Incorrect memory view after running self-modifying code

[wonderkun]

*Describe the bug

I notice that there are some bugs when unicorn emulates self-modifying code. So qiling-framework run shellcode with SMC will crash.

Here is a example in this issue https://github.com/unicorn-engine/unicorn/issues/820 .

I found a fix for this issue https://github.com/alxchk/unicorn/commit/a195b312141850d9a7fd892625db2bbec242ff7c, but this commit isn't merged to unicorn which version used by qiling framework. And I don't know whether this commit fix this problem.

Sample Code

ql = Qiling(["x8664_freebsd/bin/x8664_hello_asm"], "x8664_freebsd", output = "dump")
ql.run()

Expected behavior A clear and concise description of what you expected to happen.

Screenshots If applicable, add screenshots to help explain your problem.

Additional context Add any other context about the problem here.

[xwings]

we need @aquynh to merge it.

@wonderkun since this is a unicorn issue, can u provide and test with latest unicorn 1.0.2 to confirm this issue.

[wonderkun]

first test

Yes, Let's test the issue https://github.com/unicorn-engine/unicorn/issues/820 first.

python: 3.8.5 unicorn: 1.0.2

Run the code as bellow:

import unicorn
import unicorn.x86_const as x86

sc = bytes.fromhex(
    "dbd0d97424f45fb8e67741bc31c9b15831471a03471a83c704e2138ba93edb742"
    "a5f52911b5f00d10c6f43b7a004012c32688d43f3c7eb6a047bcfed868603ceb7"
    "48560fffb59a5da8b20872dd8f90f9ad1e901e6520b1b0fd7b1132d1f7182c363"
    "dd3c78cc9e201dd32486cd1c091a8d63ae4c024c6fe16561c8b8cf0d72b69003b"
    "adfa0ef0baa512076fde2f8c8e31a6d6b495e28dd58c4e63eacf30dc4e9bdd09e"
    "3c689a39e8c4954170424cd83bef47a0d38fa50609d5708d1720bc6ef22d2b1f0"
    "1e77ed64a22b4210ffda64e0175064e0e7460ca6d7ad862648a641aff7f0917a8"
    "e3b3eec91f12168c2a6f227b61e9d2c6db1664d5b5bf2bb3b0c8388c3cc0a0ea9"
    "c85ca43187344d08b94352419618ff394ff7d2bb777cd31102425e90423679cca"
    "c0ddb5bb2bb7124244495a4b42c95a4f4acc6ccac08bbe9b284a8a11fae2912c8"
    "b0959d08e283f51a92a2e4e44f31286ebdb2ae8efe4170e5e511b2590ed4cb993"
    "1614311fda3c8b573d46323595f0c85c5fe98bc05"
)

def main():
    uc = unicorn.Uc(unicorn.UC_ARCH_X86, unicorn.UC_MODE_32)
    uc.mem_map(0x1000, 0x2000)
    uc.mem_write(0x1000, sc)
    uc.reg_write(x86.UC_X86_REG_ESP, 0x2000)
    uc.emu_start(0x1000, 0, count=0x166)
    out = uc.mem_read(0x1000, 0x2000)

    for x in range(len(sc)):
        print("0x%08x: 0x%02x => 0x%02x" % (x, sc[x],out[x]))
    print(str(out))

if __name__ == "__main__":
    main()

the output is the same as issue https://github.com/unicorn-engine/unicorn/issues/820

0x0000000f: 0x58 => 0x58
0x00000010: 0x31 => 0x31
0x00000011: 0x47 => 0x47
0x00000012: 0x1a => 0x1a
0x00000013: 0x03 => 0x03
0x00000014: 0x47 => 0x47
0x00000015: 0x1a => 0x1a
0x00000016: 0x83 => 0x83
0x00000017: 0xc7 => 0xc7
0x00000018: 0x04 => 0x04
0x00000019: 0xe2 => 0xe2  ; the loop instruction
0x0000001a: 0x13 => 0xf5  ; the correct immediate after decoding
0x0000001b: 0x8b => 0x8b  ; definitely not the `cld` instruction (see `x64dbg` screenshot)
0x0000001c: 0xa9 => 0xa9  ; definitely not a `call` instruction
0x0000001d: 0x3e => 0x3e
0x0000001e: 0xdb => 0x00  ; interestingly enough this byte is correct!
0x0000001f: 0x74 => 0x77  ; ^ that's the first byte of the 2nd xor
0x00000020: 0x2a => 0xc1
0x00000021: 0x5f => 0xa5
0x00000022: 0x52 => 0x89  ; and so is this one (first byte, 3rd xor)
0x00000023: 0x91 => 0xeb
0x00000024: 0x1b => 0xb7

second test

Here is a self-modify shellcode

The real asm code dump from memory is like this:

let's run it in unicorn.

import unicorn
import unicorn.x86_const as x86

sc = bytes.fromhex(
            'bb1c577352d9cbd97424f45e2bc9b13a83c604315e10035e10fea2a899437'
            '50f1aa85ffb063a072f8e738bacf88fdddaee6c760dedbe7b7e1871cd1512'
            'db0c1a84d38fbcbcfe46673ddcd74fc0e1a21bff31234b8a52df3d1e823a2'
            'a96360381fb0d38818232246c9e4e037805ccb144263da113512d546a2943'
            '9a0b661895c947be24fbb8e30928bd805e3e1057998c1280977d382cbf248'
            '02fe04d6745f3b2896164ace56c5db05a4975a618b822883e2773102c5ce7'
            'c7d5fe6cbd54fa5815db17894753f9989842fba365fe29c48be52e75c93e2'
            '834091d18c39340cab8dd3033eae87fe460bd1d55d775310e76ebc9d0676d'
            '9583c3845f7bd9f64784e2b6c1'
)

# print(sc)

def main():
    uc = unicorn.Uc(unicorn.UC_ARCH_X86, unicorn.UC_MODE_32)
    uc.mem_map(0x1000, 0x2000)
    uc.mem_write(0x1000, sc)
    uc.reg_write(x86.UC_X86_REG_ESP, 0x2000)
    uc.emu_start(0x1000,0,count=0x1b)
    out = uc.mem_read(0x1000, 0x2000)

    for x in range(len(sc)):
        print("0x%08x: 0x%02x => 0x%02x" % (x+0x08048054, sc[x],out[x]))
    print(str(out))

if __name__ == "__main__":
    main()

the output is not correct:

0x08048054: 0xbb => 0xbb
0x08048055: 0x1c => 0x1c
0x08048056: 0x57 => 0x57
0x08048057: 0x73 => 0x73
0x08048058: 0x52 => 0x52
0x08048059: 0xd9 => 0xd9
0x0804805a: 0xcb => 0xcb
0x0804805b: 0xd9 => 0xd9
0x0804805c: 0x74 => 0x74
0x0804805d: 0x24 => 0x24
0x0804805e: 0xf4 => 0xf4
0x0804805f: 0x5e => 0x5e
0x08048060: 0x2b => 0x2b
0x08048061: 0xc9 => 0xc9
0x08048062: 0xb1 => 0xb1
0x08048063: 0x3a => 0x3a
0x08048064: 0x83 => 0x83
0x08048065: 0xc6 => 0xc6
0x08048066: 0x04 => 0x04
0x08048067: 0x31 => 0x31
0x08048068: 0x5e => 0x5e
0x08048069: 0x10 => 0x10
0x0804806a: 0x03 => 0x03
0x0804806b: 0x5e => 0x5e
0x0804806c: 0x10 => 0x10
0x0804806d: 0xfe => 0xe2  // loop 
0x0804806e: 0xa2 => 0xf5  
0x0804806f: 0xa8 => 0xa8   // error decode result ! 
0x08048070: 0x99 => 0x99  // error decode result !  ....
0x08048071: 0x43 => 0xbd
0x08048072: 0x75 => 0x39
0x08048073: 0x0f => 0x13
0x08048074: 0x1a => 0xf6
0x08048075: 0xa8 => 0x13
0x08048076: 0x5f => 0xd9
0x08048077: 0xfb => 0xd4
0x08048078: 0x06 => 0xe4
0x08048079: 0x3a => 0xf4
0x0804807a: 0x07 => 0x58
0x0804807b: 0x2f => 0x2b
0x0804807c: 0x8e => 0x49
0x0804807d: 0x73 => 0xb1
0x0804807e: 0x8b => 0x33
0x0804807f: 0xac => 0x83
0x08048080: 0xf8 => 0xe8
0x08048081: 0x8f => 0x8f
0x08048082: 0xdd => 0xdd
0x08048083: 0xda => 0xda
0x08048084: 0xee => 0xee
0x08048085: 0x6c => 0x6c
0x08048086: 0x76 => 0x76
0x08048087: 0x0d => 0x0d
0x08048088: 0xed => 0xed
0x08048089: 0xbe => 0xbe
0x0804808a: 0x7b => 0x7b
0x0804808b: 0x7e => 0x7e
0x0804808c: 0x18 => 0x18
0x0804808d: 0x71 => 0x71
0x0804808e: 0xcd => 0xcd
0x0804808f: 0x15 => 0x15
0x08048090: 0x12 => 0x12
0x08048091: 0xdb => 0xdb
0x08048092: 0x0c => 0x0c

@aquynh @xwings

[wonderkun]

@xwings Anyone pay attention to this issue? I am supposed to fix it with your help.

[xwings]

If this is a issue with Unicorn (very likely it is)

1. the issue should be raise in Unicorn, community there will take a look into it.
2. Since this is a open source project, will be be able to raise a PR in Unicorn ?
3. Test and confirm the fix the for the PR u mention is working. and get @aquynh to merge it ?

Finally, I cannot fix Unicorn :(

[wonderkun]

Thank you very much! I will send a PR to unicron.