Ath0 not detected on Alix 2D13

phil.davis · Apr 6, 2014, 8:13 AM

Time to start playing with 2.2, so I upgraded an Alix 2D13 from 2.1.1-RELEASE to 2.2. It sat at the interface mismatch prompt on the console. It is running:
2.2-ALPHA (i386)
built on Sat Apr 5 19:36:44 CDT 2014
FreeBSD 10.0-STABLE

This one has a WiFi card in it from NetGate - KIT-ALIX-5004MP-DUAL - it did not find ath0 :(
I rebooted, and tried Interfaces->Assign but there is no ath0 listed. (Better plug in an AP to the LAN so the family and visitors get their internet back)
Error installing Sudo during re-install all packages on the console. I tried to install Sudo from the webGUI, and also got these errors about unsigned PBI. Same styler of thing for OpenVPN Client Export Utility. Cron auto-installed fine. So packages with PBIs expect them to be signed, but they are not.

Beginning package installation for sudo .
Downloading package configuration file... done.
Saving updated package information... done.
Downloading sudo and its dependencies... 
Checking for package installation... 
 Downloading https://files.pfsense.org/packages/10/All/sudo-1.8.10p2-i386.pbi ...  (extracting)
 ERROR: No digital signature! If you are *SURE* you trust this PBI, re-install with --no-checksig option.
of sudo-1.8.10p2-i386 failed!

Installation aborted.Removing package...
Starting package deletion for sudo-1.8.10p2-i386...done.
Removing sudo components...
Menu items... done.
Loading package instructions...
Include file sudo.inc could not be found for inclusion.
Removing package instructions...done.
Auxiliary files... done.
Package XML... done.
Configuration... done.
done.
Failed to install package.

Installation halted.

Beginning package installation for OpenVPN Client Export Utility .
Downloading package configuration file... done.
Saving updated package information... done.
Downloading OpenVPN Client Export Utility and its dependencies... 
Checking for package installation... 
 Downloading https://files.pfsense.org/packages/10/All/zip-3.0-i386.pbi ...  (extracting)
 ERROR: No digital signature! If you are *SURE* you trust this PBI, re-install with --no-checksig option.
of zip-3.0-i386.pbi p7zip-9.20.1_2-i386 failed!

Installation aborted.Removing package...
Starting package deletion for zip-3.0-i386...done.
Starting package deletion for p7zip-9.20.1_2-i386...done.
Removing OpenVPN Client Export Utility components...
Tabs items... done.
Loading package instructions...
Include file openvpn-client-export.inc could not be found for inclusion.
Deinstall commands... 
Not executing custom deinstall hook because an include is missing.
Removing package instructions...done.
Auxiliary files... done.
Package XML... done.
Configuration... done.
done.
Failed to install package.

Installation halted.

It has an OpenVPN client making a site-to-site connection to the office. It dies on startup with this in the log:

Apr 6 13:43:48 	openvpn[46044]: OpenVPN 2.3.2 i386-portbld-freebsd10.0 [SSL (OpenSSL)] [LZO] [eurephia] [MH] [IPv6] built on Mar 19 2014
Apr 6 13:43:48 	openvpn[46044]: NOTE: the current --script-security setting may allow this configuration to call user-defined scripts
Apr 6 13:43:49 	openvpn[46044]: TUN/TAP device /dev/tun1 opened
Apr 6 13:43:49 	openvpn[46044]: do_ifconfig, tt->ipv6=1, tt->did_ifconfig_ipv6_setup=0
Apr 6 13:43:49 	openvpn[46044]: /sbin/ifconfig ovpnc1 10.49.255.2 10.49.255.1 mtu 1500 netmask 255.255.255.255 up
Apr 6 13:43:49 	openvpn[46044]: FreeBSD ifconfig failed: external program exited with error status: 1
Apr 6 13:43:49 	openvpn[46044]: Exiting due to fatal error

Messages like this will cause endless questions on the forum:

Apr  6 13:37:28 system[270]: [WARNING] The maximum number of processes has been reached. Please review your configuration and consider raising 'process.max'

IMHO it will be worth sorting out why these pop out and getting rid of them - rather than having 'normal' warnings.

Console output from first and second boot attached for reference.
That's all so far - will get out an old AP for now so I can leave this running and see what else happens…
Alix2D13-22-upgrade-first-boot.txt
Alix2D13-22-upgrade-second-boot.txt

eri-- · Apr 8, 2014, 10:01 AM

Can you show a pciconf -lv
Also did you try to boot without acpi or ahci?

GruensFroeschli · Apr 8, 2014, 10:11 AM

Could be the same as this:
https://forum.pfsense.org/index.php?topic=74571.msg407448#msg407448

phil.davis · Apr 8, 2014, 11:12 AM

@ermal:

Can you show a pciconf -lv
Also did you try to boot without acpi or ahci?

$ pciconf -lv
hostb0@pci0:0:1:0:	class=0x060000 card=0x20801022 chip=0x20801022 rev=0x33 hdr=0x00
    class      = bridge
    subclass   = HOST-PCI
none0@pci0:0:1:2:	class=0x101000 card=0x20821022 chip=0x20821022 rev=0x00 hdr=0x00
    class      = encrypt/decrypt
vr0@pci0:0:9:0:	class=0x020000 card=0x01061106 chip=0x30531106 rev=0x96 hdr=0x00
    class      = network
    subclass   = ethernet
vr1@pci0:0:10:0:	class=0x020000 card=0x01061106 chip=0x30531106 rev=0x96 hdr=0x00
    class      = network
    subclass   = ethernet
vr2@pci0:0:11:0:	class=0x020000 card=0x01061106 chip=0x30531106 rev=0x96 hdr=0x00
    class      = network
    subclass   = ethernet
none1@pci0:0:12:0:	class=0x020000 card=0x1012185f chip=0x0013168c rev=0x01 hdr=0x00
    class      = network
    subclass   = ethernet
isab0@pci0:0:15:0:	class=0x060100 card=0x20901022 chip=0x20901022 rev=0x03 hdr=0x00
    class      = bridge
    subclass   = PCI-ISA
atapci0@pci0:0:15:2:	class=0x010180 card=0x209a1022 chip=0x209a1022 rev=0x01 hdr=0x00
    class      = mass storage
    subclass   = ATA
ohci0@pci0:0:15:4:	class=0x0c0310 card=0x20941022 chip=0x20941022 rev=0x02 hdr=0x00
    class      = serial bus
    subclass   = USB
ehci0@pci0:0:15:5:	class=0x0c0320 card=0x20951022 chip=0x20951022 rev=0x02 hdr=0x00
    class      = serial bus
    subclass   = USB

I haven't tried any different boot options. I simply upgraded from 2.1.1-RELEASE to 2.2. The ath0 wireless was working fine on 2.1.1-RELEASE.
I am surprised these things are not upwardly compatible and I expected that any previously-working not-too-old hardware and its driver config from 8.3 would "just work the same" on a later FreeBSD 10.
Let me know anything you would like me to do or show…

eri-- · Apr 8, 2014, 3:38 PM

Try tomorrow snapshots please.

phil.davis · Apr 8, 2014, 5:14 PM

Yes, will report back tomorrow.

Just a comment about this process now. We cannot see anything of what is happening to pfSense-tools repo. We are flying blind. Yes, that is the way it is with Microsoft products, we try things and report apparent bugs and some time later a "magic fix" (might) appear. But this is an open source project!

I used to debug OpenVMS device drivers. I do have some skills and if I could look at what is in pfSense-tools then I could help spot what some of the issues are. Sure, Ermal can do this stuff himself behind the scenes, but IMHO it is useful to have other sets of eyes to help in the debugging process.

And then people will quickly assume that some security agency is working with ESF to insert "bonus code" into pfSense. Actually I don't think that will be the case, but unless people can see the pfSense mods made to "pf" and other standard FreeBSD components, then they will naturally start to think these things.

Come on ESF, sort out a way for community members to see the code.

eri-- · Apr 9, 2014, 10:49 AM

This was mostly related to kernel config handling.

The tools repo issues is being handled slowly but being handled.

charliem · Apr 9, 2014, 1:08 PM

@phil.davis:

Just a comment about this process now. We cannot see anything of what is happening to pfSense-tools repo. We are flying blind. Yes, that is the way it is with Microsoft products, we try things and report apparent bugs and some time later a "magic fix" (might) appear. But this is an open source project!

I used to debug OpenVMS device drivers. I do have some skills and if I could look at what is in pfSense-tools then I could help spot what some of the issues are. Sure, Ermal can do this stuff himself behind the scenes, but IMHO it is useful to have other sets of eyes to help in the debugging process.

And then people will quickly assume that some security agency is working with ESF to insert "bonus code" into pfSense. Actually I don't think that will be the case, but unless people can see the pfSense mods made to "pf" and other standard FreeBSD components, then they will naturally start to think these things.

Come on ESF, sort out a way for community members to see the code.

Phil, these comments distill my concerns exactly, too bad it's buried in this thread (not that these comments make any difference apparently).

A less nefarious concern than three letter agencies and 'bonus code' is simply documentation. When a newbie like me say, tries to dig into a problem using available freebsd documentation, it's very frustrating to find that pfsense contains mods that are not reflected in the docs. Often pfsense relies on these changes, yet they are undocumented. I'm thinking of my on-going issues with state killing in 2.1 (pfctl mods) and of course the new tracker numbers.

I recall a post where someone said in effect ' … if we wanted to close it, we took away the wrong repo', referring to removing the pfsense-tools repo rather than pfsense repo. Well, IMHO pfsense is effectively closed until the tools repo is available again.

eri-- · Apr 9, 2014, 1:33 PM

Please send this discussion on its own thread not related to this.

This is about ath0.

phil.davis · Apr 10, 2014, 12:29 AM

I will be a couple of days before being able to try this. I am away and something happened at home - the home laptop is not getting out through the setup I left behind so I can't TeamViewer in to it. And OpenVPN client is not working on that pfSense 2.2 box, so it did not make a site-to-site connection out to the intended place, so I can't get back to it over that. Bit hard for the wife to troubleshoot - she is getting internet from a WiFi ADSL device until I get back.

bman212121 · Apr 11, 2014, 1:03 PM

Just want to throw it out there, if anyone does get their ath0 wireless N card working, it would be handy to know any steps you did to set it up. I went to test my atheros 5416 chipset card (Dlink DWA-552) and am happy to report it no longer causes a kernel panic. (anything FreeBSD 8 and it's not happy) I set it up as the WAN interface on a test box and it automatically connected itself to an open network, got an IP and appears to have internet. I wasn't able to get it to connect to a secured network however and it didn't look like the rescan option worked as it couldn't find any networks. Haven't had time to investigate why yet.

This was done on the April 7th snapshot..

phil.davis · Apr 12, 2014, 8:20 AM

2.2-ALPHA (i386)
built on Wed Apr 9 09:01:59 CDT 2014
FreeBSD 10.0-STABLE

Still no "ath0" available in Interfaces->Assign on my Alix 2D13.

phil.davis · Apr 14, 2014, 5:16 AM

2.2-ALPHA (i386)
built on Sun Apr 13 15:35:58 CDT 2014
FreeBSD 10.0-STABLE

ath0 is back on Alix 2D13! Very good. I have set it up in Access Point mode, enabled DHCP and added firewall rules and it is working as expected. I am posting across that now.

Raul Ramos · Apr 14, 2014, 7:25 AM

Hi good morning,

@phil.davis, have you support to N Wireless?

Thanks

phil.davis · Apr 14, 2014, 8:57 AM

I don't know about that. I think the card in my Alix 2D13 is only b/g

GruensFroeschli · Apr 14, 2014, 9:12 AM

I tested an N card and it's not recognised anymore.
It was recognised in an earlier version.

Is still shown as:
none2@pci0:4:0:0: class=0x028000 card=0x3123168c chip=0x0033168c rev=0x01 hdr=0x00
class = network

eri-- · Apr 14, 2014, 10:46 AM

Can you tell me which vrsion it was recognized before?

Its a PCI card and nothing should have changed there afaik.

I see it even supported by ath AR9300_DEVID_AR9580_PCIE

GruensFroeschli · Apr 14, 2014, 11:49 AM

The march 27 version.
It seems these files aren't on the snapshot server anymore.

PhilipG · Apr 14, 2014, 11:57 AM

A March 27th image was the last time I saw if_cdce.ko; I suspect many modules have been missing since then.

eri-- · Apr 14, 2014, 1:08 PM

Those are inside the kernel nowdays.

And if_cdce iirc is for USB interfaces while this is a PCIe one.
Judging on the other seeing the ath interface for assignment this is a bit strange happening.

Though can you try with the next set that will come out?