fdo-mirrors/NetworkManager
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/NetworkManager/NetworkManager.git synced 2025-12-21 11:00:07 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
NetworkManager/libnm/nm-libnm-utils.c

911 lines
25 KiB
C
Raw Permalink Normal View History

all: unify comment style for SPDX-License-Identifier tag Our coding style recommends C style comments (/* */) instead of C++ (//). Also, systemd (which we partly fork) uses C style comments for the SPDX-License-Identifier. Unify the style. $ sed -i '1 s#// SPDX-License-Identifier: \([^ ]\+\)$#/* SPDX-License-Identifier: \1 */#' -- $(git ls-files -- '*.[hc]' '*.[hc]pp')
2020-09-29 16:42:22 +02:00
/* SPDX-License-Identifier: LGPL-2.1+ */
libnm: add testable libnm/nm-libnm-utils.c file Previously, internal parts of libnm were not testable. Instead, add "libnm/nm-libnm-utils.c" and "libnm/libnm-utils.la" to contain code that can be statically linked with a new test "libnm/tests/test-general".
2017-05-19 10:19:25 +02:00
/*
all: unify format of our Copyright source code comments ```bash readarray -d '' FILES < <( git ls-files -z \ ':(exclude)po' \ ':(exclude)shared/c-rbtree' \ ':(exclude)shared/c-list' \ ':(exclude)shared/c-siphash' \ ':(exclude)shared/c-stdaux' \ ':(exclude)shared/n-acd' \ ':(exclude)shared/n-dhcp4' \ ':(exclude)src/systemd/src' \ ':(exclude)shared/systemd/src' \ ':(exclude)m4' \ ':(exclude)COPYING*' ) sed \ -e 's/^\(--\|#\| \*\) *\(([cC]) *\)\?Copyright \+\(\(([cC])\) \+\)\?\(\(20\|19\)[0-9][0-9]\) *[-–] *\(\(20\|19\)[0-9][0-9]\) \+\([^ ].*\)$/\1 C1pyright#\5 - \7#\9/' \ -e 's/^\(--\|#\| \*\) *\(([cC]) *\)\?Copyright \+\(\(([cC])\) \+\)\?\(\(20\|19\)[0-9][0-9]\) *[,] *\(\(20\|19\)[0-9][0-9]\) \+\([^ ].*\)$/\1 C2pyright#\5, \7#\9/' \ -e 's/^\(--\|#\| \*\) *\(([cC]) *\)\?Copyright \+\(\(([cC])\) \+\)\?\(\(20\|19\)[0-9][0-9]\) \+\([^ ].*\)$/\1 C3pyright#\5#\7/' \ -e 's/^Copyright \(\(20\|19\)[0-9][0-9]\) \+\([^ ].*\)$/C4pyright#\1#\3/' \ -i \ "${FILES[@]}" echo ">>> untouched Copyright lines" git grep Copyright "${FILES[@]}" echo ">>> Copyright lines with unusual extra" git grep '\<C[0-9]pyright#' "${FILES[@]}" | grep -i reserved sed \ -e 's/\<C[0-9]pyright#\([^#]*\)#\(.*\)$/Copyright (C) \1 \2/' \ -i \ "${FILES[@]}" ``` https://gitlab.freedesktop.org/NetworkManager/NetworkManager/merge_requests/298
2019-10-01 09:20:35 +02:00
* Copyright (C) 2007 - 2008 Novell, Inc.
* Copyright (C) 2007 - 2018 Red Hat, Inc.
libnm: add testable libnm/nm-libnm-utils.c file Previously, internal parts of libnm were not testable. Instead, add "libnm/nm-libnm-utils.c" and "libnm/libnm-utils.la" to contain code that can be statically linked with a new test "libnm/tests/test-general".
2017-05-19 10:19:25 +02:00
*/
#include "nm-default.h"
#include "nm-libnm-utils.h"
libnm: print timestamp in LIBNM_CLIENT_DEBUG debug logging It's useful, because it's easy to get overwhelemed by the logging output. The timestamp makes it easier to keep track. Also, it allows to see how long things take.
2019-10-29 16:50:08 +01:00
#include "nm-glib-aux/nm-time-utils.h"
libnm: move nm_permission_to_client()/nm_permission_result_to_client() to nm-libnm-utils.c It's nicely trivial and independent. Move it to a separate place, to avoid cluttering the more complicated code and to make it testable. Also, use binary search to find the value by string.
2019-10-29 16:55:15 +01:00
#include "nm-libnm-core-intern/nm-common-macros.h"
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
#include "nm-object.h"
libnm: print timestamp in LIBNM_CLIENT_DEBUG debug logging It's useful, because it's easy to get overwhelemed by the logging output. The timestamp makes it easier to keep track. Also, it allows to see how long things take.
2019-10-29 16:50:08 +01:00
libnm: add testable libnm/nm-libnm-utils.c file Previously, internal parts of libnm were not testable. Instead, add "libnm/nm-libnm-utils.c" and "libnm/libnm-utils.la" to contain code that can be statically linked with a new test "libnm/tests/test-general".
2017-05-19 10:19:25 +02:00
/*****************************************************************************/
libnm: add logging NML_DBUS_LOG*() for debugging D-Bus for NMClient Commonly, a library (like libnm) is not supposed to log anything. Logging is not a suitable way to notify the calling application about anything. When something of importance happens, then the application must be notified via the library's API. However, logging can be very useful for debugging to see what is going on. Add a logging macro that by default does nothing, but can be turned on via an environment variable "LIBNM_CLIENT_DEBUG=debug". Another point is that libnm relies on the server side NetworkManager D-Bus interface to be in an expected manner. For example, we require a D-Bus object "org.freedesktop.NetworkManager" to be present and certain D-Bus interfaces implemented. However libnm should treat NetworkManager as external and untrusted component. That means, we cannot assert against the expectations we have. There are two reasons for this: - a bug in NetworkManager, dbus-daemon or else may cause such errors. This must not trigger an assertion failure in the client application, at least not unless requested. - libnm must be forward and backward compatible against a different NetworkManager server version. That is only possibly by ignoring anything that is unexpected. Asserting by default might prevent to implement API changes, both on libnm and server side. Note that we also don't notify the calling application via dedicated API. On the one hand, these things *can* happen. On the other hand, what would the calling appication do about it anyway? libnm by default must just behave gracefully and pretend all is good. For testing, development and debugging that is however not useful. We want the user to opt in to strict API validation. The user will be able to do that by setting "LIBNM_CLIENT_DEBUG=warning", which causes API violations being logged with g_warning(). These are assertions when running with G_DEBUG=fatal-warnings. This is inspired by GDBus' G_DBUS_DEBUG variable. Note that LIBNM_CLIENT_DEBUG environment variables is undocumented, unstable API. It's used for debugging and testing of the current libnm version at hand. There is no guaranteed stable behavior how a different libnm version might behave.
2019-10-14 08:06:33 +02:00
volatile int _nml_dbus_log_level = 0;
int
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
_nml_dbus_log_level_init(void)
libnm: add logging NML_DBUS_LOG*() for debugging D-Bus for NMClient Commonly, a library (like libnm) is not supposed to log anything. Logging is not a suitable way to notify the calling application about anything. When something of importance happens, then the application must be notified via the library's API. However, logging can be very useful for debugging to see what is going on. Add a logging macro that by default does nothing, but can be turned on via an environment variable "LIBNM_CLIENT_DEBUG=debug". Another point is that libnm relies on the server side NetworkManager D-Bus interface to be in an expected manner. For example, we require a D-Bus object "org.freedesktop.NetworkManager" to be present and certain D-Bus interfaces implemented. However libnm should treat NetworkManager as external and untrusted component. That means, we cannot assert against the expectations we have. There are two reasons for this: - a bug in NetworkManager, dbus-daemon or else may cause such errors. This must not trigger an assertion failure in the client application, at least not unless requested. - libnm must be forward and backward compatible against a different NetworkManager server version. That is only possibly by ignoring anything that is unexpected. Asserting by default might prevent to implement API changes, both on libnm and server side. Note that we also don't notify the calling application via dedicated API. On the one hand, these things *can* happen. On the other hand, what would the calling appication do about it anyway? libnm by default must just behave gracefully and pretend all is good. For testing, development and debugging that is however not useful. We want the user to opt in to strict API validation. The user will be able to do that by setting "LIBNM_CLIENT_DEBUG=warning", which causes API violations being logged with g_warning(). These are assertions when running with G_DEBUG=fatal-warnings. This is inspired by GDBus' G_DBUS_DEBUG variable. Note that LIBNM_CLIENT_DEBUG environment variables is undocumented, unstable API. It's used for debugging and testing of the current libnm version at hand. There is no guaranteed stable behavior how a different libnm version might behave.
2019-10-14 08:06:33 +02:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
const GDebugKey keys[] = {
{"trace", _NML_DBUS_LOG_LEVEL_TRACE},
{"debug", _NML_DBUS_LOG_LEVEL_DEBUG},
{"warning", _NML_DBUS_LOG_LEVEL_WARN},
{"error", _NML_DBUS_LOG_LEVEL_ERROR},
{"stdout", NML_DBUS_LOG_STDOUT},
};
int l;
l = _NML_DBUS_LOG_LEVEL_INITIALIZED
| nm_utils_parse_debug_string(g_getenv("LIBNM_CLIENT_DEBUG"), keys, G_N_ELEMENTS(keys));
if (!g_atomic_int_compare_and_exchange(&_nml_dbus_log_level, 0, l))
l = g_atomic_int_get(&_nml_dbus_log_level);
nm_assert(l & _NML_DBUS_LOG_LEVEL_INITIALIZED);
return l;
libnm: add logging NML_DBUS_LOG*() for debugging D-Bus for NMClient Commonly, a library (like libnm) is not supposed to log anything. Logging is not a suitable way to notify the calling application about anything. When something of importance happens, then the application must be notified via the library's API. However, logging can be very useful for debugging to see what is going on. Add a logging macro that by default does nothing, but can be turned on via an environment variable "LIBNM_CLIENT_DEBUG=debug". Another point is that libnm relies on the server side NetworkManager D-Bus interface to be in an expected manner. For example, we require a D-Bus object "org.freedesktop.NetworkManager" to be present and certain D-Bus interfaces implemented. However libnm should treat NetworkManager as external and untrusted component. That means, we cannot assert against the expectations we have. There are two reasons for this: - a bug in NetworkManager, dbus-daemon or else may cause such errors. This must not trigger an assertion failure in the client application, at least not unless requested. - libnm must be forward and backward compatible against a different NetworkManager server version. That is only possibly by ignoring anything that is unexpected. Asserting by default might prevent to implement API changes, both on libnm and server side. Note that we also don't notify the calling application via dedicated API. On the one hand, these things *can* happen. On the other hand, what would the calling appication do about it anyway? libnm by default must just behave gracefully and pretend all is good. For testing, development and debugging that is however not useful. We want the user to opt in to strict API validation. The user will be able to do that by setting "LIBNM_CLIENT_DEBUG=warning", which causes API violations being logged with g_warning(). These are assertions when running with G_DEBUG=fatal-warnings. This is inspired by GDBus' G_DBUS_DEBUG variable. Note that LIBNM_CLIENT_DEBUG environment variables is undocumented, unstable API. It's used for debugging and testing of the current libnm version at hand. There is no guaranteed stable behavior how a different libnm version might behave.
2019-10-14 08:06:33 +02:00
}
void
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
_nml_dbus_log(NMLDBusLogLevel level, gboolean use_stdout, const char *fmt, ...)
{
NMLDBusLogLevel configured_log_level;
gs_free char * msg = NULL;
va_list args;
const char * prefix = "";
gint64 ts;
/* we only call _nml_dbus_log() after nml_dbus_log_enabled(), which already does
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* an atomic access to the variable. Since the value is only initialized once and
* never changes, we can just access it without additional locking. */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
configured_log_level = _nml_dbus_log_level;
nm_assert(level & configured_log_level);
va_start(args, fmt);
msg = g_strdup_vprintf(fmt, args);
va_end(args);
switch (level) {
case NML_DBUS_LOG_LEVEL_TRACE:
prefix = "<trace> ";
break;
case NML_DBUS_LOG_LEVEL_DEBUG:
prefix = "<debug> ";
break;
case NML_DBUS_LOG_LEVEL_WARN:
prefix = "<warn > ";
if (NM_FLAGS_HAS(configured_log_level, _NML_DBUS_LOG_LEVEL_WARN)) {
g_warning("libnm-dbus: %s%s", prefix, msg);
return;
}
break;
case NML_DBUS_LOG_LEVEL_ERROR:
prefix = "<error> ";
if (NM_FLAGS_HAS(configured_log_level, _NML_DBUS_LOG_LEVEL_ERROR)) {
g_critical("libnm-dbus: %s%s", prefix, msg);
return;
}
if (NM_FLAGS_HAS(configured_log_level, _NML_DBUS_LOG_LEVEL_WARN)) {
g_warning("libnm-dbus: %s%s", prefix, msg);
return;
}
break;
default:
break;
}
ts = nm_utils_clock_gettime_nsec(CLOCK_BOOTTIME);
if (use_stdout) {
g_print("libnm-dbus: %s[%" G_GINT64_FORMAT ".%05" G_GINT64_FORMAT "] %s\n",
prefix,
ts / NM_UTILS_NSEC_PER_SEC,
(ts / (NM_UTILS_NSEC_PER_SEC / 10000)) % 10000,
msg);
} else {
g_printerr("libnm-dbus: %s[%" G_GINT64_FORMAT ".%05" G_GINT64_FORMAT "] %s\n",
prefix,
ts / NM_UTILS_NSEC_PER_SEC,
(ts / (NM_UTILS_NSEC_PER_SEC / 10000)) % 10000,
msg);
}
libnm: add logging NML_DBUS_LOG*() for debugging D-Bus for NMClient Commonly, a library (like libnm) is not supposed to log anything. Logging is not a suitable way to notify the calling application about anything. When something of importance happens, then the application must be notified via the library's API. However, logging can be very useful for debugging to see what is going on. Add a logging macro that by default does nothing, but can be turned on via an environment variable "LIBNM_CLIENT_DEBUG=debug". Another point is that libnm relies on the server side NetworkManager D-Bus interface to be in an expected manner. For example, we require a D-Bus object "org.freedesktop.NetworkManager" to be present and certain D-Bus interfaces implemented. However libnm should treat NetworkManager as external and untrusted component. That means, we cannot assert against the expectations we have. There are two reasons for this: - a bug in NetworkManager, dbus-daemon or else may cause such errors. This must not trigger an assertion failure in the client application, at least not unless requested. - libnm must be forward and backward compatible against a different NetworkManager server version. That is only possibly by ignoring anything that is unexpected. Asserting by default might prevent to implement API changes, both on libnm and server side. Note that we also don't notify the calling application via dedicated API. On the one hand, these things *can* happen. On the other hand, what would the calling appication do about it anyway? libnm by default must just behave gracefully and pretend all is good. For testing, development and debugging that is however not useful. We want the user to opt in to strict API validation. The user will be able to do that by setting "LIBNM_CLIENT_DEBUG=warning", which causes API violations being logged with g_warning(). These are assertions when running with G_DEBUG=fatal-warnings. This is inspired by GDBus' G_DBUS_DEBUG variable. Note that LIBNM_CLIENT_DEBUG environment variables is undocumented, unstable API. It's used for debugging and testing of the current libnm version at hand. There is no guaranteed stable behavior how a different libnm version might behave.
2019-10-14 08:06:33 +02:00
}
/*****************************************************************************/
libnm: move wincaps_to_dash() function to "nm-libnm-utils.h" We will need it, if only for testing/asserting.
2019-10-15 12:53:53 +02:00
/* Stolen from dbus-glib */
char *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_utils_wincaps_to_dash(const char *caps)
libnm: move wincaps_to_dash() function to "nm-libnm-utils.h" We will need it, if only for testing/asserting.
2019-10-15 12:53:53 +02:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
const char *p;
GString * str;
str = g_string_new(NULL);
p = caps;
while (*p) {
if (g_ascii_isupper(*p)) {
if (str->len > 0 && (str->len < 2 || str->str[str->len - 2] != '-'))
g_string_append_c(str, '-');
g_string_append_c(str, g_ascii_tolower(*p));
} else
g_string_append_c(str, *p);
++p;
}
return g_string_free(str, FALSE);
libnm: move wincaps_to_dash() function to "nm-libnm-utils.h" We will need it, if only for testing/asserting.
2019-10-15 12:53:53 +02:00
}
/*****************************************************************************/
libnm/utils: split out vendor fixup The hwdb generally contains the strings of rather poor quality, especially when it comes to sensibly presenting them to the user and they need various cleanups. While the following patches add fixups, this one splits out vendor fixups, because it turns out that a different set of fixups is needed than for products.
2018-02-23 14:52:00 +01:00
static char *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
_fixup_string(const char * desc,
const char *const *ignored_phrases,
const char *const *ignored_words,
gboolean square_brackets_sensible)
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
char * desc_full;
gboolean in_paren = FALSE;
char * p, *q;
int i;
if (!desc || !desc[0])
return NULL;
/* restore original non-UTF-8-safe text. */
desc_full = nm_utils_str_utf8safe_unescape_cp(desc, NM_UTILS_STR_UTF8_SAFE_FLAG_NONE);
/* replace all invalid UTF-8 bytes with space. */
p = desc_full;
while (!g_utf8_validate(p, -1, (const char **) &q)) {
/* the byte is invalid UTF-8. Replace it with space and proceed. */
*q = ' ';
p = q + 1;
}
/* replace '_', ',', ASCII control characters and parentheses, with space. */
for (p = desc_full; p[0]; p++) {
if (*p == '(')
in_paren = TRUE;
if (NM_IN_SET(*p, '_', ',') || *p < ' ' || in_paren)
*p = ' ';
if (*p == ')')
in_paren = FALSE;
}
/* Attempt to shorten ID by ignoring certain phrases */
for (i = 0; ignored_phrases[i]; i++) {
p = strstr(desc_full, ignored_phrases[i]);
if (p) {
const char *eow = &p[strlen(ignored_phrases[i])];
/* require that the phrase is delimited by space, or
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* at the beginning or end of the description. */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
if ((p == desc_full || p[-1] == ' ') && NM_IN_SET(eow[0], '\0', ' '))
memmove(p, eow, strlen(eow) + 1); /* +1 for the \0 */
}
}
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
/* Attempt to shorten ID by ignoring certain individual words.
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* - word-split the description at spaces
* - coalesce multiple spaces
* - skip over ignored_words */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
p = desc_full;
q = desc_full;
for (;;) {
char *eow;
gsize l;
/* skip leading spaces. */
while (p[0] == ' ')
p++;
if (!p[0])
break;
/* split leading word on first space */
eow = strchr(p, ' ');
if (eow)
*eow = '\0';
if (nm_utils_strv_find_first((char **) ignored_words, -1, p) >= 0)
goto next;
l = strlen(p);
if (q != p) {
if (q != desc_full)
*q++ = ' ';
memmove(q, p, l);
}
q += l;
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
next:
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
if (!eow)
break;
p = eow + 1;
}
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
*q++ = '\0';
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
p = strchr(desc_full, '[');
if (p == desc_full) {
/* All we're left with is in square brackets.
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* Always prefer that to a blank string.*/
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
square_brackets_sensible = TRUE;
}
if (square_brackets_sensible) {
/* If there's a [<string>] that survived the substitution, then the string
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* is a short form that is generally preferable. */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
q = strchr(desc_full, ']');
if (p && q > p) {
p++;
memmove(desc_full, p, q - p);
desc_full[q - p] = '\0';
}
} else {
/* [<string>] sometimes contains the preferred human-readable name, but
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* mostly it's utterly useless. Sigh. Drop it. */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
if (p) {
if (p > desc_full && p[-1] == ' ')
p--;
*p = '\0';
}
}
if (!desc_full[0]) {
g_free(desc_full);
return NULL;
}
return desc_full;
libnm/utils: split out vendor fixup The hwdb generally contains the strings of rather poor quality, especially when it comes to sensibly presenting them to the user and they need various cleanups. While the following patches add fixups, this one splits out vendor fixups, because it turns out that a different set of fixups is needed than for products.
2018-02-23 14:52:00 +01:00
}
char *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_utils_fixup_vendor_string(const char *desc)
libnm/utils: split out vendor fixup The hwdb generally contains the strings of rather poor quality, especially when it comes to sensibly presenting them to the user and they need various cleanups. While the following patches add fixups, this one splits out vendor fixups, because it turns out that a different set of fixups is needed than for products.
2018-02-23 14:52:00 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
static const char *const IGNORED_PHRASES[] = {
"Access Systems",
"Business Mobile Networks BV",
"Communications & Multimedia",
"Company of Japan",
"Computer Co.",
"Computer Corp.",
"Computer Corporation",
"Computer Inc.",
"Computer, Inc.",
"Information and Communication Products",
"Macao Commercial Offshore",
"Mobile Phones",
"(M) Son",
"Multimedia Internet Technology",
"Technology Group Ltd.",
"Wireless Networks",
"Wireless Solutions",
NULL,
};
static const char *const IGNORED_WORDS[] = {
"AB",
"AG",
"A/S",
"ASA",
"B.V.",
"Chips",
"Co.",
"Co",
"Communications",
"Components",
"Computers",
"Computertechnik",
"corp.",
"Corp.",
"Corp",
"Corporation",
"Design",
"Electronics",
"Enterprise",
"Enterprises",
"Europe",
"GmbH",
"Hardware",
"[hex]",
"Holdings",
"Inc.",
"Inc",
"INC.",
"Incorporated",
"Instruments",
"International",
"Intl.",
"Labs",
"Limited.",
"Limited",
"Ltd.",
"Ltd",
"Microelectronics",
"Microsystems",
"MSM",
"Multimedia",
"Networks",
"Norway",
"Optical",
"PCS",
"Semiconductor",
"Systems",
"Systemtechnik",
"Techcenter",
"Technik",
"Technologies",
"Technology",
"TECHNOLOGY",
"Telephonics",
"USA",
"WCDMA",
NULL,
};
char *desc_full;
char *p;
desc_full = _fixup_string(desc, IGNORED_PHRASES, IGNORED_WORDS, TRUE);
if (!desc_full)
return NULL;
/* Chop off everything after a slash. */
for (p = desc_full; *p; p++) {
if ((p[0] == ' ' && p[1] == '/') || p[0] == '/') {
p[0] = '\0';
break;
}
}
nm_assert(g_utf8_validate(desc_full, -1, NULL));
return desc_full;
libnm/utils: split out vendor fixup The hwdb generally contains the strings of rather poor quality, especially when it comes to sensibly presenting them to the user and they need various cleanups. While the following patches add fixups, this one splits out vendor fixups, because it turns out that a different set of fixups is needed than for products.
2018-02-23 14:52:00 +01:00
}
char *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_utils_fixup_product_string(const char *desc)
libnm/utils: split out vendor fixup The hwdb generally contains the strings of rather poor quality, especially when it comes to sensibly presenting them to the user and they need various cleanups. While the following patches add fixups, this one splits out vendor fixups, because it turns out that a different set of fixups is needed than for products.
2018-02-23 14:52:00 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
static const char *const IGNORED_PHRASES[] = {
"100/10 MBit",
"10/100 Mbps",
"1.0 GbE",
"10 GbE",
"10 Gigabit",
"10 Mbps",
"1/10 Gigabit",
"150 Mbps",
"2.5 GbE",
"54 Mbps",
"Attached Port",
"+ BT",
"\"CDC Subset\"",
"CE Media Processor",
"Controller Area Network",
"Converged Network",
"DEC-Tulip compatible",
"Dish Adapter",
"Double 108 Mbps",
"Dual Band",
"Dual Port",
"Embedded UTP",
"Ethernet Connection",
"Ethernet Pro 100",
"Express Module",
"Fabric Adapter",
"Fast Ethernet",
"for 10GBASE-T",
"for 10GbE backplane",
"for 10GbE QSFP+",
"for 10GbE SFP+",
"for 1GbE",
"for 20GbE backplane",
"for 25GbE backplane",
"for 25GbE SFP28",
"for 40GbE backplane",
"for 40GbE QSFP+",
"G Adapter",
"Gigabit Desktop Network",
"Gigabit Ethernet",
"Gigabit or",
"Host Interface",
"Host Virtual Interface",
"IEEE 802.11a/b/g",
"IEEE 802.11g",
"IEEE 802.11G",
"IEEE 802.11n",
"MAC + PHY",
"Mini Card",
"Mini Wireless",
"multicore SoC",
"Multi Function",
"N Draft 11n Wireless",
"Network Connection",
"Network Everywhere",
"N Wireless",
"N+ Wireless",
"OCT To Fast Ethernet Converter",
"PC Card",
"PCI Express",
"Platform Controller Hub",
"Plus Bluetooth",
"Quad Gigabit",
"rev 1",
"rev 17",
"rev 2",
"rev A",
"rev B",
"rev F",
"TO Ethernet",
"Turbo Wireless Adapter",
"Unified Wire",
"USB 1.1",
"USB 2.0",
"Virtual media for",
"WiFi Link",
"+ WiMAX",
"WiMAX/WiFi Link",
"Wireless G",
"Wireless G+",
"Wireless Lan",
"Wireless Mini adapter",
"Wireless Mini Adapter",
"Wireless N",
"with 1000-BASE-T interface",
"with CX4 copper interface",
"with Range Amplifier",
"with SR-XFP optical interface",
"w/ Upgradable Antenna",
NULL,
};
static const char *const IGNORED_WORDS[] = {
"1000BaseSX",
"1000BASE-T",
"1000Base-ZX",
"100/10M",
"100baseFx",
"100Base-MII",
"100Base-T",
"100BaseT4",
"100Base-TX",
"100BaseTX",
"100GbE",
"100Mbps",
"100MBps",
"10/100",
"10/100/1000",
"10/100/1000Base-T",
"10/100/1000BASE-T",
"10/100BaseT",
"10/100baseTX",
"10/100BaseTX",
"10/100/BNC",
"10/100M",
"10/20-Gigabit",
"10/25/40/50GbE",
"10/40G",
"10base-FL",
"10BaseT",
"10BASE-T",
"10G",
"10Gb",
"10Gb/25Gb",
"10Gb/25Gb/40Gb/50Gb",
"10Gbase-T",
"10GBase-T",
"10GBASE-T",
"10GbE",
"10Gbps",
"10-Giga",
"10-Gigabit",
"10mbps",
"10Mbps",
"1/10GbE",
"1/10-Gigabit",
"11b/g/n",
"11g",
"150Mbps",
"16Gbps/10Gbps",
"1GbE",
"1x2:2",
"20GbE",
"25Gb",
"25GbE",
"2-Port",
"2x3:3",
"3G",
"3G/4G",
"3x3:3",
"40GbE",
"4G",
"54g",
"54M",
"54Mbps",
"56k",
"5G",
"802.11",
"802.11a/b/g",
"802.11abg",
"802.11a/b/g/n",
"802.11abgn",
"802.11ac",
"802.11ad",
"802.11a/g",
"802.11b",
"802.11b/g",
"802.11bg",
"802.11b/g/n",
"802.11bgn",
"802.11b/g/n-draft",
"802.11g",
"802.11n",
"802.11N",
"802.11n/b/g",
"802.11ng",
"802AIN",
"802UIG-1",
"adapter",
"Adapter",
"adaptor",
"ADSL",
"Basic",
"CAN-Bus",
"card",
"Card",
"Cardbus",
"CardBus",
"CDMA",
"CNA",
"Composite",
"controller",
"Controller",
"Copper",
"DB",
"Desktop",
"device",
"Device",
"dongle",
"driver",
"Dual-band",
"Dual-Protocol",
"EISA",
"Enhanced",
"ethernet.",
"ethernet",
"Ethernet",
"Ethernet/RNDIS",
"ExpressModule",
"family",
"Family",
"Fast/Gigabit",
"Fiber",
"gigabit",
"Gigabit",
"G-NIC",
"Hi-Gain",
"Hi-Speed",
"HSDPA",
"HSUPA",
"integrated",
"Integrated",
"interface",
"LAN",
"LAN+Winmodem",
"Laptop",
"LTE",
"LTE/UMTS/GSM",
"MAC",
"Micro",
"Mini-Card",
"Mini-USB",
"misprogrammed",
"modem",
"Modem",
"Modem/Networkcard",
"Module",
"Multimode",
"Multithreaded",
"Name:",
"net",
"network",
"Network",
"n/g/b",
"NIC",
"Notebook",
"OEM",
"PCI",
"PCI64",
"PCIe",
"PCI-E",
"PCI-Express",
"PCI-X",
"PCMCIA",
"PDA",
"PnP",
"RDMA",
"RJ-45",
"Series",
"Server",
"SoC",
"Switch",
"Technologies",
"TOE",
"USB",
"USB2.0",
"USB/Ethernet",
"UTP",
"UTP/Coax",
"v1",
"v1.1",
"v2",
"V2.0",
"v3",
"v4",
"wifi",
"Wi-Fi",
"WiFi",
"wireless",
"Wireless",
"Wireless-150N",
"Wireless-300N",
"Wireless-G",
"Wireless-N",
"WLAN",
NULL,
};
char *desc_full;
char *p;
desc_full = _fixup_string(desc, IGNORED_PHRASES, IGNORED_WORDS, FALSE);
if (!desc_full)
return NULL;
/* Chop off everything after a '-'. */
for (p = desc_full; *p; p++) {
if (p[0] == ' ' && p[1] == '-' && p[2] == ' ') {
p[0] = '\0';
break;
}
}
nm_assert(g_utf8_validate(desc_full, -1, NULL));
return desc_full;
libnm: move fixup_desc_string() to nm-libnm-utils.c
2017-05-19 10:32:13 +02:00
}
libnm: move nm_permission_to_client()/nm_permission_result_to_client() to nm-libnm-utils.c It's nicely trivial and independent. Move it to a separate place, to avoid cluttering the more complicated code and to make it testable. Also, use binary search to find the value by string.
2019-10-29 16:55:15 +01:00
/*****************************************************************************/
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
const NMLDBusMetaIface *const _nml_dbus_meta_ifaces[] = {
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
&_nml_dbus_meta_iface_nm,
&_nml_dbus_meta_iface_nm_accesspoint,
&_nml_dbus_meta_iface_nm_agentmanager,
&_nml_dbus_meta_iface_nm_checkpoint,
&_nml_dbus_meta_iface_nm_connection_active,
&_nml_dbus_meta_iface_nm_dhcp4config,
&_nml_dbus_meta_iface_nm_dhcp6config,
&_nml_dbus_meta_iface_nm_device,
&_nml_dbus_meta_iface_nm_device_adsl,
&_nml_dbus_meta_iface_nm_device_bluetooth,
&_nml_dbus_meta_iface_nm_device_bond,
&_nml_dbus_meta_iface_nm_device_bridge,
&_nml_dbus_meta_iface_nm_device_dummy,
&_nml_dbus_meta_iface_nm_device_generic,
&_nml_dbus_meta_iface_nm_device_iptunnel,
&_nml_dbus_meta_iface_nm_device_infiniband,
&_nml_dbus_meta_iface_nm_device_lowpan,
&_nml_dbus_meta_iface_nm_device_macsec,
&_nml_dbus_meta_iface_nm_device_macvlan,
&_nml_dbus_meta_iface_nm_device_modem,
&_nml_dbus_meta_iface_nm_device_olpcmesh,
&_nml_dbus_meta_iface_nm_device_ovsbridge,
&_nml_dbus_meta_iface_nm_device_ovsinterface,
&_nml_dbus_meta_iface_nm_device_ovsport,
&_nml_dbus_meta_iface_nm_device_ppp,
&_nml_dbus_meta_iface_nm_device_statistics,
&_nml_dbus_meta_iface_nm_device_team,
&_nml_dbus_meta_iface_nm_device_tun,
&_nml_dbus_meta_iface_nm_device_veth,
&_nml_dbus_meta_iface_nm_device_vlan,
&_nml_dbus_meta_iface_nm_device_vrf,
&_nml_dbus_meta_iface_nm_device_vxlan,
&_nml_dbus_meta_iface_nm_device_wifip2p,
&_nml_dbus_meta_iface_nm_device_wireguard,
&_nml_dbus_meta_iface_nm_device_wired,
&_nml_dbus_meta_iface_nm_device_wireless,
&_nml_dbus_meta_iface_nm_device_wpan,
&_nml_dbus_meta_iface_nm_dnsmanager,
&_nml_dbus_meta_iface_nm_ip4config,
&_nml_dbus_meta_iface_nm_ip6config,
&_nml_dbus_meta_iface_nm_settings,
&_nml_dbus_meta_iface_nm_settings_connection,
&_nml_dbus_meta_iface_nm_vpn_connection,
&_nml_dbus_meta_iface_nm_wifip2ppeer,
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
};
#define COMMON_PREFIX "org.freedesktop.NetworkManager"
static int
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
_strcmp_common_prefix(gconstpointer a, gconstpointer b, gpointer user_data)
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
const NMLDBusMetaIface *iface = a;
const char * dbus_iface_name = b;
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_assert(g_str_has_prefix(iface->dbus_iface_name, COMMON_PREFIX));
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
return strcmp(&iface->dbus_iface_name[NM_STRLEN(COMMON_PREFIX)], dbus_iface_name);
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
}
const NMLDBusMetaIface *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nml_dbus_meta_iface_get(const char *dbus_iface_name)
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
gssize idx;
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_assert(dbus_iface_name);
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
G_STATIC_ASSERT_EXPR(G_STRUCT_OFFSET(NMLDBusMetaIface, dbus_iface_name) == 0);
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
/* we assume that NetworkManager only uses unique interface names. E.g. one
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* interface name always has one particular meaning (and offers one set of
* properties, signals and methods). This is a convenient assumption, and
* we sure would never violate it when extending NM's D-Bus API. */
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
if (NM_STR_HAS_PREFIX(dbus_iface_name, COMMON_PREFIX)) {
/* optimize, that in fact all our interfaces have the same prefix. */
idx = nm_utils_ptrarray_find_binary_search((gconstpointer *) _nml_dbus_meta_ifaces,
G_N_ELEMENTS(_nml_dbus_meta_ifaces),
&dbus_iface_name[NM_STRLEN(COMMON_PREFIX)],
_strcmp_common_prefix,
NULL,
NULL,
NULL);
} else
return NULL;
if (idx < 0)
return NULL;
return _nml_dbus_meta_ifaces[idx];
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
}
const NMLDBusMetaProperty *
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nml_dbus_meta_property_get(const NMLDBusMetaIface *meta_iface,
const char * dbus_property_name,
guint * out_idx)
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
gssize idx;
nm_assert(meta_iface);
nm_assert(dbus_property_name);
idx = nm_utils_array_find_binary_search(meta_iface->dbus_properties,
sizeof(meta_iface->dbus_properties[0]),
meta_iface->n_dbus_properties,
&dbus_property_name,
nm_strcmp_p_with_data,
NULL);
if (idx < 0) {
NM_SET_OUT(out_idx, meta_iface->n_dbus_properties);
return NULL;
}
NM_SET_OUT(out_idx, idx);
return &meta_iface->dbus_properties[idx];
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
}
void
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
_nml_dbus_meta_class_init_with_properties_impl(GObjectClass * object_class,
const NMLDBusMetaIface *const *meta_ifaces)
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
int i_iface;
nm_assert(G_IS_OBJECT_CLASS(object_class));
nm_assert(meta_ifaces);
nm_assert(meta_ifaces[0]);
for (i_iface = 0; meta_ifaces[i_iface]; i_iface++) {
const NMLDBusMetaIface *meta_iface = meta_ifaces[i_iface];
guint8 * reverse_idx;
guint8 i;
nm_assert(g_type_is_a(meta_iface->get_type_fcn(), G_OBJECT_CLASS_TYPE(object_class)));
nm_assert(meta_iface->n_obj_properties > 0);
nm_assert(meta_iface->obj_properties);
nm_assert(meta_iface->obj_properties_reverse_idx[0] == 0);
nm_assert(meta_iface->obj_properties == meta_ifaces[0]->obj_properties);
if (i_iface == 0)
g_object_class_install_properties(object_class,
meta_iface->n_obj_properties,
(GParamSpec **) meta_iface->obj_properties);
reverse_idx = (guint8 *) meta_iface->obj_properties_reverse_idx;
for (i = 0; i < meta_iface->n_obj_properties; i++)
reverse_idx[i] = 0xFFu;
for (i = 0; i < meta_iface->n_dbus_properties; i++) {
const NMLDBusMetaProperty *mpr = &meta_iface->dbus_properties[i];
if (mpr->obj_properties_idx != 0 && !mpr->obj_property_no_reverse_idx) {
nm_assert(mpr->obj_properties_idx < meta_iface->n_obj_properties);
nm_assert(reverse_idx[mpr->obj_properties_idx] == 0xFFu);
reverse_idx[mpr->obj_properties_idx] = i;
}
}
}
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
}
gboolean
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
nm_utils_g_param_spec_is_default(const GParamSpec *pspec)
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
{
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
g_return_val_if_fail(pspec, FALSE);
if (pspec->value_type == G_TYPE_BOOLEAN)
return ((((GParamSpecBoolean *) pspec)->default_value) == FALSE);
if (pspec->value_type == G_TYPE_UCHAR)
return ((((GParamSpecUChar *) pspec)->default_value) == 0u);
if (pspec->value_type == G_TYPE_INT)
return ((((GParamSpecInt *) pspec)->default_value) == 0);
if (pspec->value_type == G_TYPE_UINT)
return ((((GParamSpecUInt *) pspec)->default_value) == 0u);
if (pspec->value_type == G_TYPE_INT64)
return ((((GParamSpecInt64 *) pspec)->default_value) == 0);
if (pspec->value_type == G_TYPE_UINT64)
return ((((GParamSpecUInt64 *) pspec)->default_value) == 0u);
if (g_type_is_a(pspec->value_type, G_TYPE_ENUM))
return ((((GParamSpecEnum *) pspec)->default_value) == 0);
if (g_type_is_a(pspec->value_type, G_TYPE_FLAGS))
return ((((GParamSpecFlags *) pspec)->default_value) == 0u);
if (pspec->value_type == G_TYPE_STRING)
return ((((GParamSpecString *) pspec)->default_value) == NULL);
if (NM_IN_SET(pspec->value_type,
G_TYPE_BYTES,
G_TYPE_PTR_ARRAY,
G_TYPE_ARRAY,
G_TYPE_HASH_TABLE,
G_TYPE_STRV)) {
/* boxed types have NULL default. */
g_return_val_if_fail(G_IS_PARAM_SPEC_BOXED(pspec), FALSE);
g_return_val_if_fail(G_TYPE_IS_BOXED(pspec->value_type), FALSE);
return TRUE;
}
if (g_type_is_a(pspec->value_type, NM_TYPE_OBJECT)) {
/* object types have NULL default. */
g_return_val_if_fail(G_IS_PARAM_SPEC_OBJECT(pspec), FALSE);
g_return_val_if_fail(G_TYPE_IS_OBJECT(pspec->value_type), FALSE);
return TRUE;
}
/* This function is only used for asserting/testing. It thus
format: replace tabs for indentation in code comments sed -i \ -e 's/^'$'\t'' \*/ */g' \ -e 's/^'$'\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t'' \*/ */g' \ -e 's/^'$'\t\t\t\t\t\t\t'' \*/ */g' \ $(git ls-files -- '*.[hc]')
2020-09-28 14:50:01 +02:00
* strictly asserts and only support argument types that we expect. */
all: reformat all with new clang-format style Run: ./contrib/scripts/nm-code-format.sh -i ./contrib/scripts/nm-code-format.sh -i Yes, it needs to run twice because the first run doesn't yet produce the final result. Signed-off-by: Antonio Cardace <acardace@redhat.com>
2020-09-28 16:03:33 +02:00
g_return_val_if_reached(FALSE);
libnm: refactor caching of D-Bus objects in NMClient No longer use GDBusObjectMangaerClient and gdbus-codegen generated classes for the NMClient cache. Instead, use GDBusConnection directly and a custom implementation (NMLDBusObject) for caching D-Bus' ObjectManager data. CHANGES ------- - This is a complete rework. I think the previous implementation was difficult to understand. There were unfixed bugs and nobody understood the code well enough to fix them. Maybe somebody out there understood the code, but I certainly did not. At least nobody provided patches to fix those issues. I do believe that this implementation is more straightforward and easier to understand. It removes a lot of layers of code. Whether this claim of simplicity is true, each reader must decide for himself/herself. Note that it is still fairly complex. - There was a lingering performance issue with large number of D-Bus objects. The patch tries hard that the implementation scales well. Of course, when we cache N objects that have N-to-M references to other, we still are fundamentally O(N*M) for runtime and memory consumption (with M being the number of references between objects). But each part should behave efficiently and well. - Play well with GMainContext. libnm code (NMClient) is generally not thread safe. However, it should work to use multiple instances in parallel, as long as each access to a NMClient is through the caller's GMainContext. This follows glib's style and effectively allows to use NMClient in a multi threaded scenario. This implies to stick to a main context upon construction and ensure that callbacks are only invoked when iterating that context. Also, NMClient itself shall never iterate the caller's context. This also means, libnm must never use g_idle_add() or g_timeout_add(), as those enqueue sources in the g_main_context_default() context. - Get ordering of messages right. All events are consistently enqueued in a GMainContext and processed strictly in order. For example, previously "nm-object.c" tried to combine signals and emit them on an idle handler. That is wrong, signals must be emitted in the right order and when they happen. Note that when using GInitable's synchronous initialization to initialize the NMClient instance, NMClient internally still operates fully asynchronously. In that case NMClient has an internal main context. - NMClient takes over most of the functionality. When using D-Bus' ObjectManager interface, one needs to handle basically the entire state of the D-Bus interface. That cannot be separated well into distinct parts, and even if you try, you just end up having closely related code in different source files. Spreading related code does not make it easier to understand, on the contrary. That means, NMClient is inherently complex as it contains most of the logic. I think that is not avoidable, but it's not as bad as it sounds. - NMClient processes D-Bus messages and state changes in separate steps. First NMClient unpacks the message (e.g. _dbus_handle_properties_changed()) and keeps track of the changed data. Then we update the GObject instances (_dbus_handle_obj_changed_dbus()) without emitting any signals yet. Finally, we emit all signals and notifications that were collected (_dbus_handle_changes_commit()). Note that for example during the initial GetManagedObjects() reply, NMClient receive a large amount of state at once. But we first apply all the changes to our GObject instances before emitting any signals. The result is that signals are always emitted in a moment when the cache is consistent. The unavoidable downside is that when you receive a property changed signal, possibly many other properties changed already and more signals are about to be emitted. - NMDeviceWifi no longer modifies the content of the cache from client side during poke_wireless_devices_with_rf_status(). The content of the cache should be determined by D-Bus alone and follow what NetworkManager service exposes. Local modifications should be avoided. - This aims to bring no API/ABI change, though it does of course bring various subtle changes in behavior. Those should be all for the better, but the goal is not to break any existing clients. This does change internal (albeit externally visible) API, like dropping NM_OBJECT_DBUS_OBJECT_MANAGER property and NMObject no longer implementing GInitableIface and GAsyncInitableIface. - Some uses of gdbus-codegen classes remain in NMVpnPluginOld, NMVpnServicePlugin and NMSecretAgentOld. These are independent of NMClient/NMObject and should be reworked separately. - While we no longer use generated classes from gdbus-codegen, we don't need more glue code than before. Also before we constructed NMPropertiesInfo and a had large amount of code to propagate properties from NMDBus* to NMObject. That got completely reworked, but did not fundamentally change. You still need about the same effort to create the NMLDBusMetaIface. Not using generated bindings did not make anything worse (which tells about the usefulness of generated code, at least in the way it was used). - NMLDBusMetaIface and other meta data is static and immutable. This avoids copying them around. Also, macros like NML_DBUS_META_PROPERTY_INIT_U() have compile time checks to ensure the property types matches. It's pretty hard to misuse them because it won't compile. - The meta data now explicitly encodes the expected D-Bus types and makes sure never to accept wrong data. That would only matter when the server (accidentally or intentionally) exposes unexpected types on D-Bus. I don't think that was previously ensured in all cases. For example, demarshal_generic() only cared about the GObject property type, it didn't know the expected D-Bus type. - Previously GDBusObjectManager would sometimes emit warnings (g_log()). Those probably indicated real bugs. In any case, it prevented us from running CI with G_DEBUG=fatal-warnings, because there would be just too many unrelated crashes. Now we log debug messages that can be enabled with "LIBNM_CLIENT_DEBUG=trace". Some of these messages can also be turned into g_warning()/g_critical() by setting LIBNM_CLIENT_DEBUG=warning,error. Together with G_DEBUG=fatal-warnings, this turns them into assertions. Note that such "assertion failures" might also happen because of a server bug (or change). Thus these are not common assertions that indicate a bug in libnm and are thus not armed unless explicitly requested. In our CI we should now always run with LIBNM_CLIENT_DEBUG=warning,error and G_DEBUG=fatal-warnings and to catch bugs. Note that currently NetworkManager has bugs in this regard, so enabling this will result in assertion failures. That should be fixed first. - Note that this changes the order in which we emit "notify:devices" and "device-added" signals. I think it makes the most sense to emit first "device-removed", then "notify:devices", and finally "device-added" signals. This changes behavior for commit 52ae28f6e5bf ('libnm: queue added/removed signals and suppress uninitialized notifications'), but I don't think that users should actually rely on the order. Still, the new order makes the most sense to me. - In NetworkManager, profiles can be invisible to the user by setting "connection.permissions". Such profiles would be hidden by NMClient's nm_client_get_connections() and their "connection-added"/"connection-removed" signals. Note that NMActiveConnection's nm_active_connection_get_connection() and NMDevice's nm_device_get_available_connections() still exposes such hidden NMRemoteConnection instances. This behavior was preserved. NUMBERS ------- I compared 3 versions of libnm. [1] 962297f9085d, current tip of nm-1-20 branch [2] 4fad8c7c642e, current master, immediate parent of this patch [3] this patch All tests were done on Fedora 31, x86_64, gcc 9.2.1-1.fc31. The libraries were build with $ ./contrib/fedora/rpm/build_clean.sh -g -w test -W debug Note that RPM build already stripped the library. --- N1) File size of libnm.so.0.1.0 in bytes. There currently seems to be a issue on Fedora 31 generating wrong ELF notes. Usually, libnm is smaller but in these tests it had large (and bogus) ELF notes. Anyway, the point is to show the relative sizes, so it doesn't matter). [1] 4075552 (102.7%) [2] 3969624 (100.0%) [3] 3705208 ( 93.3%) --- N2) `size /usr/lib64/libnm.so.0.1.0`: text data bss dec hex filename [1] 1314569 (102.0%) 69980 ( 94.8%) 10632 ( 80.4%) 1395181 (101.4%) 1549ed /usr/lib64/libnm.so.0.1.0 [2] 1288410 (100.0%) 73796 (100.0%) 13224 (100.0%) 1375430 (100.0%) 14fcc6 /usr/lib64/libnm.so.0.1.0 [3] 1229066 ( 95.4%) 65248 ( 88.4%) 13400 (101.3%) 1307714 ( 95.1%) 13f442 /usr/lib64/libnm.so.0.1.0 --- N3) Performance test with test-client.py. With checkout of [2], run ``` prepare_checkout() { rm -rf /tmp/nm-test && \ git checkout -B test 4fad8c7c642e && \ git clean -fdx && \ ./autogen.sh --prefix=/tmp/nm-test && \ make -j 5 install && \ make -j 5 check-local-clients-tests-test-client } prepare_test() { NM_TEST_REGENERATE=1 NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v } do_test() { for i in {1..10}; do NM_TEST_CLIENT_BUILDDIR="/data/src/NetworkManager" NM_TEST_CLIENT_NMCLI_PATH=/usr/bin/nmcli python3 ./clients/tests/test-client.py -v || return -1 done echo "done!" } prepare_checkout prepare_test time do_test ``` [1] real 2m14.497s (101.3%) user 5m26.651s (100.3%) sys 1m40.453s (101.4%) [2] real 2m12.800s (100.0%) user 5m25.619s (100.0%) sys 1m39.065s (100.0%) [3] real 1m54.915s ( 86.5%) user 4m18.585s ( 79.4%) sys 1m32.066s ( 92.9%) --- N4) Performance. Run NetworkManager from build [2] and setup a large number of profiles (551 profiles and 515 devices, mostly unrealized). This setup is already at the edge of what NetworkManager currently can handle. Of course, that is a different issue. Here we just check how long plain `nmcli` takes on the system. ``` do_cleanup() { for UUID in $(nmcli -g NAME,UUID connection show | sed -n 's/^xx-c-.*:\([^:]\+\)$/\1/p'); do nmcli connection delete uuid "$UUID" done for DEVICE in $(nmcli -g DEVICE device status | grep '^xx-i-'); do nmcli device delete "$DEVICE" done } do_setup() { do_cleanup for i in {1..30}; do nmcli connection add type bond autoconnect no con-name xx-c-bond-$i ifname xx-i-bond-$i ipv4.method disabled ipv6.method ignore for j in $(seq $i 30); do nmcli connection add type vlan autoconnect no con-name xx-c-vlan-$i-$j vlan.id $j ifname xx-i-vlan-$i-$j vlan.parent xx-i-bond-$i ipv4.method disabled ipv6.method ignore done done systemctl restart NetworkManager.service sleep 5 } do_test() { perf stat -r 50 -B nmcli 1>/dev/null } do_test ``` [1] Performance counter stats for 'nmcli' (50 runs): 456.33 msec task-clock:u # 1.093 CPUs utilized ( +- 0.44% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,900 page-faults:u # 0.013 M/sec ( +- 0.02% ) 1,408,675,453 cycles:u # 3.087 GHz ( +- 0.48% ) 1,594,741,060 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 368,744,018 branches:u # 808.061 M/sec ( +- 0.02% ) 4,566,058 branch-misses:u # 1.24% of all branches ( +- 0.76% ) 0.41761 +- 0.00282 seconds time elapsed ( +- 0.68% ) [2] Performance counter stats for 'nmcli' (50 runs): 477.99 msec task-clock:u # 1.088 CPUs utilized ( +- 0.36% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 5,948 page-faults:u # 0.012 M/sec ( +- 0.03% ) 1,471,133,482 cycles:u # 3.078 GHz ( +- 0.36% ) 1,655,275,369 instructions:u # 1.13 insn per cycle ( +- 0.02% ) 382,595,152 branches:u # 800.433 M/sec ( +- 0.02% ) 4,746,070 branch-misses:u # 1.24% of all branches ( +- 0.49% ) 0.43923 +- 0.00242 seconds time elapsed ( +- 0.55% ) [3] Performance counter stats for 'nmcli' (50 runs): 352.36 msec task-clock:u # 1.027 CPUs utilized ( +- 0.32% ) 0 context-switches:u # 0.000 K/sec 0 cpu-migrations:u # 0.000 K/sec 4,790 page-faults:u # 0.014 M/sec ( +- 0.26% ) 1,092,341,186 cycles:u # 3.100 GHz ( +- 0.26% ) 1,209,045,283 instructions:u # 1.11 insn per cycle ( +- 0.02% ) 281,708,462 branches:u # 799.499 M/sec ( +- 0.01% ) 3,101,031 branch-misses:u # 1.10% of all branches ( +- 0.61% ) 0.34296 +- 0.00120 seconds time elapsed ( +- 0.35% ) --- N5) same setup as N4), but run `PAGER= /bin/time -v nmcli`: [1] Command being timed: "nmcli" User time (seconds): 0.42 System time (seconds): 0.04 Percent of CPU this job got: 107% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.43 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34456 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6128 Voluntary context switches: 1298 Involuntary context switches: 1106 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [2] Command being timed: "nmcli" User time (seconds): 0.44 System time (seconds): 0.04 Percent of CPU this job got: 108% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.44 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 34452 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 6169 Voluntary context switches: 1849 Involuntary context switches: 142 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 [3] Command being timed: "nmcli" User time (seconds): 0.32 System time (seconds): 0.02 Percent of CPU this job got: 102% Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.34 Average shared text size (kbytes): 0 Average unshared data size (kbytes): 0 Average stack size (kbytes): 0 Average total size (kbytes): 0 Maximum resident set size (kbytes): 29196 Average resident set size (kbytes): 0 Major (requiring I/O) page faults: 0 Minor (reclaiming a frame) page faults: 5059 Voluntary context switches: 919 Involuntary context switches: 685 Swaps: 0 File system inputs: 0 File system outputs: 0 Socket messages sent: 0 Socket messages received: 0 Signals delivered: 0 Page size (bytes): 4096 Exit status: 0 --- N6) same setup as N4), but run `nmcli monitor` and look at `ps aux` for the RSS size. USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND [1] me 1492900 21.0 0.2 461348 33248 pts/10 Sl+ 15:02 0:00 nmcli monitor [2] me 1490721 5.0 0.2 461496 33548 pts/10 Sl+ 15:00 0:00 nmcli monitor [3] me 1495801 16.5 0.1 459476 28692 pts/10 Sl+ 15:04 0:00 nmcli monitor
2019-10-30 11:42:58 +01:00
}
libnm: add nm_utils_print() function libnm supports verbose debug logging by setting "LIBNM_CLIENT_DEBUG" environment variable. That mechanism uses g_printerr() (or g_print()). When testing an application it's useful to combine printf debugging with this debug logging. However, python's print() statement is additionally buffered and not in sync with the logging functions that libnm uses. As far as I see, g_print() and g_printerr() is not accessible via introspections/pygobject, probably because these are variadic functions. Add nm_utils_print() to libnm. This ensures to use the same logging mechanism as libnm.
2020-11-04 16:53:25 +01:00
/*****************************************************************************/
/**
* nm_utils_print:
* @output_mode: if 1 it uses g_print(). If 2, it uses g_printerr().
* If 0, it uses either g_print() or g_printerr(), depending
* on LIBNM_CLIENT_DEBUG (and the "stdout" flag).
* @msg: the message to print. The function does not append
* a trailing newline.
*
* The only purpose of this function is to give access to g_print()
* or g_printerr() from pygobject. libnm can do debug logging by
* setting LIBNM_CLIENT_DEBUG and uses thereby g_printerr() or
* g_print(). A plain "print()" function in python is not in sync
* with these functions (it implements additional buffering). By
* using nm_utils_print(), the same logging mechanisms can be used.
*
* Since: 1.30
*/
void
nm_utils_print(int output_mode, const char *msg)
{
gboolean use_stdout;
g_return_if_fail(msg);
if (output_mode == 0) {
nml_dbus_log_enabled_full(NML_DBUS_LOG_LEVEL_ANY, &use_stdout);
output_mode = use_stdout ? 1 : 2;
}
if (output_mode == 1)
g_print("%s", msg);
else if (output_mode == 2)
g_printerr("%s", msg);
else
g_return_if_reached();
}
Reference in a new issue Copy permalink