GHSA-J842-XGM4-WF88

Vulnerability from github – Published: 2025-11-21 18:03 – Updated: 2025-11-21 22:18
VLAI?
Summary
MLX has Wild Pointer Dereference in load_gguf()
Details

Summary

Segmentation fault in mlx::core::load_gguf() when loading malicious GGUF files. Untrusted pointer from external gguflib library is dereferenced without validation, causing application crash.

Environment: - OS: Ubuntu 20.04.6 LTS - Compiler: Clang 19.1.7

Vulnerability

Location: mlx/io/gguf.cpp - Function extract_tensor_data() at lines 59-79 - Vulnerable memcpy at lines 64-67 - Called from load_arrays() at line 177

The Bug:

std::tuple<allocator::Buffer, Dtype> extract_tensor_data(gguf_tensor* tensor) {
  std::optional<Dtype> equivalent_dtype = gguf_type_to_dtype(tensor->type);
  if (equivalent_dtype.has_value()) {
    allocator::Buffer buffer = allocator::malloc(tensor->bsize);
    memcpy(
        buffer.raw_ptr(),
        tensor->weights_data,  // untrusted pointer from gguflib
        tensor->num_weights * equivalent_dtype.value().size());
    return {buffer, equivalent_dtype.value()};
  }
  // ...
}

Possible Fix

std::tuple<allocator::Buffer, Dtype> extract_tensor_data(gguf_tensor* tensor) {
  std::optional<Dtype> equivalent_dtype = gguf_type_to_dtype(tensor->type);
  if (equivalent_dtype.has_value()) {
    // FIX: Validate pointer
    if (!tensor->weights_data) {
      throw std::runtime_error("[load_gguf] NULL tensor data pointer");
    }

    allocator::Buffer buffer = allocator::malloc(tensor->bsize);
    memcpy(
        buffer.raw_ptr(),
        tensor->weights_data,
        tensor->num_weights * equivalent_dtype.value().size());
    return {buffer, equivalent_dtype.value()};
  }
  // ...
}

PoC

# Install MLX
pip install mlx

python3 -c "import mlx.core as mx; mx.load('exploit.gguf', format='gguf')"

Download the poc file there, or let me know how I can send it to you.

AddressSanitizer Output (with instrumented build):

AddressSanitizer:DEADLYSIGNAL
=================================================================
==5855==ERROR: AddressSanitizer: SEGV on unknown address 0x7fc432f64bc0 (pc 0x7fc430841c12 bp 0x7ffc04847ab0 sp 0x7ffc04847268 T0)
==5855==The signal is caused by a READ memory access.
    #0 0x7fc430841c12  /build/glibc-B3wQXB/glibc-2.31/string/../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:312
    #1 0x55aac829756b in __asan_memcpy (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x9ef56b) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)
    #2 0x55aacaa6e8dc in mlx::core::extract_tensor_data(gguf_tensor*) /home/user1/mlx/mlx/io/gguf.cpp:64:5
    #3 0x55aacaa773fc in mlx::core::load_arrays[abi:cxx11](gguf_ctx*) /home/user1/mlx/mlx/io/gguf.cpp:226:35
    #4 0x55aacaa782a9 in mlx::core::load_gguf(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&, std::variant<std::monostate, mlx::core::Stream, mlx::core::Device>) /home/user1/mlx/mlx/io/gguf.cpp:250:17
    #5 0x55aac82dc696 in LLVMFuzzerTestOneInput /home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf.cpp:49:19
    #6 0x55aac81e25c6 in fuzzer::Fuzzer::ExecuteCallback(unsigned char const*, unsigned long) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x93a5c6) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)
    #7 0x55aac81cc738 in fuzzer::RunOneTest(fuzzer::Fuzzer*, char const*, unsigned long) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x924738) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)
    #8 0x55aac81d220a in fuzzer::FuzzerDriver(int*, char***, int (*)(unsigned char const*, unsigned long)) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x92a20a) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)
    #9 0x55aac81fbb82 in main (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x953b82) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)
    #10 0x7fc4307aa082 in __libc_start_main /build/glibc-B3wQXB/glibc-2.31/csu/../csu/libc-start.c:308:16
    #11 0x55aac81c73ed in _start (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x91f3ed) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)

==5855==Register values:
rax = 0x0000502000000098  rbx = 0xfafafafa0000fa00  rcx = 0x00000a047fff8013  rdx = 0x0000000000000008
rdi = 0x0000502000000098  rsi = 0x00007fc432f64bc0  rbp = 0x00007ffc04847ab0  rsp = 0x00007ffc04847268
 r8 = 0x00000a0400000013   r9 = 0x0000000000000000  r10 = 0x00000a0400000013  r11 = 0x0000000000000000
r12 = 0x00000a047fff8010  r13 = 0xffffffffffffffc7  r14 = 0x00007fc42dd00280  r15 = 0x00000ff885ba0050
AddressSanitizer can not provide additional info.
SUMMARY: AddressSanitizer: SEGV /build/glibc-B3wQXB/glibc-2.31/string/../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:312
==5855==ABORTING

Impact

  • Attack vector: Malicious GGUF file (model weights, typically from untrusted sources)
  • Affects: MLX users on all platforms who call the vulnerable method with unsanitized input.
  • Result: Segmentation fault (uncatchable by exception handlers)

Credits:

  • Markiyan Melnyk (ARIMLABS)
  • Mykyta Mudryi (ARIMLABS)
  • Markiyan Chaklosh (ARIMLABS)
Show details on source website

{
  "affected": [
    {
      "database_specific": {
        "last_known_affected_version_range": "\u003c= 0.29.3"
      },
      "package": {
        "ecosystem": "PyPI",
        "name": "mlx"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.29.4"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2025-62609"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-476"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2025-11-21T18:03:20Z",
    "nvd_published_at": "2025-11-21T19:16:02Z",
    "severity": "MODERATE"
  },
  "details": "## Summary\n\nSegmentation fault in `mlx::core::load_gguf()` when loading malicious GGUF files. Untrusted pointer from external gguflib library is dereferenced without validation, causing application crash.\n\nEnvironment:\n- OS: Ubuntu 20.04.6 LTS\n- Compiler: Clang 19.1.7\n\n## Vulnerability\n\n**Location**: `mlx/io/gguf.cpp`\n- Function `extract_tensor_data()` at lines 59-79\n- Vulnerable memcpy at lines 64-67\n- Called from `load_arrays()` at line 177\n\n**The Bug**:\n```cpp\nstd::tuple\u003callocator::Buffer, Dtype\u003e extract_tensor_data(gguf_tensor* tensor) {\n  std::optional\u003cDtype\u003e equivalent_dtype = gguf_type_to_dtype(tensor-\u003etype);\n  if (equivalent_dtype.has_value()) {\n    allocator::Buffer buffer = allocator::malloc(tensor-\u003ebsize);\n    memcpy(\n        buffer.raw_ptr(),\n        tensor-\u003eweights_data,  // untrusted pointer from gguflib\n        tensor-\u003enum_weights * equivalent_dtype.value().size());\n    return {buffer, equivalent_dtype.value()};\n  }\n  // ...\n}\n```\n\n## Possible Fix\n\n```cpp\nstd::tuple\u003callocator::Buffer, Dtype\u003e extract_tensor_data(gguf_tensor* tensor) {\n  std::optional\u003cDtype\u003e equivalent_dtype = gguf_type_to_dtype(tensor-\u003etype);\n  if (equivalent_dtype.has_value()) {\n    // FIX: Validate pointer\n    if (!tensor-\u003eweights_data) {\n      throw std::runtime_error(\"[load_gguf] NULL tensor data pointer\");\n    }\n\n    allocator::Buffer buffer = allocator::malloc(tensor-\u003ebsize);\n    memcpy(\n        buffer.raw_ptr(),\n        tensor-\u003eweights_data,\n        tensor-\u003enum_weights * equivalent_dtype.value().size());\n    return {buffer, equivalent_dtype.value()};\n  }\n  // ...\n}\n```\n\n## PoC\n\n```bash\n# Install MLX\npip install mlx\n\npython3 -c \"import mlx.core as mx; mx.load(\u0027exploit.gguf\u0027, format=\u0027gguf\u0027)\"\n```\n\nDownload the poc file [there](https://drive.google.com/file/d/1t9Z2RJGn-oHmluKWebU077UWSMGXeuVp/view?usp=sharing), or let me know how I can send it to you.\n\n**AddressSanitizer Output (with instrumented build)**:\n```\nAddressSanitizer:DEADLYSIGNAL\n=================================================================\n==5855==ERROR: AddressSanitizer: SEGV on unknown address 0x7fc432f64bc0 (pc 0x7fc430841c12 bp 0x7ffc04847ab0 sp 0x7ffc04847268 T0)\n==5855==The signal is caused by a READ memory access.\n    #0 0x7fc430841c12  /build/glibc-B3wQXB/glibc-2.31/string/../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:312\n    #1 0x55aac829756b in __asan_memcpy (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x9ef56b) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n    #2 0x55aacaa6e8dc in mlx::core::extract_tensor_data(gguf_tensor*) /home/user1/mlx/mlx/io/gguf.cpp:64:5\n    #3 0x55aacaa773fc in mlx::core::load_arrays[abi:cxx11](gguf_ctx*) /home/user1/mlx/mlx/io/gguf.cpp:226:35\n    #4 0x55aacaa782a9 in mlx::core::load_gguf(std::__cxx11::basic_string\u003cchar, std::char_traits\u003cchar\u003e, std::allocator\u003cchar\u003e\u003e const\u0026, std::variant\u003cstd::monostate, mlx::core::Stream, mlx::core::Device\u003e) /home/user1/mlx/mlx/io/gguf.cpp:250:17\n    #5 0x55aac82dc696 in LLVMFuzzerTestOneInput /home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf.cpp:49:19\n    #6 0x55aac81e25c6 in fuzzer::Fuzzer::ExecuteCallback(unsigned char const*, unsigned long) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x93a5c6) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n    #7 0x55aac81cc738 in fuzzer::RunOneTest(fuzzer::Fuzzer*, char const*, unsigned long) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x924738) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n    #8 0x55aac81d220a in fuzzer::FuzzerDriver(int*, char***, int (*)(unsigned char const*, unsigned long)) (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x92a20a) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n    #9 0x55aac81fbb82 in main (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x953b82) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n    #10 0x7fc4307aa082 in __libc_start_main /build/glibc-B3wQXB/glibc-2.31/csu/../csu/libc-start.c:308:16\n    #11 0x55aac81c73ed in _start (/home/user1/mlx/fuzz/load_gguf/fuzz_load_gguf+0x91f3ed) (BuildId: 57467f1ce96052757daeef4b04739be7f23c5f1f)\n\n==5855==Register values:\nrax = 0x0000502000000098  rbx = 0xfafafafa0000fa00  rcx = 0x00000a047fff8013  rdx = 0x0000000000000008\nrdi = 0x0000502000000098  rsi = 0x00007fc432f64bc0  rbp = 0x00007ffc04847ab0  rsp = 0x00007ffc04847268\n r8 = 0x00000a0400000013   r9 = 0x0000000000000000  r10 = 0x00000a0400000013  r11 = 0x0000000000000000\nr12 = 0x00000a047fff8010  r13 = 0xffffffffffffffc7  r14 = 0x00007fc42dd00280  r15 = 0x00000ff885ba0050\nAddressSanitizer can not provide additional info.\nSUMMARY: AddressSanitizer: SEGV /build/glibc-B3wQXB/glibc-2.31/string/../sysdeps/x86_64/multiarch/memmove-vec-unaligned-erms.S:312\n==5855==ABORTING\n```\n\n## Impact\n\n- **Attack vector**: Malicious GGUF file (model weights, typically from untrusted sources)\n- **Affects**: MLX users on all platforms who call the vulnerable method with unsanitized input.\n- **Result**: Segmentation fault (uncatchable by exception handlers)\n\n---\nCredits:\n\n- Markiyan Melnyk (ARIMLABS)\n- Mykyta Mudryi (ARIMLABS)\n- Markiyan Chaklosh (ARIMLABS)",
  "id": "GHSA-j842-xgm4-wf88",
  "modified": "2025-11-21T22:18:18Z",
  "published": "2025-11-21T18:03:20Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/ml-explore/mlx/security/advisories/GHSA-j842-xgm4-wf88"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-62609"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/ml-explore/mlx"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:L/SC:N/SI:N/SA:N/E:P",
      "type": "CVSS_V4"
    }
  ],
  "summary": "MLX has Wild Pointer Dereference in load_gguf()"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Sightings

Author Source Type Date

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.


Loading…

Detection rules are retrieved from Rulezet.

Loading…

Loading…