Very Computer

by **Venka** » Fri, 20 Jun 2003 18:06:43

Hi All,

If my understanding is right the equivalent data type of CriticalSection of
Win32 API
in Linux(all other flavors of unix/POSIX standard) is pthread_mutex_t.

In Windows we have 4 methods for critical section like
InitializeCriticalSection
EnterCriticalSection
LeaveCriticalSection
DeleteCriticalSection

and the equivalent in Linux(POSIX) are
pthread_mutex_init
pthread_mutex_lock
pthread_mutex_unlock

but i want to know what is the equivalent of DeleteCriticalSection in Linux?

Regards
Venkat

by **David Schwart** » Fri, 20 Jun 2003 18:19:47

Quote:> Hi All,

> If my understanding is right the equivalent data type of CriticalSection
of
> Win32 API
> in Linux(all other flavors of unix/POSIX standard) is pthread_mutex_t.

> In Windows we have 4 methods for critical section like
> InitializeCriticalSection
> EnterCriticalSection
> LeaveCriticalSection
> DeleteCriticalSection

> and the equivalent in Linux(POSIX) are
> pthread_mutex_init
> pthread_mutex_lock
> pthread_mutex_unlock

> but i want to know what is the equivalent of DeleteCriticalSection in

Linux?

pthread_mutex_destroy.

DS

by **Timur Aydi** » Sat, 21 Jun 2003 07:15:39

> Hi All,

> If my understanding is right the equivalent data type of CriticalSection of
> Win32 API
> in Linux(all other flavors of unix/POSIX standard) is pthread_mutex_t.

There is one important difference between these two:

The win32 critical sections can be recursively locked by the same
thread, while the pthread_mutex_t cannot. An application relying on
the recursive nature of Critical Sections will deadlock under Linux.

--
Timur Aydin

by **Kevin Easto** » Sat, 21 Jun 2003 08:27:00

>> Hi All,

>> If my understanding is right the equivalent data type of CriticalSection of
>> Win32 API
>> in Linux(all other flavors of unix/POSIX standard) is pthread_mutex_t.

> There is one important difference between these two:

> The win32 critical sections can be recursively locked by the same
> thread, while the pthread_mutex_t cannot. An application relying on
> the recursive nature of Critical Sections will deadlock under Linux.

You can specify a recursive mutex at creation time, if that's what you
need.

- Kevin.

by **Timur Aydi** » Sat, 21 Jun 2003 16:33:00

> You can specify a recursive mutex at creation time, if that's what you
> need.

Last time I checked, this was a non-portable extension of glibc and
not startardized by POSIX. Has this changed?

--
Timur Aydin

by **Thomas Borg Salli** » Thu, 26 Jun 2003 23:14:19

> Last time I checked, this was a non-portable extension of glibc and
> not startardized by POSIX. Has this changed?

IEEE POSIX 1003.1c -- draft 10 is the final and approved version.

There's lots of discussions about this on comp.programming.threads.

Best regards,
Thomas.
-linux/g++/ACE/TAO/Qt-developer for hire!

Very Computer

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux

CriticalSection-NT & Mutex-Linux