipart/form-data; boundary=----WebKitFormBoundarycc7YgAPEIHvgE9Bf */
if (!strncmp(buf, "multipart/form-data", 19)) {
s->multipart_form_data = 1;
s->state = MT_LOOK_BOUND_IN;
s->mp = 2;
p = strstr(buf, "boundary=");
if (p) {
p += 9;
s->mime_boundary[m++] = '\x0d';
s->mime_boundary[m++] = '\x0a';
s->mime_boundary[m++] = '-';
s->mime_boundary[m++] = '-';
while (m < sizeof(s->mime_boundary) - 1 &&
*p && *p != ' ')
s->mime_boundary[m++] = *p++;
s->mime_boundary[m] = '\0';
lwsl_notice("boundary '%s'\n", s->mime_boundary);
}
}
}
return s;
}
static int
lws_urldecode_s_process(struct lws_urldecode_stateful *s, const char *in, int len)
{
int n, m, hit = 0;
char c, was_end = 0;
while (len--) {
if (s->pos == s->out_len - s->mp - 1) {
if (s->output(s->data, s->name, &s->out, s->pos, 0))
return -1;
was_end = s->pos;
s->pos = 0;
}
switch (s->state) {
/* states for url arg style */
case US_NAME:
s->inside_quote = 0;
if (*in == '=') {
s->name[s->pos] = '\0';
s->pos = 0;
s->state = US_IDLE;
in++;
continue;
}
if (*in == '&') {
s->name[s->pos] = '\0';
if (s->output(s->data, s->name, &s->out, s->pos, 1))
return -1;
s->pos = 0;
s->state = US_IDLE;
in++;
continue;
}
if (s->pos >= sizeof(s->name) - 1) {
lwsl_notice("Name too long\n");
return -1;
}
s->name[s->pos++] = *in++;
break;
case US_IDLE:
if (*in == '%') {
s->state++;
in++;
continue;
}
if (*in == '&') {
s->out[s->pos] = '\0';
if (s->output(s->data, s->name, &s->out, s->pos, 1))
return -1;
s->pos = 0;
s->state = US_NAME;
in++;
continue;
}
if (*in == '+') {
in++;
s->out[s->pos++] = ' ';
continue;
}
s->out[s->pos++] = *in++;
break;
case US_PC1:
n = char_to_hex(*in);
if (n < 0)
return -1;
in++;
s->sum = n << 4;
s->state++;
break;
case US_PC2:
n = char_to_hex(*in);
if (n < 0)
return -1;
in++;
s->out[s->pos++] = s->sum | n;
s->state = US_IDLE;
break;
/* states for multipart / mime style */
case MT_LOOK_BOUND_IN:
retry_as_first:
if (*in == s->mime_boundary[s->mp] &&
s->mime_boundary[s->mp]) {
in++;
s->mp++;
if (!s->mime_boundary[s->mp]) {
s->mp = 0;
s->state = MT_IGNORE1;
if (s->pos || was_end)
if (s->output(s->data, s->name,
&s->out, s->pos, 1))
return -1;
s->pos = 0;
s->content_disp[0] = '\0';
s->name[0] = '\0';
s->content_disp_filename[0] = '\0';
s->boundary_real_crlf = 1;
}
continue;
}
if (s->mp) {
n = 0;
if (!s->boundary_real_crlf)
n = 2;
memcpy(s->out + s->pos, s->mime_boundary + n, s->mp - n);
s->pos += s->mp;
s->mp = 0;
goto retry_as_first;
}
s->out[s->pos++] = *in;
in++;
s->mp = 0;
break;
case MT_HNAME:
m = 0;
c =*in;
if (c >= 'A' && c <= 'Z')
c += 'a' - 'A';
for (n = 0; n < ARRAY_SIZE(mp_hdr); n++)
if (c == mp_hdr[n][s->mp]) {
m++;
hit = n;
}
in++;
if (!m) {
s->mp = 0;
continue;
}
s->mp++;
if (m != 1)
continue;
if (mp_hdr[hit][s->mp])
continue;
s->mp = 0;
s->temp[0] = '\0';
s->subname = 0;
if (hit == 2)
s->state = MT_LOOK_BOUND_IN;
else
s->state += hit + 1;
break;
case MT_DISP:
/* form-data; name="file"; filename="t.txt" */
if (*in == '\x0d') {
// lwsl_notice("disp: '%s', '%s', '%s'\n",
// s->content_disp, s->name,
// s->content_disp_filename);
if (s->content_disp_filename[0])
if (s->output(s->data, s->name,
&s->out, s->pos, LWS_UFS_OPEN))
return -1;
s->state = MT_IGNORE2;
goto done;
}
if (*in == ';') {
s->subname = 1;
s->temp[0] = '\0';
s->mp = 0;
goto done;
}
if (*in == '\"') {
s->inside_quote ^= 1;
goto done;
}
if (s->subname) {
if (*in == '=') {
s->temp[s->mp] = '\0';
s->subname = 0;
s->mp = 0;
goto done;
}
if (s->mp < sizeof(s->temp) - 1 &&
(*in != ' ' || s->inside_quote))
s->temp[s->mp++] = *in;
goto done;
}
if (!s->temp[0]) {
if (s->mp < sizeof(s->content_disp) - 1)
s->content_disp[s->mp++] = *in;
s->content_disp[s->mp] = '\0';
goto done;
}
if (!strcmp(s->temp, "name")) {
if (s->mp < sizeof(s->name) - 1)
s->name[s->mp++] = *in;
s->name[s->mp] = '\0';
goto done;
}
if (!strcmp(s->temp, "filename")) {
if (s->mp < sizeof(s->content_disp_filename) - 1)
s->content_disp_filename[s->mp++] = *in;
s->content_disp_filename[s->mp] = '\0';
goto done;
}
done:
in++;
break;
case MT_TYPE:
if (*in == '\x0d')
s->state = MT_IGNORE2;
else {
if (s->mp < sizeof(s->content_type) - 1)
s->content_type[s->mp++] = *in;
s->content_type[s->mp] = '\0';
}
in++;
break;
case MT_IGNORE1:
if (*in == '\x0d')
s->state = MT_IGNORE2;
in++;
break;
case MT_IGNORE2:
s->mp = 0;
if (*in == '\x0a')
s->state = MT_HNAME;
in++;
break;
}
}
return 0;
}
static int
lws_urldecode_s_destroy(struct lws_urldecode_stateful *s)
{
int ret = 0;
if (s->state != US_IDLE)
ret = -1;
if (!ret)
if (s->output(s->data, s->name, &s->out, s->pos, 1))
ret = -1;
lws_free(s);
return ret;
}
struct lws_spa {
struct lws_urldecode_stateful *s;
lws_spa_fileupload_cb opt_cb;
const char * const *param_names;
int count_params;
char **params;
int *param_length;
void *opt_data;
char *storage;
char *end;
int max_storage;
};
static int
lws_urldecode_spa_lookup(struct lws_spa *spa,
const char *name)
{
int n;
for (n = 0; n < spa->count_params; n++)
if (!strcmp(spa->param_names[n], name))
return n;
return -1;
}
static int
lws_urldecode_spa_cb(void *data, const char *name, char **buf, int len,
int final)
{
struct lws_spa *spa =
(struct lws_spa *)data;
int n;
if (spa->s->content_disp_filename[0]) {
if (spa->opt_cb) {
n = spa->opt_cb(spa->opt_data, name,
spa->s->content_disp_filename,
*buf, len, final);
if (n < 0)
return -1;
}
return 0;
}
n = lws_urldecode_spa_lookup(spa, name);
if (n == -1 || !len) /* unrecognized */
return 0;
if (!spa->params[n])
spa->params[n] = *buf;
if ((*buf) + len >= spa->end) {
lwsl_notice("%s: exceeded storage\n", __func__);
return -1;
}
spa->param_length[n] += len;
/* move it on inside storage */
(*buf) += len;
*((*buf)++) = '\0';
spa->s->out_len -= len + 1;
return 0;
}
LWS_VISIBLE LWS_EXTERN struct lws_spa *
lws_spa_create(struct lws *wsi, const char * const *param_names,
int count_params, int max_storage,
lws_spa_fileupload_cb opt_cb, void *opt_data)
{
struct lws_spa *spa = lws_zalloc(sizeof(*spa));
if (!spa)
return NULL;
spa->param_names = param_names;
spa->count_params = count_params;
spa->max_storage = max_storage;
spa->opt_cb = opt_cb;
spa->opt_data = opt_data;
spa->storage = lws_malloc(max_storage);
if (!spa->storage)
goto bail2;
spa->end = spa->storage + max_storage - 1;
spa->params = lws_zalloc(sizeof(char *) * count_params);
if (!spa->params)
goto bail3;
spa->s = lws_urldecode_s_create(wsi, spa->storage, max_storage, spa,
lws_urldecode_spa_cb);
if (!spa->s)
goto bail4;
spa->param_length = lws_zalloc(sizeof(int) * count_params);
if (!spa->param_length)
goto bail5;
lwsl_notice("%s: Created SPA %p\n", __func__, spa);
return spa;
bail5:
lws_urldecode_s_destroy(spa->s);
bail4:
lws_free(spa->params);
bail3:
lws_free(spa->storage);
bail2:
lws_free(spa);
return NULL;
}
LWS_VISIBLE LWS_EXTERN int
lws_spa_process(struct lws_spa *ludspa, const char *in, int len)
{
if (!ludspa) {
lwsl_err("%s: NULL spa\n", __func__);
return -1;
}
return lws_urldecode_s_process(ludspa->s, in, len);
}
LWS_VISIBLE LWS_EXTERN int
lws_spa_get_length(struct lws_spa *ludspa, int n)
{
if (n >= ludspa->count_params)
return 0;
return ludspa->param_length[n];
}
LWS_VISIBLE LWS_EXTERN const char *
lws_spa_get_string(struct lws_spa *ludspa, int n)
{
if (n >= ludspa->count_params)
return NULL;
return ludspa->params[n];
}
LWS_VISIBLE LWS_EXTERN int
lws_spa_finalize(struct lws_spa *spa)
{
if (spa->s) {
lws_urldecode_s_destroy(spa->s);
spa->s = NULL;
}
return 0;
}
LWS_VISIBLE LWS_EXTERN int
lws_spa_destroy(struct lws_spa *spa)
{
int n = 0;
lwsl_notice("%s: destroy spa %p\n", __func__, spa);
if (spa->s)
lws_urldecode_s_destroy(spa->s);
lwsl_debug("%s\n", __func__);
lws_free(spa->param_length);
lws_free(spa->params);
lws_free(spa->storage);
lws_free(spa);
return n;
}
LWS_VISIBLE LWS_EXTERN int
lws_chunked_html_process(struct lws_process_html_args *args,
struct lws_process_html_state *s)
{
char *sp, buffer[32];
const char *pc;
int old_len, n;
/* do replacements */
sp = args->p;
old_len = args->len;
args->len = 0;
s->start = sp;
while (sp < args->p + old_len) {
if (args->len + 7 >= args->max_len) {
lwsl_err("Used up interpret padding\n");
return -1;
}
if ((!s->pos && *sp == '$') || s->pos) {
int hits = 0, hit = 0;
if (!s->pos)
s->start = sp;
s->swallow[s->pos++] = *sp;
if (s->pos == sizeof(s->swallow) - 1)
goto skip;
for (n = 0; n < s->count_vars; n++)
if (!strncmp(s->swallow, s->vars[n], s->pos)) {
hits++;
hit = n;
}
if (!hits) {
skip:
s->swallow[s->pos] = '\0';
memcpy(s->start, s->swallow, s->pos);
args->len++;
s->pos = 0;
sp = s->start + 1;
continue;
}
if (hits == 1 && s->pos == strlen(s->vars[hit])) {
pc = s->replace(s->data, hit);
if (!pc)
pc = "NULL";
n = strlen(pc);
s->swallow[s->pos] = '\0';
if (n != s->pos) {
memmove(s->start + n,
s->start + s->pos,
old_len - (sp - args->p));
old_len += (n - s->pos) + 1;
}
memcpy(s->start, pc, n);
args->len++;
sp = s->start + 1;
s->pos = 0;
}
sp++;
continue;
}
args->len++;
sp++;
}
/* no space left for final chunk trailer */
if (args->final && args->len + 7 >= args->max_len)
return -1;
n = sprintf(buffer, "%X\x0d\x0a", args->len);
args->p -= n;
memcpy(args->p, buffer, n);
args->len += n;
if (args->final) {
sp = args->p + args->len;
*sp++ = '\x0d';
*sp++ = '\x0a';
*sp++ = '0';
*sp++ = '\x0d';
*sp++ = '\x0a';
*sp++ = '\x0d';
*sp++ = '\x0a';
args->len += 7;
} else {
sp = args->p + args->len;
*sp++ = '\x0d';
*sp++ = '\x0a';
args->len += 2;
}
return 0;
}