{ "schema_version": "1.6.1", "id": "RPI-CVE-2025-38734", "modified": "2026-10-09T11:39:44Z", "upstream": [ "CVE-2025-38734" ], "published": "2025-09-05T17:22:05Z", "summary": "net/smc: fix UAF on smcsk after smc_listen_out()", "details": "In the Linux kernel, the following vulnerability has been resolved:\n\nnet/smc: fix UAF on smcsk after smc_listen_out()\n\nBPF CI testing report a UAF issue:\n\n [ 16.446633] BUG: kernel NULL pointer dereference, address: 000000000000003 0\n [ 16.447134] #PF: supervisor read access in kernel mod e\n [ 16.447516] #PF: error_code(0x0000) - not-present pag e\n [ 16.447878] PGD 0 P4D 0\n [ 16.448063] Oops: Oops: 0000 [#1] PREEMPT SMP NOPT I\n [ 16.448409] CPU: 0 UID: 0 PID: 9 Comm: kworker/0:1 Tainted: G OE 6.13.0-rc3-g89e8a75fda73-dirty #4 2\n [ 16.449124] Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODUL E\n [ 16.449502] Hardware name: QEMU Ubuntu 24.04 PC (i440FX + PIIX, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/201 4\n [ 16.450201] Workqueue: smc_hs_wq smc_listen_wor k\n [ 16.450531] RIP: 0010:smc_listen_work+0xc02/0x159 0\n [ 16.452158] RSP: 0018:ffffb5ab40053d98 EFLAGS: 0001024 6\n [ 16.452526] RAX: 0000000000000001 RBX: 0000000000000002 RCX: 000000000000030 0\n [ 16.452994] RDX: 0000000000000280 RSI: 00003513840053f0 RDI: 000000000000000 0\n [ 16.453492] RBP: ffffa097808e3800 R08: ffffa09782dba1e0 R09: 000000000000000 5\n [ 16.453987] R10: 0000000000000000 R11: 0000000000000000 R12: ffffa0978274640 0\n [ 16.454497] R13: 0000000000000000 R14: 0000000000000000 R15: ffffa09782d4092 0\n [ 16.454996] FS: 0000000000000000(0000) GS:ffffa097bbc00000(0000) knlGS:000000000000000 0\n [ 16.455557] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003 3\n [ 16.455961] CR2: 0000000000000030 CR3: 0000000102788004 CR4: 0000000000770ef 0\n [ 16.456459] PKRU: 5555555 4\n [ 16.456654] Call Trace :\n [ 16.456832] \n [ 16.456989] ? __die+0x23/0x7 0\n [ 16.457215] ? page_fault_oops+0x180/0x4c 0\n [ 16.457508] ? __lock_acquire+0x3e6/0x249 0\n [ 16.457801] ? exc_page_fault+0x68/0x20 0\n [ 16.458080] ? asm_exc_page_fault+0x26/0x3 0\n [ 16.458389] ? smc_listen_work+0xc02/0x159 0\n [ 16.458689] ? smc_listen_work+0xc02/0x159 0\n [ 16.458987] ? lock_is_held_type+0x8f/0x10 0\n [ 16.459284] process_one_work+0x1ea/0x6d 0\n [ 16.459570] worker_thread+0x1c3/0x38 0\n [ 16.459839] ? __pfx_worker_thread+0x10/0x1 0\n [ 16.460144] kthread+0xe0/0x11 0\n [ 16.460372] ? __pfx_kthread+0x10/0x1 0\n [ 16.460640] ret_from_fork+0x31/0x5 0\n [ 16.460896] ? __pfx_kthread+0x10/0x1 0\n [ 16.461166] ret_from_fork_asm+0x1a/0x3 0\n [ 16.461453] \n [ 16.461616] Modules linked in: bpf_testmod(OE) [last unloaded: bpf_testmod(OE) ]\n [ 16.462134] CR2: 000000000000003 0\n [ 16.462380] ---[ end trace 0000000000000000 ]---\n [ 16.462710] RIP: 0010:smc_listen_work+0xc02/0x1590\n\nThe direct cause of this issue is that after smc_listen_out_connected(),\nnewclcsock->sk may be NULL since it will releases the smcsk. Therefore,\nif the application closes the socket immediately after accept,\nnewclcsock->sk can be NULL. A possible execution order could be as\nfollows:\n\nsmc_listen_work | userspace\n-----------------------------------------------------------------\nlock_sock(sk) |\nsmc_listen_out_connected() |\n| \\- smc_listen_out |\n| | \\- release_sock |\n | |- sk->sk_data_ready() |\n | fd = accept();\n | close(fd);\n | \\- socket->sk = NULL;\n/* newclcsock->sk is NULL now */\nSMC_STAT_SERV_SUCC_INC(sock_net(newclcsock->sk))\n\nSince smc_listen_out_connected() will not fail, simply swapping the order\nof the code can easily fix this issue.", "severity": [ { "type": "CVSS_V3", "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/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": null, "compiled_filter": false } } } } ], "references": [ { "type": "ADVISORY", "url": "https://www.cve.org/CVERecord?id=CVE-2025-38734" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/070b4af44c4b6e4c35fb1ca7001a6a88fd2d318f" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/2e765ba0ee0eae35688b443e97108308a716773e" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/85545f1525f9fa9bf44fec77ba011024f15da342" }, { "type": "FIX", "url": "https://git.kernel.org/stable/c/d9cef55ed49117bd63695446fb84b4b91815c0b4" } ], "database_specific": { "source": "https://git.kernel.org/pub/scm/linux/security/vulns.git", "upstream_modified": "2026-08-05T12:19:11Z", "severity": { "source": "kernel CNA", "cvss_v3_score": 9.8, "cvss_v3_severity": "Critical", "scenario": "AV:N - The vulnerable `smc_listen_work()` handler is scheduled and driven entirely by an inbound TCP connection carrying the SMC experimental option and the subsequent CLC handshake messages from a remote peer, making this a net/ stack bug reachable via received packets.\nAC:L - The attacker can open unlimited concurrent SMC connections and retry the race indefinitely, and in the loopback-ism/local-peer scenario controls both sides outright (the connect and the accept()+close()); the race was reproduced reliably enough to be caught by CI.\nPR:N - The SMC CLC handshake performs no authentication whatsoever \u2014 the connection is accepted and the vulnerable worker runs before any credential check, so an unauthenticated remote peer reaches the code.\nUI:N - No deliberate victim action is required; the server application's normal accept()/close() sequence is what races the worker, and the attacker triggers the handshake itself.\nS:U - The use-after-free and resulting corruption are confined to kernel memory within the same security authority; no VM, IOMMU, or sandbox boundary is crossed.\nC:H - This is a use-after-free read of a freed `struct socket`, and the stale value is dereferenced as a `struct sock *` and chased through `sk_net` \u2014 with the freed slab reclaimed under attacker influence this yields arbitrary kernel-memory reads.\nI:H - The pointer read from the freed object is used as the base for `this_cpu_inc()` in SMC_STAT_SERV_SUCC_INC, turning the UAF into an attacker-influenceable kernel write primitive suitable for heap grooming and control-flow hijacking.\nA:H - The bug reproducibly oopses the smc_hs_wq kworker with a supervisor page fault (confirmed in the commit's trace), and a remote attacker can trigger it repeatedly to panic or destabilize the system." } } }