linux.git/net/sched, branch v6.5

net/sched: fix a qdisc modification with ambiguous command request

2023-08-23T08:44:48+00:00

When replacing an existing root qdisc, with one that is of the same kind, the
request boils down to essentially a parameterization change  i.e not one that
requires allocation and grafting of a new qdisc. syzbot was able to create a
scenario which resulted in a taprio qdisc replacing an existing taprio qdisc
with a combination of NLM_F_CREATE, NLM_F_REPLACE and NLM_F_EXCL leading to
create and graft scenario.
The fix ensures that only when the qdisc kinds are different that we should
allow a create and graft, otherwise it goes into the "change" codepath.

While at it, fix the code and comments to improve readability.

While syzbot was able to create the issue, it did not zone on the root cause.
Analysis from Vladimir Oltean  helped narrow it down.

v1->V2 changes:
- remove "inline" function definition (Vladmir)
- remove extrenous braces in branches (Vladmir)
- change inline function names (Pedro)
- Run tdc tests (Victor)
v2->v3 changes:
- dont break else/if (Simon)

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: syzbot+a3618a167af2021433cd@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/20230816225759.g25x76kmgzya2gei@skbuf/T/
Tested-by: Vladimir Oltean 
Tested-by: Victor Nogueira 
Reviewed-by: Pedro Tammela 
Reviewed-by: Victor Nogueira 
Signed-off-by: Jamal Hadi Salim 
Signed-off-by: David S. Miller

net/sched: cls_route: No longer copy tcf_result on update to avoid use-after-free

2023-08-01T03:10:37+00:00

When route4_change() is called on an existing filter, the whole
tcf_result struct is always copied into the new instance of the filter.

This causes a problem when updating a filter bound to a class,
as tcf_unbind_filter() is always called on the old instance in the
success path, decreasing filter_cnt of the still referenced class
and allowing it to be deleted, leading to a use-after-free.

Fix this by no longer copying the tcf_result struct from the old filter.

Fixes: 1109c00547fc ("net: sched: RCU cls_route")
Reported-by: valis 
Reported-by: Bing-Jhong Billy Jheng 
Signed-off-by: valis 
Signed-off-by: Jamal Hadi Salim 
Reviewed-by: Victor Nogueira 
Reviewed-by: Pedro Tammela 
Reviewed-by: M A Ramdhan 
Link: https://lore.kernel.org/r/20230729123202.72406-4-jhs@mojatatu.com
Signed-off-by: Jakub Kicinski

net/sched: cls_fw: No longer copy tcf_result on update to avoid use-after-free

2023-08-01T03:10:36+00:00

When fw_change() is called on an existing filter, the whole
tcf_result struct is always copied into the new instance of the filter.

This causes a problem when updating a filter bound to a class,
as tcf_unbind_filter() is always called on the old instance in the
success path, decreasing filter_cnt of the still referenced class
and allowing it to be deleted, leading to a use-after-free.

Fix this by no longer copying the tcf_result struct from the old filter.

Fixes: e35a8ee5993b ("net: sched: fw use RCU")
Reported-by: valis 
Reported-by: Bing-Jhong Billy Jheng 
Signed-off-by: valis 
Signed-off-by: Jamal Hadi Salim 
Reviewed-by: Victor Nogueira 
Reviewed-by: Pedro Tammela 
Reviewed-by: M A Ramdhan 
Link: https://lore.kernel.org/r/20230729123202.72406-3-jhs@mojatatu.com
Signed-off-by: Jakub Kicinski

net/sched: cls_u32: No longer copy tcf_result on update to avoid use-after-free

2023-08-01T03:10:36+00:00

When u32_change() is called on an existing filter, the whole
tcf_result struct is always copied into the new instance of the filter.

This causes a problem when updating a filter bound to a class,
as tcf_unbind_filter() is always called on the old instance in the
success path, decreasing filter_cnt of the still referenced class
and allowing it to be deleted, leading to a use-after-free.

Fix this by no longer copying the tcf_result struct from the old filter.

Fixes: de5df63228fc ("net: sched: cls_u32 changes to knode must appear atomic to readers")
Reported-by: valis 
Reported-by: M A Ramdhan 
Signed-off-by: valis 
Signed-off-by: Jamal Hadi Salim 
Reviewed-by: Victor Nogueira 
Reviewed-by: Pedro Tammela 
Reviewed-by: M A Ramdhan 
Link: https://lore.kernel.org/r/20230729123202.72406-2-jhs@mojatatu.com
Signed-off-by: Jakub Kicinski

net/sched: taprio: Limit TCA_TAPRIO_ATTR_SCHED_CYCLE_TIME to INT_MAX.

2023-07-31T08:12:27+00:00

syzkaller found zero division error [0] in div_s64_rem() called from
get_cycle_time_elapsed(), where sched->cycle_time is the divisor.

We have tests in parse_taprio_schedule() so that cycle_time will never
be 0, and actually cycle_time is not 0 in get_cycle_time_elapsed().

The problem is that the types of divisor are different; cycle_time is
s64, but the argument of div_s64_rem() is s32.

syzkaller fed this input and 0x100000000 is cast to s32 to be 0.

  @TCA_TAPRIO_ATTR_SCHED_CYCLE_TIME={0xc, 0x8, 0x100000000}

We use s64 for cycle_time to cast it to ktime_t, so let's keep it and
set max for cycle_time.

While at it, we prevent overflow in setup_txtime() and add another
test in parse_taprio_schedule() to check if cycle_time overflows.

Also, we add a new tdc test case for this issue.

[0]:
divide error: 0000 [#1] PREEMPT SMP KASAN NOPTI
CPU: 1 PID: 103 Comm: kworker/1:3 Not tainted 6.5.0-rc1-00330-g60cc1f7d0605 #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Workqueue: ipv6_addrconf addrconf_dad_work
RIP: 0010:div_s64_rem include/linux/math64.h:42 [inline]
RIP: 0010:get_cycle_time_elapsed net/sched/sch_taprio.c:223 [inline]
RIP: 0010:find_entry_to_transmit+0x252/0x7e0 net/sched/sch_taprio.c:344
Code: 3c 02 00 0f 85 5e 05 00 00 48 8b 4c 24 08 4d 8b bd 40 01 00 00 48 8b 7c 24 48 48 89 c8 4c 29 f8 48 63 f7 48 99 48 89 74 24 70 <48> f7 fe 48 29 d1 48 8d 04 0f 49 89 cc 48 89 44 24 20 49 8d 85 10
RSP: 0018:ffffc90000acf260 EFLAGS: 00010206
RAX: 177450e0347560cf RBX: 0000000000000000 RCX: 177450e0347560cf
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000100000000
RBP: 0000000000000056 R08: 0000000000000000 R09: ffffed10020a0934
R10: ffff8880105049a7 R11: ffff88806cf3a520 R12: ffff888010504800
R13: ffff88800c00d800 R14: ffff8880105049a0 R15: 0000000000000000
FS:  0000000000000000(0000) GS:ffff88806cf00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f0edf84f0e8 CR3: 000000000d73c002 CR4: 0000000000770ee0
PKRU: 55555554
Call Trace:
 
 get_packet_txtime net/sched/sch_taprio.c:508 [inline]
 taprio_enqueue_one+0x900/0xff0 net/sched/sch_taprio.c:577
 taprio_enqueue+0x378/0xae0 net/sched/sch_taprio.c:658
 dev_qdisc_enqueue+0x46/0x170 net/core/dev.c:3732
 __dev_xmit_skb net/core/dev.c:3821 [inline]
 __dev_queue_xmit+0x1b2f/0x3000 net/core/dev.c:4169
 dev_queue_xmit include/linux/netdevice.h:3088 [inline]
 neigh_resolve_output net/core/neighbour.c:1552 [inline]
 neigh_resolve_output+0x4a7/0x780 net/core/neighbour.c:1532
 neigh_output include/net/neighbour.h:544 [inline]
 ip6_finish_output2+0x924/0x17d0 net/ipv6/ip6_output.c:135
 __ip6_finish_output+0x620/0xaa0 net/ipv6/ip6_output.c:196
 ip6_finish_output net/ipv6/ip6_output.c:207 [inline]
 NF_HOOK_COND include/linux/netfilter.h:292 [inline]
 ip6_output+0x206/0x410 net/ipv6/ip6_output.c:228
 dst_output include/net/dst.h:458 [inline]
 NF_HOOK.constprop.0+0xea/0x260 include/linux/netfilter.h:303
 ndisc_send_skb+0x872/0xe80 net/ipv6/ndisc.c:508
 ndisc_send_ns+0xb5/0x130 net/ipv6/ndisc.c:666
 addrconf_dad_work+0xc14/0x13f0 net/ipv6/addrconf.c:4175
 process_one_work+0x92c/0x13a0 kernel/workqueue.c:2597
 worker_thread+0x60f/0x1240 kernel/workqueue.c:2748
 kthread+0x2fe/0x3f0 kernel/kthread.c:389
 ret_from_fork+0x2c/0x50 arch/x86/entry/entry_64.S:308
 
Modules linked in:

Fixes: 4cfd5779bd6e ("taprio: Add support for txtime-assist mode")
Reported-by: syzkaller 
Signed-off-by: Kuniyuki Iwashima 
Co-developed-by: Eric Dumazet 
Co-developed-by: Pedro Tammela 
Acked-by: Vinicius Costa Gomes 
Signed-off-by: David S. Miller

net: annotate data-races around sk->sk_{rcv|snd}timeo

2023-07-29T17:13:41+00:00

sk_getsockopt() runs without locks, we must add annotations
to sk->sk_rcvtimeo and sk->sk_sndtimeo.

In the future we might allow fetching these fields before
we lock the socket in TCP fast path.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: sched: cls_u32: Fix match key mis-addressing

2023-07-29T01:05:04+00:00

A match entry is uniquely identified with an "address" or "path" in the
form of: hashtable ID(12b):bucketid(8b):nodeid(12b).

When creating table match entries all of hash table id, bucket id and
node (match entry id) are needed to be either specified by the user or
reasonable in-kernel defaults are used. The in-kernel default for a table id is
0x800(omnipresent root table); for bucketid it is 0x0. Prior to this fix there
was none for a nodeid i.e. the code assumed that the user passed the correct
nodeid and if the user passes a nodeid of 0 (as Mingi Cho did) then that is what
was used. But nodeid of 0 is reserved for identifying the table. This is not
a problem until we dump. The dump code notices that the nodeid is zero and
assumes it is referencing a table and therefore references table struct
tc_u_hnode instead of what was created i.e match entry struct tc_u_knode.

Ming does an equivalent of:
tc filter add dev dummy0 parent 10: prio 1 handle 0x1000 \
protocol ip u32 match ip src 10.0.0.1/32 classid 10:1 action ok

Essentially specifying a table id 0, bucketid 1 and nodeid of zero
Tableid 0 is remapped to the default of 0x800.
Bucketid 1 is ignored and defaults to 0x00.
Nodeid was assumed to be what Ming passed - 0x000

dumping before fix shows:
~$ tc filter ls dev dummy0 parent 10:
filter protocol ip pref 1 u32 chain 0
filter protocol ip pref 1 u32 chain 0 fh 800: ht divisor 1
filter protocol ip pref 1 u32 chain 0 fh 800: ht divisor -30591

Note that the last line reports a table instead of a match entry
(you can tell this because it says "ht divisor...").
As a result of reporting the wrong data type (misinterpretting of struct
tc_u_knode as being struct tc_u_hnode) the divisor is reported with value
of -30591. Ming identified this as part of the heap address
(physmap_base is 0xffff8880 (-30591 - 1)).

The fix is to ensure that when table entry matches are added and no
nodeid is specified (i.e nodeid == 0) then we get the next available
nodeid from the table's pool.

After the fix, this is what the dump shows:
$ tc filter ls dev dummy0 parent 10:
filter protocol ip pref 1 u32 chain 0
filter protocol ip pref 1 u32 chain 0 fh 800: ht divisor 1
filter protocol ip pref 1 u32 chain 0 fh 800::800 order 2048 key ht 800 bkt 0 flowid 10:1 not_in_hw
match 0a000001/ffffffff at 12
action order 1: gact action pass
random type none pass val 0
index 1 ref 1 bind 1

Reported-by: Mingi Cho
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Jamal Hadi Salim
Link: https://lore.kernel.org/r/20230726135151.416917-1-jhs@mojatatu.com
Signed-off-by: Jakub Kicinski

net: flower: fix stack-out-of-bounds in fl_set_key_cfm()

2023-07-28T03:01:29+00:00

Typical misuse of

	nla_parse_nested(array, XXX_MAX, ...);

array must be declared as

	struct nlattr *array[XXX_MAX + 1];

v2: Based on feedbacks from Ido Schimmel and Zahari Doychev,
I also changed TCA_FLOWER_KEY_CFM_OPT_MAX and cfm_opt_policy
definitions.

syzbot reported:

BUG: KASAN: stack-out-of-bounds in __nla_validate_parse+0x136/0x2bd0 lib/nlattr.c:588
Write of size 32 at addr ffffc90003a0ee20 by task syz-executor296/5014

CPU: 0 PID: 5014 Comm: syz-executor296 Not tainted 6.5.0-rc2-syzkaller-00307-gd192f5382581 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/12/2023
Call Trace:

__dump_stack lib/dump_stack.c:88 [inline]
dump_stack_lvl+0x1e7/0x2d0 lib/dump_stack.c:106
print_address_description mm/kasan/report.c:364 [inline]
print_report+0x163/0x540 mm/kasan/report.c:475
kasan_report+0x175/0x1b0 mm/kasan/report.c:588
kasan_check_range+0x27e/0x290 mm/kasan/generic.c:187
__asan_memset+0x23/0x40 mm/kasan/shadow.c:84
__nla_validate_parse+0x136/0x2bd0 lib/nlattr.c:588
__nla_parse+0x40/0x50 lib/nlattr.c:700
nla_parse_nested include/net/netlink.h:1262 [inline]
fl_set_key_cfm+0x1e3/0x440 net/sched/cls_flower.c:1718
fl_set_key+0x2168/0x6620 net/sched/cls_flower.c:1884
fl_tmplt_create+0x1fe/0x510 net/sched/cls_flower.c:2666
tc_chain_tmplt_add net/sched/cls_api.c:2959 [inline]
tc_ctl_chain+0x131d/0x1ac0 net/sched/cls_api.c:3068
rtnetlink_rcv_msg+0x82b/0xf50 net/core/rtnetlink.c:6424
netlink_rcv_skb+0x1df/0x430 net/netlink/af_netlink.c:2549
netlink_unicast_kernel net/netlink/af_netlink.c:1339 [inline]
netlink_unicast+0x7c3/0x990 net/netlink/af_netlink.c:1365
netlink_sendmsg+0xa2a/0xd60 net/netlink/af_netlink.c:1914
sock_sendmsg_nosec net/socket.c:725 [inline]
sock_sendmsg net/socket.c:748 [inline]
____sys_sendmsg+0x592/0x890 net/socket.c:2494
___sys_sendmsg net/socket.c:2548 [inline]
__sys_sendmsg+0x2b0/0x3a0 net/socket.c:2577
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x41/0xc0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f54c6150759
Code: 48 83 c4 28 c3 e8 d7 19 00 00 0f 1f 80 00 00 00 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007ffe06c30578 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
RAX: ffffffffffffffda RBX: 00007f54c619902d RCX: 00007f54c6150759
RDX: 0000000000000000 RSI: 0000000020000280 RDI: 0000000000000003
RBP: 00007ffe06c30590 R08: 0000000000000000 R09: 00007ffe06c305f0
R10: 0000000000000000 R11: 0000000000000246 R12: 00007f54c61c35f0
R13: 00007ffe06c30778 R14: 0000000000000001 R15: 0000000000000001


The buggy address belongs to stack of task syz-executor296/5014
and is located at offset 32 in frame:
fl_set_key_cfm+0x0/0x440 net/sched/cls_flower.c:374

This frame has 1 object:
[32, 56) 'nla_cfm_opt'

The buggy address belongs to the virtual mapping at
[ffffc90003a08000, ffffc90003a11000) created by:
copy_process+0x5c8/0x4290 kernel/fork.c:2330

Fixes: 7cfffd5fed3e ("net: flower: add support for matching cfm fields")
Reported-by: syzbot 
Signed-off-by: Eric Dumazet 
Cc: Simon Horman 
Reviewed-by: Ido Schimmel 
Reviewed-by: Zahari Doychev 
Link: https://lore.kernel.org/r/20230726145815.943910-1-edumazet@google.com
Signed-off-by: Jakub Kicinski

net/sched: mqprio: Add length check for TCA_MQPRIO_{MAX/MIN}_RATE64

2023-07-27T05:08:14+00:00

The nla_for_each_nested parsing in function mqprio_parse_nlattr() does
not check the length of the nested attribute. This can lead to an
out-of-attribute read and allow a malformed nlattr (e.g., length 0) to
be viewed as 8 byte integer and passed to priv->max_rate/min_rate.

This patch adds the check based on nla_len() when check the nla_type(),
which ensures that the length of these two attribute must equals
sizeof(u64).

Fixes: 4e8b86c06269 ("mqprio: Introduce new hardware offload mode and shaper in mqprio")
Reviewed-by: Victor Nogueira 
Signed-off-by: Lin Ma 
Link: https://lore.kernel.org/r/20230725024227.426561-1-linma@zju.edu.cn
Signed-off-by: Jakub Kicinski

net: sched: cls_flower: Undo tcf_bind_filter in case of an error

2023-07-17T06:33:39+00:00

If TCA_FLOWER_CLASSID is specified in the netlink message, the code will
call tcf_bind_filter. However, if any error occurs after that, the code
should undo this by calling tcf_unbind_filter.

Fixes: 77b9900ef53a ("tc: introduce Flower classifier")
Signed-off-by: Victor Nogueira 
Acked-by: Jamal Hadi Salim 
Reviewed-by: Pedro Tammela 
Reviewed-by: Simon Horman 
Signed-off-by: David S. Miller