DHCPCD Problems on eduroam (uni network)

m4ttj · 2021-04-28 19:48:48

Hi, I'm using NetworkManager, dhcpcd, and iwd to connect to my university's WiFi network eduroam. It uses 8021x authentication with PEAP and so I have to configure authentication via iwd conf files. This all seems to be working as I can successfully connect but every time (maybe not every?) dhcpcd needs to renew my IP lease it seems to lose connection, despite the lease seemingly going through okay. This is all fixed by using systemd to restart the dhcpcd service but obviously this isn't ideal. Any ideas? dhcpcd logs: https://pastebin.com/agwZnznh (with debug enabled) NetworkManager logs: https://pastebin.com/AsR0XnUy (don't seem helpful) iwd outputs nothing at the default verbosity when this occurs.

In dhcpcd I notice something about dhcp_envoption 119 but this doesn't seem relevant as it's DNS related and DNS seems to be working fine.

Note: this only seems to occur on large networks - on my home network it is fine but this also occurs on another large (separate) university network.

Last edited by m4ttj (2021-04-28 19:51:09)

rsmarples · 2021-04-29 12:39:58

Option 119 is Domain Search List - https://tools.ietf.org/html/rfc3397

The chances are the DHCP server has it ASCII encoded which is not RFC compliant.
It should be DOMAIN encoded.

You can adjust dhcpcd.conf to hide the DHCP server error by stating it can be encoded differently:

# RFC definition for domain search
# define 119	array domain		domain_search

# DHCP server definition for domain search
define 119	string		domain_search

m4ttj · 2021-04-29 13:34:04

rsmarples wrote:

You can adjust dhcpcd.conf to hide the DHCP server error by stating it can be encoded differently:

Done. They shouldn't have been causing disconnects like those I experience right? It must be something else?

rsmarples wrote:

The chances are the DHCP server has it ASCII encoded which is not RFC compliant.

If it helps, I wouldn't be surprised to find out this isn't the only thing that isn't RFC compliant.

rsmarples · 2021-04-29 13:48:12

m4ttj wrote:

rsmarples wrote:
You can adjust dhcpcd.conf to hide the DHCP server error by stating it can be encoded differently:
Done. They shouldn't have been causing disconnects like those I experience right? It must be something else?

Correct.

It's probably because you are running dhcpcd and NetworkManager via systemd.
Run only one or the other via systemd.

I know which I would run, but then I am heavily biased

m4ttj · 2021-04-29 13:56:25

rsmarples wrote:

I know which I would run, but then I am heavily biased

Was confused for a moment until I looked up your username - it's cool that you look on the forums! Thanks for your help!

rsmarples wrote:

It's probably because you are running dhcpcd and NetworkManager via systemd.
Run only one or the other via systemd.

As in both are `systemctl enable` and so systemd starting them both is a problem? I've disabled dhcpcd I'm afraid as I use NetworkManager for my entire network setup (and it should spawn dhcpcd anyway right?). Will I still be able to do `systemctl status dhcpcd` to check the logs when spawned by NetworkManager.

I occasionally noticed double default routes when I ran `ip route` something like

default via 10.154.160.1 dev wlan0 proto dhcp metric 600 
default via 10.154.160.1 dev wlan0 proto dhcp src 10.154.161.145 metric 3003

is this normal or is this a byproduct of two both being enabled?

rsmarples · 2021-04-29 14:05:50

Byproduct of running them both.
I can't answer the logging question as I'm not a systemd user.
However, dhcpcd will always log to syslog daemon loglevel, so look there.

Or, enable file logging in dhcpcd

logfile /var/log/dhcpcd.log

Scimmia · 2021-04-29 14:11:21

m4ttj wrote:

rsmarples wrote:
I know which I would run, but then I am heavily biased
Was confused for a moment until I looked up your username - it's cool that you look on the forums! Thanks for your help!
rsmarples wrote:
It's probably because you are running dhcpcd and NetworkManager via systemd.
Run only one or the other via systemd.
As in both are `systemctl enable` and so systemd starting them both is a problem? I've disabled dhcpcd I'm afraid as I use NetworkManager for my entire network setup (and it should spawn dhcpcd anyway right?). Will I still be able to do `systemctl status dhcpcd` to check the logs when spawned by NetworkManager.

NetworkManager hasn't supported dhcpcd for a long time now. It will use either dhclient or it's internal dhcp client.

Last edited by Scimmia (2021-04-29 14:11:51)

rsmarples · 2021-04-29 14:15:55

Scimmia wrote:

NetworkManager hasn't supported dhcpcd for a long time now. It will use either dhclient or it's internal dhcp client.

I know for a fact NetworkManager-1.30+ supports dhcpcd-9.3.3+

https://gitlab.freedesktop.org/NetworkM … ef02d78f71

Scimmia · 2021-04-29 14:17:27

Not in Arch it doesn't.

https://github.com/archlinux/svntogit-p … GBUILD#L63

Last edited by Scimmia (2021-04-29 14:18:31)

rsmarples · 2021-04-29 14:24:56

So the good news is that can now be changed

Lone_Wolf · 2021-04-30 19:37:21

I created a feature request for that, https://bugs.archlinux.org/task/70666

m4ttj · 2021-05-03 20:44:22

Although it seems (partially?) fixed now I still get random disconnects. dmesg seems to have something to say though:

[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Queue 4 is active on fifo 1 and stuck for 10000 ms. SW [69, 91] HW [69, 91] FH TRB=0x0c0104054
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Microcode SW error detected.  Restarting 0x2000000.
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Start IWL Error Log Dump:
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Status: 0x00000040, count: 6
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Loaded firmware version: 36.ad812ee0.0 8265-36.ucode
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000084 | NMI_INTERRUPT_UNKNOWN       
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x200002F2 | trm_hw_status0
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | trm_hw_status1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00024A8C | branchlink2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0003AC02 | interruptlink1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00013430 | interruptlink2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | data1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000080 | data2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x07830000 | data3
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x4B40390D | beacon time
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0xD094C7EA | tsf low
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000111 | tsf hi
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | time gp1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x000D14A9 | time gp2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000001 | uCode revision type
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000024 | uCode version major
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0xAD812EE0 | uCode version minor
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000230 | hw version
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00C89000 | board version
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0000001C | hcmd
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x02322042 | isr0
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00800000 | isr1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x08001802 | isr2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x004000C0 | isr3
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | isr4
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00F6019C | last cmd Id
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | wait_event
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00004306 | l2p_control
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00001420 | l2p_duration
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000003 | l2p_mhvalid
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x000000FE | l2p_addr_match
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0000000D | lmpm_pmg_sel
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x14100651 | timestamp
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0034E8F0 | flow_handler
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | ADVANCED_SYSASSERT
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac branchlink1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac branchlink2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac interruptlink1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac interruptlink2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac data1
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac data2
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac data3
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac major
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | umac minor
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | frame pointer
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | stack pointer
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | last host cmd
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000000 | isr status reg
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Fseq Registers:
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x68E9BFA6 | FSEQ_ERROR_CODE
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0xDFF42C00 | FSEQ_TOP_INIT_VERSION
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x841B68AD | FSEQ_CNVIO_INIT_VERSION
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0000A10B | FSEQ_OTP_VERSION
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x5530B46D | FSEQ_TOP_CONTENT_VERSION
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x3C1B66DF | FSEQ_ALIVE_TOKEN
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0xC9CC3DAB | FSEQ_CNVI_ID
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x4AB71496 | FSEQ_CNVR_ID
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x00000010 | CNVI_AUX_MISC_CHIP
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0BADCAFE | CNVR_AUX_MISC_CHIP
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0BADCAFE | CNVR_SCU_SD_REGS_SD_REG_DIG_DCDC_VTRIM
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: 0x0BADCAFE | CNVR_SCU_SD_REGS_SD_REG_ACTIVE_VDIG_MIRROR
[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Collecting data: trigger 2 fired.
[Mon May  3 21:43:26 2021] ieee80211 phy0: Hardware restart was requested

Any ideas?

Last edited by m4ttj (2021-05-03 20:44:57)

seth · 2021-05-04 05:58:02

[Mon May  3 21:43:26 2021] iwlwifi 0000:02:00.0: Microcode SW error detected.  Restarting 0x2000000.
…
[Mon May  3 21:43:26 2021] ieee80211 phy0: Hardware restart was requested

https://bbs.archlinux.org/viewtopic.php?id=256583
https://bbs.archlinux.org/viewtopic.php?id=260750

https://wireless.wiki.kernel.org/en/use … iffer_mode ?

m4ttj · 2021-05-05 22:57:36

seth wrote:

https://bbs.archlinux.org/viewtopic.php?id=256583
https://bbs.archlinux.org/viewtopic.php?id=260750
https://wireless.wiki.kernel.org/en/use … iffer_mode ?

None those really offer any help. I'm running the latest clear linux kernel (still occurred on mainline) and everything is up-to-date. The second link you sent suggested that power management is disabled on the latest kernels and so I haven't added that which seems to be the solution for others.

I do have the Intel Corporation Wireless 8265 / 8275 though.

Last edited by m4ttj (2021-05-05 23:01:05)

seth · 2021-05-06 08:12:41

The first links are to illustrate that there's an exisiting pattern of problem reports you might be interested in (and for me to bump if we solve it here…) and

the last link actually wrote:

About the monitor / sniffer mode
Our devices support monitor mode. When you have VHT APs around, you should load the iwlwifi module with
amsdu_size=3
This will put lots of pressure on the memory subsystem, but it will allow you to hear 12K long packets. You may see firmware crashes in case you didn't set that module parameter.

m4ttj · 2021-05-06 11:11:55

seth wrote:

The first links are to illustrate that there's an exisiting pattern of problem reports you might be interested in (and for me to bump if we solve it here…) and
the last link actually wrote:
About the monitor / sniffer mode
Our devices support monitor mode. When you have VHT APs around, you should load the iwlwifi module with
amsdu_size=3
This will put lots of pressure on the memory subsystem, but it will allow you to hear 12K long packets. You may see firmware crashes in case you didn't set that module parameter.

Ah - I understand. Is there anything I can do to help?

seth · 2021-05-06 13:33:46

the last link actually wrote:

About the monitor / sniffer mode
Our devices support monitor mode. When you have VHT APs around, you should load the iwlwifi module with
amsdu_size=3
This will put lots of pressure on the memory subsystem, but it will allow you to hear 12K long packets. You may see firmware crashes in case you didn't set that module parameter.

Arch Linux

#1 2021-04-28 19:48:48

DHCPCD Problems on eduroam (uni network)

#2 2021-04-29 12:39:58

Re: DHCPCD Problems on eduroam (uni network)

#3 2021-04-29 13:34:04

Re: DHCPCD Problems on eduroam (uni network)

#4 2021-04-29 13:48:12

Re: DHCPCD Problems on eduroam (uni network)

#5 2021-04-29 13:56:25

Re: DHCPCD Problems on eduroam (uni network)

#6 2021-04-29 14:05:50

Re: DHCPCD Problems on eduroam (uni network)

#7 2021-04-29 14:11:21

Re: DHCPCD Problems on eduroam (uni network)

#8 2021-04-29 14:15:55

Re: DHCPCD Problems on eduroam (uni network)

#9 2021-04-29 14:17:27

Re: DHCPCD Problems on eduroam (uni network)

#10 2021-04-29 14:24:56

Re: DHCPCD Problems on eduroam (uni network)

#11 2021-04-30 19:37:21

Re: DHCPCD Problems on eduroam (uni network)

#12 2021-05-03 20:44:22

Re: DHCPCD Problems on eduroam (uni network)

#13 2021-05-04 05:58:02

Re: DHCPCD Problems on eduroam (uni network)

#14 2021-05-05 22:57:36

Re: DHCPCD Problems on eduroam (uni network)

#15 2021-05-06 08:12:41

Re: DHCPCD Problems on eduroam (uni network)

About the monitor / sniffer mode

#16 2021-05-06 11:11:55

Re: DHCPCD Problems on eduroam (uni network)

About the monitor / sniffer mode

#17 2021-05-06 13:33:46

Re: DHCPCD Problems on eduroam (uni network)

About the monitor / sniffer mode

Board footer