[syzbot] KMSAN: uninit-value in can_send

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 4a3e741a3d6a x86: fortify: kmsan: fix KMSAN fortify builds
git tree: https://github.com/google/kmsan.git master
console+strace: https://syzkaller.appspot.com/x/log.txt?x=14247636880000
kernel config: https://syzkaller.appspot.com/x/.config?x=c19210a0c25eebb
dashboard link: https://syzkaller.appspot.com/bug?extid=d168ec0caca4697e03b1
compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project.git 610139d2d9ce6746b3c617fb3e2f7886272d26ff), GNU ld (GNU Binutils for Debian) 2.35.2
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=13e16e86880000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1535a2f6880000

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/fbb1997bc1e0/disk-4a3e741a.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/d5dd2e1efaa4/vmlinux-4a3e741a.xz

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+d168ec...@syzkaller.appspotmail.com

=====================================================
BUG: KMSAN: uninit-value in can_is_canxl_skb include/linux/can/skb.h:128 [inline]
BUG: KMSAN: uninit-value in can_send+0x269/0x1100 net/can/af_can.c:205
can_is_canxl_skb include/linux/can/skb.h:128 [inline]
can_send+0x269/0x1100 net/can/af_can.c:205
j1939_send_one+0x40f/0x4d0 net/can/j1939/main.c:352
j1939_xtp_do_tx_ctl+0x69f/0x9e0 net/can/j1939/transport.c:664
j1939_tp_tx_ctl net/can/j1939/transport.c:672 [inline]
j1939_session_tx_rts net/can/j1939/transport.c:740 [inline]
j1939_xtp_txnext_transmiter net/can/j1939/transport.c:880 [inline]
j1939_tp_txtimer+0x35bb/0x4520 net/can/j1939/transport.c:1158
__run_hrtimer+0x298/0x910 kernel/time/hrtimer.c:1685
__hrtimer_run_queues kernel/time/hrtimer.c:1749 [inline]
hrtimer_run_softirq+0x4b0/0x870 kernel/time/hrtimer.c:1766
__do_softirq+0x1c5/0x7b9 kernel/softirq.c:571
invoke_softirq+0x8f/0x100 kernel/softirq.c:445
__irq_exit_rcu+0x5a/0x110 kernel/softirq.c:650
irq_exit_rcu+0xe/0x10 kernel/softirq.c:662
sysvec_apic_timer_interrupt+0x9a/0xc0 arch/x86/kernel/apic/apic.c:1107
asm_sysvec_apic_timer_interrupt+0x1b/0x20 arch/x86/include/asm/idtentry.h:649
__raw_spin_unlock_irqrestore include/linux/spinlock_api_smp.h:152 [inline]
_raw_spin_unlock_irqrestore+0x2f/0x50 kernel/locking/spinlock.c:194
unlock_hrtimer_base kernel/time/hrtimer.c:1017 [inline]
hrtimer_start_range_ns+0xaba/0xb50 kernel/time/hrtimer.c:1301
hrtimer_start include/linux/hrtimer.h:418 [inline]
j1939_tp_schedule_txtimer+0xbe/0x100 net/can/j1939/transport.c:697
j1939_sk_send_loop net/can/j1939/socket.c:1143 [inline]
j1939_sk_sendmsg+0x1c2c/0x25d0 net/can/j1939/socket.c:1256
sock_sendmsg_nosec net/socket.c:714 [inline]
sock_sendmsg net/socket.c:734 [inline]
____sys_sendmsg+0xa8e/0xe70 net/socket.c:2482
___sys_sendmsg+0x2a1/0x3f0 net/socket.c:2536
__sys_sendmsg net/socket.c:2565 [inline]
__do_sys_sendmsg net/socket.c:2574 [inline]
__se_sys_sendmsg net/socket.c:2572 [inline]
__x64_sys_sendmsg+0x367/0x540 net/socket.c:2572
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd

Uninit was created at:
slab_post_alloc_hook mm/slab.h:742 [inline]
slab_alloc_node mm/slub.c:3398 [inline]
__kmem_cache_alloc_node+0x6ee/0xc90 mm/slub.c:3437
__do_kmalloc_node mm/slab_common.c:954 [inline]
__kmalloc_node_track_caller+0x117/0x3d0 mm/slab_common.c:975
kmalloc_reserve net/core/skbuff.c:437 [inline]
__alloc_skb+0x34a/0xca0 net/core/skbuff.c:509
alloc_skb include/linux/skbuff.h:1267 [inline]
j1939_tp_tx_dat_new net/can/j1939/transport.c:593 [inline]
j1939_xtp_do_tx_ctl+0xa3/0x9e0 net/can/j1939/transport.c:654
j1939_tp_tx_ctl net/can/j1939/transport.c:672 [inline]
j1939_session_tx_rts net/can/j1939/transport.c:740 [inline]
j1939_xtp_txnext_transmiter net/can/j1939/transport.c:880 [inline]
j1939_tp_txtimer+0x35bb/0x4520 net/can/j1939/transport.c:1158
__run_hrtimer+0x298/0x910 kernel/time/hrtimer.c:1685
__hrtimer_run_queues kernel/time/hrtimer.c:1749 [inline]
hrtimer_run_softirq+0x4b0/0x870 kernel/time/hrtimer.c:1766
__do_softirq+0x1c5/0x7b9 kernel/softirq.c:571

CPU: 0 PID: 3506 Comm: syz-executor289 Not tainted 6.1.0-rc2-syzkaller-61955-g4a3e741a3d6a #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/11/2022
=====================================================


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
syzbot can test patches for this issue, for details see:
https://goo.gl/tpsmEJ#testing-patches

[Oliver Hartkopp]

Good catch!

That's a correct issue caused by only filling the CAN frame header with
can_id and len in j1939_send_one():

https://elixir.bootlin.com/linux/v6.0.6/source/net/can/j1939/main.c#L345

@Oleksij: Can you please (zero)initialize the unused elements in struct
can_frame (namely __pad, __res0, len8_dlc) in j1939_send_one()?

Or do you want me to create a patch for it?

Thanks and best regards,
Oliver

[Oleksij Rempel]

Hi Oliver,

On Thu, Nov 03, 2022 at 10:21:15PM +0100, Oliver Hartkopp wrote:
> Good catch!
>
> That's a correct issue caused by only filling the CAN frame header with
> can_id and len in j1939_send_one():
>
> https://elixir.bootlin.com/linux/v6.0.6/source/net/can/j1939/main.c#L345
>
> @Oleksij: Can you please (zero)initialize the unused elements in struct
> can_frame (namely __pad, __res0, len8_dlc) in j1939_send_one()?
>
> Or do you want me to create a patch for it?

I'll not be able to make it today. I'll be glad if you could do it

Thanks and best regards,
Oleksij

--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |