Correct the incorrect netmask calculation logic leading to incorrect
network range comparisons in some cases.
Fixes: #6956
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Right now, when there's active DHCPv6 leases and more than one address per host
(e.g. because ULA is used together with GUA prefixes), the two IPv6 addresses
will be printed on one line (which may or may not get broken up depending on
the length of the addresses, which also looks inconsistent).
Putting each address on a separate line makes it much easier to read the
addresses (IMHO).
Signed-off-by: David Härdeman <david@hardeman.nu>
Instead of using a second, semitransparent quadliteral of the
same station color to highlight the main 20 MHz channel of a
station with a bandwidht wider than 20 MHz, use a semitransparent,
black svg gradient with the same shape and size.
This has the benefit that it is harder to mistake for a second,
overlain station.
The gradient is black because this way it can be reused for
all stations, instead of having to create a separate one for
each station color.
Signed-off-by: Frank Löffler <frank.loeffler@uni-jena.de>
* luci-mod-status: add missing vht40 channels
As of now, channel_analysis is missing 40 MHz VHT channels.
What it does for those right now is
display them as 20 MHz channels (both in the plot and the table below),
but it uses the center frequency of the 40 MHz channel, which means
their (wrong) 20 MHz band in the plot is offset by 2.
This fixes#6419. Note that this only occurs for **other** stations.
The code for local_wifi is different and already complete.
Signed-off-by: Frank Löffler <frank.loeffler@uni-jena.de>
Add missing ACL entry for storage index page.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
luci-mod-status: expand storage index page with mount points
Expand storage index page with mount points. For custom mounts point we
use the device name and we reference the mount point between ().
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
luci-mod-status: ensure each storage getMountPoints result is unique
Signed-off-by: Paul Donald <newtwen@gmail.com>
Closes#2767
As of now, channel analysis highlights the complete frequency spectrum of a base station (all channels used by the station, which might be more or fewer depending on channel bonding). This means in particular, that in the plot, a station at one main channel might look identical to another station at a different channel, because both might use the same bonded channel spectrum.
One such example is two 80 MHz stations on channels 100 and 108 respectively: both will be plotted from channel 100 up to 115, as both will use those those 80 MHz. This is not incorrect, but it makes it more difficult to see which station is really where, as they might as well switch back to 20 or 40 MHz at times, or for different clients, and where they are then is currently not visible in the plot.
This patch adds another trapez to the plot at the main 20 MHz channel of a station, with width 20 MHz, in case the station uses a wider general width. This is visible, because those polygons are partially transparent. It also moves the x-position of the label to the main channel of the station (which now is marked: this would have looked odd before).
This does not highlight the primary 40 MHz channel for 80 MHz channels and wider. It should be possible be compute this from the knowledge in luci of the primary 20 MHz channel, the center channel and the bandwidth, assuming ac logic, but I decided against adding it, to keep the plot less cluttered.
Signed-off-by: Frank Löffler <frank.loeffler@uni-jena.de>
Use carrier state to distinguish connected/no link states and simply
print "Connected" if no speed indication is available.
Fixes: #6663
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
When attempting to resolve VLAN devices, protect against potential deep
recursion due to invalid bridge configs referencing themselves.
Fixes: #6648
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Add an ID attribute to the dynamically generated lease status tables to
persist row ordering choice across reloads.
Fixes: #6640
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Enable support for the 6GHz band now that the code supports it. For that it
just needs to be added to the "bands" object.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Try to use the new luci/getBuiltinEthernetPorts RPC call to enumerate known
ports and fall back to manual board.json parsing if the call is unavailable
yet. The fallback code will be dropped in a while when everything settled.
Ref: #6534, #6538
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Improve channel graph for 5GHz radio by using dotted lines to identify
the intermediate channel, full line to identify the different channel
band and bold line to identify the different section of the wifi
channels.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Channel tag for 5GHz radios is wrong and is confusing for any user
that wants to better identify the used channel from the graph.
Add some additional logic to better position the channel tag.
Fixes: #6419
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
When the new result is empty, including when another scan is running in the background, the table was emptied, but the graph persisted.
With this change, the list will persist and missing APs will be shown faded.
Signed-off-by: Milad Mohtashamirad <miladmohtashamirad@gmail.com>
As there is now a page in luci-mod-dsl that shows detailed DSl status,
it is no longer necessary to have the entire information in the overview
page. Only show the most important data on the overview instead.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
Currently, non-existing values (null or undefined) are formatted like
zero values. Add explicit handling to allow the user to distinguish
between those and actual zero values.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
With the material theme, there is a gap between the tab view and the
content. The css class 'cbi-section' is missing here in the data-tab div.
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
The neighbour display parses the output of 'ip -4 neigh show' e.g.
192.168.219.95 dev eth1 lladdr 38:b4:d3:c9:b2:0c REACHABLE
192.168.219.200 dev eth1 lladdr 04:c4:61:12:f2:d2 STALE
Using regexp /^([0-9a-f:.]+) (.+) (\S+)$/ Unfortunately there's a
space character that sneaks in at the end of line, so let's make this
less brittle by accepting lines that may end with spaces e.g.
/^([0-9a-f:.]+) (.+) (\S+) *$/
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Handle HE channel width for channel analysis status page and correctly
draw local interface band width.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
In refactoring the code there was an error and the sort function wasn't
correctly reworked with diff not correctly dropped.
Drop that and correctly sort the WiFi ap if the channel is different.
Fixes: 75dcb09754 ("luci-mod-status: improve channel_analysis page")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Instead of scraping the LuCI version from Lua sources, fetch it via ubus
in order to be independent from the Lua runtime.
Fixes: #6154
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Many user complained problem with using pool with wifi scan. This comes
from the limitation that some wifi driver have problems with scanning
nearby wifi and keeping traffic.
Fix this by doing the wifi scan only one time on page load and provide a
button to refresh the channels manually. The original implementation is
preserved as the user can simply reenable the poll referesh from the ui.
While at it also sort the table by channel instead of by signal quality
to better track the most used channels in the table.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
[rewrap commit message]
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
The `network.getDSLModemType()` function returns a promise, so handle it
accordingly.
Fixes: 45ab2cd6be ("luci-mod-status: use network.getDSLModemType()")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
This matches what luci-mod-network does for the DSL modem configuration.
Since this is based on a common uci config, it may also prevent issues as
fixed with 111c551c "luci-base: fix DSL feature detection" in the future.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Do not expose the iptables status page as menu item when nftables is present
on the system. Instead add a warning banner to the nftables status page
directing the user to the hidden iptables status page when we encounter
legacy rules on the system.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Added code to display fcdn-name in DHCPv4 lease table.
(based on code in DHCPv6 lease table)
Signed-off-by: Max S Kash <asukms@ya.ru>
Indentation adjused and wrapped commit message
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
