fdo-mirrors/NetworkManager
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/NetworkManager/NetworkManager.git synced 2025-12-20 05:50:06 +01:00
Code Issues Projects Releases Packages Wiki Activity Actions
NetworkManager/src/devices/nm-device-bond.c

679 lines
22 KiB
C
Raw Normal View History

all: SPDX header conversion $ find * -type f |xargs perl contrib/scripts/spdx.pl $ git rm contrib/scripts/spdx.pl
2019-09-10 11:19:01 +02:00
// SPDX-License-Identifier: GPL-2.0+
all: manually drop code comments with file description
2019-09-25 13:13:40 +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) 2011 - 2018 Red Hat, Inc.
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
*/
all: cleanup includes and let "nm-default.h" include "config.h" - All internal source files (except "examples", which are not internal) should include "config.h" first. As also all internal source files should include "nm-default.h", let "config.h" be included by "nm-default.h" and include "nm-default.h" as first in every source file. We already wanted to include "nm-default.h" before other headers because it might contains some fixes (like "nm-glib.h" compatibility) that is required first. - After including "nm-default.h", we optinally allow for including the corresponding header file for the source file at hand. The idea is to ensure that each header file is self contained. - Don't include "config.h" or "nm-default.h" in any header file (except "nm-sd-adapt.h"). Public headers anyway must not include these headers, and internal headers are never included after "nm-default.h", as of the first previous point. - Include all internal headers with quotes instead of angle brackets. In practice it doesn't matter, because in our public headers we must include other headers with angle brackets. As we use our public headers also to compile our interal source files, effectively the result must be the same. Still do it for consistency. - Except for <config.h> itself. Include it with angle brackets as suggested by https://www.gnu.org/software/autoconf/manual/autoconf.html#Configuration-Headers
2016-02-19 14:57:48 +01:00
#include "nm-default.h"
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core: refactor private data in "src" - use _NM_GET_PRIVATE() and _NM_GET_PRIVATE_PTR() everywhere. - reorder statements, to have GObject related functions (init, dispose, constructed) at the bottom of each file and in a consistent order w.r.t. each other. - unify whitespaces in signal and properties declarations. - use NM_GOBJECT_PROPERTIES_DEFINE() and _notify() - drop unused signal slots in class structures - drop unused header files for device factories
2016-09-29 13:49:01 +02:00
#include "nm-device-bond.h"
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
#include <stdlib.h>
bond: avoid setting "active_slave" option without interface enslaved Kernel will reject setting "active_slave", if the interface is not enslaved or not up. We already handle that by setting the option whenever we enslave an interface. However, we also must not set it initially, otherwise we get an ugly error log message: NetworkManager[939]: <debug> [1594709143.7459] platform-linux: sysctl: setting net:/sys/class/net/bond99/bonding/active_slave to eth1 (current value is ) NetworkManager[939]: <error> [1594709143.7459] platform-linux: sysctl: failed to set bonding/active_slave to eth1: (22) Invalid argument NetworkManager[939]: <warn> [1594709143.7460] device (bond99): failed to set bonding attribute active_slave to eth1 ... kernel: bond99: (slave eth1): Device is not bonding slave kernel: bond99: option active_slave: invalid value (eth1) See-also: https://bugzilla.redhat.com/show_bug.cgi?id=1856640 https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/577
2020-07-14 09:16:51 +02:00
#include <net/if.h>
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
#include "NetworkManagerUtils.h"
#include "nm-device-private.h"
build: don't add subdirectories to include search path but require qualified include Keep the include paths clean and separate. We use directories to group source files together. That makes sense (I guess), but then we should use this grouping also when including files. Thus require to #include files with their path relative to "src/". Also, we build various artifacts from the "src/" tree. Instead of having individual CFLAGS for each artifact in Makefile.am, the CFLAGS should be unified. Previously, the CFLAGS for each artifact differ and are inconsistent in which paths they add to the search path. Fix the inconsistency by just don't add the paths at all.
2016-11-21 00:43:52 +01:00
#include "platform/nm-platform.h"
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
#include "nm-device-factory.h"
libnm-core: add nm-core-types.h, remove cross-includes Add nm-core-types.h, typedefing all of the GObject types in libnm-core; this is needed so that nm-setting.h can reference NMConnection in addition to nm-connection.h referencing NMSetting. Removing the cross-includes from the various headers causes lots of fallout elsewhere. (In particular, nm-utils.h used to include nm-connection.h, which included every setting header, so any file that included nm-utils.h automatically got most of the rest of libnm-core without needing to pay attention to specifics.) Fix this up by including nm-core-internal.h from those files that are now missing includes.
2014-10-21 22:09:52 -04:00
#include "nm-core-internal.h"
bond: allow overriding the MTU Set the MTU if 802-3-ethernet.mtu is present and non-zero. The ifcfg-rh writer sets it if there's a MTU key. https://bugzilla.redhat.com/show_bug.cgi?id=1177582
2015-01-06 19:09:46 +01:00
#include "nm-ip4-config.h"
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
#include "nm-device-logging.h"
_LOG_DECLARE_SELF(NMDeviceBond);
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core: refactor private data in "src" - use _NM_GET_PRIVATE() and _NM_GET_PRIVATE_PTR() everywhere. - reorder statements, to have GObject related functions (init, dispose, constructed) at the bottom of each file and in a consistent order w.r.t. each other. - unify whitespaces in signal and properties declarations. - use NM_GOBJECT_PROPERTIES_DEFINE() and _notify() - drop unused signal slots in class structures - drop unused header files for device factories
2016-09-29 13:49:01 +02:00
/*****************************************************************************/
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core: refactor private data in "src" - use _NM_GET_PRIVATE() and _NM_GET_PRIVATE_PTR() everywhere. - reorder statements, to have GObject related functions (init, dispose, constructed) at the bottom of each file and in a consistent order w.r.t. each other. - unify whitespaces in signal and properties declarations. - use NM_GOBJECT_PROPERTIES_DEFINE() and _notify() - drop unused signal slots in class structures - drop unused header files for device factories
2016-09-29 13:49:01 +02:00
struct _NMDeviceBond {
NMDevice parent;
};
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core: refactor private data in "src" - use _NM_GET_PRIVATE() and _NM_GET_PRIVATE_PTR() everywhere. - reorder statements, to have GObject related functions (init, dispose, constructed) at the bottom of each file and in a consistent order w.r.t. each other. - unify whitespaces in signal and properties declarations. - use NM_GOBJECT_PROPERTIES_DEFINE() and _notify() - drop unused signal slots in class structures - drop unused header files for device factories
2016-09-29 13:49:01 +02:00
struct _NMDeviceBondClass {
NMDeviceClass parent;
};
G_DEFINE_TYPE (NMDeviceBond, nm_device_bond, NM_TYPE_DEVICE)
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
all: modify line separator comments to be 80 chars wide sed 's#^/\*\{5\}\*\+/$#/*****************************************************************************/#' $(git grep -l '\*\{5\}' | grep '\.[hc]$') -i
2016-10-02 18:22:50 +02:00
/*****************************************************************************/
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
device: use NMDeviceCapabilities enum for device capabilities
2015-04-17 15:15:38 +02:00
static NMDeviceCapabilities
core: remove "real_" prefix from method implementation names The idea was copied from gtk, but it's only used there in cases where the method's wrapper function and default implementation would otherwise have the same name, which never happens in NM because our method implementations aren't prefixed with the type name, so it's just noise here.
2012-09-27 12:12:15 -04:00
get_generic_capabilities (NMDevice *dev)
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
{
core: earlier software capability detection We need to know whether we can create interfaces of any given NMDevice subclass or not. So don't rely on just the NMPlatformLink for that information, because we won't have a platform link for software devices before we create them.
2014-09-24 17:46:15 -05:00
return NM_DEVICE_CAP_CARRIER_DETECT | NM_DEVICE_CAP_IS_SOFTWARE;
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
}
static gboolean
core: remove "real_" prefix from method implementation names The idea was copied from gtk, but it's only used there in cases where the method's wrapper function and default implementation would otherwise have the same name, which never happens in NM because our method implementations aren't prefixed with the type name, so it's just noise here.
2012-09-27 12:12:15 -04:00
complete_connection (NMDevice *device,
NMConnection *connection,
const char *specific_object,
core: avoid clone of all-connections list for nm_utils_complete_generic() NMSettings exposes a cached list of all connection. We don't need to clone it. Note that this is not save against concurrent modification, meaning, add/remove of connections in NMSettings will invalidate the list. However, it wasn't save against that previously either, because altough we cloned the container (GSList), we didn't take an additional reference to the elements. This is purely a performance optimization, we don't need to clone the list. Also, since the original list is of type "NMConnection *const*", use that type insistently, instead of dependent API requiring GSList. IMO, GSList is anyway not a very nice API for many use cases because it requires an additional slice allocation for each element. It's slower, and often less convenient to use.
2018-03-14 08:57:42 +01:00
NMConnection *const*existing_connections,
core: remove "real_" prefix from method implementation names The idea was copied from gtk, but it's only used there in cases where the method's wrapper function and default implementation would otherwise have the same name, which never happens in NM because our method implementations aren't prefixed with the type name, so it's just noise here.
2012-09-27 12:12:15 -04:00
GError **error)
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
{
all: stop using virtual interface-name properties The virtual :interface-name properties (eg, NMDeviceBond:interface-name) are deprecated in favor of NMSettingConnection:interface-name, and nm_connection_verify() ensures that their values are kept in sync. So (a) there is no need to set those properties when we can just set NMSettingConnection:interface-name instead, and (b) we can replace any calls to the setting-specific get_interface_name() methods with nm_connection_get_interface_name() or nm_setting_connection_get_interface_name().
2014-08-05 14:13:42 -04:00
NMSettingBond *s_bond;
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
device: don't use platform singleton getter in device subclasses Reduce the use of NM_PLATFORM_GET / nm_platform_get() to get the platform singleton instance. For one, this is a step towards supporting namespaces, where we need to use different NMNetns/NMPlatform instances depending on in which namespace the device lives. Also, we should reduce our use of singletons. They are difficult to coordinate on shutdown. Instead there should be a clear order of dependencies, expressed by owning a reference to those singelton instances. We already own a reference to the platform singelton, so use it and avoid NM_PLATFORM_GET. (cherry picked from commit 94d9ee129d85ff926a10de78fedcc6b57cdcdb19)
2017-04-18 12:09:02 +02:00
nm_utils_complete_generic (nm_device_get_platform (device),
utils: inject platform into nm_utils_complete_generic()
2016-03-08 13:57:20 +01:00
connection,
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
NM_SETTING_BOND_SETTING_NAME,
existing_connections,
NULL,
core: refactor nm_utils_complete_generic() not to use a dynamic format string For NMDeviceWifi and NMDeviceWimax, the printf format string for nm_utils_complete_generic() was created based on ssid/nsp. Since these input strings are untrusted, this is a serious bug. Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-25 16:21:59 +02:00
_("Bond connection"),
core: abstract out the duplicated default-ifname-generating code NMDeviceBond, NMDeviceBridge, and NMDeviceTeam all used basically the same code to generate a default interface name. Move it into nm_utils_complete_generic().
2014-08-05 17:11:57 -04:00
"bond",
utils: add ifname argument to nm_utils_complete_generic() It's a common thing to complete a connection with an interface name; adding it to the common path is goint to save as a few tens of lines later on.
2019-05-07 10:15:10 +02:00
NULL,
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
TRUE);
s_bond = nm_connection_get_setting_bond (connection);
if (!s_bond) {
s_bond = (NMSettingBond *) nm_setting_bond_new ();
nm_connection_add_setting (connection, NM_SETTING (s_bond));
}
return TRUE;
}
all: modify line separator comments to be 80 chars wide sed 's#^/\*\{5\}\*\+/$#/*****************************************************************************/#' $(git grep -l '\*\{5\}' | grep '\.[hc]$') -i
2016-10-02 18:22:50 +02:00
/*****************************************************************************/
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core: use nm_platform_master_*_option() for bonds Acked-by: Dan Winship <danw@gnome.org>
2013-08-28 10:38:46 +02:00
static gboolean
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
_set_bond_attr (NMDevice *device, const char *attr, const char *value)
core: use nm_platform_master_*_option() for bonds Acked-by: Dan Winship <danw@gnome.org>
2013-08-28 10:38:46 +02:00
{
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
NMDeviceBond *self = NM_DEVICE_BOND (device);
bond: use platform functions for writing sysfs Refactor writing the bonding options, to use again the platform function for the sysfs options. Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 20:15:03 +02:00
int ifindex = nm_device_get_ifindex (device);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
gboolean ret;
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
ret = nm_platform_sysctl_master_set_option (nm_device_get_platform (device),
ifindex,
attr,
value);
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
if (!ret)
logging: remove LOGD_HW alias for LOGD_PLATFORM Since commit 1495853e01299d1eb33d4eb10b91e37dcf855cc8, LOGD_HW is renamed to LOGD_PLATFORM. Remove the internal usage of the deprecated name.
2016-10-06 21:28:40 +02:00
_LOGW (LOGD_PLATFORM, "failed to set bonding attribute '%s' to '%s'", attr, value);
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
return ret;
core: use nm_platform_master_*_option() for bonds Acked-by: Dan Winship <danw@gnome.org>
2013-08-28 10:38:46 +02:00
}
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
#define _set_bond_attr_take(device, attr, value) \
G_STMT_START { \
gs_free char *_tmp = (value); \
\
_set_bond_attr (device, NM_SETTING_BOND_OPTION_ARP_IP_TARGET, _tmp); \
} G_STMT_END
#define _set_bond_attr_printf(device, attr, fmt, ...) \
_set_bond_attr_take ((device), (attr), g_strdup_printf (fmt, __VA_ARGS__))
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
static gboolean
bond: ignore miimon option only when it is zero The default value for miimon, when missing in the setting, is 0 if arp_interval is != 0, and 100 otherwise. So, when generating a connection, let's ignore miimon=0 (which means that miimon is disabled) and accept any other value. Adding miimon=100 does not cause any harm to the connection assumption. While at it, slightly improve the code: ignore_if_zero() is not useful for 'updelay','downdelay','arp_interval' because zero is their default value, so introduce a new function that checks if the value is the default (and specially handles 'miimon'). Reported-by: Taketo Kabe <rkabe@vega.pgw.jp> https://bugzilla.redhat.com/show_bug.cgi?id=1463077
2017-06-21 11:19:03 +02:00
ignore_option (NMSettingBond *s_bond, const char *option, const char *value)
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
{
bond: ignore miimon option only when it is zero The default value for miimon, when missing in the setting, is 0 if arp_interval is != 0, and 100 otherwise. So, when generating a connection, let's ignore miimon=0 (which means that miimon is disabled) and accept any other value. Adding miimon=100 does not cause any harm to the connection assumption. While at it, slightly improve the code: ignore_if_zero() is not useful for 'updelay','downdelay','arp_interval' because zero is their default value, so introduce a new function that checks if the value is the default (and specially handles 'miimon'). Reported-by: Taketo Kabe <rkabe@vega.pgw.jp> https://bugzilla.redhat.com/show_bug.cgi?id=1463077
2017-06-21 11:19:03 +02:00
const char *defvalue;
if (nm_streq0 (option, NM_SETTING_BOND_OPTION_MIIMON)) {
/* The default value for miimon, when missing in the setting, is
* 0 if arp_interval is != 0, and 100 otherwise. So, let's ignore
* miimon=0 (which means that miimon is disabled) and accept any
* other value. Adding miimon=100 does not cause any harm.
*/
defvalue = "0";
} else
defvalue = nm_setting_bond_get_option_default (s_bond, option);
return nm_streq0 (value, defvalue);
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
}
static void
update_connection (NMDevice *device, NMConnection *connection)
{
NMSettingBond *s_bond = nm_connection_get_setting_bond (connection);
int ifindex = nm_device_get_ifindex (device);
bond: add only supported options to the generated connection Upstream commit [1] changed in the kernel the default value of tlb_dynamic_lb bond from 1 to 0 when the mode is not tlb. This is not wrong, as the option value doesn't really matter for other modes, but it breaks the connection matching because we read back a 0 value when we expect a default of 1. Fix this in a generic way by ignoring altogether options that are not relevant for the current bond mode, because they are removed from the connection during normalization. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8b426dc54cf4056984bab7dfa48c92ee79a46434 https://bugzilla.redhat.com/show_bug.cgi?id=1457909
2017-06-05 14:48:08 +02:00
NMBondMode mode = NM_BOND_MODE_UNKNOWN;
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
const char **options;
if (!s_bond) {
s_bond = (NMSettingBond *) nm_setting_bond_new ();
nm_connection_add_setting (connection, (NMSetting *) s_bond);
}
/* Read bond options from sysfs and update the Bond setting to match */
options = nm_setting_bond_get_valid_options (s_bond);
bond: add only supported options to the generated connection Upstream commit [1] changed in the kernel the default value of tlb_dynamic_lb bond from 1 to 0 when the mode is not tlb. This is not wrong, as the option value doesn't really matter for other modes, but it breaks the connection matching because we read back a 0 value when we expect a default of 1. Fix this in a generic way by ignoring altogether options that are not relevant for the current bond mode, because they are removed from the connection during normalization. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8b426dc54cf4056984bab7dfa48c92ee79a46434 https://bugzilla.redhat.com/show_bug.cgi?id=1457909
2017-06-05 14:48:08 +02:00
for (; *options; options++) {
bond: fix re-assuming of connections When a value of a TYPE_BOTH option is read back from kernel it contains both string and numeric values ("balance-rr 0"), so we must chop off the number before adding the option to the setting. Also change the default values of options to the string form so that the option matching logic works.
2016-03-16 11:22:07 +01:00
char *p;
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
gs_free char *value = nm_platform_sysctl_master_get_option (nm_device_get_platform (device),
ifindex,
*options);
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
bond: fix crash in update_connection() The value read from sysfs can be NULL. Fixes: 2324410a75dda09d4c46362d29fb84048bfe9fcd
2017-02-07 11:05:01 +01:00
if ( value
&& _nm_setting_bond_get_option_type (s_bond, *options) == NM_BOND_OPTION_TYPE_BOTH) {
bond: fix re-assuming of connections When a value of a TYPE_BOTH option is read back from kernel it contains both string and numeric values ("balance-rr 0"), so we must chop off the number before adding the option to the setting. Also change the default values of options to the string form so that the option matching logic works.
2016-03-16 11:22:07 +01:00
p = strchr (value, ' ');
if (p)
*p = '\0';
}
bond: small cleanups * Use an enum instead of a string, is faster for comparisons. * Add debug assertions * Have NMBondMode enum correspond to Kernel numbering
2020-04-09 19:18:27 +02:00
if (mode == NM_BOND_MODE_UNKNOWN) {
if (value && nm_streq (*options, NM_SETTING_BOND_OPTION_MODE))
mode = _nm_setting_bond_mode_from_string (value);
if (mode == NM_BOND_MODE_UNKNOWN)
continue;
}
bond: add only supported options to the generated connection Upstream commit [1] changed in the kernel the default value of tlb_dynamic_lb bond from 1 to 0 when the mode is not tlb. This is not wrong, as the option value doesn't really matter for other modes, but it breaks the connection matching because we read back a 0 value when we expect a default of 1. Fix this in a generic way by ignoring altogether options that are not relevant for the current bond mode, because they are removed from the connection during normalization. [1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=8b426dc54cf4056984bab7dfa48c92ee79a46434 https://bugzilla.redhat.com/show_bug.cgi?id=1457909
2017-06-05 14:48:08 +02:00
if (!_nm_setting_bond_option_supported (*options, mode))
continue;
bond: fix re-assuming of connections When a value of a TYPE_BOTH option is read back from kernel it contains both string and numeric values ("balance-rr 0"), so we must chop off the number before adding the option to the setting. Also change the default values of options to the string form so that the option matching logic works.
2016-03-16 11:22:07 +01:00
if ( value
&& value[0]
bond: ignore miimon option only when it is zero The default value for miimon, when missing in the setting, is 0 if arp_interval is != 0, and 100 otherwise. So, when generating a connection, let's ignore miimon=0 (which means that miimon is disabled) and accept any other value. Adding miimon=100 does not cause any harm to the connection assumption. While at it, slightly improve the code: ignore_if_zero() is not useful for 'updelay','downdelay','arp_interval' because zero is their default value, so introduce a new function that checks if the value is the default (and specially handles 'miimon'). Reported-by: Taketo Kabe <rkabe@vega.pgw.jp> https://bugzilla.redhat.com/show_bug.cgi?id=1463077
2017-06-21 11:19:03 +02:00
&& !ignore_option (s_bond, *options, value)) {
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
/* Replace " " with "," for arp_ip_targets from the kernel */
bond: use defines for sysfs attribute names Previously we used defines for bond option names and used string literals for their attribute names in sysfs. But they are the same by definition so let's use defines also for attributes.
2016-07-05 22:24:47 +02:00
if (strcmp (*options, NM_SETTING_BOND_OPTION_ARP_IP_TARGET) == 0) {
bond: fix re-assuming of connections When a value of a TYPE_BOTH option is read back from kernel it contains both string and numeric values ("balance-rr 0"), so we must chop off the number before adding the option to the setting. Also change the default values of options to the string form so that the option matching logic works.
2016-03-16 11:22:07 +01:00
for (p = value; *p; p++) {
core: implement update_connection() for bonds
2013-08-28 11:47:44 +02:00
if (*p == ' ')
*p = ',';
}
}
nm_setting_bond_add_option (s_bond, *options, value);
}
}
}
core: refactor update_slave_connection() as virtual function of master device Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-06-18 20:17:57 +02:00
static gboolean
master_update_slave_connection (NMDevice *self,
NMDevice *slave,
NMConnection *connection,
GError **error)
{
g_object_set (nm_connection_get_setting_connection (connection),
NM_SETTING_CONNECTION_MASTER, nm_device_get_iface (self),
NM_SETTING_CONNECTION_SLAVE_TYPE, NM_SETTING_BOND_SETTING_NAME,
NULL);
return TRUE;
}
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
static void
bond: use platform functions for writing sysfs Refactor writing the bonding options, to use again the platform function for the sysfs options. Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 20:15:03 +02:00
set_arp_targets (NMDevice *device,
bond: improve compatibility check of options and modes We print an error when the write of a bond options fails as this is considered an effect of a wrong configuration (or a bug in the checks done by NM) that the user should notice. But not all options are supported in all bonding modes and so we ignore some unsupported options for the current mode to avoid populating logs with useless errors. Improve the code there by using a more generic approach and synchronize the mode/option compatibility table with kernel (file drivers/net/bonding/bond_options.c). https://bugzilla.gnome.org/show_bug.cgi?id=767776 https://bugzilla.redhat.com/show_bug.cgi?id=1352131
2016-07-04 16:25:39 +02:00
NMBondMode mode,
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
const char *cur_arp_ip_target,
const char *new_arp_ip_target)
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
{
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
gs_unref_ptrarray GPtrArray *free_list = NULL;
gs_free const char **cur_strv = NULL;
gs_free const char **new_strv = NULL;
gsize cur_len;
gsize new_len;
all: replace g_strsplit_set() by nm_utils_strsplit_set*()
2019-04-04 13:06:00 +02:00
gsize i;
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
gsize j;
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
cur_strv = nm_utils_strsplit_set_full (cur_arp_ip_target, NM_ASCII_SPACES, NM_UTILS_STRSPLIT_SET_FLAGS_STRSTRIP);
new_strv = nm_utils_bond_option_arp_ip_targets_split (new_arp_ip_target);
cur_len = NM_PTRARRAY_LEN (cur_strv);
new_len = NM_PTRARRAY_LEN (new_strv);
bond: use platform functions for writing sysfs Refactor writing the bonding options, to use again the platform function for the sysfs options. Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 20:15:03 +02:00
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
if (new_len > 0) {
for (j = 0, i = 0; i < new_len; i++) {
const char *s;
in_addr_t a4;
s = new_strv[i];
if (nm_utils_parse_inaddr_bin (AF_INET, s, NULL, &a4)) {
char sbuf[INET_ADDRSTRLEN];
_nm_utils_inet4_ntop (a4, sbuf);
if (!nm_streq (s, sbuf)) {
if (!free_list)
free_list = g_ptr_array_new_with_free_func (g_free);
s = g_strdup (sbuf);
g_ptr_array_add (free_list, (gpointer) s);
}
}
if (nm_utils_strv_find_first ((char **) new_strv, i, s) < 0)
new_strv[j++] = s;
}
new_strv[j] = NULL;
new_len = j;
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
}
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
if ( cur_len == 0
&& new_len == 0)
return;
if (_nm_utils_strv_equal ((char **) cur_strv, (char **) new_strv))
return;
for (i = 0; i < cur_len; i++)
_set_bond_attr_printf (device, NM_SETTING_BOND_OPTION_ARP_IP_TARGET, "-%s", cur_strv[i]);
for (i = 0; i < new_len; i++)
_set_bond_attr_printf (device, NM_SETTING_BOND_OPTION_ARP_IP_TARGET, "+%s", new_strv[i]);
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
}
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
/*
* Sets bond attribute stored in the option hashtable or
* the default value if no value was set.
*/
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
static void
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (NMDevice *device,
NMSettingBond *s_bond,
const char *opt)
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
{
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
NMDeviceBond *self = NM_DEVICE_BOND (device);
bond: log only skipped bond options if they are set in the profile (cherry picked from commit 3a25b3bfc7bbfe93c66e2cdac98ea1868f0bbed7)
2020-07-03 12:43:34 +02:00
const char *value;
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
bond: log only skipped bond options if they are set in the profile (cherry picked from commit 3a25b3bfc7bbfe93c66e2cdac98ea1868f0bbed7)
2020-07-03 12:43:34 +02:00
value = nm_setting_bond_get_option_or_default (s_bond, opt);
if (!value) {
if ( _LOGT_ENABLED (LOGD_BOND)
&& nm_setting_bond_get_option_by_name (s_bond, opt))
_LOGT (LOGD_BOND, "bond option '%s' not set as it conflicts with other options", opt);
return;
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
}
bond: log only skipped bond options if they are set in the profile (cherry picked from commit 3a25b3bfc7bbfe93c66e2cdac98ea1868f0bbed7)
2020-07-03 12:43:34 +02:00
_set_bond_attr (device, opt, value);
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
}
bond: avoid setting "active_slave" option without interface enslaved Kernel will reject setting "active_slave", if the interface is not enslaved or not up. We already handle that by setting the option whenever we enslave an interface. However, we also must not set it initially, otherwise we get an ugly error log message: NetworkManager[939]: <debug> [1594709143.7459] platform-linux: sysctl: setting net:/sys/class/net/bond99/bonding/active_slave to eth1 (current value is ) NetworkManager[939]: <error> [1594709143.7459] platform-linux: sysctl: failed to set bonding/active_slave to eth1: (22) Invalid argument NetworkManager[939]: <warn> [1594709143.7460] device (bond99): failed to set bonding attribute active_slave to eth1 ... kernel: bond99: (slave eth1): Device is not bonding slave kernel: bond99: option active_slave: invalid value (eth1) See-also: https://bugzilla.redhat.com/show_bug.cgi?id=1856640 https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/577
2020-07-14 09:16:51 +02:00
static void
set_bond_attr_active_slave (NMDevice *device, NMSettingBond *s_bond)
{
NMDeviceBond *self = NM_DEVICE_BOND (device);
const NMPlatformLink *plink;
const char *value;
const char *error_reason;
int ifindex;
value = nm_setting_bond_get_option_or_default (s_bond, NM_SETTING_BOND_OPTION_ACTIVE_SLAVE);
if (!value)
return;
if (!nm_str_is_empty (value)) {
ifindex = nm_device_get_ifindex (device);
plink = nm_platform_link_get_by_ifname (nm_device_get_platform (device), value);
if (!plink)
error_reason = "does not exist";
else if (plink->master != ifindex)
error_reason = "is not yet enslaved";
else if (!NM_FLAGS_HAS (plink->n_ifi_flags, IFF_UP))
error_reason = "is not up";
else
error_reason = NULL;
if (error_reason) {
_LOGT (LOGD_BOND, "bond option 'active_slave' not set as device \"%s\" %s", value, error_reason);
return;
}
}
_set_bond_attr (device, NM_SETTING_BOND_OPTION_ACTIVE_SLAVE, value);
}
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
static gboolean
apply_bonding_config (NMDeviceBond *self)
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
{
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
NMDevice *device = NM_DEVICE (self);
bond: use platform functions for writing sysfs Refactor writing the bonding options, to use again the platform function for the sysfs options. Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 20:15:03 +02:00
int ifindex = nm_device_get_ifindex (device);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
NMSettingBond *s_bond;
bond: improve compatibility check of options and modes We print an error when the write of a bond options fails as this is considered an effect of a wrong configuration (or a bug in the checks done by NM) that the user should notice. But not all options are supported in all bonding modes and so we ignore some unsupported options for the current mode to avoid populating logs with useless errors. Improve the code there by using a more generic approach and synchronize the mode/option compatibility table with kernel (file drivers/net/bonding/bond_options.c). https://bugzilla.gnome.org/show_bug.cgi?id=767776 https://bugzilla.redhat.com/show_bug.cgi?id=1352131
2016-07-04 16:25:39 +02:00
NMBondMode mode;
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
const char *mode_str;
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
gs_free char *cur_arp_ip_target = NULL;
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
core: don't cast return value of nm_device_get_applied_setting()
2018-10-22 13:06:27 +02:00
s_bond = nm_device_get_applied_setting (device, NM_TYPE_SETTING_BOND);
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
g_return_val_if_fail (s_bond, FALSE);
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
mode_str = nm_setting_bond_get_option_or_default (s_bond, NM_SETTING_BOND_OPTION_MODE);
bond: improve compatibility check of options and modes We print an error when the write of a bond options fails as this is considered an effect of a wrong configuration (or a bug in the checks done by NM) that the user should notice. But not all options are supported in all bonding modes and so we ignore some unsupported options for the current mode to avoid populating logs with useless errors. Improve the code there by using a more generic approach and synchronize the mode/option compatibility table with kernel (file drivers/net/bonding/bond_options.c). https://bugzilla.gnome.org/show_bug.cgi?id=767776 https://bugzilla.redhat.com/show_bug.cgi?id=1352131
2016-07-04 16:25:39 +02:00
mode = _nm_setting_bond_mode_from_string (mode_str);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
g_return_val_if_fail (mode != NM_BOND_MODE_UNKNOWN, FALSE);
bond: improve compatibility check of options and modes We print an error when the write of a bond options fails as this is considered an effect of a wrong configuration (or a bug in the checks done by NM) that the user should notice. But not all options are supported in all bonding modes and so we ignore some unsupported options for the current mode to avoid populating logs with useless errors. Improve the code there by using a more generic approach and synchronize the mode/option compatibility table with kernel (file drivers/net/bonding/bond_options.c). https://bugzilla.gnome.org/show_bug.cgi?id=767776 https://bugzilla.redhat.com/show_bug.cgi?id=1352131
2016-07-04 16:25:39 +02:00
/* Set mode first, as some other options (e.g. arp_interval) are valid
* only for certain modes.
*/
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_MODE);
bond: use defines for sysfs attribute names Previously we used defines for bond option names and used string literals for their attribute names in sysfs. But they are the same by definition so let's use defines also for attributes.
2016-07-05 22:24:47 +02:00
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_MIIMON);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_UPDELAY);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_DOWNDELAY);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_ARP_INTERVAL);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_ARP_VALIDATE);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_PRIMARY);
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
bond: improve compatibility check of options and modes We print an error when the write of a bond options fails as this is considered an effect of a wrong configuration (or a bug in the checks done by NM) that the user should notice. But not all options are supported in all bonding modes and so we ignore some unsupported options for the current mode to avoid populating logs with useless errors. Improve the code there by using a more generic approach and synchronize the mode/option compatibility table with kernel (file drivers/net/bonding/bond_options.c). https://bugzilla.gnome.org/show_bug.cgi?id=767776 https://bugzilla.redhat.com/show_bug.cgi?id=1352131
2016-07-04 16:25:39 +02:00
/* ARP targets: clear and initialize the list */
device/bond: rework setting of arp_ip_target bond options - the arp_ip_target option in the settings might not have normalized IP addresses or duplicates. If there would be duplicates, setting them twice would fail with EINVAL. Hence, first normalize them and make them unique. - if what we want to set is identical to what is already set, don't do anything. (cherry picked from commit 6a923a5d5786f55f3aee14cf8d0b1c7f8a9cb76f)
2020-06-30 15:33:10 +02:00
cur_arp_ip_target = nm_platform_sysctl_master_get_option (nm_device_get_platform (device),
ifindex,
NM_SETTING_BOND_OPTION_ARP_IP_TARGET);
set_arp_targets (device,
mode,
cur_arp_ip_target,
nm_setting_bond_get_option_or_default (s_bond, NM_SETTING_BOND_OPTION_ARP_IP_TARGET));
bond: add some missing options https://bugzilla.redhat.com/show_bug.cgi?id=1299103
2016-03-15 17:37:06 +01:00
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_AD_ACTOR_SYSTEM);
bond: avoid setting "active_slave" option without interface enslaved Kernel will reject setting "active_slave", if the interface is not enslaved or not up. We already handle that by setting the option whenever we enslave an interface. However, we also must not set it initially, otherwise we get an ugly error log message: NetworkManager[939]: <debug> [1594709143.7459] platform-linux: sysctl: setting net:/sys/class/net/bond99/bonding/active_slave to eth1 (current value is ) NetworkManager[939]: <error> [1594709143.7459] platform-linux: sysctl: failed to set bonding/active_slave to eth1: (22) Invalid argument NetworkManager[939]: <warn> [1594709143.7460] device (bond99): failed to set bonding attribute active_slave to eth1 ... kernel: bond99: (slave eth1): Device is not bonding slave kernel: bond99: option active_slave: invalid value (eth1) See-also: https://bugzilla.redhat.com/show_bug.cgi?id=1856640 https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/577
2020-07-14 09:16:51 +02:00
set_bond_attr_active_slave (device, s_bond);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_AD_ACTOR_SYS_PRIO);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_AD_SELECT);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_AD_USER_PORT_KEY);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_ALL_SLAVES_ACTIVE);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_ARP_ALL_TARGETS);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_FAIL_OVER_MAC);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_LACP_RATE);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_LP_INTERVAL);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_MIN_LINKS);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_PACKETS_PER_SLAVE);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_PRIMARY_RESELECT);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_RESEND_IGMP);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_TLB_DYNAMIC_LB);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_USE_CARRIER);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_XMIT_HASH_POLICY);
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_NUM_GRAT_ARP);
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
return TRUE;
trivial: move vpn/bonding functions out of nm-system As nm-system only contained auxiliary functions which fit in their respective modules better, it can be safely removed.
2013-07-05 16:41:51 +02:00
}
bond: set bond device options during activation
2012-02-29 12:39:57 -06:00
static NMActStageReturn
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
act_stage1_prepare (NMDevice *device, NMDeviceStateReason *out_failure_reason)
bond: set bond device options during activation
2012-02-29 12:39:57 -06:00
{
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
NMDeviceBond *self = NM_DEVICE_BOND (device);
bond: set bond device options during activation
2012-02-29 12:39:57 -06:00
NMActStageReturn ret = NM_ACT_STAGE_RETURN_SUCCESS;
bond: handle bond options more gracefully Support new bonding options and set them carefully. The options cannot be set arbitrarily because they interfere with each other. This commit is forward-ported from rhel-6.5, see patch rh901662-bond-more-options.patch, originally written by Dan Williams. https://bugzilla.redhat.com/show_bug.cgi?id=901662 https://bugzilla.redhat.com/show_bug.cgi?id=905532 Co-Authored-By: Dan Williams <dcbw@redhat.com> Signed-off-by: Thomas Haller <thaller@redhat.com>
2013-10-01 18:27:25 +02:00
/* Interface must be down to set bond options */
device/bond: cleanup act-stage return values in NMDeviceBond's act_stage1_prepare()
2019-08-21 16:57:55 +02:00
nm_device_take_down (device, TRUE);
if (!apply_bonding_config (self))
ret = NM_ACT_STAGE_RETURN_FAILURE;
else {
if (!nm_device_hw_addr_set_cloned (device,
nm_device_get_applied_connection (device),
FALSE))
ret = NM_ACT_STAGE_RETURN_FAILURE;
}
nm_device_bring_up (device, TRUE, NULL);
bond: set bond device options during activation
2012-02-29 12:39:57 -06:00
return ret;
}
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
static gboolean
core: make assumed activations go through all the stages Assumed slave connections need to be added to their master devices, which didn't used to happen because the devices activating assumed connections jumped directly to stage3, bypassing all the master/slave handling stuff. Instead, make all assumed connections go through all activation stages, but make sure that things which touch the device don't get done for assumed connections. This requires moving the master/slave code out of the override-able class methods because we need to call the master/slave code for assumed connections, but we don't want to call the override-able class activation methods.
2013-11-07 01:08:02 -06:00
enslave_slave (NMDevice *device,
NMDevice *slave,
NMConnection *connection,
gboolean configure)
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
{
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
NMDeviceBond *self = NM_DEVICE_BOND (device);
device: don't bother passing no_firmware uselessly It's actually confusing in cases it's not handled. Besides, nm_device_bring_up() is just happy about NULL argument.
2018-05-22 10:30:37 +02:00
gboolean success = TRUE;
core: rework master/slave handling
2012-11-14 14:05:30 -06:00
const char *slave_iface = nm_device_get_ip_iface (slave);
bond: set the active slave only after enslavement The kernel doesn't accept an interface to be set as active_slave until the interface is enslaved to the bond. Delay the initialization of the property.
2017-03-01 14:55:51 +01:00
NMConnection *master_con;
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
core: improve handling of NPAR/SR-IOV devices (rh #804527) Use the new kernel physical_port_id interface property to recognize when two devices are just virtual devices sharing the same physical port, and refuse to bond/team multiple slaves on the same port.
2013-10-11 14:59:26 -04:00
nm_device_master_check_slave_physical_port (device, slave, LOGD_BOND);
core: make assumed activations go through all the stages Assumed slave connections need to be added to their master devices, which didn't used to happen because the devices activating assumed connections jumped directly to stage3, bypassing all the master/slave handling stuff. Instead, make all assumed connections go through all activation stages, but make sure that things which touch the device don't get done for assumed connections. This requires moving the master/slave code out of the override-able class methods because we need to call the master/slave code for assumed connections, but we don't want to call the override-able class activation methods.
2013-11-07 01:08:02 -06:00
if (configure) {
nm_device_take_down (slave, TRUE);
device: don't use platform singleton getter in device subclasses Reduce the use of NM_PLATFORM_GET / nm_platform_get() to get the platform singleton instance. For one, this is a step towards supporting namespaces, where we need to use different NMNetns/NMPlatform instances depending on in which namespace the device lives. Also, we should reduce our use of singletons. They are difficult to coordinate on shutdown. Instead there should be a clear order of dependencies, expressed by owning a reference to those singelton instances. We already own a reference to the platform singelton, so use it and avoid NM_PLATFORM_GET. (cherry picked from commit 94d9ee129d85ff926a10de78fedcc6b57cdcdb19)
2017-04-18 12:09:02 +02:00
success = nm_platform_link_enslave (nm_device_get_platform (device),
platform: add self argument to platform functions Most nm_platform_*() functions operate on the platform singleton nm_platform_get(). That made sense because the NMPlatform instance was mainly to hook fake platform for testing. While the implicit argument saved some typing, I think explicit is better. Especially, because NMPlatform could become a more usable object then just a hook for testing. With this change, NMPlatform instances can be used individually, not only as a singleton instance. Before this change, the constructor of NMLinuxPlatform could not call any nm_platform_*() functions because the singleton was not yet initialized. We could only instantiate an incomplete instance, register it via nm_platform_setup(), and then complete initialization via singleton->setup(). With this change, we can create and fully initialize NMPlatform instances before/without setting them up them as singleton. Also, currently there is no clear distinction between functions that operate on the NMPlatform instance, and functions that can be used stand-alone (e.g. nm_platform_ip4_address_to_string()). The latter can not be mocked for testing. With this change, the distinction becomes obvious. That is also useful because it becomes clearer which functions make use of the platform cache and which not. Inside nm-linux-platform.c, continue the pattern that the self instance is named @platform. That makes sense because its type is NMPlatform, and not NMLinuxPlatform what we would expect from a paramter named @self. This is a major diff that causes some pain when rebasing. Try to rebase to the parent commit of this commit as a first step. Then rebase on top of this commit using merge-strategy "ours".
2015-04-18 12:36:09 +02:00
nm_device_get_ip_ifindex (device),
core: make assumed activations go through all the stages Assumed slave connections need to be added to their master devices, which didn't used to happen because the devices activating assumed connections jumped directly to stage3, bypassing all the master/slave handling stuff. Instead, make all assumed connections go through all activation stages, but make sure that things which touch the device don't get done for assumed connections. This requires moving the master/slave code out of the override-able class methods because we need to call the master/slave code for assumed connections, but we don't want to call the override-able class activation methods.
2013-11-07 01:08:02 -06:00
nm_device_get_ip_ifindex (slave));
device: don't bother passing no_firmware uselessly It's actually confusing in cases it's not handled. Besides, nm_device_bring_up() is just happy about NULL argument.
2018-05-22 10:30:37 +02:00
nm_device_bring_up (slave, TRUE, NULL);
core: rework master/slave handling
2012-11-14 14:05:30 -06:00
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
if (!success)
return FALSE;
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
_LOGI (LOGD_BOND, "enslaved bond slave %s", slave_iface);
bond: set the active slave only after enslavement The kernel doesn't accept an interface to be set as active_slave until the interface is enslaved to the bond. Delay the initialization of the property.
2017-03-01 14:55:51 +01:00
/* The active_slave option can be set only after the interface is enslaved */
master_con = nm_device_get_applied_connection (device);
if (master_con) {
NMSettingBond *s_bond = nm_connection_get_setting_bond (master_con);
const char *active;
if (s_bond) {
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
active = nm_setting_bond_get_option_or_default (s_bond,
NM_SETTING_BOND_OPTION_ACTIVE_SLAVE);
if (nm_streq0 (active, nm_device_get_iface (slave))) {
device: don't use platform singleton getter in device subclasses Reduce the use of NM_PLATFORM_GET / nm_platform_get() to get the platform singleton instance. For one, this is a step towards supporting namespaces, where we need to use different NMNetns/NMPlatform instances depending on in which namespace the device lives. Also, we should reduce our use of singletons. They are difficult to coordinate on shutdown. Instead there should be a clear order of dependencies, expressed by owning a reference to those singelton instances. We already own a reference to the platform singelton, so use it and avoid NM_PLATFORM_GET. (cherry picked from commit 94d9ee129d85ff926a10de78fedcc6b57cdcdb19)
2017-04-18 12:09:02 +02:00
nm_platform_sysctl_master_set_option (nm_device_get_platform (device),
bond: set the active slave only after enslavement The kernel doesn't accept an interface to be set as active_slave until the interface is enslaved to the bond. Delay the initialization of the property.
2017-03-01 14:55:51 +01:00
nm_device_get_ifindex (device),
bond: avoid setting "active_slave" option without interface enslaved Kernel will reject setting "active_slave", if the interface is not enslaved or not up. We already handle that by setting the option whenever we enslave an interface. However, we also must not set it initially, otherwise we get an ugly error log message: NetworkManager[939]: <debug> [1594709143.7459] platform-linux: sysctl: setting net:/sys/class/net/bond99/bonding/active_slave to eth1 (current value is ) NetworkManager[939]: <error> [1594709143.7459] platform-linux: sysctl: failed to set bonding/active_slave to eth1: (22) Invalid argument NetworkManager[939]: <warn> [1594709143.7460] device (bond99): failed to set bonding attribute active_slave to eth1 ... kernel: bond99: (slave eth1): Device is not bonding slave kernel: bond99: option active_slave: invalid value (eth1) See-also: https://bugzilla.redhat.com/show_bug.cgi?id=1856640 https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/577
2020-07-14 09:16:51 +02:00
NM_SETTING_BOND_OPTION_ACTIVE_SLAVE,
bond: set the active slave only after enslavement The kernel doesn't accept an interface to be set as active_slave until the interface is enslaved to the bond. Delay the initialization of the property.
2017-03-01 14:55:51 +01:00
active);
_LOGD (LOGD_BOND, "setting slave %s as active one for master %s",
active, nm_device_get_iface (device));
}
}
}
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
} else
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
_LOGI (LOGD_BOND, "bond slave %s was enslaved", slave_iface);
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
return TRUE;
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
}
device: don't return error from release_slave() release_slave() should do the right thing and handle errors as good as it can. There is no value in propagating the error and it's not clear what the caller should do in face of a failure during release.
2015-12-02 09:56:17 +01:00
static void
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
release_slave (NMDevice *device,
NMDevice *slave,
gboolean configure)
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
{
logging: use new logging macros in NMDevice subclasses Signed-off-by: Thomas Haller <thaller@redhat.com>
2014-08-02 15:14:26 +02:00
NMDeviceBond *self = NM_DEVICE_BOND (device);
device: don't bother passing no_firmware uselessly It's actually confusing in cases it's not handled. Besides, nm_device_bring_up() is just happy about NULL argument.
2018-05-22 10:30:37 +02:00
gboolean success;
bond: support the ethernet.cloned-mac-address property
2016-10-21 14:37:05 +02:00
gs_free char *address = NULL;
device: fix crash releasing destroyed slave I encountered this on a WIP branch, but I think it can happen under regular conditions. I think there is no error condition here, and we should do nothing if we have no ifindex. <debug> [1561653068.2192] platform: signal: link removed: 1699: test1p <DOWN;broadcast,multicast> mtu 1500 master 1698 arp 1 veth* init addrgenmode none addr D6:14:45:97:06:75 brd FF:FF:FF:FF:FF:FF driver veth rx:0,0 tx:38,5606 ... <info> [1561653068.2617] device (test1): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') ... <trace> [1561653068.2635] device[0x564058c73750] (test1p): sys-iface-state: external -> removed <debug> [1561653068.2635] device[0x564058c73750] (test1p): unrealize (ifindex 1699) <debug> [1561653068.2636] device[0x564058c73750] (test1p): parent: clear <trace> [1561653068.2636] device[0x564058b98eb0] (vethbr): mtu: commit-mtu... <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!by-type,!user-explicit,!user-settings,!user-udev,!is-slave=0x10/0x1479/unmanaged/unrealized], set-unmanaged [platform-init=0x10]) <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!user-settings=0x10/0x51/unmanaged/unrealized], forget [parent,by-type,user-explicit,user-udev,external-down,is-slave=0x1c2c]) <info> [1561653068.2639] device (test1p): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') <debug> [1561653068.2640] device[0x564058c73750] (test1p): deactivating device (reason 'unmanaged') [3] <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip4-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip6-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'dhcp6' not pending (expected) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'autoconf6' not pending (expected) <debug> [1561653068.2640] rules-manager: sync <debug> [1561653068.2640] device[0x564058c73750] (test1p): set metered value 0 <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip4-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip6-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2644] device[0x564058b98eb0] (vethbr): slave test1p state change 100 (activated) -> 10 (unmanaged) <trace> [1561653068.2644] device[0x564058b98eb0] (vethbr): master: release one slave 0x564058c73750/test1p ((src/platform/nm-platform.c:2002)): assertion '<dropped>' failed backtrace: ... #3 0x0000564057fb713e _nm_g_return_if_fail_warning (NetworkManager) #4 0x000056405808b37c release_slave (NetworkManager) #5 0x0000564058079aef nm_device_master_release_one_slave (NetworkManager) #6 0x00005640580844d7 slave_state_changed (NetworkManager) #7 0x00007efc24833fae ffi_call_unix64 (libffi.so.6) #8 0x00007efc2483396f ffi_call (libffi.so.6) #9 0x00007efc29b836e5 g_cclosure_marshal_generic (libgobject-2.0.so.0) #10 0x00007efc29b82c1d g_closure_invoke (libgobject-2.0.so.0) #11 0x00007efc29b96173 signal_emit_unlocked_R (libgobject-2.0.so.0) #12 0x00007efc29b9f29a g_signal_emit_valist (libgobject-2.0.so.0) #13 0x00007efc29b9f893 g_signal_emit (libgobject-2.0.so.0) #14 0x000056405807ab20 _set_state_full (NetworkManager) #15 0x000056405807d803 nm_device_unrealize (NetworkManager) #16 0x0000564057f6072c _platform_link_cb_idle (NetworkManager) #17 0x00007efc296a01db g_idle_dispatch (libglib-2.0.so.0) ...
2019-06-28 17:31:40 +02:00
int ifindex_slave;
device: fix releasing slaves Not all masters type have a platform link and so it's wrong to check for it to decide whether the slave should be really released. Move the check to master devices that need it (bond, bridge and team). OVS ports don't need the check because they don't call to platform to remove a slave. https://bugzilla.redhat.com/show_bug.cgi?id=1733709
2019-07-31 11:40:35 +02:00
int ifindex;
device: avoid unnecessary check for existing device in release_slave() implementations
2019-08-01 14:56:07 +02:00
if (configure) {
ifindex = nm_device_get_ifindex (device);
if ( ifindex <= 0
|| !nm_platform_link_get (nm_device_get_platform (device), ifindex))
configure = FALSE;
}
device: fix crash releasing destroyed slave I encountered this on a WIP branch, but I think it can happen under regular conditions. I think there is no error condition here, and we should do nothing if we have no ifindex. <debug> [1561653068.2192] platform: signal: link removed: 1699: test1p <DOWN;broadcast,multicast> mtu 1500 master 1698 arp 1 veth* init addrgenmode none addr D6:14:45:97:06:75 brd FF:FF:FF:FF:FF:FF driver veth rx:0,0 tx:38,5606 ... <info> [1561653068.2617] device (test1): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') ... <trace> [1561653068.2635] device[0x564058c73750] (test1p): sys-iface-state: external -> removed <debug> [1561653068.2635] device[0x564058c73750] (test1p): unrealize (ifindex 1699) <debug> [1561653068.2636] device[0x564058c73750] (test1p): parent: clear <trace> [1561653068.2636] device[0x564058b98eb0] (vethbr): mtu: commit-mtu... <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!by-type,!user-explicit,!user-settings,!user-udev,!is-slave=0x10/0x1479/unmanaged/unrealized], set-unmanaged [platform-init=0x10]) <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!user-settings=0x10/0x51/unmanaged/unrealized], forget [parent,by-type,user-explicit,user-udev,external-down,is-slave=0x1c2c]) <info> [1561653068.2639] device (test1p): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') <debug> [1561653068.2640] device[0x564058c73750] (test1p): deactivating device (reason 'unmanaged') [3] <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip4-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip6-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'dhcp6' not pending (expected) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'autoconf6' not pending (expected) <debug> [1561653068.2640] rules-manager: sync <debug> [1561653068.2640] device[0x564058c73750] (test1p): set metered value 0 <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip4-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip6-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2644] device[0x564058b98eb0] (vethbr): slave test1p state change 100 (activated) -> 10 (unmanaged) <trace> [1561653068.2644] device[0x564058b98eb0] (vethbr): master: release one slave 0x564058c73750/test1p ((src/platform/nm-platform.c:2002)): assertion '<dropped>' failed backtrace: ... #3 0x0000564057fb713e _nm_g_return_if_fail_warning (NetworkManager) #4 0x000056405808b37c release_slave (NetworkManager) #5 0x0000564058079aef nm_device_master_release_one_slave (NetworkManager) #6 0x00005640580844d7 slave_state_changed (NetworkManager) #7 0x00007efc24833fae ffi_call_unix64 (libffi.so.6) #8 0x00007efc2483396f ffi_call (libffi.so.6) #9 0x00007efc29b836e5 g_cclosure_marshal_generic (libgobject-2.0.so.0) #10 0x00007efc29b82c1d g_closure_invoke (libgobject-2.0.so.0) #11 0x00007efc29b96173 signal_emit_unlocked_R (libgobject-2.0.so.0) #12 0x00007efc29b9f29a g_signal_emit_valist (libgobject-2.0.so.0) #13 0x00007efc29b9f893 g_signal_emit (libgobject-2.0.so.0) #14 0x000056405807ab20 _set_state_full (NetworkManager) #15 0x000056405807d803 nm_device_unrealize (NetworkManager) #16 0x0000564057f6072c _platform_link_cb_idle (NetworkManager) #17 0x00007efc296a01db g_idle_dispatch (libglib-2.0.so.0) ...
2019-06-28 17:31:40 +02:00
ifindex_slave = nm_device_get_ip_ifindex (slave);
if (ifindex_slave <= 0)
core,libnm: add VRF support Add VRF support to the daemon. When the device we are activating is a VRF or a VRF's slave, put routes in the table specified by the VRF connection. Also, introduce a VRF device type in libnm.
2019-12-05 10:36:54 +01:00
_LOGD (LOGD_BOND, "bond slave %s is already released", nm_device_get_ip_iface (slave));
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
if (configure) {
bond: support the ethernet.cloned-mac-address property
2016-10-21 14:37:05 +02:00
/* When the last slave is released the bond MAC will be set to a random
* value by kernel; remember the current one and restore it afterwards.
*/
address = g_strdup (nm_device_get_hw_address (device));
device: fix crash releasing destroyed slave I encountered this on a WIP branch, but I think it can happen under regular conditions. I think there is no error condition here, and we should do nothing if we have no ifindex. <debug> [1561653068.2192] platform: signal: link removed: 1699: test1p <DOWN;broadcast,multicast> mtu 1500 master 1698 arp 1 veth* init addrgenmode none addr D6:14:45:97:06:75 brd FF:FF:FF:FF:FF:FF driver veth rx:0,0 tx:38,5606 ... <info> [1561653068.2617] device (test1): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') ... <trace> [1561653068.2635] device[0x564058c73750] (test1p): sys-iface-state: external -> removed <debug> [1561653068.2635] device[0x564058c73750] (test1p): unrealize (ifindex 1699) <debug> [1561653068.2636] device[0x564058c73750] (test1p): parent: clear <trace> [1561653068.2636] device[0x564058b98eb0] (vethbr): mtu: commit-mtu... <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!by-type,!user-explicit,!user-settings,!user-udev,!is-slave=0x10/0x1479/unmanaged/unrealized], set-unmanaged [platform-init=0x10]) <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!user-settings=0x10/0x51/unmanaged/unrealized], forget [parent,by-type,user-explicit,user-udev,external-down,is-slave=0x1c2c]) <info> [1561653068.2639] device (test1p): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') <debug> [1561653068.2640] device[0x564058c73750] (test1p): deactivating device (reason 'unmanaged') [3] <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip4-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip6-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'dhcp6' not pending (expected) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'autoconf6' not pending (expected) <debug> [1561653068.2640] rules-manager: sync <debug> [1561653068.2640] device[0x564058c73750] (test1p): set metered value 0 <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip4-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip6-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2644] device[0x564058b98eb0] (vethbr): slave test1p state change 100 (activated) -> 10 (unmanaged) <trace> [1561653068.2644] device[0x564058b98eb0] (vethbr): master: release one slave 0x564058c73750/test1p ((src/platform/nm-platform.c:2002)): assertion '<dropped>' failed backtrace: ... #3 0x0000564057fb713e _nm_g_return_if_fail_warning (NetworkManager) #4 0x000056405808b37c release_slave (NetworkManager) #5 0x0000564058079aef nm_device_master_release_one_slave (NetworkManager) #6 0x00005640580844d7 slave_state_changed (NetworkManager) #7 0x00007efc24833fae ffi_call_unix64 (libffi.so.6) #8 0x00007efc2483396f ffi_call (libffi.so.6) #9 0x00007efc29b836e5 g_cclosure_marshal_generic (libgobject-2.0.so.0) #10 0x00007efc29b82c1d g_closure_invoke (libgobject-2.0.so.0) #11 0x00007efc29b96173 signal_emit_unlocked_R (libgobject-2.0.so.0) #12 0x00007efc29b9f29a g_signal_emit_valist (libgobject-2.0.so.0) #13 0x00007efc29b9f893 g_signal_emit (libgobject-2.0.so.0) #14 0x000056405807ab20 _set_state_full (NetworkManager) #15 0x000056405807d803 nm_device_unrealize (NetworkManager) #16 0x0000564057f6072c _platform_link_cb_idle (NetworkManager) #17 0x00007efc296a01db g_idle_dispatch (libglib-2.0.so.0) ...
2019-06-28 17:31:40 +02:00
if (ifindex_slave > 0) {
success = nm_platform_link_release (nm_device_get_platform (device),
nm_device_get_ip_ifindex (device),
ifindex_slave);
if (success) {
_LOGI (LOGD_BOND, "released bond slave %s",
nm_device_get_ip_iface (slave));
} else {
_LOGW (LOGD_BOND, "failed to release bond slave %s",
nm_device_get_ip_iface (slave));
}
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
}
device: don't use platform singleton getter in device subclasses Reduce the use of NM_PLATFORM_GET / nm_platform_get() to get the platform singleton instance. For one, this is a step towards supporting namespaces, where we need to use different NMNetns/NMPlatform instances depending on in which namespace the device lives. Also, we should reduce our use of singletons. They are difficult to coordinate on shutdown. Instead there should be a clear order of dependencies, expressed by owning a reference to those singelton instances. We already own a reference to the platform singelton, so use it and avoid NM_PLATFORM_GET. (cherry picked from commit 94d9ee129d85ff926a10de78fedcc6b57cdcdb19)
2017-04-18 12:09:02 +02:00
nm_platform_process_events (nm_device_get_platform (device));
bond: support the ethernet.cloned-mac-address property
2016-10-21 14:37:05 +02:00
if (nm_device_update_hw_address (device))
nm_device_hw_addr_set (device, address, "restore", FALSE);
devices: observe externally-caused master/slave changes (rh #1066706) If a link's "master" property changes unexpectedly (ie, from outside NM), update the master and slave NMDevices to reflect it, without making any changes to them.
2014-02-25 16:44:01 -05:00
/* Kernel bonding code "closes" the slave when releasing it, (which clears
* IFF_UP), so we must bring it back up here to ensure carrier changes and
* other state is noticed by the now-released slave.
*/
device: fix crash releasing destroyed slave I encountered this on a WIP branch, but I think it can happen under regular conditions. I think there is no error condition here, and we should do nothing if we have no ifindex. <debug> [1561653068.2192] platform: signal: link removed: 1699: test1p <DOWN;broadcast,multicast> mtu 1500 master 1698 arp 1 veth* init addrgenmode none addr D6:14:45:97:06:75 brd FF:FF:FF:FF:FF:FF driver veth rx:0,0 tx:38,5606 ... <info> [1561653068.2617] device (test1): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') ... <trace> [1561653068.2635] device[0x564058c73750] (test1p): sys-iface-state: external -> removed <debug> [1561653068.2635] device[0x564058c73750] (test1p): unrealize (ifindex 1699) <debug> [1561653068.2636] device[0x564058c73750] (test1p): parent: clear <trace> [1561653068.2636] device[0x564058b98eb0] (vethbr): mtu: commit-mtu... <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!by-type,!user-explicit,!user-settings,!user-udev,!is-slave=0x10/0x1479/unmanaged/unrealized], set-unmanaged [platform-init=0x10]) <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!user-settings=0x10/0x51/unmanaged/unrealized], forget [parent,by-type,user-explicit,user-udev,external-down,is-slave=0x1c2c]) <info> [1561653068.2639] device (test1p): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') <debug> [1561653068.2640] device[0x564058c73750] (test1p): deactivating device (reason 'unmanaged') [3] <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip4-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip6-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'dhcp6' not pending (expected) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'autoconf6' not pending (expected) <debug> [1561653068.2640] rules-manager: sync <debug> [1561653068.2640] device[0x564058c73750] (test1p): set metered value 0 <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip4-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip6-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2644] device[0x564058b98eb0] (vethbr): slave test1p state change 100 (activated) -> 10 (unmanaged) <trace> [1561653068.2644] device[0x564058b98eb0] (vethbr): master: release one slave 0x564058c73750/test1p ((src/platform/nm-platform.c:2002)): assertion '<dropped>' failed backtrace: ... #3 0x0000564057fb713e _nm_g_return_if_fail_warning (NetworkManager) #4 0x000056405808b37c release_slave (NetworkManager) #5 0x0000564058079aef nm_device_master_release_one_slave (NetworkManager) #6 0x00005640580844d7 slave_state_changed (NetworkManager) #7 0x00007efc24833fae ffi_call_unix64 (libffi.so.6) #8 0x00007efc2483396f ffi_call (libffi.so.6) #9 0x00007efc29b836e5 g_cclosure_marshal_generic (libgobject-2.0.so.0) #10 0x00007efc29b82c1d g_closure_invoke (libgobject-2.0.so.0) #11 0x00007efc29b96173 signal_emit_unlocked_R (libgobject-2.0.so.0) #12 0x00007efc29b9f29a g_signal_emit_valist (libgobject-2.0.so.0) #13 0x00007efc29b9f893 g_signal_emit (libgobject-2.0.so.0) #14 0x000056405807ab20 _set_state_full (NetworkManager) #15 0x000056405807d803 nm_device_unrealize (NetworkManager) #16 0x0000564057f6072c _platform_link_cb_idle (NetworkManager) #17 0x00007efc296a01db g_idle_dispatch (libglib-2.0.so.0) ...
2019-06-28 17:31:40 +02:00
if (ifindex_slave > 0) {
if (!nm_device_bring_up (slave, TRUE, NULL))
_LOGW (LOGD_BOND, "released bond slave could not be brought up.");
}
device: cleanup handling master/slave relationships in NMDevice I found the handling of the master-device very confusing because it was unclear who sets priv->master, and when it should be set. Now: - Setting priv->master (in a slave) always goes together with adding the master to priv->slaves (in the master). Previously, this was done at separate places, so it was not clear if master and slave always agree on their relationship -- in fact, they did not. - There are now three basic functions which do the enslaving/releasing: (1) nm_device_master_add_slave() (2) nm_device_master_enslave_slave() (3) nm_device_master_release_one_slave() Step 3/release basically undoes the 1/add and 2/enslave steps. - completing the enslaving/releasing is now done by (1) nm_device_slave_notify_enslave() (2) nm_device_slave_notify_release() These functions also emit signals like NM_DEVICE_MASTER. - Derived classes no longer emit NM_DEVICE_SLAVES notification. Instead the notification is emited together with NM_DEVICE_MASTER, whenever a slaves changes state. Also, NM_DEVICE_SLAVES list now only exposes slaves that are actually @is_enslaved.
2015-12-02 10:53:16 +01:00
} else {
device: fix crash releasing destroyed slave I encountered this on a WIP branch, but I think it can happen under regular conditions. I think there is no error condition here, and we should do nothing if we have no ifindex. <debug> [1561653068.2192] platform: signal: link removed: 1699: test1p <DOWN;broadcast,multicast> mtu 1500 master 1698 arp 1 veth* init addrgenmode none addr D6:14:45:97:06:75 brd FF:FF:FF:FF:FF:FF driver veth rx:0,0 tx:38,5606 ... <info> [1561653068.2617] device (test1): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') ... <trace> [1561653068.2635] device[0x564058c73750] (test1p): sys-iface-state: external -> removed <debug> [1561653068.2635] device[0x564058c73750] (test1p): unrealize (ifindex 1699) <debug> [1561653068.2636] device[0x564058c73750] (test1p): parent: clear <trace> [1561653068.2636] device[0x564058b98eb0] (vethbr): mtu: commit-mtu... <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!by-type,!user-explicit,!user-settings,!user-udev,!is-slave=0x10/0x1479/unmanaged/unrealized], set-unmanaged [platform-init=0x10]) <debug> [1561653068.2639] device[0x564058c73750] (test1p): unmanaged: flags set to [platform-init,!sleeping,!user-settings=0x10/0x51/unmanaged/unrealized], forget [parent,by-type,user-explicit,user-udev,external-down,is-slave=0x1c2c]) <info> [1561653068.2639] device (test1p): state change: activated -> unmanaged (reason 'unmanaged', sys-iface-state: 'removed') <debug> [1561653068.2640] device[0x564058c73750] (test1p): deactivating device (reason 'unmanaged') [3] <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip4-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): ip6-state: set to 0 (none) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'dhcp6' not pending (expected) <trace> [1561653068.2640] device[0x564058c73750] (test1p): remove_pending_action (0): 'autoconf6' not pending (expected) <debug> [1561653068.2640] rules-manager: sync <debug> [1561653068.2640] device[0x564058c73750] (test1p): set metered value 0 <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip4-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2641] device[0x564058c73750] (test1p): ip6-config: update (commit=1, new-config=(nil)) <debug> [1561653068.2644] device[0x564058b98eb0] (vethbr): slave test1p state change 100 (activated) -> 10 (unmanaged) <trace> [1561653068.2644] device[0x564058b98eb0] (vethbr): master: release one slave 0x564058c73750/test1p ((src/platform/nm-platform.c:2002)): assertion '<dropped>' failed backtrace: ... #3 0x0000564057fb713e _nm_g_return_if_fail_warning (NetworkManager) #4 0x000056405808b37c release_slave (NetworkManager) #5 0x0000564058079aef nm_device_master_release_one_slave (NetworkManager) #6 0x00005640580844d7 slave_state_changed (NetworkManager) #7 0x00007efc24833fae ffi_call_unix64 (libffi.so.6) #8 0x00007efc2483396f ffi_call (libffi.so.6) #9 0x00007efc29b836e5 g_cclosure_marshal_generic (libgobject-2.0.so.0) #10 0x00007efc29b82c1d g_closure_invoke (libgobject-2.0.so.0) #11 0x00007efc29b96173 signal_emit_unlocked_R (libgobject-2.0.so.0) #12 0x00007efc29b9f29a g_signal_emit_valist (libgobject-2.0.so.0) #13 0x00007efc29b9f893 g_signal_emit (libgobject-2.0.so.0) #14 0x000056405807ab20 _set_state_full (NetworkManager) #15 0x000056405807d803 nm_device_unrealize (NetworkManager) #16 0x0000564057f6072c _platform_link_cb_idle (NetworkManager) #17 0x00007efc296a01db g_idle_dispatch (libglib-2.0.so.0) ...
2019-06-28 17:31:40 +02:00
if (ifindex_slave > 0) {
_LOGI (LOGD_BOND, "bond slave %s was released",
nm_device_get_ip_iface (slave));
}
bond: set slave IFF_UP after releasing it The kernel bonding code calls "closes" (which clears IFF_UP) the slave and restores some of its attributes (MAC address, MTU, etc), but doesn't bring it back up. This breaks things like carrier detection, and indeed when the device is closed, its carrier is also cleared, which leads us to think the device is not available for activation. To ensure that further events are noticed by the device, and that its carrier state is accurately represented, make sure the device is still IFF_UP after it has been released.
2013-01-21 14:33:40 -06:00
}
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
}
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
static gboolean
create_and_realize (NMDevice *device,
NMConnection *connection,
NMDevice *parent,
platform: return pointer to NMPlatformLink object for add functions Let the link-add functions return the internal pointer to the platform link object. Similar to link-get, which doesn't copy the link either. Also adjust the sole users of the add-functions (create-and-realize) to take the pointer. Eventually we still copy the returned data, because accessing platform can invalidate the returned pointer. Thus we don't actually safe any copying by this (at least every use of the function currently leads to the data being copied). Still change it, because I think the API of NMPlatform should look like that.
2015-12-09 15:13:57 +01:00
const NMPlatformLink **out_plink,
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
GError **error)
{
const char *iface = nm_device_get_iface (device);
platform: merge NMPlatformError with nm-error Platform had it's own scheme for reporting errors: NMPlatformError. Before, NMPlatformError indicated success via zero, negative integer values are numbers from <errno.h>, and positive integer values are platform specific codes. This changes now according to nm-error: success is still zero. Negative values indicate a failure, where the numeric value is either from <errno.h> or one of our error codes. The meaning of positive values depends on the functions. Most functions can only report an error reason (negative) and success (zero). For such functions, positive values should never be returned (but the caller should anticipate them). For some functions, positive values could mean additional information (but still success). That depends. This is also what systemd does, except that systemd only returns (negative) integers from <errno.h>, while we merge our own error codes into the range of <errno.h>. The advantage is to get rid of one way how to signal errors. The other advantage is, that these error codes are compatible with all other nm-errno values. For example, previously negative values indicated error codes from <errno.h>, but it did not entail error codes from netlink.
2018-12-22 14:13:05 +01:00
int r;
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
g_assert (iface);
platform: merge NMPlatformError with nm-error Platform had it's own scheme for reporting errors: NMPlatformError. Before, NMPlatformError indicated success via zero, negative integer values are numbers from <errno.h>, and positive integer values are platform specific codes. This changes now according to nm-error: success is still zero. Negative values indicate a failure, where the numeric value is either from <errno.h> or one of our error codes. The meaning of positive values depends on the functions. Most functions can only report an error reason (negative) and success (zero). For such functions, positive values should never be returned (but the caller should anticipate them). For some functions, positive values could mean additional information (but still success). That depends. This is also what systemd does, except that systemd only returns (negative) integers from <errno.h>, while we merge our own error codes into the range of <errno.h>. The advantage is to get rid of one way how to signal errors. The other advantage is, that these error codes are compatible with all other nm-errno values. For example, previously negative values indicated error codes from <errno.h>, but it did not entail error codes from netlink.
2018-12-22 14:13:05 +01:00
r = nm_platform_link_bond_add (nm_device_get_platform (device), iface, out_plink);
if (r < 0) {
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
g_set_error (error, NM_DEVICE_ERROR, NM_DEVICE_ERROR_CREATION_FAILED,
"Failed to create bond interface '%s' for '%s': %s",
iface,
nm_connection_get_id (connection),
platform: merge NMPlatformError with nm-error Platform had it's own scheme for reporting errors: NMPlatformError. Before, NMPlatformError indicated success via zero, negative integer values are numbers from <errno.h>, and positive integer values are platform specific codes. This changes now according to nm-error: success is still zero. Negative values indicate a failure, where the numeric value is either from <errno.h> or one of our error codes. The meaning of positive values depends on the functions. Most functions can only report an error reason (negative) and success (zero). For such functions, positive values should never be returned (but the caller should anticipate them). For some functions, positive values could mean additional information (but still success). That depends. This is also what systemd does, except that systemd only returns (negative) integers from <errno.h>, while we merge our own error codes into the range of <errno.h>. The advantage is to get rid of one way how to signal errors. The other advantage is, that these error codes are compatible with all other nm-errno values. For example, previously negative values indicated error codes from <errno.h>, but it did not entail error codes from netlink.
2018-12-22 14:13:05 +01:00
nm_strerror (r));
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
return FALSE;
}
return TRUE;
}
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
static gboolean
check_changed_options (NMSettingBond *s_a, NMSettingBond *s_b, GError **error)
{
bond: fix can_reapply_change() false positives can_reapply_change() would wrongly return true for unsupported reapply values because it used 'nm_setting_bond_get_option_default()' that is ill-named because it returns the overriden option other than its default value. https://bugzilla.redhat.com/show_bug.cgi?id=1847814 Fixes: 9bd07336ef16 ('bond: bond options logic rework')
2020-08-04 18:19:47 +02:00
const char **option_list;
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
bond: fix can_reapply_change() false positives can_reapply_change() would wrongly return true for unsupported reapply values because it used 'nm_setting_bond_get_option_default()' that is ill-named because it returns the overriden option other than its default value. https://bugzilla.redhat.com/show_bug.cgi?id=1847814 Fixes: 9bd07336ef16 ('bond: bond options logic rework')
2020-08-04 18:19:47 +02:00
option_list = nm_setting_bond_get_valid_options (NULL);
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
bond: fix can_reapply_change() false positives can_reapply_change() would wrongly return true for unsupported reapply values because it used 'nm_setting_bond_get_option_default()' that is ill-named because it returns the overriden option other than its default value. https://bugzilla.redhat.com/show_bug.cgi?id=1847814 Fixes: 9bd07336ef16 ('bond: bond options logic rework')
2020-08-04 18:19:47 +02:00
for (; *option_list; ++option_list) {
const char *name = *option_list;
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
/* We support changes to these */
if (NM_IN_STRSET (name,
NM_SETTING_BOND_OPTION_ACTIVE_SLAVE,
NM_SETTING_BOND_OPTION_PRIMARY)) {
continue;
}
/* Reject any other changes */
bond: fix can_reapply_change() false positives can_reapply_change() would wrongly return true for unsupported reapply values because it used 'nm_setting_bond_get_option_default()' that is ill-named because it returns the overriden option other than its default value. https://bugzilla.redhat.com/show_bug.cgi?id=1847814 Fixes: 9bd07336ef16 ('bond: bond options logic rework')
2020-08-04 18:19:47 +02:00
if (!nm_streq0 (nm_setting_bond_get_option_normalized (s_a, name),
nm_setting_bond_get_option_normalized (s_b, name))) {
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
g_set_error (error,
NM_DEVICE_ERROR,
NM_DEVICE_ERROR_INCOMPATIBLE_CONNECTION,
"Can't reapply '%s' bond option",
name);
return FALSE;
}
}
return TRUE;
}
static gboolean
can_reapply_change (NMDevice *device,
const char *setting_name,
NMSetting *s_old,
NMSetting *s_new,
GHashTable *diffs,
GError **error)
{
NMDeviceClass *device_class;
/* Only handle bond setting here, delegate other settings to parent class */
if (nm_streq (setting_name, NM_SETTING_BOND_SETTING_NAME)) {
if (!nm_device_hash_check_invalid_keys (diffs,
NM_SETTING_BOND_SETTING_NAME,
error,
NM_SETTING_BOND_OPTIONS))
return FALSE;
bond: fix can_reapply_change() false positives can_reapply_change() would wrongly return true for unsupported reapply values because it used 'nm_setting_bond_get_option_default()' that is ill-named because it returns the overriden option other than its default value. https://bugzilla.redhat.com/show_bug.cgi?id=1847814 Fixes: 9bd07336ef16 ('bond: bond options logic rework')
2020-08-04 18:19:47 +02:00
return check_changed_options (NM_SETTING_BOND (s_old), NM_SETTING_BOND (s_new), error);
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
}
device_class = NM_DEVICE_CLASS (nm_device_bond_parent_class);
return device_class->can_reapply_change (device,
setting_name,
s_old,
s_new,
diffs,
error);
}
static void
reapply_connection (NMDevice *device, NMConnection *con_old, NMConnection *con_new)
{
NMDeviceBond *self = NM_DEVICE_BOND (device);
const char *value;
NMSettingBond *s_bond;
NMBondMode mode;
NM_DEVICE_CLASS (nm_device_bond_parent_class)->reapply_connection (device,
con_old,
con_new);
_LOGD (LOGD_BOND, "reapplying bond settings");
s_bond = nm_connection_get_setting_bond (con_new);
g_return_if_fail (s_bond);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
value = nm_setting_bond_get_option_or_default (s_bond, NM_SETTING_BOND_OPTION_MODE);
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
mode = _nm_setting_bond_mode_from_string (value);
g_return_if_fail (mode != NM_BOND_MODE_UNKNOWN);
bond: bond options logic rework Add '_nm_setting_bond_get_option_or_default()' and move all the custom policies applied by NM for bond options in there. One such example of a custom policy is to set 'miimon' to 0 (instead of its default value of 100) if 'arp_interval' is explicitly enabled and 'miimon' is not. This means removing every piece of logic from nm_setting_bond_add_option() which used to clear out 'arp_interval' and 'arp_ip_target' if 'miimon' was set or clear out 'miimon' along with 'downdelay', 'updelay' and 'miimon' if 'arp_interval' was set. This behaviour is a bug since the kernel allow setting any combination of this options for bonds and NetworkManager should not limit the user to do so. Also use 'set_bond_attr_or_default()' instead of 'set_bond_attr()' as the former calls '_nm_setting_bond_get_option_or_default()' to implement the right logic to retrieve bond options according to current bond configuration.
2020-02-21 11:41:07 +01:00
set_bond_attr_or_default (device, s_bond, NM_SETTING_BOND_OPTION_PRIMARY);
bond: avoid setting "active_slave" option without interface enslaved Kernel will reject setting "active_slave", if the interface is not enslaved or not up. We already handle that by setting the option whenever we enslave an interface. However, we also must not set it initially, otherwise we get an ugly error log message: NetworkManager[939]: <debug> [1594709143.7459] platform-linux: sysctl: setting net:/sys/class/net/bond99/bonding/active_slave to eth1 (current value is ) NetworkManager[939]: <error> [1594709143.7459] platform-linux: sysctl: failed to set bonding/active_slave to eth1: (22) Invalid argument NetworkManager[939]: <warn> [1594709143.7460] device (bond99): failed to set bonding attribute active_slave to eth1 ... kernel: bond99: (slave eth1): Device is not bonding slave kernel: bond99: option active_slave: invalid value (eth1) See-also: https://bugzilla.redhat.com/show_bug.cgi?id=1856640 https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/577
2020-07-14 09:16:51 +02:00
set_bond_attr_active_slave (device, s_bond);
device: bond: implement connection reapply Only a couple of bond options (active_slave, primary) are supported at for now.
2017-03-03 12:08:39 +01:00
}
all: modify line separator comments to be 80 chars wide sed 's#^/\*\{5\}\*\+/$#/*****************************************************************************/#' $(git grep -l '\*\{5\}' | grep '\.[hc]$') -i
2016-10-02 18:22:50 +02:00
/*****************************************************************************/
core: move slave handling to device subclasses Let the master control how the slave gets enslaved and released since that's dependent on the virtual interface type of the master.
2012-02-28 18:35:30 -06:00
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
static void
nm_device_bond_init (NMDeviceBond * self)
{
device: transform NM_DEVICE_IS_MASTER gobject property to field in NMDeviceClass We don't need this flexibility of having a full fledged GObject property for is-master. The property value only depends on the device's class.
2017-06-01 23:08:30 +02:00
nm_assert (nm_device_is_master (NM_DEVICE (self)));
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
}
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API Previously, we used the generated GDBusInterfaceSkeleton types and glued them via the NMExportedObject base class to our NM types. We also used GDBusObjectManagerServer. Don't do that anymore. The resulting code was more complicated despite (or because?) using generated classes. It was hard to understand, complex, had ordering-issues, and had a runtime and memory overhead. This patch refactors this entirely and uses the lower layer API GDBusConnection directly. It replaces the generated code, GDBusInterfaceSkeleton, and GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager and static descriptor instances of type GDBusInterfaceInfo. This adds a net plus of more then 1300 lines of hand written code. I claim that this implementation is easier to understand. Note that previously we also required extensive and complex glue code to bind our objects to the generated skeleton objects. Instead, now glue our objects directly to GDBusConnection. The result is more immediate and gets rid of layers of code in between. Now that the D-Bus glue us more under our control, we can address issus and bottlenecks better, instead of adding code to bend the generated skeletons to our needs. Note that the current implementation now only supports one D-Bus connection. That was effectively the case already, although there were places (and still are) where the code pretends it could also support connections from a private socket. We dropped private socket support mainly because it was unused, untested and buggy, but also because GDBusObjectManagerServer could not export the same objects on multiple connections. Now, it would be rather straight forward to fix that and re-introduce ObjectManager on each private connection. But this commit doesn't do that yet, and the new code intentionally supports only one D-Bus connection. Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start() succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough for the moment. It could be easily extended later, for example with polling whether the system bus appears (like was done previously). Also, restart of D-Bus daemon isn't supported either -- just like before. Note how NMDBusManager now implements the ObjectManager D-Bus interface directly. Also, this fixes race issues in the server, by no longer delaying PropertiesChanged signals. NMExportedObject would collect changed properties and send the signal out in idle_emit_properties_changed() on idle. This messes up the ordering of change events w.r.t. other signals and events on the bus. Note that not only NMExportedObject messed up the ordering. Also the generated code would hook into notify() and process change events in and idle handle, exhibiting the same ordering issue too. No longer do that. PropertiesChanged signals will be sent right away by hooking into dispatch_properties_changed(). This means, changing a property in quick succession will no longer be combined and is guaranteed to emit signals for each individual state. Quite possibly we emit now more PropertiesChanged signals then before. However, we are now able to group a set of changes by using standard g_object_freeze_notify()/g_object_thaw_notify(). We probably should make more use of that. Also, now that our signals are all handled in the right order, we might find places where we still emit them in the wrong order. But that is then due to the order in which our GObjects emit signals, not due to an ill behavior of the D-Bus glue. Possibly we need to identify such ordering issues and fix them. Numbers (for contrib/rpm --without debug on x86_64): - the patch changes the code size of NetworkManager by - 2809360 bytes + 2537528 bytes (-9.7%) - Runtime measurements are harder because there is a large variance during testing. In other words, the numbers are not reproducible. Currently, the implementation performs no caching of GVariants at all, but it would be rather simple to add it, if that turns out to be useful. Anyway, without strong claim, it seems that the new form tends to perform slightly better. That would be no surprise. $ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done) - real 1m39.355s + real 1m37.432s $ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done) - real 0m26.843s + real 0m25.281s - Regarding RSS size, just looking at the processes in similar conditions, doesn't give a large difference. On my system they consume about 19MB RSS. It seems that the new version has a slightly smaller RSS size. - 19356 RSS + 18660 RSS
2018-02-26 13:51:52 +01:00
static const NMDBusInterfaceInfoExtended interface_info_device_bond = {
.parent = NM_DEFINE_GDBUS_INTERFACE_INFO_INIT (
NM_DBUS_INTERFACE_DEVICE_BOND,
.signals = NM_DEFINE_GDBUS_SIGNAL_INFOS (
&nm_signal_info_property_changed_legacy,
),
.properties = NM_DEFINE_GDBUS_PROPERTY_INFOS (
NM_DEFINE_DBUS_PROPERTY_INFO_EXTENDED_READABLE_L ("HwAddress", "s", NM_DEVICE_HW_ADDRESS),
NM_DEFINE_DBUS_PROPERTY_INFO_EXTENDED_READABLE_L ("Carrier", "b", NM_DEVICE_CARRIER),
NM_DEFINE_DBUS_PROPERTY_INFO_EXTENDED_READABLE_L ("Slaves", "ao", NM_DEVICE_SLAVES),
),
),
.legacy_property_changed = TRUE,
};
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
static void
nm_device_bond_class_init (NMDeviceBondClass *klass)
{
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API Previously, we used the generated GDBusInterfaceSkeleton types and glued them via the NMExportedObject base class to our NM types. We also used GDBusObjectManagerServer. Don't do that anymore. The resulting code was more complicated despite (or because?) using generated classes. It was hard to understand, complex, had ordering-issues, and had a runtime and memory overhead. This patch refactors this entirely and uses the lower layer API GDBusConnection directly. It replaces the generated code, GDBusInterfaceSkeleton, and GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager and static descriptor instances of type GDBusInterfaceInfo. This adds a net plus of more then 1300 lines of hand written code. I claim that this implementation is easier to understand. Note that previously we also required extensive and complex glue code to bind our objects to the generated skeleton objects. Instead, now glue our objects directly to GDBusConnection. The result is more immediate and gets rid of layers of code in between. Now that the D-Bus glue us more under our control, we can address issus and bottlenecks better, instead of adding code to bend the generated skeletons to our needs. Note that the current implementation now only supports one D-Bus connection. That was effectively the case already, although there were places (and still are) where the code pretends it could also support connections from a private socket. We dropped private socket support mainly because it was unused, untested and buggy, but also because GDBusObjectManagerServer could not export the same objects on multiple connections. Now, it would be rather straight forward to fix that and re-introduce ObjectManager on each private connection. But this commit doesn't do that yet, and the new code intentionally supports only one D-Bus connection. Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start() succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough for the moment. It could be easily extended later, for example with polling whether the system bus appears (like was done previously). Also, restart of D-Bus daemon isn't supported either -- just like before. Note how NMDBusManager now implements the ObjectManager D-Bus interface directly. Also, this fixes race issues in the server, by no longer delaying PropertiesChanged signals. NMExportedObject would collect changed properties and send the signal out in idle_emit_properties_changed() on idle. This messes up the ordering of change events w.r.t. other signals and events on the bus. Note that not only NMExportedObject messed up the ordering. Also the generated code would hook into notify() and process change events in and idle handle, exhibiting the same ordering issue too. No longer do that. PropertiesChanged signals will be sent right away by hooking into dispatch_properties_changed(). This means, changing a property in quick succession will no longer be combined and is guaranteed to emit signals for each individual state. Quite possibly we emit now more PropertiesChanged signals then before. However, we are now able to group a set of changes by using standard g_object_freeze_notify()/g_object_thaw_notify(). We probably should make more use of that. Also, now that our signals are all handled in the right order, we might find places where we still emit them in the wrong order. But that is then due to the order in which our GObjects emit signals, not due to an ill behavior of the D-Bus glue. Possibly we need to identify such ordering issues and fix them. Numbers (for contrib/rpm --without debug on x86_64): - the patch changes the code size of NetworkManager by - 2809360 bytes + 2537528 bytes (-9.7%) - Runtime measurements are harder because there is a large variance during testing. In other words, the numbers are not reproducible. Currently, the implementation performs no caching of GVariants at all, but it would be rather simple to add it, if that turns out to be useful. Anyway, without strong claim, it seems that the new form tends to perform slightly better. That would be no surprise. $ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done) - real 1m39.355s + real 1m37.432s $ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done) - real 0m26.843s + real 0m25.281s - Regarding RSS size, just looking at the processes in similar conditions, doesn't give a large difference. On my system they consume about 19MB RSS. It seems that the new version has a slightly smaller RSS size. - 19356 RSS + 18660 RSS
2018-02-26 13:51:52 +01:00
NMDBusObjectClass *dbus_object_class = NM_DBUS_OBJECT_CLASS (klass);
device/trivial: rename parent-class variable in device class constructor The majority of device implementations name their parent-class variable "device_class". That also makes more sense as it is more consistant. E.g. "parent" sounds like it's the direct parent, but that is not the crucial point here. The crucial point at this place, is that we access the NMDeviceClass typed pointer. Rename.
2018-07-10 07:45:35 +02:00
NMDeviceClass *device_class = NM_DEVICE_CLASS (klass);
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
core/dbus: rework D-Bus implementation to use lower layer GDBusConnection API Previously, we used the generated GDBusInterfaceSkeleton types and glued them via the NMExportedObject base class to our NM types. We also used GDBusObjectManagerServer. Don't do that anymore. The resulting code was more complicated despite (or because?) using generated classes. It was hard to understand, complex, had ordering-issues, and had a runtime and memory overhead. This patch refactors this entirely and uses the lower layer API GDBusConnection directly. It replaces the generated code, GDBusInterfaceSkeleton, and GDBusObjectManagerServer. All this is now done by NMDbusObject and NMDBusManager and static descriptor instances of type GDBusInterfaceInfo. This adds a net plus of more then 1300 lines of hand written code. I claim that this implementation is easier to understand. Note that previously we also required extensive and complex glue code to bind our objects to the generated skeleton objects. Instead, now glue our objects directly to GDBusConnection. The result is more immediate and gets rid of layers of code in between. Now that the D-Bus glue us more under our control, we can address issus and bottlenecks better, instead of adding code to bend the generated skeletons to our needs. Note that the current implementation now only supports one D-Bus connection. That was effectively the case already, although there were places (and still are) where the code pretends it could also support connections from a private socket. We dropped private socket support mainly because it was unused, untested and buggy, but also because GDBusObjectManagerServer could not export the same objects on multiple connections. Now, it would be rather straight forward to fix that and re-introduce ObjectManager on each private connection. But this commit doesn't do that yet, and the new code intentionally supports only one D-Bus connection. Also, the D-Bus startup was simplified. There is no retry, either nm_dbus_manager_start() succeeds, or it detects the initrd case. In the initrd case, bus manager never tries to connect to D-Bus. Since the initrd scenario is not yet used/tested, this is good enough for the moment. It could be easily extended later, for example with polling whether the system bus appears (like was done previously). Also, restart of D-Bus daemon isn't supported either -- just like before. Note how NMDBusManager now implements the ObjectManager D-Bus interface directly. Also, this fixes race issues in the server, by no longer delaying PropertiesChanged signals. NMExportedObject would collect changed properties and send the signal out in idle_emit_properties_changed() on idle. This messes up the ordering of change events w.r.t. other signals and events on the bus. Note that not only NMExportedObject messed up the ordering. Also the generated code would hook into notify() and process change events in and idle handle, exhibiting the same ordering issue too. No longer do that. PropertiesChanged signals will be sent right away by hooking into dispatch_properties_changed(). This means, changing a property in quick succession will no longer be combined and is guaranteed to emit signals for each individual state. Quite possibly we emit now more PropertiesChanged signals then before. However, we are now able to group a set of changes by using standard g_object_freeze_notify()/g_object_thaw_notify(). We probably should make more use of that. Also, now that our signals are all handled in the right order, we might find places where we still emit them in the wrong order. But that is then due to the order in which our GObjects emit signals, not due to an ill behavior of the D-Bus glue. Possibly we need to identify such ordering issues and fix them. Numbers (for contrib/rpm --without debug on x86_64): - the patch changes the code size of NetworkManager by - 2809360 bytes + 2537528 bytes (-9.7%) - Runtime measurements are harder because there is a large variance during testing. In other words, the numbers are not reproducible. Currently, the implementation performs no caching of GVariants at all, but it would be rather simple to add it, if that turns out to be useful. Anyway, without strong claim, it seems that the new form tends to perform slightly better. That would be no surprise. $ time (for i in {1..1000}; do nmcli >/dev/null || break; echo -n .; done) - real 1m39.355s + real 1m37.432s $ time (for i in {1..2000}; do busctl call org.freedesktop.NetworkManager /org/freedesktop org.freedesktop.DBus.ObjectManager GetManagedObjects > /dev/null || break; echo -n .; done) - real 0m26.843s + real 0m25.281s - Regarding RSS size, just looking at the processes in similar conditions, doesn't give a large difference. On my system they consume about 19MB RSS. It seems that the new version has a slightly smaller RSS size. - 19356 RSS + 18660 RSS
2018-02-26 13:51:52 +01:00
dbus_object_class->interface_infos = NM_DBUS_INTERFACE_INFOS (&interface_info_device_bond);
device: replace NM_DEVICE_CLASS_DECLARE_TYPES() macro by explicit initialization It seems to me the NM_DEVICE_CLASS_DECLARE_TYPES() macro confuses more than helping. Let's explicitly initialize the two fields, albeit with another helper macro NM_DEVICE_DEFINE_LINK_TYPES() to get the list of link-types right. For consistency, also leave nop-lines like device_class->connection_type_supported = NULL; device_class->link_types = NM_DEVICE_DEFINE_LINK_TYPES (); because all NMDevice class init methods should have this same boiler plate code and to make it explicit that this is intended. And there are only 3 occurences where this actually comes into play.
2018-07-10 09:26:42 +02:00
device_class->connection_type_supported = NM_SETTING_BOND_SETTING_NAME;
core: give better error reason why device is incompatible with profile Note the special error codes NM_UTILS_ERROR_CONNECTION_AVAILABLE_*. This will be used to determine, whether the profile is fundamentally incompatible with the device, or whether just some other properties mismatch. That information will be importand during a plain `nmcli connection up`, where NetworkManager searches all devices for a device to activate. If no device is found (and multiple errors happened), we want to show the error that is most likely relevant for the user. Also note, how NMDevice's check_connection_compatible() uses the new class field "device_class->connection_type_check_compatible" to simplify checks for compatible profiles. The error reason is still unused.
2018-06-27 17:00:55 +02:00
device_class->connection_type_check_compatible = NM_SETTING_BOND_SETTING_NAME;
device: replace NM_DEVICE_CLASS_DECLARE_TYPES() macro by explicit initialization It seems to me the NM_DEVICE_CLASS_DECLARE_TYPES() macro confuses more than helping. Let's explicitly initialize the two fields, albeit with another helper macro NM_DEVICE_DEFINE_LINK_TYPES() to get the list of link-types right. For consistency, also leave nop-lines like device_class->connection_type_supported = NULL; device_class->link_types = NM_DEVICE_DEFINE_LINK_TYPES (); because all NMDevice class init methods should have this same boiler plate code and to make it explicit that this is intended. And there are only 3 occurences where this actually comes into play.
2018-07-10 09:26:42 +02:00
device_class->link_types = NM_DEVICE_DEFINE_LINK_TYPES (NM_LINK_TYPE_BOND);
device/trivial: rename parent-class variable in device class constructor The majority of device implementations name their parent-class variable "device_class". That also makes more sense as it is more consistant. E.g. "parent" sounds like it's the direct parent, but that is not the crucial point here. The crucial point at this place, is that we access the NMDeviceClass typed pointer. Rename.
2018-07-10 07:45:35 +02:00
device_class->is_master = TRUE;
device_class->get_generic_capabilities = get_generic_capabilities;
device_class->complete_connection = complete_connection;
device_class->update_connection = update_connection;
device_class->master_update_slave_connection = master_update_slave_connection;
device_class->create_and_realize = create_and_realize;
device_class->act_stage1_prepare = act_stage1_prepare;
device_class->get_configured_mtu = nm_device_get_configured_mtu_for_wired;
device_class->enslave_slave = enslave_slave;
device_class->release_slave = release_slave;
device_class->can_reapply_change = can_reapply_change;
device_class->reapply_connection = reapply_connection;
bond: move bond master interfaces to NMDeviceBond This way clients know it's a bond master, and it allows us to more cleanly separate the code for bonding and non-bonding cases.
2012-02-15 17:31:37 -06:00
}
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
all: modify line separator comments to be 80 chars wide sed 's#^/\*\{5\}\*\+/$#/*****************************************************************************/#' $(git grep -l '\*\{5\}' | grep '\.[hc]$') -i
2016-10-02 18:22:50 +02:00
/*****************************************************************************/
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
device: rename internal device factories Instead of NMBondFactory, call it NMBondDeviceFactory.
2016-10-07 17:00:59 +02:00
#define NM_TYPE_BOND_DEVICE_FACTORY (nm_bond_device_factory_get_type ())
#define NM_BOND_DEVICE_FACTORY(obj) (G_TYPE_CHECK_INSTANCE_CAST ((obj), NM_TYPE_BOND_DEVICE_FACTORY, NMBondDeviceFactory))
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
static NMDevice *
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
create_device (NMDeviceFactory *factory,
const char *iface,
device: pass NMPlatformLink instance as const pointer
2016-01-10 15:13:20 +01:00
const NMPlatformLink *plink,
core: split device creation and device setup (bgo #737458) Future patches will create devices long before they are backed by kernel resources, so we need to split NMDevice object creation from actual setup based on the backing resources. This patch combines the NMDeviceFactory's new_link() and create_virtual_device_for_connection() class methods into a single create_device() method that simply creates an unrealized NMDevice object; this method is not expected to fail unless the device is supposed to be ignored. This also means that the NMDevice 'platform-device' property is removed, because a platform link object may not be available at NMDevice object creation time. After the device is created, it is then "realized" at some later time from a platform link (for existing/hardware devices via the realize() method) or from an NMConnection (for newly created software devices via the create_and_realize() NMDeviceClass methods). https://bugzilla.gnome.org/show_bug.cgi?id=737458
2014-09-05 08:50:02 -05:00
NMConnection *connection,
gboolean *out_ignore)
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
{
return (NMDevice *) g_object_new (NM_TYPE_DEVICE_BOND,
trivial: fix whitespace
2016-01-08 16:11:41 +01:00
NM_DEVICE_IFACE, iface,
NM_DEVICE_DRIVER, "bonding",
NM_DEVICE_TYPE_DESC, "Bond",
NM_DEVICE_DEVICE_TYPE, NM_DEVICE_TYPE_BOND,
NM_DEVICE_LINK_TYPE, NM_LINK_TYPE_BOND,
NULL);
bond: port to internal device factory
2014-09-08 10:15:02 -05:00
}
core: let device plugins advertise supported link and setting types Instead of looping over all plugins and asking each plugin whether it can handle a link or a connection, have them advertise the link and connection types they support, and use that when creating new devices.
2014-09-17 14:17:30 -05:00
NM_DEVICE_FACTORY_DEFINE_INTERNAL (BOND, Bond, bond,
NM_DEVICE_FACTORY_DECLARE_LINK_TYPES (NM_LINK_TYPE_BOND)
NM_DEVICE_FACTORY_DECLARE_SETTING_TYPES (NM_SETTING_BOND_SETTING_NAME),
device: make NMDeviceFactory a class instead of an interface An interface would make sense to allow the actual device-factory to inherit from another type. However, glib interfaces make code much harder to follow and less efficient. The device factory shall be a very simple type with meta data about supported device types and the ability to create device instances. There is no need to make this an interface implementation, instead just let the factories inherit from NM_TYPE_DEVICE_FACTORY directly.
2016-10-07 16:05:43 +02:00
factory_class->create_device = create_device;
core: refactor private data in "src" - use _NM_GET_PRIVATE() and _NM_GET_PRIVATE_PTR() everywhere. - reorder statements, to have GObject related functions (init, dispose, constructed) at the bottom of each file and in a consistent order w.r.t. each other. - unify whitespaces in signal and properties declarations. - use NM_GOBJECT_PROPERTIES_DEFINE() and _notify() - drop unused signal slots in class structures - drop unused header files for device factories
2016-09-29 13:49:01 +02:00
);
Reference in a new issue Copy permalink