deadlock

From: uDAy (arakumar_at_gmail.com)
Date: 05/31/05

  • Next message: Jan M. STANKOVSKY: "SUMMARY: SunRays over Routers"
    Date: Tue, 31 May 2005 16:45:50 +0530
    To: sunmanagers@sunmanagers.org
    
    

    Hi,

    My application is getting stuck and the pstack shows that

    1. One thread is stuck in free system call
    2. Another thread is stuck in malloc system call..
    3. Another thread is stuck in user defined lock.

    I seriously dont understand the possible reason for this...
    this has happend two times in a week but the application was running for
    past 5 months...

    Can u tell me the possible reasons for this.

    I am sending the pstack output... and also the truss output collected after
    the stuck...

    Thank you in advance,

    **************pstack *********************

    ----------------- lwp# 1 / thread# 1 --------------------
    ff21f1b8 lwp_sema_wait (1b2b5a8)
    ff129ac4 _park (1b2b5a8, ff14e000, 0, 1b2b4e8, 24d84, 0) + 114
    ff12978c _swtch (1b2b4e8, 1b2b4e8, ff14e000, 5, 1000, 1802be0) + 424
    ff12b148 _mutex_adaptive_lock (ff1598e4, 4c00, 1000, fffeffff, 1, 4d58) +
    160
    ff12ae80 _cmutex_lock (ff240600, ff14e000, a9f0818, ff1c2b3c, 0, 0) + 70
    ff1c2b3c free (faa0050, ff9bafb8, 11, ff9baf00, 231b58, ff9bad34) + 18
    001fad50 HssFn_Memfree (102, ff9bafb8, 11, ff9baf00, ff9baf40, 0) + 2c
    001ec75c ras_indicate (11d50140, faa0050, ff9bb0e8, ff9bb220, 11f05900, 0) +
    3464
    0031133c ras_peerMessageHandler (11d50140, ff9bb0e8, ff9bb068, 4, 11433820,
    46) + 544
    0030f05c ras_peerEventHandler (0, ff9bb0e8, 11433820, 46, ff9bb0fc, 0) + b0
    0026bd80 H323Fn_IvkRasChannelReceive (28811d4, ff9bb220, 0, 26ba98, 0, 0) +
    2e8
    0024c85c HssIvk_Receive (28811d4, ff9bb220, 0, 4, 0, ffbee888) + 134
    000411fc GkMnFn_ProcessRasMsg (ffbee828, ff9bbae4, 40800, 0, 0, 0) + 218
    000402f8 GkMnFn_ProcessMessageFromAllocator (ffbee828, ffbee828, 0, 0,
    ffbef0c0, ff9bba08) + 80
    0003f30c main (ffddc998, ffbef24c, 0, 10eebb0, 0, 0) + 1d4c
    0003d420 _start (0, 0, 0, 0, 0, 0) + 5c
    ----------------- lwp# 2 / thread# 2 --------------------
    ff21eb08 signotifywait ()
    ff12ed54 _dynamiclwps (ff14e000, 0, ff3e95a0, 0, 0, ff3b61d8) + 1c
    ff132030 thr_yield (0, 0, 0, 0, 0, 0) + 8c
    ----------------- lwp# 3 / thread# 4 --------------------
    ff21f1b8 lwp_sema_wait (fee0be30)
    ff129ac4 _park (fee0be30, ff14e000, 0, fee0bd70, 24d84, fed09d70) + 114
    ff12978c _swtch (fee0bd70, fee0bd70, ff14e000, 5, 1000, 0) + 424
    ff12b148 _mutex_adaptive_lock (ff1598e4, 4c00, 1000, fffeffff, 1, 4d58) +
    160
    ff12bbf8 pthread_mutex_lock (0, ff14e000, 46ec00, 1431d38, 0, 10f8fd9) + f8
    000f7a6c GkCliFn_readStatistics (f6b2c, f6da0, f6c00, ff23fc70, 0, 0) + 48
    000f6b2c GkCli_main (0, ff115d10, 1, ff15ad8c, 0, 2) + 108
    ff13b744 _thread_start (0, 0, 0, 0, 0, 0) + 40
    ----------------- lwp# 4 / thread# 5 --------------------
    ff21f1b8 lwp_sema_wait (fed09e30)
    ff129ac4 _park (fed09e30, ff14e000, 0, fed09d70, 24d84, 1b2b4e8) + 114
    ff12978c _swtch (fed09d70, fed09d70, ff14e000, 5, 1000, 0) + 424
    ff12b148 _mutex_adaptive_lock (ff1598e4, 4c00, 1000, fffeffff, 1, 4d58) +
    160
    ff12ae80 _cmutex_lock (ff240600, ff14e000, 0, ff1c1ca8, 0, 0) + 70
    ff1c1ca8 malloc (31, ffbeffa8, 6f6e6600, 7efefeff, 81010100, ff00) + 18
    ff1cf2f4 strdup (ffbeff77, 46954c, b4000, 0, 0, 0) + c
    000b3fb0 GkConfFn_OpenConfig (1a, 410450, fed09cd8, fed09cd4, 410400,
    fed09c51) + c4
    000c54d0 HaGkRpmIf_StoreRegistrations (0, ff105d10, 1, ff15ad8c, 0, 2) + 114
    ff13b744 _thread_start (0, 0, 0, 0, 0, 0) + 40
    ----------------- lwp# 5 --------------------------------
    ff21c6ac door (0, 0, 0, 0, ff0f5d10, 4)
    ff136bac _sc_door_func (1a, ff14f688, ff14f6a0, 3, ff14e000, 1) + 54
    ff12a740 _lwp_start (ff0f5d70, 0, 6000, ff105b74, 0, 0) + 18
    ff132030 thr_yield (0, 0, 0, 0, 0, 0) + 8c
    ----------------- lwp# 6 --------------------------------
    ff13978c lwp_cond_wait (ff155548, ff155558, ff14edb0)
    ff1290ac _age (3e, ff14ed9c, ff14e000, 3, ff14e000, 1) + 74
    ff12a740 _lwp_start (ff0e3d70, 0, 4000, ff00fc2c, 0, 0) + 18
    ff132030 thr_yield (0, 0, 0, 0, 0, 0) + 8c
    -------------------------- thread# 3 --------------------
    ff12ddbc _reap_wait (ff1529e0, 20520, 0, ff14e000, 0, 0) + 38
    ff12db14 _reaper (ff14ee30, ff154740, ff1529e0, ff14ee08, 1, fe400000) + 38
    ff13b744 _thread_start (0, 0, 0, 0, 0, 0) + 40

    =========End of pstack ===========

    =========Start of truss ===========

    bash# truss -p 28050
    6427
    truss: non-existent process ignored: 28050
    lwp_sema_wait(0x01B2B5A8) (sleeping...)
    signotifywait() (sleeping...)
    lwp_sema_wait(0xFEE0BE30) (sleeping...)
    lwp_sema_wait(0xFED09E30) (sleeping...)
    door_return(0x00000000, 0, 0x00000000, 0) (sleeping...)
    lwp_cond_wait(0xFF155548, 0xFF155558, 0xFF14EDB0) (sleeping...)

    =========End of truss ===========

    --
    A.Uday Kumar,
    Software Engineer,
    Flextronics Software Systems,
    Bangalore.
    Mobile:+91-94483-81052.
    Office:080-51069077.
    _______________________________________________
    sunmanagers mailing list
    sunmanagers@sunmanagers.org
    http://www.sunmanagers.org/mailman/listinfo/sunmanagers
    

  • Next message: Jan M. STANKOVSKY: "SUMMARY: SunRays over Routers"

    Relevant Pages

    • Fw: Solaris Multithreaded Application Hang
      ... The output of the pstack shows that it got ... Please let us know the possible reason for the stuck. ... damage arising from the use of the information transmitted ...
      (SunManagers)
    • Re: "route" problem
      ... It's stuck right there? ... The very next action taken here is to open a PF_ROUTE socket. ... do a pstack on the stuck process to see exactly where it is, ... sounds like some kind of kernel problem to me, ...
      (comp.unix.solaris)
    • Re: Thread debugging utilities.
      ... debugging on Solaris. ... Say application is stuck at some point, how can I figure out whether ... there is some thread deadlock? ...
      (comp.programming.threads)
    • Re: Thread debugging utilities.
      ... debugging on Solaris. ... Say application is stuck at some point, how can I figure out whether ... there is some thread deadlock? ...
      (comp.programming.threads)
    • Re: locked thread - how to investigate?
      ... With truss is there any way to focus on a ... the threads are being consumed which causes more consumption of the ... Perhaps you mean a thread stuck in a CPU loop? ...
      (comp.unix.solaris)