net: sched: refactor block offloads counter usage

Without rtnl lock protection filters can no longer safely manage block
offloads counter themselves. Refactor cls API to protect block offloadcnt
with tcf_block->cb_lock that is already used to protect driver callback
list and nooffloaddevcnt counter. The counter can be modified by concurrent
tasks by new functions that execute block callbacks (which is safe with
previous patch that changed its type to atomic_t), however, block
bind/unbind code that checks the counter value takes cb_lock in write mode
to exclude any concurrent modifications. This approach prevents race
conditions between bind/unbind and callback execution code but allows for
concurrency for tc rule update path.

Move block offload counter, filter in hardware counter and filter flags
management from classifiers into cls hardware offloads API. Make functions
tcf_block_offload_{inc|dec}() and tc_cls_offload_cnt_update() to be cls API
private. Implement following new cls API to be used instead:

  tc_setup_cb_add() - non-destructive filter add. If filter that wasn't
  already in hardware is successfully offloaded, increment block offloads
  counter, set filter in hardware counter and flag. On failure, previously
  offloaded filter is considered to be intact and offloads counter is not
  decremented.

  tc_setup_cb_replace() - destructive filter replace. Release existing
  filter block offload counter and reset its in hardware counter and flag.
  Set new filter in hardware counter and flag. On failure, previously
  offloaded filter is considered to be destroyed and offload counter is
  decremented.

  tc_setup_cb_destroy() - filter destroy. Unconditionally decrement block
  offloads counter.

  tc_setup_cb_reoffload() - reoffload filter to single cb. Execute cb() and
  call tc_cls_offload_cnt_update() if cb() didn't return an error.

Refactor all offload-capable classifiers to atomically offload filters to
hardware, change block offload counter, and set filter in hardware counter
and flag by means of the new cls API functions.

Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 14 insertions, 24 deletions

diff --git a/net/sched/cls_flower.c b/net/sched/cls_flower.c
index 054123742e32..cb816bbbd376 100644
--- a/net/sched/cls_flower.c
+++ b/net/sched/cls_flower.c
@@ -419,10 +419,10 @@ static void fl_hw_destroy_filter(struct tcf_proto *tp, struct cls_fl_filter *f,
 	cls_flower.command = FLOW_CLS_DESTROY;
 	cls_flower.cookie = (unsigned long) f;
 
-	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+	tc_setup_cb_destroy(block, tp, TC_SETUP_CLSFLOWER, &cls_flower, false,
+			    &f->flags, &f->in_hw_count, true);
 	spin_lock(&tp->lock);
 	list_del_init(&f->hw_list);
-	tcf_block_offload_dec(block, &f->flags);
 	spin_unlock(&tp->lock);
 
 	if (!rtnl_held)
@@ -466,18 +466,13 @@ static int fl_hw_replace_filter(struct tcf_proto *tp,
 		goto errout;
 	}
 
-	err = tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, skip_sw);
+	err = tc_setup_cb_add(block, tp, TC_SETUP_CLSFLOWER, &cls_flower,
+			      skip_sw, &f->flags, &f->in_hw_count, true);
 	kfree(cls_flower.rule);
 
-	if (err < 0) {
+	if (err) {
 		fl_hw_destroy_filter(tp, f, true, NULL);
 		goto errout;
-	} else if (err > 0) {
-		f->in_hw_count = err;
-		err = 0;
-		spin_lock(&tp->lock);
-		tcf_block_offload_inc(block, &f->flags);
-		spin_unlock(&tp->lock);
 	}
 
 	if (skip_sw && !(f->flags & TCA_CLS_FLAGS_IN_HW)) {
@@ -509,7 +504,7 @@ static void fl_hw_update_stats(struct tcf_proto *tp, struct cls_fl_filter *f,
 	cls_flower.cookie = (unsigned long) f;
 	cls_flower.classid = f->res.classid;
 
-	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
 
 	tcf_exts_stats_update(&f->exts, cls_flower.stats.bytes,
 			      cls_flower.stats.pkts,
@@ -1844,21 +1839,16 @@ static int fl_reoffload(struct tcf_proto *tp, bool add, flow_setup_cb_t *cb,
 
 		cls_flower.classid = f->res.classid;
 
-		err = cb(TC_SETUP_CLSFLOWER, &cls_flower, cb_priv);
+		err = tc_setup_cb_reoffload(block, tp, add, cb,
+					    TC_SETUP_CLSFLOWER, &cls_flower,
+					    cb_priv, &f->flags,
+					    &f->in_hw_count);
 		kfree(cls_flower.rule);
 
 		if (err) {
-			if (add && tc_skip_sw(f->flags)) {
-				__fl_put(f);
-				return err;
-			}
-			goto next_flow;
+			__fl_put(f);
+			return err;
 		}
-
-		spin_lock(&tp->lock);
-		tc_cls_offload_cnt_update(block, &f->in_hw_count, &f->flags,
-					  add);
-		spin_unlock(&tp->lock);
 next_flow:
 		__fl_put(f);
 	}
@@ -1886,7 +1876,7 @@ static int fl_hw_create_tmplt(struct tcf_chain *chain,
 	/* We don't care if driver (any of them) fails to handle this
 	 * call. It serves just as a hint for it.
 	 */
-	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
 	kfree(cls_flower.rule);
 
 	return 0;
@@ -1902,7 +1892,7 @@ static void fl_hw_destroy_tmplt(struct tcf_chain *chain,
 	cls_flower.command = FLOW_CLS_TMPLT_DESTROY;
 	cls_flower.cookie = (unsigned long) tmplt;
 
-	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false);
+	tc_setup_cb_call(block, TC_SETUP_CLSFLOWER, &cls_flower, false, true);
 }
 
 static void *fl_tmplt_create(struct net *net, struct tcf_chain *chain,