linux_dsm_epyc7002/net/sched
Vlad Buslov e3ae1f96ac net: sched: sch_sfb: don't call qdisc_put() while holding tree lock
Recent changes that removed rtnl dependency from rules update path of tc
also made tcf_block_put() function sleeping. This function is called from
ops->destroy() of several Qdisc implementations, which in turn is called by
qdisc_put(). Some Qdiscs call qdisc_put() while holding sch tree spinlock,
which results sleeping-while-atomic BUG.

Steps to reproduce for sfb:

tc qdisc add dev ens1f0 handle 1: root sfb
tc qdisc add dev ens1f0 parent 1:10 handle 50: sfq perturb 10
tc qdisc change dev ens1f0 root handle 1: sfb

Resulting dmesg:

[ 7265.938717] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:909
[ 7265.940152] in_atomic(): 1, irqs_disabled(): 0, pid: 28579, name: tc
[ 7265.941455] INFO: lockdep is turned off.
[ 7265.942744] CPU: 11 PID: 28579 Comm: tc Tainted: G        W         5.3.0-rc8+ #721
[ 7265.944065] Hardware name: Supermicro SYS-2028TP-DECR/X10DRT-P, BIOS 2.0b 03/30/2017
[ 7265.945396] Call Trace:
[ 7265.946709]  dump_stack+0x85/0xc0
[ 7265.947994]  ___might_sleep.cold+0xac/0xbc
[ 7265.949282]  __mutex_lock+0x5b/0x960
[ 7265.950543]  ? tcf_chain0_head_change_cb_del.isra.0+0x1b/0xf0
[ 7265.951803]  ? tcf_chain0_head_change_cb_del.isra.0+0x1b/0xf0
[ 7265.953022]  tcf_chain0_head_change_cb_del.isra.0+0x1b/0xf0
[ 7265.954248]  tcf_block_put_ext.part.0+0x21/0x50
[ 7265.955478]  tcf_block_put+0x50/0x70
[ 7265.956694]  sfq_destroy+0x15/0x50 [sch_sfq]
[ 7265.957898]  qdisc_destroy+0x5f/0x160
[ 7265.959099]  sfb_change+0x175/0x330 [sch_sfb]
[ 7265.960304]  tc_modify_qdisc+0x324/0x840
[ 7265.961503]  rtnetlink_rcv_msg+0x170/0x4b0
[ 7265.962692]  ? netlink_deliver_tap+0x95/0x400
[ 7265.963876]  ? rtnl_dellink+0x2d0/0x2d0
[ 7265.965064]  netlink_rcv_skb+0x49/0x110
[ 7265.966251]  netlink_unicast+0x171/0x200
[ 7265.967427]  netlink_sendmsg+0x224/0x3f0
[ 7265.968595]  sock_sendmsg+0x5e/0x60
[ 7265.969753]  ___sys_sendmsg+0x2ae/0x330
[ 7265.970916]  ? ___sys_recvmsg+0x159/0x1f0
[ 7265.972074]  ? do_wp_page+0x9c/0x790
[ 7265.973233]  ? __handle_mm_fault+0xcd3/0x19e0
[ 7265.974407]  __sys_sendmsg+0x59/0xa0
[ 7265.975591]  do_syscall_64+0x5c/0xb0
[ 7265.976753]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 7265.977938] RIP: 0033:0x7f229069f7b8
[ 7265.979117] Code: 89 02 48 c7 c0 ff ff ff ff eb bb 0f 1f 80 00 00 00 00 f3 0f 1e fa 48 8d 05 65 8f 0c 00 8b 00 85 c0 75 17 b8 2e 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 58 c3 0f 1f 80 00 00 00 00 48 83 ec 28 89 5
4
[ 7265.981681] RSP: 002b:00007ffd7ed2d158 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
[ 7265.983001] RAX: ffffffffffffffda RBX: 000000005d813ca1 RCX: 00007f229069f7b8
[ 7265.984336] RDX: 0000000000000000 RSI: 00007ffd7ed2d1c0 RDI: 0000000000000003
[ 7265.985682] RBP: 0000000000000000 R08: 0000000000000001 R09: 000000000165c9a0
[ 7265.987021] R10: 0000000000404eda R11: 0000000000000246 R12: 0000000000000001
[ 7265.988309] R13: 000000000047f640 R14: 0000000000000000 R15: 0000000000000000

In sfb_change() function use qdisc_purge_queue() instead of
qdisc_tree_flush_backlog() to properly reset old child Qdisc and save
pointer to it into local temporary variable. Put reference to Qdisc after
sch tree lock is released in order not to call potentially sleeping cls API
in atomic section. This is safe to do because Qdisc has already been reset
by qdisc_purge_queue() inside sch tree lock critical section.

Reported-by: syzbot+ac54455281db908c581e@syzkaller.appspotmail.com
Fixes: c266f64dbf ("net: sched: protect block state with mutex")
Suggested-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2019-09-27 12:13:55 +02:00
..
act_api.c net_sched: add policy validation for action attributes 2019-09-21 19:33:13 -07:00
act_bpf.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_connmark.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_csum.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_ct.c netfilter: update include directives. 2019-09-13 12:33:06 +02:00
act_ctinfo.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_gact.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_ife.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_ipt.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_meta_mark.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
act_meta_skbprio.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
act_meta_skbtcindex.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
act_mirred.c net: sched: use get_dev() action API in flow_action infra 2019-09-16 09:18:03 +02:00
act_mpls.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_nat.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_pedit.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_police.c net_sched: act_police: add 2 new attributes to support police 64bit rate and peakrate 2019-09-06 15:02:16 +02:00
act_sample.c net/sched: act_sample: don't push mac header on ip6gre ingress 2019-09-20 17:01:59 -07:00
act_simple.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_skbedit.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_skbmod.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_tunnel_key.c net_sched: fix a NULL pointer deref in ipt action 2019-08-27 15:05:58 -07:00
act_vlan.c Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net 2019-09-02 11:20:17 -07:00
cls_api.c net: sched: fix possible crash in tcf_action_destroy() 2019-09-24 16:33:57 +02:00
cls_basic.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_bpf.c net: sched: refactor block offloads counter usage 2019-08-26 14:17:43 -07:00
cls_cgroup.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_flow.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_flower.c net: sched: flower: don't take rtnl lock for cls hw offloads API 2019-08-26 14:17:43 -07:00
cls_fw.c net: sched: remove NET_CLS_IND config option 2019-06-15 14:06:13 -07:00
cls_matchall.c net: sched: cls_matchall: cleanup flow_action before deallocating 2019-08-30 15:12:05 -07:00
cls_route.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_rsvp6.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_rsvp.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_rsvp.h treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
cls_tcindex.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
cls_u32.c net: sched: refactor block offloads counter usage 2019-08-26 14:17:43 -07:00
em_canid.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 11 2019-05-21 11:28:45 +02:00
em_cmp.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
em_ipset.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
em_ipt.c net: sched: em_ipt: add support for addrtype matching 2019-06-29 11:15:12 -07:00
em_meta.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
em_nbyte.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
em_text.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
em_u32.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
ematch.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
Kconfig net: Fix Kconfig indentation 2019-09-26 08:56:17 +02:00
Makefile net/sched: Introduce action ct 2019-07-09 12:11:59 -07:00
sch_api.c net_sched: add max len check for TCA_KIND 2019-09-21 19:18:51 -07:00
sch_atm.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
sch_blackhole.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_cake.c netlink: make validation more configurable for future strictness 2019-04-27 17:07:21 -04:00
sch_cbq.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_cbs.c net/sched: cbs: Fix not adding cbs instance to list 2019-09-26 09:03:03 +02:00
sch_choke.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
sch_codel.c net: sched: Fix a possible null-pointer dereference in dequeue_func() 2019-07-29 09:46:58 -07:00
sch_drr.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
sch_dsmark.c treewide: Add SPDX license identifier for more missed files 2019-05-21 10:50:45 +02:00
sch_etf.c etf: Add skip_sock_check 2019-06-28 14:45:33 -07:00
sch_fifo.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_fq_codel.c fq_codel: remove set but not used variables 'prev_ecn_mark' and 'prev_drop_count' 2019-08-08 22:32:19 -07:00
sch_fq.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_generic.c net_sched: let qdisc_put() accept NULL pointer 2019-09-15 20:54:14 +02:00
sch_gred.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_hfsc.c netlink: make validation more configurable for future strictness 2019-04-27 17:07:21 -04:00
sch_hhf.c sch_hhf: ensure quantum and hhf_non_hh_weight are non-zero 2019-09-10 18:31:00 +01:00
sch_htb.c net: sched: sch_htb: don't call qdisc_put() while holding tree lock 2019-09-27 12:13:55 +02:00
sch_ingress.c net: flow_offload: rename TCF_BLOCK_BINDER_TYPE_* to FLOW_BLOCK_BINDER_TYPE_* 2019-07-09 14:38:50 -07:00
sch_mq.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
sch_mqprio.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
sch_multiq.c net: sched: multiq: don't call qdisc_put() while holding tree lock 2019-09-27 12:13:55 +02:00
sch_netem.c sch_netem: fix rcu splat in netem_enqueue() 2019-09-27 10:29:11 +02:00
sch_pie.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 235 2019-06-19 17:09:07 +02:00
sch_plug.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_prio.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_qfq.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500 2019-06-19 17:09:55 +02:00
sch_red.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_sfb.c net: sched: sch_sfb: don't call qdisc_put() while holding tree lock 2019-09-27 12:13:55 +02:00
sch_sfq.c net_sched: unset TCQ_F_CAN_BYPASS when adding filters 2019-07-17 13:34:09 -07:00
sch_skbprio.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_taprio.c taprio: Add support for hardware offloading 2019-09-16 21:32:57 +02:00
sch_tbf.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00
sch_teql.c treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 152 2019-05-30 11:26:32 -07:00