[ale] Re: perl new IPC::Semaphore and/or C semget()

[Geoffrey]

Geoffrey wrote:
> I've done quite a bit of work with SysV ipcs, and some with perl.  I 
> can't recall, or locate the answer to a question that's bugging me.
> 
> When using new IPC::Semaphore() semget() or semget(2), is the process of 
> getting the semaphore atomic?  Assuming usage as in:
> 
> my $sem = new IPC::Semaphore($semkey, 1, IPC_EXCL | IPC_CREAT | 0666);
> 
> or similar use of semget(2) in C?
> 
> The point is, does anyone see the flaw in using exclusive creation of a 
> semaphore as the semaphore?

It appears that it is atomic as the man page for semget states:

        The  presence  in  semflg  of  the  fields  IPC_CREAT  and
        IPC_EXCL  plays  the  same role, with respect to the exis?
        tence of the semaphore set, as the presence of O_CREAT and
        O_EXCL  in  the  mode argument of the open(2) system call:
        i.e. the semget function  fails  if  semflg  asserts  both
        IPC_CREAT  and IPC_EXCL and a semaphore set already exists
        for key.

Don't know how I missed that.

-- 
Until later, Geoffrey                     Registered Linux User #108567
Building secure systems in spite of Microsoft