{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2024-35979",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2024-35979"
  ],
  "published": "2024-05-20T09:42:41Z",
  "summary": "raid1: fix use-after-free for original bio in raid1_write_request()",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nraid1: fix use-after-free for original bio in raid1_write_request()\n\nr1_bio->bios[] is used to record new bios that will be issued to\nunderlying disks, however, in raid1_write_request(), r1_bio->bios[]\nwill set to the original bio temporarily. Meanwhile, if blocked rdev\nis set, free_r1bio() will be called causing that all r1_bio->bios[]\nto be freed:\n\nraid1_write_request()\n r1_bio = alloc_r1bio(mddev, bio); -> r1_bio->bios[] is NULL\n for (i = 0;  i < disks; i++) -> for each rdev in conf\n  // first rdev is normal\n  r1_bio->bios[0] = bio; -> set to original bio\n  // second rdev is blocked\n  if (test_bit(Blocked, &rdev->flags))\n   break\n\n if (blocked_rdev)\n  free_r1bio()\n   put_all_bios()\n    bio_put(r1_bio->bios[0]) -> original bio is freed\n\nTest scripts:\n\nmdadm -CR /dev/md0 -l1 -n4 /dev/sd[abcd] --assume-clean\nfio -filename=/dev/md0 -ioengine=libaio -rw=write -bs=4k -numjobs=1 \\\n    -iodepth=128 -name=test -direct=1\necho blocked > /sys/block/md0/md/rd2/state\n\nTest result:\n\nBUG bio-264 (Not tainted): Object already free\n-----------------------------------------------------------------------------\n\nAllocated in mempool_alloc_slab+0x24/0x50 age=1 cpu=1 pid=869\n kmem_cache_alloc+0x324/0x480\n mempool_alloc_slab+0x24/0x50\n mempool_alloc+0x6e/0x220\n bio_alloc_bioset+0x1af/0x4d0\n blkdev_direct_IO+0x164/0x8a0\n blkdev_write_iter+0x309/0x440\n aio_write+0x139/0x2f0\n io_submit_one+0x5ca/0xb70\n __do_sys_io_submit+0x86/0x270\n __x64_sys_io_submit+0x22/0x30\n do_syscall_64+0xb1/0x210\n entry_SYSCALL_64_after_hwframe+0x6c/0x74\nFreed in mempool_free_slab+0x1f/0x30 age=1 cpu=1 pid=869\n kmem_cache_free+0x28c/0x550\n mempool_free_slab+0x1f/0x30\n mempool_free+0x40/0x100\n bio_free+0x59/0x80\n bio_put+0xf0/0x220\n free_r1bio+0x74/0xb0\n raid1_make_request+0xadf/0x1150\n md_handle_request+0xc7/0x3b0\n md_submit_bio+0x76/0x130\n __submit_bio+0xd8/0x1d0\n submit_bio_noacct_nocheck+0x1eb/0x5c0\n submit_bio_noacct+0x169/0xd40\n submit_bio+0xee/0x1d0\n blkdev_direct_IO+0x322/0x8a0\n blkdev_write_iter+0x309/0x440\n aio_write+0x139/0x2f0\n\nSince that bios for underlying disks are not allocated yet, fix this\nproblem by using mempool_free() directly to free the r1_bio.",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:H/I:H/A:H"
    }
  ],
  "affected": [
    {
      "package": {
        "ecosystem": "Raspberry Pi OS:12",
        "name": "linux"
      },
      "versions": [
        "1:6.6.20-1+rpt1"
      ],
      "ecosystem_specific": {
        "binaries": [
          {
            "binary_name": "linux-image-6.6.20+rpt-rpi-2712",
            "binary_version": "1:6.6.20-1+rpt1"
          },
          {
            "binary_name": "linux-image-6.6.20+rpt-rpi-v6",
            "binary_version": "1:6.6.20-1+rpt1"
          },
          {
            "binary_name": "linux-image-6.6.20+rpt-rpi-v7",
            "binary_version": "1:6.6.20-1+rpt1"
          },
          {
            "binary_name": "linux-image-6.6.20+rpt-rpi-v7l",
            "binary_version": "1:6.6.20-1+rpt1"
          },
          {
            "binary_name": "linux-image-6.6.20+rpt-rpi-v8",
            "binary_version": "1:6.6.20-1+rpt1"
          }
        ]
      },
      "database_specific": {
        "per_version": {
          "1:6.6.20-1+rpt1": {
            "linux_commit": "6f16847710cc0502450788b9f12f0a14d3429668",
            "upstream_version": "6.6.20",
            "assessment": "commit ancestry (strak rule)",
            "upstream_fixed": "6.6.28"
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2024-35979"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/3f28d49a328fe20926995d5fbdc92da665596268"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/f423f41b7679c09abb26d2bd54be5cbef23c9446"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/fcf3f7e2fc8a53a6140beee46ec782a4c88e4744"
    }
  ],
  "database_specific": {
    "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git",
    "upstream_modified": "2026-08-05T11:52:36Z",
    "severity": {
      "source": "kernel CNA",
      "cvss_v3_score": 7.8,
      "cvss_v3_severity": "High",
      "scenario": "AV:L - The bug is in the md/raid1 block driver and is reached only through local I/O submission \u2014 a write to /dev/mdX (O_DIRECT/AIO/io_uring) or to a filesystem hosted on the array \u2014 via md_submit_bio() \u2192 md_handle_request() \u2192 raid1_make_request() \u2192 raid1_write_request(). No remote or network protocol reaches this code.\nAC:L - Once any member rdev carries the Blocked/BlockedBadBlocks flag (set by raid1_error() on a failing member, by an unacknowledged bad block, or by a one-line sysfs write), every subsequent write to the array deterministically takes the blocked_rdev path and calls free_r1bio() on an r1bio whose bios[] entries still alias the caller's original bio \u2014 no race must be won, and the upstream reproducer triggers it reliably on the first write batch.\nPR:L - The freed object is the submitting task's own bio, and any unprivileged user who can write to a file on a raid1-backed filesystem (or to the array device) supplies the bio that gets prematurely bio_put()'d; no capability is checked anywhere on md_submit_bio() \u2192 raid1_write_request(), and the Blocked state is ordinary array state rather than a privilege the attacker must hold.\nUI:N - The attacker's own write syscalls drive the entire path; no action by another user or administrator is needed at exploitation time.\nS:U - The corruption is confined to kernel heap objects managed by the same kernel security authority; no VM, IOMMU, or sandbox boundary is crossed.\nC:H - The original bio is returned to its bioset mempool and then kept in use as r1_bio->master_bio: it is re-read by init_r1bio(), bio_split()/bio_chain(), and bio_alloc_clone(), which copies bi_bdev, bi_iter and the bi_io_vec pointer out of freed memory \u2014 after reallocation with attacker-sprayed data this yields disk I/O into or out of attacker-chosen kernel memory, an arbitrary-read primitive.\nI:H - md_wait_for_blocked_rdev() sleeps with the freed bio still referenced, giving an unbounded window to reclaim the slab object; the clones built from it write attacker-controlled vectors, and call_bio_endio() finally invokes bio->bi_end_io from freed memory, giving a function-pointer hijack plus a double-free when the original submitter completes the same bio.\nA:H - This is a use-after-free/double-free on a mempool-backed slab object that KASAN/SLUB flags as \"Object already free\"; in practice it panics the kernel or corrupts the block layer's bio state, taking the storage stack and the machine down."
    }
  }
}