wifi: cfg80211: cancel wiphy_work before freeing wiphy

A wiphy_work can be queued from the moment the wiphy is allocated and
initialized (i.e. wiphy_new_nm). When a wiphy_work is queued, the
rdev::wiphy_work is getting queued.

If wiphy_free is called before the rdev::wiphy_work had a chance to run,
the wiphy memory will be freed, and then when it eventally gets to run
it'll use invalid memory.

Fix this by canceling the work before freeing the wiphy.

Fixes: a3ee4dc84c4e ("wifi: cfg80211: add a work abstraction with special semantics")
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://patch.msgid.link/20250306123626.efd1d19f6e07.I48229f96f4067ef73f5b87302335e2fd750136c9@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

diff --git a/net/wireless/core.c b/net/wireless/core.c
index 12b780de8779ce656a2ce78ae9d3886ab2924fbd..828e2987263351dbd57e6ebd5805f2bffa6c2533 100644 (file)
--- a/net/wireless/core.c
+++ b/net/wireless/core.c
@@ -1191,6 +1191,13 @@ void cfg80211_dev_free(struct cfg80211_registered_device *rdev)
 {
        struct cfg80211_internal_bss *scan, *tmp;
        struct cfg80211_beacon_registration *reg, *treg;
+       unsigned long flags;
+
+       spin_lock_irqsave(&rdev->wiphy_work_lock, flags);
+       WARN_ON(!list_empty(&rdev->wiphy_work_list));
+       spin_unlock_irqrestore(&rdev->wiphy_work_lock, flags);
+       cancel_work_sync(&rdev->wiphy_work);
+
        rfkill_destroy(rdev->wiphy.rfkill);
        list_for_each_entry_safe(reg, treg, &rdev->beacon_registrations, list) {
                list_del(&reg->list);