danicoin/external/unbound/libunbound/unbound.h
2014-10-05 23:44:31 +02:00

573 lines
20 KiB
C

/*
* unbound.h - unbound validating resolver public API
*
* Copyright (c) 2007, NLnet Labs. All rights reserved.
*
* This software is open source.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution.
*
* Neither the name of the NLNET LABS nor the names of its contributors may
* be used to endorse or promote products derived from this software without
* specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED
* TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
/**
* \file
*
* This file contains functions to resolve DNS queries and
* validate the answers. Synchonously and asynchronously.
*
* Several ways to use this interface from an application wishing
* to perform (validated) DNS lookups.
*
* All start with
* ctx = ub_ctx_create();
* err = ub_ctx_add_ta(ctx, "...");
* err = ub_ctx_add_ta(ctx, "...");
* ... some lookups
* ... call ub_ctx_delete(ctx); when you want to stop.
*
* Application not threaded. Blocking.
* int err = ub_resolve(ctx, "www.example.com", ...
* if(err) fprintf(stderr, "lookup error: %s\n", ub_strerror(err));
* ... use the answer
*
* Application not threaded. Non-blocking ('asynchronous').
* err = ub_resolve_async(ctx, "www.example.com", ... my_callback);
* ... application resumes processing ...
* ... and when either ub_poll(ctx) is true
* ... or when the file descriptor ub_fd(ctx) is readable,
* ... or whenever, the app calls ...
* ub_process(ctx);
* ... if no result is ready, the app resumes processing above,
* ... or process() calls my_callback() with results.
*
* ... if the application has nothing more to do, wait for answer
* ub_wait(ctx);
*
* Application threaded. Blocking.
* Blocking, same as above. The current thread does the work.
* Multiple threads can use the *same context*, each does work and uses
* shared cache data from the context.
*
* Application threaded. Non-blocking ('asynchronous').
* ... setup threaded-asynchronous config option
* err = ub_ctx_async(ctx, 1);
* ... same as async for non-threaded
* ... the callbacks are called in the thread that calls process(ctx)
*
* Openssl needs to have locking in place, and the application must set
* it up, because a mere library cannot do this, use the calls
* CRYPTO_set_id_callback and CRYPTO_set_locking_callback.
*
* If no threading is compiled in, the above async example uses fork(2) to
* create a process to perform the work. The forked process exits when the
* calling process exits, or ctx_delete() is called.
* Otherwise, for asynchronous with threading, a worker thread is created.
*
* The blocking calls use shared ctx-cache when threaded. Thus
* ub_resolve() and ub_resolve_async() && ub_wait() are
* not the same. The first makes the current thread do the work, setting
* up buffers, etc, to perform the work (but using shared cache data).
* The second calls another worker thread (or process) to perform the work.
* And no buffers need to be set up, but a context-switch happens.
*/
#ifndef _UB_UNBOUND_H
#define _UB_UNBOUND_H
#ifdef __cplusplus
extern "C" {
#endif
/** the version of this header file */
#define UNBOUND_VERSION_MAJOR @UNBOUND_VERSION_MAJOR@
#define UNBOUND_VERSION_MINOR @UNBOUND_VERSION_MINOR@
#define UNBOUND_VERSION_MICRO @UNBOUND_VERSION_MICRO@
/**
* The validation context is created to hold the resolver status,
* validation keys and a small cache (containing messages, rrsets,
* roundtrip times, trusted keys, lameness information).
*
* Its contents are internally defined.
*/
struct ub_ctx;
/**
* The validation and resolution results.
* Allocated by the resolver, and need to be freed by the application
* with ub_resolve_free().
*/
struct ub_result {
/** The original question, name text string. */
char* qname;
/** the type asked for */
int qtype;
/** the class asked for */
int qclass;
/**
* a list of network order DNS rdata items, terminated with a
* NULL pointer, so that data[0] is the first result entry,
* data[1] the second, and the last entry is NULL.
* If there was no data, data[0] is NULL.
*/
char** data;
/** the length in bytes of the data items, len[i] for data[i] */
int* len;
/**
* canonical name for the result (the final cname).
* zero terminated string.
* May be NULL if no canonical name exists.
*/
char* canonname;
/**
* DNS RCODE for the result. May contain additional error code if
* there was no data due to an error. 0 (NOERROR) if okay.
*/
int rcode;
/**
* The DNS answer packet. Network formatted. Can contain DNSSEC types.
*/
void* answer_packet;
/** length of the answer packet in octets. */
int answer_len;
/**
* If there is any data, this is true.
* If false, there was no data (nxdomain may be true, rcode can be set).
*/
int havedata;
/**
* If there was no data, and the domain did not exist, this is true.
* If it is false, and there was no data, then the domain name
* is purported to exist, but the requested data type is not available.
*/
int nxdomain;
/**
* True, if the result is validated securely.
* False, if validation failed or domain queried has no security info.
*
* It is possible to get a result with no data (havedata is false),
* and secure is true. This means that the non-existance of the data
* was cryptographically proven (with signatures).
*/
int secure;
/**
* If the result was not secure (secure==0), and this result is due
* to a security failure, bogus is true.
* This means the data has been actively tampered with, signatures
* failed, expected signatures were not present, timestamps on
* signatures were out of date and so on.
*
* If !secure and !bogus, this can happen if the data is not secure
* because security is disabled for that domain name.
* This means the data is from a domain where data is not signed.
*/
int bogus;
/**
* If the result is bogus this contains a string (zero terminated)
* that describes the failure. There may be other errors as well
* as the one described, the description may not be perfectly accurate.
* Is NULL if the result is not bogus.
*/
char* why_bogus;
/**
* TTL for the result, in seconds. If the security is bogus, then
* you also cannot trust this value.
*/
int ttl;
};
/**
* Callback for results of async queries.
* The readable function definition looks like:
* void my_callback(void* my_arg, int err, struct ub_result* result);
* It is called with
* void* my_arg: your pointer to a (struct of) data of your choice,
* or NULL.
* int err: if 0 all is OK, otherwise an error occured and no results
* are forthcoming.
* struct result: pointer to more detailed result structure.
* This structure is allocated on the heap and needs to be
* freed with ub_resolve_free(result);
*/
typedef void (*ub_callback_t)(void*, int, struct ub_result*);
/**
* Create a resolving and validation context.
* The information from /etc/resolv.conf and /etc/hosts is not utilised by
* default. Use ub_ctx_resolvconf and ub_ctx_hosts to read them.
* @return a new context. default initialisation.
* returns NULL on error.
*/
struct ub_ctx* ub_ctx_create(void);
/**
* Destroy a validation context and free all its resources.
* Outstanding async queries are killed and callbacks are not called for them.
* @param ctx: context to delete.
*/
void ub_ctx_delete(struct ub_ctx* ctx);
/**
* Set an option for the context.
* @param ctx: context.
* @param opt: option name from the unbound.conf config file format.
* (not all settings applicable). The name includes the trailing ':'
* for example ub_ctx_set_option(ctx, "logfile:", "mylog.txt");
* This is a power-users interface that lets you specify all sorts
* of options.
* For some specific options, such as adding trust anchors, special
* routines exist.
* @param val: value of the option.
* @return: 0 if OK, else error.
*/
int ub_ctx_set_option(struct ub_ctx* ctx, const char* opt, const char* val);
/**
* Get an option from the context.
* @param ctx: context.
* @param opt: option name from the unbound.conf config file format.
* (not all settings applicable). The name excludes the trailing ':'
* for example ub_ctx_get_option(ctx, "logfile", &result);
* This is a power-users interface that lets you specify all sorts
* of options.
* @param str: the string is malloced and returned here. NULL on error.
* The caller must free() the string. In cases with multiple
* entries (auto-trust-anchor-file), a newline delimited list is
* returned in the string.
* @return 0 if OK else an error code (malloc failure, syntax error).
*/
int ub_ctx_get_option(struct ub_ctx* ctx, const char* opt, char** str);
/**
* setup configuration for the given context.
* @param ctx: context.
* @param fname: unbound config file (not all settings applicable).
* This is a power-users interface that lets you specify all sorts
* of options.
* For some specific options, such as adding trust anchors, special
* routines exist.
* @return: 0 if OK, else error.
*/
int ub_ctx_config(struct ub_ctx* ctx, const char* fname);
/**
* Set machine to forward DNS queries to, the caching resolver to use.
* IP4 or IP6 address. Forwards all DNS requests to that machine, which
* is expected to run a recursive resolver. If the proxy is not
* DNSSEC-capable, validation may fail. Can be called several times, in
* that case the addresses are used as backup servers.
*
* To read the list of nameservers from /etc/resolv.conf (from DHCP or so),
* use the call ub_ctx_resolvconf.
*
* @param ctx: context.
* At this time it is only possible to set configuration before the
* first resolve is done.
* @param addr: address, IP4 or IP6 in string format.
* If the addr is NULL, forwarding is disabled.
* @return 0 if OK, else error.
*/
int ub_ctx_set_fwd(struct ub_ctx* ctx, const char* addr);
/**
* Read list of nameservers to use from the filename given.
* Usually "/etc/resolv.conf". Uses those nameservers as caching proxies.
* If they do not support DNSSEC, validation may fail.
*
* Only nameservers are picked up, the searchdomain, ndots and other
* settings from resolv.conf(5) are ignored.
*
* @param ctx: context.
* At this time it is only possible to set configuration before the
* first resolve is done.
* @param fname: file name string. If NULL "/etc/resolv.conf" is used.
* @return 0 if OK, else error.
*/
int ub_ctx_resolvconf(struct ub_ctx* ctx, const char* fname);
/**
* Read list of hosts from the filename given.
* Usually "/etc/hosts".
* These addresses are not flagged as DNSSEC secure when queried for.
*
* @param ctx: context.
* At this time it is only possible to set configuration before the
* first resolve is done.
* @param fname: file name string. If NULL "/etc/hosts" is used.
* @return 0 if OK, else error.
*/
int ub_ctx_hosts(struct ub_ctx* ctx, const char* fname);
/**
* Add a trust anchor to the given context.
* The trust anchor is a string, on one line, that holds a valid DNSKEY or
* DS RR.
* @param ctx: context.
* At this time it is only possible to add trusted keys before the
* first resolve is done.
* @param ta: string, with zone-format RR on one line.
* [domainname] [TTL optional] [type] [class optional] [rdata contents]
* @return 0 if OK, else error.
*/
int ub_ctx_add_ta(struct ub_ctx* ctx, const char* ta);
/**
* Add trust anchors to the given context.
* Pass name of a file with DS and DNSKEY records (like from dig or drill).
* @param ctx: context.
* At this time it is only possible to add trusted keys before the
* first resolve is done.
* @param fname: filename of file with keyfile with trust anchors.
* @return 0 if OK, else error.
*/
int ub_ctx_add_ta_file(struct ub_ctx* ctx, const char* fname);
/**
* Add trust anchors to the given context.
* Pass the name of a bind-style config file with trusted-keys{}.
* @param ctx: context.
* At this time it is only possible to add trusted keys before the
* first resolve is done.
* @param fname: filename of file with bind-style config entries with trust
* anchors.
* @return 0 if OK, else error.
*/
int ub_ctx_trustedkeys(struct ub_ctx* ctx, const char* fname);
/**
* Set debug output (and error output) to the specified stream.
* Pass NULL to disable. Default is stderr.
* @param ctx: context.
* @param out: FILE* out file stream to log to.
* Type void* to avoid stdio dependency of this header file.
* @return 0 if OK, else error.
*/
int ub_ctx_debugout(struct ub_ctx* ctx, void* out);
/**
* Set debug verbosity for the context
* Output is directed to stderr.
* @param ctx: context.
* @param d: debug level, 0 is off, 1 is very minimal, 2 is detailed,
* and 3 is lots.
* @return 0 if OK, else error.
*/
int ub_ctx_debuglevel(struct ub_ctx* ctx, int d);
/**
* Set a context behaviour for asynchronous action.
* @param ctx: context.
* @param dothread: if true, enables threading and a call to resolve_async()
* creates a thread to handle work in the background.
* If false, a process is forked to handle work in the background.
* Changes to this setting after async() calls have been made have
* no effect (delete and re-create the context to change).
* @return 0 if OK, else error.
*/
int ub_ctx_async(struct ub_ctx* ctx, int dothread);
/**
* Poll a context to see if it has any new results
* Do not poll in a loop, instead extract the fd below to poll for readiness,
* and then check, or wait using the wait routine.
* @param ctx: context.
* @return: 0 if nothing to read, or nonzero if a result is available.
* If nonzero, call ctx_process() to do callbacks.
*/
int ub_poll(struct ub_ctx* ctx);
/**
* Wait for a context to finish with results. Calls ub_process() after
* the wait for you. After the wait, there are no more outstanding
* asynchronous queries.
* @param ctx: context.
* @return: 0 if OK, else error.
*/
int ub_wait(struct ub_ctx* ctx);
/**
* Get file descriptor. Wait for it to become readable, at this point
* answers are returned from the asynchronous validating resolver.
* Then call the ub_process to continue processing.
* This routine works immediately after context creation, the fd
* does not change.
* @param ctx: context.
* @return: -1 on error, or file descriptor to use select(2) with.
*/
int ub_fd(struct ub_ctx* ctx);
/**
* Call this routine to continue processing results from the validating
* resolver (when the fd becomes readable).
* Will perform necessary callbacks.
* @param ctx: context
* @return: 0 if OK, else error.
*/
int ub_process(struct ub_ctx* ctx);
/**
* Perform resolution and validation of the target name.
* @param ctx: context.
* The context is finalized, and can no longer accept config changes.
* @param name: domain name in text format (a zero terminated text string).
* @param rrtype: type of RR in host order, 1 is A (address).
* @param rrclass: class of RR in host order, 1 is IN (for internet).
* @param result: the result data is returned in a newly allocated result
* structure. May be NULL on return, return value is set to an error
* in that case (out of memory).
* @return 0 if OK, else error.
*/
int ub_resolve(struct ub_ctx* ctx, const char* name, int rrtype,
int rrclass, struct ub_result** result);
/**
* Perform resolution and validation of the target name.
* Asynchronous, after a while, the callback will be called with your
* data and the result.
* @param ctx: context.
* If no thread or process has been created yet to perform the
* work in the background, it is created now.
* The context is finalized, and can no longer accept config changes.
* @param name: domain name in text format (a string).
* @param rrtype: type of RR in host order, 1 is A.
* @param rrclass: class of RR in host order, 1 is IN (for internet).
* @param mydata: this data is your own data (you can pass NULL),
* and is passed on to the callback function.
* @param callback: this is called on completion of the resolution.
* It is called as:
* void callback(void* mydata, int err, struct ub_result* result)
* with mydata: the same as passed here, you may pass NULL,
* with err: is 0 when a result has been found.
* with result: a newly allocated result structure.
* The result may be NULL, in that case err is set.
*
* If an error happens during processing, your callback will be called
* with error set to a nonzero value (and result==NULL).
* @param async_id: if you pass a non-NULL value, an identifier number is
* returned for the query as it is in progress. It can be used to
* cancel the query.
* @return 0 if OK, else error.
*/
int ub_resolve_async(struct ub_ctx* ctx, const char* name, int rrtype,
int rrclass, void* mydata, ub_callback_t callback, int* async_id);
/**
* Cancel an async query in progress.
* Its callback will not be called.
*
* @param ctx: context.
* @param async_id: which query to cancel.
* @return 0 if OK, else error.
* This routine can return an error if the async_id passed does not exist
* or has already been delivered. If another thread is processing results
* at the same time, the result may be delivered at the same time and the
* cancel fails with an error. Also the cancel can fail due to a system
* error, no memory or socket failures.
*/
int ub_cancel(struct ub_ctx* ctx, int async_id);
/**
* Free storage associated with a result structure.
* @param result: to free
*/
void ub_resolve_free(struct ub_result* result);
/**
* Convert error value to a human readable string.
* @param err: error code from one of the ub_val* functions.
* @return pointer to constant text string, zero terminated.
*/
const char* ub_strerror(int err);
/**
* Debug routine. Print the local zone information to debug output.
* @param ctx: context. Is finalized by the routine.
* @return 0 if OK, else error.
*/
int ub_ctx_print_local_zones(struct ub_ctx* ctx);
/**
* Add a new zone with the zonetype to the local authority info of the
* library.
* @param ctx: context. Is finalized by the routine.
* @param zone_name: name of the zone in text, "example.com"
* If it already exists, the type is updated.
* @param zone_type: type of the zone (like for unbound.conf) in text.
* @return 0 if OK, else error.
*/
int ub_ctx_zone_add(struct ub_ctx* ctx, const char *zone_name,
const char *zone_type);
/**
* Remove zone from local authority info of the library.
* @param ctx: context. Is finalized by the routine.
* @param zone_name: name of the zone in text, "example.com"
* If it does not exist, nothing happens.
* @return 0 if OK, else error.
*/
int ub_ctx_zone_remove(struct ub_ctx* ctx, const char *zone_name);
/**
* Add localdata to the library local authority info.
* Similar to local-data config statement.
* @param ctx: context. Is finalized by the routine.
* @param data: the resource record in text format, for example
* "www.example.com IN A 127.0.0.1"
* @return 0 if OK, else error.
*/
int ub_ctx_data_add(struct ub_ctx* ctx, const char *data);
/**
* Remove localdata from the library local authority info.
* @param ctx: context. Is finalized by the routine.
* @param data: the name to delete all data from, like "www.example.com".
* @return 0 if OK, else error.
*/
int ub_ctx_data_remove(struct ub_ctx* ctx, const char *data);
/**
* Get a version string from the libunbound implementation.
* @return a static constant string with the version number.
*/
const char* ub_version(void);
#ifdef __cplusplus
}
#endif
#endif /* _UB_UNBOUND_H */