LNXCORE/v8/include/libwebsockets/lws-misc.h

/*
 * libwebsockets - small server side websockets and web server implementation
 *
 * Copyright (C) 2010 - 2021 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.
 */

#if defined(LWS_WITH_SPAWN)

#if defined(WIN32) || defined(_WIN32)
#else
#include <sys/wait.h>
#include <sys/times.h>
#endif
#endif

#if defined(__OpenBSD__)
#include <sys/siginfo.h>
#endif

/** \defgroup misc Miscellaneous APIs
* ##Miscellaneous APIs
*
* Various APIs outside of other categories
*/
///@{

struct lws_buflist;

/**
 * lws_buflist_append_segment(): add buffer to buflist at head
 *
 * \param head: list head
 * \param buf: buffer to stash
 * \param len: length of buffer to stash
 *
 * Returns -1 on OOM, 1 if this was the first segment on the list, and 0 if
 * it was a subsequent segment.
 */
LWS_VISIBLE LWS_EXTERN int LWS_WARN_UNUSED_RESULT
lws_buflist_append_segment(struct lws_buflist **head, const uint8_t *buf,
			   size_t len);
/**
 * lws_buflist_next_segment_len(): number of bytes left in current segment
 *
 * \param head: list head
 * \param buf: if non-NULL, *buf is written with the address of the start of
 *		the remaining data in the segment
 *
 * Returns the number of bytes left in the current segment.  0 indicates
 * that the buflist is empty (there are no segments on the buflist).
 */
LWS_VISIBLE LWS_EXTERN size_t
lws_buflist_next_segment_len(struct lws_buflist **head, uint8_t **buf);

/**
 * lws_buflist_use_segment(): remove len bytes from the current segment
 *
 * \param head: list head
 * \param len: number of bytes to mark as used
 *
 * If len is less than the remaining length of the current segment, the position
 * in the current segment is simply advanced and it returns.
 *
 * If len uses up the remaining length of the current segment, then the segment
 * is deleted and the list head moves to the next segment if any.
 *
 * Returns the number of bytes left in the current segment.  0 indicates
 * that the buflist is empty (there are no segments on the buflist).
 */
LWS_VISIBLE LWS_EXTERN size_t
lws_buflist_use_segment(struct lws_buflist **head, size_t len);

/**
 * lws_buflist_total_len(): Get the total size of the buflist
 *
 * \param head: list head
 *
 * Returns the total number of bytes held on all segments of the buflist
 */
LWS_VISIBLE LWS_EXTERN size_t
lws_buflist_total_len(struct lws_buflist **head);

/**
 * lws_buflist_linear_copy(): copy everything out as one without consuming
 *
 * \param head: list head
 * \param ofs: start offset into buflist in bytes
 * \param buf: buffer to copy linearly into
 * \param len: length of buffer available
 *
 * Returns -1 if len is too small, or bytes copied.  Happy to do partial
 * copies, returns 0 when there are no more bytes to copy.
 */
LWS_VISIBLE LWS_EXTERN int
lws_buflist_linear_copy(struct lws_buflist **head, size_t ofs, uint8_t *buf,
			size_t len);

/**
 * lws_buflist_linear_use(): copy and consume from buflist head
 *
 * \param head: list head
 * \param buf: buffer to copy linearly into
 * \param len: length of buffer available
 *
 * Copies a possibly fragmented buflist from the head into the linear output
 * buffer \p buf for up to length \p len, and consumes the buflist content that
 * was copied out.
 *
 * Since it was consumed, calling again will resume copying out and consuming
 * from as far as it got the first time.
 *
 * Returns the number of bytes written into \p buf.
 */
LWS_VISIBLE LWS_EXTERN int
lws_buflist_linear_use(struct lws_buflist **head, uint8_t *buf, size_t len);

/**
 * lws_buflist_fragment_use(): copy and consume <= 1 frag from buflist head
 *
 * \param head: list head
 * \param buf: buffer to copy linearly into
 * \param len: length of buffer available
 * \param frag_first: pointer to char written on exit to if this is start of frag
 * \param frag_fin: pointer to char written on exit to if this is end of frag
 *
 * Copies all or part of the fragment at the start of a buflist from the head
 * into the output buffer \p buf for up to length \p len, and consumes the
 * buflist content that was copied out.
 *
 * Since it was consumed, calling again will resume copying out and consuming
 * from as far as it got the first time.
 *
 * Returns the number of bytes written into \p buf.
 */
LWS_VISIBLE LWS_EXTERN int
lws_buflist_fragment_use(struct lws_buflist **head, uint8_t *buf,
			 size_t len, char *frag_first, char *frag_fin);

/**
 * lws_buflist_destroy_all_segments(): free all segments on the list
 *
 * \param head: list head
 *
 * This frees everything on the list unconditionally.  *head is always
 * NULL after this.
 */
LWS_VISIBLE LWS_EXTERN void
lws_buflist_destroy_all_segments(struct lws_buflist **head);

/**
 * lws_buflist_describe(): debug helper logging buflist status
 *
 * \param head: list head
 * \param id: pointer shown in debug list
 * \param reason: reason string show in debug list
 *
 * Iterates through the buflist segments showing position and size.
 * This only exists when lws was built in debug mode
 */
LWS_VISIBLE LWS_EXTERN void
lws_buflist_describe(struct lws_buflist **head, void *id, const char *reason);


/*
 * Optional helpers for closely-managed stream flow control.  These are useful
 * when there is no memory for large rx buffers and instead tx credit is being
 * used to regulate the server sending data.
 *
 * When combined with stateful consumption-on-demand, this can be very effective
 * at managing data flows through restricted circumstances.  These helpers
 * implement a golden implementation that can be bound to a stream in its priv
 * data.
 *
 * The helper is sophisticated enough to contain a buflist to manage overflows
 * on heap and preferentially drain it.  RX goes through heap to guarantee the
 * consumer can exit cleanly at any time.
 */

enum {
	LWSDLOFLOW_STATE_READ, /* default, we want input */
	LWSDLOFLOW_STATE_READ_COMPLETED, /* we do not need further rx, every-
					  * thing is locally buffered or used */
	LWSDLOFLOW_STATE_READ_FAILED, /* operation has fatal error */
};

struct lws_ss_handle;

typedef struct lws_flow {
	lws_dll2_t			list;

	struct lws_ss_handle		*h;
	struct lws_buflist		*bl;

	const uint8_t			*data;
	size_t				len;		/* bytes left in data */
	uint32_t			blseglen;	/* bytes issued */
	int32_t				window;

	uint8_t				state;
} lws_flow_t;

/**
 * lws_flow_feed() - consume waiting data if ready for it
 *
 * \param flow: pointer to the flow struct managing waiting data
 *
 * This will bring out waiting data from the flow buflist when it is needed.
 */
LWS_VISIBLE LWS_EXTERN lws_stateful_ret_t
lws_flow_feed(lws_flow_t *flow);

/**
 * lws_flow_req() - request remote data if we have run low
 *
 * \param flow: pointer to the flow struct managing waiting data
 *
 * When the estimated remote tx credit is below flow->window, accounting for
 * what is in the buflist, add to the peer tx credit so it can send us more.
 */
LWS_VISIBLE LWS_EXTERN lws_stateful_ret_t
lws_flow_req(lws_flow_t *flow);

/**
 * lws_ptr_diff(): helper to report distance between pointers as an int
 *
 * \param head: the pointer with the larger address
 * \param tail: the pointer with the smaller address
 *
 * This helper gives you an int representing the number of bytes further
 * forward the first pointer is compared to the second pointer.
 */
#define lws_ptr_diff(head, tail) \
			((int)((char *)(head) - (char *)(tail)))

#define lws_ptr_diff_size_t(head, tail) \
			((size_t)(ssize_t)((char *)(head) - (char *)(tail)))

/**
 * lws_snprintf(): snprintf that truncates the returned length too
 *
 * \param str: destination buffer
 * \param size: bytes left in destination buffer
 * \param format: format string
 * \param ...: args for format
 *
 * This lets you correctly truncate buffers by concatenating lengths, if you
 * reach the limit the reported length doesn't exceed the limit.
 */
LWS_VISIBLE LWS_EXTERN int
lws_snprintf(char *str, size_t size, const char *format, ...) LWS_FORMAT(3);

/**
 * lws_strncpy(): strncpy that guarantees NUL on truncated copy
 *
 * \param dest: destination buffer
 * \param src: source buffer
 * \param size: bytes left in destination buffer
 *
 * This lets you correctly truncate buffers by concatenating lengths, if you
 * reach the limit the reported length doesn't exceed the limit.
 */
LWS_VISIBLE LWS_EXTERN char *
lws_strncpy(char *dest, const char *src, size_t size);

/*
 * Variation where we want to use the smaller of two lengths, useful when the
 * source string is not NUL terminated
 */
#define lws_strnncpy(dest, src, size1, destsize) \
	lws_strncpy(dest, src, (size_t)(size1 + 1) < (size_t)(destsize) ? \
				(size_t)(size1 + 1) : (size_t)(destsize))

/**
 * lws_nstrstr(): like strstr for length-based strings without terminating NUL
 *
 * \param buf: the string to search
 * \param len: the length of the string to search
 * \param name: the substring to search for
 * \param nl: the length of name
 *
 * Returns NULL if \p name is not present in \p buf.  Otherwise returns the
 * address of the first instance of \p name in \p buf.
 *
 * Neither buf nor name need to be NUL-terminated.
 */
LWS_VISIBLE LWS_EXTERN const char *
lws_nstrstr(const char *buf, size_t len, const char *name, size_t nl);

/**
 * lws_json_simple_find(): dumb JSON string parser
 *
 * \param buf: the JSON to search
 * \param len: the length of the JSON to search
 * \param name: the name field to search the JSON for, eg, "\"myname\":"
 * \param alen: set to the length of the argument part if non-NULL return
 *
 * Either returns NULL if \p name is not present in buf, or returns a pointer
 * to the argument body of the first instance of \p name, and sets *alen to the
 * length of the argument body.
 *
 * This can cheaply handle fishing out, eg, myarg from {"myname": "myarg"} by
 * searching for "\"myname\":".  It will return a pointer to myarg and set *alen
 * to 5.  It equally handles args like "myname": true, or "myname":false, and
 * null or numbers are all returned as delimited strings.
 *
 * Anything more complicated like the value is a subobject or array, you should
 * parse it using a full parser like lejp.  This is suitable is the JSON is
 * and will remain short and simple, and contains well-known names amongst other
 * extensible JSON members.
 */
LWS_VISIBLE LWS_EXTERN const char *
lws_json_simple_find(const char *buf, size_t len, const char *name, size_t *alen);

/**
 * lws_json_simple_strcmp(): dumb JSON string comparison
 *
 * \param buf: the JSON to search
 * \param len: the length of the JSON to search
 * \param name: the name field to search the JSON for, eg, "\"myname\":"
 * \param comp: return a strcmp of this and the discovered argument
 *
 * Helper that combines lws_json_simple_find() with strcmp() if it was found.
 * If the \p name was not found, returns -1.  Otherwise returns a strcmp()
 * between what was found and \p comp, ie, return 0 if they match or something
 * else if they don't.
 *
 * If the JSON is relatively simple and you want to target constrained
 * devices, this can be a good choice.  If the JSON may be complex, you
 * should use a full JSON parser.
 */
LWS_VISIBLE LWS_EXTERN int
lws_json_simple_strcmp(const char *buf, size_t len, const char *name, const char *comp);

/**
 * lws_hex_len_to_byte_array(): convert hex string like 0123456789ab into byte data
 *
 * \param h: incoming hex string
 * \param hlen: number of chars to process at \p h
 * \param dest: array to fill with binary decodes of hex pairs from h
 * \param max: maximum number of bytes dest can hold, must be at least half
 *		the size of strlen(h)
 *
 * This converts hex strings into an array of 8-bit representations, ie the
 * input "abcd" produces two bytes of value 0xab and 0xcd.
 *
 * Returns number of bytes produced into \p dest, or -1 on error.
 *
 * Errors include non-hex chars and an odd count of hex chars in the input
 * string.
 */
LWS_VISIBLE LWS_EXTERN int
lws_hex_len_to_byte_array(const char *h, size_t hlen, uint8_t *dest, int max);

/**
 * lws_hex_to_byte_array(): convert hex string like 0123456789ab into byte data
 *
 * \param h: incoming NUL-terminated hex string
 * \param dest: array to fill with binary decodes of hex pairs from h
 * \param max: maximum number of bytes dest can hold, must be at least half
 *		the size of strlen(h)
 *
 * This converts hex strings into an array of 8-bit representations, ie the
 * input "abcd" produces two bytes of value 0xab and 0xcd.
 *
 * Returns number of bytes produced into \p dest, or -1 on error.
 *
 * Errors include non-hex chars and an odd count of hex chars in the input
 * string.
 */
LWS_VISIBLE LWS_EXTERN int
lws_hex_to_byte_array(const char *h, uint8_t *dest, int max);

/**
 * lws_hex_from_byte_array(): render byte array as hex char string
 *
 * \param src: incoming binary source array
 * \param slen: length of src in bytes
 * \param dest: array to fill with hex chars representing src
 * \param len: max extent of dest
 *
 * This converts binary data of length slen at src, into a hex string at dest
 * of maximum length len.  Even if truncated, the result will be NUL-terminated.
 */
LWS_VISIBLE LWS_EXTERN void
lws_hex_from_byte_array(const uint8_t *src, size_t slen, char *dest, size_t len);

/**
 * lws_hex_random(): generate len - 1 or - 2 characters of random ascii hex
 *
 * \param context: the lws_context used to get the random
 * \param dest: destination for hex ascii chars
 * \param len: the number of bytes the buffer dest points to can hold
 *
 * This creates random ascii-hex strings up to a given length, with a
 * terminating NUL.
 *
 * There will not be any characters produced that are not 0-9, a-f, so it's
 * safe to go straight into, eg, JSON.
 */
LWS_VISIBLE LWS_EXTERN int
lws_hex_random(struct lws_context *context, char *dest, size_t len);

/*
 * lws_timingsafe_bcmp(): constant time memcmp
 *
 * \param a: first buffer
 * \param b: second buffer
 * \param len: count of bytes to compare
 *
 * Return 0 if the two buffers are the same, else nonzero.
 *
 * Always compares all of the buffer before returning, so it can't be used as
 * a timing oracle.
 */

LWS_VISIBLE LWS_EXTERN int
lws_timingsafe_bcmp(const void *a, const void *b, uint32_t len);

/**
 * lws_get_random(): fill a buffer with platform random data
 *
 * \param context: the lws context
 * \param buf: buffer to fill
 * \param len: how much to fill
 *
 * Fills buf with len bytes of random.  Returns the number of bytes set, if
 * not equal to len, then getting the random failed.
 */
LWS_VISIBLE LWS_EXTERN size_t
lws_get_random(struct lws_context *context, void *buf, size_t len);
/**
 * lws_daemonize(): make current process run in the background
 *
 * \param _lock_path: the filepath to write the lock file
 *
 * Spawn lws as a background process, taking care of various things
 */
LWS_VISIBLE LWS_EXTERN int LWS_WARN_UNUSED_RESULT
lws_daemonize(const char *_lock_path);
/**
 * lws_get_library_version(): return string describing the version of lws
 *
 * On unix, also includes the git describe
 */
LWS_VISIBLE LWS_EXTERN const char * LWS_WARN_UNUSED_RESULT
lws_get_library_version(void);

/**
 * lws_wsi_user() - get the user data associated with the connection
 * \param wsi: lws connection
 *
 * Not normally needed since it's passed into the callback
 */
LWS_VISIBLE LWS_EXTERN void *
lws_wsi_user(struct lws *wsi);

/**
 * lws_wsi_tsi() - get the service thread index the wsi is bound to
 * \param wsi: lws connection
 *
 * Only useful is LWS_MAX_SMP > 1
 */
LWS_VISIBLE LWS_EXTERN int
lws_wsi_tsi(struct lws *wsi);

/**
 * lws_set_wsi_user() - set the user data associated with the client connection
 * \param wsi: lws connection
 * \param user: user data
 *
 * By default lws allocates this and it's not legal to externally set it
 * yourself.  However client connections may have it set externally when the
 * connection is created... if so, this api can be used to modify it at
 * runtime additionally.
 */
LWS_VISIBLE LWS_EXTERN void
lws_set_wsi_user(struct lws *wsi, void *user);

/**
 * lws_parse_uri:	cut up prot:/ads:port/path into pieces
 *			Notice it does so by dropping '\0' into input string
 *			and the leading / on the path is consequently lost
 *
 * \param p:			incoming uri string.. will get written to
 * \param prot:		result pointer for protocol part (https://)
 * \param ads:		result pointer for address part
 * \param port:		result pointer for port part
 * \param path:		result pointer for path part
 *
 * You may also refer to unix socket addresses, using a '+' at the start of
 * the address.  In this case, the address should end with ':', which is
 * treated as the separator between the address and path (the normal separator
 * '/' is a valid part of the socket path).  Eg,
 *
 * http://+/var/run/mysocket:/my/path
 *
 * If the first character after the + is '@', it's interpreted by lws client
 * processing as meaning to use linux abstract namespace sockets, the @ is
 * replaced with a '\0' before use.
 */
LWS_VISIBLE LWS_EXTERN int LWS_WARN_UNUSED_RESULT
lws_parse_uri(char *p, const char **prot, const char **ads, int *port,
	      const char **path);
/**
 * lws_cmdline_option():	simple commandline parser
 *
 * \param argc:		count of argument strings
 * \param argv:		argument strings
 * \param val:		string to find
 *
 * Returns NULL if the string \p val is not found in the arguments.
 *
 * If it is found, then it returns a pointer to the next character after \p val.
 * So if \p val is "-d", then for the commandlines "myapp -d15" and
 * "myapp -d 15", in both cases the return will point to the "15".
 *
 * In the case there is no argument, like "myapp -d", the return will
 * either point to the '\\0' at the end of -d, or to the start of the
 * next argument, ie, will be non-NULL.
 */
LWS_VISIBLE LWS_EXTERN const char *
lws_cmdline_option(int argc, const char **argv, const char *val);

/**
 * lws_cmdline_option_handle_builtin(): apply standard cmdline options
 *
 * \param argc:		count of argument strings
 * \param argv:		argument strings
 * \param info:		context creation info
 *
 * Applies standard options to the context creation info to save them having
 * to be (unevenly) copied into the minimal examples.
 *
 * Applies default log levels that can be overriden by -d
 */
LWS_VISIBLE LWS_EXTERN void
lws_cmdline_option_handle_builtin(int argc, const char **argv,
				  struct lws_context_creation_info *info);

/**
 * lws_now_secs(): return seconds since 1970-1-1
 */
LWS_VISIBLE LWS_EXTERN unsigned long
lws_now_secs(void);

/**
 * lws_now_usecs(): return useconds since 1970-1-1
 */
LWS_VISIBLE LWS_EXTERN lws_usec_t
lws_now_usecs(void);

/**
 * lws_get_context - Allow getting lws_context from a Websocket connection
 * instance
 *
 * With this function, users can access context in the callback function.
 * Otherwise users may have to declare context as a global variable.
 *
 * \param wsi:	Websocket connection instance
 */
LWS_VISIBLE LWS_EXTERN struct lws_context * LWS_WARN_UNUSED_RESULT
lws_get_context(const struct lws *wsi);

/**
 * lws_get_vhost_listen_port - Find out the port number a vhost is listening on
 *
 * In the case you passed 0 for the port number at context creation time, you
 * can discover the port number that was actually chosen for the vhost using
 * this api.
 *
 * \param vhost:	Vhost to get listen port from
 */
LWS_VISIBLE LWS_EXTERN int LWS_WARN_UNUSED_RESULT
lws_get_vhost_listen_port(struct lws_vhost *vhost);

/**
 * lws_get_count_threads(): how many service threads the context uses
 *
 * \param context: the lws context
 *
 * By default this is always 1, if you asked for more than lws can handle it
 * will clip the number of threads.  So you can use this to find out how many
 * threads are actually in use.
 */
LWS_VISIBLE LWS_EXTERN int LWS_WARN_UNUSED_RESULT
lws_get_count_threads(struct lws_context *context);

/**
 * lws_get_parent() - get parent wsi or NULL
 * \param wsi: lws connection
 *
 * Specialized wsi like cgi stdin/out/err are associated to a parent wsi,
 * this allows you to get their parent.
 */
LWS_VISIBLE LWS_EXTERN struct lws * LWS_WARN_UNUSED_RESULT
lws_get_parent(const struct lws *wsi);

/**
 * lws_get_child() - get child wsi or NULL
 * \param wsi: lws connection
 *
 * Allows you to find a related wsi from the parent wsi.
 */
LWS_VISIBLE LWS_EXTERN struct lws * LWS_WARN_UNUSED_RESULT
lws_get_child(const struct lws *wsi);

/**
 * lws_get_effective_uid_gid() - find out eventual uid and gid while still root
 *
 * \param context: lws context
 * \param uid: pointer to uid result
 * \param gid: pointer to gid result
 *
 * This helper allows you to find out what the uid and gid for the process will
 * be set to after the privileges are dropped, beforehand.  So while still root,
 * eg in LWS_CALLBACK_PROTOCOL_INIT, you can arrange things like cache dir
 * and subdir creation / permissions down /var/cache dynamically.
 */
LWS_VISIBLE LWS_EXTERN void
lws_get_effective_uid_gid(struct lws_context *context, uid_t *uid, gid_t *gid);

/**
 * lws_get_udp() - get wsi's udp struct
 *
 * \param wsi: lws connection
 *
 * Returns NULL or pointer to the wsi's UDP-specific information
 */
LWS_VISIBLE LWS_EXTERN const struct lws_udp * LWS_WARN_UNUSED_RESULT
lws_get_udp(const struct lws *wsi);

LWS_VISIBLE LWS_EXTERN void *
lws_get_opaque_parent_data(const struct lws *wsi);

LWS_VISIBLE LWS_EXTERN void
lws_set_opaque_parent_data(struct lws *wsi, void *data);

LWS_VISIBLE LWS_EXTERN void *
lws_get_opaque_user_data(const struct lws *wsi);

LWS_VISIBLE LWS_EXTERN void
lws_set_opaque_user_data(struct lws *wsi, void *data);

LWS_VISIBLE LWS_EXTERN int
lws_get_child_pending_on_writable(const struct lws *wsi);

LWS_VISIBLE LWS_EXTERN void
lws_clear_child_pending_on_writable(struct lws *wsi);

LWS_VISIBLE LWS_EXTERN int
lws_get_close_length(struct lws *wsi);

LWS_VISIBLE LWS_EXTERN unsigned char *
lws_get_close_payload(struct lws *wsi);

/**
 * lws_get_network_wsi() - Returns wsi that has the tcp connection for this wsi
 *
 * \param wsi: wsi you have
 *
 * Returns wsi that has the tcp connection (which may be the incoming wsi)
 *
 * HTTP/1 connections will always return the incoming wsi
 * HTTP/2 connections may return a different wsi that has the tcp connection
 */
LWS_VISIBLE LWS_EXTERN
struct lws *lws_get_network_wsi(struct lws *wsi);

/**
 * lws_set_allocator() - custom allocator support
 *
 * \param realloc
 *
 * Allows you to replace the allocator (and deallocator) used by lws
 */
LWS_VISIBLE LWS_EXTERN void
lws_set_allocator(void *(*realloc)(void *ptr, size_t size, const char *reason));

enum {
	/*
	 * Flags for enable and disable rxflow with reason bitmap and with
	 * backwards-compatible single bool
	 */
	LWS_RXFLOW_REASON_USER_BOOL		= (1 << 0),
	LWS_RXFLOW_REASON_HTTP_RXBUFFER		= (1 << 6),
	LWS_RXFLOW_REASON_H2_PPS_PENDING	= (1 << 7),

	LWS_RXFLOW_REASON_APPLIES		= (1 << 14),
	LWS_RXFLOW_REASON_APPLIES_ENABLE_BIT	= (1 << 13),
	LWS_RXFLOW_REASON_APPLIES_ENABLE	= LWS_RXFLOW_REASON_APPLIES |
						  LWS_RXFLOW_REASON_APPLIES_ENABLE_BIT,
	LWS_RXFLOW_REASON_APPLIES_DISABLE	= LWS_RXFLOW_REASON_APPLIES,
	LWS_RXFLOW_REASON_FLAG_PROCESS_NOW	= (1 << 12),

};

/**
 * lws_rx_flow_control() - Enable and disable socket servicing for
 *				received packets.
 *
 * If the output side of a server process becomes choked, this allows flow
 * control for the input side.
 *
 * \param wsi:	Websocket connection instance to get callback for
 * \param enable:	0 = disable read servicing for this connection, 1 = enable
 *
 * If you need more than one additive reason for rxflow control, you can give
 * iLWS_RXFLOW_REASON_APPLIES_ENABLE or _DISABLE together with one or more of
 * b5..b0 set to idicate which bits to enable or disable.  If any bits are
 * enabled, rx on the connection is suppressed.
 *
 * LWS_RXFLOW_REASON_FLAG_PROCESS_NOW  flag may also be given to force any change
 * in rxflowbstatus to benapplied immediately, this should be used when you are
 * changing a wsi flow control state from outside a callback on that wsi.
 */
LWS_VISIBLE LWS_EXTERN int
lws_rx_flow_control(struct lws *wsi, int enable);

/**
 * lws_rx_flow_allow_all_protocol() - Allow all connections with this protocol to receive
 *
 * When the user server code realizes it can accept more input, it can
 * call this to have the RX flow restriction removed from all connections using
 * the given protocol.
 * \param context:	lws_context
 * \param protocol:	all connections using this protocol will be allowed to receive
 */
LWS_VISIBLE LWS_EXTERN void
lws_rx_flow_allow_all_protocol(const struct lws_context *context,
			       const struct lws_protocols *protocol);

/**
 * lws_remaining_packet_payload() - Bytes to come before "overall"
 *					      rx fragment is complete
 * \param wsi:		Websocket instance (available from user callback)
 *
 * This tracks how many bytes are left in the current ws fragment, according
 * to the ws length given in the fragment header.
 *
 * If the message was in a single fragment, and there is no compression, this
 * is the same as "how much data is left to read for this message".
 *
 * However, if the message is being sent in multiple fragments, this will
 * reflect the unread amount of the current **fragment**, not the message.  With
 * ws, it is legal to not know the length of the message before it completes.
 *
 * Additionally if the message is sent via the negotiated permessage-deflate
 * extension, zero is always reported.  You should use lws_is_final_fragment()
 * to find out if you have completed the message... in compressed case, it holds
 * back reporting the final fragment until it's also the final decompressed
 * block issued.
 */
LWS_VISIBLE LWS_EXTERN size_t
lws_remaining_packet_payload(struct lws *wsi);

#if defined(LWS_WITH_DIR)

typedef enum {
	LDOT_UNKNOWN,
	LDOT_FILE,
	LDOT_DIR,
	LDOT_LINK,
	LDOT_FIFO,
	LDOTT_SOCKET,
	LDOT_CHAR,
	LDOT_BLOCK
} lws_dir_obj_type_t;

struct lws_dir_entry {
	const char *name;
	lws_dir_obj_type_t type;
};

typedef int
lws_dir_callback_function(const char *dirpath, void *user,
			  struct lws_dir_entry *lde);

/**
 * lws_dir() - get a callback for everything in a directory
 *
 * \param dirpath: the directory to scan
 * \param user: pointer to give to callback
 * \param cb: callback to receive information on each file or dir
 *
 * Calls \p cb (with \p user) for every object in dirpath.
 *
 * This wraps whether it's using POSIX apis, or libuv (as needed for windows,
 * since it refuses to support POSIX apis for this).
 */
LWS_VISIBLE LWS_EXTERN int
lws_dir(const char *dirpath, void *user, lws_dir_callback_function cb);

/**
 * lws_dir_rm_rf_cb() - callback for lws_dir that performs recursive rm -rf
 *
 * \param dirpath: directory we are at in lws_dir
 * \param user: ignored
 * \param lde: lws_dir info on the file or directory we are at
 *
 * This is a readymade rm -rf callback for use with lws_dir.  It recursively
 * removes everything below the starting dir and then the starting dir itself.
 * Works on linux, OSX and Windows at least.
 */
LWS_VISIBLE LWS_EXTERN int
lws_dir_rm_rf_cb(const char *dirpath, void *user, struct lws_dir_entry *lde);

/*
 * We pass every file in the base dir through a filter, and call back on the
 * ones that match.  Directories are ignored.
 *
 * The original path filter string may look like, eg, "sai-*.deb" or "*.txt"
 */

typedef int (*lws_dir_glob_cb_t)(void *data, const char *path);

typedef struct lws_dir_glob {
	const char		*filter;
	lws_dir_glob_cb_t	cb;
	void			*user;
} lws_dir_glob_t;

/**
 * lws_dir_glob_cb() - callback for lws_dir that performs filename globbing
 *
 * \param dirpath: directory we are at in lws_dir
 * \param user: pointer to your prepared lws_dir_glob_cb_t
 * \param lde: lws_dir info on the file or directory we are at
 *
 * \p user is prepared with an `lws_dir_glob_t` containing a callback for paths
 * that pass the filtering, a user pointer to pass to that callback, and a
 * glob string like "*.txt".  It may not contain directories, the lws_dir musr
 * be started at the correct dir.
 *
 * Only the base path passed to lws_dir is scanned, it does not look in subdirs.
 */
LWS_VISIBLE LWS_EXTERN int
lws_dir_glob_cb(const char *dirpath, void *user, struct lws_dir_entry *lde);

#endif

/**
 * lws_get_allocated_heap() - if the platform supports it, returns amount of
 *				heap allocated by lws itself
 *
 * On glibc currently, this reports the total amount of current logical heap
 * allocation, found by tracking the amount allocated by lws_malloc() and
 * friends and accounting for freed allocations via lws_free().
 *
 * This is useful for confirming where processwide heap allocations actually
 * come from... this number represents all lws internal allocations, for
 * fd tables, wsi allocations, ah, etc combined.  It doesn't include allocations
 * from user code, since lws_malloc() etc are not exported from the library.
 *
 * On other platforms, it always returns 0.
 */
size_t lws_get_allocated_heap(void);

/**
 * lws_get_tsi() - Get thread service index wsi belong to
 * \param wsi:  websocket connection to check
 *
 * Returns more than zero (or zero if only one service thread as is the default).
 */
LWS_VISIBLE LWS_EXTERN int
lws_get_tsi(struct lws *wsi);

/**
 * lws_is_ssl() - Find out if connection is using SSL
 * \param wsi:	websocket connection to check
 *
 * Returns nonzero if the wsi is inside a tls tunnel, else zero.
 */
LWS_VISIBLE LWS_EXTERN int
lws_is_ssl(struct lws *wsi);
/**
 * lws_is_cgi() - find out if this wsi is running a cgi process
 *
 * \param wsi: lws connection
 */
LWS_VISIBLE LWS_EXTERN int
lws_is_cgi(struct lws *wsi);

/**
 * lws_tls_jit_trust_blob_queury_skid() - walk jit trust blob for skid
 *
 * \param _blob: the start of the blob in memory
 * \param blen: the length of the blob in memory
 * \param skid: the SKID we are looking for
 * \param skid_len: the length of the SKID we are looking for
 * \param prpder: result pointer to receive a pointer to the matching DER
 * \param prder_len: result pointer to receive matching DER length
 *
 * Helper to scan a JIT Trust blob in memory for a trusted CA cert matching
 * a given SKID.  Returns 0 if found and *prpder and *prder_len are set, else
 * nonzero.
 */
LWS_VISIBLE LWS_EXTERN int
lws_tls_jit_trust_blob_queury_skid(const void *_blob, size_t blen,
				   const uint8_t *skid, size_t skid_len,
				   const uint8_t **prpder, size_t *prder_len);

/**
 * lws_open() - platform-specific wrapper for open that prepares the fd
 *
 * \param __file: the filepath to open
 * \param __oflag: option flags
 *
 * This is a wrapper around platform open() that sets options on the fd
 * according to lws policy.  Currently that is FD_CLOEXEC to stop the opened
 * fd being available to any child process forked by user code.
 */
LWS_VISIBLE LWS_EXTERN int
lws_open(const char *__file, int __oflag, ...);

struct lws_wifi_scan { /* generic wlan scan item */
	struct lws_wifi_scan *next;
	char ssid[32];
	int32_t rssi; /* divide by .count to get db */
	uint8_t bssid[6];
	uint8_t count;
	uint8_t channel;
	uint8_t authmode;
};

#if defined(LWS_WITH_TLS) && !defined(LWS_WITH_MBEDTLS)
/**
 * lws_get_ssl() - Return wsi's SSL context structure
 * \param wsi:	websocket connection
 *
 * Returns pointer to the SSL library's context structure
 */
LWS_VISIBLE LWS_EXTERN SSL*
lws_get_ssl(struct lws *wsi);
#endif

LWS_VISIBLE LWS_EXTERN void
lws_explicit_bzero(void *p, size_t len);

typedef struct lws_humanize_unit {
	const char *name; /* array ends with NULL name */
	uint64_t factor;
} lws_humanize_unit_t;

LWS_VISIBLE extern const lws_humanize_unit_t humanize_schema_si[7];
LWS_VISIBLE extern const lws_humanize_unit_t humanize_schema_si_bytes[7];
LWS_VISIBLE extern const lws_humanize_unit_t humanize_schema_us[8];

#if defined(_DEBUG)
void
lws_assert_fourcc(uint32_t fourcc, uint32_t expected);
#else
#define lws_assert_fourcc(_a, _b) do { } while (0);
#endif

/**
 * lws_humanize() - Convert possibly large number to human-readable uints
 *
 * \param buf: result string buffer
 * \param len: remaining length in \p buf
 * \param value: the uint64_t value to represent
 * \param schema: and array of scaling factors and units
 *
 * This produces a concise string representation of \p value, referencing the
 * schema \p schema of scaling factors and units to find the smallest way to
 * render it.
 *
 * Three schema are exported from lws for general use, humanize_schema_si, which
 * represents as, eg, "  22.130Gi" or " 128      "; humanize_schema_si_bytes
 * which is the same but shows, eg, "  22.130GiB", and humanize_schema_us,
 * which represents a count of us as a human-readable time like "  14.350min",
 * or "  1.500d".
 *
 * You can produce your own schema.
 */

LWS_VISIBLE LWS_EXTERN int
lws_humanize(char *buf, size_t len, uint64_t value,
	     const lws_humanize_unit_t *schema);

LWS_VISIBLE LWS_EXTERN void
lws_ser_wu16be(uint8_t *b, uint16_t u);

LWS_VISIBLE LWS_EXTERN void
lws_ser_wu32be(uint8_t *b, uint32_t u32);

LWS_VISIBLE LWS_EXTERN void
lws_ser_wu64be(uint8_t *b, uint64_t u64);

LWS_VISIBLE LWS_EXTERN uint16_t
lws_ser_ru16be(const uint8_t *b);

LWS_VISIBLE LWS_EXTERN uint32_t
lws_ser_ru32be(const uint8_t *b);

LWS_VISIBLE LWS_EXTERN uint64_t
lws_ser_ru64be(const uint8_t *b);

LWS_VISIBLE LWS_EXTERN int
lws_vbi_encode(uint64_t value, void *buf);

LWS_VISIBLE LWS_EXTERN int
lws_vbi_decode(const void *buf, uint64_t *value, size_t len);

///@}

#if defined(LWS_WITH_SPAWN)

/* opaque internal struct */
struct lws_spawn_piped;

#if defined(WIN32)
struct _lws_siginfo_t {
	int retcode;
};
typedef struct _lws_siginfo_t siginfo_t;
#endif

typedef void (*lsp_cb_t)(void *opaque, lws_usec_t *accounting, siginfo_t *si,
			 int we_killed_him);


/**
 * lws_spawn_piped_info - details given to create a spawned pipe
 *
 * \p owner: lws_dll2_owner_t that lists all active spawns, or NULL
 * \p vh: vhost to bind stdwsi to... from opt_parent if given
 * \p opt_parent: optional parent wsi for stdwsi
 * \p exec_array: argv for process to spawn
 * \p env_array: environment for spawned process, NULL ends env list
 * \p protocol_name: NULL, or vhost protocol name to bind stdwsi to
 * \p chroot_path: NULL, or chroot patch for child process
 * \p wd: working directory to cd to after fork, NULL defaults to /tmp
 * \p plsp: NULL, or pointer to the outer lsp pointer so it can be set NULL when destroyed
 * \p opaque: pointer passed to the reap callback, if any
 * \p timeout: optional us-resolution timeout, or zero
 * \p reap_cb: callback when child process has been reaped and the lsp destroyed
 * \p tsi: tsi to bind stdwsi to... from opt_parent if given
 */
struct lws_spawn_piped_info {
	struct lws_dll2_owner		*owner;
	struct lws_vhost		*vh;
	struct lws			*opt_parent;

	const char * const		*exec_array;
	const char			**env_array;
	const char			*protocol_name;
	const char			*chroot_path;
	const char			*wd;

	struct lws_spawn_piped		**plsp;

	void				*opaque;

	lsp_cb_t			reap_cb;

	lws_usec_t			timeout_us;
	int				max_log_lines;
	int				tsi;

	const struct lws_role_ops	*ops; /* NULL is raw file */

	uint8_t				disable_ctrlc;
};

/**
 * lws_spawn_piped() - spawn a child process with stdxxx redirected
 *
 * \p lspi: info struct describing details of spawn to create
 *
 * This spawns a child process managed in the lsp object and with attributes
 * set in the arguments.  The stdin/out/err streams are redirected to pipes
 * which are instantiated into wsi that become child wsi of \p parent if non-
 * NULL.  .opaque_user_data on the stdwsi created is set to point to the
 * lsp object, so this can be recovered easily in the protocol handler.
 *
 * If \p owner is non-NULL, successful spawns join the given dll2 owner in the
 * original process.
 *
 * If \p timeout is non-zero, successful spawns register a sul with the us-
 * resolution timeout to callback \p timeout_cb, in the original process.
 *
 * Returns 0 if the spawn went OK or nonzero if it failed and was cleaned up.
 * The spawned process continues asynchronously and this will return after
 * starting it if all went well.
 */
LWS_VISIBLE LWS_EXTERN struct lws_spawn_piped *
lws_spawn_piped(const struct lws_spawn_piped_info *lspi);

/*
 * lws_spawn_piped_kill_child_process() - attempt to kill child process
 *
 * \p lsp: child object to kill
 *
 * Attempts to signal the child process in \p lsp to terminate.
 */
LWS_VISIBLE LWS_EXTERN int
lws_spawn_piped_kill_child_process(struct lws_spawn_piped *lsp);

/**
 * lws_spawn_stdwsi_closed() - inform the spawn one of its stdxxx pipes closed
 *
 * \p lsp: the spawn object
 * \p wsi: the wsi that is closing
 *
 * When you notice one of the spawn stdxxx pipes closed, inform the spawn
 * instance using this api.  When it sees all three have closed, it will
 * automatically try to reap the child process.
 *
 * This is the mechanism whereby the spawn object can understand its child
 * has closed.
 */
LWS_VISIBLE LWS_EXTERN void
lws_spawn_stdwsi_closed(struct lws_spawn_piped *lsp, struct lws *wsi);

/**
 * lws_spawn_get_stdfd() - return std channel index for stdwsi
 *
 * \p wsi: the wsi
 *
 * If you know wsi is a stdwsi from a spawn, you can determine its original
 * channel index / fd before the pipes replaced the default fds.  It will return
 * one of 0 (STDIN), 1 (STDOUT) or 2 (STDERR).  You can handle all three in the
 * same protocol handler and then disambiguate them using this api.
 */
LWS_VISIBLE LWS_EXTERN int
lws_spawn_get_stdfd(struct lws *wsi);

#endif

struct lws_fsmount {
	const char	*layers_path;	/* where layers live */
	const char	*overlay_path;	/* where overlay instantiations live */

	char		mp[256];	/* mountpoint path */
	char		ovname[64];	/* unique name for mount instance */
	char		distro[64];	/* unique name for layer source */

#if defined(__linux__)
	const char	*layers[4];	/* distro layers, like "base", "env" */
#endif
};

/**
 * lws_fsmount_mount() - Mounts an overlayfs stack of layers
 *
 * \p fsm: struct lws_fsmount specifying the mount layout
 *
 * This api is able to assemble up to 4 layer directories on to a mountpoint
 * using overlayfs mount (Linux only).
 *
 * Set fsm.layers_path to the base dir where the layers themselves live, the
 * entries in fsm.layers[] specifies the relative path to the layer, comprising
 * fsm.layers_path/fsm.distro/fsm.layers[], with [0] being the deepest, earliest
 * layer and the rest being progressively on top of [0]; NULL indicates the
 * layer is unused.
 *
 * fsm.overlay_path is the base path of the overlayfs instantiations... empty
 * dirs must exist at
 *
 * fsm.overlay_path/overlays/fsm.ovname/work
 * fsm.overlay_path/overlays/fsm.ovname/session
 *
 * Set fsm.mp to the path of an already-existing empty dir that will be the
 * mountpoint, this can be whereever you like.
 *
 * Overlayfs merges the union of all the contributing layers at the mountpoint,
 * the mount is writeable but the layer themselves are immutable, all additions
 * and changes are stored in
 *
 * fsm.overlay_path/overlays/fsm.ovname/session
 *
 * Returns 0 if mounted OK, nonzero if errors.
 *
 * Retain fsm for use with unmounting.
 */
LWS_VISIBLE LWS_EXTERN int
lws_fsmount_mount(struct lws_fsmount *fsm);

/**
 * lws_fsmount_unmount() - Unmounts an overlayfs dir
 *
 * \p fsm: struct lws_fsmount specifying the mount layout
 *
 * Unmounts the mountpoint in fsm.mp.
 *
 * Delete fsm.overlay_path/overlays/fsm.ovname/session to permanently eradicate
 * all changes from the time the mountpoint was in use.
 *
 * Returns 0 if unmounted OK.
 */
LWS_VISIBLE LWS_EXTERN int
lws_fsmount_unmount(struct lws_fsmount *fsm);

#define LWS_MINILEX_FAIL -1
#define LWS_MINILEX_CONTINUE 0
#define LWS_MINILEX_MATCH 1

/**
 * lws_minilex_parse() - stateful matching vs lws minilex tables
 *
 * \p lex: the start of the precomputed minilex table
 * \p ps: pointer to the int16_t that holds the parsing state (init to 0)
 * \p c: the next incoming character to parse
 * \p match: pointer to take the match
 *
 * Returns either
 *
 *  - LWS_MINILEX_FAIL if there is no way to match the characters seen,
 * this is sticky for additional characters until the *ps is reset to 0.
 *
 *  - LWS_MINILEX_CONTINUE if the character could be part of a match but more
 *    are required to see if it can match
 *
 *  - LWS_MINILEX_MATCH and *match is set to the match index if there is a
 *    valid match.
 *
 * In cases where the match is ambiguous, eg, we saw "right" and the possible
 * matches are "right" or "right-on", LWS_MINILEX_CONTINUE is returned.  To
 * allow it to match on the complete-but-ambiguous token, if the caller sees
 * a delimiter it can call lws_minilex_parse() again with c == 0.  This will
 * either return LWS_MINILEX_MATCH and set *match to the smaller ambiguous
 * match, or return LWS_MINILEX_FAIL.
 */
LWS_VISIBLE LWS_EXTERN int
lws_minilex_parse(const uint8_t *lex, int16_t *ps, const uint8_t c,
			int *match);

/*
 * Reports the number of significant bits (from the left) that is needed to
 * represent u.  So if u is 0x80, result is 8.
 */

LWS_VISIBLE LWS_EXTERN unsigned int
lws_sigbits(uintptr_t u);

/**
 * lws_wol() - broadcast a magic WOL packet to MAC, optionally binding to if IP
 *
 * \p ctx: The lws context
 * \p ip_or_NULL: The IP address to bind to at the client side, to send the
 *                magic packet on.  If NULL, the system chooses, probably the
 *                interface with the default route.
 * \p mac_6_bytes: Points to a 6-byte MAC address to direct the magic packet to
 */
LWS_VISIBLE LWS_EXTERN int
lws_wol(struct lws_context *ctx, const char *ip_or_NULL, uint8_t *mac_6_bytes);