1
0
Fork 0
mirror of https://github.com/warmcat/libwebsockets.git synced 2025-03-16 00:00:07 +01:00
libwebsockets/lib/core-net/output.c
Andy Green 3549a94ce6 roles: compress role ops structs
role ops are usually only sparsely filled, there are currently 20
function pointers but several roles only fill in two.  No single
role has more than 14 of the ops.  On a 32/64 bit build this part
of the ops struct takes a fixed 80 / 160 bytes then.

First reduce the type of the callback reason part from uint16_t to
uint8_t, this saves 12 bytes unconditionally.

Change to a separate function pointer array with a nybble index
array, it costs 10 bytes for the index and a pointer to the
separate array, for 32-bit the cost is

2 + (4 x ops_used)

and for 64-bit

6 + (8 x ops_used)

for 2 x ops_used it means 32-bit: 10 vs 80 / 64-bit: 22 vs 160

For a typical system with h1 (9), h2 (14), listen (2), netlink (2),
pipe (1), raw_skt (3), ws (12), == 43 ops_used out of 140, it means
the .rodata for this reduced from 32-bit: 560 -> 174 (386 byte
saving) and 64-bit: 1120 -> 350 (770 byte saving)

This doesn't account for the changed function ops calling code, two
ways were tried, a preprocessor macro and explicit functions

For an x86_64 gcc 10 build with most options, release mode,
.text + .rodata

before patch:          553282
accessor macro:        552714 (568 byte saving)
accessor functions:    553674 (392 bytes worse than without patch)

therefore we went with the macros
2020-11-28 10:58:38 +00:00

415 lines
11 KiB
C

/*
* libwebsockets - small server side websockets and web server implementation
*
* Copyright (C) 2010 - 2019 Andy Green <andy@warmcat.com>
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to
* deal in the Software without restriction, including without limitation the
* rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
* sell copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
* FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
* IN THE SOFTWARE.
*/
#include "private-lib-core.h"
/*
* notice this returns number of bytes consumed, or -1
*/
int
lws_issue_raw(struct lws *wsi, unsigned char *buf, size_t len)
{
struct lws_context *context = lws_get_context(wsi);
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
size_t real_len = len;
unsigned int n, m;
/*
* If you're looking to dump data being sent down the tls tunnel, see
* lws_ssl_capable_write() in lib/tls/mbedtls/mbedtls-ssl.c or
* lib/tls/openssl/openssl-ssl.c.
*
* There's also a corresponding lws_ssl_capable_read() in those files
* where you can enable a dump of decrypted data as soon as it was
* read.
*/
/*
* Detect if we got called twice without going through the
* event loop to handle pending. Since that guarantees extending any
* existing buflist_out it's inefficient.
*/
if (0 && buf && wsi->could_have_pending) {
lwsl_hexdump_level(LLL_INFO, buf, len);
lwsl_info("** %p: vh: %s, prot: %s, role %s: "
"Inefficient back-to-back write of %lu detected...\n",
wsi, wsi->a.vhost ? wsi->a.vhost->name : "no vhost",
wsi->a.protocol->name, wsi->role_ops->name,
(unsigned long)len);
}
lws_stats_bump(pt, LWSSTATS_C_API_WRITE, 1);
/* just ignore sends after we cleared the truncation buffer */
if (lwsi_state(wsi) == LRS_FLUSHING_BEFORE_CLOSE &&
!lws_has_buffered_out(wsi)
#if defined(LWS_WITH_HTTP_STREAM_COMPRESSION)
&& !wsi->http.comp_ctx.may_have_more
#endif
)
return (int)len;
if (buf && lws_has_buffered_out(wsi)) {
lwsl_info("** %p: vh: %s, prot: %s, incr buflist_out by %lu\n",
wsi, wsi->a.vhost ? wsi->a.vhost->name : "no vhost",
wsi->a.protocol->name, (unsigned long)len);
/*
* already buflist ahead of this, add it on the tail of the
* buflist, then ignore it for now and act like we're flushing
* the buflist...
*/
if (lws_buflist_append_segment(&wsi->buflist_out, buf, len))
return -1;
buf = NULL;
len = 0;
}
if (wsi->buflist_out) {
/* we have to drain the earliest buflist_out stuff first */
len = lws_buflist_next_segment_len(&wsi->buflist_out, &buf);
real_len = len;
lwsl_debug("%s: draining %d\n", __func__, (int)len);
}
if (!len || !buf)
return 0;
if (!wsi->mux_substream && !lws_socket_is_valid(wsi->desc.sockfd))
lwsl_err("%s: invalid sock %p\n", __func__, wsi);
/* limit sending */
if (wsi->a.protocol->tx_packet_size)
n = (int)wsi->a.protocol->tx_packet_size;
else {
n = (int)wsi->a.protocol->rx_buffer_size;
if (!n)
n = context->pt_serv_buf_size;
}
n += LWS_PRE + 4;
if (n > len)
n = (int)len;
/* nope, send it on the socket directly */
m = lws_ssl_capable_write(wsi, buf, n);
lwsl_info("%s: ssl_capable_write (%d) says %d\n", __func__, n, m);
/* something got written, it can have been truncated now */
wsi->could_have_pending = 1;
switch (m) {
case LWS_SSL_CAPABLE_ERROR:
/* we're going to close, let close know sends aren't possible */
wsi->socket_is_permanently_unusable = 1;
return -1;
case LWS_SSL_CAPABLE_MORE_SERVICE:
/*
* nothing got sent, not fatal. Retry the whole thing later,
* ie, implying treat it was a truncated send so it gets
* retried
*/
m = 0;
break;
}
if ((int)m < 0)
m = 0;
/*
* we were sending this from buflist_out? Then not sending everything
* is a small matter of advancing ourselves only by the amount we did
* send in the buflist.
*/
if (lws_has_buffered_out(wsi)) {
if (m) {
lwsl_info("%p partial adv %d (vs %ld)\n", wsi, m,
(long)real_len);
lws_buflist_use_segment(&wsi->buflist_out, m);
}
if (!lws_has_buffered_out(wsi)) {
lwsl_info("%s: wsi %p: buflist_out flushed\n",
__func__, wsi);
m = (int)real_len;
if (lwsi_state(wsi) == LRS_FLUSHING_BEFORE_CLOSE) {
lwsl_info("*%p signalling to close now\n", wsi);
return -1; /* retry closing now */
}
if (wsi->close_when_buffered_out_drained) {
wsi->close_when_buffered_out_drained = 0;
return -1;
}
#if defined(LWS_ROLE_H1) || defined(LWS_ROLE_H2)
#if defined(LWS_WITH_SERVER)
if (wsi->http.deferred_transaction_completed) {
lwsl_notice("%s: partial completed, doing "
"deferred transaction completed\n",
__func__);
wsi->http.deferred_transaction_completed = 0;
return lws_http_transaction_completed(wsi) ?
-1 : (int)real_len;
}
#endif
#endif
#if defined(LWS_ROLE_WS)
/* Since buflist_out flushed, we're not inside a frame any more */
if (wsi->ws)
wsi->ws->inside_frame = 0;
#endif
}
/* always callback on writeable */
lws_callback_on_writable(wsi);
return m;
}
#if defined(LWS_WITH_HTTP_STREAM_COMPRESSION)
if (wsi->http.comp_ctx.may_have_more)
lws_callback_on_writable(wsi);
#endif
if (m == real_len)
/* what we just sent went out cleanly */
return m;
/*
* We were not able to send everything... and we were not sending from
* an existing buflist_out. So we are starting a fresh buflist_out, by
* buffering the unsent remainder on it.
* (it will get first priority next time the socket is writable).
*/
lwsl_debug("%p new partial sent %d from %lu total\n", wsi, m,
(unsigned long)real_len);
if (lws_buflist_append_segment(&wsi->buflist_out, buf + m,
real_len - m) < 0)
return -1;
lws_stats_bump(pt, LWSSTATS_C_WRITE_PARTIALS, 1);
lws_stats_bump(pt, LWSSTATS_B_PARTIALS_ACCEPTED_PARTS, m);
#if defined(LWS_WITH_UDP)
if (lws_wsi_is_udp(wsi))
/* stash original destination for fulfilling UDP partials */
wsi->udp->sa46_pending = wsi->udp->sa46;
#endif
/* since something buffered, force it to get another chance to send */
lws_callback_on_writable(wsi);
return (int)real_len;
}
int
lws_write(struct lws *wsi, unsigned char *buf, size_t len,
enum lws_write_protocol wp)
{
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
#if defined(LWS_WITH_DETAILED_LATENCY)
lws_usec_t us;
#endif
int m;
lws_stats_bump(pt, LWSSTATS_C_API_LWS_WRITE, 1);
if ((int)len < 0) {
lwsl_err("%s: suspicious len int %d, ulong %lu\n", __func__,
(int)len, (unsigned long)len);
return -1;
}
lws_stats_bump(pt, LWSSTATS_B_WRITE, len);
#ifdef LWS_WITH_ACCESS_LOG
wsi->http.access_log.sent += len;
#endif
#if defined(LWS_WITH_SERVER_STATUS)
if (wsi->a.vhost)
wsi->a.vhost->conn_stats.tx += len;
#endif
#if defined(LWS_WITH_DETAILED_LATENCY)
us = lws_now_usecs();
#endif
assert(wsi->role_ops);
if (!lws_rops_fidx(wsi->role_ops, LWS_ROPS_write_role_protocol))
return lws_issue_raw(wsi, buf, len);
m = lws_rops_func_fidx(wsi->role_ops, LWS_ROPS_write_role_protocol).
write_role_protocol(wsi, buf, len, &wp);
if (m < 0)
return m;
#if defined(LWS_WITH_DETAILED_LATENCY)
if (wsi->a.context->detailed_latency_cb) {
wsi->detlat.req_size = len;
wsi->detlat.acc_size = m;
wsi->detlat.type = LDLT_WRITE;
if (wsi->detlat.earliest_write_req_pre_write)
wsi->detlat.latencies[LAT_DUR_PROXY_PROXY_REQ_TO_WRITE] =
us - wsi->detlat.earliest_write_req_pre_write;
else
wsi->detlat.latencies[LAT_DUR_PROXY_PROXY_REQ_TO_WRITE] = 0;
wsi->detlat.latencies[LAT_DUR_USERCB] = lws_now_usecs() - us;
lws_det_lat_cb(wsi->a.context, &wsi->detlat);
}
#endif
return m;
}
int
lws_ssl_capable_read_no_ssl(struct lws *wsi, unsigned char *buf, int len)
{
struct lws_context *context = wsi->a.context;
struct lws_context_per_thread *pt = &context->pt[(int)wsi->tsi];
int n = 0;
lws_stats_bump(pt, LWSSTATS_C_API_READ, 1);
errno = 0;
#if defined(LWS_WITH_UDP)
if (lws_wsi_is_udp(wsi)) {
socklen_t slt = sizeof(wsi->udp->sa46);
n = recvfrom(wsi->desc.sockfd, (char *)buf, len, 0,
sa46_sockaddr(&wsi->udp->sa46), &slt);
} else
#endif
n = recv(wsi->desc.sockfd, (char *)buf, len, 0);
if (n >= 0) {
if (!n && wsi->unix_skt)
return LWS_SSL_CAPABLE_ERROR;
/*
* See https://libwebsockets.org/
* pipermail/libwebsockets/2019-March/007857.html
*/
if (!n)
return LWS_SSL_CAPABLE_ERROR;
#if defined(LWS_WITH_SERVER_STATUS)
if (wsi->a.vhost)
wsi->a.vhost->conn_stats.rx += n;
#endif
lws_stats_bump(pt, LWSSTATS_B_READ, n);
return n;
}
if (LWS_ERRNO == LWS_EAGAIN ||
LWS_ERRNO == LWS_EWOULDBLOCK ||
LWS_ERRNO == LWS_EINTR)
return LWS_SSL_CAPABLE_MORE_SERVICE;
lwsl_info("error on reading from skt : %d\n", LWS_ERRNO);
return LWS_SSL_CAPABLE_ERROR;
}
int
lws_ssl_capable_write_no_ssl(struct lws *wsi, unsigned char *buf, int len)
{
int n = 0;
#if defined(LWS_PLAT_OPTEE)
ssize_t send(int sockfd, const void *buf, size_t len, int flags);
#endif
#if defined(LWS_WITH_UDP)
if (lws_wsi_is_udp(wsi)) {
if (wsi->a.context->udp_loss_sim_tx_pc) {
uint16_t u16;
/*
* We should randomly drop some of these
*/
if (lws_get_random(wsi->a.context, &u16, 2) == 2 &&
((u16 * 100) / 0xffff) <=
wsi->a.context->udp_loss_sim_tx_pc) {
lwsl_warn("%s: dropping udp tx\n", __func__);
/* pretend it was sent */
n = len;
goto post_send;
}
}
if (lws_has_buffered_out(wsi))
n = sendto(wsi->desc.sockfd, (const char *)buf,
len, 0, sa46_sockaddr(&wsi->udp->sa46_pending),
sa46_socklen(&wsi->udp->sa46_pending));
else
n = sendto(wsi->desc.sockfd, (const char *)buf,
len, 0, sa46_sockaddr(&wsi->udp->sa46),
sa46_socklen(&wsi->udp->sa46));
} else
#endif
if (wsi->role_ops->file_handle)
n = write((int)(lws_intptr_t)wsi->desc.filefd, buf, len);
else
n = send(wsi->desc.sockfd, (char *)buf, len, MSG_NOSIGNAL);
// lwsl_info("%s: sent len %d result %d", __func__, len, n);
#if defined(LWS_WITH_UDP)
post_send:
#endif
if (n >= 0)
return n;
if (LWS_ERRNO == LWS_EAGAIN ||
LWS_ERRNO == LWS_EWOULDBLOCK ||
LWS_ERRNO == LWS_EINTR) {
if (LWS_ERRNO == LWS_EWOULDBLOCK) {
lws_set_blocking_send(wsi);
}
return LWS_SSL_CAPABLE_MORE_SERVICE;
}
lwsl_debug("ERROR writing len %d to skt fd %d err %d / errno %d\n",
len, wsi->desc.sockfd, n, LWS_ERRNO);
return LWS_SSL_CAPABLE_ERROR;
}
int
lws_ssl_pending_no_ssl(struct lws *wsi)
{
(void)wsi;
#if defined(LWS_PLAT_FREERTOS)
return 100;
#else
return 0;
#endif
}