Re: [PATCH v2] rcu/kvfree: Prevents cache growing when the backoff_page_cache_fill is set

[Uladzislau Rezki]

On Wed, Apr 12, 2023 at 02:36:30PM +0200, Uladzislau Rezki wrote:
> On Wed, Apr 12, 2023 at 04:18:20AM +0000, Zhang, Qiang1 wrote:
> > > On Thu, Apr 06, 2023 at 11:11:37PM +0000, Zhang, Qiang1 wrote:
> > > > >>On Thu, Apr 06, 2023 at 06:37:53AM +0200, Uladzislau Rezki wrote:
> > > > > On Thu, Apr 06, 2023 at 08:12:38AM +0800, Zqiang wrote:
> > > > > > Currently, in kfree_rcu_shrink_scan(), the drain_page_cache() is
> > > > > > executed before kfree_rcu_monitor() to drain page cache, if the bnode
> > > > > > structure's->gp_snap has done, the kvfree_rcu_bulk() will fill the
> > > > > > page cache again in kfree_rcu_monitor(), this commit add a check
> > > > > > for krcp structure's->backoff_page_cache_fill in put_cached_bnode(),
> > > > > > if the krcp structure's->backoff_page_cache_fill is set, prevent page
> > > > > > cache growing.
> > > > > > 
> > > > > > Signed-off-by: Zqiang <qiang1.zhang@xxxxxxxxx>
> > > > > > ---
> > > > > >  kernel/rcu/tree.c | 2 ++
> > > > > >  1 file changed, 2 insertions(+)
> > > > > > 
> > > > > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > > > > > index 9cc0a7766fd2..f25430ae1936 100644
> > > > > > --- a/kernel/rcu/tree.c
> > > > > > +++ b/kernel/rcu/tree.c
> > > > > > @@ -2907,6 +2907,8 @@ static inline bool
> > > > > >  put_cached_bnode(struct kfree_rcu_cpu *krcp,
> > > > > >  	struct kvfree_rcu_bulk_data *bnode)
> > > > > >  {
> > > > > > +	if (atomic_read(&krcp->backoff_page_cache_fill))
> > > > > > +		return false;
> > > > > >  	// Check the limit.
> > > > > >  	if (krcp->nr_bkv_objs >= rcu_min_cached_objs)
> > > > > >  		return false;
> > > > > > -- 
> > > > > > 2.32.0
> > > > > >
> > > > > Reviewed-by: Uladzislau Rezki (Sony) <urezki@xxxxxxxxx>
> > > > >
> > > > >Thank you both!
> > > > >
> > > > >One question, though.  Might it be better to instead modify the "for"
> > > > >loop in fill_page_cache_func() to start at krcp->nr_bkv_objs instead
> > > > >of starting at zero?  That way, we still provide a single page under
> > > > >low-memory conditions, but provide rcu_min_cached_objs of them if memory
> > > > >is plentiful.
> > > > >
> > > > >Alternatively, if we really don't want to allow any pages at all under
> > > > >low-memory conditions, shouldn't the fill_page_cache_func() set nr_pages
> > > > >to zero (instead of the current 1) when the krcp->backoff_page_cache_fill
> > > > >flag is set?  
> > > > 
> > > > Hi, Paul
> > > > 
> > > > If the krcp->backoff_page_cache_fill is true, the put_cached_bnode () return false,
> > > > the allocated single page will also be freed in fill_page_cache_func().
> > > > 
> > > > or it would be better not to allocate under memory pressure.
> > > 
> > > That was my thought.  ;-)
> > > 
> > > > How about like this?
> > > > 
> > > > diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> > > > index 9cc0a7766fd2..94aedbc3da36 100644
> > > > --- a/kernel/rcu/tree.c
> > > > +++ b/kernel/rcu/tree.c
> > > > @@ -2907,6 +2907,8 @@ static inline bool
> > > >  put_cached_bnode(struct kfree_rcu_cpu *krcp,
> > > >         struct kvfree_rcu_bulk_data *bnode)
> > > >  {
> > > > +       if (atomic_read(&krcp->backoff_page_cache_fill))
> > > > +               return false;
> > > >         // Check the limit.
> > > >         if (krcp->nr_bkv_objs >= rcu_min_cached_objs)
> > > >                 return false;
> > > > @@ -3220,7 +3222,7 @@ static void fill_page_cache_func(struct work_struct *work)
> > > >         int i;
> > > > 
> > > >         nr_pages = atomic_read(&krcp->backoff_page_cache_fill) ?
> > > > -               1 : rcu_min_cached_objs;
> > > > +               0 : rcu_min_cached_objs;
> > > > 
> > > >         for (i = 0; i < nr_pages; i++) {
> > > 
> > > The other question is why this loop does not allow for any pages
> > > that might already be allocated, thus perhaps looking like this:
> > > 
> > > 		for (i = krcp->nr_bkv_objs; i < nr_pages; i++) {
> > > 
> > > Or do we somehow know that krcp->nr_bkv_objs is equal to zero?  (I am not
> > > seeing this, but I do feel the need to ask.)
> > > 
> > >Usually we start from zero, this is when a ptr. was not added into
> > >a bulk array, due to no memory reason for a single argument and no
> > >cache pages anymore for double argument.
> > >
> > >In the fill page function, the limit is checked by the put_cached_bnode() itself
> > >so it stops prefetch once nr_bkv_objs contains desired value.
> > >
> > 
> > If the krcp->nr_bkv_objs is updated in kfree_rcu_work() and happens before invoke fill_page_cache_func(),
> > when invoke fill_page_cache_func(), we start from zero,  will allocate page and hold krcp->lock,
> > fill krcp->bkvcache,  but if krcp->nr_bkv_objs already equal to rcu_min_cached_objs,  this page will
> > be freed and exit loop,  this allocate page seems like a meaningless operation. 
> > 
> > I also want to ask if starting from krcp->nr_bkv_objs is necessary?
> > 
> At least it does not break anything. The example like you described can
> occur. So starting from the krcp->nr_bkv_objs is worth to do.
> 
> So, if it happens it would be good of you could simulate it and update
> the commit message accordingly.
> 
A small nit, the krcp->nr_bkv_objs can not be accessed without a lock.
So it must then accessed using READ_ONCE()/WRITE_ONCE() helpers.

--
Uladzislau Rezki