[cpp-threads] Review comments on N2176 WRT dependency ordering

[Peter Dimov]

Paul E. McKenney wrote:
> On Tue, Apr 17, 2007 at 02:38:29PM +0300, Peter Dimov wrote:

>> In the dependency discussion paper, you write that a single level of
>> indirection is not enough for the Linux kernel, but I don't see how
>> you could make a multi-level primitive work on an Alpha.
>
> This depends on how the multi-level list was created. [...]

Yes, you are right. atomic_load_address will probably always work in 
practice for such dependencies, but I'm not sure that I can specify it so 
that it will also work in theory.

> Why not just use atomic_load_address() on the index?  Then given a
> static array whose indexes are dynamically filled in:
>
> i = atomic_load_address(myindex);
> r1 = myarray[i];

This could work in practice, but again, I'm not sure that I can specify it 
in a formal way since the object whose visibility needs to be ensured is not 
known until after the second line.

The explicit per-object fence can handle these scenarios - in principle - 
since the user specifies the dependent object explicitly.

i = atomic_load_relaxed( myindex );

dependency_fence( &myarray[ i ] );
dependency_fence( &myarray2[ i * 2 ] );

r1 = myarray[ i ].foo;
r2 = myarray[ i ].bar;
r3 = myarray[ i*2 ].baz;