[cpp-threads] Web site updated
Chris Thomasson
cristom at comcast.net
Fri Feb 16 05:15:43 GMT 2007
> On Tue, Feb 13, 2007 at 10:48:43PM -0000, Boehm, Hans wrote:
>> > From: Paul E. McKenney [mailto:paulmck at linux.vnet.ibm.com]
[...]
> I am concerned about algorithms that interact via signal handlers.
So am I.
> The mainline code interacts with a signal handler on the same
> CPU, and the signal handler does the multi-CPU operations. I am
> willing to exclude non-atomic_${TYPEKEY} types in this case.
Sure; I hope the Standard will be flexible enough (e.g.; well, C++ is a
syslang?) to allow programmers to deal with signals. Well, I have a
full-blown prototype of an atomic reference counted pointer in IA32 assembly
language. It signal safe in about 80% of its API:
http://appcore.home.comcast.net/vzoom/refcount/
(100% full source code...)
http://appcore.home.comcast.net/vzoom/refcount/doc/
(pre-alpha fairly detailed per-API function documentation...)
I dream of a day that I can use C++ to create my counting algorithm... Well,
assembly language is perfect! When C++ compiler gets the 'standard' ability
to be as stable as an assembler wrt. lock-free programming in general...
Well, good times all around indeed!!!
:^)
Any thoughts?
More information about the cpp-threads
mailing list