- 12 Apr, 2024 24 commits
-
-
Eric Biggers authored
Since sha256_transform_rorx() uses ymm registers, execute vzeroupper before returning from it. This is necessary to avoid reducing the performance of SSE code. Fixes: d34a4600 ("crypto: sha256 - Optimized sha256 x86_64 routine using AVX2's RORX instructions") Signed-off-by:
Eric Biggers <ebiggers@google.com> Acked-by:
Tim Chen <tim.c.chen@linux.intel.com> Signed-off-by:
Herbert Xu <herbert@gondor.apana.org.au>
-
Eric Biggers authored
Since nh_avx2() uses ymm registers, execute vzeroupper before returning from it. This is necessary to avoid reducing the performance of SSE code. Fixes: 0f961f9f ("crypto: x86/nhpoly1305 - add AVX2 accelerated NHPoly1305") Signed-off-by:
-
Tom Zanussi authored
If some cpus are offlined, or if the node mask is smaller than expected, the 'nonexistent cpu' warning in rebalance_wq_table() may be erroneously triggered. Use cpumask_weight() to make sure we only iterate over the exact number of cpus in the mask. Also use num_possible_cpus() instead of num_online_cpus() to make sure all slots in the wq table are initialized. Signed-off-by:
Tom Zanussi <tom.zanussi@linux.intel.com> Signed-off-by:
-
Stefan Berger authored
Enable the x509 parser to accept NIST P521 certificates and add the OID for ansip521r1, which is the identifier for NIST P521. Cc: David Howells <dhowells@redhat.com> Tested-by:
Lukas Wunner <lukas@wunner.de> Reviewed-by:
Jarkko Sakkinen <jarkko@kernel.org> Signed-off-by:
Stefan Berger <stefanb@linux.ibm.com> Signed-off-by:
-
Stefan Berger authored
Adjust the calculation of the maximum signature size for support of NIST P521. While existing curves may prepend a 0 byte to their coordinates (to make the number positive), NIST P521 will not do this since only the first bit in the most significant byte is used. If the encoding of the x & y coordinates requires at least 128 bytes then an additional byte is needed for the encoding of the length. Take this into account when calculating the maximum signature size. Reviewed-by:
-
Stefan Berger authored
Register NIST P521 as an akcipher and extend the testmgr with NIST P521-specific test vectors. Add a module alias so the module gets automatically loaded by the crypto subsystem when the curve is needed. Tested-by:
-
Stefan Berger authored
In cases where 'keylen' was referring to the size of the buffer used by a curve's digits, it does not reflect the purpose of the variable anymore once NIST P521 is used. What it refers to then is the size of the buffer, which may be a few bytes larger than the size a coordinate of a key. Therefore, rename keylen to bufsize where appropriate. Tested-by:
-
Stefan Berger authored
Replace the usage of ndigits with nbits where precise space calculations are needed, such as in ecdsa_max_size where the length of a coordinate is determined. Tested-by:
-
Stefan Berger authored
Add the parameters for the NIST P521 curve and define a new curve ID for it. Make the curve available in ecc_get_curve. Tested-by:
-
Stefan Berger authored
In ecc_point_mult use the number of bits of the NIST P521 curve + 2. The change is required specifically for NIST P521 to pass mathematical tests on the public key. Tested-by:
-
Stefan Berger authored
Implement vli_mmod_fast_521 following the description for how to calculate the modulus for NIST P521 in the NIST publication "Recommendations for Discrete Logarithm-Based Cryptography: Elliptic Curve Domain Parameters" section G.1.4. NIST p521 requires 9 64bit digits, so increase the ECC_MAX_DIGITS so that the vli digit array provides enough elements to fit the larger integers required by this curve. Tested-by:
-
Stefan Berger authored
Add the number of bits a curve has to the ecc_curve definition to be able to derive the number of bytes a curve requires for its coordinates from it. It also allows one to identify a curve by its particular size. Set the number of bits on all curve definitions. Tested-by:
Vitaly Chikunov <vt@altlinux.org> Signed-off-by:
-
Stefan Berger authored
res.x has been calculated by ecc_point_mult_shamir, which uses 'mod curve_prime' on res.x and therefore p > res.x with 'p' being the curve_prime. Further, it is true that for the NIST curves p > n with 'n' being the 'curve_order' and therefore the following may be true as well: p > res.x >= n. If res.x >= n then res.x mod n can be calculated by iteratively sub- tracting n from res.x until res.x < n. For NIST P192/256/384 this can be done in a single subtraction. This can also be done in a single subtraction for NIST P521. The mathematical reason why a single subtraction is sufficient is due to the values of 'p' and 'n' of the NIST curves where the following holds true: note: max(res.x) = p - 1 max(res.x) - n < n p - 1 - n < n p - 1 < 2n => holds true for the NIST curves Tested-by: -
Stefan Berger authored
In preparation for support of NIST P521, adjust the basic tests on the length of the provided key parameters to only ensure that the length of the x plus y coordinates parameter array is not an odd number and that each coordinate fits into an array of 'ndigits' digits. Mathematical tests on the key's parameters are then done in ecc_is_pubkey_valid_full rejecting invalid keys. The change is necessary since NIST P521 keys do not have keys with coordinates that each require 'full' digits (= all bits in u64 used). NIST P521 only requires 2 bytes (9 bits) in the most significant digit unlike NIST P192/256/384 that each require multiple 'full' digits. Tested-by:
-
Stefan Berger authored
For NIST P192/256/384 the public key's x and y parameters could be copied directly from a given array since both parameters filled 'ndigits' of digits (a 'digit' is a u64). For support of NIST P521 the key parameters need to have leading zeros prepended to the most significant digit since only 2 bytes of the most significant digit are provided. Therefore, implement ecc_digits_from_bytes to convert a byte array into an array of digits and use this function in ecdsa_set_pub_key where an input byte array needs to be converted into digits. Suggested-by:
-
Stefan Berger authored
Replace hard-coded numbers with ECC_CURVE_NIST_P192/256/384_DIGITS where possible. Tested-by:
-
Akhil R authored
Add support for Tegra Security Engine which can accelerate various crypto algorithms. The Engine has two separate instances within for AES and HASH algorithms respectively. The driver registers two crypto engines - one for AES and another for HASH algorithms and these operate independently and both uses the host1x bus. Additionally, it provides hardware-assisted key protection for up to 15 symmetric keys which it can use for the cipher operations. Signed-off-by:
Akhil R <akhilrajeev@nvidia.com> Signed-off-by:
-
Akhil R authored
Add Tegra Security Engine details to the SID table in host1x driver. These entries are required to be in place to configure the stream ID for SE. Register writes to stream ID registers fail otherwise. Signed-off-by:
Mikko Perttunen <mperttunen@nvidia.com> Signed-off-by:
-
Akhil R authored
Add DT binding document for Tegra Security Engine. The AES and HASH algorithms are handled independently by separate engines within the Security Engine. These engines are registered as two separate crypto engine drivers. Signed-off-by:
Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org> Signed-off-by:
-
Herbert Xu authored
As the old padata code can execute in softirq context, disable softirqs for the new padata_do_mutithreaded code too as otherwise lockdep will get antsy. Reported-by: syzbot+0cb5bb0f4bf9e79db3b3@syzkaller.appspotmail.com Signed-off-by:
Daniel Jordan <daniel.m.jordan@oracle.com> Signed-off-by:
-
Arnd Bergmann authored
When both ACPI and OF are disabled, the dev_vdata variable is unused: drivers/crypto/ccp/sp-platform.c:33:34: error: unused variable 'dev_vdata' [-Werror,-Wunused-const-variable] This is not a useful configuration, and there is not much point in saving a few bytes when only one of the two is enabled, so just remove all these ifdef checks and rely on of_match_node() and acpi_match_device() returning NULL when these subsystems are disabled. Fixes: 6c506343 ("crypto: ccp - Add ACPI support") Signed-off-by:
Arnd Bergmann <arnd@arndb.de> Acked-by:
Tom Lendacky <thomas.lendacky@amd.com> Signed-off-by:
-
Colin Ian King authored
There is a spelling mistake in a dev_err message. Fix it. Signed-off-by:
Colin Ian King <colin.i.king@gmail.com> Acked-by:
Giovanni Cabiddu <giovanni.cabiddu@intel.com> Signed-off-by:
-
Roman Smirnov authored
The for loop will be executed at least once, so i > 0. If the loop is interrupted before i is incremented (e.g., when checking len for NULL), i will not be checked. Found by Linux Verification Center (linuxtesting.org) with Svace. Signed-off-by:
Roman Smirnov <r.smirnov@omp.ru> Reviewed-by:
Sergey Shtylyov <s.shtylyov@omp.ru> Signed-off-by:
-
Roman Smirnov authored
With the current state of the code, ecc_get_curve() cannot return NULL in crypto_ecdh_shared_secret() and ecc_make_pub_key(). This is conditioned by the fact that they are only called from ecdh_compute_value(), which implements the kpp_alg::{generate_public_key,compute_shared_secret}() methods. Also ecdh implements the kpp_alg::init() method, which is called before the other methods and sets ecdh_ctx::curve_id to a valid value. Signed-off-by:
-
- 05 Apr, 2024 13 commits
-
-
Thorsten Blum authored
The PDF is also available via https. Signed-off-by:
Thorsten Blum <thorsten.blum@toblux.com> Signed-off-by:
-
Thorsten Blum authored
s/in// Signed-off-by:
-
Eric Biggers authored
Add an AES-XTS implementation "xts-aes-vaes-avx10_512" for x86_64 CPUs with the VAES, VPCLMULQDQ, and either AVX10/512 or AVX512BW + AVX512VL extensions. This implementation uses zmm registers to operate on four AES blocks at a time. The assembly code is instantiated using a macro so that most of the source code is shared with other implementations. To avoid downclocking on older Intel CPU models, an exclusion list is used to prevent this 512-bit implementation from being used by default on some CPU models. They will use xts-aes-vaes-avx10_256 instead. For now, this exclusion list is simply coded into aesni-intel_glue.c. It may make sense to eventually move it into a more central location. xts-aes-vaes-avx10_512 is slightly faster than xts-aes-vaes-avx10_256 on some current CPUs. E.g., on AMD Zen 4, AES-256-XTS decryption throughput increases by 13% with 4096-byte inputs, or 14% with 512-byte inputs. On Intel Sapphire Rapids, AES-256-XTS decryption throughput increases by 2% with 4096-byte inputs, or 3% with 512-byte inputs. Future CPUs may provide stronger 512-bit support, in which case a larger benefit should be seen. Signed-off-by:
-
Eric Biggers authored
Add an AES-XTS implementation "xts-aes-vaes-avx10_256" for x86_64 CPUs with the VAES, VPCLMULQDQ, and either AVX10/256 or AVX512BW + AVX512VL extensions. This implementation avoids using zmm registers, instead using ymm registers to operate on two AES blocks at a time. The assembly code is instantiated using a macro so that most of the source code is shared with other implementations. This is the optimal implementation on CPUs that support VAES and AVX512 but where the zmm registers should not be used due to downclocking effects, for example Intel's Ice Lake. It should also be the optimal implementation on future CPUs that support AVX10/256 but not AVX10/512. The performance is slightly better than that of xts-aes-vaes-avx2, which uses the same 256-bit vector length, due to factors such as being able to use ymm16-ymm31 to cache the AES round keys, and being able to use the vpternlogd instruction to do XORs more efficiently. For example, on Ice Lake, the throughput of decrypting 4096-byte messages with AES-256-XTS is 6.6% higher with xts-aes-vaes-avx10_256 than with xts-aes-vaes-avx2. While this is a small improvement, it is straightforward to provide this implementation (xts-aes-vaes-avx10_256) as long as we are providing xts-aes-vaes-avx2 and xts-aes-vaes-avx10_512 anyway, due to the way the _aes_xts_crypt macro is structured. Signed-off-by:
-
Eric Biggers authored
Add an AES-XTS implementation "xts-aes-vaes-avx2" for x86_64 CPUs with the VAES, VPCLMULQDQ, and AVX2 extensions, but not AVX512 or AVX10. This implementation uses ymm registers to operate on two AES blocks at a time. The assembly code is instantiated using a macro so that most of the source code is shared with other implementations. This is the optimal implementation on AMD Zen 3. It should also be the optimal implementation on Intel Alder Lake, which similarly supports VAES but not AVX512. Comparing to xts-aes-aesni-avx on Zen 3, xts-aes-vaes-avx2 provides 70% higher AES-256-XTS decryption throughput with 4096-byte messages, or 23% higher with 512-byte messages. A large improvement is also seen with CPUs that do support AVX512 (e.g., 98% higher AES-256-XTS decryption throughput on Ice Lake with 4096-byte messages), though the following patches add AVX512 optimized implementations to get a bit more performance on those CPUs. Signed-off-by:
-
Eric Biggers authored
Add an AES-XTS implementation "xts-aes-aesni-avx" for x86_64 CPUs that have the AES-NI and AVX extensions but not VAES. It's similar to the existing xts-aes-aesni in that uses xmm registers to operate on one AES block at a time. It differs from xts-aes-aesni in the following ways: - It uses the VEX-coded (non-destructive) instructions from AVX. This improves performance slightly. - It incorporates some additional optimizations such as interleaving the tweak computation with AES en/decryption, handling single-page messages more efficiently, and caching the first round key. - It supports only 64-bit (x86_64). - It's generated by an assembly macro that will also be used to generate VAES-based implementations. The performance improvement over xts-aes-aesni varies from small to large, depending on the CPU and other factors such as the size of the messages en/decrypted. For example, the following increases in AES-256-XTS decryption throughput are seen on the following CPUs: | 4096-byte messages | 512-byte messages | ----------------------+--------------------+-------------------+ Intel Skylake | 6% | 31% | Intel Cascade Lake | 4% | 26% | AMD Zen 1 | 61% | 73% | AMD Zen 2 | 36% | 59% | (The above CPUs don't support VAES, so they can't use VAES instead.) While this isn't as large an improvement as what VAES provides, this still seems worthwhile. This implementation is fairly easy to provide based on the assembly macro that's needed for VAES anyway, and it will be the best implementation on a large number of CPUs (very roughly, the CPUs launched by Intel and AMD from 2011 to 2018). This makes the existing xts-aes-aesni *mostly* obsolete. For now, leave it in place to support 32-bit kernels and also CPUs like Intel Westmere that support AES-NI but not AVX. (We could potentially remove it anyway and just rely on the indirect acceleration via ecb-aes-aesni in those cases, but that change will need to be considered separately.) Signed-off-by: -
Eric Biggers authored
Add an assembly file aes-xts-avx-x86_64.S which contains a macro that expands into AES-XTS implementations for x86_64 CPUs that support at least AES-NI and AVX, optionally also taking advantage of VAES, VPCLMULQDQ, and AVX512 or AVX10. This patch doesn't expand the macro at all. Later patches will do so, adding each implementation individually so that the motivation and use case for each individual implementation can be fully presented. The file also provides a function aes_xts_encrypt_iv() which handles the encryption of the IV (tweak), using AES-NI and AVX. Signed-off-by:
-
Eric Biggers authored
Add config symbols AS_VAES and AS_VPCLMULQDQ that expose whether the assembler supports the vector AES and carryless multiplication cryptographic extensions. Reviewed-by:
Ingo Molnar <mingo@kernel.org> Signed-off-by:
-
Joachim Vandersmissen authored
private_key is overwritten with the key parameter passed in by the caller (if present), or alternatively a newly generated private key. However, it is possible that the caller provides a key (or the newly generated key) which is shorter than the previous key. In that scenario, some key material from the previous key would not be overwritten. The easiest solution is to explicitly zeroize the entire private_key array first. Note that this patch slightly changes the behavior of this function: previously, if the ecc_gen_privkey failed, the old private_key would remain. Now, the private_key is always zeroized. This behavior is consistent with the case where params.key is set and ecc_is_key_valid fails. Signed-off-by:
Joachim Vandersmissen <git@jvdsn.com> Signed-off-by:
-
Joel Granados authored
This commit comes at the tail end of a greater effort to remove the empty elements at the end of the ctl_table arrays (sentinels) which will reduce the overall build time size of the kernel and run time memory bloat by ~64 bytes per sentinel (further information Link : https://lore.kernel.org/all/ZO5Yx5JFogGi%2FcBo@bombadil.infradead.org/) Remove sentinel from crypto_sysctl_table Signed-off-by:
Joel Granados <j.granados@samsung.com> Signed-off-by:
-
Thorsten Blum authored
Replace memzero_explicit() and kvfree() with kvfree_sensitive() to fix the following Coccinelle/coccicheck warning reported by kfree_sensitive.cocci: WARNING opportunity for kfree_sensitive/kvfree_sensitive Signed-off-by:
-
Animesh Agarwal authored
Convert the OMAP SoC SHA crypto Module bindings to DT Schema. Signed-off-by:
Animesh Agarwal <animeshagarwal28@gmail.com> Reviewed-by:
Conor Dooley <conor.dooley@microchip.com> Signed-off-by:
-
Gustavo A. R. Silva authored
-Wflex-array-member-not-at-end is coming in GCC-14, and we are getting ready to enable it globally. Use the `__struct_group()` helper to separate the flexible array from the rest of the members in flexible `struct qat_alg_buf_list`, through tagged `struct qat_alg_buf_list_hdr`, and avoid embedding the flexible-array member in the middle of `struct qat_alg_fixed_buf_list`. Also, use `container_of()` whenever we need to retrieve a pointer to the flexible structure. So, with these changes, fix the following warnings: drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] drivers/crypto/intel/qat/qat_common/qat_bl.h:25:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] Link: https://github.com/KSPP/linux/issues/202Signed-off-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Acked-by:
-
- 02 Apr, 2024 3 commits
-
-
Uwe Kleine-König authored
There are considerations to drop platform_driver_probe() as a concept that isn't relevant any more today. It comes with an added complexity that makes many users hold it wrong. (E.g. this driver should have mark the driver struct with __refdata.) Convert the driver to the more usual module_platform_driver(). This fixes a W=1 build warning: WARNING: modpost: drivers/char/hw_random/mxc-rnga: section mismatch in reference: mxc_rnga_driver+0x10 (section: .data) -> mxc_rnga_remove (section: .exit.text) with CONFIG_HW_RANDOM_MXC_RNGA=m. Signed-off-by:
Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Signed-off-by:
-
Chang S. Bae authored
The aesni_set_key() implementation has no error case, yet its prototype specifies to return an error code. Modify the function prototype to return void and adjust the related code. Signed-off-by:
Chang S. Bae <chang.seok.bae@intel.com> Reviewed-by:
-
Chang S. Bae authored
aes_expandkey() already includes an AES key size check. If AES-NI is unusable, invoke the function without the size check. Also, use aes_check_keylen() instead of open code. Signed-off-by:
-
