Bug 1087838

Description of problem:
netgroup caching in nscd has several problems that have been fixed upstream. For example, one is caused by improper usage of extend_alloca, causing nscd to segfault if a long netgroup entry is added to /etc/netgroup. If it doesn't segfault, the improperly extended buffer contains garbage or is missing part of the netgroup data, causing all kinds of weird behaviour (e.g. group membership not being seen by sudo and others).

This has been fixed upstream with this commit: https://sourceware.org/git/?p=glibc.git;a=commitdiff_plain;h=980cb5180e1b71224a57ca52b995c959b7148c09

Version-Release number of selected component (if applicable):
nscd-2.12-1.132.el6

How reproducible:
always

Steps to Reproduce:
1. add long entry (>1024 bytes) to /etc/netgroup
2. enable netgroup caching in nscd
3. clear netgroup cache
4. getent netgroup longnetgroupentry

Actual results:
nscd segfaults or returns incomplete netgroup entry

[root@de-do-rzcfx211 ~]# rm -f /var/db/nscd/netgroup ; nscd -d
Tue 15 Apr 2014 02:33:31 PM CEST - 23139: handle_request: request received (Version = 2) from PID 23149
Tue 15 Apr 2014 02:33:31 PM CEST - 23139:       GETFDNETGR
Tue 15 Apr 2014 02:33:31 PM CEST - 23139: provide access to FD 13, for netgroup
Tue 15 Apr 2014 02:33:31 PM CEST - 23139: handle_request: request received (Version = 2) from PID 23149
Tue 15 Apr 2014 02:33:31 PM CEST - 23139:       GETNETGRENT (test)
Tue 15 Apr 2014 02:33:31 PM CEST - 23139: Haven't found "test" in netgroup cache!
Segmentation fault

Expected results:
nscd doesn't segfault and returns complete netgroup entry

Workaround:
Disable netgroup caching in /etc/nscd.conf.