ghsa-4297-r96p-57pp
Vulnerability from github
Published
2025-04-16 15:34
Modified
2025-04-16 15:34
Details

In the Linux kernel, the following vulnerability has been resolved:

mm/gup: reject FOLL_SPLIT_PMD with hugetlb VMAs

Patch series "mm: fixes for device-exclusive entries (hmm)", v2.

Discussing the PageTail() call in make_device_exclusive_range() with Willy, I recently discovered [1] that device-exclusive handling does not properly work with THP, making the hmm-tests selftests fail if THPs are enabled on the system.

Looking into more details, I found that hugetlb is not properly fenced, and I realized that something that was bugging me for longer -- how device-exclusive entries interact with mapcounts -- completely breaks migration/swapout/split/hwpoison handling of these folios while they have device-exclusive PTEs.

The program below can be used to allocate 1 GiB worth of pages and making them device-exclusive on a kernel with CONFIG_TEST_HMM.

Once they are device-exclusive, these folios cannot get swapped out (proc$pid/smaps_rollup will always indicate 1 GiB RSS no matter how much one forces memory reclaim), and when having a memory block onlined to ZONE_MOVABLE, trying to offline it will loop forever and complain about failed migration of a page that should be movable.

echo offline > /sys/devices/system/memory/memory136/state

echo online_movable > /sys/devices/system/memory/memory136/state

./hmm-swap &

... wait until everything is device-exclusive

echo offline > /sys/devices/system/memory/memory136/state

[ 285.193431][T14882] page: refcount:2 mapcount:0 mapping:0000000000000000 index:0x7f20671f7 pfn:0x442b6a [ 285.196618][T14882] memcg:ffff888179298000 [ 285.198085][T14882] anon flags: 0x5fff0000002091c(referenced|uptodate| dirty|active|owner_2|swapbacked|node=1|zone=3|lastcpupid=0x7ff) [ 285.201734][T14882] raw: ... [ 285.204464][T14882] raw: ... [ 285.207196][T14882] page dumped because: migration failure [ 285.209072][T14882] page_owner tracks the page as allocated [ 285.210915][T14882] page last allocated via order 0, migratetype Movable, gfp_mask 0x140dca(GFP_HIGHUSER_MOVABLE|__GFP_COMP|__GFP_ZERO), id 14926, tgid 14926 (hmm-swap), ts 254506295376, free_ts 227402023774 [ 285.216765][T14882] post_alloc_hook+0x197/0x1b0 [ 285.218874][T14882] get_page_from_freelist+0x76e/0x3280 [ 285.220864][T14882] __alloc_frozen_pages_noprof+0x38e/0x2740 [ 285.223302][T14882] alloc_pages_mpol+0x1fc/0x540 [ 285.225130][T14882] folio_alloc_mpol_noprof+0x36/0x340 [ 285.227222][T14882] vma_alloc_folio_noprof+0xee/0x1a0 [ 285.229074][T14882] __handle_mm_fault+0x2b38/0x56a0 [ 285.230822][T14882] handle_mm_fault+0x368/0x9f0 ...

This series fixes all issues I found so far. There is no easy way to fix without a bigger rework/cleanup. I have a bunch of cleanups on top (some previous sent, some the result of the discussion in v1) that I will send out separately once this landed and I get to it.

I wish we could just use some special present PROT_NONE PTEs instead of these (non-present, non-none) fake-swap entries; but that just results in the same problem we keep having (lack of spare PTE bits), and staring at other similar fake-swap entries, that ship has sailed.

With this series, make_device_exclusive() doesn't actually belong into mm/rmap.c anymore, but I'll leave moving that for another day.

I only tested this series with the hmm-tests selftests due to lack of HW, so I'd appreciate some testing, especially if the interaction between two GPUs wanting a device-exclusive entry works as expected.

include

include

include

include

include

include

include

include

include

include

define HMM_DMIRROR_EXCLUSIVE _IOWR('H', 0x05, struct hmm_dmirror_cmd)

struct hmm_dmirror_cmd { __u64 addr; __u64 ptr; __u64 npages; __u64 cpages; __u64 faults; };

const size_t size = 1 * 1024 * 1024 * 1024ul; const size_t chunk_size = 2 * 1024 * 1024ul;

int m ---truncated---

Show details on source website

To clipboard 

{
  "affected": [],
  "aliases": [
    "CVE-2025-22034"
  ],
  "database_specific": {
    "cwe_ids": [],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2025-04-16T15:15:56Z",
    "severity": null
  },
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nmm/gup: reject FOLL_SPLIT_PMD with hugetlb VMAs\n\nPatch series \"mm: fixes for device-exclusive entries (hmm)\", v2.\n\nDiscussing the PageTail() call in make_device_exclusive_range() with\nWilly, I recently discovered [1] that device-exclusive handling does not\nproperly work with THP, making the hmm-tests selftests fail if THPs are\nenabled on the system.\n\nLooking into more details, I found that hugetlb is not properly fenced,\nand I realized that something that was bugging me for longer -- how\ndevice-exclusive entries interact with mapcounts -- completely breaks\nmigration/swapout/split/hwpoison handling of these folios while they have\ndevice-exclusive PTEs.\n\nThe program below can be used to allocate 1 GiB worth of pages and making\nthem device-exclusive on a kernel with CONFIG_TEST_HMM.\n\nOnce they are device-exclusive, these folios cannot get swapped out\n(proc$pid/smaps_rollup will always indicate 1 GiB RSS no matter how much\none forces memory reclaim), and when having a memory block onlined to\nZONE_MOVABLE, trying to offline it will loop forever and complain about\nfailed migration of a page that should be movable.\n\n# echo offline \u003e /sys/devices/system/memory/memory136/state\n# echo online_movable \u003e /sys/devices/system/memory/memory136/state\n# ./hmm-swap \u0026\n... wait until everything is device-exclusive\n# echo offline \u003e /sys/devices/system/memory/memory136/state\n[  285.193431][T14882] page: refcount:2 mapcount:0 mapping:0000000000000000\n  index:0x7f20671f7 pfn:0x442b6a\n[  285.196618][T14882] memcg:ffff888179298000\n[  285.198085][T14882] anon flags: 0x5fff0000002091c(referenced|uptodate|\n  dirty|active|owner_2|swapbacked|node=1|zone=3|lastcpupid=0x7ff)\n[  285.201734][T14882] raw: ...\n[  285.204464][T14882] raw: ...\n[  285.207196][T14882] page dumped because: migration failure\n[  285.209072][T14882] page_owner tracks the page as allocated\n[  285.210915][T14882] page last allocated via order 0, migratetype\n  Movable, gfp_mask 0x140dca(GFP_HIGHUSER_MOVABLE|__GFP_COMP|__GFP_ZERO),\n  id 14926, tgid 14926 (hmm-swap), ts 254506295376, free_ts 227402023774\n[  285.216765][T14882]  post_alloc_hook+0x197/0x1b0\n[  285.218874][T14882]  get_page_from_freelist+0x76e/0x3280\n[  285.220864][T14882]  __alloc_frozen_pages_noprof+0x38e/0x2740\n[  285.223302][T14882]  alloc_pages_mpol+0x1fc/0x540\n[  285.225130][T14882]  folio_alloc_mpol_noprof+0x36/0x340\n[  285.227222][T14882]  vma_alloc_folio_noprof+0xee/0x1a0\n[  285.229074][T14882]  __handle_mm_fault+0x2b38/0x56a0\n[  285.230822][T14882]  handle_mm_fault+0x368/0x9f0\n...\n\nThis series fixes all issues I found so far.  There is no easy way to fix\nwithout a bigger rework/cleanup.  I have a bunch of cleanups on top (some\nprevious sent, some the result of the discussion in v1) that I will send\nout separately once this landed and I get to it.\n\nI wish we could just use some special present PROT_NONE PTEs instead of\nthese (non-present, non-none) fake-swap entries; but that just results in\nthe same problem we keep having (lack of spare PTE bits), and staring at\nother similar fake-swap entries, that ship has sailed.\n\nWith this series, make_device_exclusive() doesn\u0027t actually belong into\nmm/rmap.c anymore, but I\u0027ll leave moving that for another day.\n\nI only tested this series with the hmm-tests selftests due to lack of HW,\nso I\u0027d appreciate some testing, especially if the interaction between two\nGPUs wanting a device-exclusive entry works as expected.\n\n\u003cprogram\u003e\n#include \u003cstdio.h\u003e\n#include \u003cfcntl.h\u003e\n#include \u003cstdint.h\u003e\n#include \u003cunistd.h\u003e\n#include \u003cstdlib.h\u003e\n#include \u003cstring.h\u003e\n#include \u003csys/mman.h\u003e\n#include \u003csys/ioctl.h\u003e\n#include \u003clinux/types.h\u003e\n#include \u003clinux/ioctl.h\u003e\n\n#define HMM_DMIRROR_EXCLUSIVE _IOWR(\u0027H\u0027, 0x05, struct hmm_dmirror_cmd)\n\nstruct hmm_dmirror_cmd {\n\t__u64 addr;\n\t__u64 ptr;\n\t__u64 npages;\n\t__u64 cpages;\n\t__u64 faults;\n};\n\nconst size_t size = 1 * 1024 * 1024 * 1024ul;\nconst size_t chunk_size = 2 * 1024 * 1024ul;\n\nint m\n---truncated---",
  "id": "GHSA-4297-r96p-57pp",
  "modified": "2025-04-16T15:34:39Z",
  "published": "2025-04-16T15:34:39Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-22034"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/2e877ff3492267def06dd50cb165dc9ab8838e7d"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/48d28417c66cce2f3b0ba773fcb6695a56eff220"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/8977752c8056a6a094a279004a49722da15bace3"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/fd900832e8440046627b60697687ab5d04398008"
    }
  ],
  "schema_version": "1.4.0",
  "severity": []
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Sightings

Author Source Type Date

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or seen somewhere by the user.
  • Confirmed: The vulnerability is confirmed from an analyst perspective.
  • Exploited: This vulnerability was exploited and seen by the user reporting the sighting.
  • Patched: This vulnerability was successfully patched by the user reporting the sighting.
  • Not exploited: This vulnerability was not exploited or seen by the user reporting the sighting.
  • Not confirmed: The user expresses doubt about the veracity of the vulnerability.
  • Not patched: This vulnerability was not successfully patched by the user reporting the sighting.