[PATCH v2] Fix double-checked locking in _res_hconf_reorder_addrs [BZ #19074]

[Torvald Riegel]

On Wed, 2015-10-07 at 16:12 +0200, Florian Weimer wrote:
> On 10/07/2015 12:28 PM, Torvald Riegel wrote:
> 
> > If you should write this page, check whether you really need the acquire
> > MO in the load in the critical section, or whether relaxed MO is
> > sufficient ;)
> 
> Acquire/release is needed because the access to the actual data needs to
> be synchronized as well.

Yes, but there's the lock used for the critical section too.  The
critical sections are ordered.  So have a look at whether you can
actually not read from initialization done in the initializing critical
section if you read that the initialization flag has been set.

> I'm struggling with the cppmem syntax to show that.  Is there a grammar
> published anywhere?

Not as far as I know.  You could look at the sources of the tool
though :)

The slides of my Cauldron tutorial have a selection of what is accepted.
Modeling a lock order with a release/acquire store/load pair can be
helpful.