head	1.11;
access;
symbols
	netbsd-11-0-RELEASE:1.7
	netbsd-11-0-RC7:1.7
	netbsd-11-0-RC6:1.7
	netbsd-11-0-RC5:1.7
	netbsd-11-0-RC4:1.7
	netbsd-11-0-RC3:1.7
	netbsd-11-0-RC2:1.7
	netbsd-11-0-RC1:1.7
	perseant-exfatfs-base-20250801:1.7
	netbsd-11:1.7.0.4
	netbsd-11-base:1.7
	netbsd-10-1-RELEASE:1.6
	perseant-exfatfs-base-20240630:1.7
	perseant-exfatfs:1.7.0.2
	perseant-exfatfs-base:1.7
	netbsd-10-0-RELEASE:1.6
	netbsd-10-0-RC6:1.6
	netbsd-10-0-RC5:1.6
	netbsd-10-0-RC4:1.6
	netbsd-10-0-RC3:1.6
	netbsd-10-0-RC2:1.6
	netbsd-10-0-RC1:1.6
	netbsd-10:1.6.0.6
	netbsd-10-base:1.6
	cjep_sun2x-base1:1.6
	cjep_sun2x:1.6.0.4
	cjep_sun2x-base:1.6
	cjep_staticlib_x-base1:1.6
	cjep_staticlib_x:1.6.0.2
	cjep_staticlib_x-base:1.6;
locks; strict;
comment	@# @;


1.11
date	2025.11.24.16.28.51;	author nia;	state Exp;
branches;
next	1.10;
commitid	WOieiTJXnjM8ePjG;

1.10
date	2025.11.23.22.48.27;	author riastradh;	state Exp;
branches;
next	1.9;
commitid	DyNWn8bpWikomJjG;

1.9
date	2025.11.23.22.44.14;	author riastradh;	state Exp;
branches;
next	1.8;
commitid	sK7Zpd26tf3XkJjG;

1.8
date	2025.11.22.22.32.39;	author riastradh;	state Exp;
branches;
next	1.7;
commitid	oq1ie9hpKwiYiBjG;

1.7
date	2023.08.08.06.27.32;	author mrg;	state Exp;
branches;
next	1.6;
commitid	fXtLeeKhU4aARWzE;

1.6
date	2020.09.08.17.35.27;	author jakllsch;	state Exp;
branches;
next	1.5;
commitid	OV3IOQDxsoBWJgnC;

1.5
date	2020.08.17.16.26.02;	author riastradh;	state Exp;
branches;
next	1.4;
commitid	mhTGtcH1evXm5rkC;

1.4
date	2020.08.16.15.52.14;	author martin;	state Exp;
branches;
next	1.3;
commitid	8tzJlBDJ67rEVikC;

1.3
date	2020.07.25.22.33.04;	author riastradh;	state Exp;
branches;
next	1.2;
commitid	EhLNwxAgajQhRvhC;

1.2
date	2020.07.01.09.58.29;	author riastradh;	state Exp;
branches;
next	1.1;
commitid	d9HFrUDJEwOfsmeC;

1.1
date	2020.06.30.20.32.11;	author riastradh;	state Exp;
branches;
next	;
commitid	FIYM0QmP24NsZheC;


desc
@@


1.11
log
@Needs the same compiler bug workaround as the kernel.
@
text
@#	$NetBSD: Makefile,v 1.10 2025/11/23 22:48:27 riastradh Exp $

.include <bsd.own.mk>

TESTSDIR=	${TESTSBASE}/sys/crypto/aes

TESTS_C=	t_aes

AFLAGS+=	-D_LOCORE

.PATH:	${NETBSDSRCDIR}/sys/crypto/aes
CPPFLAGS+=	-I${NETBSDSRCDIR}/sys

SRCS.t_aes+=	t_aes.c

SRCS.t_aes+=	aes_bear.c
SRCS.t_aes+=	aes_ct.c
SRCS.t_aes+=	aes_ct_dec.c
SRCS.t_aes+=	aes_ct_enc.c

SRCS.t_aes+=	aes_bear64.c
SRCS.t_aes+=	aes_ct64.c
SRCS.t_aes+=	aes_ct64_dec.c
SRCS.t_aes+=	aes_ct64_enc.c

SRCS.t_aes+=	aes_keysched.c
SRCS.t_aes+=	aes_selftest.c

.if !empty(MACHINE_ARCH:Mearmv7*) || !empty(MACHINE_ARCH:Maarch64*)

.PATH:	${NETBSDSRCDIR}/sys/crypto/aes/arch/arm
CPPFLAGS+=	-I${NETBSDSRCDIR}/sys/crypto/aes/arch/arm

.if !empty(MACHINE_ARCH:Maarch64*)	# XXX no AESE/AESD in 32-bit mode yet
SRCS.t_aes+=	aes_armv8.c
SRCS.t_aes+=	aes_armv8_64.S
.endif

CLANG_NO_NONPORTABLE_VECTOR_INITIALIZATION= ${${ACTIVE_CC} == "clang":? -Wno-nonportable-vector-initialization :}

SRCS.t_aes+=	aes_neon.c
COPTS.aes_neon.c+=${CLANG_NO_NONPORTABLE_VECTOR_INITIALIZATION}
SRCS.t_aes+=	aes_neon_impl.c
SRCS.t_aes+=	aes_neon_subr.c
COPTS.aes_neon_subr.c+=${CLANG_NO_NONPORTABLE_VECTOR_INITIALIZATION}
.if !empty(MACHINE_ARCH:Mearmv7*)
SRCS.t_aes+=	aes_neon_32.S
.endif

.if !empty(MACHINE_ARCH:Mearmv7*) && empty(MACHINE_ARCH:Mearmv7hf*)
COPTS.aes_neon.c+=	-mfloat-abi=softfp -mfpu=neon
COPTS.aes_neon_subr.c+=	-mfloat-abi=softfp -mfpu=neon
AOPTS.aes_neon_32.S+=	-D__SOFTFP__
.endif

.endif				# earmv7/aarch64

.if ${MACHINE_ARCH} == "i386" || ${MACHINE_ARCH} == "x86_64"

.PATH:	${NETBSDSRCDIR}/sys/crypto/aes/arch/x86
CPPFLAGS+=	-I${NETBSDSRCDIR}/sys/crypto/aes/arch/x86

.if ${MACHINE_ARCH} == "x86_64"		# XXX no AES-NI in 32-bit mode yet
SRCS.t_aes+=	aes_ni.c
SRCS.t_aes+=	aes_ni_64.S
.endif

SRCS.t_aes+=	aes_sse2_4x32.c
SRCS.t_aes+=	aes_sse2_4x32_dec.c
SRCS.t_aes+=	aes_sse2_4x32_enc.c
SRCS.t_aes+=	aes_sse2_4x32_impl.c
SRCS.t_aes+=	aes_sse2_4x32_subr.c
COPTS.aes_sse2_4x32.c+=		-msse -msse2
COPTS.aes_sse2_4x32_dec.c+=	-msse -msse2
COPTS.aes_sse2_4x32_enc.c+=	-msse -msse2
COPTS.aes_sse2_4x32_subr.c+=	-msse -msse2

SRCS.t_aes+=	aes_ssse3.c
SRCS.t_aes+=	aes_ssse3_impl.c
SRCS.t_aes+=	aes_ssse3_subr.c
COPTS.aes_ssse3.c+=		-msse -msse2 -msse3 -mssse3
COPTS.aes_ssse3_subr.c+=	-msse -msse2 -msse3 -mssse3

SRCS.t_aes+=	aes_via.c

.endif				# x86

WARNS=		5

# Many of these are probably GCC bugs, see
#   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110878
COPTS.aes_armv8.c+=	${CC_WNO_STRINGOP_OVERREAD} ${CC_WNO_STRINGOP_OVERFLOW}
COPTS.aes_bear.c+=	${CC_WNO_STRINGOP_OVERFLOW} ${CC_WNO_ARRAY_BOUNDS}
COPTS.aes_bear64.c+=	${CC_WNO_STRINGOP_OVERFLOW} ${CC_WNO_ARRAY_BOUNDS} \
			${CC_WNO_STRINGOP_OVERREAD}
COPTS.aes_neon_subr.c+=	${CC_WNO_ARRAY_BOUNDS}

COPTS.aes_ni.c+=	${CC_WNO_STRINGOP_OVERREAD} ${CC_WNO_STRINGOP_OVERFLOW}
COPTS.aes_sse2_4x32_subr.c+=	${CC_WNO_ARRAY_BOUNDS}
COPTS.aes_ssse3_subr.c+=${CC_WNO_ARRAY_BOUNDS}
COPTS.aes_via.c+=	${CC_WNO_ARRAY_BOUNDS}

.include <bsd.test.mk>
@


1.10
log
@aes(9): Rewrite x86 SSE2 implementation.

This computes eight AES_k instances simultaneously, using the
bitsliced 32-bit aes_ct logic which computes two blocks at a time in
uint32_t arithmetic, vectorized four ways.

Previously, the SSE2 code was a very naive adaptation of aes_ct64,
which computes four blocks at a time in uint64_t arithmetic, without
any 2x vectorization -- I did it at the time because:

(a) it was easier to get working,
(b) it only affects really old hardware with neither AES-NI nor SSSE3
    which are both much much faster.

But it was bugging me that this was a kind of dumb use of SSE2.

Substantially reduces stack usage (from ~1200 bytes to ~800 bytes)
and should approximately double throughput for CBC decryption and for
XTS encryption/decryption.

I also tried a 2x64 version but cursory performance measurements
didn't reveal much benefit over 4x32.  (If anyone is interested in
doing more serious performance measurements, on ancient hardware for
which it might matter, I also have the 2x64 code around.)

Prompted by:

PR kern/59774: bearssl 32-bit AES is too slow, want 64-bit optimized
version in kernel
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.9 2025/11/23 22:44:14 riastradh Exp $
d94 2
a95 1
COPTS.aes_bear64.c+=	${CC_WNO_STRINGOP_OVERFLOW} ${CC_WNO_ARRAY_BOUNDS}
@


1.9
log
@aes(9): New 64-bit bitsliced implementation.

Derived from BearSSL's aes_ct64 code.  Compared to the aes_ct code,
on machines with native 64-bit integer arithmetic, aes_ct64 should
have approximately:

- the same throughput for:
  . CBC encryption,
  . CCM encryption/decryption, and
  . CBC-MAC;

- double the throughput for:
  . CBC decryption,
  . XTS encryption/decryption.

(aes_ct computes AES on two blocks at a time; aes_ct64 computes it on
four blocks at a time, with roughly the same number of instructions.
CBC encryption and CBC-MAC are inherently sequential; CCM, being a
combination of CTR and CBC-MAC, can only really be parallelized two
ways, so having four ways available doesn't help; and CBC decryption
and XTS admit parallelism limited only by the size of the inputs.)

Enable with `options AES_BEAR64'.  Should be a reasonable default on
all platforms with 64-bit integer registers.  Caveat: uses about 1200
bytes of stack space.  (Could approximately halve that, like the
BearSSL aes_ct code, at some speed cost which I haven't measured --
by moving the br_aes_ct64_skey_expand logic into add_round_key in
aes_ct64_{enc,dec}.c.)

PR kern/59774: bearssl 32-bit AES is too slow, want 64-bit optimized
version in kernel
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.8 2025/11/22 22:32:39 riastradh Exp $
d68 9
a76 9
SRCS.t_aes+=	aes_sse2.c
SRCS.t_aes+=	aes_sse2_dec.c
SRCS.t_aes+=	aes_sse2_enc.c
SRCS.t_aes+=	aes_sse2_impl.c
SRCS.t_aes+=	aes_sse2_subr.c
COPTS.aes_sse2.c+=	-msse -msse2
COPTS.aes_sse2_dec.c+=	-msse -msse2
COPTS.aes_sse2_enc.c+=	-msse -msse2
COPTS.aes_sse2_subr.c+=	-msse -msse2
d98 1
a98 1
COPTS.aes_sse2_subr.c+=	${CC_WNO_ARRAY_BOUNDS}
@


1.8
log
@aes(9): New aes_keysched_enc/dec.

These implement the standard key schedule.  They are named
independently of any particular AES implementation, so that:

(a) we can swap between the BearSSL aes_ct and aes_ct64 code without
    changing all the callers who don't care which one they get, and

(b) we could push it into the aes_impl abstraction if we wanted.

This eliminates all br_aes_* references outside aes_bear.c, aes_ct*.c,
and the new aes_keysched.c wrappers.

Preparation for:

PR kern/59774: bearssl 32-bit AES is too slow, want 64-bit optimized
version in kernel
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.7 2023/08/08 06:27:32 mrg Exp $
d20 6
d94 1
@


1.7
log
@introduce new GCC 12 warning disables and use them in a few places

this introduces 4 new warning disable flags:

   CC_WNO_MISSING_TEMPLATE_KEYWORD
   CC_WNO_REGISTER
   CC_WNO_STRINGOP_OVERREAD
   CC_WNO_ARRAY_BOUNDS

and documents them in README.warnings.  of these, the string op
and array bounds are both problematic (real bugs) and also spurious
(not real bugs), and the other 2 are mostly temporary for older
3rd party code.

add some new uses of CC_WNO_STRINGOP_OVERFLOW.

fix m68k build for gallium and GCC 12.
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.6 2020/09/08 17:35:27 jakllsch Exp $
d20 1
@


1.6
log
@Acknowledge clang warning for NEON cipher code on aarch64eb

We've already made the nonportable vector initializations portable; the
code works on aarch64eb.
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.5 2020/08/17 16:26:02 riastradh Exp $
d83 11
@


1.5
log
@Make the AES and ChaCha NEON tests work in softfloat userland.

(`Softfloat' here refers to the ABI, which of course may be running
on a CPU with NEON.)
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.4 2020/08/16 15:52:14 martin Exp $
d32 2
d35 1
d38 1
@


1.4
log
@Restrict the NEON code to v7hf - the softfloat toolchain does not like
it (nor is it likely to work if there is no FPU present).
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.3 2020/07/25 22:33:04 riastradh Exp $
d22 1
a22 1
.if !empty(MACHINE_ARCH:Mearmv7hf*) || !empty(MACHINE_ARCH:Maarch64*)
d35 1
a35 1
.if !empty(MACHINE_ARCH:Mearmv7hf*)
d39 6
@


1.3
log
@Implement AES-CCM with ARMv8.5-AES.
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.2 2020/07/01 09:58:29 riastradh Exp $
d22 1
a22 1
.if !empty(MACHINE_ARCH:Mearmv7*) || !empty(MACHINE_ARCH:Maarch64*)
d35 1
a35 1
.if !empty(MACHINE_ARCH:Mearmv7*)
@


1.2
log
@Pass the requisite -msse options for i386.
@
text
@d1 1
a1 1
#	$NetBSD: Makefile,v 1.1 2020/06/30 20:32:11 riastradh Exp $
d9 2
@


1.1
log
@New test sys/crypto/aes/t_aes.

Runs aes_selftest on all kernel AES implementations supported on the
current hardware, not just the preferred one.
@
text
@d1 1
a1 1
#	$NetBSD$
d54 4
d62 2
a63 1
COPTS.aes_ssse3.c+=	-msse3 -mssse3
@