This is the mail archive of the cygwin@sourceware.cygnus.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

concurrency problem (crash) of gnuwin32 b18 in nt 4.0 sp3


i'm using gnuwin32 b18 on windows nt, with service pack 3 installed.
the problem is this: i have a long running program that sleeps for a minute or
so and then wakes up to perform an activity.  the code for this is a simple
batch file:

  @echo off
  :repeater
  call do_the_real_work.bat
  sleep 108
  goto :repeater

this runs fine, for days and days if necessary, until i have another shell
window open that also uses gnuwin32.  for example, if i perform a compile in
another cmd window, i will often get a crash that resembles this:

( 1000) In cygwin_except_handler
( 1000) Exception trapped!
( 1000) exception C0000005 at 10011EB3
( 1000) exception: ax 0 bx 0 cx 2E92970 dx 3
( 1000) exception: si 0 di A0000FC bp 241F8E4 sp 241F8D8
( 1000) exception is: STATUS_ACCESS_VIOLATION
( 1000) Stack trace:
( 1000) frame 0: sp = 0x241F70C, pc = 0x1000CEC2
( 1000) frame 1: sp = 0x241F728, pc = 0x77F94512
( 1000) frame 2: sp = 0x241F74C, pc = 0x77F88EEB
( 1000) frame 3: sp = 0x241F7D8, pc = 0x77F76266
( 1000) frame 4: sp = 0x241F8E4, pc = 0x10020B3A
( 1000) frame 5: sp = 0x241F904, pc = 0x1003056B
( 1000) frame 6: sp = 0x241F918, pc = 0x1002700B
( 1000) frame 7: sp = 0x241F93C, pc = 0x10027070
( 1000) frame 8: sp = 0x241F950, pc = 0x10015DD2
( 1000) frame 9: sp = 0x241FA30, pc = 0x10024259
( 1000) frame 10: sp = 0x241FA3C, pc = 0x1000BBF7
( 1000) frame 11: sp = 0x241FF94, pc = 0x1000C113
( 1000) frame 12: sp = 0x241FFA0, pc = 0x40F301
( 1000) frame 13: sp = 0x241FFB0, pc = 0x40103B
( 1000) frame 14: sp = 0x241FFC0, pc = 0x77F1B304
( 1000) frame 15: sp = 0x241FFF0, pc = 0x0
( 1000) End of stack trace

which terminates either my sleeping program, the program i just started (a
"diff" in the above case), or both.

  the problem seems to point to some shared resource collision; the sleeper
program and the diff are executing in entirely different command (cmd) shells,
and i can think of no other way that they would conflict with each other.
  i have done the obligatory search for a previously reported bug of this
nature and am really surprised i couldn't find one.  thus this message...
  has anyone else experienced this?  i'm not sure it's confined to windows nt,
but i haven't had an example of win95 blowing up recently.
thanks,
chris.

-- 
__________________ chosen by the Nechung Oracle program... __________________ 
 
This is the way the world ends. Not with a bang but a whimper.
                         - T. S. Eliot, "The Hollow Men", 1925
 
_____________ not necessarily my opinions, not necessarily not. _____________
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request@cygnus.com" with one line of text: "help".


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]