Ciphers, block modes, name it, are grouped together and sorted.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

On Thu, Mar 27, 2008 at 03:40:36PM +0100, Bodo Eggert wrote:
> Kamalesh Babulal <kamalesh@linux.vnet.ibm.com> wrote:
> 
> > This patch cleanups the crypto code, replaces the init() and fini()
> > with the <algorithm name>_init/_fini
> 
> This part ist OK.
> 
> > or init/fini_<algorithm name> (if the 
> > <algorithm name>_init/_fini exist)
> 
> Having init_foo and foo_init won't be a good thing, will it? I'd start
> confusing them.
> 
> What about foo_modinit instead?

Thanks for the suggestion, the init() is replaced with

	<algorithm name>_mod_init ()

and fini () is replaced with <algorithm name>_mod_fini.
 
Signed-off-by: Kamalesh Babulal <kamalesh@linux.vnet.ibm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The key expansion routine could be get little more generic, become
a kernel doc entry and then get exported.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Tested-by: Stefan Hellermann <stefan@the2masters.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Implement CTS wrapper for CBC mode required for support of AES
encryption support for Kerberos (rfc3962).

Signed-off-by: Kevin Coffman <kwc@citi.umich.edu>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

replace all:
big_endian_variable = cpu_to_beX(beX_to_cpu(big_endian_variable) +
					expression_in_cpu_byteorder);
with:
	beX_add_cpu(&big_endian_variable, expression_in_cpu_byteorder);

Signed-off-by: Marcin Slusarz <marcin.slusarz@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Roel Kluin <12o3l@tiscali.nl>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The third test vector of ECB-XTEA-ENC fails for me all other
are fine. I could not find a RFC or something else where they
are defined. The test vector has not been modified since git
started recording histrory. The implementation is very close
(not to say equal) to what is available as Public Domain (they
recommend 64 rounds and the in kernel uses 32). Therefore I
belive that there is typo somewhere and tcrypt reported always
*fail* instead of *okey*.
This patch replaces input + result of the third test vector with
result + input from the third decryption vector. The key is the
same, the other three test vectors are also the reverse.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Currently the tcrypt module is about 2 MiB on x86-32. The
main reason for the huge size is the data segment which contains
all the test vectors for each algorithm. The test vectors are
staticly allocated in an array and the size of the array has been
drastically increased by the merge of the Salsa20 test vectors.

With a hint from Benedigt Spranger I found a way how I could
convert those fixed-length arrays to strings which are flexible
in size. VIM and regex were also very helpfull :)
So, I am talking about a shrinking of ~97% on x86-32:

   text    data     bss     dec     hex filename
  18309 2039708      20 2058037  1f6735 tcrypt-b4.ko
  45628   23516      80   69224   10e68 tcrypt.ko

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The test routines (test_{cipher,hash,aead}) are makeing a copy
of the test template and are processing the encryption process
in place. This patch changes the creation of the copy so it will
work even if the source address of the input data isn't an array
inside of the template but a pointer.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Jan Engelhardt <jengelh@computergmbh.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The speed templates as it look always the same. The key size
is repeated for each block size and we test always the same
block size. The addition of one inner loop makes it possible
to get rid of the struct and it is possible to use a tiny
u8 array :)

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Some crypto ciphers which are impleneted support similar key sizes
(16,24 & 32 byte). They can be grouped together and use a common
templatte instead of their own which contains the same data.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Rename sha512 to sha512_generic and add a MODULE_ALIAS for sha512
so all sha512 implementations can be loaded automatically.

Keep the broken tabs so git recognizes this as a rename.

Signed-off-by: Jan Glauber <jang@linux.vnet.ibm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Alexey Dobriyan <adobriyan@sw.ru>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

'ack' is currently a simple integer that flags whether or not a client is done
touching fields in the given descriptor.  It is effectively just a single bit
of information.  Converting this to a flags parameter allows the other bits to
be put to use to control completion actions, like dma-unmap, and capture
results, like xor-zero-sum == 0.

Changes are one of:
1/ convert all open-coded ->ack manipulations to use async_tx_ack
   and async_tx_test_ack.
2/ set the ack bit at prep time where possible
3/ make drivers store the flags at prep time
4/ add flags to the device_prep_dma_interrupt prototype

Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

Shrink struct dma_async_tx_descriptor and introduce
async_tx_channel_switch to properly inject a channel switch interrupt in
the descriptor stream.  This simplifies the locking model as drivers no
longer need to handle dma_async_tx_descriptor.lock.

Acked-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

The kernel crashes when ipsec passes a udp packet of about 14XX bytes
of data to aes-xcbc-mac.

It seems the first xxxx bytes of the data are in first sg entry,
and remaining xx bytes are in next sg entry. But we don't 
check next sg entry to see if we need to go look the page up.

I noticed in hmac.c, we do a scatterwalk_sg_next(), to do this check
and possible lookup, thus xcbc.c needs to use this routine too.

A 15-hour run of an ipsec stress test sending streams of tcp and
udp packets of various sizes,  using this patch and 
aes-xcbc-mac completed successfully, so hopefully this fixes the
problem.
 
Signed-off-by: Joy Latten <latten@austin.ibm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

If the channel cannot perform the operation in one call to
->device_prep_dma_zero_sum, then fallback to the xor+page_is_zero path.
This only affects users with arrays larger than 16 devices on iop13xx or
32 devices on iop3xx.

Cc: <stable@kernel.org>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

Signed-off-by: Dan Williams <dan.j.williams@intel.com>

The previous patch to move chainiv and eseqiv into blkcipher created
a section mismatch for the chainiv exit function which was also called
from __init.  This patch removes the __exit marking on it.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

When using aes-xcbc-mac for authentication in IPsec, 
the kernel crashes. It seems this algorithm doesn't 
account for the space IPsec may make in scatterlist for authtag.
Thus when crypto_xcbc_digest_update2() gets called,
nbytes may be less than sg[i].length. 
Since nbytes is an unsigned number, it wraps
at the end of the loop allowing us to go back 
into loop and causing crash in memcpy.

I used update function in digest.c to model this fix.
Please let me know if it looks ok.

Signed-off-by: Joy Latten <latten@austin.ibm.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The XTS blockmode uses a copy of the IV which is saved on the stack
and may or may not be properly aligned. If it is not, it will break
hardware cipher like the geode or padlock.
This patch encrypts the IV in place so we don't have to worry about
alignment.

Signed-off-by: Sebastian Siewior <sebastian@breakpoint.cc>
Tested-by: Stefan Hellermann <stefan@the2masters.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Every file should include the headers containing the externs for its 
global code (in this case for struct crypto_{init,exit}_digest_ops()).

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The authenc algorithm requires BLKCIPHER to be present.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

For compatibility with dm-crypt initramfs setups it is useful to merge
chainiv/seqiv into the crypto_blkcipher module.  Since they're required
by most algorithms anyway this is an acceptable trade-off.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

This patch fixes the following build error caused by commit 
3631c650:

<--  snip  -->

...
  LD      .tmp_vmlinux1
crypto/built-in.o: In function `skcipher_null_crypt':
crypto_null.c:(.text+0x3d14): undefined reference to `blkcipher_walk_virt'
crypto_null.c:(.text+0x3d14): relocation truncated to fit: R_MIPS_26 against `blkcipher_walk_virt'
crypto/built-in.o: In function `$L32':
crypto_null.c:(.text+0x3d54): undefined reference to `blkcipher_walk_done'
crypto_null.c:(.text+0x3d54): relocation truncated to fit: R_MIPS_26 against `blkcipher_walk_done'
crypto/built-in.o:(.data+0x2e8): undefined reference to `crypto_blkcipher_type'
make[1]: *** [.tmp_vmlinux1] Error 1

<--  snip  -->

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Building latest git fails with the following error:
	ERROR: "crypto_alloc_ablkcipher" [crypto/tcrypt.ko] undefined!
This appears to happen because CONFIG_CRYPTO_TEST is set while
CONFIG_CRYPTO_BLKCIPHER is not.
The following patch fixes the problem for me.

Signed-off-by: Frederik Deweerdt <frederik.deweerdt@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Convert instances of ERR_PTR(PTR_ERR(p)) to ERR_CAST(p) using:

perl -spi -e 's/ERR_PTR[(]PTR_ERR[(](.*)[)][)]/ERR_CAST(\1)/' `grep -rl 'ERR_PTR[(]*PTR_ERR' fs crypto net security`

Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

The source and destination addresses are included to allow channel
selection based on address alignment.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Haavard Skinnemoen <hskinnemoen@atmel.com>

Pass a full set of flags to drivers' per-operation 'prep' routines.
Currently the only flag passed is DMA_PREP_INTERRUPT.  The expectation is
that arch-specific async_tx_find_channel() implementations can exploit this
capability to find the best channel for an operation.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Shannon Nelson <shannon.nelson@intel.com>
Reviewed-by: Haavard Skinnemoen <hskinnemoen@atmel.com>

The tx_set_src and tx_set_dest methods were originally implemented to allow
an array of addresses to be passed down from async_xor to the dmaengine
driver while minimizing stack overhead.  Removing these methods allows
drivers to have all transaction parameters available at 'prep' time, saves
two function pointers in struct dma_async_tx_descriptor, and reduces the
number of indirect branches..

A consequence of moving this data to the 'prep' routine is that
multi-source routines like async_xor need temporary storage to convert an
array of linear addresses into an array of dma addresses.  In order to keep
the same stack footprint of the previous implementation the input array is
reused as storage for the dma addresses.  This requires that
sizeof(dma_addr_t) be less than or equal to sizeof(void *).  As a
consequence CONFIG_DMADEVICES now depends on !CONFIG_HIGHMEM64G.  It also
requires that drivers be able to make descriptor resources available when
the 'prep' routine is polled.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Shannon Nelson <shannon.nelson@intel.com>

Remove the unused ASYNC_TX_ASSUME_COHERENT flag.  Async_tx is
meant to hide the difference between asynchronous hardware and synchronous
software operations, this flag requires clients to understand cache
coherency consequences of the async path.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Reviewed-by: Haavard Skinnemoen <hskinnemoen@atmel.com>

single list_head variable initialized with LIST_HEAD_INIT could almost
always can be replaced with LIST_HEAD declaration, this shrinks the code
and looks better.

Signed-off-by: Denis Cheng <crquan@gmail.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>

do_async_xor must be compiled away on !HAS_DMA archs.

Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>

Bloat-o-meter shows rather high readings for cast6...

crypto/cast6.c:
  cast6_setkey  | -1310
  cast6_encrypt | -4567
  cast6_decrypt | -4561
 3 functions changed, 10438 bytes removed, diff: -10438

crypto/cast6.c:
  W    | +659
  Q    | +308
  QBAR | +316
 3 functions changed, 1283 bytes added, diff: +1283

crypto/cast6.o:
 6 functions changed, 1283 bytes added, 10438 bytes removed, diff: -9155

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Currently the gcm(aes) tests have to be taken together with all other
algorithms.  This patch makes it available by itself at number 106.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

When setting the digest size xcbc tests to see if the underlying algorithm
is a hash.  This is silly because we don't allow it to be a hash and we've
specifically requested for a cipher.

This patch removes the bogus test.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

When the underlying algorithm has a block size other than 16 we abort
without freeing it.  In fact, we try to return the algorithm itself
as an error!

This patch plugs the leak and makes it return -EINVAL instead.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

The axbuf buffer is used by test_aead and therefore should be zeroed
there instead of in test_hash.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Tan Swee Heng <thesweeheng@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>