src/sys/opencrypto, branch stable/14

ktls CBC decrypt: Only increment iovec index when an entry is used

2026-06-30T17:00:22+00:00

If an mbuf in the chain was skipped because it only contained bytes
from the header, the iovec index ('i') was incremented even though the
entry was not populated.  Only increment 'i' when an iovec entry is
consumed.

Add a new type of KTLS receive test which writes a single TLS record
via two separate write(2) calls over a TCP_NODELAY socket to trigger
a split in the mbuf chain in the kernel.  Test various split locations
including after the "plain" TLS header (5 bytes), after the full TLS
header, in the middle of the data payload, just before the start of
the trailer, and in the middle of the trailer.  These tests are also
run against all supported ciphers, not just CBC.  The 'header' test
for CBC ciphersuites was able to trigger the bug.

Approved by:	so
Security:	FreeBSD-SA-26:46.ktls
Security:	CVE-2026-49423
Sponsored by:	Chelsio Communications

ktls: Remove the socket parameter to ktls_ocf_try()

2024-07-23T13:01:30+00:00

The socket is unused, and not passing it means that there's less to
think about when considering how KTLS is synchronized with the rest of
the socket code.  No functional change intended.

Reviewed by:	gallatin
MFC after:	2 weeks
Sponsored by:	Klara, Inc.
Sponsored by:	Stormshield
Differential Revision:	https://reviews.freebsd.org/D45675

(cherry picked from commit 5dfca6c375d530908eedb7f103681c2493cf0ca3)

sys: Remove $FreeBSD$: one-line sh pattern

2023-08-16T17:54:58+00:00

Remove /^\s*#[#!]?\s*\$FreeBSD\$.*$\n/

sys: Remove $FreeBSD$: one-line .c pattern

2023-08-16T17:54:36+00:00

Remove /^[\s*]*__FBSDID\("\$FreeBSD\$"\);?\s*\n/

sys: Remove $FreeBSD$: one-line .c comment pattern

2023-08-16T17:54:24+00:00

Remove /^/[*/]\s*\$FreeBSD\$.*\n/

sys: Remove $FreeBSD$: one-line .h pattern

2023-08-16T17:54:18+00:00

Remove /^\s*\*+\s*\$FreeBSD\$.*$\n/

sys: Remove $FreeBSD$: two-line .h pattern

2023-08-16T17:54:11+00:00

Remove /^\s*\*\n \*\s+\$FreeBSD\$$\n/

opencrypto: Respect alignment constraints in xor_and_encrypt()

2023-07-27T19:44:52+00:00

Copy operands to an aligned buffer before performing operations which
require alignment.  Otherwise it's possible for this code to trigger an
alignment fault on armv7.

Reviewed by:	jhb
MFC after:	2 weeks
Sponsored by:	Klara, Inc.
Sponsored by:	Stormshield
Differential Revision:	https://reviews.freebsd.org/D41211

crypto: Advance the correct pointer in crypto_cursor_copydata()

2023-06-12T16:52:24+00:00

PR:		271766
Reported by:	Michael Laß 
MFC after:	1 week
Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D40468

opencrypto: Handle end-of-cursor conditions in crypto_cursor_segment()

2023-06-12T16:52:24+00:00

Some consumers, e.g., swcr_encdec(), may call crypto_cursor_segment()
after having advanced the cursor to the end of the buffer.  In this case
I believe the right behaviour is to return NULL and a length of 0.

When this occurs with a CRYPTO_BUF_VMPAGE buffer, the cc_vmpage pointer
will point past the end of the page pointer array, so
crypto_cursor_segment() ends up dereferencing a random pointer before
the function returns a length of 0.  The uio-backed cursor has
a similar problem.

Address this by keeping track of the residual buffer length and
returning immediately once the length is zero.

PR:		271766
Reported by:	Andrew "RhodiumToad" Gierth 
Reviewed by:	jhb
MFC after:	1 week
Sponsored by:	The FreeBSD Foundation
Differential Revision:	https://reviews.freebsd.org/D40428