Mail Archives: cygwin/2002/12/11/03:48:55
Sorry, I forgot to mention:
Your work around is not portable. The type of sem_t is not necessary a
pointer.
Portable would be something like:
memset(&sem, 0, sizeof(sem));
Thomas
Thomas Pfaff wrote:
> Just add a breakpoint after the sem_init call.
> To avoid multiple initializations sem_init checks if the semaphore is
> already initialized and will return EBUSY in that case.
> The check is done via IsBadWritePtr which will trigger a SEGV if the
> pointer is invalid and not NULL.
> The SEGV is handled internally and will not cause the program to stop.
>
> Thomas
>
> Oreste Bernardi wrote:
>
>> I installed
>> gdb 20020718-1.
>>
>> I tried again and I see same behavior.
>>
>> By the way, is gdb 20020718-1 experimental version or stable version ?
>>
>> Oreste
>>
>>
>> ----- Original Message ----- From: <lhall AT pop DOT ma DOT ultranet DOT com>
>> To: <oreste DOT bernardi AT st DOT com>; <cygwin AT cygwin DOT com>
>> Sent: Tuesday, December 10, 2002 5:03 PM
>> Subject: RE: SIGSEGV when called sem_init() with gdb.
>>
>>
>> Rerun setup and update your gdb. If that doesn't help, let the list
>> know.
>>
>> Larry
>>
>> Original Message:
>> -----------------
>> From: Oreste Bernardi oreste DOT bernardi AT st DOT com
>> Date: Tue, 10 Dec 2002 16:01:16 +0100
>> To: cygwin AT cygwin DOT com
>> Subject: SIGSEGV when called sem_init() with gdb.
>>
>>
>> Dear, all
>>
>> #include <semaphore.h>
>> int main () {
>> int p;
>> sem_t sem;
>> int i=0,d;
>> //sem=(sem_t)NULL; //With this There isn't any problem
>> sem_init (&sem,0,1);
>> sem_post (&sem);
>> sem_wait(&sem);
>> return i;
>> }
>>
>> cygwin 1.3.17-1
>> gcc 3.2-3
>> gdb 20010428-3
>>
>> When executing this source code with gdb I get an SIGSEGV signal when
>> sem_init() function is called.
>> If I continue on, program works fine.
>> If program is not executed from gdb, it works without any error.
>> If I initialize it with NULL, I don't get any error.
>>
>> I think that is the same behavior as described by following mail:
>> http://sources.redhat.com/ml/cygwin/2002-10/msg00956.html
>> Is this true ?
>>
>> To avoid annoying gdb error message, is my solution ("sem=(sem_t)NULL")
>> reliable and portable ?
>>
>> Best Regards
>> Oreste Bernardi
>>
>>
>> --
>> Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
>> Bug reporting: http://cygwin.com/bugs.html
>> Documentation: http://cygwin.com/docs.html
>> FAQ: http://cygwin.com/faq/
>>
>>
>> --------------------------------------------------------------------
>> mail2web - Check your email from the web at
>> http://mail2web.com/ .
>>
>>
>>
>>
>
>
>
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Bug reporting: http://cygwin.com/bugs.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -