aboutsummaryrefslogtreecommitdiffstatshomepage
path: root/src/crypto/zinc (unfollow)
Commit message (Collapse)AuthorFilesLines
2019-12-12global: fix up spellingJosh Soref1-1/+1
Signed-off-by: Josh Soref <jsoref@gmail.com>
2019-12-06chacha20poly1305: double check the sgmiter logic with testJason A. Donenfeld1-8/+59
2019-12-05crypto: use new assembler macros for 5.5Jason A. Donenfeld5-14/+14
2019-12-05chacha20poly1305: port to sgmitter for 5.5Jason A. Donenfeld2-108/+137
I'm not totally comfortable with these changes yet, and it'll require some more scrutiny. But it's a start.
2019-06-03blake2s: spacingJason A. Donenfeld2-123/+123
2019-06-02curve25519: not all linkers support bmi2 and adxJason A. Donenfeld2-6/+48
2019-05-31blake2s: add ssse3 to nobsJason A. Donenfeld1-1/+2
2019-05-31blake2s: do not use xgetbv for ssse3 detectionJason A. Donenfeld1-3/+1
2019-05-29zinc: update copyrightJason A. Donenfeld2-2/+2
2019-05-29blake2s: shorten ssse3 loopSamuel Neves1-857/+66
This (mostly) preserves the performance (as measured on Haswell and *lake) of last commit, but it drastically reduces code size. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2019-05-29blake2s,chacha: latency tweakSamuel Neves5-618/+982
In every odd-numbered round, instead of operating over the state x00 x01 x02 x03 x05 x06 x07 x04 x10 x11 x08 x09 x15 x12 x13 x14 we operate over the rotated state x03 x00 x01 x02 x04 x05 x06 x07 x09 x10 x11 x08 x14 x15 x12 x13 The advantage here is that this requires no changes to the 'x04 x05 x06 x07' row, which is in the critical path. This results in a noticeable latency improvement of roughly R cycles, for R diagonal rounds in the primitive. In the case of BLAKE2s, which I also moved from requiring AVX to only requiring SSSE3, we save approximately 30 cycles per compression function call on Haswell and Skylake. In other words, this is an improvement of ~0.6 cpb. This idea was pointed out to me by Shunsuke Shimizu, though it appears to have been around for longer. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2019-05-29zinc: arm64: use cpu_get_elf_hwcap accessor for 5.2Jason A. Donenfeld2-2/+2
2019-03-27blake2s: remove outlen parameter from finalJason A. Donenfeld1-6/+5
2019-03-27blake2s: simplifySamuel Neves1-38/+10
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2019-02-03noise: store clamped key instead of raw keyJason A. Donenfeld4-14/+7
2019-02-03chacha20poly1305: permit unaligned strides on certain platformsJason A. Donenfeld1-18/+14
The map allocations required to fix this are mostly slower than unaligned paths. Reported-by: Louis Sautier <sbraz@gentoo.org>
2019-01-23global: normalize -> clampJason A. Donenfeld4-17/+10
2019-01-07global: update copyrightJason A. Donenfeld31-31/+31
2018-12-07chacha20: do not define unused asm functionJason A. Donenfeld1-4/+2
This causes RAP to be unhappy, and we're not using it anyway. Reported-by: Ivan J. <parazyd@dyne.org>
2018-12-07chacha20,poly1305: simplify perlasm fancinessJason A. Donenfeld3-75/+69
2018-11-19chacha20,poly1305: do not use xlateJason A. Donenfeld3-1496/+73
2018-11-17poly1305: make frame pointers for auxiliary callsSamuel Neves1-31/+43
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-15chacha20,poly1305: don't do compiler testing in generator and remove xor helperJason A. Donenfeld2-30/+39
2018-11-15poly1305: cleanup leftover debugging changesJason A. Donenfeld1-3/+3
2018-11-15poly1305: only export neon symbols when in useJason A. Donenfeld1-2/+6
2018-11-15chacha20,poly1305: fix up for win64Samuel Neves2-27/+29
These don't help us, but it is important to keep this working for when it's re-added to cryptogams. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-15perlasm: avoid rep retJason A. Donenfeld1-1/+1
The original hardcodes returns as .byte 0xf3,0xc3, aka "rep ret". We replace this by "ret". "rep ret" was meant to help with AMD K8 chips, cf. http://repzret.org/p/repzret. It makes no sense to continue to use this kludge for code that won't even run on ancient AMD chips.
2018-11-15poly1305: specialize to wireguardJason A. Donenfeld1-11/+20
2018-11-15chacha20: specialize to wireguardJason A. Donenfeld2-20/+38
2018-11-15perlasm: cleanup whitespaceJason A. Donenfeld1-5/+5
2018-11-15poly1305: adjust to kernelSamuel Neves1-220/+291
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: cleaner function declarationsSamuel Neves1-23/+23
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: normalize namesSamuel Neves1-71/+71
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: fixup win64 stack offsetsSamuel Neves1-129/+129
We don't need to do this for kernel purposes, but it's polite to leave things unbroken. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: simplify stack unwinding on ChaCha20_ctr32Samuel Neves1-10/+8
objtool did not quite understand the stack arithmetic employed here. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: use DRAP idiomSamuel Neves1-236/+235
This effectively means swapping the usage of %r9 and %r10 globally. Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: add hchacha_ssse3Samuel Neves1-0/+39
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20: begin adapting to kernel settingSamuel Neves2-68/+116
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20,poly1305: switch to perlasm originals on x86_64Samuel Neves5-5424/+9596
Signed-off-by: Samuel Neves <sneves@dei.uc.pt>
2018-11-14chacha20,poly1305: use CONFIG_KERNEL_MODE_NEON in .pl on armJason A. Donenfeld4-8/+11
While Andy is right to desire a separation between compiler defines and project defines, there are simply too many odd kernel configurations and we require testing for CONFIG_KERNEL_MODE_NEON.
2018-11-14chacha20,poly1305: switch to perlasm originals on mips and armJason A. Donenfeld11-6103/+5563
We also separate out Eric Biggers' Cortex A7 implementation into its own file.
2018-11-13global: various formatting tweeksJason A. Donenfeld2-2/+1
2018-10-27curve25519-x86_64: this was relicensed to BSD-3-Clause upstreamJason A. Donenfeld1-1/+1
2018-10-27poly1305-donna64: mark large constants as ULLJason A. Donenfeld1-24/+24
2018-10-07crypto: clean up remaining .h->.cJason A. Donenfeld8-10/+10
2018-10-07crypto: use BIT(i) & bitmap instead of (bitmap >> i) & 1Jason A. Donenfeld1-2/+2
Pros: clearer if you're not familiar with the shift idiom, uses kernel macro. Cons: doesn't work any more if the lvalue ever ceases to be a bool. Neutral: generates the same machine code. Suggested-by: Sultan Alsawaf <sultanxda@gmail.com>
2018-10-07crypto: disable broken implementations in selftestsJason A. Donenfeld1-9/+8
2018-10-06crypto: test all SIMD combinationsJason A. Donenfeld20-40/+82
2018-10-06global: rename include'd C files to be .cJason A. Donenfeld19-28/+28
This is done by 259 other files in the kernel tree: linux $ rg '#include.*\.c' -l | wc -l 259 Suggested-by: Sultan Alsawaf <sultanxda@gmail.com>
2018-10-04curve25519-arm: rearrange multiplications for better in-order performanceJason A. Donenfeld1-12/+12
Suggested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>