luci-mod-network: The expand-hosts seems to also affect the AAAA DNS records.

[adelton]

• This PR is not from my main or master branch 💩, but a separate branch ✅
• Each commit has a valid ✒️ Signed-off-by: <my@email.address> row (via git commit --signoff)
• Each commit and PR title has a valid 📝 <package name>: title first line subject for packages
• Incremented 🆙 any PKG_VERSION in the Makefile
• Tested on: (architecture, openwrt version, browser) ✅
• ( Preferred ) Mention: @ the original code author for feedback
• ( Preferred ) Screenshot or mp4 of changes:
• ( Optional ) Closes: e.g. openwrt/luci#issue-number
• ( Optional ) Depends on: e.g. openwrt/packages#pr-number in sister repo
• Description: (describe the changes proposed in this PR)

I've spent quite some time trying to figure out why

dig @192.168.1.1 -t a laptop.lan

returns the A record and

dig @192.168.1.1 -t aaaa laptop.lan

does not return the AAAA record, even if I saw the IPv6 address assigned to the host and listed in the Active DHCPv6 Leases.

I finally found out that I have the Expand hosts unchecked (because I prefer to use the full hostnames) and that it does not affect the A records but does affect the AAAA records. It simply did not occur to me to just check

dig @192.168.1.1 -t aaaa laptop

and I thought something about AAAA records is broken.

If there was a note in the UI mentioning the relationship of the Expand hosts to the IPv6 addresses and AAAA records, it would have saved me nontrivial amount of time. Adding a note to help others who might find themselves in the same situation.

[systemcrash]

Some nuance like this is worth investigating. A problem more likely lies in how dnsmasq works. The setting you mention is a plausible offender. But your amendment invalidates an existing translation. It's better to add a new follow-on string.

E.g.

... + ' ' +_('your string');

Because a host has IPs via DHCP doesn't mean a dns lookup will automatically work. You likely need to assign it a static IP under host names in DHCP.

[adelton]

Thanks for the hint about preserving the existing translations, updated in 6b1c47c.

I specifically kept the wording a little bit vague WRT to when the AAAA records get actually created. But on a stock OpenWrt 23.05.5 (r24106-10cc5fcd00) I see

$ for i in openwrt openwrt.lan laptop laptop.lan ; do for t in a aaaa ; do echo "$i $t: $( dig +short @192.168.1.1 -t $t $i )" ; done ; done
openwrt a: 192.168.1.1
openwrt aaaa: fddd:7eab:3d2a::1
openwrt.lan a: 192.168.1.1
openwrt.lan aaaa: fddd:7eab:3d2a::1
laptop a: 192.168.1.142
laptop aaaa: fddd:7eab:3d2a::bb3
laptop.lan a: 192.168.1.142
laptop.lan aaaa: fddd:7eab:3d2a::bb3

and when I uncheck the Expand hosts and Save & Apply, I see

$ for i in openwrt openwrt.lan laptop laptop.lan ; do for t in a aaaa ; do echo "$i $t: $( dig +short @192.168.1.1 -t $t $i )" ; done ; done
openwrt a: 192.168.1.1
openwrt aaaa: fddd:7eab:3d2a::1
openwrt.lan a: 
openwrt.lan aaaa: 
laptop a: 192.168.1.142
laptop aaaa: fddd:7eab:3d2a::bb3
laptop.lan a: 192.168.1.142
laptop.lan aaaa: 

[systemcrash]

Can you show me the content of your /etc/hosts file, and your /etc/config/dhcp file? (redact or mutate anything considered sensitive)

[adelton]

The vanilla content after resetting to defaults is

root@OpenWrt:~# cat /etc/hosts
127.0.0.1 localhost

::1     localhost ip6-localhost ip6-loopback
ff02::1 ip6-allnodes
ff02::2 ip6-allrouters
root@OpenWrt:~# cat /etc/config/dhcp

config dnsmasq
	option domainneeded '1'
	option boguspriv '1'
	option filterwin2k '0'
	option localise_queries '1'
	option rebind_protection '1'
	option rebind_localhost '1'
	option local '/lan/'
	option domain 'lan'
	option expandhosts '1'
	option nonegcache '0'
	option cachesize '1000'
	option authoritative '1'
	option readethers '1'
	option leasefile '/tmp/dhcp.leases'
	option resolvfile '/tmp/resolv.conf.d/resolv.conf.auto'
	option nonwildcard '1'
	option localservice '1'
	option ednspacket_max '1232'
	option filter_aaaa '0'
	option filter_a '0'

config dhcp 'lan'
	option interface 'lan'
	option start '100'
	option limit '150'
	option leasetime '12h'
	option dhcpv4 'server'
	option dhcpv6 'server'
	option ra 'server'
	option ra_slaac '1'
	list ra_flags 'managed-config'
	list ra_flags 'other-config'

config dhcp 'wan'
	option interface 'wan'
	option ignore '1'

config odhcpd 'odhcpd'
	option maindhcp '0'
	option leasefile '/tmp/hosts/odhcpd'
	option leasetrigger '/usr/sbin/odhcpd-update'
	option loglevel '4'

When I uncheck that Expand hosts and Save & Apply, the /etc/hosts is the same and the diff in /etc/config/dhcp is

--- dhcp.orig	2025-01-06 08:23:26.949500182 +0100
+++ dhcp	2025-01-06 08:23:07.720432382 +0100
@@ -1,25 +1,18 @@
 
 config dnsmasq
 	option domainneeded '1'
-	option boguspriv '1'
-	option filterwin2k '0'
 	option localise_queries '1'
 	option rebind_protection '1'
 	option rebind_localhost '1'
 	option local '/lan/'
 	option domain 'lan'
-	option expandhosts '1'
-	option nonegcache '0'
 	option cachesize '1000'
 	option authoritative '1'
 	option readethers '1'
 	option leasefile '/tmp/dhcp.leases'
 	option resolvfile '/tmp/resolv.conf.d/resolv.conf.auto'
-	option nonwildcard '1'
 	option localservice '1'
 	option ednspacket_max '1232'
-	option filter_aaaa '0'
-	option filter_a '0'
 
 config dhcp 'lan'
 	option interface 'lan'

[systemcrash]

It's possible your tests may have hit cached entries. Also, the dnsmasq manpage has this to say about host-records (Unlike hosts-files, names are not expanded, even when --expand-hosts is in effect):

--host-record=<name>[,<name>....],[<IPv4-address>],[<IPv6-address>][,<TTL>]
Add A, AAAA and PTR records to the DNS. This adds one or more names to the DNS with associated IPv4 (A) and IPv6 (AAAA) records. A name may appear in more than one --host-record and therefore be assigned more than one address. Only the first address creates a PTR record linking the address to the name. This is the same rule as is used reading hosts-files. --host-record options are considered to be read before host-files, so a name appearing there inhibits PTR-record creation if it appears in hosts-file also. Unlike hosts-files, names are not expanded, even when --expand-hosts is in effect. Short and long names may appear in the same --host-record, eg. --host-record=laptop,laptop.thekelleys.org,192.168.0.1,1234::100

These are defined by hostrecord types in the dhcp config. They're not in the GUI yet.

Compare with host entries, which correspond to --dhcp-host and do receive the suffix.

So the question is: where do your entries come from? e.g. laptop above?

[adelton]

I see laptop mentioned in

/var/dhcp.leases:1727138289 90:2e:16:aa:bb:cc 192.168.1.142 laptop 01:90:2e:16:aa:bb:cc
/var/hosts/odhcpd:fd7f:422f:4895::bb3	laptop
/var/hosts/odhcpd:# br-lan 0004804054c45add8a5b5fdf7aac01234567 667b932a laptop 1727138286 bb3 128 fd7f:422f:4895::bb3/128

These records do not change when I toggle Expand hosts back and forth (and Save & Apply), the mtimes stay the same, while the dig output changes. (The IPv6 addresses changed since the first time I showed the dig outputs above because I've reset the router to defaults which reset ULA prefix.)

[systemcrash]

Then this seems related not to 'dns' entries per sé, but DHCP entries. Wording here is important, if we are going to put a message that users will read. In which case, let's revise it to:

Also affects DHCP(v6)-derived names

as stated in the dnsmasq manpage.

There's a big but here: there might be undefined behaviour or a bug in dnsmasq regarding this. I don't see why AAAA should be absent when this is changed, and A get treated.

[systemcrash]

Actually, what is the content of your /etc/resolv.conf in both cases? Since the lines in the odhcpd files are controlled not by dnsmasq but odhcpd which uses

https://linux.die.net/man/3/dn_expand

[adelton]

Then this seems related not to 'dns' entries per sé, but DHCP entries. Wording here is important, if we are going to put a message that users will read. In which case, let's revise it to:

Also affects DHCP(v6)-derived names

as stated in the dnsmasq manpage.

Updated, rebased on master.

There's a big but here: there might be undefined behaviour or a bug in dnsmasq regarding this. I don't see why AAAA should be absent when this is changed, and A get treated.

I don't disagree. :-)

Actually, what is the content of your /etc/resolv.conf in both cases? Since the lines in the odhcpd files are controlled not by dnsmasq but odhcpd which uses https://linux.die.net/man/3/dn_expand

The content of the /etc/resolv.conf is

search lan
nameserver 127.0.0.1
nameserver ::1

and the Expand hosts checkbox does not seem to have any effect on that content.

[systemcrash]

The content of the /etc/resolv.conf is

search lan
nameserver 127.0.0.1
nameserver ::1

and the Expand hosts checkbox does not seem to have any effect on that content.

Is it possible that your resolv file gets clobbered after changing the setting? That search lan bit might be interpreted differently by various daemons.

[adelton]

Can you elaborate on that "clobbered"? (Non-native speaker here.)

[systemcrash]

Replaced with something completely different?

[adelton]

Ah. No, the content of /etc/resolv.conf does not seem to get affected at all by changing the expandhosts value:

root@OpenWrt:~# grep expand /etc/config/dhcp 
	option expandhosts '1'
root@OpenWrt:~# cat /etc/resolv.conf 
search lan
nameserver 127.0.0.1
nameserver ::1
root@OpenWrt:~# sha256sum /etc/resolv.conf
7d6806759821bb1b18ad3ce81edec5f082c2c0dda30f17907016591fb4d6e84c  /etc/resolv.conf

root@OpenWrt:~# uci del dhcp.cfg01411c.expandhosts
root@OpenWrt:~# uci commit
root@OpenWrt:~# grep expand /etc/config/dhcp 
root@OpenWrt:~# sha256sum /etc/resolv.conf
7d6806759821bb1b18ad3ce81edec5f082c2c0dda30f17907016591fb4d6e84c  /etc/resolv.conf

root@OpenWrt:~# uci set dhcp.cfg01411c.expandhosts=0
root@OpenWrt:~# uci commit
root@OpenWrt:~# grep expand /etc/config/dhcp 
	option expandhosts '0'
root@OpenWrt:~# sha256sum /etc/resolv.conf
7d6806759821bb1b18ad3ce81edec5f082c2c0dda30f17907016591fb4d6e84c  /etc/resolv.conf
root@OpenWrt:~# 

This was on a freshly installed OpenWrt 24.10.0-rc5 (r28304-6dacba30a7) x86_64 VM.