CVE-2025-39758
Vulnérabilité critique dans Linux (Linux)
Description
In the Linux kernel, the following vulnerability has been resolved: RDMA/siw: Fix the sendmsg byte count in siw_tcp_sendpages Ever since commit c2ff29e99a76 ("siw: Inline do_tcp_sendpages()"), we have been doing this: static int siw_tcp_sendpages(struct socket *s, struct page **page, int offset, size_t size) [...] /* Calculate the number of bytes we need to push, for this page * specifically */ size_t bytes = min_t(size_t, PAGE_SIZE - offset, size); /* If we can't splice it, then copy it in, as normal */ if (!sendpage_ok(page[i])) msg.msg_flags &= ~MSG_SPLICE_PAGES; /* Set the bvec pointing to the page, with len $bytes */ bvec_set_page(&bvec, page[i], bytes, offset); /* Set the iter to $size, aka the size of the whole sendpages (!!!) */ iov_iter_bvec(&msg.msg_iter, ITER_SOURCE, &bvec, 1, size); try_page_again: lock_sock(sk); /* Sendmsg with $size size (!!!) */ rv = tcp_sendmsg_locked(sk, &msg, size); This means we've been sending oversized iov_iters and tcp_sendmsg calls for a while. This has a been a benign bug because sendpage_ok() always returned true. With the recent slab allocator changes being slowly introduced into next (that disallow sendpage on large kmalloc allocations), we have recently hit out-of-bounds crashes, due to slight differences in iov_iter behavior between the MSG_SPLICE_PAGES and "regular" copy paths: (MSG_SPLICE_PAGES) skb_splice_from_iter iov_iter_extract_pages iov_iter_extract_bvec_pages uses i->nr_segs to correctly stop in its tracks before OoB'ing everywhere skb_splice_from_iter gets a "short" read (!MSG_SPLICE_PAGES) skb_copy_to_page_nocache copy=iov_iter_count [...] copy_from_iter /* this doesn't help */ if (unlikely(iter->count < len)) len = iter->count; iterate_bvec ... and we run off the bvecs Fix this by properly setting the iov_iter's byte count, plus sending the correct byte count to tcp_sendmsg_locked.
Gravité et risque d'exploitation
CVE-2025-39758 obtient un score CVSS de 9.8/10, soit une gravité critique. Selon le modèle EPSS, la probabilité qu'elle soit exploitée dans les 30 prochains jours est de 0.42 %, ce qui la place devant 34.6 % des CVE référencées.
Que faire ?
- Vérifiez si Linux (versions <edf82bc8150570167a33a7d54627d66614cbf841, <=6.12.*, a19a58ab12695abb91a4e8ecf62cc0bc5badefbe, <=6.16.*, <b47c4b7cb78b6382c6d1af0bfc0d8218e4f445cc, <6.5, 6.17, 0, <=6.15.*, 6.6.103, <c18646248fed07683d4cee8a8af933fc4fe83c0d, 6.5, <=6.6.*, c2ff29e99a764769eb2ce3a1a5585013633ee9a6, <5661fdd218c2799001b88c17acd19f4395e4488e, 6.12.43, <42ebc16d9d2563f1a1ce0f05b643ee68d54fabf8, 6.16.2, <=*, <673cf582fd788af12cdacfb62a6a593083542481, 6.15.11) est présent sur votre parc.
- Appliquez le correctif ou la version corrigée indiqués dans les références de l'éditeur.
- À défaut de correctif, limitez l'exposition (filtrage réseau, désactivation du composant).
- Automatisez la détection : les agents CVE Heimdall (gratuits, open source) corrèlent l'inventaire de vos serveurs avec la base CVE.
Références
- https://git.kernel.org/stable/c/b47c4b7cb78b6382c6d1af0bfc0d8218e4f445cc
- https://git.kernel.org/stable/c/5661fdd218c2799001b88c17acd19f4395e4488e
- https://git.kernel.org/stable/c/673cf582fd788af12cdacfb62a6a593083542481
- https://git.kernel.org/stable/c/42ebc16d9d2563f1a1ce0f05b643ee68d54fabf8
- https://git.kernel.org/stable/c/edf82bc8150570167a33a7d54627d66614cbf841
- https://git.kernel.org/stable/c/c18646248fed07683d4cee8a8af933fc4fe83c0d
