[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference decwet::windows-nt

Title:Windows NT
Notice:See note 15.0 for HCL location
Moderator:TARKIN::LIN.com::FOLEY
Created:Thu Oct 31 1991
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:6086
Total number of notes:31449

5990.0. "?PANIC_STACK_SWITCH caused by CD Player?" by RICKS::OPP () Thu May 08 1997 08:48

    	Alpha PC systems running NT V4.0 SP2 have been observed to crash
    if the eject button on the CD Player application is clicked two or
    three times.  The STOP error is a PANIC_STACK_SWITCH, as described 
    in the attached QAR.  Will this problem be resolved in Service Pack
    3?  
    
    Thanks,
    
    Greg
    
    
    
            <<< RICKS::SYS$SYSDEVICE:[NOTES$LIBRARY]APC164.NOTE;1 >>>
                                -< AlphaPC164 >-
================================================================================
Note 79.0           QAR 042: WNT BugCheck via CD Player eject          2 replies
GREGOR::OPP                                        1773 lines  11-APR-1997 11:48
--------------------------------------------------------------------------------

     System Type:  PC164LX              QAR Number:  039
     Entered by:  Greg Opp              Date Entered:  11-APR-1997
     Software:  WNT V4.0 SP2            First Observed:  10-APR-1997
     Firmware:  AlphaBIOS 5.60 Proto    System:  PERKED, 466 MHz, 128 MB
     Hardware:  Motherboard Rev. B01    SROM: 2.1-23
                S/N: KA711VLNPJ


     Problem Summary

        While using the Multimedia CD Player and attempting to unload an
     audio CD, I accidentally doubled-clicked on the eject button.  The
     CD drawer opened and then immediately closed.  So, I clicked the 
     virtual eject button a third time.  Soon afterwards, PERKED crashed 
     with a PANIC_STACK_SWITCH STOP error.  

        According to Des Gordon's "Interpreting System Crashes on Windows
     NT", this crash message:

        indicates a panic switch to the kernel stack because of stack
        overflow.  This error may be caused by the kernel-mode driver
        using too much stack space.  Or data corruption has occurred
        in the kernel.  

     While other SIG engineers have observed this behavior and attributed
     it to specific models of CD-ROM drives, this crash strongly suggests
     that a Windows NT driver deficiency is also involved.  


     Test Description

        There are no additions to the description provided in the Problem
     Summary.

     
     Problem Description 

        When PERKED crashed, the following "blue screen of death" informa-
     tion was recorded in the system log book.

        STOP 0x0000002B (0x80082428, 0x00000003, 0xF173BFF0, 0xF1740000)
        PANIC_STACK_SWITCH      Microsoft Windows NT [0xf000565]

        Machine State at call to BugCheck  PC: 800BBE04  PSR:  1E

        Callee SP       Return RA       DLL Base               Name
        ---------       ---------       --------        ---------------------
        806CDE00        800BBE04        80080000        ntoskrnl.exe
        806CE000        8008242C        80080000        ntoskrnl.exe
        806CE010        F1045008        80080000        ntoskrnl.exe
        806CE050        00000000        F1040000        Cdfs.sys
        806CE050        00000004        00000000
        806CE050        00000004        00000000
        806CE050        00000004        00000000
        806CE050        00000004        00000000

        IRQL:  7        DPC Active:  False



     Tentative Conclusion 

        This crash appears to be caused by the CDFS.SYS driver consuming all
     the available stack space.  However, since the system apparently 
     "paniced" and switched to the kernel stack, the stack overflow is not
     obvious (but presumably retained in the dump file).  If it's not the
     CD file system driver, then some other component of the Windows NT 
     operating system is presumably responsible for this failure, since
     hardware does not directly use the stack.  


     Attachments & References

        A WinDbg log file from a crash dump analysis session is attached.
     It clearly shows the final stages of the progression to BugCheck and
     the registers and current process at that time.  

[ WinDbg log file deleted. ]
T.RTitleUserPersonal
Name
DateLines