Ahsan Atta [Tue, 22 Dec 2015 19:20:34 +0000 (11:20 -0800)]
crypto: qat - Rename dh895xcc mmp firmware
Resending again. The fw name suppoed to be 895xcc instead of 895xxcc.
Sorry for the noise.
Rename dh895xcc mmp fw to make it consistent with other mmp images.
Signed-off-by: Ahsan Atta <ahsan.atta@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Dan Streetman [Mon, 21 Dec 2015 17:52:10 +0000 (12:52 -0500)]
crypto: 842 - remove WARN inside printk
Remove the WARN() from the beN_to_cpu macro, which is used as a param to a
pr_debug() call. With a certain kernel config, this printk-in-printk
results in the no_printk() macro trying to recursively call the
no_printk() macro, and since macros can't recursively call themselves
a build error results.
Reported-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 17:13:08 +0000 (18:13 +0100)]
crypto: atmel-aes - add debug facilities to monitor register accesses.
This feature should not be enabled in release but can be usefull for
developers who need to monitor register accesses at some specific places.
Set the AES_FLAGS_DUMP_REG flag inside dd->flags to start monitoring the
I/O accesses, clear it to stop monitoring.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 17:13:05 +0000 (18:13 +0100)]
crypto: atmel-aes - fix the counter overflow in CTR mode
Depending on its hardware version, the AES IP provides either a 16 or a
32 bit counter. However the CTR mode expects the size of the counter to be
the same as the size of the cipher block, ie 128 bits for AES.
This patch detects and handles counter overflows.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 17:13:03 +0000 (18:13 +0100)]
crypto: atmel-aes - create sections to regroup functions by usage
This patch only creates sections to regroup functions by usage.
This will help to integrate the GCM support patch later by making the
difference between shared/common and specific code. Hence current
sections are:
- Shared functions: common code which will be reused by the GCM support.
- CPU transfer: handles transfers monitored by the CPU (PIO accesses).
- DMA transfer: handles transfers monitored by the DMA controller.
- AES async block ciphers: dedicated to the already supported block ciphers
- Probe functions: used to register all crypto algorithms.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 17:13:00 +0000 (18:13 +0100)]
crypto: atmel-aes - improve performances of data transfer
This patch totally reworks data transfer.
1 - DMA
The new code now fully supports scatter-gather lists hence reducing the
number of interrupts in some cases. Also buffer alignments are better
managed to avoid useless copies.
2 - CPU
The new code allows to use PIO accesses even when transferring more than
one AES block, so futher patches could tune the DMA threshold
(ATMEL_AES_DMA_THRESHOLD).
Moreover, CPU transfers now have a chance to be processed synchronously,
hence reducing the latency by avoiding context switches when possible
(less interrupts to process, less scheduling of the 'done' task).
Indeed the 'DATA READY' bit is polled only one time in the Interrupt
Status Register before enabling then waiting for the associated interrupt.
In some condition, this single poll is enough as the data have already
been processed by the AES hardware and so are ready.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 16:48:44 +0000 (17:48 +0100)]
crypto: atmel-aes - reduce latency of DMA completion
atmel_aes_dma_callback() now directly calls the 'resume' callback instead
of scheduling the done task, which in turn only calls the very same
'resume' callback.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch introduces a new callback 'resume' in the struct atmel_aes_dev.
This callback is run to resume/complete the processing of the crypto
request when woken up by I/O events such as AES interrupts or DMA
completion.
This callback will help implementing the GCM mode support in further
patches.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 16:48:39 +0000 (17:48 +0100)]
crypto: atmel-aes - make crypto request queue management more generic
This patch changes atmel_aes_handle_queue() to make it more generic.
The function argument is now a pointer to struct crypto_async_request,
which is the common base of struct ablkcipher_request and
struct aead_request.
Also this patch introduces struct atmel_aes_base_ctx which will be the
common base of all the transformation contexts.
Hence the very same queue will be used to manage both block cipher and
AEAD requests (such as gcm and authenc implemented in further patches).
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch changes the signature of atmel_aes_write_ctrl() to make it more
generic. This will be used by future patches when implementing new block
cipher modes such as GCM.
Especially atmel_aes_hw_init() is now called outside
atmel_aes_write_ctrl(): this allows to call atmel_aes_write_ctrl() many
times, still initializing the hardware only once.
Indeed, the support of GCM will require to update the Mode Register and
the IV when processing a single request.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 16:48:37 +0000 (17:48 +0100)]
crypto: atmel-aes - propagate error from atmel_aes_hw_version_init()
Before this patch atmel_aes_hw_version_init() had no returned value.
However it calls atmel_aes_hw_init(), which may fail. So check the
returned code of atmel_aes_hw_init() and propagate error if needed.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 16:48:35 +0000 (17:48 +0100)]
crypto: atmel-aes - fix unregistration order of crypto algorithms
This dummy patch fixes atmel_aes_unregister_algs() so crypto algorithms
are unregistered in the reverse order they were registered by
atmel_aes_register_algs().
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Cyrille Pitchen [Thu, 17 Dec 2015 16:48:34 +0000 (17:48 +0100)]
crypto: atmel-aes - change algorithm priorities
Increase the algorithm priorities so the hardware acceleration is now
preferred to the software computation: the "aes-generice" driver uses 100
as priority.
Signed-off-by: Cyrille Pitchen <cyrille.pitchen@atmel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Pingchao Yang [Wed, 16 Dec 2015 06:09:50 +0000 (14:09 +0800)]
crypto: qat - fix some timeout tests
Change the timeout condition since the times value would be -1 after
running MAX_RETRY_TIMES.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Yang Pingchao <pingchao.yang@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Pingchao Yang [Wed, 16 Dec 2015 02:39:40 +0000 (10:39 +0800)]
crypto: qat - fix CTX_ENABLES bits shift direction issue
AE CTX bits should be 8-15 in CTX_ENABLES, so the mask
value 0xff should be left shifted 0x8.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Yang Pingchao <pingchao.yang@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
crypto: rsa-pkcs1pad - don't allocate buffer on stack
Avoid the s390 compile "warning: 'pkcs1pad_encrypt_sign_complete'
uses dynamic stack allocation" reported by kbuild test robot. Don't
use a flat zero-filled buffer, instead zero the contents of the SGL.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Rahul Pathak [Mon, 14 Dec 2015 08:45:23 +0000 (08:45 +0000)]
crypto: omap - Removed unused variable "err"
Removed unused variable "err" and directly return "0"
Reported by coccicheck -
./drivers/crypto/omap-aes.c:542:5-8: Unneeded variable: "err". Return "0" on line 551
./drivers/crypto/omap-des.c:530:5-8: Unneeded variable: "err". Return "0" on line 539
Signed-off-by: Rahul Pathak <rpathak@visteon.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Haren Myneni [Sun, 13 Dec 2015 11:30:41 +0000 (03:30 -0800)]
crypto: nx-842 - Mask XERS0 bit in return value
NX842 coprocessor sets 3rd bit in CR register with XER[S0] which is
nothing to do with NX request. Since this bit can be set with other
valuable return status, mast this bit.
One of other bits (INITIATED, BUSY or REJECTED) will be returned for
any given NX request.
Signed-off-by: Haren Myneni <haren@us.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tadeusz Struk [Thu, 10 Dec 2015 22:23:03 +0000 (14:23 -0800)]
crypto: qat - uint8_t is not large enough for accel_id
accel_id has to be large enough to hold ADF_MAX_DEVICES + 1
(which is > 1025) so uint8_t is too small.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Sam Protsenko [Thu, 10 Dec 2015 16:06:59 +0000 (18:06 +0200)]
crypto: omap-des - Fix "schedule while atomic" bug
When using DES module the next bug appears:
BUG: scheduling while atomic: kworker/0:1/63/0x00000102
With backtrace as follows:
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< cut here >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
[<c0012294>] (dump_backtrace) from [<c00124ac>] (show_stack+0x18/0x1c)
[<c0012494>] (show_stack) from [<c0752554>] (dump_stack+0x84/0xc4)
[<c07524d0>] (dump_stack) from [<c0750218>] (__schedule_bug+0x54/0x64)
[<c07501c4>] (__schedule_bug) from [<c07548a4>] (__schedule+0x4ac/0x53c)
[<c07543f8>] (__schedule) from [<c075496c>] (schedule+0x38/0x88)
[<c0754934>] (schedule) from [<c03c3984>] (rpm_resume+0x158/0x59c)
[<c03c382c>] (rpm_resume) from [<c03c3e1c>] (__pm_runtime_resume+0x54/0x6c)
[<c03c3dc8>] (__pm_runtime_resume) from [<c0568ff8>] (omap_des_handle_queue+0x154/0x7bc)
[<c0568ea4>] (omap_des_handle_queue) from [<c05696b8>] (omap_des_crypt+0x58/0xbc)
[<c0569660>] (omap_des_crypt) from [<c0569730>] (omap_des_cbc_decrypt+0x14/0x18)
[<c056971c>] (omap_des_cbc_decrypt) from [<c0297534>] (authenc_verify_ahash_done+0xe0/0xe8)
[<c0297454>] (authenc_verify_ahash_done) from [<c056a330>] (omap_sham_finish_req+0x58/0xa8)
[<c056a2d8>] (omap_sham_finish_req) from [<c056b714>] (omap_sham_done_task+0x1c0/0x1e0)
[<c056b554>] (omap_sham_done_task) from [<c003e53c>] (tasklet_action+0x80/0x118)
[<c003e4bc>] (tasklet_action) from [<c003e740>] (__do_softirq+0x11c/0x260)
[<c003e624>] (__do_softirq) from [<c003eb64>] (irq_exit+0xc0/0xfc)
[<c003eaa4>] (irq_exit) from [<c000f1c4>] (handle_IRQ+0x4c/0x98)
[<c000f178>] (handle_IRQ) from [<c0008568>] (gic_handle_irq+0x34/0x64)
[<c0008534>] (gic_handle_irq) from [<c0758540>] (__irq_svc+0x40/0x70)
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< cut here >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Insight was seen in drivers/crypto/omap-sham.c driver.
All credits for this patch go to Grygorii Strashko.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Zeng Xin [Thu, 10 Dec 2015 05:38:30 +0000 (21:38 -0800)]
crypto: qat - enable VF irq after guest exits ungracefully
The VF bundle interrupt is not triggered any more in
the case when guest is shut down with sample app running.
Need to clear the flag interrupt bit when restarting to fix
this irrecoverable state.
Signed-off-by: Zeng Xin <xin.zeng@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Paul Gortmaker [Wed, 9 Dec 2015 20:05:28 +0000 (15:05 -0500)]
crypto: asymmetric_keys - signature.c does not need <module.h>
This file does not contain any modular related function calls. So get
rid of module.h since it drags in a lot of other headers and adds to
the preprocessing load. It does export some symbols though, so we'll
need to ensure it has export.h present instead.
Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: "David S. Miller" <davem@davemloft.net> Cc: linux-crypto@vger.kernel.org Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Harvijay Saini [Wed, 9 Dec 2015 19:59:45 +0000 (11:59 -0800)]
crypto: qat - ring returning retry even though ring has BW
When many threads submit multiple requests they get blocked until all
responses are processed, which prevents them from submitting more requests
even though there is space on the rings.
To fix this we need to decrement the inflight counter early to in the callback.
Signed-off-by: Harvijay Saini <harvijayx.saini@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Arnd Bergmann [Tue, 8 Dec 2015 15:24:22 +0000 (16:24 +0100)]
crypto: sahara - fix debug output for 64-bit dma_addr_t
The sahara_dump_descriptors and sahara_dump_links functions attempt
to print a dma_addr_t value with a 0x%08x format string, which
produces a warning when dma_addr_t is 64-bit wide:
drivers/crypto/sahara.c:419:120: warning: format '%x' expects argument of type 'unsigned int', but argument 5 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=]
This changes the code to use the %pad format string that is meant
for dma_addr_t, which avoids the warning and gives us the correct
output in all configurations.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The sahara hardware uses DMA descriptors with 32-bit addresses, but
dma_addr_t is variable size depending on whether we want to support
any devices that use 64-bit DMA addresses in hardware.
This means that the definition of the DMA descriptor structure is wrong,
and we helpfully get a compiler warning about them too:
drivers/crypto/sahara.c:423:372: warning: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'dma_addr_t {aka long long unsigned int}' [-Wformat=]
This changes the definition of the sahara_hw_desc and sahara_hw_link
structures to only contain fixed-length members, which is required
to make the driver work on ARM LPAE mode, and avoids most of the
gcc warnings we get.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
crypto: chacha20poly1305 - Skip encryption/decryption for 0-len
If the length of the plaintext is zero, there's no need to waste cycles
on encryption and decryption. Using the chacha20poly1305 construction
for zero-length plaintexts is a common way of using a shared encryption
key for AAD authentication.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds PKCS#1 v1.5 standard RSA padding as a separate template.
This way an RSA cipher with padding can be obtained by instantiating
"pkcs1pad(rsa)". The reason for adding this is that RSA is almost
never used without this padding (or OAEP) so it will be needed for
either certificate work in the kernel or the userspace, and I also hear
that it is likely implemented by hardware RSA in which case hardware
implementations of the whole of pkcs1pad(rsa) can be provided.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
crypto: akcipher - add akcipher declarations needed by templates.
Add a struct akcipher_instance and struct akcipher_spawn similar to
how AEAD declares them and the macros for converting to/from
crypto_instance/crypto_spawn. Also add register functions to
avoid exposing crypto_akcipher_type.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Tadeusz Struk [Sat, 5 Dec 2015 00:56:28 +0000 (16:56 -0800)]
crypto: qat - move isr files to qat common so that they can be reused
Move qat_isr.c and qat_isrvf.c files to qat_common dir
so that they can be reused by all devices.
Remove adf_drv.h files because thay are not longer needed.
Move adf_dev_configure() function to qat_common so it can be reused.
Also some minor updates to common code for multidevice.
Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Thierry Reding [Wed, 2 Dec 2015 16:16:36 +0000 (17:16 +0100)]
crypto: n2 - Use platform_register/unregister_drivers()
These new helpers simplify implementing multi-driver modules and
properly handle failure to register one driver by unregistering all
previously registered drivers.
Signed-off-by: Thierry Reding <treding@nvidia.com> Acked-by: David S. Miller <davem@davemloft.net> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
IV size was zero on CBC and CTR modes,
causing a bug triggered by skcipher.
Fixing this adding a correct size.
Signed-off-by: Leonidas Da Silva Barbosa <leosilva@linux.vnet.ibm.com> Signed-off-by: Paulo Smorigo <pfsmorigo@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Wang, Rui Y [Sun, 29 Nov 2015 14:45:34 +0000 (22:45 +0800)]
crypto: cryptd - Assign statesize properly
cryptd_create_hash() fails by returning -EINVAL. It is because after 8996eafdc ("crypto: ahash - ensure statesize is non-zero") all ahash
drivers must have a non-zero statesize.
This patch fixes the problem by properly assigning the statesize.
Signed-off-by: Rui Wang <rui.y.wang@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Wang, Rui Y [Sun, 29 Nov 2015 14:45:33 +0000 (22:45 +0800)]
crypto: ghash-clmulni - Fix load failure
ghash_clmulni_intel fails to load on Linux 4.3+ with the following message:
"modprobe: ERROR: could not insert 'ghash_clmulni_intel': Invalid argument"
After 8996eafdc ("crypto: ahash - ensure statesize is non-zero") all ahash
drivers are required to implement import()/export(), and must have a non-
zero statesize.
This patch has been tested with the algif_hash interface. The calculated
digest values, after several rounds of import()s and export()s, match those
calculated by tcrypt.
Signed-off-by: Rui Wang <rui.y.wang@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Jiri Slaby [Fri, 27 Nov 2015 15:50:43 +0000 (16:50 +0100)]
hwrng: core - sleep interruptible in read
hwrng kthread can be waiting via hwrng_fillfn for some data from a rng
like virtio-rng:
hwrng D ffff880093e17798 0 382 2 0x00000000
...
Call Trace:
[<ffffffff817339c6>] wait_for_completion_killable+0x96/0x210
[<ffffffffa00aa1b7>] virtio_read+0x57/0xf0 [virtio_rng]
[<ffffffff814f4a35>] hwrng_fillfn+0x75/0x130
[<ffffffff810aa243>] kthread+0xf3/0x110
And this is indeed unkillable. So use mutex_lock_interruptible
instead of mutex_lock in rng_dev_read and exit immediatelly when
interrupted. And possibly return already read data, if any (as POSIX
allows).
v2: use ERESTARTSYS instead of EINTR
Signed-off-by: Jiri Slaby <jslaby@suse.cz> Cc: Matt Mackall <mpm@selenic.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: <linux-crypto@vger.kernel.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
LABBE Corentin [Thu, 19 Nov 2015 12:38:18 +0000 (13:38 +0100)]
crypto: picoxcell - set [src|dst]_nents and nents as signed int
The unsigned int variables [src|dst]_nents and nents can be assigned
signed value (-EINVAL) from sg_nents_for_len().
Furthermore they are used only by dma_map_sg and dma_unmap_sg which wait
for an signed int, so they must be set as int.
Fixes: f051f95eb47b ("crypto: picoxcell - check return value of sg_nents_for_len") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
LABBE Corentin [Thu, 19 Nov 2015 12:38:17 +0000 (13:38 +0100)]
crypto: sahara - set nb_[in|out]_sg as signed int
The two unsigned int variables nb_in_sg and nb_out_sg can be assigned
signed value (-EINVAL) from sg_nents_for_len().
Furthermore they are used only by dma_map_sg and dma_unmap_sg which wait
for an signed int, so they must be set as int.
Fixes: 6c2b74d4774f ("crypto: sahara - check return value of sg_nents_for_len") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Aaro Koskinen [Wed, 18 Nov 2015 19:59:01 +0000 (21:59 +0200)]
hwrng: omap3-rom - convert timer to delayed work
We cannot put the HW RNG to idle using a timer because we cannot disable
clocks from atomic context. Use a delayed work instead.
Fixes a warning with CONFIG_DEBUG_MUTEXES on Nokia N900 during boot.
Reported-by: Sebastian Reichel <sre@kernel.org> Signed-off-by: Aaro Koskinen <aaro.koskinen@iki.fi> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Arnd Bergmann [Tue, 27 Jan 2015 21:34:04 +0000 (22:34 +0100)]
crypto: atmel: fix bogus select
The Atmel at91 crypto driver unconditionally selects AT_HDMAC,
which results in a Kconfig warning if that driver is not enabled:
warning: (CRYPTO_DEV_ATMEL_AES) selects AT_HDMAC which has unmet direct dependencies (DMADEVICES && ARCH_AT91)
The crypto driver itself does not actually have a dependency
on a particular dma engine, other than this being the one that
is used in at91.
Removing the 'select' gets rid of the warning, but can cause
the driver to be unusable if the HDMAC is not enabled at the
same time. To work around that, this patch clarifies the runtime
dependency to be 'AT_HDMAC || AT_XDMAC', but adds an alternative
for COMPILE_TEST, which lets the driver get build on all systems.
The ARCH_AT91 dependency is implied by AT_XDMAC || AT_HDMAC now
and no longer needs to be listed separately.
Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Arnd Bergmann [Tue, 17 Nov 2015 09:22:06 +0000 (10:22 +0100)]
crypto: atmel - fix 64-bit warnings
The atmel AES driver assumes that 'int' and 'size_t' are the same
type in multiple locations, which the compiler warns about when
building it for 64-bit systems:
In file included from ../drivers/crypto/atmel-aes.c:17:0:
drivers/crypto/atmel-aes.c: In function 'atmel_aes_sg_copy':
include/linux/kernel.h:724:17: warning: comparison of distinct pointer types lacks a cast
drivers/crypto/atmel-aes.c:448:11: note: in expansion of macro 'min'
drivers/crypto/atmel-aes.c: In function 'atmel_aes_crypt_dma_stop':
include/linux/kern_levels.h:4:18: warning: format '%u' expects argument of type 'unsigned int', but argument 2 has type 'size_t {aka long unsigned int}' [-Wformat=]
This changes the format strings to use the %z modifier when printing
a size_t, and makes sure that we use the correct size_t type where
needed. In case of sg_dma_len(), the type of the result depends
on CONFIG_NEED_SG_DMA_LENGTH, so we have to use min_t to get it to
work in all configurations.
LABBE Corentin [Mon, 16 Nov 2015 08:35:54 +0000 (09:35 +0100)]
crypto: sun4i-ss - add missing statesize
sun4i-ss implementaton of md5/sha1 is via ahash algorithms.
Commit 8996eafdcbad ("crypto: ahash - ensure statesize is non-zero")
made impossible to load them without giving statesize. This patch
specifiy statesize for sha1 and md5.
crypto: rsa - only require output buffers as big as needed.
rhe RSA operations explicitly left-align the integers being written
skipping any leading zero bytes, but still require the output buffers to
include just enough space for the integer + the leading zero bytes.
Since the size of integer + the leading zero bytes (i.e. the key modulus
size) can now be obtained more easily through crypto_akcipher_maxsize
change the operations to only require as big a buffer as actually needed
if the caller has that information. The semantics for request->dst_len
don't change.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
lib/mpi: only require buffers as big as needed for the integer
Since mpi_write_to_sgl and mpi_read_buffer explicitly left-align the
integers being written it makes no sense to require a buffer big enough for
the number + the leading zero bytes which are not written. The error
returned also doesn't convey any information. So instead require only the
size needed and return -EOVERFLOW to signal when buffer too short.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
LABBE Corentin [Wed, 4 Nov 2015 20:13:37 +0000 (21:13 +0100)]
crypto: picoxcell - check return value of sg_nents_for_len
The sg_nents_for_len() function could fail, this patch add a check for
its return value.
In the same time, we remove sg_count() as it is used as an alias of
sg_nents_for_len.
LABBE Corentin [Fri, 23 Oct 2015 12:10:36 +0000 (14:10 +0200)]
crypto: algif - Change some variable to size_t
Some variable are set as int but store only positive values.
Furthermore there are used in operation/function that wait for unsigned
value.
This patch set them as size_t.
Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
David Gstir [Sun, 15 Nov 2015 16:14:42 +0000 (17:14 +0100)]
crypto: talitos - Fix timing leak in ESP ICV verification
Using non-constant time memcmp() makes the verification of the authentication
tag in the decrypt path vulnerable to timing attacks. Fix this by using
crypto_memneq() instead.
Cc: stable@vger.kernel.org Signed-off-by: David Gstir <david@sigma-star.at> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
projects / karo-tx-linux.git / log