● CVE-2024-26584 / High Severity

CVE-2024-26584: Linux Kernel TLS Subsystem Use-After-Free Teardown

How an asynchronous cryptographic callback race condition in net/tls/tls_sw.c freed network pages while still queued for zero-copy transmission.

💡 Explication en Termes Simples (ELI5)

Imagine you order a meal for delivery. The chef starts cooking, but you call and cancel the order. Because the kitchen staff wasn't notified properly, the delivery driver picks up an empty plate, drives to a new customer's house, and serves them whatever was left on the counter, contaminating the new customer's food.

Concepts Clés et Termes du Sous-système

Kernel TLS (kTLS): Linux kernel facility to perform symmetric TLS encryption/decryption directly inside the network socket layer for maximum throughput.
Zero-Copy Networking: Transmitting data directly from application buffers to the network card without intermediate CPU memory copying.
Asynchronous Crypto (aead_request): Offloading cryptographic AES-GCM operations to asynchronous hardware accelerators.
Slab Corruption: Corrupting Linux kernel slab cache structures (`kmalloc-512`), destabilizing kernel execution.

Mécanique d'Exécution Étape par Étape

Step 1

1. Open kTLS Socket

The attacker creates a TLS socket and enables kernel encryption offload via setsockopt(TCP_ULP, "tls").

Step 2

2. Submit Asynchronous Zero-Copy Payload

The attacker sends data using MSG_ZEROCOPY, queuing pages for hardware crypto.

Step 3

3. Abrupt Socket Teardown

The attacker closes the socket before the crypto accelerator completes its async callback.

Step 4

4. Kernel Memory Corruption

The cleanup handler frees memory pages that the delayed hardware callback subsequently overwrites.

Code Source : Faille Critique vs. Correctif Sécurisé

Fourni en code source de haut niveau lisible (sans assembleur brut ni diff binaire).

FAILLE NON CORRIGÉE

// VULNERABLE: net/tls/tls_sw.c before patch
static void tls_encrypt_done(void *data, int err) {
    struct tls_context *ctx = data;
    struct tls_sw_context_tx *ctx_tx = tls_sw_ctx_tx(ctx);

    // ROOT CAUSE:
    // Asynchronous completion callback assumes socket context is still locked!
    // If the socket was closed while crypto was in flight, ctx_tx is already freed!
    clear_bit(TLS_TX_SYNC_MORE, &ctx_tx->tx_bitmask);
    tls_free_open_rec(ctx);
}

CORRECTIF RENFORCÉ

// SECURE: net/tls/tls_sw.c patch
static void tls_encrypt_done(void *data, int err) {
    struct tls_context *ctx = data;
    
    // 1. Verify context reference counter before dereferencing context pointers
    if (!refcount_inc_not_zero(&ctx->refcount)) {
        return; // Socket is already dying, abort callback safely!
    }
    
    struct tls_sw_context_tx *ctx_tx = tls_sw_ctx_tx(ctx);
    clear_bit(TLS_TX_SYNC_MORE, &ctx_tx->tx_bitmask);
    tls_free_open_rec(ctx);
    
    // 2. Drop reference cleanly
    refcount_dec(&ctx->refcount);
}

Checklist d'Ingénierie et de Durcissement Système

Use `refcount_inc_not_zero()` on shared context handles inside asynchronous completion handlers.
Synchronously drain or cancel all in-flight asynchronous hardware crypto jobs before releasing socket memory.
Verify thread-safety and lock ownership inside callback routines.
Audit zero-copy socket implementations for lifetime and reference races.

← Parcourir l'Annuaire de Sécurité Tous les Bulletins de Sécurité →