When the DRBG is initialized, the core is looked up using the DRBG name.
The name that can be used for the lookup is registered in
cra_driver_name. The cra_name value contains stdrng.
Thus, the lookup code must use crypto_tfm_alg_driver_name to obtain the
precise DRBG name and select the correct DRBG.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CTR DRBG requires the update function to be called twice when
generating a random number. In both cases, update function must process
the additional information string by using the DF function. As the DF
produces the same result in both cases, we can save one invocation of
the DF function when the first DF function result is reused.
The result of the DF function is stored in the scratchpad storage. The
patch ensures that the scratchpad is not cleared when we want to reuse
the DF result. For achieving this, the CTR DRBG update function must
know by whom and in which scenario it is called. This information is
provided with the reseed parameter to the update function.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The initial format strings caused warnings on several architectures. The
updated format strings now match the variable types.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Reported-by: Randy Dunlap <rdunlap@infradead.org>
CC: Joe Perches <joe@perches.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The structure used to construct the module description line was marked
problematic by the sparse code analysis tool. The module line
description now does not contain any ifdefs to prevent error reports
from sparse.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch removes the build-time test that ensures at least one RNG
is set. Instead we will simply not build drbg if no options are set
through Kconfig.
This also fixes a typo in the name of the Kconfig option CRYTPO_DRBG
(should be CRYPTO_DRBG).
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The DRBG-style linked list to manage input data that is fed into the
cipher invocations is replaced with the kernel linked list
implementation.
The change is transparent to users of the interfaces offered by the
DRBG. Therefore, no changes to the testmgr code is needed.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
For the CTR DRBG, the drbg_state->scratchpad temp buffer (i.e. the
memory location immediately before the drbg_state->tfm variable
is the buffer that the BCC function operates on. BCC operates
blockwise. Making the temp buffer drbg_statelen(drbg) in size is
sufficient when the DRBG state length is a multiple of the block
size. For AES192 this is not the case and the length for temp is
insufficient (yes, that also means for such ciphers, the final
output of all BCC rounds are truncated before used to update the
state of the DRBG!!).
The patch enlarges the temp buffer from drbg_statelen to
drbg_statelen + drbg_blocklen to have sufficient space.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Print the driver name that is being tested. The driver name can be
inferred parsing /proc/crypto but having it in the output is
clearer
Signed-off-by: Luca Clementi <luca.clementi@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Per further discussion with NIST, the requirements for FIPS state that
we only need to panic the system on failed kernel module signature checks
for crypto subsystem modules. This moves the fips-mode-only module
signature check out of the generic module loading code, into the crypto
subsystem, at points where we can catch both algorithm module loads and
mode module loads. At the same time, make CONFIG_CRYPTO_FIPS dependent on
CONFIG_MODULE_SIG, as this is entirely necessary for FIPS mode.
v2: remove extraneous blank line, perform checks in static inline
function, drop no longer necessary fips.h include.
CC: "David S. Miller" <davem@davemloft.net>
CC: Rusty Russell <rusty@rustcorp.com.au>
CC: Stephan Mueller <stephan.mueller@atsec.com>
Signed-off-by: Jarod Wilson <jarod@redhat.com>
Acked-by: Neil Horman <nhorman@tuxdriver.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
As reported by a static code analyzer, the code for the ordering of
the linked list can be simplified.
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
kvfree() helper is now available, use it instead of open code it.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This adds 4 test vectors for GHASH (of which one for chunked mode), making
a total of 5.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The DRBG test code implements the CAVS test approach.
As discussed for the test vectors, all DRBG types are covered with
testing. However, not every backend cipher is covered with testing. To
prevent the testmgr from logging missing testing, the NULL test is
registered for all backend ciphers not covered with specific test cases.
All currently implemented DRBG types and backend ciphers are defined
in SP800-90A. Therefore, the fips_allowed flag is set for all.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
All types of the DRBG (CTR, HMAC, Hash) are covered with test vectors.
In addition, all permutations of use cases of the DRBG are covered:
* with and without predition resistance
* with and without additional information string
* with and without personalization string
As the DRBG implementation is agnositc of the specific backend cipher,
only test vectors for one specific backend cipher is used. For example:
the Hash DRBG uses the same code paths irrespectively of using SHA-256
or SHA-512. Thus, the test vectors for SHA-256 cover the testing of all
DRBG code paths of SHA-512.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The different DRBG types of CTR, Hash, HMAC can be enabled or disabled
at compile time. At least one DRBG type shall be selected.
The default is the HMAC DRBG as its code base is smallest.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This is a clean-room implementation of the DRBG defined in SP800-90A.
All three viable DRBGs defined in the standard are implemented:
* HMAC: This is the leanest DRBG and compiled per default
* Hash: The more complex DRBG can be enabled at compile time
* CTR: The most complex DRBG can also be enabled at compile time
The DRBG implementation offers the following:
* All three DRBG types are implemented with a derivation function.
* All DRBG types are available with and without prediction resistance.
* All SHA types of SHA-1, SHA-256, SHA-384, SHA-512 are available for
the HMAC and Hash DRBGs.
* All AES types of AES-128, AES-192 and AES-256 are available for the
CTR DRBG.
* A self test is implemented with drbg_healthcheck().
* The FIPS 140-2 continuous self test is implemented.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
zswap allocates one LZO context per online cpu.
Using vmalloc() for small (16KB) memory areas has drawback of slowing
down /proc/vmallocinfo and /proc/meminfo reads, TLB pressure and poor
NUMA locality, as default NUMA policy at boot time is to interleave
pages :
edumazet:~# grep lzo /proc/vmallocinfo | head -4
0xffffc90006062000-0xffffc90006067000 20480 lzo_init+0x1b/0x30 pages=4 vmalloc N0=2 N1=2
0xffffc90006067000-0xffffc9000606c000 20480 lzo_init+0x1b/0x30 pages=4 vmalloc N0=2 N1=2
0xffffc9000606c000-0xffffc90006071000 20480 lzo_init+0x1b/0x30 pages=4 vmalloc N0=2 N1=2
0xffffc90006071000-0xffffc90006076000 20480 lzo_init+0x1b/0x30 pages=4 vmalloc N0=2 N1=2
This patch tries a regular kmalloc() and fallback to vmalloc in case
memory is too fragmented.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iEYEABECAAYFAlOTY+wACgkQuseO5dulBZXrIgCdFZyXRojufLLKikWEvHjZ3/k5
KsQAnimtcge+62/IX7YwDjWS+xg9Wt3m
=yPrI
-----END PGP SIGNATURE-----
Merge tag 'llvmlinux-for-v3.16' of git://git.linuxfoundation.org/llvmlinux/kernel
Pull LLVM patches from Behan Webster:
"Next set of patches to support compiling the kernel with clang.
They've been soaking in linux-next since the last merge window.
More still in the works for the next merge window..."
* tag 'llvmlinux-for-v3.16' of git://git.linuxfoundation.org/llvmlinux/kernel:
arm, unwind, LLVMLinux: Enable clang to be used for unwinding the stack
ARM: LLVMLinux: Change "extern inline" to "static inline" in glue-cache.h
all: LLVMLinux: Change DWARF flag to support gcc and clang
net: netfilter: LLVMLinux: vlais-netfilter
crypto: LLVMLinux: aligned-attribute.patch
__attribute__((aligned)) applies the default alignment for the largest scalar
type for the target ABI. gcc allows it to be applied inline to a defined type.
Clang only allows it to be applied to a type definition (PR11071).
Making it into 2 lines makes it more readable and works with both compilers.
Author: Mark Charlebois <charlebm@gmail.com>
Signed-off-by: Mark Charlebois <charlebm@gmail.com>
Signed-off-by: Behan Webster <behanw@converseincode.com>
Test vectors were taken from existing test for
CBC(DES3_EDE). Associated data has been added to test vectors.
HMAC computed with Crypto++ has been used. Following algos have
been covered.
(a) "authenc(hmac(sha1),cbc(des))"
(b) "authenc(hmac(sha1),cbc(des3_ede))"
(c) "authenc(hmac(sha224),cbc(des))"
(d) "authenc(hmac(sha224),cbc(des3_ede))"
(e) "authenc(hmac(sha256),cbc(des))"
(f) "authenc(hmac(sha256),cbc(des3_ede))"
(g) "authenc(hmac(sha384),cbc(des))"
(h) "authenc(hmac(sha384),cbc(des3_ede))"
(i) "authenc(hmac(sha512),cbc(des))"
(j) "authenc(hmac(sha512),cbc(des3_ede))"
Signed-off-by: Vakul Garg <vakul@freescale.com>
[NiteshNarayanLal@freescale.com: added hooks for the missing algorithms test and tested the patch]
Signed-off-by: Nitesh Lal <NiteshNarayanLal@freescale.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
With DMA-API debug enabled testmgr triggers a "DMA-API: device driver maps memory from stack" warning, when tested on a crypto HW accelerator.
Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Although the existing hash walk interface has already been used
by a number of ahash crypto drivers, it turns out that none of
them were really asynchronous. They were all essentially polling
for completion.
That's why nobody has noticed until now that the walk interface
couldn't work with a real asynchronous driver since the memory
is mapped using kmap_atomic.
As we now have a use-case for a real ahash implementation on x86,
this patch creates a minimal ahash walk interface. Basically it
just calls kmap instead of kmap_atomic and does away with the
crypto_yield call. Real ahash crypto drivers don't need to yield
since by definition they won't be hogging the CPU.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
CRYPTO_USER requires CAP_NET_ADMIN for all operations. Most information
provided by CRYPTO_MSG_GETALG is also accessible through /proc/modules
and AF_ALG. CRYPTO_MSG_GETALG should not require CAP_NET_ADMIN so that
processes without CAP_NET_ADMIN can use CRYPTO_MSG_GETALG to get cipher
details, such as cipher priorities, for AF_ALG.
Signed-off-by: Matthias-Christian Ott <ott@mirix.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Fix leakage of memory for struct aead_request that is allocated via
aead_request_alloc() but not released via aead_request_free().
Reported by Coverity - CID 1163869.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Fix a potential memory leak in the error handling of test_aead_speed(). In case
crypto_alloc_aead() fails, the function returns without going through the
centralized cleanup path. Reported by Coverity - CID 1163870.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Fix a potential memory leak in the error handling of test_aead_speed(). In case
the size check on the associate data length parameter fails, the function goes
through the wrong exit label. Reported by Coverity - CID 1163870.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at>
Acked-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
It is possible by passing a netlink socket to a more privileged
executable and then to fool that executable into writing to the socket
data that happens to be valid netlink message to do something that
privileged executable did not intend to do.
To keep this from happening replace bare capable and ns_capable calls
with netlink_capable, netlink_net_calls and netlink_ns_capable calls.
Which act the same as the previous calls except they verify that the
opener of the socket had the desired permissions as well.
Reported-by: Andy Lutomirski <luto@amacapital.net>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Patch adds large test-vectors for SHA algorithms for better code coverage in
optimized assembly implementations. Empty test-vectors are also added, as some
crypto drivers appear to have special case handling for empty input.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This adds test cases for SHA-1, SHA-224, SHA-256 and AES-CCM with an input size
that is an exact multiple of the block size. The reason is that some
implementations use a different code path for these cases.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This git patch adds x86_64 AVX2 optimization of SHA1
transform to crypto support. The patch has been tested with 3.14.0-rc1
kernel.
On a Haswell desktop, with turbo disabled and all cpus running
at maximum frequency, tcrypt shows AVX2 performance improvement
from 3% for 256 bytes update to 16% for 1024 bytes update over
AVX implementation.
This patch adds sha1_avx2_transform(), the glue, build and
configuration changes needed for AVX2 optimization of
SHA1 transform to crypto support.
sha1-ssse3 is one module which adds the necessary optimization
support (SSSE3/AVX/AVX2) for the low-level SHA1 transform function.
With better optimization support, transform function is overridden
as the case may be. In the case of AVX2, due to performance reasons
across datablock sizes, the AVX or AVX2 transform function is used
at run-time as it suits best. The Makefile change therefore appends
the necessary objects to the linkage. Due to this, the patch merely
appends AVX2 transform to the existing build mix and Kconfig support
and leaves the configuration build support as is.
Signed-off-by: Chandramouli Narayanan <mouli@linux.intel.com>
Reviewed-by: Marek Vasut <marex@denx.de>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The crypto algorithm modules utilizing the crypto daemon could
be used early when the system start up. Using module_init
does not guarantee that the daemon's work queue is initialized
when the cypto alorithm depending on crypto_wq starts. It is necessary
to initialize the crypto work queue earlier at the subsystem
init time to make sure that it is initialized
when used.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add test vectors for aead with null encryption and md5,
respectively sha1 authentication.
Input data is taken from test vectors listed in RFC2410.
Signed-off-by: Horia Geanta <horia.geanta@freescale.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
These defines might be needed by crypto drivers.
Signed-off-by: Horia Geanta <horia.geanta@freescale.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The ahash_def_finup() can make use of the request save/restore functions,
thus make it so. This simplifies the code a little and unifies the code
paths.
Note that the same remark about free()ing the req->priv applies here, the
req->priv can only be free()'d after the original request was restored.
Finally, squash a bug in the invocation of completion in the ASYNC path.
In both ahash_def_finup_done{1,2}, the function areq->base.complete(X, err);
was called with X=areq->base.data . This is incorrect , as X=&areq->base
is the correct value. By analysis of the data structures, we see the areq is
of type 'struct ahash_request' , areq->base is of type 'struct crypto_async_request'
and areq->base.completion is of type crypto_completion_t, which is defined in
include/linux/crypto.h as:
typedef void (*crypto_completion_t)(struct crypto_async_request *req, int err);
This is one lead that the X should be &areq->base . Next up, we can inspect
other code which calls the completion callback to give us kind-of statistical
idea of how this callback is used. We can try:
$ git grep base\.complete\( drivers/crypto/
Finally, by inspecting ahash_request_set_callback() implementation defined
in include/crypto/hash.h , we observe that the .data entry of 'struct
crypto_async_request' is intended for arbitrary data, not for completion
argument.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: David S. Miller <davem@davemloft.net>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Shawn Guo <shawn.guo@linaro.org>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The functions to save original request within a newly adjusted request
and it's counterpart to restore the original request can be re-used by
more code in the crypto/ahash.c file. Pull these functions out from the
code so they're available.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: David S. Miller <davem@davemloft.net>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Shawn Guo <shawn.guo@linaro.org>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Add documentation for the pointer voodoo that is happening in crypto/ahash.c
in ahash_op_unaligned(). This code is quite confusing, so add a beefy chunk
of documentation.
Moreover, make sure the mangled request is completely restored after finishing
this unaligned operation. This means restoring all of .result, .base.data
and .base.complete .
Also, remove the crypto_completion_t complete = ... line present in the
ahash_op_unaligned_done() function. This type actually declares a function
pointer, which is very confusing.
Finally, yet very important nonetheless, make sure the req->priv is free()'d
only after the original request is restored in ahash_op_unaligned_done().
The req->priv data must not be free()'d before that in ahash_op_unaligned_finish(),
since we would be accessing previously free()'d data in ahash_op_unaligned_done()
and cause corruption.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: David S. Miller <davem@davemloft.net>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Shawn Guo <shawn.guo@linaro.org>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This adds the function blkcipher_aead_walk_virt_block, which allows the caller
to use the blkcipher walk API to handle the input and output scatterlists.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In order to allow other uses of the blkcipher walk API than the blkcipher
algos themselves, this patch copies some of the transform data members to the
walk struct so the transform is only accessed at walk init time.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
We added the soft module dependency of crc32c module alias
to generic crc32c module so other hardware accelerated crc32c
modules could get loaded and used before the generic version.
We also renamed the crypto/crc32c.c containing the generic
crc32c crypto computation to crypto/crc32c_generic.c according
to convention.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When finishing the ahash request, the ahash_op_unaligned_done() will
call complete() on the request. Yet, this will not call the correct
complete callback. The correct complete callback was previously stored
in the requests' private data, as seen in ahash_op_unaligned(). This
patch restores the correct complete callback and .data field of the
request before calling complete() on it.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: David S. Miller <davem@davemloft.net>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Adding simple speed tests for a range of block sizes for AEAD crypto
algorithms.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Commit fe8c8a1268 introduced a possible build error for archs
that do not have CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS set. :/
Fix this up by bringing else braces outside of the ifdef.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Fixes: fe8c8a1268 ("crypto: more robust crypto_memneq")
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Acked-By: Cesar Eduardo Barros <cesarb@cesarb.eti.br>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
A kernel with enabled lockdep complains about the wrong usage of
rcu_dereference() under a rcu_read_lock_bh() protected region.
===============================
[ INFO: suspicious RCU usage. ]
3.13.0-rc1+ #126 Not tainted
-------------------------------
linux/crypto/pcrypt.c:81 suspicious rcu_dereference_check() usage!
other info that might help us debug this:
rcu_scheduler_active = 1, debug_locks = 1
1 lock held by cryptomgr_test/153:
#0: (rcu_read_lock_bh){.+....}, at: [<ffffffff812c8075>] pcrypt_do_parallel.isra.2+0x5/0x200
Fix that by using rcu_dereference_bh() instead.
Signed-off-by: Mathias Krause <minipli@googlemail.com>
Cc: "David S. Miller" <davem@davemloft.net>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Stephan Mueller
Herbert Xu
Luca Clementi
Jarod Wilson
Eric Dumazet
Ard Biesheuvel
Jussi Kivilinna
Linus Torvalds
Mark Charlebois
Nitesh Lal
Tadeusz Struk
Matthias-Christian Ott
Christian Engelmayer
Eric W. Biederman
Peter Zijlstra
chandramouli narayanan
Tim Chen
Horia Geanta
Marek Vasut
Daniel Borkmann
Mathias Krause