IV42221: SYSTEM HANG IN SODEQUEUE() AND SOPUTONFREELIST() APPLIES TO AIX 7100-01

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• System hang with the following stack.
  .
  --- CPU #0 ---
  pvthread+000200 STACK:
  [004B16C8]krlock+000248 (F1000A0480110E00,
  F1000A00232EB288 [??])
  [00534138]slock_krlock_acquire+000258 (??, ??, ??)
  [00534778]slock+000318 (??, ??)
  [000090C4].disable_lock+0000C0 ()
  [00469548]soputonfreelist+0009A8 (??)
  [043BB204]in_pcbdetach+0002E4 (??)
  [044455C4]tcp_close+000164 (??)
  [04445CFC]tcp_drop+0000FC (??, ??)
  ..
  --- CPU #12 ---
  pvthread+0BAB00 STACK:
  [005346C4]slock+000264 (F1000E00010B4300,
  F000000030853110 [??])
  [000090C4].disable_lock+0000C0 ()
  [004E6D34]sodequeue+000374 (??, ??, ??, ??)
  [004E6388]k_soaccept+000548 (??, ??, ??, ??, ??)
  [004E56A0]accept1+000200 (??, ??, ??, ??, ??)
  [00003850]ovlya_addr_sc_flih_main+000130 ()
  [kdb_get_virtual_memory] no real storage @ 368386B8
  [D0230AA8]D0230AA8 ()
  [kdb_read_mem] no real storage @ FFFFFFFFFFF8EF0
  

Local fix

• N/A
  

Problem summary

• System may hang because of a deadlock between sodequeue() and
  soputonfreelist(). This may happen when there is one thread
  trying to accept a connection and another thread is trying to
  drop that same connection because of an error. The stack traces
  of two threads would look like this:
  
  --- CPU #0 ---
  pvthread+000200 STACK:
   004B16C8 krlock+000248 (F1000A0480110E00,
  F1000A00232EB288  ?? )
   00534138 slock_krlock_acquire+000258 (??, ??, ??)
   00534778 slock+000318 (??, ??)
   000090C4 .disable_lock+0000C0 ()
   00469548 soputonfreelist+0009A8 (??)
   043BB204 in_pcbdetach+0002E4 (??)
   044455C4 tcp_close+000164 (??)
   04445CFC tcp_drop+0000FC (??, ??)
  ..
  --- CPU #12 ---
  pvthread+0BAB00 STACK:
   005346C4 slock+000264 (F1000E00010B4300,
  F000000030853110  ?? )
   000090C4 .disable_lock+0000C0 ()
   004E6D34 sodequeue+000374 (??, ??, ??, ??)
   004E6388 k_soaccept+000548 (??, ??, ??, ??, ??)
   004E56A0 accept1+000200 (??, ??, ??, ??, ??)
   00003850 ovlya_addr_sc_flih_main+000130 ()
  ..
  

Problem conclusion

• Fixed a lock order issue in socket code to avoid the deadlock.
  

Temporary fix

Comments

• 6100-06 - use AIX APAR IV21636
  6100-07 - use AIX APAR IV34802
  6100-08 - use AIX APAR IV42094
  7100-00 - use AIX APAR IV42247
  7100-01 - use AIX APAR IV42221
  7100-02 - use AIX APAR IV42147
  

APAR Information

• APAR is sysrouted FROM one or more of the following:

  IV21636

• APAR is sysrouted TO one or more of the following:

Fix information

• Fixed component name

  AIX V7.1

• Fixed component ID

  5765H4000

Applicable component levels

• R710 PSY U860705

     UP13/07/30 I 1000