Hello,

I am trying to install libnetconf2 as part of netopeer-GUI.
I am running on RedHat7.2 and have installed netopeer-GUI successfully before on the same device.

after updating the libyang and libnetconf2 libraries and following the install steps I am facing an issue whenever I run the 'make' command in libnetconf2

I am getting the below error:

/var/www/netopeer-gui/install/netopeerguid/libnetconf2/src/messages_server.c: In function "nc_err_libyang":

/var/www/netopeer-gui/install/netopeerguid/libnetconf2/src/messages_server.c:471:14: error: "LYVE_INVALATTR" undeclared (first use in this function) case LYVE_INVALATTR:

/var/www/netopeer-gui/install/netopeerguid/libnetconf2/src/messages_server.c:471:14: note: each undeclared identifier is reported only once for each function it appears in make[2]: *** [CMakeFiles/netconf2.dir/src/messages_server.c.o] Error 1 make[1]: *** [CMakeFiles/netconf2.dir/all] Error 2 make: *** [all] Error 2

Let me know if you need any further information from my end.

Thank you for your support.

Maged

Hi,
This requirement is for asynchronous mode support for NETCONF session, now, client must wait for the previous response before sending a new request in one session. (the new request is blocked if the previous response is not received and sent to client.)

i have a yang module ,it has "when" synatx, I find "relove_when" return error, i wonder how to judge this variable(type) value ? or it does not support ?

when "/oc-platform:components/oc-platform:component/" +
"oc-platform:state/oc-platform:type = 'OPTICAL_CHANNEL'"

When condition "/openconfig-platform:components/openconfig-platform:component/openconfig-platform:state/openconfig-platform:type = 'OPTICAL_CHANNEL'" not satisfied.

The test we did had a TLS call-home configuration where the NC client didn't accept the netopeer2-servers certificate (by mistake though), and the client side closed the SSL session. This loop was running continuously every second.
Depending on the timing, the netopeer2-servers side, can fail at any of the following functions with the given error.

nc_read()/nc_write()/nc_read_poll()
...
ERROR: Call Home client "tls_ch_client" session terminated, reconnecting...
VERBOSE: Trying to connect to 192.168.16.2.
VERBOSE: Successfully connected to 192.168.16.2:4335 over IPv4.
netopeer2-server[3900]: Resolving unresolved data nodes and their constraints...
netopeer2-server[3900]: All data nodes and constraints resolved.
ERROR: Call Home client "tls_ch_client" session 242 established.
ERROR: Session 242: communication socket unexpectedly closed (OpenSSL).
netopeer2-server[3900]: Session 242: thread 3 event session terminated.
netopeer2-server[3900]: Generated new event (netconf-session-start).
netopeer2-server[3900]: Generated new event (netconf-session-end).
ERROR: Call Home client "tls_ch_client" session terminated, reconnecting...
VERBOSE: Trying to connect to 192.168.16.2.
VERBOSE: Successfully connected to 192.168.16.2:4335 over IPv4.
....
VERBOSE: Trying to connect to 192.168.16.2.
VERBOSE: Successfully connected to 192.168.16.2:4335 over IPv4.
netopeer2-server[3900]: Resolving unresolved data nodes and their constraints...
netopeer2-server[3900]: All data nodes and constraints resolved.
ERROR: Call Home client "tls_ch_client" session 325 established.
ERROR: Session 325: communication socket unexpectedly closed (OpenSSL).
netopeer2-server[3900]: Session 325: thread 2 event session terminated.
netopeer2-server[3900]: Generated new event (netconf-session-start).
netopeer2-server[3900]: Generated new event (netconf-session-end).
ERROR: Call Home client "tls_ch_client" session terminated, reconnecting...
VERBOSE: Trying to connect to 192.168.16.2.
VERBOSE: Successfully connected to 192.168.16.2:4335 over IPv4.
netopeer2-server[3900]: Resolving unresolved data nodes and their constraints...
netopeer2-server[3900]: All data nodes and constraints resolved.
ERROR: Session 326: reading from the TLS session failed (SSL code 5).
ERROR: Failed to SSL_shutdown
VERBOSE: Trying to connect to 192.168.16.2.
VERBOSE: Successfully connected to 192.168.16.2:4335 over IPv4.
netopeer2-server[3900]: Resolving unresolved data nodes and their constraints...
netopeer2-server[3900]: All data nodes and constraints resolved.
ERROR: Call Home client "tls_ch_client" session 327 established.
ERROR: Session 327: communication socket unexpectedly closed (OpenSSL).
netopeer2-server[3900]: Session 327: thread 2 event session terminated.

....
this continues....

This is acceptable behaviour.
However, we observed the netopeer2-server has a number of opened sockets:

 2497 root     23712 S    /usr/sbin/netopeer2-server -v 1

root@box:/proc/2497/fd# for i in `seq 10`; do fds=`ls -1 | wc -l`; echo "FD count $fds"; sleep 2; done
FD count 667
FD count 669
FD count 669
FD count 670
FD count 672
FD count 672
FD count 673
FD count 675
FD count 675
FD count 676

root@box:/proc/2497/fd# ls -l | head -n 50
lrwx------    1 root     root            64 Mar 19 12:26 0 -> /dev/null
lrwx------    1 root     root            64 Mar 19 12:26 1 -> /dev/null
lrwx------    1 root     root            64 Mar 19 12:26 10 -> socket:[2826]
lrwx------    1 root     root            64 Mar 19 12:26 100 -> socket:[5035]
lrwx------    1 root     root            64 Mar 19 12:26 101 -> socket:[5057]
lrwx------    1 root     root            64 Mar 19 12:26 102 -> socket:[5061]
lrwx------    1 root     root            64 Mar 19 12:26 103 -> socket:[5062]
lrwx------    1 root     root            64 Mar 19 12:26 104 -> socket:[5073]
lrwx------    1 root     root            64 Mar 19 12:26 105 -> socket:[5074]
lrwx------    1 root     root            64 Mar 19 12:26 106 -> socket:[5075]
lrwx------    1 root     root            64 Mar 19 12:26 107 -> socket:[5089]
lrwx------    1 root     root            64 Mar 19 12:26 108 -> socket:[5090]
lrwx------    1 root     root            64 Mar 19 12:26 109 -> socket:[5091]
lrwx------    1 root     root            64 Mar 19 12:26 11 -> socket:[2829]
lrwx------    1 root     root            64 Mar 19 12:26 110 -> socket:[5102]
lrwx------    1 root     root            64 Mar 19 12:26 111 -> socket:[5103]
lrwx------    1 root     root            64 Mar 19 12:26 112 -> socket:[5123]
lrwx------    1 root     root            64 Mar 19 12:26 113 -> socket:[5129]
...
lrwx------    1 root     root            64 Mar 19 12:26 140 -> socket:[5359]
lrwx------    1 root     root            64 Mar 19 12:26 141 -> socket:[5369]
lrwx------    1 root     root            64 Mar 19 12:26 142 -> socket:[5372]

This count eventually grew upto 1024 & the system was frozen (no more fds could be opened).

Some analysis:
After above failure in SSL layer, the server goes to nc_session_free(struct nc_session *session, void (*data_free)(void *)). This function cleans the SSL context, however it doesn't closes the socket which is set in the session->ti.tls by SSL_set_fd(). Then the call-home thread kiks in and nc_sock_connect(const char* host, uint16_t port) create another socket.

I have created the pull request: #52. Still, there can be other reasons why this is done this way, so I'm open for changes.

Regards,
Darshana.

Will I be able to use libnetconf2 on Mac OS X? Is there intent to support this on Mac OS X?

Thanks,

Is anybody using this or can provide an example of how to use? Or tell me if I'm using it inappropriately. I am trying to establish a NETCONF session between a server and a client on the same machine.

In my client, when I replace a working SSH connect:

    s = nc_connect_ssh("localhost", 0, NULL);

with:

    int sock[2];
    socketpair(AF_UNIX, SOCK_STREAM, 0, sock);
    s = nc_connect_inout(sock[0], sock[1], NULL);

NULL is returned and libnetconf2 reports: Missing <session-id> in server's <hello>. Debug logging seems to show messages flowing in both directions, it just looks like the server reply fails validation. Any information is appreciated. Thanks,

-- Kevin

Hi michalvasko,
Can you merge change "server session FEATURE allow to define an SSH passwd auth callback" to master?
We will use this function.

Thanks, Dongyang

Client side is able to get schemas in different ways:

1. via get-schema from the server
2. from a local directory
3. via callback (#27)

It would be usefult to set a different priorities to these methods to change libnetconf behavior.

Client side code currently parses <hello> content to get information about the server's schemas and to create libyang context. This way all the schemas are implemented on the client side while they are just imported on the server side.

If the server supports ietf-yang-library the primary way to get info about the server's schemas should be to get the yang-library data and create context according to it. yang-library data include information if the schema is imported or implemented so the context on the client side will be really the same as on the server side.

From the devel branch v 0.9.9.

ly_verb is defined as
void ly_verb(LY_LOG_LEVEL level);

so the following two lines produce errors

libnetconf2/src/src/session_client.c: In function 'nc_ctx_fill_cpblts':
libnetconf2/src/src/session_client.c:481:18: error: void value not ignored as it ought to be
verb = ly_verb(LY_LLSILENT);
^
libnetconf2/src/src/session_client.c: In function 'nc_ctx_fill_yl':
libnetconf2/src/src/session_client.c:676:18: error: void value not ignored as it ought to be
verb = ly_verb(LY_LLSILENT);
^

While trying to compile libnetconf2 using OpenSSL 1.1:

Downloads/libnetconf2-master/src/session.c: In function ‘nc_thread_destroy’:
Downloads/libnetconf2-master/src/session.c:1210:5: warning: ‘ERR_remove_thread_state’ is deprecated (declared at /usr/local/include/openssl/err.h:246) [-Wdeprecated-declarations]
Downloads/libnetconf2-master/src/session.c: At top level:
Downloads/libnetconf2-master/src/session.c:1049:1: warning: ‘tls_dyn_create_func’ defined but not used [-Wunused-function]
Downloads/libnetconf2-master/src/session.c:1064:1: warning: ‘tls_dyn_lock_func’ defined but not used [-Wunused-function]
Downloads/libnetconf2-master/src/session.c:1076:1: warning: ‘tls_dyn_destroy_func’ defined but not used [-Wunused-function]
[ 33%] Building C object CMakeFiles/netconf2.dir/src/session_client.c.o
[ 38%] Building C object CMakeFiles/netconf2.dir/src/session_server.c.o
[ 44%] Building C object CMakeFiles/netconf2.dir/src/time.c.o
[ 50%] Building C object CMakeFiles/netconf2.dir/src/session_client_ssh.c.o
[ 55%] Building C object CMakeFiles/netconf2.dir/src/session_server_ssh.c.o
[ 61%] Building C object CMakeFiles/netconf2.dir/src/session_client_tls.c.o
Downloads/libnetconf2-master/src/session_client_tls.c: In function ‘tlsauth_verify_callback’:
Downloads/libnetconf2-master/src/session_client_tls.c:40:20: error: storage size of ‘store_ctx’ isn’t known
Downloads/libnetconf2-master/src/session_client_tls.c:41:17: error: storage size of ‘obj’ isn’t known
Downloads/libnetconf2-master/src/session_client_tls.c:74:9: warning: ‘X509_CRL_get_nextUpdate’ is deprecated (declared at /usr/local/include/openssl/x509.h:709) [-Wdeprecated-declarations]
Downloads/libnetconf2-master/src/session_client_tls.c:80:13: warning: implicit declaration of function ‘X509_OBJECT_free_contents’ [-Wimplicit-function-declaration]
Downloads/libnetconf2-master/src/session_client_tls.c:116:41: error: dereferencing pointer to incomplete type
Downloads/libnetconf2-master/src/session_client_tls.c:41:17: warning: unused variable ‘obj’ [-Wunused-variable]
Downloads/libnetconf2-master/src/session_client_tls.c:40:20: warning: unused variable ‘store_ctx’ [-Wunused-variable]
Downloads/libnetconf2-master/src/session_client_tls.c: In function ‘nc_client_tls_update_opts’:
Downloads/libnetconf2-master/src/session_client_tls.c:396:9: warning: ‘TLSv1_2_client_method’ is deprecated (declared at /usr/local/include/openssl/ssl.h:1610) [-Wdeprecated-declarations]
Downloads/libnetconf2-master/src/session_client_tls.c:434:24: error: dereferencing pointer to incomplete type
make[2]: *** [CMakeFiles/netconf2.dir/src/session_client_tls.c.o] Error 1
make[1]: *** [CMakeFiles/netconf2.dir/all] Error 2
make: *** [all] Error 2
[ 5%] Building C object CMakeFiles/netconf2.dir/src/session_client_tls.c.o
Downloads/libnetconf2-master/src/session_client_tls.c: In function ‘tlsauth_verify_callback’:
Downloads/libnetconf2-master/src/session_client_tls.c:40:20: error: storage size of ‘store_ctx’ isn’t known
Downloads/libnetconf2-master/src/session_client_tls.c:41:17: error: storage size of ‘obj’ isn’t known
Downloads/libnetconf2-master/src/session_client_tls.c:74:9: warning: ‘X509_CRL_get_nextUpdate’ is deprecated (declared at /usr/local/include/openssl/x509.h:709) [-Wdeprecated-declarations]
Downloads/libnetconf2-master/src/session_client_tls.c:80:13: warning: implicit declaration of function ‘X509_OBJECT_free_contents’ [-Wimplicit-function-declaration]
Downloads/libnetconf2-master/src/session_client_tls.c:116:41: error: dereferencing pointer to incomplete type
Downloads/libnetconf2-master/src/session_client_tls.c:41:17: warning: unused variable ‘obj’ [-Wunused-variable]
Downloads/libnetconf2-master/src/session_client_tls.c:40:20: warning: unused variable ‘store_ctx’ [-Wunused-variable]
Downloads/libnetconf2-master/src/session_client_tls.c: In function ‘nc_client_tls_update_opts’:
Downloads/libnetconf2-master/src/session_client_tls.c:396:9: warning: ‘TLSv1_2_client_method’ is deprecated (declared at /usr/local/include/openssl/ssl.h:1610) [-Wdeprecated-declarations]
Downloads/libnetconf2-master/src/session_client_tls.c:434:24: error: dereferencing pointer to incomplete type
make[2]: *** [CMakeFiles/netconf2.dir/src/session_client_tls.c.o] Error 1
make[1]: *** [CMakeFiles/netconf2.dir/all] Error 2
make: *** [all] Error 2

Hi

Its currently not possible to save a private key to a tmp file.
The problem seems to be in:

https://github.com/CESNET/libnetconf2/blob/master/src/session_server_ssh.c#L56

here the file is open with the "read" option but some lines below the code is trying to write to the file.

Best Regards
Nuno

Working from devel stream.

A YANG file with embedded XML in the text strings is sent over get-schema YANG encoding without XML escapes. This causes the client to fail with a malformed XML.

I am trying to run " libnetconf2/python/examples/editconfig.py" but getting below issue,please help me in this.I am able to run get & serverinfo python examples:

~/libnetconf2/python/examples$ ./editconfig1.py
hostname: localhost
port : 830
username: vagrant
vagrant@localhost password :
libyang[1]: Previous warning generated by XPath subexpression[3] "map-type = 'ietf-x509-cert-to-name:specified'".
Traceback (most recent call last):
File "./editconfig1.py", line 66, in
print(data.print_mem(ly.LYD_XML, ly.LYP_FORMAT | ly.LYP_WITHSIBLINGS))
NameError: name 'data' is not defined

If the dependent library paths are different than default, SSH_LIBRARY library path should be included before loading -lssh_threads library.

File : CMakeLists.txt

dependencies - libssh
if(ENABLE_SSH)
find_package(LibSSH 0.6.4 REQUIRED)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DNC_ENABLED_SSH ${LIBSSH_DEFINITIONS}")
target_link_libraries(netconf2 ${LIBSSH_LIBRARIES} -lssh_threads -lcrypt)
include_directories(${LIBSSH_INCLUDE_DIRS})
endif()

dependencies - libssh
if(ENABLE_SSH)
find_package(LibSSH 0.6.4 REQUIRED)
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DNC_ENABLED_SSH ${LIBSSH_DEFINITIONS}")
target_link_libraries(netconf2 ${LIBSSH_LIBRARIES} -L${SSH_LIBRARY} -lssh_threads -lcrypt)
include_directories(${LIBSSH_INCLUDE_DIRS})
endif()

It is possible for server to modify its supported schemas (YANG 1.1 RFC) and the client may also want to support the new context. libnetconf2 does not allow this currently.

The implementation can add one new function (*_update), which would compare its current module-set-id with the one received from the server and if they would not match, try to update the context. If an application required it already, it can implement this completely outside libnetconf2 (it can access the context). Nevertheless, libnetconf2 already implements several ways of retrieving schemas, at least some of which the application would duplicate.

I tried to write the main function

ly_ctx = ly_ctx_new(TESTS_DIR"/../schemas");
nc_server_init(ly_ctx);
snode = ly_ctx_get_node(ly_ctx, NULL, "/ietf-netconf:get");
nc_set_rpc_callback(snode, op_get);
ly_ctx_get_module(....) and start ssh server.

I use the MG-SOFT client to connect ssh server.
client send rpc package:

<rpc xmlns="urn:ietf:params:xml:ns:netconf:base:1.0" essage-id="2">
  <get>
    <filter>
      <yanglib:modules-state xmlns:yanglib="urn:ietf:params:xml:ns:yang:ietf-yang-library"/>
    </filter>
  </get>
</rpc>

in op_get function:

lyd_print_file(stdout, root, LYD_XML, LYP_WITHSIBLINGS);

print:

<get xmlns="urn:ietf:params:xml:ns:netconf:base:1.0">
<data>
<modules-state xmlns="urn:ietf:params:xml:ns:yang:ietf-yang-library">
<module>
<name>ietf-yang-metadata</name>
<revision>2016-08-05</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-yang-metadata</namespace>
<conformance-type>import</conformance-type>
</module>

<module>
<name>yang</name>
<revision>2017-02-20</revision>
<namespace>urn:ietf:params:xml:ns:yang:1</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>ietf-inet-types</name>
<revision>2013-07-15</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-inet-types</namespace>
<conformance-type>import</conformance-type>
</module>

<module>
<name>ietf-yang-types</name>
<revision>2013-07-15</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-yang-types</namespace>
<conformance-type>import</conformance-type>
</module>

<module>
<name>ietf-yang-library</name>
<revision>2016-06-21</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-yang-library</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>ietf-netconf</name>
<revision>2011-06-01</revision>
<namespace>urn:ietf:params:xml:ns:netconf:base:1.0</namespace>
<feature>writable-running</feature>
<feature>candidate</feature>
<feature>rollback-on-error</feature>
<feature>validate</feature>
<feature>startup</feature>
<feature>xpath</feature>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>ietf-netconf-monitoring</name>
<revision>2010-10-04</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-netconf-monitoring</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>ietf-netconf-with-defaults</name>
<revision>2011-06-01</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-netconf-with-defaults</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>notifications</name>
<revision>2008-07-14</revision>
<namespace>urn:ietf:params:xml:ns:netconf:notification:1.0</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>nc-notifications</name>
<revision>2008-07-14</revision>
<namespace>urn:ietf:params:xml:ns:netmod:notification</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>ietf-netconf-notifications</name>
<revision>2012-02-06</revision>
<namespace>urn:ietf:params:xml:ns:yang:ietf-netconf-notifications</namespace>
<conformance-type>implement</conformance-type>
</module>

<module>
<name>if</name>
<revision>2017-06-21</revision>
<schema>file:///opt/libnetconf2/schemas/if.yang</schema>
<namespace>utstar:tn705e:if</namespace>
<conformance-type>implement</conformance-type>
</module>

<module-set-id>13</module-set-id>
</modules-state>
</data>
</get>

after I call nc_server_reply_data(root, nc_wd, NC_PARAMTYPE_FREE) server SSH hangs.

I do not know what's wrong. can you help me? thank you!

Hello,

If you send an invalid RPC such as the one below, the latest libnetconf2 crashes.

<?xml version="1.0" encoding="utf-8"?>
<rpc xmlns="urn:ietf:params:xml:ns:netconf:base:1.0" message-id="5">
  <blah-blah/>
</rpc>

The error seems to originate in src/messages_server.c:470 where it attempts to get the error message path. Call to ly_errpath can return NULL which isn't verified in the switch case for LYVE_INELEM which crashes with SIGSEGV.

For now, I've patched those lines with:

        case LYVE_INELEM:
            str = ly_errmsg(ctx);
            if (str) {
                if (strstr(str, "Unknown")) {
                    e = nc_err(NC_ERR_UNKNOWN_ELEM, NC_ERR_TYPE_PROT, ly_errpath(ctx));
                } else if (strstr(str, "Unsupported")) {
                    e = nc_err(NC_ERR_OP_NOT_SUPPORTED, NC_ERR_TYPE_APP);
                    return e;
                }
            } else {
                e = nc_err(NC_ERR_UNKNOWN_ELEM, NC_ERR_TYPE_PROT, ly_errpath(ctx));
            }
            break;

I could be wrong. Could you please have a look at it instead?

Thanks,
Anandu

Is there interest in supporting other transports such as QUIC?

Summary:
If the client uses a prefix for the NETCONF namespace then the rpc-reply element tag is not in the NETCONF namespace

Details:
The Python library ncclient sends an rpc beginning as follows. Note that there is an 'nc' prefix on the rpc tag:
<nc:rpc xmlns:nc="urn:ietf:params:xml:ns:netconf:base:1.0" message-id="urn:uuid:b9e28b4b-6bf2-4606-9e5a-9bc08074d609"><nc:edit-config>

The reply from libnetconf2 is as follows:
<rpc-reply xmlns:nc="urn:ietf:params:xml:ns:netconf:base:1.0" message-id="urn:uuid:b9e28b4b-6bf2-4606-9e5a-9bc08074d609"><ok/></rpc-reply>

Note that there is no 'nc' prefix on the rpc-reply tag. The response is not recognized and ncclient times out. I believe that it is because the rpc-reply element tag is not in the NETCONF namespace.

I think that the correct response would need to have a prefix on the rpc-reply tag:
<nc:rpc-reply xmlns:nc="urn:ietf:params:xml:ns:netconf:base:1.0"

NETCONF seems to require that the rpc attributes are echoed back verbatim.

I made an attempt to fix it, however I don't suppose my change is very clean. I have done limited testing so far, but I was able to get further with ncclient.

Thanks
fix_reply_namespace .patch.txt

As per my understanding, at many places monotonic clock should be used instead of real time clock.

Like in the main loop of nc_ps_poll. Timeouts are affected using real time clock when changing TOD of the system.

Also, last_rpc & idle timeout calculations for call home connections should also be using monotonic clocks.

I'm saying this because I was using a very slightly modified version of nc_ps_poll in my code. On changing the system time backwards (say by 2 hours), thread having nc_ps_lock got stuck in a loop for 2 hours, and other threads were deprived of the lock during that time.
Changing to monotonic clock in nc_ps_poll solved the issue.

I thought to change nc_gettimespec to use monotonic clock, but it's being used for pthread_mutex_timedlock & pthread_cond_timedwait which need absolute time.

So selective changes need to be made.

I have a netconf server which listens on port 830, and accept client connection and process RPC.
For first session nc_ps_add_session() is successfull and session is added to the poll context, but nc_ps_add_session() is failing always with timeout error for second session.

SESSION 1 :

New incomming session -->
Received Hello message from client. Adding session to poll context.
Successfully added a new session to poll context.

SESSION 2:
New incomming session
Received Hello message from client. Adding session to poll context.
ERROR: nc_ps_add_session: failed to wait for a pollsession condition (Connection timed out).
Failed to add new session to poll context.Retrying..
ERROR: nc_ps_add_session: failed to wait for a pollsession condition (Connection timed out).
Failed to add new session to poll context.Retrying..
ERROR: nc_ps_add_session: failed to wait for a pollsession condition (Connection timed out).
Failed to add new session to poll context.Retrying..
ERROR: nc_ps_add_session: failed to wait for a pollsession condition (Connection timed out).
Failed to add new session to poll context.Retrying..
ERROR: nc_ps_add_session: failed to wait for a pollsession condition (Connection timed out).
Failed to add new session to poll context.Retrying..
Failed to add new session to poll context. Clossing SSh session.

EDIT:
If the issue is fixed, please point the commit ID .

Hello,

I am facing a couple of issues in parsing a module to connect to a NETCONF server using netopeergui.

The capability is "urn:yumaworks:params:xml:ns:netconf:config-id?id=401", which according to my understanding is allowed per RFC 6020.

However whenever I try to connect I receive the following error: "lib ERROR: Unknown capability "urn:yumaworks:params:xml:ns:netconf:config-id?id=401" could not be parsed.

After some basic tracing methods I found out the this error occurs in the session_client.c script, which dictates that the capability has to have "module= " in its body, which is not the case. This consequently interrupts the whole connection process.

Is there a way that I can edit/overwrite that without causing other issues that might occur accordingly?

Thanks,

Maged

Hi,
I have met a question about handling RPC request. I want to use one common callback for all my RPC requests (in my case, just dispatch all RPC request to sysrepo with sr_rpc_send). But now I must set the callback function pointer to lys_node for all RPC, so I think we should allow users to give a global callback for handling RPC request, and if this global callback is not set, we can get callback function from lys_node to make sure current API still works.
If you agree with this idea, I can work on it and push the code to you.

CMake Error at CMakeLists.txt:36 (include):
include could not find load file:

CPackConfig.cmake

am I just missing a step that creates this file or is it missing?

hi，I just learned libnetconf2 library. so I do not know how to make it support：
writable-running
:candidate
:confirmed-commit:1.1

can you help me？thanks！

Hi, it seems that the recent commit which fixes handlig of decimal64 list keys actually works. Here's how a get from the CLI ends up:

Thread 6 "netopeer2-serve" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xb4afeb40 (LWP 1078)]
0xb7668b60 in resolve_partial_json_data_nodeid (nodeid=0xb6514c20 "/czechlight-roadm-v2:roadm/configuration/channel[center-frequency='195.3']/center-frequency", 
    llist_value=0xb4afe05c "195.3", '0' <repeats 12 times>, "1136868377216160297393798828125", '0' <repeats 152 times>..., start=0x30303030, options=1, 
    parsed=0xb4afdf9c) at /home/jkt/work/prog/buildroot/out-epia-geode/build/libyang-custom/src/resolve.c:2230
2230    /home/jkt/work/prog/buildroot/out-epia-geode/build/libyang-custom/src/resolve.c: No such file or directory.
(gdb) bt
#0  0xb7668b60 in resolve_partial_json_data_nodeid (nodeid=0xb6514c20 "/czechlight-roadm-v2:roadm/configuration/channel[center-frequency='195.3']/center-frequency", 
    llist_value=0xb4afe05c "195.3", '0' <repeats 12 times>, "1136868377216160297393798828125", '0' <repeats 152 times>..., start=0x30303030, options=1, 
    parsed=0xb4afdf9c) at /home/jkt/work/prog/buildroot/out-epia-geode/build/libyang-custom/src/resolve.c:2230
#1  0xb76d234e in lyd_new_path (data_tree=0x30303030, ctx=0x88c0420, 
    path=0xb6514c20 "/czechlight-roadm-v2:roadm/configuration/channel[center-frequency='195.3']/center-frequency", value=0xb4afe05c, 
    value_type=LYD_ANYDATA_CONSTSTRING, options=1) at /home/jkt/work/prog/buildroot/out-epia-geode/build/libyang-custom/src/tree_data.c:1229
#2  0x08057951 in opget_build_subtree_from_sysrepo (ds=0x892d320, root=0xb4afe110, subtree_xpath=0xb6503520 "/czechlight-roadm-v2:roadm")
    at /home/jkt/work/prog/buildroot/out-epia-geode/build/netopeer2-server-custom/server/op_get_config.c:62
#3  0x30303030 in ?? ()
#4  0x30303030 in ?? ()
#5  0x30303030 in ?? ()
#6  0x30303030 in ?? ()
#7  0x30303030 in ?? ()
#8  0x30303030 in ?? ()
#9  0x30303030 in ?? ()
#10 0x30303030 in ?? ()
#11 0x30303030 in ?? ()
#12 0x30303030 in ?? ()
#13 0x30303030 in ?? ()
#14 0x30303030 in ?? ()
#15 0x30303030 in ?? ()
#16 0x30303030 in ?? ()
#17 0x30303030 in ?? ()
#18 0x30303030 in ?? ()
#19 0x30303030 in ?? ()
#20 0x30303030 in ?? ()
#21 0x30303030 in ?? ()
#22 0x30303030 in ?? ()
#23 0x30303030 in ?? ()
#24 0x30303030 in ?? ()
#25 0x30303030 in ?? ()
#26 0x30303030 in ?? ()
#27 0x30303030 in ?? ()
#28 0x00000000 in ?? ()

A major part of the stack trace looks like garbage.

This is with freshly rebuild libyang/sysrepo/libnetconf2/netopeer2 from devel branches.

Hi ,
we find all netconf packet without the following data:

can you add the data before all neconf packet.

RFC 6241's example describe as :
The following example shows a capability exchange. Data sent by the
NETCONF client are marked with "C:", and data sent by the NETCONF
server are marked with "S:".
S:
S:
S:
S:
S: urn:ietf:params:netconf:base:1.1
S:
S:
S: urn:ietf:params:ns:netconf:capability:startup:1.0
S:
S:
S: 4
S:
S: ]]>]]>
C:
C:
C:
C:
C: urn:ietf:params:netconf:base:1.1
C:
C:
C:
C: ]]>]]>

Thanks~

I was trying to get USERAUTH BANNER on login.

The code has an API named nc_server_ssh_endpt_set_banner which sets SSH banner for the Netconf server to send to a client when connecting. The API doesn't seems to work in my application.
Attached are client-side logs with debug logging. I set the banner string as This is a banner!!

Also, as I searched over the internet for underlying library (libssh) used here, it does not have support for the warning/MOTD/issue banner. This is based on the absence of any code for sending of the SSH_MSG_USERAUTH_BANNER / SSH2_MSG_USERAUTH_BANNER message from the server.
Libssh seems to use “banner” to mean protocol version (i.e. the string initially exchanged by client and server to determine compatibility SSHv1/SSHv2). Though the same has been supported in libssh2.

If this feature not covered in libnetconf2, is there any plan to support the same in near future or upgrade the libssh library to libssh2 to make such small feature work or any workaround that I can use here.
Client Logs.txt

libnetconf2 uses <hello> message to announce all YANG modules as NETCONF capabilities. But this should apply only to 1.0 modules. For 1.1 modules, there should be only ietf-yang-library capability with module-set-id parameter and the list of 1.1 schemas should be available via ietf-yang-library data. So, libnetconf2 need to update

1. how it announces list of schemas from the server side
2. how it gets the schemas for the libyang context on the client side

Occasionally on low powered (?) hardware, when nc_session_free signals ch_lock, the timedwait does not exit in a timely manner and nc_session_free completes thus destroying the mutex and condition variable. The pthread_cond_timedwait call in nc_server_ch_client_thread_session_cond_wait never exits and the callhome retry never happens as the session is effectively not closed.

pthread_cond_timedwait behavior when the underlying condition variable is destroyed is undefined.

linux used is 4.1.38 on aarch64 with 2 cores.
libnetconf2 is current to nov 2017. No further commits were found that were relevant to this issue.

This was exhibited with ncclient (master, 0.5.3 is inadequate for callhome) and netopeer2 server where on client program shutdown the session was closed without stopping the tcp listener first.

A possible tested solution is attached.

A better solution may include another cond var/mutex for signalling back to nc_session_free that the session is no longer in use and destruction can occur.

ssh_callhome_race.patch.gz

we are seeing a crash when the call home connection is closed from the other end, the ch_lock and ch_cond parameters are NULL and nc_session_free tries to grab a lock on NULL ch_lock. I am pasting the stack trace below. from the code analysis it looks like the locks are not being allocated for the child sessions on the parent call home session. Please advise.

#0 __GI___pthread_mutex_lock (mutex=0x0) at ../nptl/pthread_mutex_lock.c:65
#1 0x00007ffff76570d6 in nc_session_free (session=0x7fffe4001250, data_free=0x407acf <free_ds>)
at session.c:622
#2 0x0000000000408e7c in np2srv_del_session_clb (session=0x7fffe4001250) at main.c:756
#3 0x000000000040a5fd in worker_thread (arg=0x6e0240) at main.c:1302
#4 0x000000000040b056 in main (argc=3, argv=0x7fffffffe178) at main.c:1541

libnetconf2 seems to handle keyboard-interactive authentication incorrectly, as it responds with a spurious SSH_MSG_USERAUTH_FAILURE when asked for keyboard-interactive authentication.
This results in handshake state error, with the an extra SSH_MSG_USERAUTH_FAILURE being delivered to an authenticated session. While OpenSSH reacts with SSH_MSG_UNIMPLEMENTED, Apache MINA SSHD throws an IllegalStateException and terminates the session.

In nc_ps_poll, line 1541, the rpc object result from nc_server_recv_rpc_io maybe leaked.

    if (ret == NC_PSPOLL_RPC) {
        ret = nc_server_recv_rpc_io(cur_session, timeout, &rpc); // not free the rpc object
        if (ret & (NC_PSPOLL_ERROR | NC_PSPOLL_BAD_RPC)) {
            if (cur_session->status != NC_STATUS_RUNNING) {
                ret |= NC_PSPOLL_SESSION_TERM | NC_PSPOLL_SESSION_ERROR;
                cur_ps_session->state = NC_PS_STATE_INVALID;
            } else {
                cur_ps_session->state = NC_PS_STATE_NONE;
            }
        } else {

Hello,

I am working on installing libnetconf2 on a docker container to effectively install the new Netopeer2GUI.
According to the install instructions, the devel branches have to be used in the time being.

However, whenever I try to install libnetconf2 (after installing libyang) I get the following error:

/libnetconf2/python/rpc.c:22:31: fatal error: libyang/swigpyrun.h: No such file or directory
 #include <libyang/swigpyrun.h>
                               ^
compilation terminated.
error: command 'x86_64-linux-gnu-gcc' failed with exit status 1
make[2]: *** [python/CMakeFiles/pyapi] Error 1
make[1]: *** [python/CMakeFiles/pyapi.dir/all] Error 2
make: *** [all] Error 2

I have attached the Dockerfile container for your reference (remove the '.txt' extension to run in linux).

Dockerfile.txt

You can just run the following command inside the folder with the Dockerfile:

docker pull ubuntu:trusty
docker build -t ubuntu:trusty .

Let me know if you need any further information from my end, or if you would want to raise the issue in libyang

Thank you,

Maged

cpblt->name may be null pointer when receive illegal packet. Similar issues maybe exist some where else.

illegal packet example, missing \r\n in first line.

<?xml version="1.0\" encoding="UTF-8"?>
        <hello xmlns="urn:ietf:params:xml:ns:netconf:base:1.0">\r\n
        <capabilities>\r\n
        <capability>urn:ietf:params:netconf:base:1.1</capability>\r\n
        <capability>urn:ietf:params:netconf:base:1.0</capability>\r\n
        </capabilities>\r\n
        </hello>]]>]]>\r\n

I encountered errors when compiling libnetconf2 on my MAC, it seems that the MAC OSX does not support pthread_rwlock_t and CLOCK_REALTIME. The error log is :

sunlinhuideAir:build sunlinhui$ make
Scanning dependencies of target netconf2
[ 4%] Building C object CMakeFiles/netconf2.dir/src/io.c.o
In file included from /Users/sunlinhui/Documents/workspace/libnetconf2/src/io.c:29:
In file included from /Users/sunlinhui/Documents/workspace/libnetconf2/src/libnetconf.h:22:
/Users/sunlinhui/Documents/workspace/libnetconf2/src/session_p.h:195:5: error:
unknown type name 'pthread_spinlock_t'; did you mean 'pthread_rwlock_t'?
pthread_spinlock_t sid_lock;
^~~~~~~~~~~~~~~~~~
pthread_rwlock_t
/usr/include/sys/_pthread/_pthread_rwlock_t.h:30:35: note: 'pthread_rwlock_t'
declared here
typedef __darwin_pthread_rwlock_t pthread_rwlock_t;
^
/Users/sunlinhui/Documents/workspace/libnetconf2/src/io.c:459:19: warning:
implicit declaration of function 'ppoll' is invalid in C99
[-Wimplicit-function-declaration]
ret = ppoll(&fds, 1, (timeout == -1 ? NULL : &ts_timeout), &...
^
1 warning and 1 error generated.
make[2]: *** [CMakeFiles/netconf2.dir/src/io.c.o] Error 1
make[1]: *** [CMakeFiles/netconf2.dir/all] Error 2
make: *** [all] Error 2

Any idea?

Calling va_start(X) is undefined when the last argument has a type which is incompatible from the type that results from the default argument promotion:

[6/27] Building C object CMakeFiles/netconf2.dir/src/messages_server.c.o
../src/messages_server.c:146:18: warning: passing an object that undergoes default argument promotion to 'va_start' has undefined behavior [-Wvarargs]
    va_start(ap, tag);
                 ^
../src/messages_server.c:126:15: note: parameter of type 'NC_ERR' (aka 'enum NC_ERROR') is declared here
nc_err(NC_ERR tag, ...)
              ^
1 warning generated.
[9/27] Building C object CMakeFiles/netconf2.dir/src/io.c.o
../src/io.c:983:18: warning: passing an object that undergoes default argument promotion to 'va_start' has undefined behavior [-Wvarargs]
    va_start(ap, type);
                 ^
../src/io.c:961:54: note: parameter of type 'NC_MSG_TYPE' (aka 'enum NC_MSG_TYPE') is declared here
nc_write_msg(struct nc_session *session, NC_MSG_TYPE type, ...)

I think that a most elegant fix is specifying the NC_MSG_TYPE enum size to be int; however, that's an obvious ABI break because it changes sizeof(NC_MSG_TYPE).

Some workarounds are possible, including adding another tag argument just for va_args. However, this is impossible if a function is exported because that, too, is an ABI break.

When libnetconf is not able to get a schema used by the server (via get-schema or from a local directory) it should, instead of just printing error/warning, use a callback provided by the caller to get the schema. Caller can have another sources (including user interaction user) from where the schema can be obtained.

There shouldn't be any issue with timeouts because in this phase the NETCONF session is actually already established and just the client side context is being set.

1. connect to the netopeer2-server from a netopeer2-cli over SSH, no password
2. immediately quit
   Versions from devel branches as of this Wednesday.

$ netopeer2-cli 
> connect --host ... --ssh --login root
Interactive SSH Authentication
Type your password:
Password: 
> quit

=================================================================
==18891==ERROR: LeakSanitizer: detected memory leaks

Direct leak of 2712 byte(s) in 32 object(s) allocated from:
    #0 0x7fbaf544db78 in malloc (/usr/lib/gcc/x86_64-pc-linux-gnu/6.3.0/libasan.so.3+0xd2b78)
    #1 0x7fbaf3fe9aad in strndup (/lib64/libc.so.6+0x83aad)
    #2 0x7fbaf4c5bdf1 in parse_cpblts /home/jkt/work/cesnet/gerrit/github/CESNET/libnetconf2/src/session.c:984
    #3 0x7fbaf4c5c7de in nc_recv_client_hello /home/jkt/work/cesnet/gerrit/github/CESNET/libnetconf2/src/session.c:1095
    #4 0x7fbaf4c5cf96 in nc_handshake /home/jkt/work/cesnet/gerrit/github/CESNET/libnetconf2/src/session.c:1201
    #5 0x7fbaf4c80e51 in nc_connect_ssh /home/jkt/work/cesnet/gerrit/github/CESNET/libnetconf2/src/session_client_ssh.c:1526
    #6 0x40d204 in cmd_connect_listen_ssh /home/jkt/work/cesnet/gerrit/github/CESNET/Netopeer2/cli/commands.c:1393
    #7 0x4130cb in cmd_connect_listen /home/jkt/work/cesnet/gerrit/github/CESNET/Netopeer2/cli/commands.c:2444
    #8 0x4132d4 in cmd_connect /home/jkt/work/cesnet/gerrit/github/CESNET/Netopeer2/cli/commands.c:2476
    #9 0x406728 in main /home/jkt/work/cesnet/gerrit/github/CESNET/Netopeer2/cli/main.c:191
    #10 0x7fbaf3f868b3 in __libc_start_main (/lib64/libc.so.6+0x208b3)
    #11 0x405928 in _start (/opt/nc/bin/netopeer2-cli+0x405928)

Hi,
Netconf server does not send capability 'urn:ietf:params:netconf:capability:notification:1.0' even if module 'nc-notifications' and 'notifications' are loaded.

When a client that has opened multiple sessions that share the same SSH connection exits without closing the sessions, the server crashes.

Program terminated with signal SIGSEGV, Segmentation fault.
in nc_session_free (libnetconf2/src/session.c:713).
712                 if (session->flags & NC_SESSION_SSH_MSG_CB) {
713 >>                  for (siter = session->ti.libssh.next; siter->status != NC_STATUS_RUNNING; siter = siter->ti.libssh.next) {
714                         if (siter->ti.libssh.next == session) {
715                             ERRINT;
716                             break;
717                         }

It looks like there are 2 threads handling both sessions and they are calling nc_session_free at the same time. I tried to add locks to prevent an obvious race condition:

diff --git a/src/session.c b/src/session.c
index 0f3b253bdd25..fefa71f8664c 100644
--- a/src/session.c
+++ b/src/session.c
@@ -552,6 +552,13 @@ nc_session_free(struct nc_session *session, void (*data_free)(void *))
         } /* else failed to lock it, too bad */
     }
 
+#ifdef NC_ENABLED_SSH
+    /* avoid concurrent free by multiple threads of sessions that share the
+     * same SSH connection. */
+    if (session->ti_type == NC_TI_LIBSSH)
+        pthread_mutex_lock(session->io_lock);
+#endif
+
     if ((session->side == NC_CLIENT) && (session->status == NC_STATUS_RUNNING)) {
         /* cleanup message queues */
         /* notifications */
@@ -750,6 +757,11 @@ nc_session_free(struct nc_session *session, void (*data_free)(void *))
     lydict_remove(session->ctx, session->username);
     lydict_remove(session->ctx, session->host);
 
+#ifdef NC_ENABLED_SSH
+    if (session->ti_type == NC_TI_LIBSSH)
+        pthread_mutex_unlock(session->io_lock);
+#endif
+
     /* final cleanup */
     if ((session->side == NC_SERVER) && session->opts.server.rpc_lock) {
         if (rpc_locked) {

But another problem is still here:

netopeer2-server[2030]: Session 4: communication socket unexpectedly closed.
netopeer2-server[2030]: Session 4: failed to write reply.

The server exits after printing that message...

Maybe I missed something.

Is there a way to get a callback for oper data using libnetconf2/netopeer2, so far I only see a way to get a callback from sysrepo when a data store config change occurred. If this currently does not exist - will it eventually be available and if so what is the time frame for it and what mechanism will be available to get a call back?

I am unable to compile libnetconf2. The error is as below.

Linking C executable cmTC_a4e41
/usr/bin/cmake -E cmake_link_script CMakeFiles/cmTC_a4e41.dir/link.txt --verbose=1
/usr/bin/cc   -Wall -Wextra -fvisibility=hidden     CMakeFiles/cmTC_a4e41.dir/CheckSymbolExists.c.o  -o cmTC_a4e41 -rdynamic
CMakeFiles/cmTC_a4e41.dir/CheckSymbolExists.c.o: In function `main':
CheckSymbolExists.c:(.text+0x16): undefined reference to `pthread_create'
collect2: error: ld returned 1 exit status
CMakeFiles/cmTC_a4e41.dir/build.make:97: recipe for target 'cmTC_a4e41' failed
gmake[1]: Leaving directory '/home/jayshankar/Downloads/libnetconf2-master/build/CMakeFiles/CMakeTmp'
gmake[1]: *** [cmTC_a4e41] Error 1
Makefile:126: recipe for target 'cmTC_a4e41/fast' failed
gmake: *** [cmTC_a4e41/fast] Error 2

Determining if the function pthread_create exists in the pthreads failed with the following output:
Change Dir: /home/jayshankar/Downloads/libnetconf2-master/build/CMakeFiles/CMakeTmp

Run Build Command:"/usr/bin/gmake" "cmTC_4cb6f/fast"
/usr/bin/gmake -f CMakeFiles/cmTC_4cb6f.dir/build.make CMakeFiles/cmTC_4cb6f.dir/build
gmake[1]: Entering directory '/home/jayshankar/Downloads/libnetconf2-master/build/CMakeFiles/CMakeTmp'
Building C object CMakeFiles/cmTC_4cb6f.dir/CheckFunctionExists.c.o
/usr/bin/cc -Wall -Wextra -fvisibility=hidden -DCHECK_FUNCTION_EXISTS=pthread_create -o CMakeFiles/cmTC_4cb6f.dir/CheckFunctionExists.c.o -c /usr/share/cmake/Modules/CheckFunctionExists.c
Linking C executable cmTC_4cb6f
/usr/bin/cmake -E cmake_link_script CMakeFiles/cmTC_4cb6f.dir/link.txt --verbose=1
/usr/bin/cc -Wall -Wextra -fvisibility=hidden -DCHECK_FUNCTION_EXISTS=pthread_create CMakeFiles/cmTC_4cb6f.dir/CheckFunctionExists.c.o -o cmTC_4cb6f -rdynamic -lpthreads
/usr/bin/ld: cannot find -lpthreads
collect2: error: ld returned 1 exit status
CMakeFiles/cmTC_4cb6f.dir/build.make:97: recipe for target 'cmTC_4cb6f' failed
gmake[1]: Leaving directory '/home/jayshankar/Downloads/libnetconf2-master/build/CMakeFiles/CMakeTmp'
gmake[1]: *** [cmTC_4cb6f] Error 1
Makefile:126: recipe for target 'cmTC_4cb6f/fast' failed
gmake: *** [cmTC_4cb6f/fast] Error 2.

In src/session.c:

#ifndef HAVE_PTHREAD_MUTEX_TIMEDLOCK
int
pthread_mutex_timedlock(pthread_mutex_t *mutex, const struct timespec *abstime)
{
    int32_t diff;
    int rc;
    struct timespec cur, dur;

    /* Try to acquire the lock and, if we fail, sleep for 5ms. */
    while ((rc = pthread_mutex_trylock(mutex)) == EBUSY) {
        nc_gettimespec_real(&cur);

        if ((diff = nc_difftimespec(&cur, abstime)) < 1) {
            /* timeout */
            break;
        } else if (diff < 5) {
            /* sleep until timeout */
            dur = *abstime;
        } else {
            /* sleep 5 ms */
            dur.tv_sec = 0;
            dur.tv_nsec = 5000000;
        }

        nanosleep(&dur, NULL);
    }

    return rc;
}
#endif

abstime passed is time since epoch + some timeout value, i.e.,

gdb$ p *abstime
$8 = {
  tv_sec = 1505354115, 
  tv_nsec = 812837304
}

Now if code goes in else case, we nanosleep for (0s+5000000ns). But if we go inside else if (diff < 5), we sleep for (1505354115s+812837304ns). So it goes in kind of infinite sleep in this else if case.

Shouldn't we be sleeping for dur = *abstime - curtime (ignore the type-casting)?

handshake does not work when multiple sessions are created for the same ssh connection, we are trying to create multiple netconf sessions on the same ssh connection. The first netconf session goes through fine, we can see the capabilities and the hello reply back from server. however, the second session does not send the hello reply message back to the client, the client keeps on waiting for the reply message which never happens. I am using 0.9-r1 release. Please advise. I am pasting the logs from netopeer server to confirm this. I am using the latest git clone of netopeer2 server.

netopeer2-server[9027]: Accepted a connection on 0.0.0.0:830.
[2017/10/13 07:19:42.992085, 1] ssh_server_connection_callback: SSH client banner: SSH-2.0-Ganymed_262
[2017/10/13 07:19:42.992542, 1] ssh_analyze_banner: Analyzing banner: SSH-2.0-Ganymed_262
[2017/10/13 07:19:44.418126, 2] ssh_packet_newkeys: Received SSH_MSG_NEWKEYS
netopeer2-server[9027]: Received an SSH message "request-service" of subtype "ssh-userauth".
netopeer2-server[9027]: Received an SSH message "request-auth" of subtype "none".
netopeer2-server[9027]: Received an SSH message "request-auth" of subtype "password".
netopeer2-server[9027]: User "root" authenticated.
netopeer2-server[9027]: Received an SSH message "request-channel-open" of subtype "session".
netopeer2-server[9027]: Received an SSH message "request-channel" of subtype "subsystem".
netopeer2-server[9027]: Session 4: sending message:
urn:ietf:params:netconf:base:1.0urn:ietf:params:netconf:base:1.1urn:ietf:params:netconf:capability:writable-running:1.0urn:ietf:params:netconf:capability:candidate:1.0urn:ietf:params:netconf:capability:rollback-on-error:1.0urn:ietf:params:netconf:capability:validate:1.1urn:ietf:params:netconf:capability:startup:1.0urn:ietf:params:netconf:capability:xpath:1.0urn:ietf:params:netconf:capability:with-defaults:1.0?basic-mode=explicit&also-supported=report-all,report-all-tagged,trim,expliciturn:ietf:params:netconf:capability:notification:1.0urn:ietf:params:netconf:capability:interleave:1.0
netopeer2-server[9027]: Session 4: sending message:
urn:ietf:params:xml:ns:yang:ietf-yang-metadata?module=ietf-yang-metadata&revision=2016-08-05urn:ietf:params:xml:ns:yang:1?module=yang&revision=2017-02-20urn:ietf:params:xml:ns:yang:ietf-inet-types?module=ietf-inet-types&revision=2013-07-15urn:ietf:params:xml:ns:yang:ietf-yang-types?module=ietf-yang-types&revision=2013-07-15urn:ietf:params:xml:ns:yang:ietf-yang-library?module=ietf-yang-library&revision=2016-06-21&module-set-id=25urn:ietf:params:xml:ns:yang:ietf-netconf-acm?module=ietf-netconf-acm&revision=2012-02-22urn:ietf:params:xml:ns:netconf:base:1.0?module=ietf-netconf&revision=2011-06-01&features=writable-running,candidate,rollback-on-error,validate,startup,xpathurn:ietf:params:xml:ns:yang:ietf-netconf-notifications?module=ietf-netconf-notifications&revision=2012-02-06
netopeer2-server[9027]: Session 4: sending message:
urn:ietf:params:xml:ns:netconf:notification:1.0?module=notifications&revision=2008-07-14urn:ietf:params:xml:ns:netmod:notification?module=nc-notifications&revision=2008-07-14http://example.net/turing-machine?module=turing-machine&amp;revision=2013-12-27urn:ietf:params:xml:ns:yang:ietf-interfaces?module=ietf-interfaces&revision=2014-05-08urn:ietf:params:xml:ns:yang:iana-if-type?module=iana-if-type&revision=2014-05-08urn:ietf:params:xml:ns:yang:ietf-ip?module=ietf-ip&revision=2014-06-16urn:ietf:params:xml:ns:yang:ietf-keystore?module=ietf-keystore&revision=2016-10-31urn:ietf:params:xml:ns:yang:ietf-x509-cert-to-name?module=ietf-x509-cert-to-name&revision=2014-12-10urn:ietf:params:xml:ns:yang:ietf-ssh-server?module=ietf-ssh-server&revision=2016-11-02
netopeer2-server[9027]: Session 4: sending message:
urn:ietf:params:xml:ns:yang:ietf-tls-server?module=ietf-tls-server&revision=2016-11-02urn:ietf:params:xml:ns:yang:ietf-netconf-server?module=ietf-netconf-server&revision=2016-11-02&features=listen,ssh-listen,tls-listen,call-home,ssh-call-home,tls-call-homeurn:ietf:params:xml:ns:yang:iana-crypt-hash?module=iana-crypt-hash&revision=2014-08-06urn:ietf:params:xml:ns:yang:ietf-system?module=ietf-system&revision=2014-08-06&features=authentication,local-usershttp://mavenir.com/ns/MainConfig/1.1?module=MainConfig&amp;revision=2016-08-08urn:ietf:params:xml:ns:yang:ietf-netconf-monitoring?module=ietf-netconf-monitoring&revision=2010-10-04urn:ietf:params:xml:ns:yang:ietf-netconf-with-defaults?module=ietf-netconf-with-defaults&revision=2011-06-014
netopeer2-server[9027]: Session 4: sending message:
]]>]]>
[2017/10/13 07:19:45.986075, 2] grow_window: growing window (channel 43:100) to 1280000 bytes
netopeer2-server[9027]: Session 4: received message:
urn:ietf:params:netconf:base:1.0


netopeer2-server[9027]: Sending session-start request.
netopeer2-server[9027]: session-start request sent, waiting for response.
netopeer2-server[9027]: session-start response received, processing.
netopeer2-server[9027]: Sending event-notification request.
netopeer2-server[9027]: event-notification request sent, waiting for response.
netopeer2-server[9027]: event-notification response received, processing.
netopeer2-server[9027]: Generated new event (netconf-session-start).
netopeer2-server[9027]: Received an SSH message "request-channel-open" of subtype "session".
netopeer2-server[9027]: Received an SSH message "request-channel" of subtype "subsystem".

Hi, I want to compile libnetconf2 on macos using following command.

cmake -DOPENSSL_ROOT_DIR=/usr/local/opt/openssl -DOPENSSL_LIBRARIES=/usr/local/opt/openssl/lib ..
make

CMake can be done with waring CMP0042.
But when I try to make it, shadow.h can't be find.

[  4%] Building C object CMakeFiles/netconf2.dir/src/session_server_ssh.c.o
/Users/dy/Source/CESNET/libnetconf2/src/session_server_ssh.c:23:10: fatal error: 'shadow.h' file not found
#include <shadow.h>
         ^
1 error generated.
make[2]: *** [CMakeFiles/netconf2.dir/src/session_server_ssh.c.o] Error 1
make[1]: *** [CMakeFiles/netconf2.dir/all] Error 2

Hi there,

Looking through the code it is not obvious if subtree filtering is supported by a feature of the server itself, or if the filter anyxml is simply passed to the callback and the callback is supposed to manage the parsing of the filter and apply it.

Some examples or description would be great, if they are available.

dmc

nc_ps_poll -> nc_server_recv_rpc_io
when parsing RPC failed,
ret = NC_PSPOLL_REPLY_ERROR | NC_PSPOLL_BAD_RPC,
rpc is not free