| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CACkBjsaFJwjC5oiw-1KXvcazywodwXo4zGYsRHwbr2gSG9WcSw@mail.gmail.com>
Date: Fri, 9 Dec 2022 15:56:53 +0800
From: Hao Sun <sunhao.th@...il.com>
To: bpf <bpf@...r.kernel.org>
Cc: Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
John Fastabend <john.fastabend@...il.com>,
Andrii Nakryiko <andrii@...nel.org>,
Martin KaFai Lau <martin.lau@...ux.dev>,
Song Liu <song@...nel.org>, Yonghong Song <yhs@...com>,
KP Singh <kpsingh@...nel.org>,
Stanislav Fomichev <sdf@...gle.com>,
Hao Luo <haoluo@...gle.com>, Jiri Olsa <jolsa@...nel.org>,
David Miller <davem@...emloft.net>,
Jakub Kicinski <kuba@...nel.org>, hawk@...nel.org,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: BUG: missing null marking in PTR_TO_BTF_ID leads to null-ptr-deref in
BPF prog
Hi,
I constructed a BPF prog that can pass the verifier but contains null-ptr-deref.
This is achieved by exploiting the following behavior in the verifier:
1. verifier allows direct struct bpf_map access but does not mark the
pointer (PTR_TO_BTF_ID) as maybe_null
2. verifier allows pointer CMP in JMP, and marks reg as non-null
if one of the pointers is not null in the corresponding path.
The following crash can be reproduced on:
HEAD commit: 2d14123617f9 Merge branch 'Document some recent core
kfunc additions'
git tree: bpf-next
console log: https://pastebin.com/raw/UkZXqdqs
kernel config: https://pastebin.com/raw/BbzHXhkV
POC: https://pastebin.com/raw/XFHVa3xV
Essentially, the program in POC creates a MAP_HASH, then loads and run
the following BPF prog:
BPF_LD_MAP_FD(BPF_REG_1, map_fd),
// verifier believe R6 is ptr_to_btf_id, after this ldx
// but R6 = bpf_map->inner_map_meta, which is NULL,
// because we don't have inner map
BPF_LDX_MEM(BPF_DW, BPF_REG_6, BPF_REG_1, 8),
BPF_MOV64_REG(BPF_REG_2, BPF_REG_10),
BPF_ALU64_IMM(BPF_ADD, BPF_REG_2, -4),
BPF_ST_MEM(BPF_W, BPF_REG_2, 0, 0),
BPF_CALL_FUNC(BPF_FUNC_map_lookup_elem),
// verifier think R0 is map_value_or_null, R6 is ptr_btf_id
// at runtime, they all equal to NULL
BPF_JMP_REG(BPF_JEQ, BPF_REG_6, BPF_REG_0, 1),
BPF_EXIT_INSN(),
// here, verifier think R0 this map_value, but it is NULL
BPF_LDX_MEM(BPF_W, BPF_REG_0, BPF_REG_0, 0),
BPF_EXIT_INSN(),
After loading and running prog, the following is reported:
general protection fault, probably for non-canonical address
0xdffffc0000000000: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000000-0x0000000000000007]
CPU: 7 PID: 6781 Comm: a.out Not tainted 6.1.0-rc7-01517-g2d14123617f9 #125
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Arch Linux
1.16.1-1-1 04/01/2014
RIP: 0010:___bpf_prog_run+0x66d/0x8fd0 kernel/bpf/core.c:1937
Code: e0 07 83 c0 01 38 d0 7c 08 84 d2 0f 85 e6 78 00 00 48 b8 00 00
00 00 00 fc ff df 4d 0f bf 66 02 4d 01 ec 4c 89 e2 48 c1 ea 03 <0f> b6
14 02 4c 89 e0 83 e0 07 83 c0 03 38 d0 7c 08 84 d2 0f 85 a7
RSP: 0018:ffffc900043377e8 EFLAGS: 00010256
RAX: dffffc0000000000 RBX: dffffc0000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: ffffffff818efa7c RDI: ffffc90001146092
RBP: ffffc900043378a0 R08: 0000000000000006 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: ffffc90001146090 R15: 0000000000000000
FS: 00007f68c01f6440(0000) GS:ffff88823bd80000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000563424ffa008 CR3: 000000001fd2a000 CR4: 0000000000750ee0
PKRU: 55555554
Call Trace:
<TASK>
__bpf_prog_run32+0x9d/0xe0 kernel/bpf/core.c:2045
bpf_dispatcher_nop_func include/linux/bpf.h:1082 [inline]
__bpf_prog_run include/linux/filter.h:600 [inline]
bpf_prog_run include/linux/filter.h:607 [inline]
bpf_test_run+0x38e/0x980 net/bpf/test_run.c:402
bpf_prog_test_run_skb+0xb69/0x1dd0 net/bpf/test_run.c:1187
bpf_prog_test_run kernel/bpf/syscall.c:3644 [inline]
__sys_bpf+0x1293/0x5840 kernel/bpf/syscall.c:4997
__do_sys_bpf kernel/bpf/syscall.c:5083 [inline]
__se_sys_bpf kernel/bpf/syscall.c:5081 [inline]
__x64_sys_bpf+0x78/0xc0 kernel/bpf/syscall.c:5081
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x38/0xb0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
RIP: 0033:0x7f68bfae4469
Code: 00 f3 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 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 8b 0d ff 49 2b 00 f7 d8 64 89 01 48
RSP: 002b:00007fff992ab838 EFLAGS: 00000246 ORIG_RAX: 0000000000000141
RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f68bfae4469
RDX: 0000000000000090 RSI: 00007fff992ac940 RDI: 000000000000000a
RBP: 00007fff992ac9e0 R08: 000000000000000a R09: 0000000300001118
R10: 0000000000000001 R11: 0000000000000246 R12: 00005634238006f0
R13: 00007fff992acac0 R14: 0000000000000000 R15: 0000000000000000
</TASK>
Modules linked in:
---[ end trace 0000000000000000 ]---
RIP: 0010:___bpf_prog_run+0x66d/0x8fd0 kernel/bpf/core.c:1937
Code: e0 07 83 c0 01 38 d0 7c 08 84 d2 0f 85 e6 78 00 00 48 b8 00 00
00 00 00 fc ff df 4d 0f bf 66 02 4d 01 ec 4c 89 e2 48 c1 ea 03 <0f> b6
14 02 4c 89 e0 83 e0 07 83 c0 03 38 d0 7c 08 84 d2 0f 85 a7
RSP: 0018:ffffc900043377e8 EFLAGS: 00010256
RAX: dffffc0000000000 RBX: dffffc0000000000 RCX: 0000000000000000
RDX: 0000000000000000 RSI: ffffffff818efa7c RDI: ffffc90001146092
RBP: ffffc900043378a0 R08: 0000000000000006 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
R13: 0000000000000000 R14: ffffc90001146090 R15: 0000000000000000
FS: 00007f68c01f6440(0000) GS:ffff88823bd80000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000563424ffa008 CR3: 000000001fd2a000 CR4: 0000000000750ee0
PKRU: 55555554
----------------
Code disassembly (best guess):
0: e0 07 loopne 0x9
2: 83 c0 01 add $0x1,%eax
5: 38 d0 cmp %dl,%al
7: 7c 08 jl 0x11
9: 84 d2 test %dl,%dl
b: 0f 85 e6 78 00 00 jne 0x78f7
11: 48 b8 00 00 00 00 00 movabs $0xdffffc0000000000,%rax
18: fc ff df
1b: 4d 0f bf 66 02 movswq 0x2(%r14),%r12
20: 4d 01 ec add %r13,%r12
23: 4c 89 e2 mov %r12,%rdx
26: 48 c1 ea 03 shr $0x3,%rdx
* 2a: 0f b6 14 02 movzbl (%rdx,%rax,1),%edx <-- trapping instruction
2e: 4c 89 e0 mov %r12,%rax
31: 83 e0 07 and $0x7,%eax
34: 83 c0 03 add $0x3,%eax
37: 38 d0 cmp %dl,%al
39: 7c 08 jl 0x43
3b: 84 d2 test %dl,%dl
3d: 0f .byte 0xf
3e: 85 .byte 0x85
3f: a7 cmpsl %es:(%rdi),%ds:(%rsi)
Powered by blists - more mailing lists