The modes VALUE_STRSPLIT_MODE_OBJLIST* and VALUE_STRSPLIT_MODE_MULTILIST* are
different. We must use the right mode.
For example, _get_fcn_match_interface_name() concatenates the interface-names
with space. So, the tokenizer of the setter must also use space as delimiter.
VALUE_STRSPLIT_MODE_MULTILIST_WITH_ESCAPE does that correctly,
VALUE_STRSPLIT_MODE_OBJLIST_WITH_ESCAPE does not.
(cherry picked from commit 758bf32640)
There should be little difference here, because the priority list is
(and was) never serialized with special characters like backslashes or
delimiters that require escaping.
Likewise, no working code actually tried to set such characters.
Still, drop the plain VALUE_STRSPLIT_MODE_STRIPPED and use
VALUE_STRSPLIT_MODE_ESCAPED_TOKENS_WITH_SPACES instead. We should have
a small set of modes that we use for splitting strings.
- merge the pointless helper function vlan_priorities_to_string()
into the only caller _get_fcn_vlan_xgress_priority_map().
- minor cleanups, like setting out-is-default if num==0, not
based on whether we have a non-empty string. There is not difference
in practice, because nm_setting_vlan_get_priority() never fails.
Hence they are identical. But nm_setting_vlan_get_priority() has
an API that allows it to fail, so we should declare the default
depending on the number of vlan priorities.
- don't allocate the temporary GString instance if we won't need it.
- only append the delimiter if needed, and not truncate it afterwards.
It might have even worse performance this way, but it feels more
correct to me.
- also cache the result of nm_setting_vlan_get_num_priorities().
NMSettingVlan's implementation is horrible and uses a GSList to
track the list of priorities. This makes it relatively expensive
to call get-num-priorities repeatedly (and pointless).
the only change in behaviour is for VALUE_STRSPLIT_MODE_MULTILIST.
Previously, we would split at " \t,", now we will also split at
the white space characters "\n\r\f".
For one, use NM_ASCII_SPACES as delimiter when reading
"MATCH_INTERFACE_NAME". Previously, it was only " \t".
I think there is no change in behavior otherwise.
We had %VALUE_STRSPLIT_MODE_MULTILIST_WITH_ESCAPE, which was used
by "match.interface-names". This uses nm_utils_strsplit_set_full()
with %NM_UTILS_STRSPLIT_SET_FLAGS_ALLOW_ESCAPING and
_nm_utils_unescape_plain().
We want eventually to use nm_utils_escaped_tokens_split() everywhere.
We already have %VALUE_STRSPLIT_MODE_ESCAPED_TOKENS, which splits the
list at ',' (and strips whitespaces at the around the delimiter). That
differs from what %VALUE_STRSPLIT_MODE_MULTILIST_WITH_ESCAPE did, which
also considered whitespace a delimiter.
So, we need a new mode %VALUE_STRSPLIT_MODE_ESCAPED_TOKENS_WITH_SPACES
which replaces the previous mode.
Note that the previous implementation did almost the same thing. In
fact, I cannot imagine examples where they behaved differently, but
my feeling is that there might be some edge cases where this changes
behavior.
The reminder of the function only does (something akin to) g_strstrip().
As we split the strings are spaces to begin with, there is nothing to
strip and we can return right away.
When splitting (and concatenating) list-typed properties,
we really should use nm_utils_escaped_tokens_split()
and nm_utils_escaped_tokens_escape*().
Make that the default, and mark all properties to opt-in to the
legacy behavior.
The modes VALUE_STRSPLIT_MODE_OBJLIST* and VALUE_STRSPLIT_MODE_MULTILIST* are
different. We must use the right mode.
For example, _get_fcn_match_interface_name() concatenates the interface-names
with space. So, the tokenizer of the setter must also use space as delimiter.
VALUE_STRSPLIT_MODE_MULTILIST_WITH_ESCAPE does that correctly,
VALUE_STRSPLIT_MODE_OBJLIST_WITH_ESCAPE does not.
Also, in nm_platform_routing_rule_cmp() always compare the routing
table field, also if l3mdev is set. For kernel, we cannot set table and
l3mdev together, hence such rules don't really exist (or if we try to
configure it, it will be rejected by kernel). But as far as
nm_platform_routing_rule_cmp() is concerned, if the table is set,
always compare it.
(cherry picked from commit b6ff02e76f)
For routes and routing rules, kernel uses a certain (not stictly defined) set
of attributes to decide whether to routes/rules are identical.
That is a problem, as different kernel versions disagree on whether
two routes/rules are the same (EEXIST) or not.
Note that when NetworkManager tries to add a rule with protocol set to
anything but RTPROT_UNSPEC, then kernel will ignore the attribute if it
doesn't have support for it. Meaning: the added rule will have a
different protocol setting then intended.
Note that NMPRulesManager will add a rule if it doesn't find it in the
platform cache so far. That means, when looking into the platform cache
we must ignore or honor the protocol like kernel does.
This does not only affect FRA_PROTOCOL, but all attributes where kernel
and NetworkManager disagrees. But the protocol is the most prominent
one, because the rules tracked by nmp_rules_manager_track_default()
specify the protocol.
(cherry picked from commit ef4f8ccf6d)
Next we will need to detect more kernel features. First refactor the
handling of these to require less code changes and be more efficient.
A plain nm_platform_kernel_support_get() only reqiures to access an
array in the common case.
The other important change is that the function no longer requires a
NMPlatform instance. This allows us to check kernel support from
anywhere. The only thing is that we require kernel support to be
initialized before calling this function. That means, an NMPlatform
instance must have detected support before.
(cherry picked from commit ee269b318e)
Also, in nm_platform_routing_rule_cmp() always compare the routing
table field, also if l3mdev is set. For kernel, we cannot set table and
l3mdev together, hence such rules don't really exist (or if we try to
configure it, it will be rejected by kernel). But as far as
nm_platform_routing_rule_cmp() is concerned, if the table is set,
always compare it.
For routes and routing rules, kernel uses a certain (not stictly defined) set
of attributes to decide whether to routes/rules are identical.
That is a problem, as different kernel versions disagree on whether
two routes/rules are the same (EEXIST) or not.
Note that when NetworkManager tries to add a rule with protocol set to
anything but RTPROT_UNSPEC, then kernel will ignore the attribute if it
doesn't have support for it. Meaning: the added rule will have a
different protocol setting then intended.
Note that NMPRulesManager will add a rule if it doesn't find it in the
platform cache so far. That means, when looking into the platform cache
we must ignore or honor the protocol like kernel does.
This does not only affect FRA_PROTOCOL, but all attributes where kernel
and NetworkManager disagrees. But the protocol is the most prominent
one, because the rules tracked by nmp_rules_manager_track_default()
specify the protocol.
Next we will need to detect more kernel features. First refactor the
handling of these to require less code changes and be more efficient.
A plain nm_platform_kernel_support_get() only reqiures to access an
array in the common case.
The other important change is that the function no longer requires a
NMPlatform instance. This allows us to check kernel support from
anywhere. The only thing is that we require kernel support to be
initialized before calling this function. That means, an NMPlatform
instance must have detected support before.
- if there is only one vlan in the list, then we can return success
early. That is, because one NMBridgeVlan instance is always valid
due to the way how users must use the API to construct the element.
- the implementation for check_normalizable is only correct, if there
are no duplicate or overlapping ranges. Assert for that. In fact,
all callers first check for errors and then for normalizable errors.
- avoid duplicate calls to nm_bridge_vlan_get_vid_range(). There are
duplicate assertions that we don't need.
- only check for pvid once per range.
- combine calls to g_hash_table_contains() and g_hash_table_add().
(cherry picked from commit a358da096f)
We don't need GPtrArray to construct an array of fixed side.
Actually, we also don't need to malloc each NMPlatformBridgeVlan
element individually. Just allocate one buffer and append them
to the end.
(cherry picked from commit 6bc8ee87af)
In some cases it is convenient to specify ranges of bridge vlans, as
already supported by iproute2 and natively by kernel. With this commit
it becomes possible to add a range in this way:
nmcli connection modify eth0-slave +bridge-port.vlans "100-200 untagged"
vlan ranges can't be PVIDs because only one PVID vlan can exist.
https://bugzilla.redhat.com/show_bug.cgi?id=1652910
(cherry picked from commit 7093515777)
CC libnm-core/tests/libnm_core_tests_test_general-test-general.o
In file included from ../shared/nm-default.h:280:0,
from ../libnm-core/tests/test-general.c:24:
../libnm-core/tests/test-general.c: In function _sock_addr_endpoint:
../libnm-core/tests/test-general.c:5911:18: error: logical not is only applied to the left hand side of comparison [-Werror=logical-not-parentheses]
g_assert (!host == (port == -1));
^
../shared/nm-utils/nm-macros-internal.h:1793:7: note: in definition of macro __NM_G_BOOLEAN_EXPR_IMPL
if (expr) \
^
/usr/include/glib-2.0/glib/gmacros.h:376:43: note: in expansion of macro _G_BOOLEAN_EXPR
#define G_LIKELY(expr) (__builtin_expect (_G_BOOLEAN_EXPR((expr)), 1))
^
/usr/include/glib-2.0/glib/gtestutils.h:116:49: note: in expansion of macro G_LIKELY
if G_LIKELY (expr) ; else \
^
../libnm-core/tests/test-general.c:5911:2: note: in expansion of macro g_assert
g_assert (!host == (port == -1));
^
Fixes: 713e879d76 ('libnm: add NMSockAddrEndpoint API')
(cherry picked from commit 1e8c08730f)
CC libnm-core/tests/libnm_core_tests_test_general-test-general.o
In file included from ../shared/nm-default.h:280:0,
from ../libnm-core/tests/test-general.c:24:
../libnm-core/tests/test-general.c: In function _sock_addr_endpoint:
../libnm-core/tests/test-general.c:5911:18: error: logical not is only applied to the left hand side of comparison [-Werror=logical-not-parentheses]
g_assert (!host == (port == -1));
^
../shared/nm-utils/nm-macros-internal.h:1793:7: note: in definition of macro __NM_G_BOOLEAN_EXPR_IMPL
if (expr) \
^
/usr/include/glib-2.0/glib/gmacros.h:376:43: note: in expansion of macro _G_BOOLEAN_EXPR
#define G_LIKELY(expr) (__builtin_expect (_G_BOOLEAN_EXPR((expr)), 1))
^
/usr/include/glib-2.0/glib/gtestutils.h:116:49: note: in expansion of macro G_LIKELY
if G_LIKELY (expr) ; else \
^
../libnm-core/tests/test-general.c:5911:2: note: in expansion of macro g_assert
g_assert (!host == (port == -1));
^
Fixes: 713e879d76 ('libnm: add NMSockAddrEndpoint API')
- if there is only one vlan in the list, then we can return success
early. That is, because one NMBridgeVlan instance is always valid
due to the way how users must use the API to construct the element.
- the implementation for check_normalizable is only correct, if there
are no duplicate or overlapping ranges. Assert for that. In fact,
all callers first check for errors and then for normalizable errors.
- avoid duplicate calls to nm_bridge_vlan_get_vid_range(). There are
duplicate assertions that we don't need.
- only check for pvid once per range.
- combine calls to g_hash_table_contains() and g_hash_table_add().
We don't need GPtrArray to construct an array of fixed side.
Actually, we also don't need to malloc each NMPlatformBridgeVlan
element individually. Just allocate one buffer and append them
to the end.
In some cases it is convenient to specify ranges of bridge vlans, as
already supported by iproute2 and natively by kernel. With this commit
it becomes possible to add a range in this way:
nmcli connection modify eth0-slave +bridge-port.vlans "100-200 untagged"
vlan ranges can't be PVIDs because only one PVID vlan can exist.
https://bugzilla.redhat.com/show_bug.cgi?id=1652910
CC src/settings/plugins/ifcfg-rh/src_settings_plugins_ifcfg_rh_libnms_ifcfg_rh_core_la-nms-ifcfg-rh-reader.lo
In file included from ../shared/nm-default.h:280:0,
from ../src/settings/plugins/ifcfg-rh/nms-ifcfg-rh-reader.c:21:
../src/settings/plugins/ifcfg-rh/nms-ifcfg-rh-reader.c: In function read_routing_rules_parse:
../src/settings/plugins/ifcfg-rh/nms-ifcfg-rh-reader.c:4309:27: error: logical not is only applied to the left hand side of comparison [-Werror=logical-not-parentheses]
nm_assert (!key_is_ipv4 == NM_STR_HAS_PREFIX (key, "ROUTING_RULE6_"));
^
../shared/nm-utils/nm-macros-internal.h:1793:7: note: in definition of macro __NM_G_BOOLEAN_EXPR_IMPL
if (expr) \
^
/usr/include/glib-2.0/glib/gmacros.h:376:43: note: in expansion of macro _G_BOOLEAN_EXPR
#define G_LIKELY(expr) (__builtin_expect (_G_BOOLEAN_EXPR((expr)), 1))
^
/usr/include/glib-2.0/glib/gtestutils.h:116:49: note: in expansion of macro G_LIKELY
if G_LIKELY (expr) ; else \
^
../shared/nm-utils/nm-macros-internal.h:973:40: note: in expansion of macro g_assert
#define nm_assert(cond) G_STMT_START { g_assert (cond); } G_STMT_END
^
../src/settings/plugins/ifcfg-rh/nms-ifcfg-rh-reader.c:4309:3: note: in expansion of macro nm_assert
nm_assert (!key_is_ipv4 == NM_STR_HAS_PREFIX (key, "ROUTING_RULE6_"));
^
Fixes: 4d46804437
(cherry picked from commit c6e6dcae70)
