We don't have ifunc support in static arm64 binaries. Until we do
disable the accelerated sha256 code in a static libmd as it uses an
ifunc.
Reported by: brd
Sponsored by: The FreeBSD Foundation
Summary:
When running on a CPU that supports the arm64 sha256 intrinsics use them
to improve perfromance of sha256 calculations.
With this changethe following improvement has been seen on an Apple M1
with FreeBS running under Parallels, with similar results on a
Neoverse-N1 r3p1.
x sha256.orig
+ sha256.arm64
+--------------------------------------------------------------------+
|++ x x|
|+++ xxx|
||A |A||
+--------------------------------------------------------------------+
N Min Max Median Avg Stddev
x 5 3.41 3.5 3.46 3.458 0.042661458
+ 5 0.47 0.54 0.5 0.504 0.027018512
Difference at 95.0% confidence
-2.954 +/- 0.0520768
-85.4251% +/- 0.826831%
(Student's t, pooled s = 0.0357071)
Reviewed by: cem
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D31284
r366344 fixed and reenabled the assembly optimized skein implementation,
but skein_block objects were not being rebuilt in no-clean builds. This
resulted in failing no-clean builds. SKEIN_USE_ASM controls which
routines come from C vs assembly, and with no explicit dependency
r366344's change to SKEIN_USE_ASM did not cause skein_block.{o,pico}
to be rebuilt.
Add a dependency on this Makefile for the skein_block objects. This
dependency is broader in scope than absolutely required (that is, the
skein_block objects will now be rebuilt on any change to this Makefile).
There are ways this could be addressed, but it is probably not worth the
additional effort or testing time to pursue them.
PR: 248221
Reported by: kevans, Jeremy Faulkner
Discussed with: kevans
Sponsored by: The FreeBSD Foundation
The assembly implementation incorrectly used logical AND instead of
bitwise AND. Fix, and re-enable in libmd.
Submitted by: Yang Zhong <yzhong@freebsdfoundation.org>
Reviewed by: cem (earlier)
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D26614
It is apparently broken when assembled by contemporary GNU as as well as
Clang IAS (which is used in the default configuration).
PR: 248221
Reported by: pizzamig
Sponsored by: The FreeBSD Foundation
Comparing the object files produced by GNU as 2.17.50 and Clang IAS
shows many immaterial changes in strtab etc., and one material change
in .text:
1bac: 4c 8b 4f 18 mov 0x18(%rdi),%r9
1bb0: eb 0e jmp 1bc0 <Skein1024_block_loop>
- 1bb2: 66 66 2e 0f 1f 84 00 data16 nopw %cs:0x0(%rax,%rax,1)
- 1bb9: 00 00 00 00
- 1bbd: 0f 1f 00 nopl (%rax)
+ 1bb2: 66 2e 0f 1f 84 00 00 nopw %cs:0x0(%rax,%rax,1)
+ 1bb9: 00 00 00
+ 1bbc: 0f 1f 40 00 nopl 0x0(%rax)
0000000000001bc0 <Skein1024_block_loop>:
Skein1024_block_loop():
1bc0: 4c 8b 47 10 mov 0x10(%rdi),%r8
1bc4: 4c 03 85 c0 00 00 00 add 0xc0(%rbp),%r8
That is, GNU as and Clang's integrated assembler use different multi-
byte NOPs for alignment (GNU as emits an 11 byte NOP + a 3 byte NOP,
while Clang IAS emits a 10 byte NOP + a 4 byte NOP).
Dependency cleanup hacks are not required, because we do not create
.depend files from GNU as.
Reviewed by: allanjude, arichardson, cem, tsoome
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D8434
Update a bunch of Makefile.depend files as
a result of adding Makefile.depend.options files
Reviewed by: bdrewery
MFC after: 1 week
Sponsored by: Juniper Networks
Differential Revision: https://reviews.freebsd.org/D22494
All of them are needed to be able to boot to single user and be able
to repair a existing FreeBSD installation so put them directly into
FreeBSD-runtime.
Reviewed by: bapt, gjb
Differential Revision: https://reviews.freebsd.org/D21503
For users who have a seperate zfs mount of /usr or /usr/lib, this will
cause dynamic loading failures when attempting to execute zfs mount on
bootup. E.g. the system won't boot.
Including <src.opts.mk> sets SHLIBDIR, so SHLIBDIR?= has no
effect. The other lib/ Makefiles solve this problem by moving the
SHLIBDIR assignment to before .include <src.opts.mk>.
Submitted by: jilles
Reviewed by: allanjude
Approved by: re (rgrimes)
Differential Revision: https://reviews.freebsd.org/D16910
This fixes building libmd on MacOS/Linux. The real fix is probably to
build it as a .S file with $CC instead. It might also be better to just
compile the C file in userspace since the compiler can the use SSE/AVX.
Reviewed By: emaste, brooks
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D16844
The assembly files use directives that only work for ELF targets so skip
them when bootstrapping on MacOS.
Reviewed By: imp
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D14247
The double compilation of the kernel sources in libmd and libcrypt is
baffling, but add yet another define hack to prevent duplicate symbols.
Add documentation and SHA2-224 test cases to libmd.
Integrate with the md5(1) command, document, and add more test cases;
self-tests pass.
It causes the 32bit compat build of libmd to fail with:
libmd/rmd160c.c:86:9: error: 'ripemd160_block' macro redefined
#define ripemd160_block ripemd160_block_x86
^
libmd/ripemd.h:122:9: note: previous definition is here
#define ripemd160_block _libmd_ripemd160_block
This patch was inspired by an opposite change made to shrink the code
for the boot loader.
On my i7-4770, it increases the skein1024 speed from 470 to 550 MB/s
Reviewed by: sbruno
MFC after: 1 month
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D7824
This follows another fix to bootstrap libmd after r313404. The
MD5FileChunk prototype is needed to build libmd, but it is
only reliably in the src tree's sys/md5.h header. Rather than
polluting the legacy build with this header for the entire build,
just symlink it in here for now as is done in the elftoolchain
build. Libmd is already referencing other src tree headers by
its used of CFLAGS+= ${SRCTOP}/sys/crypto/sha2. This, and
other uses of CFLAGS+= ${SRCTOP}/sys..., may later change to
be in the legacy mechanism.
Reported by: bde, ian, sjg
Tested by: ian
C99 allows array function parameters to use the static keyword for their
sizes. This tells the compiler that the parameter will have at least the
specified size, and calling code will fail to compile if that guarantee is
not met. However, this syntax is not legal in C++.
This commit reverts r300824, which worked around the problem for
sys/sys/md5.h only, and introduces a new macro: min_size(). min_size(x) can
be used in headers as a static array size, but will still compile in C++
mode.
Reviewed by: cem, ed
MFC after: 4 weeks
Sponsored by: Spectra Logic Corp
Differential Revision: https://reviews.freebsd.org/D8277
Old versions of gas produce an invalid section index. That is ignored by
old versions of ld, but prevents a link with lld.
Submitted by: Rafael Ávila de Espíndola (earlier version)
Reviewed by: allanjude
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D6789
Connect it to userland (libmd, libcrypt, sbin/md5) and kernel (crypto.ko)
Support for skein as a ZFS checksum algorithm was introduced in r289422
but is disconnected because FreeBSD lacked a Skein implementation.
A further commit will enable it in ZFS.
Reviewed by: cem
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D6166
This implements SHA-512/256, which generates a 256 bit hash by
calculating the SHA-512 then truncating the result. A different initial
value is used, making the result different from the first 256 bits of
the SHA-512 of the same input. SHA-512 is ~50% faster than SHA-256 on
64bit platforms, so the result is a faster 256 bit hash.
The main goal of this implementation is to enable support for this
faster hashing algorithm in ZFS. The feature was introduced into ZFS
in r289422, but is disconnected because SHA-512/256 support was missing.
A further commit will enable it in ZFS.
This is the follow on to r292782
Reviewed by: cem
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D6061
These are no longer needed after the recent 'beforebuild: depend' changes
and hooking DIRDEPS_BUILD into a subset of FAST_DEPEND which supports
skipping 'make depend'.
Sponsored by: EMC / Isilon Storage Division
- Remove unneeded fstat()/lseek() calls.
- Return NULL and set errno to EINVAL on negative length.
- Fix small style problems and expand variable names.
After this change, it is possible to use this code for some irregular
files. For example, 'md5 /dev/md0' should now succeed.
Differential Revision: https://reviews.freebsd.org/D4748
Suggested by: bde
Reviewed by: bde, allanjude, delphij
cperciva's libmd implementation is 5-30% faster
The same was done for SHA256 previously in r263218
cperciva's implementation was lacking SHA-384 which I implemented, validated against OpenSSL and the NIST documentation
Extend sbin/md5 to create sha384(1)
Chase dependancies on sys/crypto/sha2/sha2.{c,h} and replace them with sha512{c.c,.h}
Reviewed by: cperciva, des, delphij
Approved by: secteam, bapt (mentor)
MFC after: 2 weeks
Sponsored by: ScaleEngine Inc.
Differential Revision: https://reviews.freebsd.org/D3929
The main problem was bitrot after elftoolchain being swapped in for the
GNU toolchain.
This also reworks how the list of 'host allowed' libraries is determined
to only allow INTERNALLIBs, which is needed for libelftc to come in.
For usr.bin/readelf use the same hack, as libelf and libdward, to bring in
the needed sys/ headers for host builds. This has not yet been a problem due
to readelf not being built as a host tool in buildworld. This is possible
in the meta build though when building the toolchain.
Sponsored by: EMC / Isilon Storage Division
This both avoids some dependencies on xinstall.host and allows
bootstrapping on older releases to work due to lack of at least 'install -l'
support.
Sponsored by: EMC / Isilon Storage Division