[OpenIndiana-discuss] SS cc multithread compilation switch (e.g., Parrot)
Udo Grabowski (IMK)
udo.grabowski at kit.edu
Wed Oct 31 13:49:21 UTC 2012
On 31/10/2012 14:24, James Carlson wrote:
> On 10/31/2012 8:34 AM, Udo Grabowski (IMK) wrote:
>> I don't know where you get your information from, this original
>> documentation says something else:
> PSARC 2001/287 is the source of most of it. I was an employee of Sun
> from 2000 through 2009, and was chair of the Platform Software
> Architecture Review Committee (PSARC) for much of that time.
>> Compiling "successfully" without _REENTRANT does NOT mean that
>> the program actually RUNS flawlessly in a multithreaded environment
>> under all conditions. One source also says that -mt changes dummy
>> locks to real locks, and I don't think you want dummy locks in
>> a mutlithreaded program...
> The old dummy-lock crap went away with Solaris 10. That's the essence
> of this comment in the libpthread man page:
> Historically, functions in this library provided POSIX
> threading support. See standards(5). This functionality now
> resides in libc(3LIB).
> It turned out that the dummy locks were *MUCH* harder to maintain (test
> and code around) than the real thing, and the real thing is just as
> cheap in terms of performance. That's why Roger Faulkner (raf) merged
> it all into libc.
> > ....
>> Or is all this documentation totally outdated and wrong ??
> It's outdated. The Sun Studio folks are a completely separate group
> from the folks who actually built Solaris. The two groups have little
> technical contact.
Thanks a lot for this info, it's a pity that this isn't documented
well, I also read the Multithread Guide
and it's all not entirely complete nor completely consistent...
confusing, at best.
Dr.Udo Grabowski Inst.f.Meteorology a.Climate Research IMK-ASF-SAT
KIT - Karlsruhe Institute of Technology http://www.kit.edu
Postfach 3640,76021 Karlsruhe,Germany T:(+49)721 608-26026 F:-926026
More information about the OpenIndiana-discuss