Re: [PATCH v2 tip/core/rcu 04/39] srcu: Check for tardy grace-period activity in cleanup_srcu_struct()

From: Josh Triplett
Date: Tue Apr 18 2017 - 15:45:35 EST


On Tue, Apr 18, 2017 at 11:34:53AM -0700, Paul E. McKenney wrote:
> On Mon, Apr 17, 2017 at 05:34:30PM -0700, Josh Triplett wrote:
> > On Mon, Apr 17, 2017 at 05:33:32PM -0700, Josh Triplett wrote:
> > > On Mon, Apr 17, 2017 at 04:44:51PM -0700, Paul E. McKenney wrote:
> > > > Users of SRCU are obliged to complete all grace-period activity before
> > > > invoking cleanup_srcu_struct(). This means that all calls to either
> > > > synchronize_srcu() or synchronize_srcu_expedited() must have returned,
> > > > and all calls to call_srcu() must have returned, and the last call to
> > > > call_srcu() must have been followed by a call to srcu_barrier().
> > > > Furthermore, the caller must have done something to prevent any
> > > > further calls to synchronize_srcu(), synchronize_srcu_expedited(),
> > > > and call_srcu().
> > > >
> > > > Therefore, if there has ever been an invocation of call_srcu() on
> > > > the srcu_struct in question, the sequence of events must be as
> > > > follows:
> > > >
> > > > 1. Prevent any further calls to call_srcu().
> > > > 2. Wait for any pre-existing call_srcu() invocations to return.
> > > > 3. Invoke srcu_barrier().
> > > > 4. It is now safe to invoke cleanup_srcu_struct().
> > > >
> > > > On the other hand, if there has ever been a call to synchronize_srcu()
> > > > or synchronize_srcu_expedited(), the sequence of events must be as
> > > > follows:
> > > >
> > > > 1. Prevent any further calls to synchronize_srcu() or
> > > > synchronize_srcu_expedited().
> > > > 2. Wait for any pre-existing synchronize_srcu() or
> > > > synchronize_srcu_expedited() invocations to return.
> > > > 3. It is now safe to invoke cleanup_srcu_struct().
> > > >
> > > > If there have been calls to all both types of functions (call_srcu()
> > > > and either of synchronize_srcu() and synchronize_srcu_expedited()), then
> > > > the caller must do the first three steps of the call_srcu() procedure
> > > > above and the first two steps of the synchronize_s*() procedure above,
> > > > and only then invoke cleanup_srcu_struct().
> > >
> > > This commit message clearly explains the correct sequence for the
> > > client, but not which aspects of this the change now enforces. Some of
> > > the steps above remain the responsibility of the caller, while the
> > > callee now checks more of them. Could you add something at the end
> > > explaining the change and what it enforces?
> >
> > More importantly, perhaps this explanation could find its way into the
> > documentation of cleanup_srcu_struct?
>
> Like this?
>
> /**
> * cleanup_srcu_struct - deconstruct a sleep-RCU structure
> * @sp: structure to clean up.
> *
> * Must invoke this only after you are finished using a given srcu_struct
> * that was initialized via init_srcu_struct(). This code does some
> * probabalistic checking, spotting late uses of srcu_read_lock(),
> * synchronize_srcu(), synchronize_srcu_expedited(), and call_srcu().
> * If any such late uses are detected, the per-CPU memory associated with
> * the srcu_struct is simply leaked and WARN_ON() is invoked. If the
> * caller frees the srcu_struct itself, a use-after-free crash will likely
> * ensue, but at least there will be a warning printed.
> */
>
> I added the following paragraph to the commit log:
>
> Note that cleanup_srcu_struct() does some probabilistic checks
> for the caller failing to follow these procedures, in which
> case cleanup_srcu_struct() does WARN_ON() and avoids freeing
> the per-CPU structures associated with the specified srcu_struct
> structure.
>
> And added your Reviewed-by, but please let me if more is needed.

Looks good to me, thanks for improving the documentation.