[Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Corinna Vinschen]

[Catching up on some older mails]

> ----- Forwarded message from "Gerrit P. Haase" -----
> From: "Gerrit P. Haase" 
> To: cygwin ML
> Subject: sem_* functions in cygwin
> Date: Sun, 21 Nov 2004 22:48:20 +0100
> 
> Hi,
> 
> nearly all sem_* functions are available, but sem_unlock is missing,
> was there a problem implementing sem_unlock() or was it just missed
> by accident?
> 
> 
> Gerrit
> ----- End forwarded message -----

I guess you're asking about sem_unlink().  It's not implemented so far
since named POSIX semaphores are implemented using named Windows semaphores. 
The SUSv3 description contains a pretty unfortunate implementation detail:

  Calls to sem_open() to recreate or reconnect to the semaphore refer
  to a new semaphore after sem_unlink() is called.

There's no way I know of, which allows to implement this using named
Windows semaphores.  At least not without adding a lot of annoying
bookkeeping overhead, possibly involving cygserver.


Corinna

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Mark Paulus]

So, does that mean that if process 1 opens a semaphore,
process 2 also grabs it, then process 1 unlinks it, and then
"reconnects" to it, that process 1 and process 2 do not have
and cannot have the same semaphore anymore, even though
they are using the same IPC_KEY?

(Or am I way confused/off base here)?

  On Thu, 09 Dec 2004 17:44:42 +0100, Corinna Vinschen wrote:

>[Catching up on some older mails]

>> ----- Forwarded message from "Gerrit P. Haase" -----
>> From: "Gerrit P. Haase" 
>> To: cygwin ML
>> Subject: sem_* functions in cygwin
>> Date: Sun, 21 Nov 2004 22:48:20 +0100
>> 
>> Hi,
>> 
>> nearly all sem_* functions are available, but sem_unlock is missing,
>> was there a problem implementing sem_unlock() or was it just missed
>> by accident?
>> 
>> 
>> Gerrit
>> ----- End forwarded message -----

>I guess you're asking about sem_unlink().  It's not implemented so far
>since named POSIX semaphores are implemented using named Windows semaphores. 
>The SUSv3 description contains a pretty unfortunate implementation detail:

>  Calls to sem_open() to recreate or reconnect to the semaphore refer
>  to a new semaphore after sem_unlink() is called.

>There's no way I know of, which allows to implement this using named
>Windows semaphores.  At least not without adding a lot of annoying
>bookkeeping overhead, possibly involving cygserver.


>Corinna

>--
>Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
>Problem reports:       http://cygwin.com/problems.html
>Documentation:         http://cygwin.com/docs.html
>FAQ:                   http://cygwin.com/faq/





--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Corinna Vinschen]

On Dec  9 09:50, Mark Paulus wrote:
> So, does that mean that if process 1 opens a semaphore,
> process 2 also grabs it, then process 1 unlinks it, and then
> "reconnects" to it, that process 1 and process 2 do not have
> and cannot have the same semaphore anymore, even though
> they are using the same IPC_KEY?
> 
> (Or am I way confused/off base here)?

Only partially.  IPC_KEYs have nothing to do with it since we're talking
about *POSIX* semaphores, not *SYSV* semaphores.

Otherwise you're right.  After some process has called sem_unlink(), any
subsequent call to sem_open with the same sempahore name connects to a new
semaphore.


Corinna

> 
>   On Thu, 09 Dec 2004 17:44:42 +0100, Corinna Vinschen wrote:
> 
> >[Catching up on some older mails]
> 
> >> ----- Forwarded message from "Gerrit P. Haase" -----
> >> From: "Gerrit P. Haase" 
> >> To: cygwin ML
> >> Subject: sem_* functions in cygwin
> >> Date: Sun, 21 Nov 2004 22:48:20 +0100
> >> 
> >> Hi,
> >> 
> >> nearly all sem_* functions are available, but sem_unlock is missing,
> >> was there a problem implementing sem_unlock() or was it just missed
> >> by accident?
> >> 
> >> 
> >> Gerrit
> >> ----- End forwarded message -----
> 
> >I guess you're asking about sem_unlink().  It's not implemented so far
> >since named POSIX semaphores are implemented using named Windows semaphores. 
> >The SUSv3 description contains a pretty unfortunate implementation detail:
> 
> >  Calls to sem_open() to recreate or reconnect to the semaphore refer
> >  to a new semaphore after sem_unlink() is called.
> 
> >There's no way I know of, which allows to implement this using named
> >Windows semaphores.  At least not without adding a lot of annoying
> >bookkeeping overhead, possibly involving cygserver.
> 
> 
> >Corinna
> 
> >--
> >Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> >Problem reports:       http://cygwin.com/problems.html
> >Documentation:         http://cygwin.com/docs.html
> >FAQ:                   http://cygwin.com/faq/
> 
> 
> 
> 
> 
> --
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> Problem reports:       http://cygwin.com/problems.html
> Documentation:         http://cygwin.com/docs.html
> FAQ:                   http://cygwin.com/faq/

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Gerrit P. Haase]

Corinna Vinschen wrote:

> [Catching up on some older mails]
> 
> 
>>----- Forwarded message from "Gerrit P. Haase" -----
>>From: "Gerrit P. Haase" 
>>To: cygwin ML
>>Subject: sem_* functions in cygwin
>>Date: Sun, 21 Nov 2004 22:48:20 +0100
>>
>>Hi,
>>
>>nearly all sem_* functions are available, but sem_unlock is missing,
>>was there a problem implementing sem_unlock() or was it just missed
>>by accident?
>>
>>
>>Gerrit
>>----- End forwarded message -----
> 
> 
> I guess you're asking about sem_unlink().  It's not implemented so far
> since named POSIX semaphores are implemented using named Windows semaphores. 
> The SUSv3 description contains a pretty unfortunate implementation detail:
> 
>   Calls to sem_open() to recreate or reconnect to the semaphore refer
>   to a new semaphore after sem_unlink() is called.
> 
> There's no way I know of, which allows to implement this using named
> Windows semaphores.  At least not without adding a lot of annoying
> bookkeeping overhead, possibly involving cygserver.

I got an undefined reference to sem_unlock().

Gerrit
-- 
=^..^=

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Gerrit P. Haase]

Gerrit P. Haase wrote:
> Corinna Vinschen wrote:
> 
>> [Catching up on some older mails]
>>
>>
>>> ----- Forwarded message from "Gerrit P. Haase" -----
>>> From: "Gerrit P. Haase" To: cygwin ML
>>> Subject: sem_* functions in cygwin
>>> Date: Sun, 21 Nov 2004 22:48:20 +0100
>>>
>>> Hi,
>>>
>>> nearly all sem_* functions are available, but sem_unlock is missing,
>>> was there a problem implementing sem_unlock() or was it just missed
>>> by accident?
>>>
>>>
>>> Gerrit
>>> ----- End forwarded message -----
>>
>>
>>
>> I guess you're asking about sem_unlink().  It's not implemented so far
>> since named POSIX semaphores are implemented using named Windows 
>> semaphores. The SUSv3 description contains a pretty unfortunate 
>> implementation detail:
>>
>>   Calls to sem_open() to recreate or reconnect to the semaphore refer
>>   to a new semaphore after sem_unlink() is called.
>>
>> There's no way I know of, which allows to implement this using named
>> Windows semaphores.  At least not without adding a lot of annoying
>> bookkeeping overhead, possibly involving cygserver.
> 
> 
> I got an undefined reference to sem_unlock().

I don't remember now which package it was, however there is a maro
used in linux which is based on ipc_unlock():
#define sem_unlock(id)  ipc_unlock(&sem_ids,id)

http://www.iglu.org.il/lxr/ident?i=sem_unlock


Gerrit
-- 
=^..^=

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [sem_* functions in cygwin]]

[Corinna Vinschen]

[Oops, I've sent this to Gerrit by PM instead of to the list. 
 Just resending it now for the records]

On Dec 12 14:54, Gerrit P. Haase wrote:
> Corinna Vinschen wrote:
> >>----- Forwarded message from "Gerrit P. Haase" -----
> >>nearly all sem_* functions are available, but sem_unlock is missing,
> >>was there a problem implementing sem_unlock() or was it just missed
> >>by accident?
> >>Gerrit
> >>----- End forwarded message -----
> >
> >I guess you're asking about sem_unlink().  It's not implemented so far
> >since named POSIX semaphores are implemented using named Windows 
> >semaphores. The SUSv3 description contains a pretty unfortunate 
> >implementation detail:
> >
> >  Calls to sem_open() to recreate or reconnect to the semaphore refer
> >  to a new semaphore after sem_unlink() is called.
> >
> >There's no way I know of, which allows to implement this using named
> >Windows semaphores.  At least not without adding a lot of annoying
> >bookkeeping overhead, possibly involving cygserver.
> 
> I got an undefined reference to sem_unlock().

There's no such function defined in SUSv3.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: [Fwd: [gp@familiehaase.de: sem_* functions in cygwin]]

[Corinna Vinschen]

On Dec 20 12:23, Gerrit P. Haase wrote:
> Gerrit P. Haase wrote:
> >I got an undefined reference to sem_unlock().
> 
> I don't remember now which package it was, however there is a maro
> used in linux which is based on ipc_unlock():
> #define sem_unlock(id)  ipc_unlock(&sem_ids,id)
> 
> http://www.iglu.org.il/lxr/ident?i=sem_unlock

That's inside the Linux kernel.  There's no such function or macro
defined on the API side of Linux.

Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/