Affected versions: IBM AIX 7.1

📖 ~4 min read  •  Source: NVD CVE-2006-0674, IBM Support Bulletin

CVE: CVE-2006-0674

NVD summary: Buffer overflow in the arp command of IBM AIX 5.3 L, 5.3, 5.2.2, 5.2 L, and 5.2 allows local users to cause a denial of service (crash) via a long iftype argument.

References: secunia.com/advisories/18773   www-1.ibm.com/support/docview.wss?rs=0&q1=IY8142   www-1.ibm.com/support/docview.wss?uid=isg1IY8147

Table of contents
  1. Symptom & Impact
  2. Environment & Reproduction
  3. Root Cause Analysis
  4. Quick Triage
  5. Step-by-Step Diagnosis
  6. Solution – Primary Fix
  7. Solution – Alternative Approaches
  8. Verification & Acceptance Criteria
  9. Rollback Plan
  10. Prevention & Hardening
  11. Related Errors & Cross-Refs
  12. References & Further Reading

Symptom & Impact

On IBM AIX 7.1 (IBM Power Systems), administrators encountering CVE-2006-0674 observe behaviour consistent with the NVD CVE-2006-0674 entry: unexpected service restarts logged through errpt, fileset integrity warnings from lppchk -v, and — for security-rated advisories — exposure to the documented threat scenario. Impact ranges from per-LPAR availability events to broader workload disruption when the affected component sits on the serving path of PowerHA, WPAR, or VIO Server roles.

Environment & Reproduction

Reproduction targets IBM AIX 7.1. Confirm release, fix level, and the installed component:

oslevel -s                          # full TL-SP-build (e.g. 7300-02-02-2420)
oslevel -r                          # recommended ML / TL
instfix -i | grep -iE 'TL|SP' | tail -20
lslpp -L | head -30                 # installed filesets
uname -aMF                          # kernel / machine / firmware
prtconf | head -40                  # hardware + LPAR config

Trigger the workflow that exposes CVE-2006-0674 — buffer overflow — patch and remediation guide while collecting:

errpt -a | head -200                # decoded error log
errpt -a -s $(date +%m%d%H%M%y) > /tmp/errpt-CVE-2006-0674.txt
alog -o -t console | tail -200      # console log
alog -o -t boot | tail -200         # boot log
tail -200 /var/adm/messages 2>/dev/null   # only if syslog routes here
# Full IBM Support evidence bundle:
snap -ac                            # writes /tmp/ibmsupt/snap.pax.Z

Root Cause Analysis

Root cause is tracked at NVD CVE-2006-0674 and the corresponding IBM Support Bulletin. IBM ships fixes as APARs delivered via Technology Levels (TL) and Service Packs (SP) for IBM AIX 7.1; running an outdated TL/SP leaves the host exposed to the failure modes referenced above. Correlate fileset and APAR state with the error log:

instfix -ivk <APAR>                 # is the APAR fix installed?
lslpp -h <fileset>                  # install/update history for a fileset
lslpp -p <fileset>                  # prerequisites
emgr -l                             # interim fixes (ifixes) currently applied
errpt -a | grep -iE 'fail|crash|panic' | head
sysdumpdev -l                       # dump device config (kernel fixes may change it)

Quick Triage

Run these on IBM AIX 7.1 to capture current state related to CVE-2006-0674:

oslevel -s && oslevel -r
instfix -i | grep -iE '<APAR>|TL|SP'
errpt | head -50
lppchk -v                           # fileset version consistency
lppchk -c                           # checksum filesets against ODM
lssrc -a | grep -v active           # subsystems not active
lsdev -C | grep -i defined          # devices in Defined state (should be Available)
genld -l 2>/dev/null | head -20     # shared library load map (leaks)

Step-by-Step Diagnosis

  1. Capture exact TL/SP and APAR status.

    oslevel -s
    oslevel -r
    instfix -ivk <APAR>
  2. Pull error log entries for the impacted window (format: mmddhhmmyy).

    errpt -a -s $(date -d '-1 hour' +%m%d%H%M%y 2>/dev/null || date +%m%d%H%M%y)
  3. Verify fileset integrity and reinstall if anything is inconsistent.

    lppchk -v
    lppchk -c
    # Re-apply a damaged fileset from media:
    installp -acgXYd /usr/sys/inst.images <fileset>
  4. Inspect SRC-managed subsystems and groups.

    lssrc -a | head
    lssrc -ls <subsystem>             # detailed state
    stopsrc -s <subsystem> && startsrc -s <subsystem>
    refresh -s <subsystem>            # re-read config without restart
  5. Stage a snap for IBM Support.

    snap -ac                          # bundles diag info to /tmp/ibmsupt
    ls -la /tmp/ibmsupt/
  6. Correlate findings with NVD CVE-2006-0674, IBM Support Bulletin, and the APAR fix list to pin the change that introduced CVE-2006-0674 — buffer overflow — patch and remediation guide.

Solution – Primary Fix

Apply the corrective Technology Level / Service Pack (or individual APAR fix) for IBM AIX 7.1 from IBM Fix Central. Recommended workflow uses NIM or local staging:

# 1. Download the fix bundle from https://www.ibm.com/support/fixcentral
# 2. Stage to a NIM master (or local /usr/sys/inst.images for standalone).
# 3. Take rollback safety net BEFORE applying:
alt_disk_copy -B -d hdisk1          # clone rootvg to a spare disk
lspv                                # confirm altinst_rootvg exists
# 4. Apply non-interactively from media path:
installp -acgXYd /usr/sys/inst.images all
# Or apply only the specific fileset:
installp -acgXYd /usr/sys/inst.images <fileset>
# Or use the smit menu (interactive):
smitty update_all
# Or via NIM from a master server:
# nim -o cust -a lpp_source=<src> -a fixes=update_all <target>
# 5. Reboot if BOS / kernel filesets were updated:
shutdown -Fr now
# 6. Verify on come-up:
oslevel -s
instfix -ivk <APAR>
lppchk -v

For an emergency single-CVE fix without a full SP, apply the IBM-supplied interim fix (ifix):

# Download the .epkg.Z ifix package matching this CVE/APAR:
emgr -p -e /tmp/ifix.epkg.Z         # preview only (no install)
emgr -X -e /tmp/ifix.epkg.Z         # install ifix (and force prereq override if listed)
emgr -l                             # confirm the ifix label is listed

Need help rolling this fix across an AIX fleet? Our IT Solutions & Services team manages AIX TL/SP rollouts with NIM, alt_disk_copy and mksysb rollback. Get in touch for a free consultation.

Solution – Alternative Approaches

If the primary TL/SP campaign is not viable, choose from these:

  • Apply only the interim fix (ifix) ahead of the next SP window:

    emgr -X -e /tmp/ifix.epkg.Z
    emgr -l
    # To remove later (by label, shown in emgr -l):
    emgr -r -L <ifix-label>
  • Isolate impacted workloads in a WPAR with restricted resources:

    lswpar
    mkwpar -n <name> -h <hostname>
    startwpar <name>
    syncwpar <name>                     # bring WPAR filesets in line with global
  • Front the affected service with PowerSC + a reverse proxy and lock down RBAC roles:

    aixpert -l high                     # apply high security profile
    aixpert -c                          # check current applied profile
    lssec -f /etc/security/user -s default -a admin -a rlogin
    setsecattr -r <role> auth=...       # tighten role authorisations
    swrole <role>                       # switch to least-priv role for ops
  • Migrate the LPAR temporarily to a TL/SP combination not affected by this CVE while a full upgrade is scheduled. Use alt_disk_copy to clone an unaffected rootvg from a sibling LPAR and boot it:

    alt_disk_copy -d hdisk2
    bootlist -m normal hdisk2           # boot from cloned rootvg
    shutdown -Fr now

Verification & Acceptance Criteria

All of these should pass after the fix:

oslevel -s                          # expected TL-SP-build
instfix -ivk <APAR>                 # APAR reported as installed
lppchk -v                           # exits 0 (no version inconsistency)
lppchk -c                           # exits 0 (no checksum mismatch)
errpt | head -20                    # no new fail/crash entries since fix
lssrc -a | grep -v active           # no unexpectedly stopped subsystems
emgr -l                             # ifixes superseded/removed as expected

The original reproduction for CVE-2006-0674 — buffer overflow — patch and remediation guide must not trigger across two consecutive runs.

Rollback Plan

Take a rootvg safety net BEFORE the fix:

# Option A: full mksysb image (NIM target, NFS, tape, or file):
mksysb -i -X /mnt/backup/$(hostname)-pre-fix.mksysb
# Option B: alt_disk_copy clone of rootvg to a spare disk:
lspv                                # find an unused hdisk
alt_disk_copy -B -d hdisk1
lspv | grep altinst_rootvg          # confirm clone exists

To revert if the fix is bad:

# Uncommitted installp updates can be rejected:
installp -r <fileset>
# Interim fixes:
emgr -r -L <ifix-label>
# Or fall back to the cloned rootvg:
bootlist -m normal hdisk1           # the altinst_rootvg disk
shutdown -Fr now
# Verify on the recovered system:
oslevel -s && lppchk -v

Prevention & Hardening

Reduce the chance of recurrence on IBM AIX 7.1:

  • Adopt a NIM-driven TL/SP cadence with a staging LPAR per release:

    # On the NIM master, build lpp_source from the latest TL:
    nim -o define -t lpp_source -a server=master -a location=/export/lpp_source/AIX73TL02 lpp_AIX73TL02
    # Push the bundle to a target:
    nim -o cust -a lpp_source=lpp_AIX73TL02 -a fixes=update_all <target>
  • Subscribe to IBM PSIRT and the AIX security bulletins index. Use My Notifications for proactive APAR alerts.

  • Apply AIXpert + PowerSC compliance profiles aligned to your workload:

    aixpert -l high
    aixpert -c
    # PowerSC compliance scans:
    pscxpert -l default                 # PowerSC default profile
  • Enable TCB (Trusted Computing Base) integrity checks:

    tcbck -y ALL                        # check whole TCB
    trustchk -p ALL                     # print trust attributes
    trustchk -n ALL                     # check and report violations only
  • Rotate mksysb backups + alt_disk_copy before each fix campaign:

    mksysb -i -X /mnt/backup/$(hostname)-$(date +%F).mksysb
    alt_disk_copy -B -d hdisk1
  • Audit security attributes for service accounts regularly:

    lssec -f /etc/security/user -s root -a admin -a rlogin -a maxage
    lssec -f /etc/security/login.cfg -s default
    chsec -f /etc/security/user -s root -a rlogin=false

Issues that commonly surface alongside CVE-2006-0674 — buffer overflow — patch and remediation guide: fileset version drift, hung subsystems, stale ODM device entries, dump-device misconfiguration after a kernel-level fix, and rootvg space exhaustion during installp. Useful triage:

lppchk -v
lssrc -a | grep -v active
lsdev -C | grep -i defined
odmget -q 'name=hdisk0' CuAt | head
sysdumpdev -l
df -g /usr /var /tmp                # rootvg filesystems

View all ibm-aix-71 tutorials on the Tutorials Hub →

Browse all common problems & solutions on the Tutorials Hub.

References & Further Reading

Primary references: NVD CVE-2006-0674, IBM Support Bulletin. Manual pages useful on IBM AIX 7.1:

man installp
man instfix
man oslevel
man emgr
man nim
man alt_disk_copy
man mksysb
man errpt
man lssrc
man aixpert

Other resources: IBM Fix Central, IBM AIX documentation, AIX security bulletins, and per-component notes in /usr/share/man/info/en_US/a_doc_lib/aixbman/ implicated in CVE-2006-0674 — buffer overflow — patch and remediation guide.