{
  "affected": [
    {
      "ranges": [
        {
          "events": [
            {
              "introduced": "d02198550423a0b695e7a24ec77153209ad45b09"
            },
            {
              "fixed": "22b365ba1af3d8c6036b8e5112fffe80998b85a0"
            },
            {
              "fixed": "db155b86d1523e85941f61efd7d7ffb594cc9a29"
            },
            {
              "fixed": "088f65e206087bf903743bd18417261d7a4c9644"
            }
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "type": "GIT"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "Linux",
        "name": "Kernel"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "6.15.0"
            },
            {
              "fixed": "6.18.30"
            }
          ],
          "type": "ECOSYSTEM"
        },
        {
          "events": [
            {
              "introduced": "6.19.0"
            },
            {
              "fixed": "7.0.7"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "database_specific": {
    "cna_assigner": "Linux",
    "osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/46xxx/CVE-2026-46290.json"
  },
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nx86/efi: Fix graceful fault handling after FPU softirq changes\n\nSince commit d02198550423 (\"x86/fpu: Improve crypto performance by\nmaking kernel-mode FPU reliably usable in softirqs\"), kernel_fpu_begin()\ncalls fpregs_lock() which uses local_bh_disable() instead of the\nprevious preempt_disable(). This sets SOFTIRQ_OFFSET in preempt_count\nduring the entire EFI runtime service call, causing in_interrupt() to\nreturn true in normal task context.\n\nThe graceful page fault handler efi_crash_gracefully_on_page_fault()\nuses in_interrupt() to bail out for faults in real interrupt context.\nWith SOFTIRQ_OFFSET now set, the handler always bails out, leaving EFI\nfirmware page faults unhandled. This escalates to die() which also sees\nin_interrupt() as true and calls panic(\"Fatal exception in interrupt\"),\nresulting in a hard system freeze. On systems with buggy firmware that\ntriggers page faults during EFI runtime calls (e.g., accessing unmapped\nmemory in GetTime()), this causes an unrecoverable hang instead of the\nexpected graceful EFI_ABORTED recovery.\n\nFix by replacing in_interrupt() with !in_task(). This preserves the\noriginal intent of bailing for interrupts or NMI faults, while no longer\nfalsely triggering from the FPU code path's local_bh_disable().\n\n[ardb: Sashiko spotted that using 'in_hardirq() || in_nmi()' leaves a\n       window where a softirq may be taken before fpregs_lock() is\n       called, but after efi_rts_work.efi_rts_id has been assigned,\n       and any page faults occurring in that window will then be\n       misidentified as having been caused by the firmware. Instead,\n       use !in_task(), which incorporates in_serving_softirq(). ]",
  "id": "CVE-2026-46290",
  "modified": "2026-07-10T03:53:44.922722172Z",
  "published": "2026-06-08T15:46:17.224Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/088f65e206087bf903743bd18417261d7a4c9644"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/22b365ba1af3d8c6036b8e5112fffe80998b85a0"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/db155b86d1523e85941f61efd7d7ffb594cc9a29"
    },
    {
      "type": "ADVISORY",
      "url": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/46xxx/CVE-2026-46290.json"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-46290"
    },
    {
      "type": "PACKAGE",
      "url": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git"
    }
  ],
  "schema_version": "1.7.5",
  "summary": "x86/efi: Fix graceful fault handling after FPU softirq changes"
}