CVE-2026-64210 (GCVE-0-2026-64210)
Vulnerability from cvelistv5
Published
2026-07-24 15:23
Modified
2026-08-05 12:40
Severity ?
7.5 (High) - CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Summary
In the Linux kernel, the following vulnerability has been resolved: net/mlx5e: xsk: Fix unlocked writing to ICOSQ During napi poll, when the affinity changes and there's still XSK work to be done, we trigger an ICOSQ interrupt on the new CPU. However, this triggering on the ICOSQ is done unprotected. There are 2 such races: A) mlx5e_trigger_irq() is called while mlx5e_xsk_alloc_rx_mpwqe() is running from a different CPU due to affinity change. This can happen because IRQ triggering is done after napi_complete_done(). At this point the NAPI can be scheduled on a different CPU. Like this: CPU A (old affinity, NAPI tail) CPU B (new affinity, fresh NAPI) ------------------------------- -------------------------------- napi_complete_done() clears SCHED mlx5e_cq_arm(...) napi_schedule_prep() sets SCHED mlx5e_napi_poll() mlx5e_xsk_alloc_rx_mpwqe() mlx5e_icosq_sync_lock() // noop memcpy 640 B UMR body advance sq->pc by 10 mlx5e_trigger_irq(&c->icosq) wqe_info[pi] = {NOP, 1} mlx5e_post_nop() advances sq->pc B) mlx5e_trigger_irq() is called on the ICOSQ when mlx5e_trigger_napi_icosq() is running. The obvious fix would be to lock the ICOSQ. But ICOSQ has an optimized locking scheme that doesn't work for this scenario. Kick the async ICOSQ instead which is always locked. This issue was noticed in the wild with the following splat: netdevice: ge-0-0-1: Bad OP in ICOSQ CQE: 0xd WARNING: drivers/net/ethernet/mellanox/mlx5/core/en_rx.c:826 [...] [...] Call Trace: <IRQ> mlx5e_napi_poll+0x11d/0x7f0 [mlx5_core] __napi_poll+0x30/0x200 ? skb_defer_free_flush+0x9c/0xc0 net_rx_action+0x2fe/0x3f0 handle_softirqs+0xd8/0x340 __irq_exit_rcu+0xbc/0xe0 common_interrupt+0x85/0xa0 </IRQ> <TASK> asm_common_interrupt+0x26/0x40 [...] ---[ end trace 0000000000000000 ]--- mlx5_core 0000:08:00.0 ge-0-0-1: Error cqe on cqn 0x548, ci 0x2022, qn 0x8f4, opcode 0xd, syndrome 0x2, vendor syndrome 0x68 00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000030: 00 00 00 00 01 00 68 02 01 00 08 f4 de 14 59 d2 WQE DUMP: WQ size 16384 WQ cur size 0, WQE index 0x1e14, len: 64 00000000: 00 00 00 01 d9 ed 80 02 00 00 00 01 d9 ed 90 02 00000010: 00 00 00 01 d9 ed a0 02 00 00 00 01 d9 ed b0 02 00000020: 00 00 00 01 d9 ed c0 02 00 00 00 01 d9 ed d0 02 00000030: 00 00 00 01 d9 ed e0 02 00 00 00 01 d9 ed f0 02 mlx5_core 0000:08:00.0 ge-0-0-1: Error cqe on cqn 0x548, ci 0x2023, qn 0x8f4, opcode 0xd, syndrome 0x5, vendor syndrome 0xf9 00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00000030: 00 00 00 00 01 00 f9 05 01 00 08 f4 de 15 cf d2
References
Impacted products
Vendor Product Version
Linux Linux Version: db05815b36cbd486c86fd002dfa81c9af6245e25
Version: db05815b36cbd486c86fd002dfa81c9af6245e25
 Create a notification for this product.
Show details on NVD website

To clipboard 

{
  "containers": {
    "cna": {
      "affected": [
        {
          "defaultStatus": "unaffected",
          "product": "Linux",
          "programFiles": [
            "drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "lessThan": "8d3b91e7d81000d295cd914d4d9d6f860252e2bf",
              "status": "affected",
              "version": "db05815b36cbd486c86fd002dfa81c9af6245e25",
              "versionType": "git"
            },
            {
              "lessThan": "c326f9c68921e2f14dfcecb2f6b4216313d50248",
              "status": "affected",
              "version": "db05815b36cbd486c86fd002dfa81c9af6245e25",
              "versionType": "git"
            }
          ]
        },
        {
          "defaultStatus": "affected",
          "product": "Linux",
          "programFiles": [
            "drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c"
          ],
          "repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
          "vendor": "Linux",
          "versions": [
            {
              "status": "affected",
              "version": "5.3"
            },
            {
              "lessThan": "5.3",
              "status": "unaffected",
              "version": "0",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "7.0.*",
              "status": "unaffected",
              "version": "7.0.11",
              "versionType": "semver"
            },
            {
              "lessThanOrEqual": "*",
              "status": "unaffected",
              "version": "7.1",
              "versionType": "original_commit_for_fix"
            }
          ]
        }
      ],
      "cpeApplicability": [
        {
          "nodes": [
            {
              "cpeMatch": [
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.0.11",
                  "versionStartIncluding": "5.3",
                  "vulnerable": true
                },
                {
                  "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                  "versionEndExcluding": "7.1",
                  "versionStartIncluding": "5.3",
                  "vulnerable": true
                }
              ],
              "negate": false,
              "operator": "OR"
            }
          ]
        }
      ],
      "descriptions": [
        {
          "lang": "en",
          "value": "In the Linux kernel, the following vulnerability has been resolved:\n\nnet/mlx5e: xsk: Fix unlocked writing to ICOSQ\n\nDuring napi poll, when the affinity changes and there\u0027s still XSK work\nto be done, we trigger an ICOSQ interrupt on the new CPU. However, this\ntriggering on the ICOSQ is done unprotected.\n\nThere are 2 such races:\n\nA) mlx5e_trigger_irq() is called while mlx5e_xsk_alloc_rx_mpwqe() is\nrunning from a different CPU due to affinity change. This can happen\nbecause IRQ triggering is done after napi_complete_done(). At this point\nthe NAPI can be scheduled on a different CPU. Like this:\n\n  CPU A (old affinity, NAPI tail)    CPU B (new affinity, fresh NAPI)\n  -------------------------------    --------------------------------\n  napi_complete_done()  clears SCHED\n  mlx5e_cq_arm(...)\n                                     napi_schedule_prep() sets SCHED\n                                     mlx5e_napi_poll()\n                                       mlx5e_xsk_alloc_rx_mpwqe()\n                                         mlx5e_icosq_sync_lock() // noop\n                                         memcpy 640 B UMR body\n                                         advance sq-\u003epc by 10\n  mlx5e_trigger_irq(\u0026c-\u003eicosq)\n    wqe_info[pi] = {NOP, 1}\n    mlx5e_post_nop() advances sq-\u003epc\n\nB) mlx5e_trigger_irq() is called on the ICOSQ when\nmlx5e_trigger_napi_icosq() is running.\n\nThe obvious fix would be to lock the ICOSQ. But ICOSQ has an optimized\nlocking scheme that doesn\u0027t work for this scenario. Kick the async ICOSQ\ninstead which is always locked.\n\nThis issue was noticed in the wild with the following splat:\n\n  netdevice: ge-0-0-1: Bad OP in ICOSQ CQE: 0xd\n  WARNING: drivers/net/ethernet/mellanox/mlx5/core/en_rx.c:826 [...]\n  [...]\n  Call Trace:\n   \u003cIRQ\u003e\n   mlx5e_napi_poll+0x11d/0x7f0 [mlx5_core]\n   __napi_poll+0x30/0x200\n   ? skb_defer_free_flush+0x9c/0xc0\n   net_rx_action+0x2fe/0x3f0\n   handle_softirqs+0xd8/0x340\n   __irq_exit_rcu+0xbc/0xe0\n   common_interrupt+0x85/0xa0\n   \u003c/IRQ\u003e\n   \u003cTASK\u003e\n   asm_common_interrupt+0x26/0x40\n  [...]\n  ---[ end trace 0000000000000000 ]---\n  mlx5_core 0000:08:00.0 ge-0-0-1: Error cqe on cqn 0x548, ci 0x2022, qn 0x8f4,\n  opcode 0xd, syndrome 0x2, vendor syndrome 0x68\n  00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000030: 00 00 00 00 01 00 68 02 01 00 08 f4 de 14 59 d2\n  WQE DUMP: WQ size 16384 WQ cur size 0, WQE index 0x1e14, len: 64\n  00000000: 00 00 00 01 d9 ed 80 02 00 00 00 01 d9 ed 90 02\n  00000010: 00 00 00 01 d9 ed a0 02 00 00 00 01 d9 ed b0 02\n  00000020: 00 00 00 01 d9 ed c0 02 00 00 00 01 d9 ed d0 02\n  00000030: 00 00 00 01 d9 ed e0 02 00 00 00 01 d9 ed f0 02\n  mlx5_core 0000:08:00.0 ge-0-0-1: Error cqe on cqn 0x548, ci 0x2023, qn 0x8f4,\n  opcode 0xd, syndrome 0x5, vendor syndrome 0xf9\n  00000000: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n  00000030: 00 00 00 00 01 00 f9 05 01 00 08 f4 de 15 cf d2"
        }
      ],
      "metrics": [
        {
          "cvssV3_1": {
            "baseScore": 7.5,
            "baseSeverity": "HIGH",
            "vectorString": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
            "version": "3.1"
          },
          "scenarios": [
            {
              "lang": "en",
              "value": "AV:N - On an internet-facing mlx5 interface using zero-copy AF_XDP, unauthenticated packets drive NIC completions into mlx5e_napi_poll and can keep XSK work busy while another NAPI run starts. AF_XDP/XDP setup is a victim-side deployment prerequisite.\nAC:L - Sustained traffic in a multicore busy-polling or IRQ-balanced deployment can repeatedly produce overlapping NAPI runs after napi_complete_done(). The attacker can retry until the shared-ICOSQ producer and descriptor race is hit.\nPR:N - The remote sender needs no account, authentication, or capability. CAP_NET_RAW and CAP_NET_ADMIN are required only by the victim-side process or administrator configuring AF_XDP/XDP.\nUI:N - After the network service is configured, packets automatically trigger the completion, NAPI, XSK refill, and ICOSQ paths without user action.\nS:U - The corrupted ICOSQ and affected receive queues remain within the host kernel and mlx5 device security authority, without crossing a VM, IOMMU, or other scope boundary.\nC:N - The race corrupts kernel-constructed queue descriptors and producer metadata but provides no path that returns host memory to the attacker or supplies an arbitrary-read primitive.\nI:N - The competing writes contain only kernel-constructed NOP and registered UMR data; no attacker-selected host address or persistent protected-data modification is established.\nA:H - The race can produce malformed ICOSQ work requests, error CQEs, queue shutdown and recovery, or an oops from inconsistent software metadata. Repeated triggering can keep the receive path and associated network service unavailable."
            }
          ]
        }
      ],
      "providerMetadata": {
        "dateUpdated": "2026-08-05T12:40:03.069Z",
        "orgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
        "shortName": "Linux"
      },
      "references": [
        {
          "url": "https://git.kernel.org/stable/c/8d3b91e7d81000d295cd914d4d9d6f860252e2bf"
        },
        {
          "url": "https://git.kernel.org/stable/c/c326f9c68921e2f14dfcecb2f6b4216313d50248"
        }
      ],
      "title": "net/mlx5e: xsk: Fix unlocked writing to ICOSQ",
      "x_generator": {
        "engine": "bippy-1.2.0"
      }
    }
  },
  "cveMetadata": {
    "assignerOrgId": "416baaa9-dc9f-4396-8d5f-8c081fb06d67",
    "assignerShortName": "Linux",
    "cveId": "CVE-2026-64210",
    "datePublished": "2026-07-24T15:23:01.328Z",
    "dateReserved": "2026-07-19T15:36:31.769Z",
    "dateUpdated": "2026-08-05T12:40:03.069Z",
    "state": "PUBLISHED"
  },
  "dataType": "CVE_RECORD",
  "dataVersion": "5.2"
}


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.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • 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.