Just a slight optimization that removes one array lookup.
The processor number is needed for other things as well so the
get/put_cpu cannot be removed.
Acked-by: Tejun Heo <tj@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Signed-off-by: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Tejun Heo <tj@kernel.org>
If we are unable to offload async_mult() or async_sum_product(), then
unmap the buffers before falling through to the synchronous path.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Testing on x86_64 with NDISKS=255 yields:
do_IRQ: modprobe near stack overflow (cur:ffff88007d19c000,sp:ffff88007d19c128)
...and eventually
general protection fault: 0000 [#1]
Moving the scribble buffers off the stack allows the test to complete
successfully.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (102 commits)
crypto: sha-s390 - Fix warnings in import function
crypto: vmac - New hash algorithm for intel_txt support
crypto: api - Do not displace newly registered algorithms
crypto: ansi_cprng - Fix module initialization
crypto: xcbc - Fix alignment calculation of xcbc_tfm_ctx
crypto: fips - Depend on ansi_cprng
crypto: blkcipher - Do not use eseqiv on stream ciphers
crypto: ctr - Use chainiv on raw counter mode
Revert crypto: fips - Select CPRNG
crypto: rng - Fix typo
crypto: talitos - add support for 36 bit addressing
crypto: talitos - align locks on cache lines
crypto: talitos - simplify hmac data size calculation
crypto: mv_cesa - Add support for Orion5X crypto engine
crypto: cryptd - Add support to access underlaying shash
crypto: gcm - Use GHASH digest algorithm
crypto: ghash - Add GHASH digest algorithm for GCM
crypto: authenc - Convert to ahash
crypto: api - Fix aligned ctx helper
crypto: hmac - Prehash ipad/opad
...
Some engines have transfer size and address alignment restrictions. Add
a per-operation alignment property to struct dma_device that the async
routines and dmatest can use to check alignment capabilities.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Channel switching is problematic for some dmaengine drivers as the
architecture precludes separating the ->prep from ->submit. In these
cases the driver can select ASYNC_TX_DISABLE_CHANNEL_SWITCH to modify
the async_tx allocator to only return channels that support all of the
required asynchronous operations.
For example MD_RAID456=y selects support for asynchronous xor, xor
validate, pq, pq validate, and memcpy. When
ASYNC_TX_DISABLE_CHANNEL_SWITCH=y any channel with all these
capabilities is marked DMA_ASYNC_TX allowing async_tx_find_channel() to
quickly locate compatible channels with the guarantee that dependency
chains will remain on one channel. When
ASYNC_TX_DISABLE_CHANNEL_SWITCH=n async_tx_find_channel() may select
channels that lead to operation chains that need to cross channel
boundaries using the async_tx channel switch capability.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Some engines optimize operation by reading ahead in the descriptor chain
such that descriptor2 may start execution before descriptor1 completes.
If descriptor2 depends on the result from descriptor1 then a fence is
required (on descriptor2) to disable this optimization. The async_tx
api could implicitly identify dependencies via the 'depend_tx'
parameter, but that would constrain cases where the dependency chain
only specifies a completion order rather than a data dependency. So,
provide an ASYNC_TX_FENCE to explicitly identify data dependencies.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
This patch adds VMAC (a fast MAC) support into crypto framework.
Signed-off-by: Shane Wang <shane.wang@intel.com>
Signed-off-by: Joseph Cihula <joseph.cihula@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We have a mechanism where newly registered algorithms of a higher
priority can displace existing instances that use a different
implementation of the same algorithm with a lower priority.
Unfortunately the same mechanism can cause a newly registered
algorithm to displace itself if it depends on an existing version
of the same algorithm.
This patch fixes this by keeping all algorithms that the newly
reigstered algorithm depends on, thus protecting them from being
removed.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Port drivers/md/raid6test/test.c to use the async raid6 recovery
routines. This is meant as a unit test for raid6 acceleration drivers. In
addition to the 16-drive test case this implements tests for the 4-disk and
5-disk special cases (dma devices can not generically handle less than 2
sources), and adds a test for the D+Q case.
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
async_raid6_2data_recov() recovers two data disk failures
async_raid6_datap_recov() recovers a data disk and the P disk
These routines are a port of the synchronous versions found in
drivers/md/raid6recov.c. The primary difference is breaking out the xor
operations into separate calls to async_xor. Two helper routines are
introduced to perform scalar multiplication where needed.
async_sum_product() multiplies two sources by scalar coefficients and
then sums (xor) the result. async_mult() simply multiplies a single
source by a scalar.
This implemention also includes, in contrast to the original
synchronous-only code, special case handling for the 4-disk and 5-disk
array cases. In these situations the default N-disk algorithm will
present 0-source or 1-source operations to dma devices. To cover for
dma devices where the minimum source count is 2 we implement 4-disk and
5-disk handling in the recovery code.
[ Impact: asynchronous raid6 recovery routines for 2data and datap cases ]
Cc: Yuri Tikhonov <yur@emcraft.com>
Cc: Ilya Yanok <yanok@emcraft.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
[ Based on an original patch by Yuri Tikhonov ]
This adds support for doing asynchronous GF multiplication by adding
two additional functions to the async_tx API:
async_gen_syndrome() does simultaneous XOR and Galois field
multiplication of sources.
async_syndrome_val() validates the given source buffers against known P
and Q values.
When a request is made to run async_pq against more than the hardware
maximum number of supported sources we need to reuse the previous
generated P and Q values as sources into the next operation. Care must
be taken to remove Q from P' and P from Q'. For example to perform a 5
source pq op with hardware that only supports 4 sources at a time the
following approach is taken:
p, q = PQ(src0, src1, src2, src3, COEF({01}, {02}, {04}, {08}))
p', q' = PQ(p, q, q, src4, COEF({00}, {01}, {00}, {10}))
p' = p + q + q + src4 = p + src4
q' = {00}*p + {01}*q + {00}*q + {10}*src4 = q + {10}*src4
Note: 4 is the minimum acceptable maxpq otherwise we punt to
synchronous-software path.
The DMA_PREP_CONTINUE flag indicates to the driver to reuse p and q as
sources (in the above manner) and fill the remaining slots up to maxpq
with the new sources/coefficients.
Note1: Some devices have native support for P+Q continuation and can skip
this extra work. Devices with this capability can advertise it with
dma_set_maxpq. It is up to each driver how to handle the
DMA_PREP_CONTINUE flag.
Note2: The api supports disabling the generation of P when generating Q,
this is ignored by the synchronous path but is implemented by some dma
devices to save unnecessary writes. In this case the continuation
algorithm is simplified to only reuse Q as a source.
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Yuri Tikhonov <yur@emcraft.com>
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
We currently walk the parent chain when waiting for a given tx to
complete however this walk may race with the driver cleanup routine.
The routines in async_raid6_recov.c may fall back to the synchronous
path at any point so we need to be prepared to call async_tx_quiesce()
(which calls dma_wait_for_async_tx). To remove the ->parent walk we
guarantee that every time a dependency is attached ->issue_pending() is
invoked, then we can simply poll the initial descriptor until
completion.
This also allows for a lighter weight 'issue pending' implementation as
there is no longer a requirement to iterate through all the channels'
->issue_pending() routines as long as operations have been submitted in
an ordered chain. async_tx_issue_pending() is added for this case.
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
If module_init and module_exit are nops then neither need to be defined.
[ Impact: pure cleanup ]
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Replace the flat zero_sum_result with a collection of flags to contain
the P (xor) zero-sum result, and the soon to be utilized Q (raid6 reed
solomon syndrome) zero-sum result. Use the SUM_CHECK_ namespace instead
of DMA_ since these flags will be used on non-dma-zero-sum enabled
platforms.
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
As struct skcipher_givcrypt_request includes struct crypto_request
at a non-zero offset, testing for NULL after converting the pointer
returned by crypto_dequeue_request does not work. This can result
in IPsec crashes when the queue is depleted.
This patch fixes it by doing the pointer conversion only when the
return value is non-NULL. In particular, we create a new function
__crypto_dequeue_request that does the pointer conversion.
Reported-by: Brad Bosch <bradbosch@comcast.net>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Return the value we got from crypto_register_alg() instead of
returning 0 in any case.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The alignment calculation of xcbc_tfm_ctx uses alg->cra_alignmask
and not alg->cra_alignmask + 1 as it should. This led to frequent
crashes during the selftest of xcbc(aes-asm) on x86_64
machines. This patch fixes this. Also we use the alignmask
of xcbc and not the alignmask of the underlying algorithm
for the alignmnent calculation in xcbc_create now.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
What about something like this? It defaults the CPRNG to m and makes FIPS
dependent on the CPRNG. That way you get a module build by default, but you can
change it to y manually during config and still satisfy the dependency, and if
you select N it disables FIPS as well. I rather like that better than making
FIPS a tristate. I just tested it out here and it seems to work well. Let me
know what you think
Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Recently we switched to using eseqiv on SMP machines in preference
over chainiv. However, eseqiv does not support stream ciphers so
they should still default to chainiv.
This patch applies the same check as done by eseqiv to weed out
the stream ciphers. In particular, all algorithms where the IV
size is not equal to the block size will now default to chainiv.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Raw counter mode only works with chainiv, which is no longer
the default IV generator on SMP machines. This broke raw counter
mode as it can no longer instantiate as a givcipher.
This patch fixes it by always picking chainiv on raw counter
mode. This is based on the diagnosis and a patch by Huang
Ying.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This reverts commit 215ccd6f55.
It causes CPRNG and everything selected by it to be built-in
whenever FIPS is enabled. The problem is that it is selecting
a tristate from a bool, which is usually not what is intended.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Correct a typo in crypto/rng.c
Signed-off-by: Christian Kujau <lists@nerdbynature.de>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
cryptd_alloc_ahash() will allocate a cryptd-ed ahash for specified
algorithm name. The new allocated one is guaranteed to be cryptd-ed
ahash, so the shash underlying can be gotten via cryptd_ahash_child().
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Remove the dedicated GHASH implementation in GCM, and uses the GHASH
digest algorithm instead. This will make GCM uses hardware accelerated
GHASH implementation automatically if available.
ahash instead of shash interface is used, because some hardware
accelerated GHASH implementation needs asynchronous interface.
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
GHASH is implemented as a shash algorithm. The actual implementation
is copied from gcm.c. This makes it possible to add
architecture/hardware accelerated GHASH implementation.
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch converts authenc to the new ahash interface.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/djbw/async_tx:
dmaengine: at_hdmac: add DMA slave transfers
dmaengine: at_hdmac: new driver for the Atmel AHB DMA Controller
dmaengine: dmatest: correct thread_count while using multiple thread per channel
dmaengine: dmatest: add a maximum number of test iterations
drivers/dma: Remove unnecessary semicolons
drivers/dma/fsldma.c: Remove unnecessary semicolons
dmaengine: move HIGHMEM64G restriction to ASYNC_TX_DMA
fsldma: do not clear bandwidth control bits on the 83xx controller
fsldma: enable external start for the 83xx controller
fsldma: use PCI Read Multiple command
This patch uses crypto_shash_export/crypto_shash_import to prehash
ipad/opad to speed up hmac. This is partly based on a similar patch
by Steffen Klassert.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
It's undefined behaviour in C to write outside the bounds of an array.
The key expansion routine takes a shortcut of creating 8 words at a
time, but this creates 4 additional words which don't fit in the array.
As everyone is hopefully now aware, GCC is at liberty to make any
assumptions and optimisations it likes in situations where it can
detect that UB has occured, up to and including nasal demons, and
as the indices being accessed in the array are trivially calculable,
it's rash to invite gcc to do take any liberties at all.
Signed-off-by: Phil Carmody <ext-phil.2.carmody@nokia.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
crypto_init_shash_ops_async() tests for setkey and not for import
before exporting the algorithms import function to ahash.
This patch fixes this.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
ahash_op_unaligned() and ahash_def_finup() allocate memory atomically,
regardless whether the request can sleep or not. This patch changes
this to use GFP_KERNEL if the request can sleep.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch provides a default export/import function for all
shash algorithms. It simply copies the descriptor context as
is done by sha1_generic.
This in essence means that all existing shash algorithms now
support export/import. This is something that will be depended
upon in implementations such as hmac. Therefore all new shash
and ahash implementations must support export/import.
For those that cannot obtain a partial result, padlock-sha's
fallback model should be used so that a partial result is always
available.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch replaces the 32-bit counters in sha512_generic with
64-bit counters. It also switches the bit count to the simpler
byte count.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch renames struct sha512_ctx and exports it as struct
sha512_state so that other sha512 implementations can use it
as the reference structure for exporting their state.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Although xcbc was converted to shash, it didn't obey the new
requirement that all hash state must be stored in the descriptor
rather than the transform.
This patch fixes this issue and also optimises away the rekeying
by precomputing K2 and K3 within setkey.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds the finup/export/import functions to the cryptd
ahash implementation. We simply invoke the underlying shash
operations.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When shash_ahash_finup encounters a null request, we end up not
calling the underlying final function. This patch fixes that.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When the alignment check was made unconditional for ahash we
may end up crashing on shash algorithms because we're always
calling alg->setkey instead of tfm->setkey.
This patch fixes it.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
If cryptd_alloc_instance() fails, the return value is uninitialized.
This patch fixes this by setting the return value.
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch exports the finup operation where available and adds
a default finup operation for ahash. The operations final, finup
and digest also will now deal with unaligned result pointers by
copying it. Finally export/import operations are will now be
exported too.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We currently use GFP_ATOMIC in the unaligned setkey function
to allocate the temporary aligned buffer. Since setkey must
be called in a sleepable context, we can use GFP_KERNEL instead.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
