[OpenAFS-devel] AIX system hang II

[Hans-Werner Paulsen]

Dear OpenAFS developers on AIX,
again we had a system hang on AIX 4.3.3 with OpenAFS 1.2.5
due to a deadlock. Now I have traces of both threads:

crash> lock
...
Threads waiting on locks:
  TSLOT  WAITING ON  HELD BY  SYMBOL OF ADDRESS
            ADDRESS    TSLOT  (if found)
...
    180  0x05549bf0      205  <.[afs.ext.iauth:DATA]+c950>
...
    205  0x05551b44      180  <.[afs.ext.iauth:DATA]+148a4>
...

crash> thread 180 205   
SLT ST    TID      PID    CPUID  POLICY PRI CPU    EVENT  PROCNAME
180 s    b4a1     9f76  unbound   other  25   0                 sh
        t_flags:
205 s    cdab     b982  unbound   other  25   0              xvscf
        t_flags:
  
crash> trace -mk 180 205
Skipping first MST
     
MST STACK TRACE:
0x2ff3b400 (excpt=00000000:40000000:6005cb77:2ff22ff8:00000106) (intpri=0)
        IAR:      .slock_ppc+25c (00150590):    cmpi   cr0,0x0,r19,0x0
        LR:       .slock_ppc+25c (00150590)
        2ff3a410: .simple_lock+64 (00009564)
        2ff3a450: .[afs.ext.iauth:rxi_Alloc]+64 (054b41b8)
        2ff3a4b0: .[afs.ext.iauth:rxepoch_Allocate]+2c (054e95ec)
        2ff3a510: .[afs.ext.iauth:rxevent_Post]+d4 (054e9d30)
        2ff3a5b0: .[afs.ext.iauth:rxi_ScheduleKeepAliveEvent]+88 (054aceb0)
        2ff3a600: .[afs.ext.iauth:rxi_KeepAliveOn]+3c (054acc94)
        2ff3a650: .[afs.ext.iauth:rx_NewCall]+3d0 (054b6d14)
        2ff3a6e0: .[afs.ext.iauth:afs_GetDCache]+1b48 (05516548)
        2ff3a990: .[afs.ext.iauth:afs_lookup]+4b8 (05522ff8)
        2ff3aae0: .[afs.ext.iauth:afs_gn_lookup]+48 (05508090)
        2ff3ab30: .[afs.ext.iauth:vn_lookup]+a0 (05503724)
        2ff3ab80: .vnop_lookup+1c (001c2958)
        2ff3abc0: .lookuppn+4d4 (001a56b4)
        2ff3ae60: .lookupname_cur+150 (001a5c90)
        2ff3b2d0: .statx+150 (001db0f0)
        2ff3b3c0: .sys_call_ret+0 (00003a90)
        0000a450: .low+0 (00000000)
        
Skipping first MST

MST STACK TRACE:
0x2ff3b400 (excpt=00000000:42000000:00074c1d:424c0000:00000106) (intpri=0)
        IAR:      .slock_ppc+25c (00150590):    cmpi   cr0,0x0,r19,0x0
        LR:       .slock_ppc+25c (00150590)
        2ff3a920: .simple_lock+64 (00009564)
        2ff3a960: .[afs.ext.iauth:rxevent_Cancel_1]+2c (054e9b40)
        2ff3a9c0: .[afs.ext.iauth:rxi_Start]+aec (054ade80)
        2ff3aad0: .[afs.ext.iauth:rxi_WritevProc]+89c (054e7ff0)
        2ff3aba0: .[afs.ext.iauth:rx_WritevProc]+40 (054e9358)
        2ff3abf0: .[afs.ext.iauth:afs_MemCacheStoreProc]+154 (05517ec4)
        2ff3ac50: .[afs.ext.iauth:afs_StoreAllSegments]+938 (0550d2a8)
        2ff3af50: .[afs.ext.iauth:afs_DoPartialWrite]+88 (0551ea14)
        2ff3afa0: .[afs.ext.iauth:afs_vm_rdwr]+cd8 (055054fc)
        2ff3b140: .[afs.ext.iauth:afs_gn_rdwr]+44c (05506774)
        2ff3b1e0: .[afs.ext.iauth:vn_rdwr]+b0 (0550268c)
        2ff3b230: .vnop_rdwr+c0 (001c259c)
        2ff3b290: .rwuio+cc (001a6318)
        2ff3b2e0: .rdwr+128 (001a6518)
        2ff3b340: .kwrite+48 (001a5f40)
        2ff3b3c0: .sys_call_ret+0 (00003a90)
        0502e000: kfork+0 (0013e00c)
        0000e000: .nodev+0 (00094b70)
IAR not in kernel segment.

-- 
Hans-Werner Paulsen		hans@MPA-Garching.MPG.DE
MPI für Astrophysik		Tel 089-30000-2602
Karl-Schwarzschild-Str. 1	Fax 089-30000-2235	
D-85741 Garching