| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
BigInt has changed size
|
|
|
|
|
|
|
|
|
|
|
| |
This allows supporting SSE2, NEON and AltiVec in a single codebase,
so drop the NEON and SSE2 code.
This new impl avoids having to do shuffles with every round and so
is about 10% faster on Skylake.
Also, fix bugs in both baseline and AVX2 implementations when the
low counter overflowed. The SSE2 and NEON code were also buggy here.
|
| |
|
|
|
|
|
|
|
| |
Plus adds os info field to disable by default for systems where
pkg-config is not commonly used.
Closes #1268
|
| |
|
| |
|
|
|
|
|
| |
This allows things to mostly work out of the box (#1702), and
allows us to write Emscripten specific code where required.
|
|
|
|
| |
Private report of a successful build on QNX 7 with these fixes.
|
|
|
|
|
|
|
| |
No convention for naming of this variable afaict
Mostly useful as an escape hatch when we for whatever reason do the
wrong thing, as in #1702 when used with emconfigure
|
|
|
|
| |
GH #1702
|
| |
|
|
|
|
|
| |
It seems otherwise VC generates a mix of SSE and AVX code resulting
in lots of transition penalties.
|
| |
|
|
|
|
| |
Disabled by default as it requires a relatively recent kernel and glibc.
|
|
|
|
|
| |
Seems to have been removed from OpenBSD which AFAIK was the only
place it occured.
|
|
|
|
| |
This is what Debian calls little endian 64-bit PPC
|
|
|
|
|
|
|
| |
RISC-V is always little endian by definition.
SPARC is technically bi-endian but basically 100% of userspace is
big endian, so assume it.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 513d19781a558fbd1ff03c7152f61b5e7f294297 removed support for
i586, put it back otherwise the following build failure is raised:
(cd /accts/mlweber1/instance-0/output/build/botan-2.7.0; PATH="/accts/mlweber1/instance-0/output/host/bin:/accts/mlweber1/instance-0/output/host/sbin:/usr/bin:/bin" ./configure.py --cpu="i586" --os=linux --cc=gcc --cc-bin="/accts/mlweber1/instance-0/output/host/bin/i586-linux-g++" --prefix=/usr --disable-static-library --enable-shared-library --without-stack-protector --with-boost --with-bzip2 --with-openssl --with-sqlite --with-lzma --with-zlib --disable-altivec --disable-neon)
INFO: ./configure.py invoked with options "--cpu=i586 --os=linux --cc=gcc --cc-bin=/accts/mlweber1/instance-0/output/host/bin/i586-linux-g++ --prefix=/usr --disable-static-library --enable-shared-library --without-stack-protector --with-boost --with-bzip2 --with-openssl --with-sqlite --with-lzma --with-zlib --disable-altivec --disable-neon"
INFO: Autodetected platform information: OS="Linux" machine="x86_64" proc="x86_64"
ERROR: Unknown or unidentifiable processor "i586"
Fixes:
- http://autobuild.buildroot.org/results/aaa2ea8c3fb5fe954c0af0061f83ad70e0a862f9
Signed-off-by: Fabrice Fontaine <[email protected]>
|
|
|
|
|
|
|
| |
They were only in the same place because of the desire to share
the sbox tables, but that can be handled by adding a dependency.
This makes it possible to disable CAST-256 while leaving CAST-128.
|
|
|
|
|
| |
Downgrade min Sphinx to 1.2 again - I checked and while Sphinx 1.2
produce many warnings it does generate usable output.
|
|\ |
|
| |
| |
| |
| |
| | |
It is the same RNG as arc4random and /dev/urandom. And arc4random seems
to be working well for iOS and macOS.
|
|/ |
|
| |
|
|\ |
|
| | |
|
| | |
|
|/ |
|
| |
|
|
|
|
| |
Easier to remember and type.
|
| |
|
|
|
|
| |
Both #1637 and #1639 change ABI
|
|
|
|
|
|
| |
It seems in practice the same key may be end up used for both
operations, so maintaining a distinction at the type level just
complicates things.
|
| |
|
| |
|
|
|
|
| |
Since RC2 has been removed since 1c0bc3cc6b no reason to have these around.
|
| |
|
|
|
|
| |
See #1624 for background
|
| |
|
|
|
|
| |
Tested in qemu
|
| |
|
| |
|
|
|
|
|
| |
Increases likelyhood that it will fail to build for an end user
which is not helpful for anyone.
|
| |
|
|
|
|
| |
From RFC 8103
|
|
|
|
|
|
|
|
|
| |
This is not a decision we should leave to end users.
And always use a random mask equal in size to the underlying field.
It was never quite clear if 80 bits was sufficient or not. But
taking a random field element is clearly the best possible situation,
and has very little additional cost.
|
|\ |
|
| |
| |
| |
| | |
Speedup of 10 to 30% depending on algo
|
| | |
|