This fixes some recent security problems in hostapd.
See here for details: https://w1.fi/security/2022-1
* CVE-2022-23303
* CVE-2022-23304
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
If authentication fails repeatedly e.g. because of a weak signal, the link
can end up in blocked state. If one of the nodes tries to establish a link
again before it is unblocked on the other side, it will block the link to
that other side. The same happens on the other side when it unblocks the
link. In that scenario, the link never recovers on its own.
To fix this, allow restarting authentication even if the link is in blocked
state, but don't initiate the attempt until the blocked period is over.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit f84053af5c)
When a ubus event handler denies a association with a non-zero return
value, the code jumps to preceeding code, creating an endless loop until
the event handler accepts the assc request.
Move the ubus handler further up the code to avoid creating such a loop.
Signed-off-by: David Bauer <mail@david-bauer.net>
They can be added as hex digit strings via the 'vendor_elements' option
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit 1818b038d7)
This makes it possible to avoid using a RADIUS server for WPA enterprise authentication
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from 98621c9782)
This allows WPA enterprise roaming in the same mobility domain without any
manual key configuration (aside from radius credentials)
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit 704ab6a002)
Using a pointer one lifter after it freed is not the best idea.
Let's not do that.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry-picked from commit 63c01ad025)
This implements the mapping recommendations from RFC8325, with an
update from RFC8622. This ensures that DSCP marked packets are properly
sorted into WMM classes.
The map can be disabled by setting iw_qos_map_set to something invalid
like 'none'
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit a5e3def182)
This feature is useful on its own even without full interworking support
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit efff3520f4)
Both hostapd and netifd attempt to add a VLAN device to a bridge.
Depending on which one wins the race, bridge vlan settings might be incomplete,
or hostapd might run into an error and refuse to service the client.
Fix this by preventing hostapd from adding interfaces to the bridge and
instead rely entirely on netifd handling this properly
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit da4be02fcd)
(cherry-picked from commit 63c01ad025)
When using htmode 'HE20' with a radio mode that uses wpa-supplicant
(like mesh or sta), it will default to 40 MHz bw if disable_ht40 is not
set. This commit fixes this behaviour.
Signed-off-by: Jesus Fernandez Manzano <jesus.manzano@galgus.net>
(cherry-picked from commit af83e3ce0f)
Use the VLAN interface instead of the bridge, to ensure that hostapd receives
untagged DHCP packets
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit 7b46377a0c)
(cherry-picked from commit f1b98fa4fa)
In setups using VLAN bridge filtering, hostapd may need to communicate using
a VLAN interface on top of the bridge, instead of using the bridge directly
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit 305c1b8d74)
- add functionality to configure RADIUS NAS-Id and Operator-Name
- add functionality to configure RADIUS accounting interval
- enable RADIUS "Chargeable User Identity"
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit 3bd6c8c728)
Airtime policy configuration is extremely useful in multiple BSS scenarios.
Since nowadays most people configure both private and guest networks (at
least), it makes sense to enable it by default, except for the most limited
of the variants.
Size of the hostapd-basic-openssl binary (mipsel 24Kc -O2):
543944 bytes (airtime policy disabled)
548040 bytes (airtime policy enabled)
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Acked-by: Daniel Golle <daniel@makrotopia.org>
(cherry-picked from commit d38f456582)
netifd is responsible for handling that, except if the vlan bridge
was provided by the config
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry-picked from commit cf45caeff1)
Make the organization (O=) of the cert configurable via uci. If not
configured, use a combination of "OpenWrt" and an unique id like it was
done before.
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
(cherry picked from commit 2c6c1501af)
In hostapd_ubus_add_bss(), ubus objects are not registered for mesh
interfaces. This provokes a segfault when accessing the ubus object in
mesh deinit.
This commit adds the same condition to hostapd_ubus_free_bss() for
discarding those mesh interfaces.
Signed-off-by: Jesus Fernandez Manzano <jesus.manzano@galgus.net>
(cherry picked from commit 5269c47e8d)
EXTRA_MOUNT variable should be reset in dnsmasq_start() rather than
just once at the beginning of the script.
Fixes: ac4e8aa2f8 ("dnsmasq: fix more dnsmasq jail issues")
Reported-by: Hartmut Birr <e9hack@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit ddc8d085f3)
* remove superflus mounts of /dev/null and /dev/urandom
* reset EXTRA_MOUNTS at the beginning of the script
* add mount according to ignore_hosts_dir
* don't add mount for file which is inside a directory already in the
EXTRA_MOUNTS list
Fixes: 59c63224e1 ("dnsmasq: rework jail mounts")
Reported-by: Hartmut Birr <e9hack@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit ac4e8aa2f8)
* split into multiple lines to improve readability
* use EXTRA_MOUNT for addnhosts instead of blindly adding /tmp/hosts
* remove no longer needed mount for /sbin/hotplug-call
* add dhcp-script.sh dependencies (jshn, ubus)
Fixes: 3a94c2ca5c ("dnsmasq: add /tmp/hosts/ to jail_mount")
Fixes: aed95c4cb8 ("dnsmasq: switch to ubus-based hotplug call")
Reported-by: Stijn Tintel <stijn@linux-ipv6.be>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 59c63224e1)
'--local' is a synonym for '--server' so let's use '--local' in the
resultant config file for uci's 'local' instead of uci's local
parameter being turned into '--server'. Slightly less confusion all
round.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit e4cfefa9fc)
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
When running multiple instances of dnsmasq, for example one being for the lan
and another for a guest network, it might not be desirable to have the same dns names
configured in both networks
Signed-off-by: João Henriques <joaoh88@gmail.com>
(cherry picked from commit e8a5670122)
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Currently, when using multiple dnsmasq instances they are all assigned
to the same Ubus instance name. This does not work, as only a single
instance can register with Ubus at a time. In the log, this leads to
`Cannot add object to UBus: Invalid argument` error messages.
Furthermore, upstream 3c93e8eb41952a9c91699386132d6fe83050e9be changes
behaviour so that instead of the log, dnsmasq exits at start instead.
With this patch, all dnsmasq instances are assigned unique names so that
they can register with Ubus concurrently. One of the enabled instances
is always assigned the previous default name "dnsmasq" to avoid breaking
backwards compatibility with other software relying on that default.
Previously, a random instance got assigned that name (while the others
produced error logs). Now, the first unnamed dnsmasq config section is
assigned the default name. If there are no unnamed dnsmasq sections the
first encountered named dnsmasq config section is assigned instead.
A similar issue exists for Dbus and was similarly addressed.
Signed-off-by: Etan Kissling <etan.kissling@gmail.com>
[tweaked commit message] dnsmasq was not crashing it is exiting
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit ba5bd8e556)
Programs like the olsr-name-plugin write hostname files to "/tmp/hosts/".
If you don't add this to the jail_mount, dnsmasq can't read it anymore.
Signed-off-by: Nick Hainke <vincent@systemli.org>
(cherry picked from commit 3a94c2ca5c)
In the aftermath of the KRACK attacks, hostapd gained an AP-side workaround
against WNM-Sleep Mode GTK/IGTK reinstallation attacks. WNM Sleep Mode is not
enabled by default on OpenWrt, but it is configurable through the option
wnm_sleep_mode. Thus, make the AP-side workaround configurable as well by
exposing the option wnm_sleep_mode_no_keys. If you use the option
wpa_disable_eapol_key_retries and have wnm_sleep_mode enabled, you might
consider using this workaround.
Signed-off-by: Timo Sigurdsson <public_timo.s@silentcreek.de>
[bump PKG_RELEASE]
Signed-off-by: Paul Spooren <mail@aparcar.org>
(cherry picked from commit bf98faaac8)
The country3 option in hostapd.conf allows the third octet of the country
string to be set. It can be used e.g. to indicate indoor or outdoor use (see
hostapd.conf for further details). Make this option configurable but optional
in OpenWrt.
Signed-off-by: Timo Sigurdsson <public_timo.s@silentcreek.de>
[bump PKG_RELEASE, rebase]
Signed-off-by: Paul Spooren <mail@aparcar.org>
(cherry picked from commit 9f09c1936a)
Signed-off-by: Paul Spooren <mail@aparcar.org>
Commit 0a7657c ("hostapd: add channel utilization as config option") added the
two new uci options bss_load_update_period and chan_util_avg_period. However,
the corresponding "config_add_int" calls for these options weren't added, so
attempting to actually use these options and change their values is bound to
fail - they always stay at their defaults. Add the missing code to actually
make these options work.
Fixes: 0a7657c ("hostapd: add channel utilization as config option")
Signed-off-by: Timo Sigurdsson <public_timo.s@silentcreek.de>
(cherry picked from commit 85ce590705)
Since upstream commit 6467de5a8840 ("Randomize z ordinates in
scalar mult when timing resistant") WolfSSL requires a RNG for
the EC key when built hardened which is the default.
Set the RNG for the EC key to fix connections for OWE clients.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit ddcb970274)
Fixes issue with merged DNS requests in 2.83/2.84 not being
retried on the firsts failed request causing lookup failures.
Also fixes the following security problem in dnsmasq:
* CVE-2021-3448:
If specifiying the source address or interface to be used
when contacting upstream name servers such as:
server=8.8.8.8@1.2.3.4, server=8.8.8.8@1.2.3.4#66 and
server=8.8.8.8@eth0 then all would use the same socket
bound to the explicitly configured port. Now only
server=8.8.8.8@1.2.3.4#66 will use the explicitly
configured port and the others random source ports.
Remove upstreamed patches and update remaining patch.
Signed-off-by: Alan Swanson <reiver@improbability.net>
[refreshed old runtime support patch]
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
(cherry picked from commit 3980daffa4)
Looks like 'openat', 'pipe2' and 'ppoll' are now needed, possibly due
to changes on libraries used by umdns now using slightly different
calls.
Found using
/etc/init.d/umdns trace
now use umdns, ie. cover all ubus call etc., then
/etc/init.d/umdns stop
find list of syscalls traced in /tmp/umdns.*.json
Fixes: FS#3355 ("UMDNS: does not start on master with seccomp")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 00a85a1634)
Now that ujail supports seccomp also on Aarch64, add missing syscall
'fstat' to the list of allowed syscalls.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit d28880cdd8)
On Linux 5.4, build WireGuard from backports. Linux 5.10 contains
wireguard in-tree.
Add in-kernel crypto libraries required by WireGuard along with
arch-specific optimizations.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
(cherry picked from commit 06351f1bd0)
(cherry picked from commit 464451d9ab)
If the environment variable FILTER is set before compilation,
compilation of the ppp-package will fail with the error message
Package ppp is missing dependencies for the following libraries:
libpcap.so.1
The reason is that the OpenWrt-patch for the Makefile only comments
out the line FILTER=y. Hence the pcap-library will be dynamically
linked if the environment variable FILTER is set elsewhere, which
causes compilation to fail. The fix consists on explicitly unsetting
the variable FILTER instead.
Signed-off-by: Eike Ritter <git@rittere.co.uk>
(cherry picked from commit 46cd0765d0)
15346de client: Always close connection with request body in case of error
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 1170655f8b)
Instead of doing uci commit and reload_config for each setting do it
only once when one of these options was changed. This should make it a
little faster when both conditions are taken.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 501221af54)
Without this change the config is only committed, but the uhttpd daemon
is not reloaded. This reload is needed to apply the config. Without the
reload of uhttpd, the ubus server is not available over http and returns
a Error 404.
This caused problems when installing luci on the snapshots and
accessing it without reloading uhttpd.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit d25d281fd6)
