1
0
Fork 0
mirror of https://github.com/warmcat/libwebsockets.git synced 2025-03-09 00:00:04 +01:00
libwebsockets/lib/roles/ws/ops-ws.c

2122 lines
56 KiB
C
Raw Permalink Normal View History

/*
* 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>
#define LWS_CPYAPP(ptr, str) { strcpy(ptr, str); ptr += strlen(str); }
/*
* client-parser.c: lws_ws_client_rx_sm() needs to be roughly kept in
* sync with changes here, esp related to ext draining
*/
int
lws_ws_rx_sm(struct lws *wsi, char already_processed, unsigned char c)
{
int callback_action = LWS_CALLBACK_RECEIVE;
2019-05-04 13:19:12 +01:00
struct lws_ext_pm_deflate_rx_ebufs pmdrx;
unsigned short close_code;
unsigned char *pp;
2019-05-04 13:19:12 +01:00
int ret = 0;
int n = 0;
#if !defined(LWS_WITHOUT_EXTENSIONS)
int rx_draining_ext = 0;
int lin;
#endif
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.token = NULL;
pmdrx.eb_in.len = 0;
pmdrx.eb_out.token = NULL;
pmdrx.eb_out.len = 0;
switch (wsi->lws_rx_parse_state) {
case LWS_RXPS_NEW:
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (wsi->ws->rx_draining_ext) {
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.token = NULL;
pmdrx.eb_in.len = 0;
pmdrx.eb_out.token = NULL;
pmdrx.eb_out.len = 0;
lws_remove_wsi_from_draining_ext_list(wsi);
rx_draining_ext = 1;
lwsl_debug("%s: doing draining flow\n", __func__);
goto drain_extension;
}
#endif
switch (wsi->ws->ietf_spec_revision) {
case 13:
/*
* no prepended frame key any more
*/
wsi->ws->all_zero_nonce = 1;
goto handle_first;
default:
lwsl_warn("lws_ws_rx_sm: unknown spec version %d\n",
wsi->ws->ietf_spec_revision);
break;
}
break;
case LWS_RXPS_04_mask_1:
wsi->ws->mask[1] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_04_mask_2;
break;
case LWS_RXPS_04_mask_2:
wsi->ws->mask[2] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_04_mask_3;
break;
case LWS_RXPS_04_mask_3:
wsi->ws->mask[3] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
/*
* start from the zero'th byte in the XOR key buffer since
* this is the start of a frame with a new key
*/
wsi->ws->mask_idx = 0;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_1;
break;
/*
* 04 logical framing from the spec (all this is masked when incoming
* and has to be unmasked)
*
* We ignore the possibility of extension data because we don't
* negotiate any extensions at the moment.
*
* 0 1 2 3
* 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
* +-+-+-+-+-------+-+-------------+-------------------------------+
* |F|R|R|R| opcode|R| Payload len | Extended payload length |
* |I|S|S|S| (4) |S| (7) | (16/63) |
* |N|V|V|V| |V| | (if payload len==126/127) |
* | |1|2|3| |4| | |
* +-+-+-+-+-------+-+-------------+ - - - - - - - - - - - - - - - +
* | Extended payload length continued, if payload len == 127 |
* + - - - - - - - - - - - - - - - +-------------------------------+
* | | Extension data |
* +-------------------------------+ - - - - - - - - - - - - - - - +
* : :
* +---------------------------------------------------------------+
* : Application data :
* +---------------------------------------------------------------+
*
* We pass payload through to userland as soon as we get it, ignoring
* FIN. It's up to userland to buffer it up if it wants to see a
* whole unfragmented block of the original size (which may be up to
* 2^63 long!)
*/
case LWS_RXPS_04_FRAME_HDR_1:
handle_first:
wsi->ws->opcode = c & 0xf;
wsi->ws->rsv = c & 0x70;
wsi->ws->final = !!((c >> 7) & 1);
wsi->ws->defeat_check_utf8 = 0;
if (((wsi->ws->opcode) & 8) && !wsi->ws->final) {
lws_close_reason(wsi, LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"frag ctl", 8);
return -1;
}
switch (wsi->ws->opcode) {
case LWSWSOPC_TEXT_FRAME:
wsi->ws->check_utf8 = lws_check_opt(
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
wsi->a.context->options,
LWS_SERVER_OPTION_VALIDATE_UTF8);
/* fallthru */
case LWSWSOPC_BINARY_FRAME:
if (wsi->ws->opcode == LWSWSOPC_BINARY_FRAME)
wsi->ws->check_utf8 = 0;
if (wsi->ws->continuation_possible) {
2018-11-23 08:47:56 +08:00
lws_close_reason(wsi,
LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"bad cont", 8);
return -1;
}
wsi->ws->rsv_first_msg = (c & 0x70);
2019-05-04 13:19:12 +01:00
#if !defined(LWS_WITHOUT_EXTENSIONS)
/*
* set the expectation that we will have to
* fake up the zlib trailer to the inflator for this
* frame
*/
wsi->ws->pmd_trailer_application = !!(c & 0x40);
#endif
wsi->ws->frame_is_binary =
wsi->ws->opcode == LWSWSOPC_BINARY_FRAME;
wsi->ws->first_fragment = 1;
wsi->ws->continuation_possible = !wsi->ws->final;
break;
case LWSWSOPC_CONTINUATION:
if (!wsi->ws->continuation_possible) {
2018-11-23 08:47:56 +08:00
lws_close_reason(wsi,
LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"bad cont", 8);
return -1;
}
break;
case LWSWSOPC_CLOSE:
wsi->ws->check_utf8 = 0;
wsi->ws->utf8 = 0;
break;
case 3:
case 4:
case 5:
case 6:
case 7:
case 0xb:
case 0xc:
case 0xd:
case 0xe:
case 0xf:
2018-11-23 08:47:56 +08:00
lws_close_reason(wsi, LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"bad opc", 7);
lwsl_info("illegal opcode\n");
return -1;
}
if (wsi->ws->owed_a_fin &&
(wsi->ws->opcode == LWSWSOPC_TEXT_FRAME ||
wsi->ws->opcode == LWSWSOPC_BINARY_FRAME)) {
lwsl_info("hey you owed us a FIN\n");
2018-11-23 08:47:56 +08:00
lws_close_reason(wsi, LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"bad fin", 7);
return -1;
}
if ((!(wsi->ws->opcode & 8)) && wsi->ws->final) {
wsi->ws->continuation_possible = 0;
wsi->ws->owed_a_fin = 0;
}
if (!wsi->ws->final)
wsi->ws->owed_a_fin = 1;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN;
if (wsi->ws->rsv &&
(
#if !defined(LWS_WITHOUT_EXTENSIONS)
!wsi->ws->count_act_ext ||
#endif
(wsi->ws->rsv & ~0x40))) {
lws_close_reason(wsi, LWS_CLOSE_STATUS_PROTOCOL_ERR,
(uint8_t *)"rsv bits", 8);
return -1;
}
break;
case LWS_RXPS_04_FRAME_HDR_LEN:
wsi->ws->this_frame_masked = !!(c & 0x80);
switch (c & 0x7f) {
case 126:
/* control frames are not allowed to have big lengths */
if (wsi->ws->opcode & 8)
goto illegal_ctl_length;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN16_2;
break;
case 127:
/* control frames are not allowed to have big lengths */
if (wsi->ws->opcode & 8)
goto illegal_ctl_length;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_8;
break;
default:
wsi->ws->rx_packet_length = c & 0x7f;
if (wsi->ws->this_frame_masked)
wsi->lws_rx_parse_state =
LWS_RXPS_07_COLLECT_FRAME_KEY_1;
else
if (wsi->ws->rx_packet_length) {
wsi->lws_rx_parse_state =
LWS_RXPS_WS_FRAME_PAYLOAD;
} else {
wsi->lws_rx_parse_state = LWS_RXPS_NEW;
goto spill;
}
break;
}
break;
case LWS_RXPS_04_FRAME_HDR_LEN16_2:
wsi->ws->rx_packet_length = (size_t)(c << 8);
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN16_1;
break;
case LWS_RXPS_04_FRAME_HDR_LEN16_1:
wsi->ws->rx_packet_length |= c;
if (wsi->ws->this_frame_masked)
wsi->lws_rx_parse_state =
LWS_RXPS_07_COLLECT_FRAME_KEY_1;
else {
wsi->lws_rx_parse_state =
LWS_RXPS_WS_FRAME_PAYLOAD;
}
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_8:
if (c & 0x80) {
lwsl_warn("b63 of length must be zero\n");
/* kill the connection */
return -1;
}
#if defined __LP64__
wsi->ws->rx_packet_length = ((size_t)c) << 56;
#else
wsi->ws->rx_packet_length = 0;
#endif
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_7;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_7:
#if defined __LP64__
wsi->ws->rx_packet_length |= ((size_t)c) << 48;
#endif
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_6;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_6:
#if defined __LP64__
wsi->ws->rx_packet_length |= ((size_t)c) << 40;
#endif
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_5;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_5:
#if defined __LP64__
wsi->ws->rx_packet_length |= ((size_t)c) << 32;
#endif
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_4;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_4:
wsi->ws->rx_packet_length |= ((size_t)c) << 24;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_3;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_3:
wsi->ws->rx_packet_length |= ((size_t)c) << 16;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_2;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_2:
wsi->ws->rx_packet_length |= ((size_t)c) << 8;
wsi->lws_rx_parse_state = LWS_RXPS_04_FRAME_HDR_LEN64_1;
break;
case LWS_RXPS_04_FRAME_HDR_LEN64_1:
wsi->ws->rx_packet_length |= ((size_t)c);
if (wsi->ws->this_frame_masked)
wsi->lws_rx_parse_state =
LWS_RXPS_07_COLLECT_FRAME_KEY_1;
else
wsi->lws_rx_parse_state = LWS_RXPS_WS_FRAME_PAYLOAD;
break;
case LWS_RXPS_07_COLLECT_FRAME_KEY_1:
wsi->ws->mask[0] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_07_COLLECT_FRAME_KEY_2;
break;
case LWS_RXPS_07_COLLECT_FRAME_KEY_2:
wsi->ws->mask[1] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_07_COLLECT_FRAME_KEY_3;
break;
case LWS_RXPS_07_COLLECT_FRAME_KEY_3:
wsi->ws->mask[2] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_07_COLLECT_FRAME_KEY_4;
break;
case LWS_RXPS_07_COLLECT_FRAME_KEY_4:
wsi->ws->mask[3] = c;
if (c)
wsi->ws->all_zero_nonce = 0;
wsi->lws_rx_parse_state = LWS_RXPS_WS_FRAME_PAYLOAD;
wsi->ws->mask_idx = 0;
if (wsi->ws->rx_packet_length == 0) {
wsi->lws_rx_parse_state = LWS_RXPS_NEW;
goto spill;
}
break;
case LWS_RXPS_WS_FRAME_PAYLOAD:
assert(wsi->ws->rx_ubuf);
if (wsi->ws->rx_ubuf_head + LWS_PRE >= wsi->ws->rx_ubuf_alloc) {
lwsl_err("Attempted overflow \n");
return -1;
}
if (!(already_processed & ALREADY_PROCESSED_IGNORE_CHAR)) {
if (wsi->ws->all_zero_nonce)
2018-11-23 08:47:56 +08:00
wsi->ws->rx_ubuf[LWS_PRE +
(wsi->ws->rx_ubuf_head++)] = c;
else
2018-11-23 08:47:56 +08:00
wsi->ws->rx_ubuf[LWS_PRE +
(wsi->ws->rx_ubuf_head++)] =
c ^ wsi->ws->mask[(wsi->ws->mask_idx++) & 3];
--wsi->ws->rx_packet_length;
}
if (!wsi->ws->rx_packet_length) {
2018-11-23 08:47:56 +08:00
lwsl_debug("%s: ws fragment length exhausted\n",
__func__);
/* spill because we have the whole frame */
wsi->lws_rx_parse_state = LWS_RXPS_NEW;
goto spill;
}
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (wsi->ws->rx_draining_ext) {
lwsl_debug("%s: UNTIL_EXHAUSTED draining\n", __func__);
goto drain_extension;
}
#endif
/*
* if there's no protocol max frame size given, we are
* supposed to default to context->pt_serv_buf_size
*/
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if (!wsi->a.protocol->rx_buffer_size &&
wsi->ws->rx_ubuf_head != wsi->a.context->pt_serv_buf_size)
break;
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if (wsi->a.protocol->rx_buffer_size &&
wsi->ws->rx_ubuf_head != wsi->a.protocol->rx_buffer_size)
break;
/* spill because we filled our rx buffer */
spill:
/*
* is this frame a control packet we should take care of at this
* layer? If so service it and hide it from the user callback
*/
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
lwsl_parser("spill on %s\n", wsi->a.protocol->name);
switch (wsi->ws->opcode) {
case LWSWSOPC_CLOSE:
if (wsi->ws->peer_has_sent_close)
break;
wsi->ws->peer_has_sent_close = 1;
pp = &wsi->ws->rx_ubuf[LWS_PRE];
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if (lws_check_opt(wsi->a.context->options,
LWS_SERVER_OPTION_VALIDATE_UTF8) &&
wsi->ws->rx_ubuf_head > 2 &&
lws_check_utf8(&wsi->ws->utf8, pp + 2,
wsi->ws->rx_ubuf_head - 2))
goto utf8_fail;
/* is this an acknowledgment of our close? */
if (lwsi_state(wsi) == LRS_AWAITING_CLOSE_ACK) {
/*
* fine he has told us he is closing too, let's
* finish our close
*/
lwsl_parser("seen client close ack\n");
return -1;
}
if (lwsi_state(wsi) == LRS_RETURNED_CLOSE)
/* if he sends us 2 CLOSE, kill him */
return -1;
if (lws_partial_buffered(wsi)) {
/*
* if we're in the middle of something,
* we can't do a normal close response and
* have to just close our end.
*/
wsi->socket_is_permanently_unusable = 1;
2018-11-23 08:47:56 +08:00
lwsl_parser("Closing on peer close "
"due to pending tx\n");
return -1;
}
if (wsi->ws->rx_ubuf_head >= 2) {
close_code = (unsigned short)((pp[0] << 8) | pp[1]);
if (close_code < 1000 ||
close_code == 1004 ||
close_code == 1005 ||
close_code == 1006 ||
close_code == 1012 ||
close_code == 1013 ||
close_code == 1014 ||
close_code == 1015 ||
(close_code >= 1016 && close_code < 3000)
) {
pp[0] = (LWS_CLOSE_STATUS_PROTOCOL_ERR >> 8) & 0xff;
pp[1] = LWS_CLOSE_STATUS_PROTOCOL_ERR & 0xff;
}
}
if (user_callback_handle_rxflow(
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
wsi->a.protocol->callback, wsi,
LWS_CALLBACK_WS_PEER_INITIATED_CLOSE,
wsi->user_space,
&wsi->ws->rx_ubuf[LWS_PRE],
wsi->ws->rx_ubuf_head))
return -1;
lwsl_parser("server sees client close packet\n");
lwsi_set_state(wsi, LRS_RETURNED_CLOSE);
/* deal with the close packet contents as a PONG */
wsi->ws->payload_is_close = 1;
goto process_as_ping;
case LWSWSOPC_PING:
lwsl_info("received %d byte ping, sending pong\n",
(int)wsi->ws->rx_ubuf_head);
2021-09-13 17:57:08 +03:00
if (wsi->ws->pong_pending_flag) {
/*
2021-09-13 17:57:08 +03:00
* there is already a pending pong payload
* we should just log and drop
*/
lwsl_parser("DROP PING since one pending\n");
goto ping_drop;
}
process_as_ping:
/* control packets can only be < 128 bytes long */
if (wsi->ws->rx_ubuf_head > 128 - 3) {
lwsl_parser("DROP PING payload too large\n");
goto ping_drop;
}
/* stash the pong payload */
2021-09-13 17:57:08 +03:00
memcpy(wsi->ws->pong_payload_buf + LWS_PRE,
&wsi->ws->rx_ubuf[LWS_PRE],
wsi->ws->rx_ubuf_head);
2021-09-13 17:57:08 +03:00
wsi->ws->pong_payload_len = (uint8_t)wsi->ws->rx_ubuf_head;
wsi->ws->pong_pending_flag = 1;
/* get it sent as soon as possible */
lws_callback_on_writable(wsi);
ping_drop:
wsi->ws->rx_ubuf_head = 0;
return 0;
case LWSWSOPC_PONG:
lwsl_info("received pong\n");
lwsl_hexdump(&wsi->ws->rx_ubuf[LWS_PRE],
wsi->ws->rx_ubuf_head);
lws_validity_confirmed(wsi);
/* issue it */
callback_action = LWS_CALLBACK_RECEIVE_PONG;
break;
case LWSWSOPC_TEXT_FRAME:
case LWSWSOPC_BINARY_FRAME:
case LWSWSOPC_CONTINUATION:
break;
default:
lwsl_parser("unknown opc %x\n", wsi->ws->opcode);
return -1;
}
/*
* No it's real payload, pass it up to the user callback.
2019-05-04 13:19:12 +01:00
*
* We have been statefully collecting it in the
* LWS_RXPS_WS_FRAME_PAYLOAD clause above.
*
* It's nicely buffered with the pre-padding taken care of
2019-05-04 13:19:12 +01:00
* so it can be sent straight out again using lws_write.
*
* However, now we have a chunk of it, we want to deal with it
* all here. Since this may be input to permessage-deflate and
* there are block limits on that for input and output, we may
* need to iterate.
*/
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.token = &wsi->ws->rx_ubuf[LWS_PRE];
pmdrx.eb_in.len = (int)wsi->ws->rx_ubuf_head;
2019-05-04 13:19:12 +01:00
/* for the non-pm-deflate case */
pmdrx.eb_out = pmdrx.eb_in;
2019-05-04 13:19:12 +01:00
if (wsi->ws->opcode == LWSWSOPC_PONG && !pmdrx.eb_in.len)
goto already_done;
#if !defined(LWS_WITHOUT_EXTENSIONS)
drain_extension:
#endif
2019-05-04 13:19:12 +01:00
do {
// lwsl_notice("%s: pmdrx.eb_in.len: %d\n", __func__,
// (int)pmdrx.eb_in.len);
if (lwsi_state(wsi) == LRS_RETURNED_CLOSE ||
lwsi_state(wsi) == LRS_AWAITING_CLOSE_ACK)
goto already_done;
n = PMDR_DID_NOTHING;
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
lin = pmdrx.eb_in.len;
//if (lin)
// lwsl_hexdump_notice(ebuf.token, ebuf.len);
lwsl_ext("%s: +++ passing %d %p to ext\n", __func__,
pmdrx.eb_in.len, pmdrx.eb_in.token);
n = lws_ext_cb_active(wsi, LWS_EXT_CB_PAYLOAD_RX, &pmdrx, 0);
lwsl_debug("%s: ext says %d / ebuf.len %d\n", __func__,
n, pmdrx.eb_out.len);
if (wsi->ws->rx_draining_ext)
already_processed &= (char)~ALREADY_PROCESSED_NO_CB;
#endif
2019-05-04 13:19:12 +01:00
/*
2019-05-04 13:19:12 +01:00
* ebuf may be pointing somewhere completely different
* now, it's the output
*/
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
if (n < 0) {
/*
* we may rely on this to get RX, just drop
* connection
*/
wsi->socket_is_permanently_unusable = 1;
return -1;
}
if (n == PMDR_DID_NOTHING)
/* ie, not PMDR_NOTHING_WE_SHOULD_DO */
2019-05-04 13:19:12 +01:00
break;
#endif
unix plat: add minimal wsi fd map option An lws context usually contains a processwide fd -> wsi lookup table. This allows any possible fd returned by a *nix type OS to be immediately converted to a wsi just by indexing an array of struct lws * the size of the highest possible fd, as found by ulimit -n or similar. This works modestly for Linux type systems where the default ulimit -n for a process is 1024, it means a 4KB or 8KB lookup table for 32-bit or 64-bit systems. However in the case your lws usage is much simpler, like one outgoing client connection and no serving, this represents increasing waste. It's made much worse if the system has a much larger default ulimit -n, eg 1M, the table is occupying 4MB or 8MB, of which you will only use one. Even so, because lws can't be sure the OS won't return a socket fd at any number up to (ulimit -n - 1), it has to allocate the whole lookup table at the moment. This patch looks to see if the context creation info is setting info->fd_limit_per_thread... if it leaves it at the default 0, then everything is as it was before this patch. However if finds that (info->fd_limit_per_thread * actual_number_of_service_threads) where the default number of service threads is 1, is less than the fd limit set by ulimit -n, lws switches to a slower lookup table scheme, which only allocates the requested number of slots. Lookups happen then by iterating the table and comparing rather than indexing the array directly, which is obviously somewhat of a performance hit. However in the case where you know lws will only have a very few wsi maximum, this method can very usefully trade off speed to be able to avoid the allocation sized by ulimit -n. minimal examples for client that can make use of this are also modified by this patch to use the smaller context allocations.
2019-05-17 01:20:07 +01:00
lwsl_debug("%s: post ext ret %d, ebuf in %d / out %d\n",
2019-05-04 13:19:12 +01:00
__func__, n, pmdrx.eb_in.len,
pmdrx.eb_out.len);
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
if (rx_draining_ext && !pmdrx.eb_out.len) {
lwsl_debug(" --- ending drain on 0 read\n");
goto already_done;
}
if (n == PMDR_HAS_PENDING)
/*
* extension had more...
* main loop will come back
*/
lws_add_wsi_to_draining_ext_list(wsi);
else
lws_remove_wsi_from_draining_ext_list(wsi);
rx_draining_ext = wsi->ws->rx_draining_ext;
#endif
if (pmdrx.eb_out.len &&
wsi->ws->check_utf8 && !wsi->ws->defeat_check_utf8) {
2019-05-04 13:19:12 +01:00
if (lws_check_utf8(&wsi->ws->utf8,
pmdrx.eb_out.token,
(size_t)pmdrx.eb_out.len)) {
2019-05-04 13:19:12 +01:00
lws_close_reason(wsi,
LWS_CLOSE_STATUS_INVALID_PAYLOAD,
(uint8_t *)"bad utf8", 8);
goto utf8_fail;
}
2019-05-04 13:19:12 +01:00
/* we are ending partway through utf-8 character? */
if (!wsi->ws->rx_packet_length &&
wsi->ws->final && wsi->ws->utf8
#if !defined(LWS_WITHOUT_EXTENSIONS)
/* if ext not negotiated, going to be UNKNOWN */
&& (n == PMDR_EMPTY_FINAL || n == PMDR_UNKNOWN)
#endif
) {
lwsl_info("FINAL utf8 error\n");
lws_close_reason(wsi,
LWS_CLOSE_STATUS_INVALID_PAYLOAD,
(uint8_t *)"partial utf8", 12);
utf8_fail:
2019-05-04 13:19:12 +01:00
lwsl_notice("utf8 error\n");
lwsl_hexdump_notice(pmdrx.eb_out.token,
(size_t)pmdrx.eb_out.len);
2019-05-04 13:19:12 +01:00
return -1;
}
}
2019-05-04 13:19:12 +01:00
/* if pmd not enabled, in == out */
2019-06-08 10:29:08 +01:00
if (n == PMDR_DID_NOTHING
#if !defined(LWS_WITHOUT_EXTENSIONS)
||
n == PMDR_NOTHING_WE_SHOULD_DO ||
2019-06-08 10:29:08 +01:00
n == PMDR_UNKNOWN
#endif
)
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.len -= pmdrx.eb_out.len;
2019-06-08 10:29:08 +01:00
if (!wsi->wsistate_pre_close &&
2019-05-04 13:19:12 +01:00
(pmdrx.eb_out.len >= 0 ||
callback_action == LWS_CALLBACK_RECEIVE_PONG ||
n == PMDR_EMPTY_FINAL)) {
if (pmdrx.eb_out.len)
pmdrx.eb_out.token[pmdrx.eb_out.len] = '\0';
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if (wsi->a.protocol->callback &&
2019-05-04 13:19:12 +01:00
!(already_processed & ALREADY_PROCESSED_NO_CB)) {
if (callback_action ==
LWS_CALLBACK_RECEIVE_PONG)
lwsl_info("Doing pong callback\n");
ret = user_callback_handle_rxflow(
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
wsi->a.protocol->callback, wsi,
2019-05-04 13:19:12 +01:00
(enum lws_callback_reasons)
callback_action,
wsi->user_space,
2019-05-04 13:19:12 +01:00
pmdrx.eb_out.token,
(size_t)pmdrx.eb_out.len);
2019-05-04 13:19:12 +01:00
}
wsi->ws->first_fragment = 0;
}
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
if (!lin)
break;
#endif
} while (pmdrx.eb_in.len
#if !defined(LWS_WITHOUT_EXTENSIONS)
|| rx_draining_ext
#endif
2019-05-04 13:19:12 +01:00
);
already_done:
wsi->ws->rx_ubuf_head = 0;
break;
}
return ret;
illegal_ctl_length:
lwsl_warn("Control frame with xtended length is illegal\n");
/* kill the connection */
return -1;
}
2020-01-02 08:32:23 +00:00
size_t
lws_remaining_packet_payload(struct lws *wsi)
{
return wsi->ws->rx_packet_length;
}
2020-01-02 08:32:23 +00:00
int lws_frame_is_binary(struct lws *wsi)
{
return wsi->ws->frame_is_binary;
}
void
lws_add_wsi_to_draining_ext_list(struct lws *wsi)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
if (wsi->ws->rx_draining_ext)
return;
lwsl_debug("%s: RX EXT DRAINING: Adding to list\n", __func__);
wsi->ws->rx_draining_ext = 1;
wsi->ws->rx_draining_ext_list = pt->ws.rx_draining_ext_list;
pt->ws.rx_draining_ext_list = wsi;
#endif
}
void
lws_remove_wsi_from_draining_ext_list(struct lws *wsi)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
struct lws **w = &pt->ws.rx_draining_ext_list;
if (!wsi->ws->rx_draining_ext)
return;
lwsl_debug("%s: RX EXT DRAINING: Removing from list\n", __func__);
wsi->ws->rx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
/* if us, point it instead to who we were pointing to */
*w = wsi->ws->rx_draining_ext_list;
break;
}
w = &((*w)->ws->rx_draining_ext_list);
}
wsi->ws->rx_draining_ext_list = NULL;
#endif
}
static int
lws_0405_frame_mask_generate(struct lws *wsi)
{
size_t n;
/* fetch the per-frame nonce */
n = lws_get_random(lws_get_context(wsi), wsi->ws->mask, 4);
if (n != 4) {
lwsl_parser("Unable to read from random device %s %d\n",
SYSTEM_RANDOM_FILEPATH, (int)n);
return 1;
}
/* start masking from first byte of masking key buffer */
wsi->ws->mask_idx = 0;
return 0;
}
int
lws_server_init_wsi_for_ws(struct lws *wsi)
{
int n;
lwsi_set_state(wsi, LRS_ESTABLISHED);
/*
* 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
*/
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
n = (int)wsi->a.protocol->rx_buffer_size;
if (!n)
n = (int)wsi->a.context->pt_serv_buf_size;
n += LWS_PRE;
wsi->ws->rx_ubuf = lws_malloc((unsigned int)n + 4 /* 0x0000ffff zlib */, "rx_ubuf");
if (!wsi->ws->rx_ubuf) {
lwsl_err("Out of Mem allocating rx buffer %d\n", n);
return 1;
}
wsi->ws->rx_ubuf_alloc = (uint32_t)n;
/* notify user code that we're ready to roll */
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if (wsi->a.protocol->callback)
if (wsi->a.protocol->callback(wsi, LWS_CALLBACK_ESTABLISHED,
wsi->user_space,
#ifdef LWS_WITH_TLS
wsi->tls.ssl,
#else
NULL,
#endif
wsi->h2_stream_carries_ws))
return 1;
lws_validity_confirmed(wsi);
lwsl_debug("ws established\n");
return 0;
}
2020-01-02 08:32:23 +00:00
int
lws_is_final_fragment(struct lws *wsi)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
2018-11-23 08:47:56 +08:00
lwsl_debug("%s: final %d, rx pk length %ld, draining %ld\n", __func__,
wsi->ws->final, (long)wsi->ws->rx_packet_length,
(long)wsi->ws->rx_draining_ext);
return wsi->ws->final && !wsi->ws->rx_packet_length &&
!wsi->ws->rx_draining_ext;
#else
return wsi->ws->final && !wsi->ws->rx_packet_length;
#endif
}
2020-01-02 08:32:23 +00:00
int
lws_is_first_fragment(struct lws *wsi)
{
return wsi->ws->first_fragment;
}
2020-01-02 08:32:23 +00:00
unsigned char
lws_get_reserved_bits(struct lws *wsi)
{
return wsi->ws->rsv;
}
2020-01-02 08:32:23 +00:00
int
lws_get_close_length(struct lws *wsi)
{
return wsi->ws->close_in_ping_buffer_len;
}
2020-01-02 08:32:23 +00:00
unsigned char *
lws_get_close_payload(struct lws *wsi)
{
return &wsi->ws->ping_payload_buf[LWS_PRE];
}
2020-01-02 08:32:23 +00:00
void
lws_close_reason(struct lws *wsi, enum lws_close_status status,
unsigned char *buf, size_t len)
{
unsigned char *p, *start;
int budget = sizeof(wsi->ws->ping_payload_buf) - LWS_PRE;
assert(lwsi_role_ws(wsi));
start = p = &wsi->ws->ping_payload_buf[LWS_PRE];
*p++ = (uint8_t)((((int)status) >> 8) & 0xff);
*p++ = (uint8_t)(((int)status) & 0xff);
if (buf)
while (len-- && p < start + budget)
*p++ = *buf++;
wsi->ws->close_in_ping_buffer_len = (uint8_t)lws_ptr_diff(p, start);
}
static int
lws_is_ws_with_ext(struct lws *wsi)
{
#if defined(LWS_WITHOUT_EXTENSIONS)
return 0;
#else
return lwsi_role_ws(wsi) && !!wsi->ws->count_act_ext;
#endif
}
static int
rops_handle_POLLIN_ws(struct lws_context_per_thread *pt, struct lws *wsi,
struct lws_pollfd *pollfd)
{
unsigned int pending = 0;
2018-11-23 08:47:56 +08:00
struct lws_tokens ebuf;
char buffered = 0;
int n = 0, m;
#if defined(LWS_WITH_HTTP2)
struct lws *wsi1;
#endif
if (!wsi->ws) {
lwsl_err("ws role wsi with no ws\n");
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
// lwsl_notice("%s: %s\n", __func__, wsi->a.protocol->name);
//lwsl_info("%s: wsistate 0x%x, pollout %d\n", __func__,
// wsi->wsistate, pollfd->revents & LWS_POLLOUT);
/*
* something went wrong with parsing the handshake, and
* we ended up back in the event loop without completing it
*/
if (lwsi_state(wsi) == LRS_PRE_WS_SERVING_ACCEPT) {
wsi->socket_is_permanently_unusable = 1;
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
ebuf.token = NULL;
ebuf.len = 0;
if (lwsi_state(wsi) == LRS_WAITING_CONNECT) {
#if defined(LWS_WITH_CLIENT)
if ((pollfd->revents & LWS_POLLOUT) &&
lws_handle_POLLOUT_event(wsi, pollfd)) {
lwsl_debug("POLLOUT event closed it\n");
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
n = lws_http_client_socket_service(wsi, pollfd);
if (n)
return LWS_HPI_RET_WSI_ALREADY_DIED;
#endif
return LWS_HPI_RET_HANDLED;
}
/* 1: something requested a callback when it was OK to write */
if ((pollfd->revents & LWS_POLLOUT) &&
lwsi_state_can_handle_POLLOUT(wsi) &&
lws_handle_POLLOUT_event(wsi, pollfd)) {
if (lwsi_state(wsi) == LRS_RETURNED_CLOSE)
lwsi_set_state(wsi, LRS_FLUSHING_BEFORE_CLOSE);
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
if (lwsi_state(wsi) == LRS_RETURNED_CLOSE ||
lwsi_state(wsi) == LRS_WAITING_TO_SEND_CLOSE) {
/*
* we stopped caring about anything except control
* packets. Force flow control off, defeat tx
* draining.
*/
lws_rx_flow_control(wsi, 1);
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (wsi->ws)
wsi->ws->tx_draining_ext = 0;
#endif
}
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
if (wsi->ws->tx_draining_ext) {
lws_handle_POLLOUT_event(wsi, pollfd);
//lwsl_notice("%s: tx drain\n", __func__);
/*
* We cannot deal with new RX until the TX ext path has
* been drained. It's because new rx will, eg, crap on
* the wsi rx buf that may be needed to retain state.
*
* TX ext drain path MUST go through event loop to avoid
* blocking.
*/
2019-05-04 13:19:12 +01:00
lws_callback_on_writable(wsi);
return LWS_HPI_RET_HANDLED;
2019-05-04 13:19:12 +01:00
}
#endif
if ((pollfd->revents & LWS_POLLIN) && lws_is_flowcontrolled(wsi)) {
/* We cannot deal with any kind of new RX because we are
* RX-flowcontrolled.
*/
lwsl_info("%s: flowcontrolled, ignoring rx\n", __func__);
if (__lws_change_pollfd(wsi, LWS_POLLIN, 0))
return -1;
return LWS_HPI_RET_HANDLED;
}
if (lws_is_flowcontrolled(wsi))
return LWS_HPI_RET_HANDLED;
#if defined(LWS_WITH_HTTP2)
if (wsi->mux_substream || wsi->upgraded_to_http2) {
wsi1 = lws_get_network_wsi(wsi);
if (wsi1 && lws_has_buffered_out(wsi1))
/* We cannot deal with any kind of new RX
* because we are dealing with a partial send
* (new RX may trigger new http_action() that
* expect to be able to send)
*/
return LWS_HPI_RET_HANDLED;
}
#endif
#if !defined(LWS_WITHOUT_EXTENSIONS)
/* 2: RX Extension needs to be drained
*/
if (wsi->ws->rx_draining_ext) {
lwsl_debug("%s: RX EXT DRAINING: Service\n", __func__);
#if defined(LWS_WITH_CLIENT)
if (lwsi_role_client(wsi)) {
n = lws_ws_client_rx_sm(wsi, 0);
if (n < 0)
/* we closed wsi */
return LWS_HPI_RET_PLEASE_CLOSE_ME;
} else
#endif
n = lws_ws_rx_sm(wsi, ALREADY_PROCESSED_IGNORE_CHAR, 0);
return LWS_HPI_RET_HANDLED;
}
if (wsi->ws->rx_draining_ext)
/*
* We have RX EXT content to drain, but can't do it
* right now. That means we cannot do anything lower
* priority either.
*/
return LWS_HPI_RET_HANDLED;
#endif
/* 3: buflist needs to be drained
*/
read:
//lws_buflist_describe(&wsi->buflist, wsi, __func__);
ebuf.len = (int)lws_buflist_next_segment_len(&wsi->buflist,
&ebuf.token);
if (ebuf.len) {
lwsl_info("draining buflist (len %d)\n", ebuf.len);
buffered = 1;
goto drain;
}
if (!(pollfd->revents & pollfd->events & LWS_POLLIN) && !wsi->http.ah)
return LWS_HPI_RET_HANDLED;
if (lws_is_flowcontrolled(wsi)) {
lwsl_info("%s: %p should be rxflow (bm 0x%x)..\n",
__func__, wsi, wsi->rxflow_bitmap);
return LWS_HPI_RET_HANDLED;
}
if (!(lwsi_role_client(wsi) &&
(lwsi_state(wsi) != LRS_ESTABLISHED &&
lwsi_state(wsi) != LRS_AWAITING_CLOSE_ACK &&
lwsi_state(wsi) != LRS_H2_WAITING_TO_SEND_HEADERS))) {
/*
* In case we are going to react to this rx by scheduling
* writes, we need to restrict the amount of rx to the size
* the protocol reported for rx buffer.
*
* Otherwise we get a situation we have to absorb possibly a
* lot of reads before we get a chance to drain them by writing
* them, eg, with echo type tests in autobahn.
*/
buffered = 0;
ebuf.token = pt->serv_buf;
if (lwsi_role_ws(wsi))
ebuf.len = (int)wsi->ws->rx_ubuf_alloc;
else
ebuf.len = (int)wsi->a.context->pt_serv_buf_size;
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
if ((unsigned int)ebuf.len > wsi->a.context->pt_serv_buf_size)
ebuf.len = (int)wsi->a.context->pt_serv_buf_size;
if ((int)pending > ebuf.len)
pending = (unsigned int)ebuf.len;
ebuf.len = lws_ssl_capable_read(wsi, ebuf.token,
(size_t)(pending ? pending :
(unsigned int)ebuf.len));
switch (ebuf.len) {
case 0:
lwsl_info("%s: zero length read\n",
__func__);
return LWS_HPI_RET_PLEASE_CLOSE_ME;
case LWS_SSL_CAPABLE_MORE_SERVICE:
lwsl_info("SSL Capable more service\n");
return LWS_HPI_RET_HANDLED;
case LWS_SSL_CAPABLE_ERROR:
lwsl_info("%s: LWS_SSL_CAPABLE_ERROR\n",
__func__);
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
/*
* coverity thinks ssl_capable_read() may read over
* 2GB. Dissuade it...
*/
ebuf.len &= 0x7fffffff;
}
drain:
/*
* give any active extensions a chance to munge the buffer
* before parse. We pass in a pointer to an lws_tokens struct
* prepared with the default buffer and content length that's in
* there. Rather than rewrite the default buffer, extensions
* that expect to grow the buffer can adapt .token to
* point to their own per-connection buffer in the extension
* user allocation. By default with no extensions or no
* extension callback handling, just the normal input buffer is
* used then so it is efficient.
*/
m = 0;
do {
/* service incoming data */
//lws_buflist_describe(&wsi->buflist, wsi, __func__);
if (ebuf.len > 0) {
#if defined(LWS_ROLE_H2)
if (lwsi_role_h2(wsi) && lwsi_state(wsi) != LRS_BODY &&
lwsi_state(wsi) != LRS_DISCARD_BODY)
n = lws_read_h2(wsi, ebuf.token,
(unsigned int)ebuf.len);
else
#endif
n = lws_read_h1(wsi, ebuf.token,
(unsigned int)ebuf.len);
if (n < 0) {
/* we closed wsi */
return LWS_HPI_RET_WSI_ALREADY_DIED;
2018-04-17 11:43:20 +08:00
}
//lws_buflist_describe(&wsi->buflist, wsi, __func__);
//lwsl_notice("%s: consuming %d / %d\n", __func__, n, ebuf.len);
2020-12-19 09:13:24 +00:00
if (ebuf.len < 0 ||
lws_buflist_aware_finished_consuming(wsi, &ebuf, n,
buffered, __func__))
return LWS_HPI_RET_PLEASE_CLOSE_ME;
}
ebuf.token = NULL;
ebuf.len = 0;
} while (m);
if (wsi->http.ah
#if defined(LWS_WITH_CLIENT)
&& !wsi->client_h2_alpn
#endif
) {
lwsl_info("%s: %p: detaching ah\n", __func__, wsi);
lws_header_table_detach(wsi, 0);
}
pending = (unsigned int)lws_ssl_pending(wsi);
2020-08-23 10:05:40 +01:00
#if defined(LWS_WITH_CLIENT)
if (!pending && (wsi->flags & LCCSCF_PRIORITIZE_READS) &&
lws_buflist_total_len(&wsi->buflist))
pending = 9999999;
#endif
if (pending) {
if (lws_is_ws_with_ext(wsi))
pending = pending > wsi->ws->rx_ubuf_alloc ?
wsi->ws->rx_ubuf_alloc : pending;
else
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
pending = pending > wsi->a.context->pt_serv_buf_size ?
wsi->a.context->pt_serv_buf_size : pending;
goto read;
}
if (buffered && /* were draining, now nothing left */
!lws_buflist_next_segment_len(&wsi->buflist, NULL)) {
lwsl_info("%s: %p flow buf: drained\n", __func__, wsi);
/* having drained the rxflow buffer, can rearm POLLIN */
#if !defined(LWS_WITH_SERVER)
n =
#endif
__lws_rx_flow_control(wsi);
/* n ignored, needed for NO_SERVER case */
}
/* n = 0 */
return LWS_HPI_RET_HANDLED;
}
int rops_handle_POLLOUT_ws(struct lws *wsi)
{
int write_type = LWS_WRITE_PONG;
#if !defined(LWS_WITHOUT_EXTENSIONS)
2019-05-04 13:19:12 +01:00
struct lws_ext_pm_deflate_rx_ebufs pmdrx;
int ret, m;
#endif
int n;
#if !defined(LWS_WITHOUT_EXTENSIONS)
lwsl_debug("%s: %s: wsi->ws->tx_draining_ext %d\n", __func__,
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
wsi->a.protocol->name, wsi->ws->tx_draining_ext);
#endif
/* Priority 3: pending control packets (pong or close)
*
* 3a: close notification packet requested from close api
*/
if (lwsi_state(wsi) == LRS_WAITING_TO_SEND_CLOSE) {
lwsl_debug("sending close packet\n");
lwsl_hexdump_debug(&wsi->ws->ping_payload_buf[LWS_PRE],
wsi->ws->close_in_ping_buffer_len);
wsi->waiting_to_send_close_frame = 0;
n = lws_write(wsi, &wsi->ws->ping_payload_buf[LWS_PRE],
wsi->ws->close_in_ping_buffer_len,
LWS_WRITE_CLOSE);
if (n >= 0) {
if (wsi->close_needs_ack) {
lwsi_set_state(wsi, LRS_AWAITING_CLOSE_ACK);
2018-11-23 08:47:56 +08:00
lws_set_timeout(wsi, PENDING_TIMEOUT_CLOSE_ACK,
5);
lwsl_debug("sent close, await ack\n");
return LWS_HP_RET_BAIL_OK;
}
wsi->close_needs_ack = 0;
lwsi_set_state(wsi, LRS_RETURNED_CLOSE);
}
return LWS_HP_RET_BAIL_DIE;
}
/* else, the send failed and we should just hang up */
2021-09-13 17:57:08 +03:00
if ((lwsi_role_ws(wsi) && wsi->ws->pong_pending_flag) ||
(lwsi_state(wsi) == LRS_RETURNED_CLOSE &&
wsi->ws->payload_is_close)) {
if (wsi->ws->payload_is_close)
write_type = LWS_WRITE_CLOSE;
else {
if (wsi->wsistate_pre_close) {
/* we started close flow, forget pong */
2021-09-13 17:57:08 +03:00
wsi->ws->pong_pending_flag = 0;
return LWS_HP_RET_BAIL_OK;
}
lwsl_info("issuing pong %d on %s\n",
2021-09-13 17:57:08 +03:00
wsi->ws->pong_payload_len, lws_wsi_tag(wsi));
}
2021-09-13 17:57:08 +03:00
n = lws_write(wsi, &wsi->ws->pong_payload_buf[LWS_PRE],
wsi->ws->pong_payload_len, (enum lws_write_protocol)write_type);
if (n < 0)
return LWS_HP_RET_BAIL_DIE;
/* well he is sent, mark him done */
2021-09-13 17:57:08 +03:00
wsi->ws->pong_pending_flag = 0;
if (wsi->ws->payload_is_close) {
// assert(0);
/* oh... a close frame was it... then we are done */
return LWS_HP_RET_BAIL_DIE;
}
/* otherwise for PING, leave POLLOUT active either way */
return LWS_HP_RET_BAIL_OK;
}
if (!wsi->socket_is_permanently_unusable &&
wsi->ws->send_check_ping) {
lwsl_info("%s: issuing ping on wsi %s: %s %s h2: %d\n", __func__,
lws_wsi_tag(wsi),
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
wsi->role_ops->name, wsi->a.protocol->name,
wsi->mux_substream);
wsi->ws->send_check_ping = 0;
n = lws_write(wsi, &wsi->ws->ping_payload_buf[LWS_PRE],
0, LWS_WRITE_PING);
if (n < 0)
return LWS_HP_RET_BAIL_DIE;
return LWS_HP_RET_BAIL_OK;
}
/* Priority 4: if we are closing, not allowed to send more data frags
* which means user callback or tx ext flush banned now
*/
if (lwsi_state(wsi) == LRS_RETURNED_CLOSE)
return LWS_HP_RET_USER_SERVICE;
#if !defined(LWS_WITHOUT_EXTENSIONS)
/* Priority 5: Tx path extension with more to send
*
* These are handled as new fragments each time around
* So while we must block new writeable callback to enforce
* payload ordering, but since they are always complete
* fragments control packets can interleave OK.
*/
if (wsi->ws->tx_draining_ext) {
lwsl_ext("SERVICING TX EXT DRAINING\n");
if (lws_write(wsi, NULL, 0, LWS_WRITE_CONTINUATION) < 0)
return LWS_HP_RET_BAIL_DIE;
/* leave POLLOUT active */
return LWS_HP_RET_BAIL_OK;
}
/* Priority 6: extensions
*/
if (!wsi->ws->extension_data_pending && !wsi->ws->tx_draining_ext) {
lwsl_ext("%s: !wsi->ws->extension_data_pending\n", __func__);
return LWS_HP_RET_USER_SERVICE;
}
/*
2019-05-04 13:19:12 +01:00
* Check in on the active extensions, see if they had pending stuff to
* spill... they need to get the first look-in otherwise sequence will
* be disordered.
*
2019-05-04 13:19:12 +01:00
* coming here with a NULL, zero-length ebuf means just spill pending
*/
ret = 1;
2020-02-28 15:50:15 +00:00
if (wsi->role_ops == &role_ops_raw_skt
#if defined(LWS_ROLE_RAW_FILE)
|| wsi->role_ops == &role_ops_raw_file
#endif
)
ret = 0;
while (ret == 1) {
/* default to nobody has more to spill */
ret = 0;
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.token = NULL;
pmdrx.eb_in.len = 0;
/* give every extension a chance to spill */
m = lws_ext_cb_active(wsi, LWS_EXT_CB_PACKET_TX_PRESEND,
2019-05-04 13:19:12 +01:00
&pmdrx, 0);
if (m < 0) {
lwsl_err("ext reports fatal error\n");
return LWS_HP_RET_BAIL_DIE;
}
if (m)
/*
* at least one extension told us he has more
* to spill, so we will go around again after
*/
ret = 1;
/* assuming they gave us something to send, send it */
2019-05-04 13:19:12 +01:00
if (pmdrx.eb_in.len) {
n = lws_issue_raw(wsi, (unsigned char *)pmdrx.eb_in.token,
(unsigned int)pmdrx.eb_in.len);
if (n < 0) {
lwsl_info("closing from POLLOUT spill\n");
return LWS_HP_RET_BAIL_DIE;
}
/*
* Keep amount spilled small to minimize chance of this
*/
2019-05-04 13:19:12 +01:00
if (n != pmdrx.eb_in.len) {
lwsl_err("Unable to spill ext %d vs %d\n",
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.len, n);
return LWS_HP_RET_BAIL_DIE;
}
} else
continue;
/* no extension has more to spill */
if (!ret)
continue;
/*
* There's more to spill from an extension, but we just sent
* something... did that leave the pipe choked?
*/
if (!lws_send_pipe_choked(wsi))
/* no we could add more */
continue;
lwsl_info("choked in POLLOUT service\n");
/*
* Yes, he's choked. Leave the POLLOUT masked on so we will
* come back here when he is unchoked. Don't call the user
* callback to enforce ordering of spilling, he'll get called
* when we come back here and there's nothing more to spill.
*/
return LWS_HP_RET_BAIL_OK;
}
wsi->ws->extension_data_pending = 0;
#endif
return LWS_HP_RET_USER_SERVICE;
}
static int
rops_service_flag_pending_ws(struct lws_context *context, int tsi)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
struct lws_context_per_thread *pt = &context->pt[tsi];
struct lws *wsi;
int forced = 0;
/* POLLIN faking (the pt lock is taken by the parent) */
/*
* 1) For all guys with already-available ext data to drain, if they are
* not flowcontrolled, fake their POLLIN status
*/
wsi = pt->ws.rx_draining_ext_list;
while (wsi && wsi->position_in_fds_table != LWS_NO_FDS_POS) {
pt->fds[wsi->position_in_fds_table].revents =
(short)((short)pt->fds[wsi->position_in_fds_table].revents |
(short)(pt->fds[wsi->position_in_fds_table].events & LWS_POLLIN));
if (pt->fds[wsi->position_in_fds_table].revents & LWS_POLLIN)
forced = 1;
wsi = wsi->ws->rx_draining_ext_list;
}
return forced;
#else
return 0;
#endif
}
static int
rops_close_via_role_protocol_ws(struct lws *wsi, enum lws_close_status reason)
{
if (!wsi->ws)
return 0;
if (!wsi->ws->close_in_ping_buffer_len && /* already a reason */
(reason == LWS_CLOSE_STATUS_NOSTATUS ||
reason == LWS_CLOSE_STATUS_NOSTATUS_CONTEXT_DESTROY))
return 0;
lwsl_debug("%s: sending close indication...\n", __func__);
/* if no prepared close reason, use 1000 and no aux data */
if (!wsi->ws->close_in_ping_buffer_len) {
wsi->ws->close_in_ping_buffer_len = 2;
wsi->ws->ping_payload_buf[LWS_PRE] = (reason >> 8) & 0xff;
wsi->ws->ping_payload_buf[LWS_PRE + 1] = reason & 0xff;
}
wsi->waiting_to_send_close_frame = 1;
wsi->close_needs_ack = 1;
lwsi_set_state(wsi, LRS_WAITING_TO_SEND_CLOSE);
__lws_set_timeout(wsi, PENDING_TIMEOUT_CLOSE_SEND, 5);
lws_callback_on_writable(wsi);
return 1;
}
static int
rops_close_role_ws(struct lws_context_per_thread *pt, struct lws *wsi)
{
if (!wsi->ws)
return 0;
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (wsi->ws->rx_draining_ext) {
struct lws **w = &pt->ws.rx_draining_ext_list;
wsi->ws->rx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
*w = wsi->ws->rx_draining_ext_list;
break;
}
w = &((*w)->ws->rx_draining_ext_list);
}
wsi->ws->rx_draining_ext_list = NULL;
}
if (wsi->ws->tx_draining_ext) {
struct lws **w = &pt->ws.tx_draining_ext_list;
lwsl_ext("%s: CLEARING tx_draining_ext\n", __func__);
wsi->ws->tx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
*w = wsi->ws->tx_draining_ext_list;
break;
}
w = &((*w)->ws->tx_draining_ext_list);
}
wsi->ws->tx_draining_ext_list = NULL;
}
#endif
lws_free_set_NULL(wsi->ws->rx_ubuf);
2021-09-13 17:57:08 +03:00
wsi->ws->pong_payload_len = 0;
wsi->ws->pong_pending_flag = 0;
/* deallocate any active extension contexts */
if (lws_ext_cb_active(wsi, LWS_EXT_CB_DESTROY, NULL, 0) < 0)
lwsl_warn("extension destruction failed\n");
return 0;
}
static int
rops_write_role_protocol_ws(struct lws *wsi, unsigned char *buf, size_t len,
enum lws_write_protocol *wp)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
fakewsi: replace with smaller substructure Currently we always reserve a fakewsi per pt so events that don't have a related actual wsi, like vhost-protocol-init or vhost cert init via protocol callback can make callbacks that look reasonable to user protocol handler code expecting a valid wsi every time. This patch splits out stuff that user callbacks often unconditionally expect to be in a wsi, like context pointer, vhost pointer etc into a substructure, which is composed into struct lws at the top of it. Internal references (struct lws is opaque, so there are only internal references) are all updated to go via the substructre, the compiler should make that a NOP. Helpers are added when fakewsi is used and referenced. If not PLAT_FREERTOS, we continue to provide a full fakewsi in the pt as before, although the helpers improve consistency by zeroing down the substructure. There is a huge amount of user code out there over the last 10 years that did not always have the minimal examples to follow, some of it does some unexpected things. If it is PLAT_FREERTOS, that is a newer thing in lws and users have the benefit of being able to follow the minimal examples' approach. For PLAT_FREERTOS we don't reserve the fakewsi in the pt any more, saving around 800 bytes. The helpers then create a struct lws_a (the substructure) on the stack, zero it down (but it is only like 4 pointers) and prepare it with whatever we know like the context. Then we cast it to a struct lws * and use it in the user protocol handler call. In this case, the remainder of the struct lws is undefined. However the amount of old protocol handlers that might touch things outside of the substructure in PLAT_FREERTOS is very limited compared to legacy lws user code and the saving is significant on constrained devices. User handlers should not be touching everything in a wsi every time anyway, there are several cases where there is no valid wsi to do the call with. Dereference of things outside the substructure should only happen when the callback reason shows there is a valid wsi bound to the activity (as in all the minimal examples).
2020-07-19 08:33:46 +01:00
struct lws_context_per_thread *pt = &wsi->a.context->pt[(int)wsi->tsi];
enum lws_write_protocol wpt;
#endif
2019-05-04 13:19:12 +01:00
struct lws_ext_pm_deflate_rx_ebufs pmdrx;
int masked7 = lwsi_role_client(wsi);
unsigned char is_masked_bit = 0;
unsigned char *dropmask = NULL;
size_t orig_len = len;
int pre = 0, n = 0;
// lwsl_err("%s: wp 0x%x len %d\n", __func__, *wp, (int)len);
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (wsi->ws->tx_draining_ext) {
/* remove us from the list */
struct lws **w = &pt->ws.tx_draining_ext_list;
lwsl_ext("%s: CLEARING tx_draining_ext\n", __func__);
wsi->ws->tx_draining_ext = 0;
/* remove us from context draining ext list */
while (*w) {
if (*w == wsi) {
*w = wsi->ws->tx_draining_ext_list;
break;
}
w = &((*w)->ws->tx_draining_ext_list);
}
wsi->ws->tx_draining_ext_list = NULL;
wpt = *wp;
2019-05-04 13:19:12 +01:00
*wp = (wsi->ws->tx_draining_stashed_wp & 0xc0) |
LWS_WRITE_CONTINUATION;
/*
* When we are just flushing (len == 0), we can trust the
* stashed wp info completely. Otherwise adjust it to the
* FIN status of the incoming packet.
*/
if (!(wpt & LWS_WRITE_NO_FIN) && len)
*wp &= (enum lws_write_protocol)~LWS_WRITE_NO_FIN;
2018-11-23 08:47:56 +08:00
lwsl_ext("FORCED draining wp to 0x%02X "
"(stashed 0x%02X, incoming 0x%02X)\n", *wp,
wsi->ws->tx_draining_stashed_wp, wpt);
// assert(0);
}
#endif
if (((*wp) & 0x1f) == LWS_WRITE_HTTP ||
((*wp) & 0x1f) == LWS_WRITE_HTTP_FINAL ||
((*wp) & 0x1f) == LWS_WRITE_HTTP_HEADERS_CONTINUATION ||
((*wp) & 0x1f) == LWS_WRITE_HTTP_HEADERS)
goto send_raw;
/* if we are continuing a frame that already had its header done */
if (wsi->ws->inside_frame) {
lwsl_debug("INSIDE FRAME\n");
goto do_more_inside_frame;
}
wsi->ws->clean_buffer = 1;
/*
* give a chance to the extensions to modify payload
* the extension may decide to produce unlimited payload erratically
* (eg, compression extension), so we require only that if he produces
* something, it will be a complete fragment of the length known at
* the time (just the fragment length known), and if he has
* more we will come back next time he is writeable and allow him to
* produce more fragments until he's drained.
*
* This allows what is sent each time it is writeable to be limited to
* a size that can be sent without partial sends or blocking, allows
* interleaving of control frames and other connection service.
*/
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.token = buf;
2019-05-04 13:19:12 +01:00
pmdrx.eb_in.len = (int)len;
/* for the non-pm-deflate case */
pmdrx.eb_out = pmdrx.eb_in;
switch ((int)*wp) {
case LWS_WRITE_PING:
case LWS_WRITE_PONG:
case LWS_WRITE_CLOSE:
break;
default:
#if !defined(LWS_WITHOUT_EXTENSIONS)
n = lws_ext_cb_active(wsi, (int)LWS_EXT_CB_PAYLOAD_TX, &pmdrx, (int)*wp);
if (n < 0)
return -1;
2019-05-04 13:19:12 +01:00
lwsl_ext("%s: defl ext ret %d, ext in remaining %d, "
"out %d compressed (wp 0x%x)\n", __func__, n,
(int)pmdrx.eb_in.len, (int)pmdrx.eb_out.len, *wp);
if (n == PMDR_HAS_PENDING) {
lwsl_ext("%s: HAS PENDING: write drain len %d "
"(wp 0x%x) SETTING tx_draining_ext "
"(remaining in %d)\n", __func__,
(int)pmdrx.eb_out.len, *wp,
(int)pmdrx.eb_in.len);
/* extension requires further draining */
wsi->ws->tx_draining_ext = 1;
2018-11-23 08:47:56 +08:00
wsi->ws->tx_draining_ext_list =
pt->ws.tx_draining_ext_list;
pt->ws.tx_draining_ext_list = wsi;
/* we must come back to do more */
lws_callback_on_writable(wsi);
/*
* keep a copy of the write type for the overall
* action that has provoked generation of these
* fragments, so the last guy can use its FIN state.
*/
wsi->ws->tx_draining_stashed_wp = (uint8_t)*wp;
2019-05-04 13:19:12 +01:00
/*
* Despite what we may have thought, this is definitely
* NOT the last fragment, because the extension asserted
* he has more coming. For example, the extension may
* be compressing, and has saved up everything until the
* end, where the output is larger than one chunk.
*
* Make sure this intermediate one doesn't actually
* go out with a FIN.
*/
*wp |= LWS_WRITE_NO_FIN;
}
#endif
2019-05-04 13:19:12 +01:00
if (pmdrx.eb_out.len && wsi->ws->stashed_write_pending) {
wsi->ws->stashed_write_pending = 0;
*wp = (unsigned int)(((*wp) & 0xc0) | (unsigned int)wsi->ws->stashed_write_type);
}
}
/*
* an extension did something we need to keep... for example, if
* compression extension, it has already updated its state according
* to this being issued
*/
if (buf != pmdrx.eb_out.token) {
/*
* ext might eat it, but not have anything to issue yet.
* In that case we have to follow his lead, but stash and
* replace the write type that was lost here the first time.
*/
2019-05-04 13:19:12 +01:00
if (len && !pmdrx.eb_out.len) {
if (!wsi->ws->stashed_write_pending)
2018-11-23 08:47:56 +08:00
wsi->ws->stashed_write_type =
(char)(*wp) & 0x3f;
wsi->ws->stashed_write_pending = 1;
return (int)len;
}
/*
* extension recreated it:
* need to buffer this if not all sent
*/
wsi->ws->clean_buffer = 0;
}
buf = pmdrx.eb_out.token;
len = (unsigned int)pmdrx.eb_out.len;
if (!buf) {
lwsl_err("null buf (%d)\n", (int)len);
return -1;
}
switch (wsi->ws->ietf_spec_revision) {
case 13:
if (masked7) {
pre += 4;
dropmask = &buf[0 - pre];
is_masked_bit = 0x80;
}
switch ((*wp) & 0xf) {
case LWS_WRITE_TEXT:
n = LWSWSOPC_TEXT_FRAME;
break;
case LWS_WRITE_BINARY:
n = LWSWSOPC_BINARY_FRAME;
break;
case LWS_WRITE_CONTINUATION:
n = LWSWSOPC_CONTINUATION;
break;
case LWS_WRITE_CLOSE:
n = LWSWSOPC_CLOSE;
break;
case LWS_WRITE_PING:
n = LWSWSOPC_PING;
break;
case LWS_WRITE_PONG:
n = LWSWSOPC_PONG;
break;
default:
lwsl_warn("lws_write: unknown write opc / wp\n");
return -1;
}
if (!((*wp) & LWS_WRITE_NO_FIN))
n |= 1 << 7;
if (len < 126) {
pre += 2;
buf[-pre] = (uint8_t)n;
buf[-pre + 1] = (unsigned char)(len | is_masked_bit);
} else {
if (len < 65536) {
pre += 4;
buf[-pre] = (uint8_t)n;
buf[-pre + 1] = (uint8_t)(126 | is_masked_bit);
buf[-pre + 2] = (unsigned char)(len >> 8);
buf[-pre + 3] = (unsigned char)len;
} else {
pre += 10;
buf[-pre] = (uint8_t)n;
buf[-pre + 1] = (uint8_t)(127 | is_masked_bit);
#if defined __LP64__
buf[-pre + 2] = (len >> 56) & 0x7f;
buf[-pre + 3] = (uint8_t)(len >> 48);
buf[-pre + 4] = (uint8_t)(len >> 40);
buf[-pre + 5] = (uint8_t)(len >> 32);
#else
buf[-pre + 2] = 0;
buf[-pre + 3] = 0;
buf[-pre + 4] = 0;
buf[-pre + 5] = 0;
#endif
buf[-pre + 6] = (unsigned char)(len >> 24);
buf[-pre + 7] = (unsigned char)(len >> 16);
buf[-pre + 8] = (unsigned char)(len >> 8);
buf[-pre + 9] = (unsigned char)len;
}
}
break;
}
do_more_inside_frame:
/*
* Deal with masking if we are in client -> server direction and
* the wp demands it
*/
if (masked7) {
if (!wsi->ws->inside_frame)
if (lws_0405_frame_mask_generate(wsi)) {
lwsl_err("frame mask generation failed\n");
return -1;
}
/*
* in v7, just mask the payload
*/
if (dropmask) { /* never set if already inside frame */
for (n = 4; n < (int)len + 4; n++)
dropmask[n] = dropmask[n] ^ wsi->ws->mask[
(wsi->ws->mask_idx++) & 3];
/* copy the frame nonce into place */
memcpy(dropmask, wsi->ws->mask, 4);
}
}
if (lwsi_role_h2_ENCAPSULATION(wsi)) {
struct lws *encap = lws_get_network_wsi(wsi);
assert(encap != wsi);
return lws_rops_func_fidx(encap->role_ops,
LWS_ROPS_write_role_protocol).
write_role_protocol(wsi, buf - pre,
len + (unsigned int)pre, wp);
}
switch ((*wp) & 0x1f) {
case LWS_WRITE_TEXT:
case LWS_WRITE_BINARY:
case LWS_WRITE_CONTINUATION:
if (!wsi->h2_stream_carries_ws) {
/*
* give any active extensions a chance to munge the
* buffer before send. We pass in a pointer to an
* lws_tokens struct prepared with the default buffer
* and content length that's in there. Rather than
* rewrite the default buffer, extensions that expect
* to grow the buffer can adapt .token to point to their
* own per-connection buffer in the extension user
* allocation. By default with no extensions or no
* extension callback handling, just the normal input
* buffer is used then so it is efficient.
*
* callback returns 1 in case it wants to spill more
* buffers
*
* This takes care of holding the buffer if send is
* incomplete, ie, if wsi->ws->clean_buffer is 0
* (meaning an extension meddled with the buffer). If
* wsi->ws->clean_buffer is 1, it will instead return
* to the user code how much OF THE USER BUFFER was
* consumed.
*/
n = lws_issue_raw_ext_access(wsi, buf - pre, len + (unsigned int)pre);
wsi->ws->inside_frame = 1;
if (n <= 0)
return n;
if (n == (int)len + pre) {
/* everything in the buffer was handled
* (or rebuffered...) */
wsi->ws->inside_frame = 0;
return (int)orig_len;
}
/*
* it is how many bytes of user buffer got sent... may
* be < orig_len in which case callback when writable
* has already been arranged and user code can call
* lws_write() again with the rest later.
*/
return n - pre;
}
break;
default:
break;
}
send_raw:
return lws_issue_raw(wsi, (unsigned char *)buf - pre, len + (unsigned int)pre);
}
static int
rops_close_kill_connection_ws(struct lws *wsi, enum lws_close_status reason)
{
/* deal with ws encapsulation in h2 */
#if defined(LWS_WITH_HTTP2)
if (wsi->mux_substream && wsi->h2_stream_carries_ws)
return lws_rops_func_fidx(&role_ops_h2,
LWS_ROPS_close_kill_connection).
close_kill_connection(wsi, reason);
return 0;
#else
return 0;
#endif
}
static int
rops_callback_on_writable_ws(struct lws *wsi)
{
#if defined(LWS_WITH_HTTP2)
if (lwsi_role_h2_ENCAPSULATION(wsi)) {
/* we know then that it has an h2 parent */
struct lws *enc = lws_rops_func_fidx(&role_ops_h2,
LWS_ROPS_encapsulation_parent).
encapsulation_parent(wsi);
assert(enc);
if (lws_rops_func_fidx(enc->role_ops,
LWS_ROPS_callback_on_writable).
callback_on_writable(wsi))
return 1;
}
#endif
return 0;
}
static int
rops_init_vhost_ws(struct lws_vhost *vh,
const struct lws_context_creation_info *info)
{
#if !defined(LWS_WITHOUT_EXTENSIONS)
#ifdef LWS_WITH_PLUGINS
2020-01-02 08:32:23 +00:00
struct lws_plugin *plugin;
int m;
if (vh->context->plugin_extension_count) {
m = 0;
while (info->extensions && info->extensions[m].callback)
m++;
/*
* give the vhost a unified list of extensions including the
* ones that came from plugins
*/
vh->ws.extensions = lws_zalloc(sizeof(struct lws_extension) *
(unsigned int)(m + vh->context->plugin_extension_count + 1),
"extensions");
if (!vh->ws.extensions)
return 1;
memcpy((struct lws_extension *)vh->ws.extensions, info->extensions,
sizeof(struct lws_extension) * (unsigned int)m);
plugin = vh->context->plugin_list;
while (plugin) {
const lws_plugin_protocol_t *plpr =
(const lws_plugin_protocol_t *)plugin->hdr;
memcpy((struct lws_extension *)&vh->ws.extensions[m],
plpr->extensions,
sizeof(struct lws_extension) *
(unsigned int)plpr->count_extensions);
m += plpr->count_extensions;
plugin = plugin->list;
}
} else
#endif
vh->ws.extensions = info->extensions;
#endif
return 0;
}
static int
rops_destroy_vhost_ws(struct lws_vhost *vh)
{
#ifdef LWS_WITH_PLUGINS
#if !defined(LWS_WITHOUT_EXTENSIONS)
if (vh->context->plugin_extension_count)
lws_free((void *)vh->ws.extensions);
#endif
#endif
return 0;
}
#if defined(LWS_WITH_HTTP_PROXY)
static int
ws_destroy_proxy_buf(struct lws_dll2 *d, void *user)
{
lws_free(d);
return 0;
}
#endif
static int
rops_destroy_role_ws(struct lws *wsi)
{
#if defined(LWS_WITH_HTTP_PROXY)
lws_dll2_foreach_safe(&wsi->ws->proxy_owner, NULL, ws_destroy_proxy_buf);
#endif
lws_free_set_NULL(wsi->ws);
return 0;
}
static int
rops_issue_keepalive_ws(struct lws *wsi, int isvalid)
{
uint64_t us;
#if defined(LWS_WITH_HTTP2)
if (lwsi_role_h2_ENCAPSULATION(wsi)) {
/* we know then that it has an h2 parent */
struct lws *enc = lws_rops_func_fidx(&role_ops_h2,
LWS_ROPS_encapsulation_parent).
encapsulation_parent(wsi);
assert(enc);
if (lws_rops_func_fidx(enc->role_ops, LWS_ROPS_issue_keepalive).
issue_keepalive(enc, isvalid))
return 1;
}
#endif
if (isvalid)
_lws_validity_confirmed_role(wsi);
else {
us = (uint64_t)lws_now_usecs();
memcpy(&wsi->ws->ping_payload_buf[LWS_PRE], &us, 8);
wsi->ws->send_check_ping = 1;
lws_callback_on_writable(wsi);
}
return 0;
}
static const lws_rops_t rops_table_ws[] = {
/* 1 */ { .init_vhost = rops_init_vhost_ws },
/* 2 */ { .destroy_vhost = rops_destroy_vhost_ws },
/* 3 */ { .service_flag_pending = rops_service_flag_pending_ws },
/* 4 */ { .handle_POLLIN = rops_handle_POLLIN_ws },
/* 5 */ { .handle_POLLOUT = rops_handle_POLLOUT_ws },
/* 6 */ { .callback_on_writable = rops_callback_on_writable_ws },
/* 7 */ { .write_role_protocol = rops_write_role_protocol_ws },
/* 8 */ { .close_via_role_protocol = rops_close_via_role_protocol_ws },
/* 9 */ { .close_role = rops_close_role_ws },
/* 10 */ { .close_kill_connection = rops_close_kill_connection_ws },
/* 11 */ { .destroy_role = rops_destroy_role_ws },
/* 12 */ { .issue_keepalive = rops_issue_keepalive_ws },
};
const struct lws_role_ops role_ops_ws = {
/* role name */ "ws",
/* alpn id */ NULL,
/* rops_table */ rops_table_ws,
/* rops_idx */ {
/* LWS_ROPS_check_upgrades */
/* LWS_ROPS_pt_init_destroy */ 0x00,
/* LWS_ROPS_init_vhost */
/* LWS_ROPS_destroy_vhost */ 0x12,
/* LWS_ROPS_service_flag_pending */
/* LWS_ROPS_handle_POLLIN */ 0x34,
/* LWS_ROPS_handle_POLLOUT */
/* LWS_ROPS_perform_user_POLLOUT */ 0x50,
/* LWS_ROPS_callback_on_writable */
/* LWS_ROPS_tx_credit */ 0x60,
/* LWS_ROPS_write_role_protocol */
/* LWS_ROPS_encapsulation_parent */ 0x70,
/* LWS_ROPS_alpn_negotiated */
/* LWS_ROPS_close_via_role_protocol */ 0x08,
/* LWS_ROPS_close_role */
/* LWS_ROPS_close_kill_connection */ 0x9a,
/* LWS_ROPS_destroy_role */
/* LWS_ROPS_adoption_bind */ 0xb0,
/* LWS_ROPS_client_bind */
/* LWS_ROPS_issue_keepalive */ 0x0c,
},
2018-11-29 08:29:48 +08:00
/* adoption_cb clnt, srv */ { LWS_CALLBACK_SERVER_NEW_CLIENT_INSTANTIATED,
LWS_CALLBACK_SERVER_NEW_CLIENT_INSTANTIATED },
/* rx_cb clnt, srv */ { LWS_CALLBACK_CLIENT_RECEIVE,
LWS_CALLBACK_RECEIVE },
/* writeable cb clnt, srv */ { LWS_CALLBACK_CLIENT_WRITEABLE,
LWS_CALLBACK_SERVER_WRITEABLE },
/* close cb clnt, srv */ { LWS_CALLBACK_CLIENT_CLOSED,
LWS_CALLBACK_CLOSED },
/* protocol_bind cb c, srv */ { LWS_CALLBACK_WS_CLIENT_BIND_PROTOCOL,
LWS_CALLBACK_WS_SERVER_BIND_PROTOCOL },
/* protocol_unbind cb c, srv */ { LWS_CALLBACK_WS_CLIENT_DROP_PROTOCOL,
LWS_CALLBACK_WS_SERVER_DROP_PROTOCOL },
/* file handles */ 0
};