MFC: r285215

remove _NORMAL flag which isn't suppose to be used w/ _alloc_ctx...

MFC: r285289

address an issue where consumers, like IPsec, can reuse the same
session in multiple threads w/o locking.. There was a single fpu
context shared per session, if multiple threads were using the session,
and both migrated away, they could corrupt each other's fpu context...

MFC: r285297

upon further examination, it turns out that _unregister_all already
provides the guarantee that no threads will be in the _newsession code..

MFC: r298332

aesni(4): Initialize error before use [1]

Reported by: Coverity [1]
CID: 1331554 [1]

Fix r314332 (MFC of r300903): do not use C99 Static array indices

The proper conditional support for that feature has not been MFC-ed yet
and GCC is quirky about it. See r314101.

This is a direct commit.

Reported by: gjb

MFC r300966: Retune SHA2 code for improved performance on CPUs with more ILP...

MFC r300903: Implement SHA-512 truncated (224 and 256 bits)

MFC r292782: Replace sys/crypto/sha2/sha2.c with lib/libmd/sha512c.c

cperciva's libmd implementation is 5-30% faster
The same was done for SHA256 previously in r263218

Approved by: secteam

MFC r263215,r263218,r285366: replace the kernel sha256 code

r263215 copy these files from lib/libmd in preperation for moving these
files into the kernel...

r263218 replace the kernel's version w/ cperciva's implementation...

r285366 Complete the move that was started w/ r263218..

Note that the last change is out of order with r282726 that I am going
to merge as well.

Many thanks to cperciva for the more efficient code and to jmg for
integrating it into the kernel.

MFC r264400,r265836:

r264400:

NO_MAN= has been deprecated in favor of MAN= for some time, go ahead
and finish the job. ncurses is now the only Makefile in the tree that
uses it since it wasn't a simple mechanical change, and will be
addressed in a future commit.

r265836:

Remove last two NO_MAN= in the tree. In both of these cases, MAN= is
what is needed.

MFC r267815:
Put the aesni_cipher_setup() and aesni_cipher_process() functions into
the file which is compiled with SSE disabled.

MFC r267767:
Add FPU_KERN_KTHR flag to fpu_kern_enter(9).
Apply the flag to padlock(4) and aesni(4).
In aesni_cipher_process(), do not leak FPU context state on error.

MFC r258399,258492:
mark aesni module _SYNC, improves performance ~27%...

Approved by: re (glebius)

MFC r257757:
fix issues w/ AES-NI on unaligned data blocks...

Approved by: re (kib)

Copy head (r256279) to stable/10 as part of the 10.0-RELEASE cycle.

Approved by: re (implicit)
Sponsored by: The FreeBSD Foundation

Use the fact that the AES-NI instructions can be pipelined to improve
performance... Use SSE2 instructions for calculating the XTS tweek
factor... Let the compiler do more work and handle register allocation
by using intrinsics, now only the key schedule is in assembly...

Replace .byte hard coded instructions w/ the proper instructions now
that both clang and gcc support them...

On my machine, pulling the code to userland I saw performance go from
~150MB/sec to 2GB/sec in XTS mode. GELI on GNOP saw a more modest
increase of about 3x due to other system overhead (geom and
opencrypto)...

These changes allow almost full disk io rate w/ geli...

Reviewed by: -current, -security
Thanks to: Mike Hamburg for the XTS tweek algorithm

Fix const propagation issues to make GCC happy.

Submitted by: Michael Butler <imb@protected-networks.net>

SipHash is a cryptographically strong pseudo-random function (a.k.a. keyed
hash function) optimized for speed on short messages returning a 64bit hash/
digest value.

SipHash is simpler and much faster than other secure MACs and competitive
in speed with popular non-cryptographic hash functions. It uses a 128-bit
key without the hidden cost of a key expansion step. SipHash iterates a
simple round function consisting of four additions, four xors, and six
rotations, interleaved with xors of message blocks for a pre-defined number
of compression and finalization rounds. The absence of secret load/store
addresses or secret branch conditions avoid timing attacks. No state is
shared between messages. Hashing is deterministic and doesn't use nonces.
It is not susceptible to length extension attacks.

Target applications include network traffic authentication, message
authentication (MAC) and hash-tables protection against hash-flooding
denial-of-service attacks.

The number of update/finalization rounds is defined during initialization:

SipHash24_Init() for the fast and reasonable strong version.
SipHash48_Init() for the strong version (half as fast).

SipHash usage is similar to other hash functions:

struct SIPHASH_CTX ctx;
char *k = "16bytes long key"
char *s = "string";
uint64_t h = 0;
SipHash24_Init(&ctx);
SipHash_SetKey(&ctx, k);
SipHash_Update(&ctx, s, strlen(s));
SipHash_Final(&h, &ctx); /* or */
h = SipHash_End(&ctx); /* or */
h = SipHash24(&ctx, k, s, strlen(s));

It was designed by Jean-Philippe Aumasson and Daniel J. Bernstein and
is described in the paper "SipHash: a fast short-input PRF", 2012.09.18:
https://131002.net/siphash/siphash.pdf
Permanent ID: b9a943a805fbfc6fde808af9fc0ecdfa

Implemented by: andre (based on the paper)
Reviewed by: cperciva

Sync with KAME.

MFC after: 1 month

Allow assert() to operate correctly when building userland code.

When porting XTS-related code from OpenBSD I forgot to update copyright (only
OpenBSD was credited in one of two commits). Fix it.

Reported by: Theo de Raadt <deraadt@cvs.openbsd.org>
Reviewed by: Damien Miller <djm@mindrot.org>

Revert previous commit...

Pointyhat to: kevlo (myself)

Prefer NULL over 0 for pointers

Add support for the extended FPU states on amd64, both for native
64bit and 32bit ABIs. As a side-effect, it enables AVX on capable
CPUs.

In particular:

- Query the CPU support for XSAVE, list of the supported extensions
and the required size of FPU save area. The hw.use_xsave tunable is
provided for disabling XSAVE, and hw.xsave_mask may be used to
select the enabled extensions.

- Remove the FPU save area from PCB and dynamically allocate the
(run-time sized) user save area on the top of the kernel stack,
right above the PCB. Reorganize the thread0 PCB initialization to
postpone it after BSP is queried for save area size.

- The dumppcb, stoppcbs and susppcbs now do not carry the FPU state as
well. FPU state is only useful for suspend, where it is saved in
dynamically allocated suspfpusave area.

- Use XSAVE and XRSTOR to save/restore FPU state, if supported and
enabled.

- Define new mcontext_t flag _MC_HASFPXSTATE, indicating that
mcontext_t has a valid pointer to out-of-struct extended FPU
state. Signal handlers are supplied with stack-allocated fpu
state. The sigreturn(2) and setcontext(2) syscall honour the flag,
allowing the signal handlers to inspect and manipilate extended
state in the interrupted context.

- The getcontext(2) never returns extended state, since there is no
place in the fixed-sized mcontext_t to place variable-sized save
area. And, since mcontext_t is embedded into ucontext_t, makes it
impossible to fix in a reasonable way. Instead of extending
getcontext(2) syscall, provide a sysarch(2) facility to query
extended FPU state.

- Add ptrace(2) support for getting and setting extended state; while
there, implement missed PT_I386_{GET,SET}XMMREGS for 32bit binaries.

- Change fpu_kern KPI to not expose struct fpu_kern_ctx layout to
consumers, making it opaque. Internally, struct fpu_kern_ctx now
contains a space for the extended state. Convert in-kernel consumers
of fpu_kern KPI both on i386 and amd64.

First version of the support for AVX was submitted by Tim Bird
<tim.bird am sony com> on behalf of Sony. This version was written
from scratch.

Tested by: pho (previous version), Yamagi Burmeister <lists yamagi org>
MFC after: 1 month

Update Copyright.

MFC after: 3 days

Improve AES-NI performance for AES-XTS:
- Operate on uint64_t types when doing XORing, etc. instead of uint8_t.
- Don't bzero() temporary block for every AES block. Do it once for entire
data block.
- AES-NI is available only on little endian architectures. Simplify code
that takes block number from IV.

Benchmarks:

Memory-backed md(4) device, software AES-XTS, 4kB sector:

# dd if=/dev/md0.eli bs=1m
59.61MB/s

Memory-backed md(4) device, old AES-NI AES-XTS, 4kB sector:

# dd if=/dev/md0.eli bs=1m
97.29MB/s

Memory-backed md(4) device, new AES-NI AES-XTS, 4kB sector:

# dd if=/dev/md0.eli bs=1m
221.26MB/s

127% performance improvement between old and new code.

Harddisk, raw speed:

# dd if=/dev/ada0 bs=1m
137.63MB/s

Harddisk, software AES-XTS, 4kB sector:

# dd if=/dev/ada0.eli bs=1m
47.83MB/s (34% of raw disk speed)

Harddisk, old AES-NI AES-XTS, 4kB sector:

# dd if=/dev/ada0.eli bs=1m
68.33MB/s (49% of raw disk speed)

Harddisk, new AES-NI AES-XTS, 4kB sector:

# dd if=/dev/ada0.eli bs=1m
108.35MB/s (78% of raw disk speed)

58% performance improvement between old and new code.

As a side-note, GELI with AES-NI using AES-CBC can achive native disk speed.

MFC after: 3 days

Fix a bug in the result of manual assembly.

Reported by: Stefan Grundmann <sg2342 googlemail com>
PR: kern/155118
MFC after: 3 days

Make private functions static.

PR: kern/43611
Submitted by: Matt Emmerton <matt at gsicomp.on.ca>
Reviewed by: kib
MFC after: 3 days

Remove DEBUG sections.

MFC after: 3 days

MFaesni r215427:
Only save FPU context when not executing in the context of the crypto
thread.

Tested by: Mike Tancsa
MFC after: 1 week

Only save FPU context when not executing in the context of the crypto
thread.

Tested by: Mike Tancsa

Change two missed instances of 'retq' in aeskeys_i386.S to 'retl', which
makes it possible to assemble this file with gas from newer binutils.

Reviewed by: kib

Fix two copy&paste bugs.

MFC after: 2 weeks

Add support for CRYPTO_AES_XTS.

MFC after: 1 week

Add support for CRD_F_KEY_EXPLICIT flag.

MFC after: 1 week

Simplify code a bit.

MFC after: 1 week

Crypto(4) driver for AESNI.

The aeskeys_{amd64,i386}.S content was mostly obtained from OpenBSD,
no objections to the license from core.

Hardware provided by: Sentex Communications
Tested by: fabient, pho (previous versions)
MFC after: 1 month

Use the fpu_kern_enter() interface to properly separate usermode FPU
context from in-kernel execution of padlock instructions and to handle
spurious FPUDNA exceptions that sometime are raised when doing padlock
calculations.

Globally mark crypto(9) kthread as using FPU.

Reviewed by: pjd
Hardware provided by: Sentex Communications
Tested by: pho
PR: amd64/135014
MFC after: 1 month

Changed to M_NOWAIT when reallocing psc_buf in padlock_sha_update(),
as we already hold the non sleepable crypto_driver_mutex.

Approved by: gnn(mentor)
Obtained from: NETASQ
MFC after: 2 weeks

identify routine takes driver_t *, not device_t *.

Connect padlock(4) to amd64 build for VIA Nano processors.

Avoid 256 integer divisions per rc4_init() call. Replace it with using
separate index variable.

It gives more then double rc4_init() performance increase on tested i386 P4.
It also gives about 15% speedup to PPTP VPN with stateless MPPE encryption
(by ng_mppc) which calls rc4_init() for every packet.

Fix a potential NULL-pointer dereference in padlock(4).

Spotted by: Coverity (via pjd)
MFC after: 1 week

Simplify session selection/allocation.

- Fix freeing session on newsession failure.
- Update copyright years.

Implify sessions freeing loop.

We don't have to drop a lock around malloc(M_NOWAIT).

When freeing session, restore its ID after zeroing memory.

Bug tracked down by: Patrick Lamaiziere <patfbsd@davenulle.org>

Sessions in-use are at the end of the queue, so use TAILQ_FOREACH_REVERSE()
when looking for them.

Idea from: Patrick Lamaiziere <patfbsd@davenulle.org>

Convert lock that protects sessions list from a mutex to a rwlock.
Now we can use read lock in fast path (padlock_process()).

Make test00 compilable again.

Commit the change from FAST_IPSEC to IPSEC. The FAST_IPSEC
option is now deprecated, as well as the KAME IPsec code.
What was FAST_IPSEC is now IPSEC.

Approved by: re
Sponsored by: Secure Computing

Integrate the Camellia Block Cipher. For more information see RFC 4132
and its bibliography.

Submitted by: Tomoyuki Okazaki <okazaki at kick dot gr dot jp>
MFC after: 1 month

Overhaul driver/subsystem api's:
o make all crypto drivers have a device_t; pseudo drivers like the s/w
crypto driver synthesize one
o change the api between the crypto subsystem and drivers to use kobj;
cryptodev_if.m defines this api
o use the fact that all crypto drivers now have a device_t to add support
for specifying which of several potential devices to use when doing
crypto operations
o add new ioctls that allow user apps to select a specific crypto device
to use (previous ioctls maintained for compatibility)
o overhaul crypto subsystem code to eliminate lots of cruft and hide
implementation details from drivers
o bring in numerous fixes from Michale Richardson/hifn; mostly for
795x parts
o add an optional mechanism for mmap'ing the hifn 795x public key h/w
to user space for use by openssl (not enabled by default)
o update crypto test tools to use new ioctl's and add cmd line options
to specify a device to use for tests

These changes will also enable much future work on improving the core
crypto subsystem; including proper load balancing and interposing code
between the core and drivers to dispatch small operations to the s/w
driver as appropriate.

These changes were instigated by the work of Michael Richardson.

Reviewed by: pjd
Approved by: re

Initialize T1 to silent gcc warning.

Approved by: cognet

Less magic.

MFC after: 3 days

GCC 3.4.6 gets confused on this file and produces bogus warning.
Shut it up.

Remove trailing spaces.

Use existing roundup2() macro.

Suggested by: njl

Remove redundant check committed by accident.

Avoid memory allocations when the given address is already 16 bytes aligned.
Such an address can be used directly in padlock's AES.
This improves speed of geli(8) significantly:

# sysctl kern.geom.zero.clear=0
# geli onetime -s 4096 gzero
# dd if=/dev/gzero.eli of=/dev/null bs=1m count=1000

Before: 113MB/s
After: 203MB/s

BTW. If sector size is set to 128kB, I can read at 276MB/s :)

Modify PADLOCK_ALIGN() macro, so when the given address is already 16 bytes
aligned, it will be used directly, not 'address + 16'.

Style fixes.

Implement support for HMAC/SHA1 and HMAC/SHA256 acceleration found in
new VIA CPUs.
For older CPUs HMAC/SHA1 and HMAC/SHA256 (and others) will still be done
in software.

Move symmetric cryptography (currently only AES-CBC 128/192/256) to
padlock_cipher.c file. Move HMAC cryptography to padlock_hash.c file.

Hardware from: Centaur Technologies

Correct few bzero()s.

MFC after: 3 days

Set ses_ictx and ses_octx to NULL after freeing them, so we won't free
them twice.
This is possible for example in situation when session is used in
authentication context, then freed and then used in encryption context
and freed - in encryption context ses_ictx and ses_octx are not touched
at newsession time, but padlock_freesession could still try to free them
when they are not NULL.

Use the already stored VIA RNG probe information
instead of probing again.
Adjust style(9) somewhat in probe.c

Reviewed by: pjd
MFC after: 1 week

Fix gratuitous compiler warning.

Reported by: Rong-en Fan <grafan@gmail.com>

- Pretend to accelerate various HMAC algorithms, so padlock(4) can be used
with fast_ipsec(4) and geli(8) authentication (comming soon).
If consumer requests only for HMAC algorithm (without encryption), return
EINVAL.
- Add support for the CRD_F_KEY_EXPLICIT flag, for both encryption and
authentication.

padlock(4) doesn't support explicitly provided keys yet.
Return an error instead of encrypting/decrypting data with a wrong key.

On padlock initialization, allocate memory with M_WAITOK.

Add VIA/ACE "PadLock" support as a crypto(9) driver.

HW donated by: Mike Tancsa <mike@sentex.net>
Most of the code obtained from: OpenBSD
MFC after: 3 days

Assert proper key size also in userland by defining KASSERT in !_KERNEL case.

Unbreak the world build (in sbin/gbde). This file is used by both
kernel and world, so KASSERT() needs to be wrapped within an #ifdef
_KERNEL / #endif pair.

Reported by: krion, tinderbox

Check key size for rijndael, as invalid key size can lead to kernel panic.
It checked other algorithms against this bug and it seems they aren't
affected.

Reported by: Mike Tancsa <mike@sentex.net>
PR: i386/84860
Reviewed by: phk, cperciva(x2)

gbde(8) is also rejndael user.

Reported by: phk

just use crypto/rijndael, and nuke opencrypto/rindael.[ch].
the two became almost identical since latest KAME merge.

Discussed with: sam

integrate rijndael-alg-fst.h into rijndael.h.

stop including rijndael-api-fst.h from rijndael.h.
this is required to integrate opencrypto into crypto.

sys/crypto/md5.[ch] is used from nowhere. So, just nuke them.

refer opencrypto/cast.h directly.

use cast128 in opencrypto to nuke duplicate code.

Start the dreaded NOFOO -> NO_FOO conversion.

OK'ed by: core

Initiate deorbit burn sequence for 80386 support in FreeBSD: Remove
80386 (I386_CPU) support from the kernel.

For variables that are only checked with defined(), don't provide
any fake value.

Use __FBSDID().

Fix a reentrancy issue in md5_calc().

avoid module name conflict with opencrypto/rijndael.c.

Reported by: tinderbox

cleanup rijndael API.
since there are naming conflicts with opencrypto, #define was
added to rename functions intend to avoid conflicts.

Obtained from: KAME

rijndael-alg-fst.[ch]:
- redo updating.

rijndael-api-fst.[ch]:
- switch to use new low level rijndael api.
- stop using u8, u16 and u32.
- space cleanup.

Tested by: gbde(8) and phk's test program

Add a testcase which validates that the same buffer can be passed to
rijndael_blockDecrypt() as both input and output.

This property is important because inside rijndael we can get away
with allocating just a 16 byte "work" buffer on the stack (which
is very cheap), whereas the calling code would need to allocate the
full sized buffer, and in all likelyhood would have to do so with
an expensive malloc(9).

- revert to old rijndael code. new rijndael code broke gbde.
- since aes-xcbc-mac and aes-ctr require functions in new
rijndael code, aes-xcbc-mac and aes-ctr are disabled for now.

Fix alignment problem on 64 bit arch.
I only tested if it doesn't break anything on i368. Since I
have no 64 bit machine, I cannot test it, actually.

Reported by: jmallett

use BF_ecb_encrypt().

Obtained from: KAME

simplify and update rijndael code.

Obtained from: KAME

use opencrypto for RMD160.

Requested by: sam

drop useless define.

use bswap32() for big endian arch.

Reported by: tinderbox via kris

RIPEMD160 support

Obtained from: KAME

switch cast128 implementation to implementation by Steve Reid;
smaller footprint.

Obtained from: KAME

one more opossite conditiion.

Reported by: "lg" <zevlg@yandex.ru>

condition of padLen check was opposite.

Reported by: "lg" <zevlg@yandex.ru>
Reviewed by: Lev Walkin <vlm@netli.com>

Correctly bzero the entire context, not just the first sizeof(void *) bytes.

Found by: Juergen Buchmueller <pullmoll@stop1984.com>

Use __FBSDID().

Assembly files put thru the C preprocessor need to have C style comments.

Remove some unnecessary casts.

make rc4 crypto support a module so other modules can depend on it

Submitted by: imp
Reviewed by: imp

Make this compilable from userland as well.

Don't panic when we can just return an error code.

Fix some really pedantic GCC warnings.

Fixed AES encryption algorithm bug

PR: kern/38465
Obtained from: Ramana Yarlagadda <ramana.yarlagadda@analog.com>

Remove macros that are defined elsewhere.

just merged cosmetic changes from KAME to ease sync between KAME and FreeBSD.
(based on freebsd4-snap-20020128)

Reviewed by: ume
MFC after: 1 week

Remove __P.

- Speedup 3DES by using assembly code for i386.
- Sync des/blowfish to more recent openssl.

Obtained from: KAME/NetBSD
MFC after: 2 weeks

off by one error in Aaron Gifford's code. KAME PR 393.

PR: kern/34242
Submitted by: Aaron D. Gifford <agifford@infowest.com>
MFC after: 1 week

fixed the cast128 calculation with a short cipher key length.
the memory was overridden when the key length was less than 16 bytes.

Obtained from: KAME
MFC after: 1 week

properly check DES weak key. KAME PR 363.

Obtained from: KAME
MFC after: 1 week

less warning

warning: cast discards qualifiers from pointer target type

Sync with recent KAME.
This work was based on kame-20010528-freebsd43-snap.tgz and some
critical problem after the snap was out were fixed.
There are many many changes since last KAME merge.

TODO:
- The definitions of SADB_* in sys/net/pfkeyv2.h are still different
from RFC2407/IANA assignment because of binary compatibility
issue. It should be fixed under 5-CURRENT.
- ip6po_m member of struct ip6_pktopts is no longer used. But, it
is still there because of binary compatibility issue. It should
be removed under 5-CURRENT.

Reviewed by: itojun
Obtained from: KAME
MFC after: 3 weeks

Kernel crypto need binary key material, not symbolic ascii.

This commit was generated by cvs2svn to compensate for changes in r67957,
which included commits to RCS files with non-trunk default branches.

Fix broken const'ness in declaration of sha1_loop().

sync with kame tree as of july00. tons of bug fixes/improvements.

API changes:
- additional IPv6 ioctls
- IPsec PF_KEY API was changed, it is mandatory to upgrade setkey(8).
(also syntax change)

Remove ~25 unneeded #include <sys/conf.h>
Remove ~60 unneeded #include <sys/malloc.h>

A simplified RC4 implementation for kernel use.

Fix bogus const-ness in declaration of sha1_loop().

Prototype fix for IPsec authentication related functions

Some of IPsec authentication related functions should have
'const' for its 2nd argument, but not now.
But if someone try to use them, and passed const data for
those functions, then much bogus compile warnings will be
generated.
So those funcs prototype should be modified.

Requested by: archie
Approved by: jkh

Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL"
is an application space macro and the applications are supposed to be free
to use it as they please (but cannot). This is consistant with the other
BSD's who made this change quite some time ago. More commits to come.

IPSEC support in the kernel.
pr_input() routines prototype is also changed to support IPSEC and IPV6
chained protocol headers.

Reviewed by: freebsd-arch, cvs-committers
Obtained from: KAME project