netdev - Re: [REGRESSION][PATCH] bpf_jit drops the ball on indirect negative mem references

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAPwCz7_tnjq32duJSDj21LpRe5TFvGGszH4joQr1DGTjNd7GcA@mail.gmail.com>
Date:	Wed, 28 Mar 2012 22:26:26 +0200
From:	Jan Seiffert <kaffeemonster@...glemail.com>
To:	Eric Dumazet <eric.dumazet@...il.com>
Cc:	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	"David S. Miller" <davem@...emloft.net>,
	Matt Evans <matt@...abs.org>
Subject: Re: [REGRESSION][PATCH] bpf_jit drops the ball on indirect negative
 mem references

2012/3/28 Eric Dumazet <eric.dumazet@...il.com>:
> On Wed, 2012-03-28 at 21:15 +0200, Jan Seiffert wrote:
>> Consider the following test program:
>>
>> #include <stdio.h>
>> #include <sys/types.h>
>> #include <sys/socket.h>
>> #include <netinet/in.h>
>> #include <pcap-bpf.h>
>>
>> #define die(x) do {perror(x); return 1;} while (0)
>> struct bpf_insn udp_filter[] = {
>>       /*   0 */ BPF_STMT(BPF_LDX|BPF_W|BPF_IMM, -1048576+(0)), /* leax        net[0] */
>>       /*   1 */ BPF_STMT(BPF_LD|BPF_B|BPF_IND, 0),             /* ldb [x+0] */
>>       /*   2 */ BPF_STMT(BPF_RET|BPF_A, 0),                    /* ret a */
>> };
>
> When this point was raised some weeks ago, we wanted to see a _real_ use
> of negative mem reference.
>
> You provide a test program but what this filter is supposed to do
> exactly ?
>

Say you have a UDP socket, and you want to filter for bogus source
addresses (drop already in kernel to save the context switch).
To have only one bpf program for ipv4 and ipv6 (you have to checked
the same bogus v4 addresses in mapped space), there is a point where
it elegant to have a negative offset saved in the X register.
This is how i found the Bug.

But use case or not, the jits behavior is different than the interpreter.

Example (not the complete program):

struct bpf_insn UDPPValid[] = {
        /*   0 */ BPF_STMT(BPF_LDX|BPF_W|BPF_IMM, -1048576+(12)),
        /*   1 */ BPF_STMT(BPF_LD|BPF_B|BPF_ABS, -1048576+(0)),
        /*   2 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xf0),
        /*   3 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x40, 22 - 4, 0),
        /*   4 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x60, 5 - 5, 39 - 5),
        /*   5 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0, 12 - 6, 0),
        /*   6 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x20010DB8, 39 - 7, 0),
        /*   7 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x20010002, 18 - 8, 0),
        /*   8 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xfffffff0),
        /*   9 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x20010010, 39 - 10, 0),
        /*  10 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xff000000),
        /*  11 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xff000000, 39 - 12, 37 - 12),
        /*  12 */ BPF_STMT(BPF_LD|BPF_W|BPF_ABS, -1048576+(12)),
        /*  13 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0, 0, 37 - 14),
        /*  14 */ BPF_STMT(BPF_LD|BPF_W|BPF_ABS, -1048576+(16)),
        /*  15 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xffff, 21 - 16, 0),
        /*  16 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0x0064FF9B, 21 - 17, 0),
        /*  17 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0, 39 - 18, 37 -
18),
        /*  18 */ BPF_STMT(BPF_LD|BPF_W|BPF_ABS, -1048576+(12)),
        /*  19 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xffff0000),
        /*  20 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0, 39 - 21, 37 -
21),
        /*  21 */ BPF_STMT(BPF_LDX|BPF_W|BPF_IMM, -1048576+(20)),
        /*  22 */ BPF_STMT(BPF_LD|BPF_W|BPF_IND, 0),
        /*  23 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xffffffff, 39 - 24,
0),
        /*  24 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xffffff00),
        /*  25 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xC0000000, 39 - 26,
0),
        /*  26 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xC0000200, 39 - 27,
0),
        /*  27 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xC6336400, 39 - 28,
0),
        /*  28 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xCB007100, 39 - 29,
0),
        /*  29 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xC0586300, 39 - 30,
0),
        /*  30 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xfffe0000),
        /*  31 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xC6120000, 39 - 32,
0),
        /*  32 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xff000000),
        /*  33 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0, 39 - 34, 0),
        /*  34 */ BPF_STMT(BPF_ALU|BPF_AND|BPF_K, 0xf0000000),
        /*  35 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xE0000000, 39 - 36,
0),
        /*  36 */ BPF_JUMP(BPF_JMP|BPF_JEQ|BPF_K, 0xF0000000, 39 - 37,
0),
        /*  37 */ BPF_STMT(BPF_LD|BPF_W|BPF_LEN, 0),
        /*  38 */ BPF_STMT(BPF_RET|BPF_A, 0),
        /*  39 */ BPF_STMT(BPF_RET|BPF_K, 0),
};

>
>

Greetings
Jan

-- 
˙qɐɥ ʇɟnɐʞǝƃ ʎɐqǝ ıǝq ɹnʇɐʇsɐʇ ǝuıǝ ɹıɯ ɥɔı sɐp lɐɯ ǝʇzʇǝl sɐp ʇsı sɐp
'ʇɯɯɐpɹǝʌ
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html