Re: [PATCH 12/30] mm: memory reserve management

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Mon, 2008-07-28 at 13:06 +0300, Pekka Enberg wrote:
> Hi Peter,
> 
> On Thu, 2008-07-24 at 16:00 +0200, Peter Zijlstra wrote:
> > +/*
> > + * alloc wrappers
> > + */
> > +
> 
> Hmm, I'm not sure I like the use of __kmalloc_track_caller() (even
> though you do add the wrappers for SLUB). The functions really are SLAB
> internals so I'd prefer to see kmalloc_reserve() moved to the
> allocators.

See below..

> > +void *___kmalloc_reserve(size_t size, gfp_t flags, int node, void *ip,
> > +			 struct mem_reserve *res, int *emerg)
> > +{
> 
> This function could use some comments...

Yes, my latest does have those.. let me paste the relevant bit:

+void *___kmalloc_reserve(size_t size, gfp_t flags, int node, void *ip,
+                        struct mem_reserve *res, int *emerg)
+{
+       void *obj;
+       gfp_t gfp;
+
+       /*
+        * Try a regular allocation, when that fails and we're not entitled
+        * to the reserves, fail.
+        */
+       gfp = flags | __GFP_NOMEMALLOC | __GFP_NOWARN;
+       obj = __kmalloc_node_track_caller(size, gfp, node, ip);
+
+       if (obj || !(gfp_to_alloc_flags(flags) & ALLOC_NO_WATERMARKS))
+               goto out;
+
+       /*
+        * If we were given a reserve to charge against, try that.
+        */
+       if (res && !mem_reserve_kmalloc_charge(res, size)) {
+               /*
+                * If we failed to charge and we're not allowed to wait for
+                * it to succeed, bail.
+                */
+               if (!(flags & __GFP_WAIT))
+                       goto out;
+
+               /*
+                * Wait for a successfull charge against the reserve. All
+                * uncharge operations against this reserve will wake us up.
+                */
+               wait_event(res->waitqueue,
+                               mem_reserve_kmalloc_charge(res, size));
+
+               /*
+                * After waiting for it, again try a regular allocation.
+                * Pressure could have lifted during our sleep. If this
+                * succeeds, uncharge the reserve.
+                */
+               obj = __kmalloc_node_track_caller(size, gfp, node, ip);
+               if (obj) {
+                       mem_reserve_kmalloc_charge(res, -size);
+                       goto out;
+               }
+       }
+
+       /*
+        * Regular allocation failed, and we've successfully charged our
+        * requested usage against the reserve. Do the emergency allocation.
+        */
+       obj = __kmalloc_node_track_caller(size, flags, node, ip);
+       WARN_ON(!obj);
+       if (emerg)
+               *emerg |= 1;
+
+out:
+       return obj;
+}


> > +	void *obj;
> > +	gfp_t gfp;
> > +
> > +	gfp = flags | __GFP_NOMEMALLOC | __GFP_NOWARN;
> > +	obj = __kmalloc_node_track_caller(size, gfp, node, ip);
> > +
> > +	if (obj || !(gfp_to_alloc_flags(flags) & ALLOC_NO_WATERMARKS))
> > +		goto out;
> > +
> > +	if (res && !mem_reserve_kmalloc_charge(res, size)) {
> > +		if (!(flags & __GFP_WAIT))
> > +			goto out;
> > +
> > +		wait_event(res->waitqueue,
> > +				mem_reserve_kmalloc_charge(res, size));
> > +
> > +		obj = __kmalloc_node_track_caller(size, gfp, node, ip);
> > +		if (obj) {
> > +			mem_reserve_kmalloc_charge(res, -size);
> 
> Why do we discharge here?

because a regular allocation succeeded.

> > +			goto out;
> > +		}
> 
> If the allocation fails, we try again (but nothing has changed, right?).
> Why?

Note the different allocation flags for the two allocations.

> > +	}
> > +
> > +	obj = __kmalloc_node_track_caller(size, flags, node, ip);
> > +	WARN_ON(!obj);
> 
> Why don't we discharge from the reserve here if !obj?

Well, this allocation should never fail:
  - we reserved memory
  - we accounted/throttle its usage

Thus this allocation should always succeed.

> > +	if (emerg)
> > +		*emerg |= 1;
> > +
> > +out:
> > +	return obj;
> > +}
> > +
> > +void __kfree_reserve(void *obj, struct mem_reserve *res, int emerg)
> 
> I don't see 'emerg' used anywhere.

Patch 19/30 has:

-       data = kmalloc_node_track_caller(size + sizeof(struct skb_shared_info),
-                       gfp_mask, node);
+       data = kmalloc_reserve(size + sizeof(struct skb_shared_info),
+                       gfp_mask, node, &net_skb_reserve, &emergency);
        if (!data)
                goto nodata;

@@ -205,6 +211,7 @@ struct sk_buff *__alloc_skb(unsigned int
         * the tail pointer in struct sk_buff!
         */
        memset(skb, 0, offsetof(struct sk_buff, tail));
+       skb->emergency = emergency;
        skb->truesize = size + sizeof(struct sk_buff);
        atomic_set(&skb->users, 1);
        skb->head = data;


> > +{
> > +	size_t size = ksize(obj);
> > +
> > +	kfree(obj);
> 
> We're trying to get rid of kfree() so I'd __kfree_reserve() could to
> mm/sl?b.c. Matt, thoughts?

My issue with moving these helpers into mm/sl?b.c is that it would
require replicating all this code 3 times. Even though the functionality
is (or should) be invariant to the actual slab implementation.

> > +	/*
> > +	 * ksize gives the full allocated size vs the requested size we used to
> > +	 * charge; however since we round up to the nearest power of two, this
> > +	 * should all work nicely.
> > +	 */
> > +	mem_reserve_kmalloc_charge(res, -size);
> > +}
> > 
>