GHSA-649R-6XHP-RM3W
Vulnerability from github – Published: 2026-08-15 15:30 – Updated: 2026-08-15 15:30In the Linux kernel, the following vulnerability has been resolved:
binfmt_misc: don't let an 'F' entry pin its own instance
An entry registered with 'F' opens its interpreter at registration time and holds that file until the entry is freed. Any entry nobody removes by hand only gets closed once the binfmt_misc superblock is shut down. If the interpreter lives on a mount that keeps that superblock alive the two pin each other:
binfmt_misc sb -> inode -> entry -> interp_file -> vfsmount -> binfmt_misc sb
TL;DR the file is never closed. Once the mount namespace is gone there is nothing left to unregister through either.
There are two ways to trigger this bug:
-
Point the interpreter at the instance itself. Its files are regular files owned by the mounter and both bm_get_inode() and simple_fill_super() leave i_op at empty_iops. So notify_change() falls back to simple_setattr() and chmod +x works. We never set SB_I_NOEXEC and so open_exec() accepts it.
-
Use the instance as an overlayfs lower layer. The overlay superblock holds a clone_private_mount() of every layer until it is destroyed and that clone is in no namespace. So umount_tree() never reaches it.
That's a DoS. And it isn't only the superblock that leaks. It pins the user namespace it was mounted in, so every iteration permanently eats one of the caller's user namespace charges.
So let's just do the sane thing. SB_I_NOEXEC makes open_exec() fail on the instance's own files and s_stack_depth makes overlayfs reject the layer before it ever takes a clone. That also covers the ecryptfs and fuse passthrough variants. What 'F' promises is unchanged.
The stable tag is narrower than the Fixes tags on purpose. Before sandboxed mounts this needed global root against the single instance everyone shares, and the change doesn't apply to those trees anyway.
Note that SB_I_NODEV is implicitly raised for userns mounts but raise it explicitly here as well.
{
"affected": [],
"aliases": [
"CVE-2026-74484"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-08-15T13:17:53Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nbinfmt_misc: don\u0027t let an \u0027F\u0027 entry pin its own instance\n\nAn entry registered with \u0027F\u0027 opens its interpreter at registration time\nand holds that file until the entry is freed. Any entry nobody removes\nby hand only gets closed once the binfmt_misc superblock is shut down.\nIf the interpreter lives on a mount that keeps that superblock alive the\ntwo pin each other:\n\n binfmt_misc sb -\u003e inode -\u003e entry -\u003e interp_file -\u003e vfsmount -\u003e binfmt_misc sb\n\nTL;DR the file is never closed. Once the mount namespace is gone there\nis nothing left to unregister through either.\n\nThere are two ways to trigger this bug:\n\n- Point the interpreter at the instance itself. Its files are regular\n files owned by the mounter and both bm_get_inode() and\n simple_fill_super() leave i_op at empty_iops. So notify_change() falls\n back to simple_setattr() and chmod +x works. We never set SB_I_NOEXEC\n and so open_exec() accepts it.\n\n- Use the instance as an overlayfs lower layer. The overlay superblock\n holds a clone_private_mount() of every layer until it is destroyed and\n that clone is in no namespace. So umount_tree() never reaches it.\n\nThat\u0027s a DoS. And it isn\u0027t only the superblock that leaks. It pins the\nuser namespace it was mounted in, so every iteration permanently eats\none of the caller\u0027s user namespace charges.\n\nSo let\u0027s just do the sane thing. SB_I_NOEXEC makes open_exec() fail on\nthe instance\u0027s own files and s_stack_depth makes overlayfs reject the\nlayer before it ever takes a clone. That also covers the ecryptfs and\nfuse passthrough variants. What \u0027F\u0027 promises is unchanged.\n\nThe stable tag is narrower than the Fixes tags on purpose. Before\nsandboxed mounts this needed global root against the single instance\neveryone shares, and the change doesn\u0027t apply to those trees anyway.\n\nNote that SB_I_NODEV is implicitly raised for userns mounts but raise it\nexplicitly here as well.",
"id": "GHSA-649r-6xhp-rm3w",
"modified": "2026-08-15T15:30:34Z",
"published": "2026-08-15T15:30:34Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-74484"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/098e92fe0f1bde5af99c8cf504f13f01ef139f85"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/1cc2decee06acc939337304e9b3f737fd5d8c4bd"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/4dad8ca637d44d5a6d5c23fa80c9e2e455198c2b"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/79055d82772b9584f259b747fe40ff56a076678d"
}
],
"schema_version": "1.4.0",
"severity": []
}
Sightings
| Author | Source | Type | Date | Other |
|---|
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.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.