Bug 1379480

Summary: Segfault when failover pool contains virtual and real home servers
Product: Red Hat Enterprise Linux 7 Reporter: Mike Ely <me>
Component: freeradiusAssignee: Nikolai Kondrashov <nikolai.kondrashov>
Status: CLOSED WORKSFORME QA Contact: BaseOS QE Security Team <qe-baseos-security>
Severity: high Docs Contact:
Priority: high    
Version: 7.2CC: dpal, ksrot, lmiksik, me, pkis
Target Milestone: rc   
Target Release: ---   
Hardware: x86_64   
OS: Unspecified   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-03-27 12:50:06 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Attachments:
Description Flags
Patch allows check to continue if virtual host listed.
none
Associated patch to .spec file
none
Patch allows check to continue if virtual host listed. (updated) none

Description Mike Ely 2016-09-26 23:26:47 UTC
Created attachment 1204992 [details]
Patch allows check to continue if virtual host listed.

Please see upstream thread here:
http://lists.freeradius.org/pipermail/freeradius-users/2016-September/085048.html

As well as resultant upstream commit here:
https://github.com/FreeRADIUS/freeradius-server/commit/c1cef7878ed2d196be1e10c9e364d031b006f65a

When configured with both virtual and real home servers, freeradius will segfault if an authentication request passes across the virtual server. This has been fixed upstream.

I'm requesting the attached patch be reviewed and added.

Thanks!

Comment 1 Mike Ely 2016-09-26 23:31:32 UTC
Created attachment 1204993 [details]
Associated patch to .spec file

Comment 3 Nikolai Kondrashov 2016-09-27 10:05:04 UTC
Thank you very much for the report, Mike. I'll see where and when we can integrate the fix.

Comment 4 Nikolai Kondrashov 2016-09-28 10:36:13 UTC
We'll work on including the fix into the next Z-stream release of RHEL7.

Comment 5 Mike Ely 2016-09-28 22:42:21 UTC
Any chance of it going in as a regular 7.2 update?

Comment 6 Nikolai Kondrashov 2016-09-29 13:26:30 UTC
Yes, we will work on getting it into a 7.2.z release too.

Comment 7 Mike Ely 2016-09-30 00:01:19 UTC
Created attachment 1206075 [details]
Patch allows check to continue if virtual host listed. (updated)

Developer added an "oh also" change that I missed in the previous patch:
https://github.com/FreeRADIUS/freeradius-server/commit/70a76e9c1b69350cc0f8436f1dbec12a1c1ac9b0

Comment 8 Mike Ely 2016-09-30 00:02:09 UTC
Fixed the patch.

Was actually asking if this change would make it into the regular (non-Z) stream.

Comment 9 Nikolai Kondrashov 2016-09-30 11:35:40 UTC
Thanks a lot on the patch update and following the developers on this, Mike.

I cannot promise an asynchronous update, unfortunately. You will have to reach support on this as a customer, explicitly.

Comment 11 Mike Ely 2016-09-30 18:14:19 UTC
Please hold work on this. The proposed fix does not work:
#0  0x00007f48dfea1416 in check_for_realm (instance=<optimized out>, request=0x7f48e7870830, 
    returnrealm=0x7f48dc18d580) at src/modules/rlm_realm/rlm_realm.c:276
276                             if (realm->acct_pool->servers[i]->ipaddr.af == AF_UNSPEC) continue;
(gdb) bt
#0  0x00007f48dfea1416 in check_for_realm (instance=<optimized out>, request=0x7f48e7870830, 
    returnrealm=0x7f48dc18d580) at src/modules/rlm_realm/rlm_realm.c:276

Comment 19 Nikolai Kondrashov 2017-02-15 15:55:54 UTC
Hi Mike, do you have any update on this issue? Have you discovered the source, a fix?

Comment 20 Mike Ely 2017-02-15 17:50:35 UTC
None at this time. I'd love it if Alan DeKok or one of his leads would participate here as the issue is pretty deep within Freeradius and they'd be best positioned to help.

Comment 21 Nikolai Kondrashov 2017-02-16 07:27:13 UTC
I see. Before we can fix this, we'll need a way to reproduce this, preferably automatic one, or at least exact instructions. Also, please reach support on this one, if you're a customer.

Comment 22 Nikolai Kondrashov 2017-02-22 13:17:11 UTC
Mike, please note that if we don't get a way to reproduce this, we won't be able to investigate it, and thus get a fix for this into next RHEL release.

Comment 23 Nikolai Kondrashov 2017-03-27 12:50:06 UTC
Closing as WORKSFORME as we have no reproducer. Please feel free to reopen, once you have one.