{"id":"CVE-2026-99093","title":"mm: don't schedule deferred kernel page table freeing while booting","summary":"In the Linux kernel, the following vulnerability has been resolved:\n\nmm: don't schedule deferred kernel page table freeing while booting\n\nBooting with a boot-time function tracer and a filter, for example\n\n  ftrace=function ftrace_filter…","severity":"none","vendor":"Linux","product":"Linux","affected":["Linux >= 84e0cd79d57f06b872154eb5d8f610584133f260 < 7a25ca170c12a9328f6fcf4410713beb516b01c0","Linux >= 922873cf4fc34124215070a8ad391bf831538bb9 < df6c704a08bc0925d5db047b0a6221f962bda4d9","Linux >= 9e4a3ec3411bb6bb59e3c1f29b75609f1e87aac4 < 2fb552b0bee70c24c092720028ab4620f5eb2646","Linux >= 6.18.53 < 6.18.57","Linux >= 7.2.7 < 7.2.11"],"published":"2026-10-26","updated":"2026-10-26","sourceUpdated":"2026-10-26T14:52:01.450Z","source":"CVEORG","sourceUrl":"https://www.cve.org/CVERecord?id=CVE-2026-99093","references":[{"url":"https://git.kernel.org/stable/c/7a25ca170c12a9328f6fcf4410713beb516b01c0"},{"url":"https://git.kernel.org/stable/c/df6c704a08bc0925d5db047b0a6221f962bda4d9"},{"url":"https://git.kernel.org/stable/c/2fb552b0bee70c24c092720028ab4620f5eb2646"}],"tags":["cve.org"],"ingestedAt":"2026-10-26T15:29:30.508Z","slug":"CVE-2026-99093","body":"## Overview\n\nIn the Linux kernel, the following vulnerability has been resolved:\n\nmm: don't schedule deferred kernel page table freeing while booting\n\nBooting with a boot-time function tracer and a filter, for example\n\n  ftrace=function ftrace_filter=pud_free_pmd_page\n\npanics on 7.3-rc4 as soon as the tracer starts:\n\n  [   23.531178] Starting tracer 'function'\n  [   23.675800] Oops: general protection fault, probably for non-canonical address 0xdffffc0000000038: 0000 [#1] SMP KASAN NOPTI\n  [   23.819917] KASAN: null-ptr-deref in range [0x00000000000001c0-0x00000000000001c7]\n  [   23.964025] CPU: 0 UID: 0 PID: 0 Comm: swapper Not tainted 7.3.0-rc4-fe2ec83746e5-with-fixes-v2+ #195 PREEMPT(undef)\n  [   24.252248] RIP: 0010:__queue_work+0xab/0xf00\n  [   25.981629] Call Trace:\n  [   26.125727]  <TASK>\n  [   26.413912]  ? pagetable_free_kernel+0x20/0x120\n  [   26.990283]  queue_work_on+0x97/0xf0\n  [   27.134382]  __cpa_collapse_large_pages+0x501/0x6f0\n  [   27.566662]  cpa_flush+0x394/0x620\n  [   27.998953]  change_page_attr_set_clr+0x321/0x4a0\n  [   29.151729]  set_memory_rox+0xa2/0xf0\n  [   29.584018]  create_trampoline+0x431/0x6f0\n  ...\n  [   44.343347] Kernel panic - not syncing: Attempted to kill the idle task!\n\nThe boot-time tracer is started from early_trace_init(), which runs\nbefore workqueue_init_early().  Making its trampoline read-only splits a\nlarge page, and CPA collapses it again right away.  The split table has\nbeen a kernel page table since commit 9e4a3ec3411b\n(\"x86/mm/pat: Allocate split page tables as kernel page tables\"), so the\ncollapse frees it through pagetable_free_kernel(), which queues work on\nsystem_percpu_wq - still NULL at that point.  That commit is correct in\nitself; it only lets CPA reach pagetable_free_kernel() before the\nworkqueue that function relies on exists.\n\nKeep putting the table on the list, but don't schedule the work while\nthe system is still booting.  A core_initcall schedules it once to free\nwhatever was queued by then.\n\n## Affected\n\n- `Linux >= 84e0cd79d57f06b872154eb5d8f610584133f260 < 7a25ca170c12a9328f6fcf4410713beb516b01c0`\n- `Linux >= 922873cf4fc34124215070a8ad391bf831538bb9 < df6c704a08bc0925d5db047b0a6221f962bda4d9`\n- `Linux >= 9e4a3ec3411bb6bb59e3c1f29b75609f1e87aac4 < 2fb552b0bee70c24c092720028ab4620f5eb2646`\n- `Linux >= 6.18.53 < 6.18.57`\n- `Linux >= 7.2.7 < 7.2.11`\n\n## Remediation\n\nRefer to the linked advisories for vendor-supplied fixes and affected version ranges.","depth":"sunlit","depthScore":3,"depthScoreParts":{"impact":2.8,"likelihood":0,"exploitation":0,"ransomware":0},"changes":[]}