Merge tag 'net-5.17-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from netfilter, bpf.

  Quite a handful of old regression fixes but most of those are
  pre-5.16.

  Current release - regressions:

   - fix memory leaks in the skb free deferral scheme if upper layer
     protocols are used, i.e. in-kernel TCP readers like TLS

  Current release - new code bugs:

   - nf_tables: fix NULL check typo in _clone() functions

   - change the default to y for Vertexcom vendor Kconfig

   - a couple of fixes to incorrect uses of ref tracking

   - two fixes for constifying netdev->dev_addr

  Previous releases - regressions:

   - bpf:
      - various verifier fixes mainly around register offset handling
        when passed to helper functions
      - fix mount source displayed for bpffs (none -> bpffs)

   - bonding:
      - fix extraction of ports for connection hash calculation
      - fix bond_xmit_broadcast return value when some devices are down

   - phy: marvell: add Marvell specific PHY loopback

   - sch_api: don't skip qdisc attach on ingress, prevent ref leak

   - htb: restore minimal packet size handling in rate control

   - sfp: fix high power modules without diagnostic monitoring

   - mscc: ocelot:
      - don't let phylink re-enable TX PAUSE on the NPI port
      - don't dereference NULL pointers with shared tc filters

   - smsc95xx: correct reset handling for LAN9514

   - cpsw: avoid alignment faults by taking NET_IP_ALIGN into account

   - phy: micrel: use kszphy_suspend/_resume for irq aware devices,
     avoid races with the interrupt

  Previous releases - always broken:

   - xdp: check prog type before updating BPF link

   - smc: resolve various races around abnormal connection termination

   - sit: allow encapsulated IPv6 traffic to be delivered locally

   - axienet: fix init/reset handling, add missing barriers, read the
     right status words, stop queues correctly

   - add missing dev_put() in sock_timestamping_bind_phc()

  Misc:

   - ipv4: prevent accidentally passing RTO_ONLINK to
     ip_route_output_key_hash() by sanitizing flags

   - ipv4: avoid quadratic behavior in netns dismantle

   - stmmac: dwmac-oxnas: add support for OX810SE

   - fsl: xgmac_mdio: add workaround for erratum A-009885"

* tag 'net-5.17-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (92 commits)
  ipv4: add net_hash_mix() dispersion to fib_info_laddrhash keys
  ipv4: avoid quadratic behavior in netns dismantle
  net/fsl: xgmac_mdio: Fix incorrect iounmap when removing module
  powerpc/fsl/dts: Enable WA for erratum A-009885 on fman3l MDIO buses
  dt-bindings: net: Document fsl,erratum-a009885
  net/fsl: xgmac_mdio: Add workaround for erratum A-009885
  net: mscc: ocelot: fix using match before it is set
  net: phy: micrel: use kszphy_suspend()/kszphy_resume for irq aware devices
  net: cpsw: avoid alignment faults by taking NET_IP_ALIGN into account
  nfc: llcp: fix NULL error pointer dereference on sendmsg() after failed bind()
  net: axienet: increase default TX ring size to 128
  net: axienet: fix for TX busy handling
  net: axienet: fix number of TX ring slots for available check
  net: axienet: Fix TX ring slot available check
  net: axienet: limit minimum TX ring size
  net: axienet: add missing memory barriers
  net: axienet: reset core on initialization prior to MDIO access
  net: axienet: Wait for PhyRstCmplt after core reset
  net: axienet: increase reset timeout
  bpf, selftests: Add ringbuf memory type confusion test
  ...

6 files changed, 28 insertions, 8 deletions

diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index 6e947cd91152..fa517ae604ad 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -316,7 +316,12 @@ enum bpf_type_flag {
 	 */
 	MEM_RDONLY		= BIT(1 + BPF_BASE_TYPE_BITS),
 
-	__BPF_TYPE_LAST_FLAG	= MEM_RDONLY,
+	/* MEM was "allocated" from a different helper, and cannot be mixed
+	 * with regular non-MEM_ALLOC'ed MEM types.
+	 */
+	MEM_ALLOC		= BIT(2 + BPF_BASE_TYPE_BITS),
+
+	__BPF_TYPE_LAST_FLAG	= MEM_ALLOC,
 };
 
 /* Max number of base types. */
@@ -400,7 +405,7 @@ enum bpf_return_type {
 	RET_PTR_TO_SOCKET_OR_NULL	= PTR_MAYBE_NULL | RET_PTR_TO_SOCKET,
 	RET_PTR_TO_TCP_SOCK_OR_NULL	= PTR_MAYBE_NULL | RET_PTR_TO_TCP_SOCK,
 	RET_PTR_TO_SOCK_COMMON_OR_NULL	= PTR_MAYBE_NULL | RET_PTR_TO_SOCK_COMMON,
-	RET_PTR_TO_ALLOC_MEM_OR_NULL	= PTR_MAYBE_NULL | RET_PTR_TO_ALLOC_MEM,
+	RET_PTR_TO_ALLOC_MEM_OR_NULL	= PTR_MAYBE_NULL | MEM_ALLOC | RET_PTR_TO_ALLOC_MEM,
 	RET_PTR_TO_BTF_ID_OR_NULL	= PTR_MAYBE_NULL | RET_PTR_TO_BTF_ID,
 
 	/* This must be the last entry. Its purpose is to ensure the enum is
diff --git a/include/linux/bpf_verifier.h b/include/linux/bpf_verifier.h
index 143401d4c9d9..e9993172f892 100644
--- a/include/linux/bpf_verifier.h
+++ b/include/linux/bpf_verifier.h
@@ -519,8 +519,8 @@ bpf_prog_offload_replace_insn(struct bpf_verifier_env *env, u32 off,
 void
 bpf_prog_offload_remove_insns(struct bpf_verifier_env *env, u32 off, u32 cnt);
 
-int check_ctx_reg(struct bpf_verifier_env *env,
-		  const struct bpf_reg_state *reg, int regno);
+int check_ptr_off_reg(struct bpf_verifier_env *env,
+		      const struct bpf_reg_state *reg, int regno);
 int check_mem_reg(struct bpf_verifier_env *env, struct bpf_reg_state *reg,
 		   u32 regno, u32 mem_size);
 
diff --git a/include/net/inet_frag.h b/include/net/inet_frag.h
index 48cc5795ceda..63540be0fc34 100644
--- a/include/net/inet_frag.h
+++ b/include/net/inet_frag.h
@@ -117,8 +117,15 @@ int fqdir_init(struct fqdir **fqdirp, struct inet_frags *f, struct net *net);
 
 static inline void fqdir_pre_exit(struct fqdir *fqdir)
 {
-	fqdir->high_thresh = 0; /* prevent creation of new frags */
-	fqdir->dead = true;
+	/* Prevent creation of new frags.
+	 * Pairs with READ_ONCE() in inet_frag_find().
+	 */
+	WRITE_ONCE(fqdir->high_thresh, 0);
+
+	/* Pairs with READ_ONCE() in inet_frag_kill(), ip_expire()
+	 * and ip6frag_expire_frag_queue().
+	 */
+	WRITE_ONCE(fqdir->dead, true);
 }
 void fqdir_exit(struct fqdir *fqdir);
 
diff --git a/include/net/ipv6_frag.h b/include/net/ipv6_frag.h
index 851029ecff13..0a4779175a52 100644
--- a/include/net/ipv6_frag.h
+++ b/include/net/ipv6_frag.h
@@ -67,7 +67,8 @@ ip6frag_expire_frag_queue(struct net *net, struct frag_queue *fq)
 	struct sk_buff *head;
 
 	rcu_read_lock();
-	if (fq->q.fqdir->dead)
+	/* Paired with the WRITE_ONCE() in fqdir_pre_exit(). */
+	if (READ_ONCE(fq->q.fqdir->dead))
 		goto out_rcu_unlock;
 	spin_lock(&fq->q.lock);
 
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index ebef45e821af..676cb8ea9e15 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -218,8 +218,10 @@ static inline int tcf_exts_init(struct tcf_exts *exts, struct net *net,
 #ifdef CONFIG_NET_CLS_ACT
 	exts->type = 0;
 	exts->nr_actions = 0;
+	/* Note: we do not own yet a reference on net.
+	 * This reference might be taken later from tcf_exts_get_net().
+	 */
 	exts->net = net;
-	netns_tracker_alloc(net, &exts->ns_tracker, GFP_KERNEL);
 	exts->actions = kcalloc(TCA_ACT_MAX_PRIO, sizeof(struct tc_action *),
 				GFP_KERNEL);
 	if (!exts->actions)
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index c11dbac5abb2..472843eedbae 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -1244,6 +1244,7 @@ struct psched_ratecfg {
 	u64	rate_bytes_ps; /* bytes per second */
 	u32	mult;
 	u16	overhead;
+	u16	mpu;
 	u8	linklayer;
 	u8	shift;
 };
@@ -1253,6 +1254,9 @@ static inline u64 psched_l2t_ns(const struct psched_ratecfg *r,
 {
 	len += r->overhead;
 
+	if (len < r->mpu)
+		len = r->mpu;
+
 	if (unlikely(r->linklayer == TC_LINKLAYER_ATM))
 		return ((u64)(DIV_ROUND_UP(len,48)*53) * r->mult) >> r->shift;
 
@@ -1275,6 +1279,7 @@ static inline void psched_ratecfg_getrate(struct tc_ratespec *res,
 	res->rate = min_t(u64, r->rate_bytes_ps, ~0U);
 
 	res->overhead = r->overhead;
+	res->mpu = r->mpu;
 	res->linklayer = (r->linklayer & TC_LINKLAYER_MASK);
 }