libspeex reports warnings in case of oddnesses with the data to stderr.
However Icecast2 uses those functions in it's codec probeing process.
This way libspeex will write messages to a posibily reused random handle.
This is clearly a bug by package maintainers of libspeex.
See the source code of this change as well as libspeex's source code
as reference.
This fixes the regression introduced by r19250 (461a537561).
The fix checks if the source client is actually known before printing it's IP-Address.
See also: #2016
As proposed in #2272, this renames the mount specific
mp3-metadata-interval to icy-metadata-interval to prevent
confusion about what it does (setting the icy metadata
interval) and to make clear it's not mp3 specific but ICY
specific.
This changes the file extension check in a way that it no longer
ignores trailing characters. This significantly reduces the risk
for false positives while matching. However this invalidates old
setups with files like foo.xsl3. However I have never files like
that in the wild.
This is based on the patch privided by ePirat in ticket #2248.
See: #2248
This replaces the add node function with the one intended for text nodes
to prevent that invalid entities are not encoded and therefore could
break the XML output.
(See d739c65e54 and #2255)
This fixes a bug where listener Referer or User-Agent strings containing
a sequence like `&T;` was not escaped properly and therefore made the
whole XML document invalid.
Injecting new XML nodes (<foo></foo>) was not possible, it seems in the
worse case all it could cause was that the XML failed to parse.
Fix#2255
There was an error how URIs were handled in our custom XSL loader
that rewrites include URIs to point to admin dir if the include
is not found in the current dir.
The check for the file would not work if the path contains chars
that are usually escaped in URIs, like a space (%20), as we get
a already encoded version of that URI in the loader and did not
decode it before checking the existence of the file.
(Fix#2249)
This works around the problem that windows ignores tailing dots.
This way you could trick Icecast2 to send a XSLT file as plain text.
Please test.
Closes: #2247
This moved the <mime-types> setting into <paths>. The code still
supports reading it from the root element but will warn the user
about this.
Also there seems to be no documentation about this setting.
Closes: #2164
* Loop over elements in input buffer.
* Most are literally copied with their contents without
inspection into the header or data buffers as appropriate.
* Some only copy the element header, to allow inspecting
children elements.
* Cluster elements are identified and used as sync points.
No probing is done for keyframes *yet*
ban and allow file support has been moved into a generic implementation
that can be re-used by later code such as proxy matching or other
blacklisting.
See: #1959
This fixes a bug that could cause the global listener count of Icecast to
be negative. This was caused due to wrong handling of errors in the
format_prepare_headers function.
Fixes: #2199
Thanks a lot to Simeon Völkel (0xBD4E031CDB4043C9) for reporting
and investigating the bug.
This let util_url_escape() handle NULL passed as parameter.
In case the parameter is NULL it will also return NULL.
This patch also does some cleanup of the code such as migration
away from int and thus avoiding future failtures.
This is a re-apply of 3baa4e46aa
as git refuses to cherry-pick.
This is a first commit to correct minor problmes reported by
Denis Denisov. There is likely another commit to correct some more
such problems.
Thank you Denis Denisov.
This cleans the libcurl usage up a bit. It moves common code used
to interact with libcurl into a single new file curl.[ch].
Notes:
* It does not alter any features nor fixes any bugs.
* Thew REVIEW about strdup(self->url) in event_url.c as been processed.
According to the manpage for curl_easy_setopt()/CURLOPT_URL
there is no need for us to copy.
* URL Auth as well as URL triggers have sent ICECAST_VERSION_STRING
as User-Agent:, this has be corrected to what is set in the config
file.
* As curl.c is now a single point for setting parameters all the TLS
parameters should be reviewed and set (based on config).
Please test!
This allows to store a history of played songs along the source
object and report it as part of the status XML.
Additional work needs to be done to make this configurable.
Also format_mp3.c needs work to support this.
A generic song changed handler should be implemented to handle this
in a nice way.
That one should also be the point to call logging_playlist().
See: #766
This adds the operation mode "strict". It does not yet have
any specific features and alias the "normal" mode.
This change is mostly to make the setting of "strict" valid
so it can be used already.
As we do not support any SSL version but TLS this renames
all TLS related config options. Docs still need to be updated.
Close still uses 'ssl' internally e.g. for variable names.
This should be changed on a later patch.
This switches the behaviour introduced with the last commit
so that now the admin folder is only checked for includes if
in the specified path none is found.
This adds support for including xslt files that are in the admin
directory in an xslt file which is in the webroot, without having to
specify the full path.
For example:
<xsl:include href="includes/head.xsl"/>
will look inside the admin directory (foo/bar/admin/includes/header.xsl)
and include it, if it's found. If not it will as usual look for it in
the directory where the xslt file is located. This enables us to have
common files like the header, footer, json stuff, etc. in the admin
directory, so we don't need to have an extra copy in the web folder.
Note:
The configuration value for the admin directory is requested
every time an xslt is requested (if not in the cache).
Additonally it should be noted that the includes are not cached.
This tries to unbreak the 100-continue thing a bit.
Proper headers are send in case of 100-continue. Also the 200 OK
was moved to the end of the stream.
The client may even be reused. But I'm not sure we only do
'End of stream' in cases we are not on an undefined state of the
client. Do we as of now support anything that give us EOS without
a fatal socket state (error or shutdown)?
At least this unbreaks 100-continue with active TLS that was broken
as hell as we sent the header in clear over the socket. Not checking
if it was transmitted completly or something. Using lowlevel socket
calls (in contrast to the connection level calls that would have
handled the TLS just fine).
While this seem sto be a huge improvement to the situation I kindly
ask you to do a lot testing on this. Both with SOURCE as well as
PUT clients.
We have answered SOURCE/PUT requests with a simple
"HTTP/1.0 200 OK\r\n\r\n" while we should answer with complet
headers. This commit corrects this.
Interoperability:
This will work for two classes of clients:
* Super dumb ones not checking the response and will just send
data and hope the socket is still open.
* Those with a real parser.
* Shoutcast clients (indepnded code path).
* All libshout, curl, wget, ... based clients.
This *could* break:
* Clients that check for the exact returned message. However
as we start with the same string this should work.
Only client that check for the string including the End-of-header
CRLF sequense may break.
Please test careful before release.
This exposes the TLS status as <tls> tag in the <listener> tag.
The tag contains a bool ("true" or "false"). TLS mode, version and
other parameters are out of the scope of this tag and may be added
as independed tags in later commits.
This fixes a bug in 8941de3273975504aaa00a35d145c3f26c832d70:
When passing NULL as client parameter to util_http_build_header()
there was a segfault before this change.
This allows the usage of RFC 2817 style upgrade commands.
This is not correctly announced yet.
Can hardly be tested as most clients do not support this.
Will be helpful with TLS support in libshout.
See: #2159, #2152
This adds to all responses:
* "Accept-Encoding: identity"
* "Connection: close"
* "Allow: GET, SOURCE" or "Allow: GET" (on admin pages)
This is to generate some infrastructure for Icecast to tell futur
source clients what is supported.
"Upgrade: TLS/1.0" should be added later when upgrade support is present.
See: #2152
This (hopefully) fixes a memory leak that caused active and
pending ypdata_t objects to be not freed up on server shutdown
or config reload.
To testers: please enable yp submission and restart and SIGHUP
your instance several times with loglevel set to warning or higher.
Each time after shutdown or SIGHUP check your log for one of the
following two warnings:
* 'active ypdata not freed up'
* 'pending ypdata not freed up'
Those warnings must not be in the log. If you spot one please report.
This adds info to the blocks about their size.
Also added two more blocks:
* for experimental features such as in private branches or
or feature branches. This is for command IDs that are going to
be merged with master some day. At merge they need to be changed
and put into the regular position.
* for private features that are not going to be merged into master.
e.g. specific features for a very specific user that are not
useful for general audiance.
This should help keeping stuff sorted and largey reducing the risk
of command ID collissions on merges of branches.
Also this should make the merger's job a little bit more easy as
there are less likely any conflicts on merge if someone else got
a number closeby and those change is within the diff's context lines.
This avoids a endless loop in a spin lock that has not been inited
yet if fserve engine was not started up by the time it was sent to
go down. This happened if icecast couldn't read the config because
the file was corrupted or non-existent.
Closes: #2144
This sets the on_demand parameter of the source correctly on
creation of a on-demand relay.
Thanks to Marvin Scholz <epirat07@gmail.com>.
Closes: #2054
This adds support to list global and per mount roles on stats.xsl.
This isn't nicest formated but works.
This also adds <source>-like <authentication> with <role>s to
root node.
Closes: #2133
This allows <resource> (former <alias>) to have only one of the
attribues source and destination. In that case other parameters
will be applied to the client but uri is not mapped.
See #2097
This allows a <role> represented by a auth_t to run in "immediate"
mode. In this mode no thread is created for this <role>. This is a
major speedup.
Closes#2124
This allows to manage <role>s via admin interface if the role supports.
Also format of admin/manageauth has been changed:
- <source> was renamed to <role>.
- mount parameter was removed.
- <role> got new parameters: type, name,
can-adduser, can-deleteuser, can-listuser.
- can-* parameters are bools ("true" or "false"). They should be used
to show or hide elements on the admin interface.
Ticket #2123 is nearly complet with this, just admin/manageauth.xsl
needs up be updated. Please close the bug in the commit that updates
admin/manageauth.xsl.
See #2123
This adds setting a URL for manegement of roles to the framework.
If no URl is given in the config file this defaults to internal
(/admin/manageauth.xsl) interface if supported by the backend.
See #2123
This exports id and name of <role>s on stats.xml output. The <role>s
are in a <authentication> container tag and mimic the format used
by the config file.
See #2123
This added a unique ID to each auth_t instance so it can be refered
to e.g. by the web interface for mangement functionallity. Mostly
stolen from connection.[ch].
See #2123
This fixes some compiler warnings. The code should be warning free
if gcc with -Wall is used but for the following:
{{{
admin.c: In function 'command_manageauth':
admin.c:790: warning: unused variable 'mountinfo'
}}}
This warning is due <role> migration. See #2123 for more details on
this warning. Please test with care. closes 2063
Old-style <authentication> within <mount> didn't work for type="url"
as well as some other parameters due to confusion between "node"
and "child" variable.
Thanks for trilliot for pointing out! Should work now.
closes#2039
This avoids if(strcmp()) constructs in favor of if(strcmp() == 0)
constructs. This is aligned with coding style and will improve
readability of the code.
Thanks to miksago. closes#2102
All metadata is now extracted from Vorbis streams and ICY streams
and exported as new <metadata> tag into admin/stats if specific
mountpoint is requested.
To testers:
- Please test with Ogg Vorbis streams.
- Please test with Ogg Vorbis + Theora streams.
- Please test chained Ogg streams.
- Please test with ICY streams.
close#2106
If there are no reistrations that are added to an event the event
doesn't need to be emited as nobody cares anyway. We will do this
simple check before we copy data into the event object to avoid
unnecessary calls to strdup() and other allocation functions.
<event> has been added and can be used within <kartoffelsalat>
both in <icecast> and <mount>.
<event> takes backend depending <option> child tags.
Currently supported backends:
- log: send message to error log.
- exec: executes a program or script.
- url: delivers the event via HTTP.
within <mount> <on-connect> and <on-disconnect> has been replaced by
<event>. Config parser can on-the-fly convert old tags.
Also <authentication type="url"> within <mount> has been fixed
for those cases with <option name="mount_add" .../> and
<option name="mount_remove" .../> which are now on-the-fly converted
by the parser to corresponding <event> tags.
Please also see TAGs added as per #2098. Some include hints for
documentation updates needed after this change. Those updates
should take place before 2.4.2.
Renamed event_config_read() into config_reread_config() and moved
it into cfgfile.c. This allowed to delete event.[ch].
event.[ch] will later be used to implement <event>.
Added {allow,deny}-all option to ACL parser. Setting this will
set {allow,deny}-{method,admin,web} to *.
Also allow boolean values (insteed of *) for
{allow,deny}-{web,all}.
renamed <alias> to <resource> as it is planned to extent the usage
of this tag. docs and default configs should be updated by those
taking core of them.
- Added <role>.
- Converted <Referer> to <referer>.
- Set id="nnn" in addition to <ID>nnn</ID>.
- Added TAGged comments about next steps.
refs #2097
svn path=/icecast/trunk/icecast/; revision=19364
This patch *replaces* the authentication system completly.
What is new:
- <authentication> in mount section is now a container object.
- <authentication> in root and mount section may hold any number of <role>-Tags.
- <role> tags:
Those tags define a 'role' and it's ACL rules.
A role is a instance of an authentication module (see below).
<role> takes the following options. All but type are optional.
- authentication related:
- type: Type of the authentication module (values: anonymous, static, legacy-password, url or htpasswd;
symbolic constants in auth.h)
- name: Name for the role. For later matching. (values: any string; default: (none))
- method: This rule is only active on the given list of HTTP methods.
(list of enum values: methods as recognized by httpp/ (e.g: get,post); default: *)
- ACL related:
- allow-method: Allowed HTTP methods.
(list of enum values: methods as recognized by httpp/ (e.g: get,post); default: get)
- deny-method: Rejected HTTP methods.
(list of enum values: methods as recognized by httpp/ (e.g: get,post); default: *)
- allow-admin: Allowed admin commands. (list of enum values: admin command; default: buildm3u)
- deny-admin: Rejected admin commands. (list of enum values: admin command; default: *)
- allow-web: Allowed web pages. (values: empty or *; default: *)
- deny-web: Rejected web pages. (values: empty or *; default: (empty))
- connections-per-user: maximum number of simultaneous connections per role and username.
This is only active on active sources. (values: unlimited or number of connections; default: unlimited)
- connection-duration: maximum time of a connection. This is only active on active sources.
(values: unlimited or number of secounds; default: unlimited)
<role> takes <option> child tags. <option> tags contain a name and a value option.
Meaning of <option> tags is up to the authentication module.
- <role>s are considered to build a stack. If a role returns with AUTH_NOMATCH the next one will be tried.
- <role>s are tested in this order: mount specific, default mount specific, global, internal fallback.
Internal fallback is set to allow web/ access via GET, POST and HEAD (only GET supported by this time)
and rejects all other requests.
- New authentication module: anonymous
This module matches all requests. No options taken.
- New authentication module: static
This module matches with a static username and password.
It takes two <option>s. One with name="username" and one with name="password" to set username and password.
This replaces old style <*-username> and <*-password> tags.
- New authentication module: legacy-password
This module matches with a statich password.
It takes one <option> with name="password" to set password.
This replaces old ICE and ICY (shoutcast compat mode) authentication.
- Parsing <authentication> in <mount> with a type set in a special way to allow 100% backward compatibility.
- Parsing of <source-password>, <admin-password>, <admin-user>, <relay-password> and <relay-user> in global
<authentication> for 100% backward compatibility.
- <alias> is now proccessed very early. This enables them to be used for all kinds of requests.
To Do List & What does not yet work:
- type="url" auth: mount_add and mount_remove.
This should be replaced by an unique feature I would call '<event>'.
- Admin commands manageauth and manageauth.xsl are disabled as they need more review:
This code needs to be ported to support multiple <role>s per <mount>.
- url authentication module can not yet return AUTH_NOMATCH.
This needs to be reviewed and discussed on how to handle this case best way.
- Default config files needs to be updated to reflect the changes.
As this is quite some political act it should be done in dicussion with the whole team
and permission of the release manager.
- Docs need to be updated to reflect the changes.
How does it work:
Code has been changed so that authentification is done early for all clients.
This allows accessing the ACL data (client->acl) from nearly everywhere in the code.
After accept() and initial client setup the request is parsed. In the next step
all <alias>es are resolved. After this the client is passed for authentication.
After authentication it is passed to the corresponding subsystem depending on kind of request.
All authentication instances have a thread running for doing the authentication.
This thread works on a queue of clients.
Hints for testers:
- Test with default config.
- Test with diffrent authentication modules in <mount>.
- Test shoutcast compatibility mode.
- Test with new style <authentication> and any amount of <role> (zero to quite some).
- Test <alias> lookup on all kinds of objects.
- Test source level credential login into the admin interface.
- Test shoucast style meta data updates.
- Test playlist generation.
Thank you for reading this long commit message. Have fun reading the full patch!
svn path=/icecast/trunk/icecast/; revision=19358
Please test this very carefully.
Some pointers what should be in the tests (NOT complet list):
- request to 'static' web/ and admin/ pages.
- requests to playlist generation.
- requests to streams.
- requests to admin/ manipulation functions.
- test everything with at least 8kB of extra headers, then reduce in 1kB (or 512B) steps.
- see if response is correct OR 500 is returned.
- run under valgrind or similar to see no buffer overflow or similiar will happen.
- take a cookie!
svn path=/icecast/trunk/icecast/; revision=19300
In case of SOURCE we are lenient and thus quite some source clients
don't send a proper content-type, especially if they only support mp3.
This was meant to be introduced in 2.4.0 already, sadly we missed it.
All source clients MUST send proper content-type after migrating to
Icecast HTTP PUT protocol.
closes#2082
svn path=/icecast/trunk/icecast/; revision=19288
Add source IP adress to startup and source exit logging, see #2016
Add mountpoint to some log lines, see #1388
svn path=/icecast/trunk/icecast/; revision=19250
In case of <changeowner> only UID and GID were changed,
supplementary groups were left in place.
This is a potential security issue only if <changeowner> is used.
New behaviour is to set UID, GID and set supplementary groups
based on the UID
Even in case of icecast remaining in supplementary group 0
this "only" gives it things like access to files that are owned
by group 0 and according to their umask. This is obviously bad,
but not as bad as UID 0 with all its other special rights.
It's a security issue and we fix immediately and recommend users to update.
PS: Cherry picking this should be fine by distros for fixing older releases.
svn path=/icecast/trunk/icecast/; revision=19137
ISO8601 compliante timestamps for statistics. Should make usage in e.g. JSON
much easier.
Added as new variables to avoid breaking backwards compatibility.
svn path=/icecast/trunk/icecast/; revision=19122
I've rethought the approach and found a much cleaner way.
We now split handling of command line arguments into two parts.
Only the critical part of getting the config file is done first (and -v as it prevents startup).
The rest (currently only -b) is deferred.
This achieves the important bit why #1886 was created.
It allows us to log error messages to stderr even if the -b argument is passed.
This is mainly for the case where the logfile or TCP port can't be opened.
svn path=/icecast/trunk/icecast/; revision=18945
The default mount is a block in the config file that contains settings for
all mount points that do not have a block in configfile themself.
This is implemented by a <mount type="default">-block.
In this case the <mount>-block MUST NOT contain a <mount-name>-subblock.
svn path=/icecast/trunk/icecast/; revision=18902
We are handling it the same as we would handle a SOURCE request.
Due to legacy code, sender MUST send proper content-type header,
if content type is not audio/mpeg!
Can be tested using real-time encoded output and piping it into
| curl -u username:password -H "Content-type: application/ogg" -T - http://localhost:8000/mountname.ogg
Note that this example has ZERO timing, so a simple 'cat *.ogg' will fail.
Whatever feeds the pipe must do it at proper timing for real-time playback!
svn path=/icecast/trunk/icecast/; revision=18888
* added ip= and agent=,
* Some cleanup && made code more uniform,
* avoid int for size_t vars.
actions: close#1422
svn path=/icecast/trunk/icecast/; revision=18798
This is currently not implemented for SOURCE and STATS clients as
I suspect to break them. This needs some more research.
close#1639, see #1870 and #1885.
svn path=/icecast/trunk/icecast/; revision=18464
This is a self-contained ebml parser. It just looks
for cluster boundaries and breaks the stream there.
svn path=/icecast/branches/icecast-webm/; revision=18297
specifying the TCP send buffer size, it seems that on at least some win32 systems,
the window size stays at 8k (even with registry settings) which could limit available
streaming bandwidth.
svn path=/icecast/trunk/icecast/; revision=15766
a non-crash type race between artist and title fields updating. Also allow
passing a url setting, mapping to StreamURL in the icy metadata block, but do
not override any existing inline setting.
svn path=/icecast/trunk/icecast/; revision=15622
issue requests (using ithe stream_auth option) to allow external engines to
determine whether a client can stream or not. Admin requests using source auth
are able to use this mechanism however source clients using the icy protocol
cannot yet.
svn path=/icecast/trunk/icecast/; revision=15621
once, only the first page is refcounted multiple times by the queue. Put some
checks in to help capture odd cases if they arise.
svn path=/icecast/trunk/icecast/; revision=15613
connect and be waiting a while before receiving stream content.
When retrieving streamlist, don't wait too long for connection to master
schedule a full stats rebuild on source disconnect.
Minor build fix for certain platforms with missing defines
svn path=/icecast/trunk/icecast/; revision=15276
the user is admin. It seems that some source clients issue metadata updates
even if they are rejected due to the mountpoint being in use.
svn path=/icecast/trunk/icecast/; revision=15264
just in a sleep loop which means most clients getting to it have to wait. We
now just start the thread when there is work to do. There is also some small
cleanup for the off_t type usage as well.
svn path=/icecast/trunk/icecast/; revision=15124
the overall time is very slow. This lag is mainly down to the sleep that
the connection and fserve threads do to prevent busy looping when idle. The
sleep has to expire even if requests are waiting.
This patch merges the tasks done by the connection and listener thread into
one thread. It allows us to remove some locking, reduce thread usage and
service new connections more quickly as the poll/select acts as the busy
CPU prevention.
svn path=/icecast/trunk/icecast/; revision=15123
We prevent a failed relay from restarting too quickly but don't prevent
connecting listeners from getting to the source_t during that time. Drop the
on-demand status in the source_t during this period so listeners are rejected.
The setting is reset from relay eventually.
Relays without matching mount sections had no stats initially.
svn path=/icecast/trunk/icecast/; revision=15122
is using them currently, so lets reduce the structure size. On win32, this
extra lock increases the handles usage considerably.
svn path=/icecast/trunk/avl/; revision=15120
need a large number of mountpoints to trigger a truncation. Allow file
serving thread to process a list of blocks for sending back to
the client. Then build the block list in the stats engine for returning.
svn path=/icecast/trunk/icecast/; revision=14812
new listeners with the reason logged. auth_t refcount was getting out of sync
which is a potential small memory leak.
svn path=/icecast/trunk/icecast/; revision=14788
details to stats engine. On YP remove, make sure we find all nodes for that
montpoint not just the first. Allow at least 20mins from a failed touch before
next add, allows YP to give a longer wait period for misconfigured setups by
issuing a TouchFreq header on rejection.
svn path=/icecast/trunk/icecast/; revision=14787
Don't add to stream directory until a minute has passed, it seems like a few sites
are triggering YP requests because they are connecting and then dropping (probably
due to timeout), so at least make sure the stream is running a little while.
svn path=/icecast/trunk/icecast/; revision=14777
was very easy for the intermediate files to clash between projects. The newer
libxml2/libxslt/libcurl also were causing some threading issues due to
library initialization
svn path=/icecast/trunk/icecast/; revision=14465
provide a mountpoint so you specify a shoutcast-mount in the listen-socket
section. Specifying this will automatically define the port+1 with the same
settings however the original shoutcast-compat approach is maintained.
svn path=/icecast/trunk/icecast/; revision=14045
ISO-8859-1 is assumed as there is no real clarity wrt the spec. In most cases people
send ASCII so it's not an issue, but for some, the extended characters they send
can cause problems with XML processing. As stats and YP require UTF8 we need to
translate them and block invalid cases.
For the moment, for non-Ogg streams only, we assume that the metadata needs converting
from ISO-8859-1. Ogg streams are UTF8 so no conversion needed. You can override the
default with a charset mount option.
svn path=/icecast/trunk/icecast/; revision=13615
them. Each listener connection for each request is checked as connected before
performing the request (so that time isn't wasted on slow authentication). Various
name/comment cleanups as well.
svn path=/icecast/trunk/icecast/; revision=13583
