UAP-AC-PRO not reachable after first bootup

For Developers
1

My uap-ac-pro stops replying to network packets after printing the
Please press button now to enter failsafe message over the network.
More closely related: https://www.mail-archive.com/openwrt-devel@lists.openwrt.org/msg55379.html

Possibly related, but not solving the issue: https://github.com/openwrt/openwrt/commit/2dffadece9a7243a236ce7d91719787a671e23d4#comments

Im on current master with the following config diff:

CONFIG_TARGET_ath79=y
CONFIG_TARGET_ath79_generic=y
CONFIG_TARGET_ath79_generic_DEVICE_ubnt_unifiac-pro=y
CONFIG_DEVEL=y
CONFIG_TOOLCHAINOPTS=y
# CONFIG_BINUTILS_USE_VERSION_2_34 is not set
CONFIG_BINUTILS_USE_VERSION_2_35_1=y
CONFIG_BINUTILS_VERSION="2.35.1"
CONFIG_BINUTILS_VERSION_2_35_1=y
CONFIG_GCC_USE_VERSION_10=y
# CONFIG_GCC_USE_VERSION_8 is not set
CONFIG_GCC_VERSION="10.2.0"
CONFIG_GCC_VERSION_10=y
# CONFIG_GDB is not set
CONFIG_PACKAGE_ATH_DYNACK=y
CONFIG_PACKAGE_ath10k-firmware-qca988x=y
# CONFIG_PACKAGE_ath10k-firmware-qca988x-ct is not set
CONFIG_PACKAGE_cgi-io=y
# CONFIG_PACKAGE_dnsmasq is not set
CONFIG_PACKAGE_kmod-ath10k=y
# CONFIG_PACKAGE_kmod-ath10k-ct is not set
CONFIG_PACKAGE_kmod-ledtrig-default-on=y
CONFIG_PACKAGE_kmod-ledtrig-heartbeat=y
CONFIG_PACKAGE_kmod-ledtrig-netdev=y
CONFIG_PACKAGE_kmod-ledtrig-timer=y
# CONFIG_PACKAGE_kmod-nls-base is not set
# CONFIG_PACKAGE_kmod-usb-core is not set
# CONFIG_PACKAGE_kmod-usb2 is not set
CONFIG_PACKAGE_libiwinfo-lua=y
CONFIG_PACKAGE_libltdl=y
CONFIG_PACKAGE_liblua=y
CONFIG_PACKAGE_liblucihttp=y
CONFIG_PACKAGE_liblucihttp-lua=y
CONFIG_PACKAGE_libubus-lua=y
CONFIG_PACKAGE_lua=y
CONFIG_PACKAGE_luci=y
CONFIG_PACKAGE_luci-app-firewall=y
CONFIG_PACKAGE_luci-app-opkg=y
CONFIG_PACKAGE_luci-base=y
CONFIG_PACKAGE_luci-compat=y
CONFIG_PACKAGE_luci-lib-base=y
CONFIG_PACKAGE_luci-lib-ip=y
CONFIG_PACKAGE_luci-lib-iptparser=y
CONFIG_PACKAGE_luci-lib-jsonc=y
CONFIG_PACKAGE_luci-lib-nixio=y
CONFIG_PACKAGE_luci-mod-admin-full=y
CONFIG_PACKAGE_luci-mod-network=y
CONFIG_PACKAGE_luci-mod-status=y
CONFIG_PACKAGE_luci-mod-system=y
CONFIG_PACKAGE_luci-proto-ipv6=y
CONFIG_PACKAGE_luci-proto-ppp=y
CONFIG_PACKAGE_luci-theme-bootstrap=y
CONFIG_PACKAGE_luci-theme-material=y
CONFIG_PACKAGE_rpcd=y
CONFIG_PACKAGE_rpcd-mod-file=y
CONFIG_PACKAGE_rpcd-mod-iwinfo=y
CONFIG_PACKAGE_rpcd-mod-luci=y
CONFIG_PACKAGE_rpcd-mod-rrdns=y
CONFIG_PACKAGE_ucert=y
CONFIG_PACKAGE_uhttpd=y
CONFIG_PACKAGE_uhttpd-mod-ubus=y
# CONFIG_PACKAGE_wpad-basic-wolfssl is not set
CONFIG_PACKAGE_wpad-wolfssl=y
CONFIG_PACKAGE_zlib=y
CONFIG_TARGET_OPTIMIZATION="-O2 -pipe -mno-branch-likely -mips32r2 -mtune=74kc"
CONFIG_TARGET_OPTIONS=y
# CONFIG_ATH10K_LEDS is not set
CONFIG_PACKAGE_kmod-hwmon-core=y

Is there anyone else using this device and with a working installation on current master?
Im going to try figuring out why it fails to boot.

[    6.044834] eth0: link up (1000Mbps/Full duplex)
[    6.054574] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready
[    6.069927] IPv6: ADDRCONF(NETDEV_CHANGE): eth0.1: link becomes ready
Press the [f] key and hit [enter] to enter failsafe mode
Press the [1], [2], [3] or [4] key and hit [enter] to [    6.110662] urandom_read: 4 callbacks suppressed
select the debug[    6.110668] random: procd: uninitialized urandom read (4 bytes read)
 level
[   10.296498] mount_root: jffs2 not ready yet, using temporary tmpfs overlay
[   10.306903] urandom-seed: Seed file not found (/etc/urandom.seed)
[   10.385689] eth0: link down
[   51.254485] random: crng init done

This is the last thing it prints, and then it keeps blinking the onboard led and locks up

[   10.028174] procd: running /etc/init.d/ubus running
[   10.041609] procd: glob failed on /etc/init.d/ubus
[   10.100554] random: ubusd: uninitialized urandom read (4 bytes read)
[   10.107704] procd: Connection to ubus failed
[   10.211963] procd: Connection to ubus failed
[   10.416250] procd: Connection to ubus failed
[   10.820736] procd: Connection to ubus failed
[   11.626669] procd: Connection to ubus failed
[   12.632740] procd: Connection to ubus failed
[   13.638863] procd: Connection to ubus failed
[   14.644944] procd: Connection to ubus failed
[   14.918610] procd: Ping
[   15.651092] procd: Connection to ubus failed
[   16.657182] procd: Connection to ubus failed
[   17.663097] procd: Connection to ubus failed
[   18.669185] procd: Connection to ubus failed
[   19.675097] procd: Connection to ubus failed
[   19.920937] procd: Ping
[   20.681869] procd: Connection to ubus failed
[   21.687099] procd: Connection to ubus failed
[   22.693185] procd: Connection to ubus failed

hmmm /etc/init.d/ubus does not even exist in my image (staging dir).

2

I think i found a part of the problem:

root@(none):/etc/config# ubus
ubus   ubusd
root@(none):/etc/config# ubusd
usock: No such file or directory
root@(none):/etc/config# mkdir /var/run/ubus
root@(none):/etc/config# ubusd

Is it possible that this is not created on first boot?

3

Maybe the same issue like x86_64's , you may find solution in my post.

1 Like
4

The only thing missing on current openwrt that is in your thread is the shadow + group+passwd file edits.
Imma try this and see if it changes something.

Edit:
Interestingly the files are changed correctly in the rootfs when in recovery console.

5

Did not work, i dont know whats going on here, either ubusd is not starting correctly during a normal system boot or procd cant connect to it properly.

Im going to try building with the default toolchain for now.
Edit: Not related to the toolchain, i ran the defaults.

Current master just does not run on this device.

19.07.4 boots fine, im going to try building this one aswell to rule out my buildhost.

6

you should see something like these in your vanilla rootfs...

############!/bin/sh
#[ "$(uci get rpcd.@rpcd[0].socket)" = "/var/run/ubus.sock"  ] && {
#    uci set rpcd.@rpcd[0].socket='/var/run/ubus/ubus.sock'
#        uci commit rpcd
#        
#}
#[ "$(uci -q get uhttpd.main.ubus_socket)" = "/var/run/ubus.sock"  ] && {
#    uci set uhttpd.main.ubus_socket='/var/run/ubus/ubus.sock'
#        uci commit uhttpd
#        
#}

fgrep -r ubus build_dir/*/root* | grep sock
1 Like
7

Okay this is exactly what i feared.

Clean installs dont work correctly, when i install 19.07.4 first and then upgrade to latest master it works correctly.
Not sure whats broken here but its super critical if not fixed before release.

8

gee... I would have assumed the opposite...

in my case, it was the 'keep settings' that trips the new commits up... ( is that device ext4? )

i'm sure they will resolve... master is master for just this purpose... usually... you just come back in a few days / week or so... and forget the whole thing happened :wink:

9

Device is jffs2.

Its been a while since they touched ubus though, so i assume noone tested fresh first installs.
Hopefully the right person comes across this forum post, otherwise i might have to send an email.

1 Like
10

i don't see ubus / ubusd?

11

Its only a diff to the official default configuration file. (generated with scripts/diffconfig.sh)
Ubus/ubusd exist and work correctly when used in recovery manually. Something just doesnt create the configs / file locations for the services correctly on the very first boot.

1 Like