On Sat, Feb 24, 2018 at 01:32:50PM +0800, Boqun Feng wrote:
> On Fri, Feb 23, 2018 at 08:37:32PM +0800, Boqun Feng wrote:
> > On Fri, Feb 23, 2018 at 12:55:20PM +0100, Peter Zijlstra wrote:
> > > On Thu, Feb 22, 2018 at 03:08:51PM +0800, Boqun Feng wrote:
> > > > @@ -1012,6 +1013,33 @@ static inline bool bfs_error(enum bfs_result res)
> > > >  	return res < 0;
> > > >  }
> > > >  
> > > > +#define DEP_NN_BIT 0
> > > > +#define DEP_RN_BIT 1
> > > > +#define DEP_NR_BIT 2
> > > > +#define DEP_RR_BIT 3
> > > > +
> > > > +#define DEP_NN_MASK (1U << (DEP_NN_BIT))
> > > > +#define DEP_RN_MASK (1U << (DEP_RN_BIT))
> > > > +#define DEP_NR_MASK (1U << (DEP_NR_BIT))
> > > > +#define DEP_RR_MASK (1U << (DEP_RR_BIT))
> > > > +
> > > > +static inline unsigned int __calc_dep_bit(int prev, int next)
> > > > +{
> > > > +	if (prev == 2 && next != 2)
> > > > +		return DEP_RN_BIT;
> > > > +	if (prev != 2 && next == 2)
> > > > +		return DEP_NR_BIT;
> > > > +	if (prev == 2 && next == 2)
> > > > +		return DEP_RR_BIT;
> > > > +	else
> > > > +		return DEP_NN_BIT;
> > > > +}
> > > > +
> > > > +static inline unsigned int calc_dep(int prev, int next)
> > > > +{
> > > > +	return 1U << __calc_dep_bit(prev, next);
> > > > +}
> > > > +
> > > >  static enum bfs_result __bfs(struct lock_list *source_entry,
> > > >  			     void *data,
> > > >  			     int (*match)(struct lock_list *entry, void *data),
> > > > @@ -1921,6 +1949,16 @@ check_prev_add(struct task_struct *curr, struct held_lock *prev,
> > > >  		if (entry->class == hlock_class(next)) {
> > > >  			if (distance == 1)
> > > >  				entry->distance = 1;
> > > > +			entry->dep |= calc_dep(prev->read, next->read);
> > > > +		}
> > > > +	}
> > > > +
> > > > +	/* Also, update the reverse dependency in @next's ->locks_before list */
> > > > +	list_for_each_entry(entry, &hlock_class(next)->locks_before, entry) {
> > > > +		if (entry->class == hlock_class(prev)) {
> > > > +			if (distance == 1)
> > > > +				entry->distance = 1;
> > > > +			entry->dep |= calc_dep(next->read, prev->read);
> > > >  			return 1;
> > > >  		}
> > > >  	}
> > > 
> > > I think it all becomes simpler if you use only 2 bits. Such that:
> > > 
> > >   bit0 is the prev R (0) or N (1) value,
> > >   bit1 is the next R (0) or N (1) value.
> > > 
> > > I think this should work because we don't care about the empty set
> > > (currently 0000) and all the complexity in patch 5 is because we can
> > > have R bits set when there's also N bits. The concequence of that is
> > > that we cannot replace ! with ~ (which is what I kept doing).
> > > 
> > > But with only 2 bits, we only track the strongest relation in the set,
> > > which is exactly what we appear to need.
> > > 
> > 
> > But if we only have RN and NR, both bits will be set, we can not check
> > whether we have NN or not. Consider we have:
> > 
> > 	A -(RR)-> B
> > 	B -(NR)-> C and B -(RN)-> C
> > 	C -(RN)-> A
> > 
> > this is not a deadlock case, but with "two bits" approach, we can not
> > differ this with:
> > 
> > 	A -(RR)-> B
> > 	B -(NN)-> C
> > 	C -(RN)-> A
> > 
> > , which is a deadlock.
> > 
> > But maybe "three bits" (NR, RN and NN bits) approach works, that is if
> > ->dep is 0, we indicates this is only RR, and is_rx() becomes:
> > 
> > 	static inline bool is_rx(u8 dep)
> > 	{
> > 		return !(dep & (NR_MASK | NN_MASK));
> > 	}
> > 
> > and is_xr() becomes:
> > 
> > 	static inline bool is_xr(u8 dep)
> > 	{
> > 		return !(dep & (RN_MASK | NN_MASK));
> > 	}
> > 
> > , with this I think your simplification with have_xr works, thanks!
> > 
> 
> Ah! I see. Actually your very first approach works, except the
> definitions of is_rx() and ir_xr() are wrong. In that approach, you
> define
> 	
> 	static inline bool is_rx(u8 dep)
> 	{
> 		return !!(dep & (DEP_RR_MASK | DEP_RN_MASK);
> 	}
> 
> , which means "whether we have a R* dependency?". But in fact, what we
> need to check is "whether we _only_ have R* dependencies?", if so and
> have_xr is true, that means we could only have a -(*R)-> A -(R*)-> if we
> pick the next dependency, and that means we should skip. So my new
> definition above works, and I think we better name it as only_rx() to
> avoid confusion? Ditto for is_xr().
> 
> I also reorder bit number for each kind of dependency, so that we have a
> simple __calc_dep_bit(), see the following:
> 
> 	/*
> 	 * DEP_*_BIT in lock_list::dep
> 	 *
> 	 * For dependency @prev -> @next:
> 	 *
> 	 *   RR: both @prev and @next are recursive read locks, i.e. ->read == 2.
> 	 *   RN: @prev is recursive and @next is non-recursive.
> 	 *   NR: @prev is a not recursive and @next is recursive.
> 	 *   NN: both @prev and @next are non-recursive.
> 	 * 
> 	 * Note that we define the value of DEP_*_BITs so that:
> 	 * 	bit0 is prev->read != 2
> 	 * 	bit1 is next->read != 2
> 	 */
> 	#define DEP_RR_BIT 0
> 	#define DEP_RN_BIT 1
> 	#define DEP_NR_BIT 2
> 	#define DEP_NN_BIT 3
> 
> 	#define DEP_RR_MASK (1U << (DEP_RR_BIT))
> 	#define DEP_RN_MASK (1U << (DEP_RN_BIT))
> 	#define DEP_NR_MASK (1U << (DEP_NR_BIT))
> 	#define DEP_NN_MASK (1U << (DEP_NN_BIT))
> 
> 	static inline unsigned int
> 	__calc_dep_bit(struct held_lock *prev, struct held_lock *next)
> 	{
> 		return (prev->read != 2) + ((next->read != 2) << 1)
> 	}
> 
> 	static inline u8 calc_dep(struct held_lock *prev, struct held_lock *next)
> 	{
> 		return 1U << __calc_dep_bit(prev, next);
> 	}
> 
>  	static inline bool only_rx(u8 dep)
>  	{
>  		return !(dep & (DEP_NR_MASK | DEP_NN_MASK));
>  	}
> 
>  	static inline bool only_xr(u8 dep)
>  	{
>  		return !(dep & (DEP_NR_MASK | DEP_NN_MASK));
>  	}
> 
> Note that we actually don't need DEP_RR_BIT, but I leave it there for
> implementation simplicity. With this, your check and set below works.
> 
> Thoughts?
> 
> Regards,
> Boqun
> 
> > > 
> > > 
> > > 	if (have_xr && is_rx(entry->dep))
> > > 		continue;
> > > 
> > > 	entry->have_xr = is_xr(entry->dep);
> > > 

Hmm.. I think this part also needs some tweak:

	/* if -> prev is *R, and we only have R* for prev -> this, * skip*/
	if (have_xr && only_rx(entry->dep))
		continue;
	
	/*
	 * we pick a *R for prev -> this only if:
	 *     prev -> this dependencies are all *R 
	 * or
	 *     -> prev is *R, and we don't have NN for prev -> this
	 */
	entry->have_xr = only_xr(entry->dep) || (have_xr && !is_nn(entry->dep));

otherwise, we will wrongly set entry->have_xr to false if have_xr is
true and we have RN for prev -> this.

Regards,
Boqun

> > > 
> > > Or did I mess that up somewhere?
> 
>