EN

Tracker / CVE-2025-38154

CVE-2025-38154

Alta 7.8

In the Linux kernel, the following vulnerability has been resolved: bpf, sockmap: Avoid using sk_socket after free when sending The sk->sk_socket is not locked or referenced in backlog thread, and during the call to skb_send_sock(), there is a race condition with the release of sk_socket. All types of sockets(tcp/udp/unix/vsock) will be affected. Race conditions: ''' CPU0 CPU1 backlog::skb_send_sock sendmsg_unlocked sock_sendmsg sock_sendmsg_nosec close(fd): ... ops->release() -> sock_map_close() sk_socket->ops = NULL free(socket) sock->ops->sendmsg ^ panic here ''' The ref of psock become 0 after sock_map_close() executed. ''' void sock_map_close() { ... if (likely(psock)) { ... // !! here we remove psock and the ref of psock become 0 sock_map_remove_links(sk, psock) psock = sk_psock_get(sk); if (unlikely(!psock)) goto no_psock; <=== Control jumps here via goto ... cancel_delayed_work_sync(&psock->work); <=== not executed sk_psock_put(sk, psock); ... } ''' Based on the fact that we already wait for the workqueue to finish in sock_map_close() if psock is held, we simply increase the psock reference count to avoid race conditions. With this patch, if the backlog thread is running, sock_map_close() will wait for the backlog thread to complete and cancel all pending work. If no backlog running, any pending work that hasn't started by then will fail when invoked by sk_psock_get(), as the psock reference count have been zeroed, and sk_psock_drop() will cancel all jobs via cancel_delayed_work_sync(). In summary, we require synchronization to coordinate the backlog thread and close() thread. The panic I catched: ''' Workqueue: events sk_psock_backlog RIP: 0010:sock_sendmsg+0x21d/0x440 RAX: 0000000000000000 RBX: ffffc9000521fad8 RCX: 0000000000000001 ... Call Trace: <TASK> ? die_addr+0x40/0xa0 ? exc_general_protection+0x14c/0x230 ? asm_exc_general_protection+0x26/0x30 ? sock_sendmsg+0x21d/0x440 ? sock_sendmsg+0x3e0/0x440 ? __pfx_sock_sendmsg+0x10/0x10 __skb_send_sock+0x543/0xb70 sk_psock_backlog+0x247/0xb80 ... '''

Prodotti e versioni affette

debian debian_linux
linux linux_kernel
linux linux_kernel · 5.15.162 → 5.15.186
linux linux_kernel · 6.1.95 → 6.1.142
linux linux_kernel · 6.10.1 → 6.12.34
linux linux_kernel · 6.13 → 6.15.3
linux linux_kernel · 6.6.35 → 6.6.94
linux linux_kernel · 6.9.6 → 6.10

Analisi

Questa pagina non è ancora indicizzabile.Finché non contiene un’analisi originale — cosa espone davvero, come verificare in due minuti se un sistema è stato toccato, cosa fare se lo è stato — la pagina resta noindex. È il database a deciderlo, non il modello di pagina.

Riferimenti