The coverage report shows what code is covered by tests and what is not
and the ratio could be tracked over time. These reports will allow us
to identify code that is not being tested and improve the test suites.
To enable the reports configure with --enable-code-coverage and execute
"make check-code-coverage". The HTML report will be generated in a
subdirectory with name libosmocore-$(PACKAGE_VERSION)-coverage/index.html
The report is generated using gcov, lcov and lcov_cobertura tools and
the OSMO_AC_CODE_COVERAGE macro. The osmo_ax_code_coverage.m4 is a copy of
ax_code_coverage.m4 taken from autoconf-archive v2018.03.13. It was
copied to avoid the additional external dependency and renamed to avoid
overwriting it in case autoconf-archive is already installed as we are
going to install it in $(datadir)/aclocal in order to be reused in other
osmocom's projects.
Closes: OS#1987
Change-Id: I6f4ffb91bd7f3dd070aa09dd16d5ad1faf130a4c
OSMO_GSUP_TO_MSGT_RESULT() is needed by osmo-hlr for osmo_gsup_req.
The others are added for completeness' sake.
Related: I3a8dff3d4a1cbe10d6ab08257a0138d6b2a082d9
Change-Id: I6e38a3bb8447f8f212f8d6f5b10a5d0df59323d7
Provide string escaping that
- returns the required buffer size, so it can be used with OSMO_STRBUF_APPEND().
- uses C compatible string constant escaping sequences.
This is intended as a replacement for all previous osmo_escape_str* and
osmo_quote_str* API. It pains me that I didn't get them right the first nor the
second time:
- The buffer functions do not return the chars needed, which is required for
allocating sufficient memory in the *_c versions of the functions.
- Because of that, these functions are accurately usable for
OSMO_STRBUF_APPEND(), producing truncated strings, for example when dumping a
GSUP message.
- They do not use the C equivalent string constant escaping: for some reason I
thought "\15" would be valid, but it should be "\x0f".
If I could, I would completely drop those mislead implementations ... but
backwards compat prohibits that.
A previous patch already provided internal static functions that accurately
return the required buffer size. Enhance these to also support C compatible
string escaping, and use them as implementation of the new functions:
osmo_escape_cstr_buf()
osmo_escape_cstr_c()
osmo_quote_cstr_buf()
osmo_quote_cstr_c()
In the tests for these, also test C string equivalence.
Naming: from API versions, it would be kind of logical to call them
osmo_escape_str_buf3() and osmo_escape_str_c2(). Since these anyway return a
different escaping, it makes sense to me to have distinct names instead.
Quasi missing are variants of the non-C-compatible weird legacy escaping that
return the required buffer size, but I refrain from adding those, because we
have enough API cruft as it is. Just always use these new cstr variants.
Change-Id: I3dfb892036e01000033dd8e7e4a6a0c32a3caa9b
Although this OSMO_DEPRECATED doesn't seem to generate a warning when compiling
code that sets .is_config_node = foo, it seems a good idea to add the
deprecation tag.
It is deprecated since commit "vty: track parent nodes also for telnet sessions"
I2b32b4fe20732728db6e9cdac7e484d96ab86dc5
Change-Id: I800507b27cb0d536c1a4c203d7f7b90eec05a69c
Keep track of parent nodes and go back hierarchically, not only for .cfg file
reading, but also for telnet VTY sessions.
A long time ago cfg file parsing was made strictly hierarchical: node exits go
back to parent nodes exactly as they were entered. However, live telnet VTY
sessions still lacked this and depended on the go_parent_cb().
From this commit on, implementing a go_parent_cb() is completely optional. The
go_parent_cb() no longer has the task to determine the correct parent node,
neither for cfg files (as already the case before this patch) nor for telnet
VTY sessions (added by this patch). Instead, a go_parent_cb() implementation
can merely take actions it requires on node exits, for example applying some
config when leaving a specific node.
The node value that is returned by the go_parent_cb() and the vty->node and
vty->index values that might be set are completely ignored; instead the
implicit parent node tracking determines the parent and node object.
As a side effect, the is_config_node() callback is no longer needed, since the
VTY now always implicitly knows when to exit back to the CONFIG_NODE.
For example, osmo_ss7_is_config_node() could now be dropped, and the
osmo_ss7_vty_go_parent() could be shortened by five switch cases, does no
longer need to set vty->node nor vty->index and could thus be shortened to:
int osmo_ss7_vty_go_parent(struct vty *vty)
{
struct osmo_ss7_asp *asp;
struct osmo_xua_server *oxs;
switch (vty->node) {
case L_CS7_ASP_NODE:
asp = vty->index;
/* If no local addr was set */
if (!asp->cfg.local.host_cnt) {
asp->cfg.local.host[0] = NULL;
asp->cfg.local.host_cnt = 1;
}
osmo_ss7_asp_restart(asp);
break;
case L_CS7_XUA_NODE:
oxs = vty->index;
/* If no local addr was set, or erased after _create(): */
if (!oxs->cfg.local.host_cnt)
osmo_ss7_xua_server_set_local_host(oxs, NULL);
if (osmo_ss7_xua_server_bind(oxs) < 0)
vty_out(vty, "%% Unable to bind xUA server to IP(s)%s", VTY_NEWLINE);
break;
}
return 0;
}
Before parent tracking, every program was required to write a go_parent_cb()
which has to return every node's parent node, basically a switch() statement
that manually traces the way back out of child nodes. If the go_parent_cb() has
errors, we may wildly jump around the node tree: a common error is to jump
right out to the top config node with one exit, even though we were N levels
deep. This kind of error has been eliminated for cfg files long ago, but still
exists for telnet VTY sessions, which this patch fixes.
This came up when I was adding multi-level config nodes to osmo-hlr to support
Distributed GSM / remote MS lookup: the config file worked fine, while vty node
tests failed to exit to the correct nodes.
Change-Id: I2b32b4fe20732728db6e9cdac7e484d96ab86dc5
Follow up for patch I3cf150cc0cc06dd36039fbde091bc71b01697322
osmo_sockaddr_str_{from,to}_32n actually use host byte order. Deprecate these
and introduce a more accurately named version ending in h.
Change-Id: Ic7fc279bf3c741811cfc002538e28e8f8560e338
When finding a char in a string, I want to be able to limit the search area by
size, not only by nul terminator.
Change-Id: I48f8ace9f51f8a06796648883afcabe3b4e8b537
A couple of times recently I've needed to copy out a substring to a buffer with
limited size. Use of strncpy() or osmo_strlcpy() are nontrivial here.
I wanted to have a dedicated function.
After I wrote that function with a test, I noticed that I had already
implemented the same thing a while ago, as osmo_print_n() :P
So here is just the test.
Change-Id: Ia716abdc1f58af6065b84f4f567388a32a7b39fc
GSUP routing was introduced when adding the E interface. Hence that was the
first realm where routing errors could occur. I did notice back then that this
message type was special: it does not convey a response to a particular message
kind -- it does not make sense, for example, to return an Updating Location
Error cause, and do that for all conceivable message types. Instead, this tells
the sender that a deeper error exists, i.e. that the desired peer is completely
gone and unreachable.
I did not foresee though that for D-GSM, there would also be arbitrary GSUP
proxy routing, and that this error is not limited to E interface semantics.
From today's point of view, adding the "_E_" in the name was a mistake.
Remove that "_E_" to yield OSMO_GSUP_MSGT_ROUTING_ERROR (with unchanged message
type discriminator), but provide a #define linking the old name
OSMO_GSUP_MSGT_E_ROUTING_ERROR to the new one.
The only visible change should be that osmo_gsup_message_type_names[] now
returns the new name without "_E_". I am not aware of any regression test
fallout from that.
Change-Id: Ic8e8bd11522d6c51ac7aaf946516cbce26bc6e1e
The osmo_escape_str_c() and osmo_quote_str_c() functions return truncated
results when characters need escaping. For example:
osmo_quote_str_c(NULL, "foo"); --> "foo"
osmo_quote_str_c(NULL, "foo\n"); --> "foo\n
osmo_quote_str_c(NULL, "foo\tbar\t\n"); --> "foo\tbar\t
Implement these _c variants using OSMO_NAME_C_IMPL() to always allocate
sufficient memory.
However, current osmo_escape_str_buf2() and osmo_quote_str_buf2() fail to
return the required buffer size (even though that information is readily
avaiable), so these don't qualify for accurate use of OSMO_NAME_C_IMPL().
Hence, move the implementations of osmo_escape_str and osmo_quote_str to an
internal static function that returns the characters needed, so that all
dynamically allocating implementations can return un-truncated results.
Of course, external callers would also benefit from escape/quote API that
accurately returns the amount of characters needed, but I am not changing
public API in this patch, on purpose, ... yet.
Change-Id: I16c08eced41bf1b7acf6e95f658068ace99ca4c8
Provide a common implementation for foo_name_c() functions that base on
foo_name_buf() functions.
char *foo_name_c(void *ctx, example_t arg)
{
OSMO_NAME_C_IMPL(ctx, 64, "ERROR", foo_name_buf, arg)
}
Rationale: the most efficient way of composing strings that have optional parts
or require loops for composition is by writing to a ready char[], and this in
turn is easiest done by using OSMO_STRBUF_* API. Using such a basic name string
implementation which typically returns a length, I often want a more convenient
version that returns a char*, which can just be inlined in a "%s" string format
-- crucially: skipping string composition when inlined in a LOGP(). This common
implementation allows saving code dup, only the function signature is needed.
Why not include the function signature in the macro? The two sets of varargs
(1: signature args, 2: function call args) are hard to do. Also, having an
explicit signature is good for readability and code grepping / ctags.
Upcoming uses: in libosmocore in the mslookup (D-GSM) implementation
(osmo_mslookup_result_name_c()), and in osmo_msc's codec negotiation
implementation (sdp_audio_codecs_name_c(), sdp_msg_name_c(), ...).
I54b6c0810f181259da307078977d9ef3d90458c9 (libosmocore)
If3ce23cd5bab15e2ab4c52ef3e4c75979dffe931 (osmo-msc)
Change-Id: Ida5ba8d9640ea641aafef0236800f6d489d3d322
Of course both v4 and v6 addresses are kept in network byte order when
represented in bytes, but when writing, I somehow must have assumed that
inet_pton() returns host byte order. Fix that mixup in the API docs:
osmo_sockaddr_str_from_32() and osmo_sockaddr_str_to_32() actually use network
byte order.
osmo_sockaddr_str_from_32n() and osmo_sockaddr_str_to_32n() actually use host
byte order, though reflecting 'n' in their name.
sockaddr_str_test: use hexdump instead of %x to show the
osmo_sockaddr_str_to_32*() conversions so that the error becomes obvious.
(Printing %x reverses the bytes again and made it look correct.)
Change-Id: I3cf150cc0cc06dd36039fbde091bc71b01697322
It's hard to figure out what color logging categories have with those ANSI
color code strings. Instead, define these OSMO_LOGCOLOR_* constants.
Naming: commonly, the logging.h header has the "LOG" prefix in the name, but it
seems saner to include the OSMO_ prefix: it seems too likely that some
libosmocore user somewhere already has defined "LOGCOLOR_RED" somewhere.
Change-Id: I03b6b1f73ae7ee61d37ff921e071a3d0881d3e9a
Currently planned user: for Distributed GSM in osmo-hlr: setting per-MSC
service addresses in VTY: replace/remove existing entries.
osmo_sockaddr_str_cmp() is useful to catch identical resulting IP addresses,
regardless of differing strings (e.g. '0::' and '::' are equal but differ in
strings).
Change-Id: I0dbc1cf707098dcda75f8e07c1b936951f9f9501
Since Icdeaea67a06da3a2f07b252e455629559ecc1829, we use stderr for
printing warnings while parsing the VTY configuration files. Make
sure we do not close() stderr. Otherwise stderr logging gets broken.
Change-Id: I6ecc85555d102f5911d50ed5ac54933c766fa84d
Fixes: Icdeaea67a06da3a2f07b252e455629559ecc1829
Setting vty->fd to 0 is a bad idea, which may cause the process
to write() warnings to its own _stdin_ (yes, it's possible).
For example, when a configuration file contains deprecated
logging commands. Let's use stderr by default.
Change-Id: Icdeaea67a06da3a2f07b252e455629559ecc1829
We shall not prevent programs from starting if their configuration
files contain deprecated 'logging level ...' commands. Just print
a warning and return CMD_SUCCESS instead of CMD_WARNING.
While writing a unit test, another funny bug has been uncovered.
Parsing of a deprecated command indeed triggers a deprecation
warning, originated from libosmovty's log_deprecated_func().
This function simply calls vty_out(), but...
Since the invocation of the vty_out() happens _before_ the VTY
is initialized, the process is actually writing that warning
to its own stdin! Most likely, because we use talloc_zero()
to allocate a new instance of struct 'vty'.
As a side effect, the evil warning magically appears in the output
of 'make check', breaking the test statistics. Let's work around
this bug for now by redirecting stdin to /dev/null.
Change-Id: Ia934581410cd41594791d4e14ee74c16abe1009a
Fixes: Ic9c1b566ec4a459f03e6319cf369691903cf9d00
Yes, we don't really need to poison stdout, as some osmo-* binaries
(like osmo-gapk) may want to use it for non-logging purposes.
This printf() call looks like a debugging leftover.
Change-Id: Ida35865b1c0bb3d3567918f8e89c6551c6b34103
The format prints IP:port separated by a colon, which of course is confusing
when the IPv6 address itself contains mostly colons. The new format adds square
braces.
cafe:face::1:42 -> [cafe:face::1]:42
The IPv4 format remains unchanged:
1.2.3.4:42
Change-Id: I161f8427729ae31be0eac719b7a4a9290715e37f
So far we had a function to allocate a new bssgp_bvc_ctx, but not
the opposite one. Let's finally introduce it, so it will be used
at least in OsmoPCU.
Please note that the new symbol has 'bssgp_' prefix, not 'btsctx_'.
Change-Id: Ia78979379dbdccd6e4628c16f00d0c06d9212172
In a multi-threaded environemnt, it's likely that each thread will have
its own, distinct set of file descriptors that it wants to watch.
Hence, let's make the osmo_fd_* functions configure not one global
list of file descriptors, but a thread-local list of file descriptors.
Change-Id: I5082ed3e500ad1a7516e1785bc57e008da2fac9a
New fields are put inside a union to keep backward compatibility with
potential older users of the struct.
Change-Id: I235635800c0de47b1e2b9ec9c7191418f6003554
Some compilers don't like declaration of enums in header files like we
do sometimes for structs:
enum gsm_band;
void foobar(enum gsm_band band);
triggers:
error: use of enum 'gsm_band' without previous declaration
Fixes: b99f4ca2d8
Related: OS#4244
Change-Id: I6c2102c763f565bbe3c8dd7e5b4e04c4a45fff67
Often, an IP address of 0.0.0.0 is considered an unset value (for clients
requiring a server address; not for listening on "any").
osmo_sockaddr_str_is_set() does return false when the port is 0, but there is
no simple way to tell whether the IP address is actually set to a server
address.
Add osmo_sockaddr_str_is_nonzero() to return false if:
- the port is zero, or
- the IP address is zero (0.0.0.0 or ::0), or
- the IP address cannot be parsed.
A practical use example: osmo-msc so far accepts an RTP IP address of 0.0.0.0
as valid. I noticed when trying to trigger error handling from a ttcn3 test.
osmo-msc can use this function to reject invalid addresses from MGCP messages.
Related: I53ddb19a70fda3deb906464e1b89c12d9b4c7cbd (osmo-msc)
Change-Id: I73cbcab90cffcdc9a5f8d5281c57c1f87b2c3550
Refuse state changes and event dispatch for FSM instances that are already
terminating.
It is assumed that refusing state changes and events after FSM termination is
seen as the sane expected behavior, hence this change in behavior is merged
without being configurable.
There is no fallout in current Osmocom code trees. fsm_dealloc_test needs a
changed expected output, since it is explicitly creating complex FSM structures
that terminate. Currently no other C test in Osmocom code needs adjusting.
Rationale:
Where multiple FSM instances are collaborating (like in osmo-bsc or osmo-msc),
a terminating FSM instance often causes events to be dispatched back to itself,
or causes state changes in FSM instances that are already terminating. That is
hard to avoid, since each FSM instance could be a cause of failure, and wants
to notify all the others of that, which in turn often choose to terminate.
Another use case: any function that dispatches events or state changes to more
than one FSM instance must be sure that after the first event dispatch, the
second FSM instance is in fact still allocated. Furthermore, if the second FSM
instance *has* terminated from the first dispatch, this often means that no
more actions should be taken. That could be done by an explicit check for
fsm->proc.terminating, but a more general solution is to do this check
internally in fsm.c.
In practice, I need this to avoid a crash in libosmo-mgcp-client, when an
on_success() event dispatch causes the MGCP endpoint FSM to deallocate. The
earlier dealloc-in-main-loop patch fixed part of it, but not all.
Change-Id: Ia81a0892f710db86bd977462730b69f0dcc78f8c
This is a simpler and more general solution to the problem so far solved by
osmo_fsm_term_safely(true). This extends use-after-free fixes to arbitrary
functions, not only FSM instances during termination.
The aim is to defer talloc_free() until back in the main loop.
Rationale: I discovered an osmo-msc use-after-free crash from an invalid
message, caused by this pattern:
void event_action()
{
osmo_fsm_inst_dispatch(foo, FOO_EVENT, NULL);
osmo_fsm_inst_dispatch(bar, BAR_EVENT, NULL);
}
Usually, FOO_EVENT takes successful action, and afterwards we also notify bar.
However, in this particular case, FOO_EVENT caused failure, and the immediate
error handling directly terminated and deallocated bar. In such a case,
dispatching BAR_EVENT causes a use-after-free; this constituted a DoS vector
just from sending messages that cause *any* failure during the first event
dispatch.
Instead, when this is enabled, we do not deallocate 'foo' until event_action()
has returned back to the main loop.
Test: duplicate fsm_dealloc_test.c using this, and print the number of items
deallocated in each test loop, to ensure the feature works. We also verify that
the deallocation safety works simply by fsm_dealloc_test.c not crashing.
We should probably follow up by refusing event dispatch and state transitions
for FSM instances that are terminating or already terminated:
see I0adc13a1a998e953b6c850efa2761350dd07e03a.
Change-Id: Ief4dba9ea587c9b4aea69993e965fbb20fb80e78
Do not return a warning and therefore fail parsing the config when the
"encapsulation framerelay-gre local-ip" command is used and FR/GRE is
disabled. Having this in the config does no harm and allows keeping the
same config if it is enabled later.
This fixes the currently failing vty tests for osmo-sgsn.
Fixes: a0c8195ad3 ("vty: Return error if cmd returns CMD_WARNING while reading cfg file")
Change-Id: Ic225232fbfca49ba868427eaf898e1f6e34e1ca8
The calculation of the beginning of a block for TCH/F, TCH/H and FACCH
can be challenging since those channels are affected by the diagonal
interleaving of the TCH channels. However, GSM 05.02 Section 7 Table 1
of 5 specifies how the blocks are distributed over the TDMA frame
interval. Lets add a mapping function that is based on that table
Related: OS#3803
Change-Id: I3d71c66f8c401f5afbad9b1c86c24580dab9e0ce
This way libosmocore build fails during configuring phase if expected
default behavior (building with libsctp support enabled and providing
osmo_sock_init2_multiaddr() API) fails. User is still provided with
--disable-libsctp option in case he doesn't need those features or his
environment doesn't provide required libsctp APIs.
Change-Id: I710c9cb1c6da0e5fc94b792df8bf60194a72208f
Those two functions are only used by osmo_sock_init2_multiaddr(), which
is only built if HAVE_LIBSCTP is defined. Avoid compiler warning about
unusued function helpers if osmo_sock_init2_multiaddr() is not being
built.
Change-Id: I52769d6b8f70af1a8bda23d60b3230a932e71fab
Since we return error at the start of the function if proto !=
IPPROTO_SCTP, it makes no sense to check for proto != IPPROTO_UDP later
on.
Fixes: CID#205088
Change-Id: Ibba7eacaa9debb77d536d47dc85170c5ee79e479
This API will be used by libosmo-netif's osmo_stream for SCTP sockets,
which in turn will be used by libosmo-sccp to support multi-homed
connections.
Related: OS#3608
Change-Id: Ic8681d9e093216c99c6bca4be81c31ef83688ed1
len provides extra information in the case the buffer was too small,
because it tells the caller "the number of characters (excluding the
terminating null byte) which would have been written to the final
string if enough space had been available" (man
snprintf).
Change-Id: Icafe559e19a92e2ae72fdd0dd2d9a394b1eda878
Without this patch, for instance in this cfg file below, go_parent_cb is
not called for nodes such as "listen" and "cs7":
"""
line vty
no login
cs7 instance 0
xua rkm routing-key-allocation dynamic-permitted
listen m3ua 2905
accept-asp-connections dynamic-permitted
local-ip 127.0.0.1
"""
Related: OS#3608
Change-Id: Ia6d88c0e63d94ba99e950da6efbc4c1871070012
log_enable_multithread() enables use of locks inside the
implementation. Lock use is disabled by default, this way only
multi-thread processes need to enable it and suffer related
complexity/performance penalties.
Locks are required around osmo_log_target_list and items inside it,
since targets can be used, modified and deleted by different threads
concurrently (for instance, user writing "logging disable" in VTY while
another thread is willing to write into that target).
Multithread apps and libraries aiming at being used in multithread apps
should update their code to use the locks introduced here when
containing code iterating over osmo_log_target_list explictly or
implicitly by obtaining a log_target (eg. osmo_log_vty2tgt()).
Related: OS#4088
Change-Id: Id7711893b34263baacac6caf4d489467053131bb
Vasil Velichkov
Pau Espin
Daniel Willmann
Harald Welte
Oliver Smith
Philipp Maier