ghsa-m8cc-c87r-9qr2
Vulnerability from github
Published
2025-10-01 09:30
Modified
2025-10-01 09:30
Details

In the Linux kernel, the following vulnerability has been resolved:

net: phy: transfer phy_config_inband() locking responsibility to phylink

Problem description

Lockdep reports a possible circular locking dependency (AB/BA) between &pl->state_mutex and &phy->lock, as follows.

phylink_resolve() // acquires &pl->state_mutex -> phylink_major_config() -> phy_config_inband() // acquires &pl->phydev->lock

whereas all the other call sites where &pl->state_mutex and &pl->phydev->lock have the locking scheme reversed. Everywhere else, &pl->phydev->lock is acquired at the top level, and &pl->state_mutex at the lower level. A clear example is phylink_bringup_phy().

The outlier is the newly introduced phy_config_inband() and the existing lock order is the correct one. To understand why it cannot be the other way around, it is sufficient to consider phylink_phy_change(), phylink's callback from the PHY device's phy->phy_link_change() virtual method, invoked by the PHY state machine.

phy_link_up() and phy_link_down(), the (indirect) callers of phylink_phy_change(), are called with &phydev->lock acquired. Then phylink_phy_change() acquires its own &pl->state_mutex, to serialize changes made to its pl->phy_state and pl->link_config. So all other instances of &pl->state_mutex and &phydev->lock must be consistent with this order.

Problem impact

I think the kernel runs a serious deadlock risk if an existing phylink_resolve() thread, which results in a phy_config_inband() call, is concurrent with a phy_link_up() or phy_link_down() call, which will deadlock on &pl->state_mutex in phylink_phy_change(). Practically speaking, the impact may be limited by the slow speed of the medium auto-negotiation protocol, which makes it unlikely for the current state to still be unresolved when a new one is detected, but I think the problem is there. Nonetheless, the problem was discovered using lockdep.

Proposed solution

Practically speaking, the phy_config_inband() requirement of having phydev->lock acquired must transfer to the caller (phylink is the only caller). There, it must bubble up until immediately before &pl->state_mutex is acquired, for the cases where that takes place.

Solution details, considerations, notes

This is the phy_config_inband() call graph:

                      sfp_upstream_ops :: connect_phy()
                      |
                      v
                      phylink_sfp_connect_phy()
                      |
                      v
                      phylink_sfp_config_phy()
                      |
                      |   sfp_upstream_ops :: module_insert()
                      |   |
                      |   v
                      |   phylink_sfp_module_insert()
                      |   |
                      |   |   sfp_upstream_ops :: module_start()
                      |   |   |
                      |   |   v
                      |   |   phylink_sfp_module_start()
                      |   |   |
                      |   v   v
                      |   phylink_sfp_config_optical()

phylink_start() | | | phylink_resume() v v | | phylink_sfp_set_config() | | | v v v phylink_mac_initial_config() | phylink_resolve() | | phylink_ethtool_ksettings_set() v v v phylink_major_config() | v phy_config_inband()

phylink_major_config() caller #1, phylink_mac_initial_config(), does not acquire &pl->state_mutex nor do its callers. It must acquire &pl->phydev->lock prior to calling phylink_major_config().

phylink_major_config() caller #2, phylink_resolve() acquires &pl->state_mutex, thus also needs to acquire &pl->phydev->lock.

phylink_major_config() caller #3, phylink_ethtool_ksettings_set(), is completely uninteresting, because it only call ---truncated---

Show details on source website


{
  "affected": [],
  "aliases": [
    "CVE-2025-39915"
  ],
  "database_specific": {
    "cwe_ids": [],
    "github_reviewed": false,
    "github_reviewed_at": null,
    "nvd_published_at": "2025-10-01T08:15:34Z",
    "severity": null
  },
  "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nnet: phy: transfer phy_config_inband() locking responsibility to phylink\n\nProblem description\n===================\n\nLockdep reports a possible circular locking dependency (AB/BA) between\n\u0026pl-\u003estate_mutex and \u0026phy-\u003elock, as follows.\n\nphylink_resolve() // acquires \u0026pl-\u003estate_mutex\n-\u003e phylink_major_config()\n   -\u003e phy_config_inband() // acquires \u0026pl-\u003ephydev-\u003elock\n\nwhereas all the other call sites where \u0026pl-\u003estate_mutex and\n\u0026pl-\u003ephydev-\u003elock have the locking scheme reversed. Everywhere else,\n\u0026pl-\u003ephydev-\u003elock is acquired at the top level, and \u0026pl-\u003estate_mutex at\nthe lower level. A clear example is phylink_bringup_phy().\n\nThe outlier is the newly introduced phy_config_inband() and the existing\nlock order is the correct one. To understand why it cannot be the other\nway around, it is sufficient to consider phylink_phy_change(), phylink\u0027s\ncallback from the PHY device\u0027s phy-\u003ephy_link_change() virtual method,\ninvoked by the PHY state machine.\n\nphy_link_up() and phy_link_down(), the (indirect) callers of\nphylink_phy_change(), are called with \u0026phydev-\u003elock acquired.\nThen phylink_phy_change() acquires its own \u0026pl-\u003estate_mutex, to\nserialize changes made to its pl-\u003ephy_state and pl-\u003elink_config.\nSo all other instances of \u0026pl-\u003estate_mutex and \u0026phydev-\u003elock must be\nconsistent with this order.\n\nProblem impact\n==============\n\nI think the kernel runs a serious deadlock risk if an existing\nphylink_resolve() thread, which results in a phy_config_inband() call,\nis concurrent with a phy_link_up() or phy_link_down() call, which will\ndeadlock on \u0026pl-\u003estate_mutex in phylink_phy_change(). Practically\nspeaking, the impact may be limited by the slow speed of the medium\nauto-negotiation protocol, which makes it unlikely for the current state\nto still be unresolved when a new one is detected, but I think the\nproblem is there. Nonetheless, the problem was discovered using lockdep.\n\nProposed solution\n=================\n\nPractically speaking, the phy_config_inband() requirement of having\nphydev-\u003elock acquired must transfer to the caller (phylink is the only\ncaller). There, it must bubble up until immediately before\n\u0026pl-\u003estate_mutex is acquired, for the cases where that takes place.\n\nSolution details, considerations, notes\n=======================================\n\nThis is the phy_config_inband() call graph:\n\n                          sfp_upstream_ops :: connect_phy()\n                          |\n                          v\n                          phylink_sfp_connect_phy()\n                          |\n                          v\n                          phylink_sfp_config_phy()\n                          |\n                          |   sfp_upstream_ops :: module_insert()\n                          |   |\n                          |   v\n                          |   phylink_sfp_module_insert()\n                          |   |\n                          |   |   sfp_upstream_ops :: module_start()\n                          |   |   |\n                          |   |   v\n                          |   |   phylink_sfp_module_start()\n                          |   |   |\n                          |   v   v\n                          |   phylink_sfp_config_optical()\n phylink_start()          |   |\n   |   phylink_resume()   v   v\n   |   |  phylink_sfp_set_config()\n   |   |  |\n   v   v  v\n phylink_mac_initial_config()\n   |   phylink_resolve()\n   |   |  phylink_ethtool_ksettings_set()\n   v   v  v\n   phylink_major_config()\n            |\n            v\n    phy_config_inband()\n\nphylink_major_config() caller #1, phylink_mac_initial_config(), does not\nacquire \u0026pl-\u003estate_mutex nor do its callers. It must acquire\n\u0026pl-\u003ephydev-\u003elock prior to calling phylink_major_config().\n\nphylink_major_config() caller #2, phylink_resolve() acquires\n\u0026pl-\u003estate_mutex, thus also needs to acquire \u0026pl-\u003ephydev-\u003elock.\n\nphylink_major_config() caller #3, phylink_ethtool_ksettings_set(), is\ncompletely uninteresting, because it only call\n---truncated---",
  "id": "GHSA-m8cc-c87r-9qr2",
  "modified": "2025-10-01T09:30:25Z",
  "published": "2025-10-01T09:30:25Z",
  "references": [
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2025-39915"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/052ac41c379c8b87629808be612a482b2d0ae283"
    },
    {
      "type": "WEB",
      "url": "https://git.kernel.org/stable/c/e2a10daba84968f6b5777d150985fd7d6abc9c84"
    }
  ],
  "schema_version": "1.4.0",
  "severity": []
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

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.


Loading…

Loading…