CVE-2026-53260

In the Linux kernel, the following vulnerability has been resolved:<br /> <br /> tcp: Add preempt_{disable,enable}_nested() in reqsk_queue_hash_req().<br /> <br /> syzbot reported a weird reqsk-&gt;rsk_refcnt underflow in<br /> __inet_csk_reqsk_queue_drop().<br /> <br /> The captured reqsk_put() in __inet_csk_reqsk_queue_drop()<br /> is called only when it successfully removes reqsk from ehash.<br /> <br /> Moreover, reqsk_timer_handler() calls another reqsk_put()<br /> after that.<br /> <br /> This indicates that the reqsk was missing both refcnts for<br /> ehash and the timer itself.<br /> <br /> Since all the syzbot reports had PREEMPT_RT enabled, the only<br /> possible scenario is that reqsk_queue_hash_req() is preempted<br /> after mod_timer() and before refcount_set(), and then the timer<br /> triggered after 1s aborts the reqsk due to its listener&amp;#39;s close().<br /> <br /> Let&amp;#39;s wrap mod_timer() and refcount_set() with<br /> preempt_disable_nested() and preempt_enable_nested().<br /> <br /> Note that inet_ehash_insert() holds the normal spin_lock()<br /> (mutex in PREEMPT_RT), so it must be called outside of<br /> preempt_disable_nested(), but this is fine.<br /> <br /> The lookup path just ignores 0 sk_refcnt entries in ehash<br /> and tries to create another reqsk, but this will fail at<br /> inet_ehash_insert().<br /> <br /> [0]:<br /> refcount_t: underflow; use-after-free.<br /> WARNING: lib/refcount.c:28 at refcount_warn_saturate+0xb2/0x110 lib/refcount.c:28, CPU#0: ktimers/0/16<br /> Modules linked in:<br /> CPU: 0 UID: 0 PID: 16 Comm: ktimers/0 Tainted: G L syzkaller #0 PREEMPT_{RT,(full)}<br /> Tainted: [L]=SOFTLOCKUP<br /> Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/18/2026<br /> RIP: 0010:refcount_warn_saturate+0xb2/0x110 lib/refcount.c:28<br /> Code: e4 7d d1 0a 67 48 0f b9 3a eb 4a e8 38 3d 23 fd 48 8d 3d e1 7d d1 0a 67 48 0f b9 3a eb 37 e8 25 3d 23 fd 48 8d 3d de 7d d1 0a 48 0f b9 3a eb 24 e8 12 3d 23 fd 48 8d 3d db 7d d1 0a 67 48 0f<br /> RSP: 0000:ffffc90000157948 EFLAGS: 00010246<br /> RAX: ffffffff84a1301b RBX: 0000000000000003 RCX: ffff88801ca98000<br /> RDX: 0000000000000100 RSI: 0000000000000000 RDI: ffffffff8f72ae00<br /> RBP: ffffffff99ae3b01 R08: ffff88801ca98000 R09: 0000000000000005<br /> R10: 0000000000000100 R11: 0000000000000004 R12: ffff8880425ef568<br /> R13: ffff8880425ef4f8 R14: ffff8880425ef578 R15: 0000000000000000<br /> FS: 0000000000000000(0000) GS:ffff888126386000(0000) knlGS:0000000000000000<br /> CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033<br /> CR2: 00007f7b46710e9c CR3: 000000000dbb6000 CR4: 00000000003526f0<br /> Call Trace:<br /> <br /> __refcount_sub_and_test include/linux/refcount.h:400 [inline]<br /> __refcount_dec_and_test include/linux/refcount.h:432 [inline]<br /> refcount_dec_and_test include/linux/refcount.h:450 [inline]<br /> reqsk_put include/net/request_sock.h:136 [inline]<br /> __inet_csk_reqsk_queue_drop+0x3ce/0x440 net/ipv4/inet_connection_sock.c:1007<br /> reqsk_timer_handler+0x651/0xdf0 net/ipv4/inet_connection_sock.c:1137<br /> call_timer_fn+0x192/0x5e0 kernel/time/timer.c:1748<br /> expire_timers kernel/time/timer.c:1799 [inline]<br /> __run_timers kernel/time/timer.c:2374 [inline]<br /> __run_timer_base+0x6a3/0x9f0 kernel/time/timer.c:2386<br /> run_timer_base kernel/time/timer.c:2395 [inline]<br /> run_timer_softirq+0x67/0x170 kernel/time/timer.c:2403<br /> handle_softirqs+0x1de/0x6d0 kernel/softirq.c:622<br /> __do_softirq kernel/softirq.c:656 [inline]<br /> run_ktimerd+0x69/0x100 kernel/softirq.c:1151<br /> smpboot_thread_fn+0x541/0xa50 kernel/smpboot.c:160<br /> kthread+0x388/0x470 kernel/kthread.c:436<br /> ret_from_fork+0x514/0xb70 arch/x86/kernel/process.c:158<br /> ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245<br />