Re: fix cond_race... was RE: src/winsup/cygwin ChangeLog thread.cc thread.h ...

[Jason Tishler <jason at tishler dot net>]

Rob,

On Sat, Oct 27, 2001 at 01:06:04PM +1000, Robert Collins wrote:
> So whats the problem again?

See below...

> $ ssh -p 2222 robertc@cc1004678-a.etntwn1.nj.home.com
> [snip]
> TERM (rxvt) = cd src/python-2.1.1

The above is prompting you for your terminal type -- enter carriage return
to default to the previous value (e.g., rxvt) or enter the appropriate
value (e.g., cygwin).

> althea[~/src/python-2.1.1]
> $ PYTHONPATH= ./python.exe -tt Lib/test/regrtest.py -l test_threadedtempfile
> test_threadedtempfile
> 1 test OK.

The above Python regression test will hang often but not every time.
You need to execute the above a few times and it should hang at with the
following gdb backtrace:

(gdb) bt
#0  0x77f82152 in _libkernel32_a_iname ()
#1  0x77e830fe in _libkernel32_a_iname ()
#2  0x77e83126 in _libkernel32_a_iname ()
#3  0x61062b34 in pthread_cond::TimedWait (this=0xa0c76f0, 
    dwMilliseconds=4294967295) at ../../../../src/winsup/cygwin/thread.cc:538
#4  0x61064623 in __pthread_cond_dowait (cond=0xa0c3edc, mutex=0xa0c3ee0, 
    waitlength=-1) at ../../../../src/winsup/cygwin/thread.cc:1769        
#5  0x6106475c in pthread_cond_wait (cond=0xa0c3edc, mutex=0xa0c3ee0)
    at ../../../../src/winsup/cygwin/thread.cc:1806                  
#6  0x61d928ad in PyThread_acquire_lock (lock=0xa0c3ed8, waitflag=1)
    at Python/thread_pthread.h:313
..

Thanks,
Jason