GHSA-3J8G-7XHM-68VV
Vulnerability from github – Published: 2026-08-10 15:33 – Updated: 2026-08-10 15:33In the Linux kernel, the following vulnerability has been resolved:
usb: core: port: Deattach Type-C connector on component unbind
connector_unbind() is the mirror of connector_bind(), but it is missing the symmetric call to typec_deattach() that connector_bind() makes via:
if (port_dev->child)
typec_attach(port_dev->connector, &port_dev->child->dev);
When a Thunderbolt dock is unplugged, two teardown paths race:
-
The component framework calls connector_unbind() first, which sets port_dev->connector = NULL without calling typec_deattach(). This leaves port->usb2_dev/port->usb3_dev in struct typec_port pointing at the USB device that is about to be freed.
-
usb_disconnect() then calls typec_deattach(port_dev->connector, ...), but port_dev->connector is already NULL, so the call is a no-op and port->usb2_dev is never cleared.
-
Concurrently, UCSI detects a PD partner-disconnect event and calls typec_unregister_partner(), which reads port->usb2_dev (now a dangling pointer to freed memory) and passes it to typec_partner_unlink_device() -> sysfs_remove_link() -> dev_name() on the freed device, corrupting the typec/UCSI partner state.
This corruption leaves the Thunderbolt tunnel in an inconsistent state on the next dock hot-plug. On affected hardware the dock's I225/igc NIC fails to enumerate: AER fires a slot reset while the igc driver is still initialising ("PCIe link lost"), and the subsequent igc_reset attempt hits igc_rd32 on an already-detached device:
igc 0000:2e:00.0 eth0: PCIe link lost, device now detached
igc: Failed to read reg 0x0!
WARNING: CPU: 9 PID: 129 at drivers/net/ethernet/intel/igc/igc_main.c:7005
igc_rd32+0xa4/0xc0 [igc]
Call Trace:
igc_disable_pcie_master+0x16/0xa0 [igc]
igc_reset_hw_base+0x14/0x170 [igc]
igc_reset+0x63/0x110 [igc]
igc_io_slot_reset+0x9e/0xd0 [igc]
report_slot_reset+0x5d/0xc0
pcie_do_recovery+0x209/0x400
aer_isr_one_error_type+0x235/0x430
aer_isr+0x4e/0x80
irq_thread+0xf4/0x1f0
- UCSI later handles the PD partner-disconnect and calls typec_unregister_partner(), which still sees the stale port->usb2_dev and tries to remove its sysfs link a second time:
kernfs: can not remove 'typec', no directory WARNING: CPU: 6 PID: 55 at fs/kernfs/dir.c:1706 kernfs_remove_by_name_ns+0xe9/0xf0 Workqueue: events ucsi_handle_connector_change [typec_ucsi] Call Trace: sysfs_remove_link+0x19/0x50 typec_unregister_partner+0x6e/0x120 [typec] ucsi_unregister_partner+0x107/0x150 [typec_ucsi] ucsi_handle_connector_change+0x3ec/0x490 [typec_ucsi] process_one_work+0x18e/0x3e0 worker_thread+0x2e3/0x420 kthread+0x10a/0x230 ret_from_fork+0x121/0x140 ret_from_fork_asm+0x1a/0x30
With worse timing the same stale pointer is dereferenced after the backing memory is freed, turning the warning into a use-after-free.
Fix the asymmetry: call typec_deattach() before clearing port_dev->connector, matching what connector_bind() does on the bind side. typec_partner_deattach() is already protected by port->partner_link_lock, so it serialises safely with the concurrent typec_unregister_partner() path.
{
"affected": [],
"aliases": [
"CVE-2026-68372"
],
"database_specific": {
"cwe_ids": [],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-08-10T13:20:30Z",
"severity": null
},
"details": "In the Linux kernel, the following vulnerability has been resolved:\n\nusb: core: port: Deattach Type-C connector on component unbind\n\nconnector_unbind() is the mirror of connector_bind(), but it is missing\nthe symmetric call to typec_deattach() that connector_bind() makes via:\n\n if (port_dev-\u003echild)\n typec_attach(port_dev-\u003econnector, \u0026port_dev-\u003echild-\u003edev);\n\nWhen a Thunderbolt dock is unplugged, two teardown paths race:\n\n1. The component framework calls connector_unbind() first, which sets\n port_dev-\u003econnector = NULL without calling typec_deattach(). This\n leaves port-\u003eusb2_dev/port-\u003eusb3_dev in struct typec_port pointing at\n the USB device that is about to be freed.\n\n2. usb_disconnect() then calls typec_deattach(port_dev-\u003econnector, ...),\n but port_dev-\u003econnector is already NULL, so the call is a no-op and\n port-\u003eusb2_dev is never cleared.\n\n3. Concurrently, UCSI detects a PD partner-disconnect event and calls\n typec_unregister_partner(), which reads port-\u003eusb2_dev (now a dangling\n pointer to freed memory) and passes it to typec_partner_unlink_device()\n -\u003e sysfs_remove_link() -\u003e dev_name() on the freed device, corrupting\n the typec/UCSI partner state.\n\nThis corruption leaves the Thunderbolt tunnel in an inconsistent state on\nthe next dock hot-plug. On affected hardware the dock\u0027s I225/igc NIC fails\nto enumerate: AER fires a slot reset while the igc driver is still\ninitialising (\"PCIe link lost\"), and the subsequent igc_reset attempt hits\nigc_rd32 on an already-detached device:\n\n igc 0000:2e:00.0 eth0: PCIe link lost, device now detached\n igc: Failed to read reg 0x0!\n WARNING: CPU: 9 PID: 129 at drivers/net/ethernet/intel/igc/igc_main.c:7005\n igc_rd32+0xa4/0xc0 [igc]\n Call Trace:\n igc_disable_pcie_master+0x16/0xa0 [igc]\n igc_reset_hw_base+0x14/0x170 [igc]\n igc_reset+0x63/0x110 [igc]\n igc_io_slot_reset+0x9e/0xd0 [igc]\n report_slot_reset+0x5d/0xc0\n pcie_do_recovery+0x209/0x400\n aer_isr_one_error_type+0x235/0x430\n aer_isr+0x4e/0x80\n irq_thread+0xf4/0x1f0\n\n4. UCSI later handles the PD partner-disconnect and calls\n typec_unregister_partner(), which still sees the stale port-\u003eusb2_dev\n and tries to remove its sysfs link a second time:\n\n kernfs: can not remove \u0027typec\u0027, no directory\n WARNING: CPU: 6 PID: 55 at fs/kernfs/dir.c:1706 kernfs_remove_by_name_ns+0xe9/0xf0\n Workqueue: events ucsi_handle_connector_change [typec_ucsi]\n Call Trace:\n sysfs_remove_link+0x19/0x50\n typec_unregister_partner+0x6e/0x120 [typec]\n ucsi_unregister_partner+0x107/0x150 [typec_ucsi]\n ucsi_handle_connector_change+0x3ec/0x490 [typec_ucsi]\n process_one_work+0x18e/0x3e0\n worker_thread+0x2e3/0x420\n kthread+0x10a/0x230\n ret_from_fork+0x121/0x140\n ret_from_fork_asm+0x1a/0x30\n\n With worse timing the same stale pointer is dereferenced after the\n backing memory is freed, turning the warning into a use-after-free.\n\nFix the asymmetry: call typec_deattach() before clearing\nport_dev-\u003econnector, matching what connector_bind() does on the bind side.\ntypec_partner_deattach() is already protected by port-\u003epartner_link_lock,\nso it serialises safely with the concurrent typec_unregister_partner() path.",
"id": "GHSA-3j8g-7xhm-68vv",
"modified": "2026-08-10T15:33:48Z",
"published": "2026-08-10T15:33:48Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-68372"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/7714fb896ed308cf13d32d317040adc4f200b8e4"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/78d361e60caf1999d51bda0e1b1004f5d39fcfbc"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/e00109b5adf71635919248e9ab6300a662e6a3e8"
},
{
"type": "WEB",
"url": "https://git.kernel.org/stable/c/e0b291fe117964037e0ba382eff4bb365d531c3a"
}
],
"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.