Re: [syzbot] [net?] memory leak in tcp_md5_do_add
From: Eric Dumazet
Date: Thu Sep 21 2023 - 23:21:02 EST
On Fri, Sep 22, 2023 at 1:15 AM Dmitry Safonov <dima@xxxxxxxxxx> wrote:
>
> Hi Eric,
>
> On 9/21/23 18:01, Dmitry Safonov wrote:
> > On 9/21/23 17:59, Eric Dumazet wrote:
> >> On Thu, Sep 21, 2023 at 6:56 PM syzbot
> >> <syzbot+68662811b3d5f6695bcb@xxxxxxxxxxxxxxxxxxxxxxxxx> wrote:
> >>>
> >>> Hello,
> >>>
> >>> syzbot found the following issue on:
> >>>
> >>> HEAD commit: ee3f96b16468 Merge tag 'nfsd-6.3-1' of git://git.kernel.or..
> >>> git tree: upstream
> >>> console output: https://syzkaller.appspot.com/x/log.txt?x=1312bba8c80000
> >>> kernel config: https://syzkaller.appspot.com/x/.config?x=f5733ca1757172ad
> >>> dashboard link: https://syzkaller.appspot.com/bug?extid=68662811b3d5f6695bcb
> >>> compiler: gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2
> >>> syz repro: https://syzkaller.appspot.com/x/repro.syz?x=105393a8c80000
> >>> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1113917f480000
> >>>
> >>> Downloadable assets:
> >>> disk image: https://storage.googleapis.com/syzbot-assets/29e7966ab711/disk-ee3f96b1.raw.xz
> >>> vmlinux: https://storage.googleapis.com/syzbot-assets/ae21b8e855de/vmlinux-ee3f96b1.xz
> >>> kernel image: https://storage.googleapis.com/syzbot-assets/803ee0425ad6/bzImage-ee3f96b1.xz
> >>>
> >>> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> >>> Reported-by: syzbot+68662811b3d5f6695bcb@xxxxxxxxxxxxxxxxxxxxxxxxx
> >>>
> >>> executing program
> >>> BUG: memory leak
> >>> unreferenced object 0xffff88810a86f7a0 (size 32):
> >>> comm "syz-executor325", pid 5099, jiffies 4294978342 (age 119.240s)
> >>> hex dump (first 32 bytes):
> >>> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> >>> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> >>> backtrace:
> >>> [<ffffffff81533d64>] kmalloc_trace+0x24/0x90 mm/slab_common.c:1061
> >>> [<ffffffff840edaa0>] kmalloc include/linux/slab.h:580 [inline]
> >>> [<ffffffff840edaa0>] tcp_md5sig_info_add net/ipv4/tcp_ipv4.c:1169 [inline]
> >>> [<ffffffff840edaa0>] tcp_md5_do_add+0xa0/0x150 net/ipv4/tcp_ipv4.c:1240
> >>> [<ffffffff84262c73>] tcp_v6_parse_md5_keys+0x253/0x4a0 net/ipv6/tcp_ipv6.c:671
> >>> [<ffffffff840c720e>] do_tcp_setsockopt+0x40e/0x1360 net/ipv4/tcp.c:3720
> >>> [<ffffffff840c81fb>] tcp_setsockopt+0x9b/0xa0 net/ipv4/tcp.c:3806
> >>> [<ffffffff83d72a8b>] __sys_setsockopt+0x1ab/0x330 net/socket.c:2274
> >>> [<ffffffff83d72c36>] __do_sys_setsockopt net/socket.c:2285 [inline]
> >>> [<ffffffff83d72c36>] __se_sys_setsockopt net/socket.c:2282 [inline]
> >>> [<ffffffff83d72c36>] __x64_sys_setsockopt+0x26/0x30 net/socket.c:2282
> >>> [<ffffffff849ad699>] do_syscall_x64 arch/x86/entry/common.c:50 [inline]
> >>> [<ffffffff849ad699>] do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
> >>> [<ffffffff84a0008b>] entry_SYSCALL_64_after_hwframe+0x63/0xcd
> >>>
> >>> BUG: memory leak
> >>> unreferenced object 0xffff88811225ccc0 (size 192):
> >>> comm "syz-executor325", pid 5099, jiffies 4294978342 (age 119.240s)
> >>> hex dump (first 32 bytes):
> >>> 00 00 00 00 00 00 00 00 22 01 00 00 00 00 ad de ........".......
> >>> 22 0a 80 00 fe 80 00 00 00 00 00 00 00 00 00 00 "...............
> >>> backtrace:
> >>> [<ffffffff8153444a>] __do_kmalloc_node mm/slab_common.c:966 [inline]
> >>> [<ffffffff8153444a>] __kmalloc+0x4a/0x120 mm/slab_common.c:980
> >>> [<ffffffff83d75c15>] kmalloc include/linux/slab.h:584 [inline]
> >>> [<ffffffff83d75c15>] sock_kmalloc net/core/sock.c:2635 [inline]
> >>> [<ffffffff83d75c15>] sock_kmalloc+0x65/0xa0 net/core/sock.c:2624
> >>> [<ffffffff840eb9bb>] __tcp_md5_do_add+0xcb/0x300 net/ipv4/tcp_ipv4.c:1212
> >>> [<ffffffff840eda67>] tcp_md5_do_add+0x67/0x150 net/ipv4/tcp_ipv4.c:1253
> >>> [<ffffffff84262c73>] tcp_v6_parse_md5_keys+0x253/0x4a0 net/ipv6/tcp_ipv6.c:671
> >>> [<ffffffff840c720e>] do_tcp_setsockopt+0x40e/0x1360 net/ipv4/tcp.c:3720
> >>> [<ffffffff840c81fb>] tcp_setsockopt+0x9b/0xa0 net/ipv4/tcp.c:3806
> >>> [<ffffffff83d72a8b>] __sys_setsockopt+0x1ab/0x330 net/socket.c:2274
> >>> [<ffffffff83d72c36>] __do_sys_setsockopt net/socket.c:2285 [inline]
> >>> [<ffffffff83d72c36>] __se_sys_setsockopt net/socket.c:2282 [inline]
> >>> [<ffffffff83d72c36>] __x64_sys_setsockopt+0x26/0x30 net/socket.c:2282
> >>> [<ffffffff849ad699>] do_syscall_x64 arch/x86/entry/common.c:50 [inline]
> >>> [<ffffffff849ad699>] do_syscall_64+0x39/0xb0 arch/x86/entry/common.c:80
> >>> [<ffffffff84a0008b>] entry_SYSCALL_64_after_hwframe+0x63/0xcd
> >>>
> >>>
> >>>
> >>> ---
> >>> 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 syzkaller@xxxxxxxxxxxxxxxx.
> >>>
> >>> syzbot will keep track of this issue. See:
> >>> https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
> >>>
> >>> If the bug is already fixed, let syzbot know by replying with:
> >>> #syz fix: exact-commit-title
> >>>
> >>> If you want syzbot to run the reproducer, reply with:
> >>> #syz test: git://repo/address.git branch-or-commit-hash
> >>> If you attach or paste a git patch, syzbot will apply it before testing.
> >>>
> >>> If you want to overwrite bug's subsystems, reply with:
> >>> #syz set subsystems: new-subsystem
> >>> (See the list of subsystem names on the web dashboard)
> >>>
> >>> If the bug is a duplicate of another bug, reply with:
> >>> #syz dup: exact-subject-of-another-report
> >>>
> >>> If you want to undo deduplication, reply with:
> >>> #syz undup
> >>
> >> Dmitry, please take a look at this bug, we need to fix it before your
> >> patch series.
> >
> > Sure, seems reasonable to me to fix before merging something on top.
>
> It seems to me that it's related to a race between RCU grace period and
> kmemleak scan period. There seems to be a patch [1] that likely fixes
> that, albeit I couldn't verify it as all my attempts to reproduce syzbot
> issue produced only unrelated to TCP-MD5 log:
>
I doubt this, looking at the repro, which seems to abuse a not often
used feature of TCP (self connect)
# https://syzkaller.appspot.com/bug?id=323165b5fe193114de7a3a6a8bd16cf3a3c36ecf
# See https://goo.gl/kgGztJ for information about syzkaller reproducers.
#{"repeat":true,"procs":1,"slowdown":1,"sandbox":"none","sandbox_arg":0,"leak":true,"netdev":true,"close_fds":true,"usb":true}
r0 = socket$inet6_tcp(0xa, 0x1, 0x0)
setsockopt$inet6_tcp_TCP_MD5SIG(r0, 0x6, 0xe,
&(0x7f0000000040)={@in6={{0xa, 0x0, 0x0, @local}}, 0x0, 0x0, 0x22,
0x0, "b05423587c18814d6b1a5f25671d09815a4687d637ffc958defc671aad3d4de8ac7d88560c759d600ab650c07ef0ef162b199da0d017fe6f0ae40cfb4e241cf9a990f20f6b8c2c070a61cfad8a2d2600"},
0xd8)
connect$inet6(r0, &(0x7f0000000180)={0xa, 0x4001, 0x0, @ipv4={'\x00',
'\xff\xff', @remote}}, 0x1c)
dup(0xffffffffffffffff)
setsockopt$SO_BINDTODEVICE(r0, 0x1, 0x19,
&(0x7f00000001c0)='ip6_vti0\x00', 0xff4a)
You could not have KMEMLEAK in the kernel, and run the repro a thousand times.
Then compare /proc/slabinfo before/after.