{
  "schema_version": "1.6.1",
  "id": "RPI-CVE-2024-35971",
  "modified": "2026-10-09T11:39:44Z",
  "upstream": [
    "CVE-2024-35971"
  ],
  "published": "2024-05-20T09:42:41Z",
  "summary": "net: ks8851: Handle softirqs at the end of IRQ thread to fix hang",
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nnet: ks8851: Handle softirqs at the end of IRQ thread to fix hang\n\nThe ks8851_irq() thread may call ks8851_rx_pkts() in case there are\nany packets in the MAC FIFO, which calls netif_rx(). This netif_rx()\nimplementation is guarded by local_bh_disable() and local_bh_enable().\nThe local_bh_enable() may call do_softirq() to run softirqs in case\nany are pending. One of the softirqs is net_rx_action, which ultimately\nreaches the driver .start_xmit callback. If that happens, the system\nhangs. The entire call chain is below:\n\nks8851_start_xmit_par from netdev_start_xmit\nnetdev_start_xmit from dev_hard_start_xmit\ndev_hard_start_xmit from sch_direct_xmit\nsch_direct_xmit from __dev_queue_xmit\n__dev_queue_xmit from __neigh_update\n__neigh_update from neigh_update\nneigh_update from arp_process.constprop.0\narp_process.constprop.0 from __netif_receive_skb_one_core\n__netif_receive_skb_one_core from process_backlog\nprocess_backlog from __napi_poll.constprop.0\n__napi_poll.constprop.0 from net_rx_action\nnet_rx_action from __do_softirq\n__do_softirq from call_with_stack\ncall_with_stack from do_softirq\ndo_softirq from __local_bh_enable_ip\n__local_bh_enable_ip from netif_rx\nnetif_rx from ks8851_irq\nks8851_irq from irq_thread_fn\nirq_thread_fn from irq_thread\nirq_thread from kthread\nkthread from ret_from_fork\n\nThe hang happens because ks8851_irq() first locks a spinlock in\nks8851_par.c ks8851_lock_par() spin_lock_irqsave(&ksp->lock, ...)\nand with that spinlock locked, calls netif_rx(). Once the execution\nreaches ks8851_start_xmit_par(), it calls ks8851_lock_par() again\nwhich attempts to claim the already locked spinlock again, and the\nhang happens.\n\nMove the do_softirq() call outside of the spinlock protected section\nof ks8851_irq() by disabling BHs around the entire spinlock protected\nsection of ks8851_irq() handler. Place local_bh_enable() outside of\nthe spinlock protected section, so that it can trigger do_softirq()\nwithout the ks8851_par.c ks8851_lock_par() spinlock being held, and\nsafely call ks8851_start_xmit_par() without attempting to lock the\nalready locked spinlock.\n\nSince ks8851_irq() is protected by local_bh_disable()/local_bh_enable()\nnow, replace netif_rx() with __netif_rx() which is not duplicating the\nlocal_bh_disable()/local_bh_enable() calls.",
  "severity": [
    {
      "type": "CVSS_V3",
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H"
    }
  ],
  "affected": [
    {
      "package": {
        "ecosystem": "Raspberry Pi OS:12",
        "name": "linux"
      },
      "versions": [
        "1:6.1.47-1+rpt4"
      ],
      "ecosystem_specific": {
        "binaries": [
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-2712",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v6",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v7",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v7l",
            "binary_version": "1:6.1.47-1+rpt4"
          },
          {
            "binary_name": "linux-image-6.1.0-rpi3-rpi-v8",
            "binary_version": "1:6.1.47-1+rpt4"
          }
        ]
      },
      "database_specific": {
        "per_version": {
          "1:6.1.47-1+rpt4": {
            "linux_commit": "655fc658a15ae7a6f37103754adb39ba52a9a14e",
            "upstream_version": "6.1.47",
            "assessment": "commit ancestry (strak rule)",
            "upstream_fixed": "6.1.87",
            "compiled_filter": false
          }
        }
      }
    }
  ],
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://www.cve.org/CVERecord?id=CVE-2024-35971"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/492337a4fbd1421b42df684ee9b34be2a2722540"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/cba376eb036c2c20077b41d47b317d8218fe754f"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/49d5d70538b6b8f2a3f8f1ac30c1f921d4a0929b"
    },
    {
      "type": "FIX",
      "url": "https://git.kernel.org/stable/c/be0384bf599cf1eb8d337517feeb732d71f75a6f"
    }
  ],
  "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": 7.5,
      "cvss_v3_severity": "High",
      "scenario": "AV:N - The deadlock is triggered purely by a received Ethernet frame whose RX processing emits an immediate reply in softirq context \u2014 ICMP echo requests, TCP SYNs eliciting SYN-ACKs, and forwarded traffic all do this and are routable from an arbitrary remote host, not just the local segment. Per net/ stack guidance, bugs reachable via received packets are Network.\nAC:L - The attacker fully controls the trigger by sending a single packet that elicits a reply (ping, ARP request, TCP SYN); no race, memory layout, or victim state is involved, and the hang even occurs under ordinary legitimate traffic.\nPR:N - The vulnerable code runs in the NIC's threaded IRQ handler on raw packet reception, entirely before any socket, authentication, or authorization check \u2014 an unauthenticated remote attacker needs no credentials on the target.\nUI:N - No victim action is required; the packet is processed automatically by the driver's interrupt thread as soon as it arrives on the wire.\nS:U - The recursive spinlock acquisition deadlocks the kernel itself, with no crossing of a virtualization, IOMMU, or sandbox boundary \u2014 impact stays within the same security authority.\nC:N - The bug is a lock-recursion deadlock with no out-of-bounds access, no use-after-free, and no leaked data; the fix only reorders BH handling and switches to __netif_rx(), confirming no information is exposed.\nI:N - No memory is corrupted and no data is modified \u2014 the CPU simply spins forever on a lock it already holds, giving no write primitive of any kind.\nA:H - Re-acquiring the non-recursive spinlock taken with spin_lock_irqsave() wedges the CPU with local interrupts disabled, producing a full system hang (unrecoverable without watchdog reset on the typical single-core embedded boards using the KS8851 parallel bus)."
    }
  }
}