- Freeradius-Users - lists.freeradius.org

EAP-TLS and IOS 13
by Ted Hyde (RSI) 04 Nov '19

04 Nov '19

I love *pple. And by love I mean exactly the opposite.....Regardless, my many thanks to all that assisted with my tribulations and blatherings regarding getting ipads and Win10 machines working with an EAP-TLS environment. As of this morning I had everything migrated, wiped, re-tested, full bare-metal automation tested and ready to deploy to the minions. I had a very happy moment. Until someone walked in with an ipad that they just upgraded to IOS 13. tl/dr: IOS13 introduces more stringent compliance for certificates (https://support.apple.com/en-us/HT210176) and that means certificates that used to work for EAP, now do not install - well that's not true. They install, they say they're verified, but the ipad does not recognize them as useful, and ONLY presents a TTLS-like connection interface (username and password, instead of certificate and identity). They simply sit there all happy and useless. (BTW, manual cert install is now an 8 page document in my library, including download, allow,accept, enable Cert Trust Settings, install, validate and....then watch do nothing.) Since the ipad does not present a tls transaction, FR3 doesn't participate. I am not using EAP-TTLS, so that module does exactly what is it supposed to do - find no verified username and reject. I've used my google-fu to get the basic idea of modifying the openssl commands to include the EKU, and sha2, but some of the other requirements I'm not sure about implementing. The "no longer than 2 years" is also a PITA. Either way, has anyone worked out a magic bullet for this yet? Amazingly, M$ is no longer on my hated list - the Win10 machines are now in the "it simply works" category! Longing to learn from the masters, yet again! Thanks, Ted. On 11/1/2019 6:56 PM, freeradius-users-request(a)lists.freeradius.org wrote: > Message: 6 > Date: Fri, 1 Nov 2019 16:22:30 -0400 > From: "Ted Hyde (RSI)"<thyde(a)rndstudio.com> > To:freeradius-users@lists.freeradius.org > Subject: Migrating FR3 instance --

3 2

Re:EAP-TLS and IOS 13 (Updated)
by Ted Hyde (RSI) 04 Nov '19

04 Nov '19

Quick update - IOS versioning strikes again; my *pple guru who lives bleeding-edge had a 13.2 pre-release in his kit and what struck me strange was that the Mode option in Wifi Security settings was available if using the "Other..." wifi setup option, wherein my test ipad (a 13.1.2 version) it was not. Short story shorter, we did get ios13 working again, but only on 13.2. any prior version newer than 12.3.2 (June?) has the missing EAP-TLS syndrome on our ipads. Unfortunately, looking at my MDM, a lot of my fleet is exactly in that range. Is it OS update season again already? This doesn't make the certificates compatible with *pple's newly released constraints; I'm still interested in available knowledge where it may exist. Thanks, Ted.

2 1

Granting/denying access according auth-method
by dump＠gmx.info 02 Nov '19

02 Nov '19

Dear list, I'm using freeradius 3 almost in standard config with mysql as storing facility. I have PEAP an EAP active. Usually EAP via TLS or TTLS/PAP is used, because the passphrases could be stored in SHA-1 encryption. PEAP is sometimes necessary for me too but it uses MS-CHAPv2 which needs cleartext-PW. Due to this I would like to restrict authentication vis MS-CHAPv2 and reactivate it when needed. I would prefer to do it via the radcheck-table for quickly activating and deactivate is easily. I read that it's possible via users-file, but I can't find any useful information/manual to me what to fill into the radcheck tables. May somebody have the idea or manual I need to follow/read? Many thanks in advance and best regards Jens

2 1

Migrating FR3 instance
by Ted Hyde (RSI) 01 Nov '19

01 Nov '19

Greets - I would like to migrate a test platform I have in a VM (FR3 3.0.12 on Debian 9) to a bare-metal deployment of same, but keep the client certificates as used in the VM instance. The bare-metal version already has other things installed, including a clean copy of FR3, so it's not as easy as simply imaging the appliance back to bare-metal. I am using eap-tls auth for a NAS, and wireless clients already have both the ca certificate and client certificates installed on them and functional. Given that this is a bare-metal install target that still has the snakeoil/testing certificates installed, is there a preferred/working method to copy the existing certs across and not destroy the entire system in the process? I expect I will have to copy the .cnf files for some items, but how does the ca structure (since I *think* I am also using CA validation in my eap module) get copied over? Thanks, Ted.

2 1

What does 'rlm' stand for in FreeRADIUS modules?
by Kamran Khan 01 Nov '19

01 Nov '19

i.e., what does 'rlm' mean in modules like rlm_perl, rlm_python etc. Thanks, Kamran.

2 2

Variable passing question between modules
by Adam Taylor 01 Nov '19

01 Nov '19

Hey all, I have been trying for a week and I have everything working on a setup I am doing except for one thing. Setting Tmp-String-0 to a string and using that string in the LDAP module. I'm trying to make the config as simple as possible so if I am doing something wrong, please let me know. I have the following in my authorize section: authorize { if (User-Name =~ /(a)ulm\.edu$$/i){ update request { Tmp-String-0 = "Employees" } preprocess mschap eap { ok = return } ldap } if (User-Name =~ /(a)warhawks\.ulm\.edu$$/i) { update request { Tmp-String-0 = "Students" } preprocess mschap eap { ok = return } ldap } } Basically if the login is XXX(a)ulm.edu<mailto:XXX@ulm.edu> set TMP-String-0 to "Employees" or if XXX(a)warhawks.ulm.edu<mailto:XXX@warhawks.ulm.edu> set to "Students" I then have the following in my ldap config: base_dn = 'ou=%{request:Tmp-String-0},ou=People,dc=ulm,dc=edu' The problem is ${request:Tmp-String-0} is always null and I get this in debug: (6) ldap: EXPAND ou=%{request:Tmp-String-0},ou=People,dc=ulm,dc=edu (6) ldap: --> ou=,ou=People,dc=ulm,dc=edu (6) ldap: Performing search in "ou=,ou=People,dc=ulm,dc=edu" with filter "(uid=pp)", scope "sub" I went to verify that it was actually setting the variable and see this: (6) update request { (6) Tmp-String-0 = "Students" (6) } # update request = noop I have tried every different way I can think of and it just will not set Tmp-String-0. I have tried both = and := as well. What am I doing wrong? Thanks, Adam Taylor

3 4

RE: OCSP Response Signed by other CA
by Mike DiBella 31 Oct '19

31 Oct '19

In testing my OCSP responder configuration I've found that the following openssl command: openssl ocsp -issuer client_cer_root.cer -cert client.cer -VAfile ocsp_signer.cer -url http://ocsp.responder.net/ocsp returns: Response verify OK client.cer: good This Update: Oct 31 11:43:47 2019 GMT Next Update: Nov 2 00:03:47 2019 GMT I see that there is a client keyword in the verify section of the eap configuration file. I'm wondering if I can I substitute this command for the default openssl verify command prototyped in the file? In my testing, it looks like openssl returns exit 0 for both verified and responder errors, so I'm wondering how freeradius parses the result of the command. From: Mike DiBella Sent: Wednesday, October 30, 2019 3:08 PM To: 'freeradius-users(a)lists.freeradius.org' <freeradius-users(a)lists.freeradius.org> Subject: OCSP Response Signed by other CA I'm attempting to add OCSP certificate verification to EAP, but my responder hosts multiple CAs and signs the responses with a cert common to all. I think freeradius is assuming that the signer of the OCSP response will be the same as the signer of the client certificate, and I can't see an option to override this behavior. I've already tried concatenating the OCSP signer to the client trust certificate file, but freeradius still fails validation with "Error: rlm_eap: SSL error error:27069070:OCSP routines:OCSP_basic_verify:root ca not trusted". Any ideas?

2 1

802.1x / EAP Assistance
by J Kephart 31 Oct '19

31 Oct '19

Good morning! We are attempting to implement 802.1x/EAP for the first time, ad we're having some trouble diagnosing what's going on in the various stages of the communications between the NAS and FR. We don't have any experience with it, so it's rather confusing. We're using FR 2.2.8, with the test certs provided. We can see that there is communication, but all attempts to authenticate a device are failing. I've included what I believe to be the relevant portion of the debug output, and I do see several error conditions. The first says that the realm LOCAL is not defined, but in looking at the config, it looks as though it is. There's also a report that there's a missing Cleartext-Password, but that is also defined in the database, so we're at a loss as to the cause of the failure. If someone can point us in the right direction, I'd truly appreciate it! rad_recv: Access-Request packet from host 146.115.19.180 port 50987, id=97, length=394 Acct-Session-Id = "5DB9E1F5-76DF7502" User-Name = "jerry" NAS-IP-Address = 192.168.185.30 NAS-Identifier = "90-3A-72-15-25-1D" NAS-Port = 1 Called-Station-Id = "90-3A-72-15-25-1D:CVGNE_W1" Calling-Station-Id = "D4-53-83-F3-C0-17" Service-Type = Framed-User Chargeable-User-Identity = "" NAS-Port-Type = Wireless-802.11 Connect-Info = "CONNECT 802.11a/n/ac" EAP-Message = 0x0208005f1900170303005400000000000000022f4cd34458205cfb0d339d2a9f6dda68b5f7a4ffbd985bbfb9ef4094114e5a1856df8479ea3c4ddc7f293487a00396643c97c4b24f51d67a7c7cc34e9bbbc1c156adc07977d4e095d6fddaa5 State = 0x957bdc849273c5bfd96a2d7f79095e2d Ruckus-SSID = "CVGNE_W1" Ruckus-Attr-14 = 0x903a7215251d Ruckus-Attr-9 = 0x000001c3 Ruckus-SCG-CBlade-IP = 3232282885 Ruckus-Attr-134 = 0x44656661756c74205a6f6e65 Ruckus-Attr-135 = 0x4356474e455f5731 Message-Authenticator = 0xad0239941dd1cf346629d1f1b23805b6 Event-Timestamp = "Oct 30 2019 15:18:14 EDT" Proxy-State = 0x3636 # Executing section authorize from file /etc/raddb/sites-enabled/default +group authorize { ++[preprocess] = ok ++update request { sql_xlat expand: select zone_migration_enabled from sites where id='%{NAS-Identifier}' -> select zone_migration_enabled from sites where id='90-3A-72-15-25-1D' rlm_sql (sql_instance2): Reserving sql socket id: 2 SQL query did not return any results rlm_sql (sql_instance2): Released sql socket id: 2 expand: %{sql_instance2: select zone_migration_enabled from sites where id='%{NAS-Identifier}'} -> ... expanding second conditional expand: %{%{sql_instance2: select zone_migration_enabled from sites where id='%{NAS-Identifier}'}:-0} -> 0 ++} # update request = noop ++? if (("%{Called-Station-Id}" =~ /^00-50-E8-/ || "%{Called-Station-Id}" =~ /^20-4C-03-/ )&& Tmp-String-2 == '1') expand: %{Called-Station-Id} -> 90-3A-72-15-25-1D:CVGNE_W1 ?? Evaluating ("%{Called-Station-Id}" =~ /^00-50-E8-/) -> FALSE expand: %{Called-Station-Id} -> 90-3A-72-15-25-1D:CVGNE_W1 ?? Evaluating ("%{Called-Station-Id}" =~ /^20-4C-03-/) -> FALSE ? Skipping (Tmp-String-2 == '1') ++? if (("%{Called-Station-Id}" =~ /^00-50-E8-/ || "%{Called-Station-Id}" =~ /^20-4C-03-/ )&& Tmp-String-2 == '1') -> FALSE ++[chap] = noop ++[mschap] = noop [suffix] No '@' in User-Name = "jerry", looking up realm NULL [suffix] No such realm "NULL" ++[suffix] = noop [eap] EAP packet type response id 8 length 95 [eap] Continuing tunnel setup. ++[eap] = ok +} # group authorize = ok Found Auth-Type = EAP # Executing group from file /etc/raddb/sites-enabled/default +group authenticate { [eap] Request found, released from the list [eap] EAP/peap [eap] processing type peap [peap] processing EAP-TLS [peap] eaptls_verify returned 7 [peap] Done initial handshake [peap] eaptls_process returned 7 [peap] EAPTLS_OK [peap] Session established. Decoding tunneled attributes. [peap] Peap state phase2 [peap] EAP type mschapv2 [peap] Got tunneled request EAP-Message = 0x020800401a0208003b31197eb5bbc69b47c5363805196ff71ff700000000000000008a49a9ade98c47831f6c39043311438b1a56945caec5a5f5006a65727279 server { [peap] Setting User-Name to jerry Sending tunneled request EAP-Message = 0x020800401a0208003b31197eb5bbc69b47c5363805196ff71ff700000000000000008a49a9ade98c47831f6c39043311438b1a56945caec5a5f5006a65727279 FreeRADIUS-Proxied-To = 127.0.0.1 User-Name = "jerry" State = 0x2b16a6c52b1ebc1c6619c5138158b3f7 server inner-tunnel { # Executing section authorize from file /etc/raddb/sites-enabled/inner-tunnel +group authorize { ++[chap] = noop ++[mschap] = noop [suffix] No '@' in User-Name = "jerry", looking up realm NULL [suffix] No such realm "NULL" ++[suffix] = noop ++update control { ++} # update control = noop [eap] EAP packet type response id 8 length 64 [eap] No EAP Start, assuming it's an on-going EAP conversation ++[eap] = updated ++[files] = noop ++[expiration] = noop ++[logintime] = noop ++[pap] = noop +} # group authorize = updated WARNING: You set Proxy-To-Realm = LOCAL, but the realm does not exist! Cancelling invalid proxy request. Found Auth-Type = EAP # Executing group from file /etc/raddb/sites-enabled/inner-tunnel +group authenticate { [eap] Request found, released from the list [eap] EAP/mschapv2 [eap] processing type mschapv2 [mschapv2] # Executing group from file /etc/raddb/sites-enabled/inner-tunnel [mschapv2] +group MS-CHAP { [mschap] No Cleartext-Password configured. Cannot create LM-Password. [mschap] No Cleartext-Password configured. Cannot create NT-Password. [mschap] Creating challenge hash with username: jerry [mschap] Client is using MS-CHAPv2 for jerry, we need NT-Password [mschap] FAILED: No NT/LM-Password. Cannot perform authentication. [mschap] FAILED: MS-CHAP2-Response is incorrect ++[mschap] = reject +} # group MS-CHAP = reject [eap] Freeing handler ++[eap] = reject +} # group authenticate = reject Failed to authenticate the user. expand: %{NAS-IP-Address} -> Login incorrect: [jerry/<via Auth-Type = EAP>] (from client Office port 0 via TLS tunnel) Using Post-Auth-Type REJECT # Executing group from file /etc/raddb/sites-enabled/inner-tunnel +group REJECT { [attr_filter.access_reject] expand: %{User-Name} -> jerry attr_filter: Matched entry DEFAULT at line 11 ++[attr_filter.access_reject] = updated +} # group REJECT = updated } # server inner-tunnel [peap] Got tunneled reply code 3 MS-CHAP-Error = "\010E=691 R=1" EAP-Message = 0x04080004 Message-Authenticator = 0x00000000000000000000000000000000 [peap] Got tunneled reply RADIUS code Access-Reject MS-CHAP-Error = "\010E=691 R=1" EAP-Message = 0x04080004 Message-Authenticator = 0x00000000000000000000000000000000 [peap] Tunneled authentication was rejected. [peap] FAILURE ++[eap] = handled +} # group authenticate = handled Sending Access-Challenge of id 97 to 146.115.19.180 port 50987 EAP-Message = 0x0109002e1900170303002334c44a5ec04ae1317ed894226db68ba139a09ceb0517c705c20507823d7bb4bc4a05c6 Message-Authenticator = 0x00000000000000000000000000000000 State = 0x957bdc849d72c5bfd96a2d7f79095e2d Proxy-State = 0x3636 Finished request 214. Many thanks! -- Jim

3 6

proxy.conf from mysql database?
by J E H A N Z A I B 31 Oct '19

31 Oct '19

Hi experts, I am currently using proxy.conf to list all the realms. is there anyway i can fetch these realms from mysql database? My proxy.conf realm xx.com { authhost = xx.xx.xx.xx:1812 accthost = xx.xx.xx.xx :1813 secret = xxsecretxx fallback = LOCAL nostrip } -- Regards, Jehanzaib

2 1

Some RLM_MODULE_INVALID events are not logged via detail
by Boris Lytochkin 31 Oct '19

31 Oct '19

Hi. We're trying to catch a bug in some vendor's equipment resulting in "Login incorrect (eap: EAP requires the State attribute to work, but no State exists in the Access-Request packet.)" error appearing in the log and subsequent Access-Reject sent from RADIUS server (version 3.0.15 with a bit of pull requests still not merged :). It seems that this configuration does not catch that particular Access-Reject into detail(ed) log: ================== detail auth_log { header = "%t (%I)" filename = ${radacctdir}/%{%{Packet-Src-IP-Address}:-%{Packet-Src-IPv6-Address}}/auth-%Y%m%d log_packet_header = yes permissions = 0640 } authenticate { Auth-Type EAP { eap { handled = 9999 } if (handled) { auth_log.post-auth # logging is done, return return } # eap module returned OK so we go a bit further perl # Access-Accept/Reject will be logged by authorize section } } post-auth { auth_log Post-Auth-Type REJECT { auth_log } ... } ================== Am I missing something? I took a quick tour though the code and failed to find a place where radiusd decides to **log** " Login incorrect (eap_tls: TLS Alert read:fatal:unknown CA):" into detail and **not to log** "Login incorrect (eap: EAP requires the State attribute to work, but no State exists in the Access-Request packet.)" p/s. Playing around with `handled` in Auth-Type EAP has nothing to do with detailed log as I see the same situation before I tweaked that part of the configuration. pp/s. Is there a way to print packet identifier as it is sent over the wire into detailed log? I made a trivial patch for that seeing no documented way exist to do that: =============== --- a/src/modules/rlm_detail/rlm_detail.c +++ b/src/modules/rlm_detail/rlm_detail.c @@ -336,6 +336,7 @@ static int detail_write(FILE *out, rlm_detail_t *inst, REQUEST *request, RADIUS_ } #endif } + WRITE("\tPacket-Identifier = %d\n", packet->id); WRITE("\tTimestamp = %ld\n", (unsigned long) request->timestamp); WRITE("\n"); =============== -- Boris Lytochkin Yandex NOC +7 (495) 739 70 00 ext. 7671

2 6