From: Pavel Tikhomirov ptikhomirov@virtuozzo.com
stable inclusion from stable-v4.19.187 commit a9a7488979a7d0abfd30bc32f370b8e0c124db33 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I5JPWC CVE: NA
--------------------------------
commit 1ffbc7ea91606e4abd10eb60de5367f1c86daf5e upstream.
Reproduce:
modprobe sch_teql tc qdisc add dev teql0 root teql0
This leads to (for instance in Centos 7 VM) OOPS:
[ 532.366633] BUG: unable to handle kernel NULL pointer dereference at 00000000000000a8 [ 532.366733] IP: [<ffffffffc06124a8>] teql_destroy+0x18/0x100 [sch_teql] [ 532.366825] PGD 80000001376d5067 PUD 137e37067 PMD 0 [ 532.366906] Oops: 0000 [#1] SMP [ 532.366987] Modules linked in: sch_teql ... [ 532.367945] CPU: 1 PID: 3026 Comm: tc Kdump: loaded Tainted: G ------------ T 3.10.0-1062.7.1.el7.x86_64 #1 [ 532.368041] Hardware name: Virtuozzo KVM, BIOS 1.11.0-2.vz7.2 04/01/2014 [ 532.368125] task: ffff8b7d37d31070 ti: ffff8b7c9fdbc000 task.ti: ffff8b7c9fdbc000 [ 532.368224] RIP: 0010:[<ffffffffc06124a8>] [<ffffffffc06124a8>] teql_destroy+0x18/0x100 [sch_teql] [ 532.368320] RSP: 0018:ffff8b7c9fdbf8e0 EFLAGS: 00010286 [ 532.368394] RAX: ffffffffc0612490 RBX: ffff8b7cb1565e00 RCX: ffff8b7d35ba2000 [ 532.368476] RDX: ffff8b7d35ba2000 RSI: 0000000000000000 RDI: ffff8b7cb1565e00 [ 532.368557] RBP: ffff8b7c9fdbf8f8 R08: ffff8b7d3fd1f140 R09: ffff8b7d3b001600 [ 532.368638] R10: ffff8b7d3b001600 R11: ffffffff84c7d65b R12: 00000000ffffffd8 [ 532.368719] R13: 0000000000008000 R14: ffff8b7d35ba2000 R15: ffff8b7c9fdbf9a8 [ 532.368800] FS: 00007f6a4e872740(0000) GS:ffff8b7d3fd00000(0000) knlGS:0000000000000000 [ 532.368885] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 532.368961] CR2: 00000000000000a8 CR3: 00000001396ee000 CR4: 00000000000206e0 [ 532.369046] Call Trace: [ 532.369159] [<ffffffff84c8192e>] qdisc_create+0x36e/0x450 [ 532.369268] [<ffffffff846a9b49>] ? ns_capable+0x29/0x50 [ 532.369366] [<ffffffff849afde2>] ? nla_parse+0x32/0x120 [ 532.369442] [<ffffffff84c81b4c>] tc_modify_qdisc+0x13c/0x610 [ 532.371508] [<ffffffff84c693e7>] rtnetlink_rcv_msg+0xa7/0x260 [ 532.372668] [<ffffffff84907b65>] ? sock_has_perm+0x75/0x90 [ 532.373790] [<ffffffff84c69340>] ? rtnl_newlink+0x890/0x890 [ 532.374914] [<ffffffff84c8da7b>] netlink_rcv_skb+0xab/0xc0 [ 532.376055] [<ffffffff84c63708>] rtnetlink_rcv+0x28/0x30 [ 532.377204] [<ffffffff84c8d400>] netlink_unicast+0x170/0x210 [ 532.378333] [<ffffffff84c8d7a8>] netlink_sendmsg+0x308/0x420 [ 532.379465] [<ffffffff84c2f3a6>] sock_sendmsg+0xb6/0xf0 [ 532.380710] [<ffffffffc034a56e>] ? __xfs_filemap_fault+0x8e/0x1d0 [xfs] [ 532.381868] [<ffffffffc034a75c>] ? xfs_filemap_fault+0x2c/0x30 [xfs] [ 532.383037] [<ffffffff847ec23a>] ? __do_fault.isra.61+0x8a/0x100 [ 532.384144] [<ffffffff84c30269>] ___sys_sendmsg+0x3e9/0x400 [ 532.385268] [<ffffffff847f3fad>] ? handle_mm_fault+0x39d/0x9b0 [ 532.386387] [<ffffffff84d88678>] ? __do_page_fault+0x238/0x500 [ 532.387472] [<ffffffff84c31921>] __sys_sendmsg+0x51/0x90 [ 532.388560] [<ffffffff84c31972>] SyS_sendmsg+0x12/0x20 [ 532.389636] [<ffffffff84d8dede>] system_call_fastpath+0x25/0x2a [ 532.390704] [<ffffffff84d8de21>] ? system_call_after_swapgs+0xae/0x146 [ 532.391753] Code: 00 00 00 00 00 00 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 66 66 66 66 90 55 48 89 e5 41 55 41 54 53 48 8b b7 48 01 00 00 48 89 fb <48> 8b 8e a8 00 00 00 48 85 c9 74 43 48 89 ca eb 0f 0f 1f 80 00 [ 532.394036] RIP [<ffffffffc06124a8>] teql_destroy+0x18/0x100 [sch_teql] [ 532.395127] RSP <ffff8b7c9fdbf8e0> [ 532.396179] CR2: 00000000000000a8
Null pointer dereference happens on master->slaves dereference in teql_destroy() as master is null-pointer.
When qdisc_create() calls teql_qdisc_init() it imediately fails after check "if (m->dev == dev)" because both devices are teql0, and it does not set qdisc_priv(sch)->m leaving it zero on error path, then qdisc_create() imediately calls teql_destroy() which does not expect zero master pointer and we get OOPS.
Fixes: 87b60cfacf9f ("net_sched: fix error recovery at qdisc creation") Signed-off-by: Pavel Tikhomirov ptikhomirov@virtuozzo.com Reviewed-by: Eric Dumazet edumazet@google.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org Signed-off-by: Lu Wei luwei32@huawei.com Reviewed-by: Yue Haibing yuehaibing@huawei.com Reviewed-by: Wei Yongjun weiyongjun1@huawei.com Signed-off-by: Yongqiang Liu liuyongqiang13@huawei.com --- net/sched/sch_teql.c | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/net/sched/sch_teql.c b/net/sched/sch_teql.c index 7ed53124069c..163364124691 100644 --- a/net/sched/sch_teql.c +++ b/net/sched/sch_teql.c @@ -138,6 +138,9 @@ teql_destroy(struct Qdisc *sch) struct teql_sched_data *dat = qdisc_priv(sch); struct teql_master *master = dat->m;
+ if (!master) + return; + prev = master->slaves; if (prev) { do {
From: Eric Dumazet edumazet@google.com
stable inclusion from stable-v4.19.207 commit 7c113506163a1ec6157927428eddd80038d2916e category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I5JPWX CVE: NA
--------------------------------
[ Upstream commit c7c5e6ff533fe1f9afef7d2fa46678987a1335a7 ]
syzbot found that forcing a big quantum attribute would crash hosts fast, essentially using this:
tc qd replace dev eth0 root fq_codel quantum 4294967295
This is because fq_codel_dequeue() would have to loop ~2^31 times in :
if (flow->deficit <= 0) { flow->deficit += q->quantum; list_move_tail(&flow->flowchain, &q->old_flows); goto begin; }
SFQ max quantum is 2^19 (half a megabyte) Lets adopt a max quantum of one megabyte for FQ_CODEL.
Fixes: 4b549a2ef4be ("fq_codel: Fair Queue Codel AQM") Signed-off-by: Eric Dumazet edumazet@google.com Reported-by: syzbot syzkaller@googlegroups.com Signed-off-by: David S. Miller davem@davemloft.net Signed-off-by: Sasha Levin sashal@kernel.org Signed-off-by: Lu Wei luwei32@huawei.com Reviewed-by: Yue Haibing yuehaibing@huawei.com Reviewed-by: Wei Yongjun weiyongjun1@huawei.com Signed-off-by: Yongqiang Liu liuyongqiang13@huawei.com --- include/uapi/linux/pkt_sched.h | 2 ++ net/sched/sch_fq_codel.c | 12 ++++++++++-- 2 files changed, 12 insertions(+), 2 deletions(-)
diff --git a/include/uapi/linux/pkt_sched.h b/include/uapi/linux/pkt_sched.h index 8975fd1a1421..24ce2aab8946 100644 --- a/include/uapi/linux/pkt_sched.h +++ b/include/uapi/linux/pkt_sched.h @@ -779,6 +779,8 @@ struct tc_codel_xstats {
/* FQ_CODEL */
+#define FQ_CODEL_QUANTUM_MAX (1 << 20) + enum { TCA_FQ_CODEL_UNSPEC, TCA_FQ_CODEL_TARGET, diff --git a/net/sched/sch_fq_codel.c b/net/sched/sch_fq_codel.c index a862d9990be7..e4f69c779b8c 100644 --- a/net/sched/sch_fq_codel.c +++ b/net/sched/sch_fq_codel.c @@ -382,6 +382,7 @@ static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt, { struct fq_codel_sched_data *q = qdisc_priv(sch); struct nlattr *tb[TCA_FQ_CODEL_MAX + 1]; + u32 quantum = 0; int err;
if (!opt) @@ -399,6 +400,13 @@ static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt, q->flows_cnt > 65536) return -EINVAL; } + if (tb[TCA_FQ_CODEL_QUANTUM]) { + quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM])); + if (quantum > FQ_CODEL_QUANTUM_MAX) { + NL_SET_ERR_MSG(extack, "Invalid quantum"); + return -EINVAL; + } + } sch_tree_lock(sch);
if (tb[TCA_FQ_CODEL_TARGET]) { @@ -425,8 +433,8 @@ static int fq_codel_change(struct Qdisc *sch, struct nlattr *opt, if (tb[TCA_FQ_CODEL_ECN]) q->cparams.ecn = !!nla_get_u32(tb[TCA_FQ_CODEL_ECN]);
- if (tb[TCA_FQ_CODEL_QUANTUM]) - q->quantum = max(256U, nla_get_u32(tb[TCA_FQ_CODEL_QUANTUM])); + if (quantum) + q->quantum = quantum;
if (tb[TCA_FQ_CODEL_DROP_BATCH_SIZE]) q->drop_batch_size = max(1U, nla_get_u32(tb[TCA_FQ_CODEL_DROP_BATCH_SIZE]));
From: Eric Dumazet edumazet@google.com
stable inclusion from stable-v4.19.240 commit acff67727dff061f107def4c8c08bacde20d0d40 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I5JNFK CVE: NA
--------------------------------
[ Upstream commit ec5b0f605b105457f257f2870acad4a5d463984b ]
While investigating a related syzbot report, I found that whenever call to tcf_exts_init() from u32_init_knode() is failing, we end up with an elevated refcount on ht->refcnt
To avoid that, only increase the refcount after all possible errors have been evaluated.
Fixes: b9a24bb76bf6 ("net_sched: properly handle failure case of tcf_exts_init()") Signed-off-by: Eric Dumazet edumazet@google.com Cc: Cong Wang xiyou.wangcong@gmail.com Cc: Jiri Pirko jiri@resnulli.us Acked-by: Jamal Hadi Salim jhs@mojatatu.com Signed-off-by: Jakub Kicinski kuba@kernel.org Signed-off-by: Sasha Levin sashal@kernel.org Signed-off-by: Lu Wei luwei32@huawei.com Reviewed-by: Yue Haibing yuehaibing@huawei.com Reviewed-by: Wei Yongjun weiyongjun1@huawei.com Signed-off-by: Yongqiang Liu liuyongqiang13@huawei.com --- net/sched/cls_u32.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/net/sched/cls_u32.c b/net/sched/cls_u32.c index 205c194bdca0..23d95db99121 100644 --- a/net/sched/cls_u32.c +++ b/net/sched/cls_u32.c @@ -877,10 +877,6 @@ static struct tc_u_knode *u32_init_knode(struct tcf_proto *tp, new->flags = n->flags; RCU_INIT_POINTER(new->ht_down, ht);
- /* bump reference count as long as we hold pointer to structure */ - if (ht) - ht->refcnt++; - #ifdef CONFIG_CLS_U32_PERF /* Statistics may be incremented by readers during update * so we must keep them in tact. When the node is later destroyed @@ -903,6 +899,10 @@ static struct tc_u_knode *u32_init_knode(struct tcf_proto *tp, return NULL; }
+ /* bump reference count as long as we hold pointer to structure */ + if (ht) + ht->refcnt++; + return new; }
-
Yongqiang Liu