GHSA-9W29-5GPQ-XMV3
Vulnerability from github – Published: 2026-10-28 12:31 – Updated: 2026-10-30 15:34In the Linux kernel, the following vulnerability has been resolved:
x86/kvm: Don't consume async #PF flags on a kernel #PF in NMI context
An async #PF "page not present" event arrives through the #PF vector, and its reason sits in the per-CPU apf_reason.flags until __kvm_handle_async_pf() reads and clears it. If an NMI arrives after the event is injected and before the #PF handler reads the flags, and the NMI handler itself takes a #PF, that fault reads and clears the flags belonging to the #PF it interrupted. It then finds interrupts disabled in the NMI context it interrupted, and panics:
Kernel panic - not syncing: Host injected async #PF in interrupt disabled region Call Trace: panic+0x52/0x60 __kvm_handle_async_pf+0xb0/0xb0 exc_page_fault+0x71/0xb0 asm_exc_page_fault+0x27/0x30 RIP: 0010:__get_user_nocheck_8+0x6/0x20 perf_callchain_user+0x147/0x1a0 get_perf_callchain+0x12d/0x1c0 ... perf_event_nmi_handler+0x26/0x50 exc_nmi+0xe8/0x110 RIP: 0010:asm_exc_page_fault+0x0/0x30
The last RIP line shows the NMI landing on the first instruction of the #PF entry. perf with user call chains takes such a fault: the sampling NMI walks the user stack with __get_user(), which faults when a saved frame pointer is not a valid user address or the walk runs off the top of the stack. The same window exists when KVM, running as an L1 hypervisor with DELIVERY_AS_PF_VMEXIT, reads the flags itself after a #PF VM-exit.
The NMI entry code saves and restores CR2 for this nesting, but not the async #PF reason. Andy Lutomirski described the race, with perf as the example, when async #PF delivery was reworked in 2020 [1]. Saving and clearing the flags across the NMI, next to CR2, was discussed there but not merged.
Since that rework, Linux has stopped setting KVM_ASYNC_PF_SEND_ALWAYS, so the host delivers "page not present" only for user mode faults and a kernel-mode fault in NMI context (including kernel-mode #MC and #DB) cannot be an async page fault. In that case, return without touching the flags so that the interrupted handler finds them intact.
Tested on v7.3-rc5 in a 4-vCPU, 2 GiB guest restored from a snapshot whose memory is served on demand through userfaultfd, so that reading it back takes async #PFs, while every thread samples hardware cycles with user call chains and points its frame pointer at an unmapped page. Without this patch the guest panicked as above in 3 of 3 runs. With it, the guest survived 3 of 3 runs of 60 seconds, and the host counted 148,441 to 153,334 kvm_async_pf_not_present events in the first 20 seconds of each. On 6.18.50 the results were the same: the guest panicked without the patch and survived with it.
{
"affected": [],
"aliases": [
"CVE-2026-99178"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-10-28T11:17:38Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nx86/kvm: Don\u0027t consume async #PF flags on a kernel #PF in NMI context\n\nAn async #PF \"page not present\" event arrives through the #PF vector,\nand its reason sits in the per-CPU apf_reason.flags until\n__kvm_handle_async_pf() reads and clears it. If an NMI arrives after\nthe event is injected and before the #PF handler reads the flags, and\nthe NMI handler itself takes a #PF, that fault reads and clears the\nflags belonging to the #PF it interrupted. It then finds interrupts\ndisabled in the NMI context it interrupted, and panics:\n\n Kernel panic - not syncing: Host injected async #PF in interrupt\ndisabled region\n Call Trace:\n \u003cNMI\u003e\n panic+0x52/0x60\n __kvm_handle_async_pf+0xb0/0xb0\n exc_page_fault+0x71/0xb0\n asm_exc_page_fault+0x27/0x30\n RIP: 0010:__get_user_nocheck_8+0x6/0x20\n perf_callchain_user+0x147/0x1a0\n get_perf_callchain+0x12d/0x1c0\n ...\n perf_event_nmi_handler+0x26/0x50\n exc_nmi+0xe8/0x110\n RIP: 0010:asm_exc_page_fault+0x0/0x30\n \u003c/NMI\u003e\n\nThe last RIP line shows the NMI landing on the first instruction of\nthe #PF entry. perf with user call chains takes such a fault: the\nsampling NMI walks the user stack with __get_user(), which faults when\na saved frame pointer is not a valid user address or the walk runs off\nthe top of the stack. The same window exists when KVM, running as an\nL1 hypervisor with DELIVERY_AS_PF_VMEXIT, reads the flags itself after\na #PF VM-exit.\n\nThe NMI entry code saves and restores CR2 for this nesting, but not the\nasync #PF reason. Andy Lutomirski described the race, with perf as the\nexample, when async #PF delivery was reworked in 2020 [1]. Saving and\nclearing the flags across the NMI, next to CR2, was discussed there but\nnot merged.\n\nSince that rework, Linux has stopped setting KVM_ASYNC_PF_SEND_ALWAYS,\nso the host delivers \"page not present\" only for user mode faults and\na kernel-mode fault in NMI context (including kernel-mode #MC and #DB)\ncannot be an async page fault. In that case, return without touching\nthe flags so that the interrupted handler finds them intact.\n\nTested on v7.3-rc5 in a 4-vCPU, 2 GiB guest restored from a snapshot\nwhose memory is served on demand through userfaultfd, so that reading\nit back takes async #PFs, while every thread samples hardware cycles\nwith user call chains and points its frame pointer at an unmapped page.\nWithout this patch the guest panicked as above in 3 of 3 runs. With it,\nthe guest survived 3 of 3 runs of 60 seconds, and the host counted\n148,441 to 153,334 kvm_async_pf_not_present events in the first 20\nseconds of each. On 6.18.50 the results were the same: the guest\npanicked without the patch and survived with it.",
"id": "GHSA-9w29-5gpq-xmv3",
"modified": "2026-10-30T15:34:23Z",
"published": "2026-10-28T12:31:07Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-99178"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/81d99abe2bd81379abfb595e403bbbee6fbfc624"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/b77e9622bbade9812e2090c66d25c372f515e921"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/f45ee1ff6358a22d7e22a4c53cb3148ea97b10d9"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/fd60b32a99627fda479c9fa1d4a82fa9f7994670"
}
],
"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.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.