Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
# When integrated in to a monolithic kernel the spl module must appear
|
|
|
|
|
# first. This ensures its module initialization function is run before
|
|
|
|
|
# any of the other module initialization functions which depend on it.
|
|
|
|
|
|
|
|
|
|
ZFS_MODULE_CFLAGS += -std=gnu99 -Wno-declaration-after-statement
|
2020-06-17 02:46:04 +03:00
|
|
|
ZFS_MODULE_CFLAGS += -Wmissing-prototypes
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
ZFS_MODULE_CFLAGS += @KERNEL_DEBUG_CFLAGS@ @NO_FORMAT_ZERO_LENGTH@
|
|
|
|
|
|
|
|
|
|
ifneq ($(KBUILD_EXTMOD),)
|
|
|
|
|
zfs_include = @abs_top_srcdir@/include
|
2022-03-31 18:20:50 +03:00
|
|
|
icp_include = @abs_srcdir@/icp/include
|
|
|
|
|
zstd_include = @abs_srcdir@/zstd/include
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
ZFS_MODULE_CFLAGS += -include @abs_top_builddir@/zfs_config.h
|
2020-06-20 05:08:59 +03:00
|
|
|
ZFS_MODULE_CFLAGS += -I@abs_top_builddir@/include
|
2022-03-31 18:20:50 +03:00
|
|
|
src = @abs_srcdir@
|
|
|
|
|
obj = @abs_builddir@
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
else
|
|
|
|
|
zfs_include = $(srctree)/include/zfs
|
2022-03-31 18:20:50 +03:00
|
|
|
icp_include = $(srctree)/$(src)/icp/include
|
|
|
|
|
zstd_include = $(srctree)/$(src)/zstd/include
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
ZFS_MODULE_CFLAGS += -include $(zfs_include)/zfs_config.h
|
|
|
|
|
endif
|
|
|
|
|
|
|
|
|
|
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/kernel
|
|
|
|
|
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/spl
|
|
|
|
|
ZFS_MODULE_CFLAGS += -I$(zfs_include)/os/linux/zfs
|
|
|
|
|
ZFS_MODULE_CFLAGS += -I$(zfs_include)
|
|
|
|
|
ZFS_MODULE_CPPFLAGS += -D_KERNEL
|
|
|
|
|
ZFS_MODULE_CPPFLAGS += @KERNEL_DEBUG_CPPFLAGS@
|
|
|
|
|
|
2022-05-11 23:26:55 +03:00
|
|
|
# KASAN enables -Werror=frame-larger-than=1024, which
|
|
|
|
|
# breaks oh so many parts of our build.
|
|
|
|
|
ifeq ($(CONFIG_KASAN),y)
|
|
|
|
|
ZFS_MODULE_CFLAGS += -Wno-error=frame-larger-than=
|
|
|
|
|
endif
|
|
|
|
|
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
ifneq ($(KBUILD_EXTMOD),)
|
|
|
|
|
@CONFIG_QAT_TRUE@ZFS_MODULE_CFLAGS += -I@QAT_SRC@/include
|
|
|
|
|
@CONFIG_QAT_TRUE@KBUILD_EXTRA_SYMBOLS += @QAT_SYMBOLS@
|
|
|
|
|
endif
|
|
|
|
|
|
2022-03-31 18:20:50 +03:00
|
|
|
asflags-y := $(ZFS_MODULE_CFLAGS) $(ZFS_MODULE_CPPFLAGS)
|
|
|
|
|
ccflags-y := $(ZFS_MODULE_CFLAGS) $(ZFS_MODULE_CPPFLAGS)
|
|
|
|
|
|
|
|
|
|
# Suppress unused-value warnings in sparc64 architecture headers
|
|
|
|
|
ccflags-$(CONFIG_SPARC64) += -Wno-unused-value
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
obj-$(CONFIG_ZFS) := spl.o zfs.o
|
|
|
|
|
|
|
|
|
|
SPL_OBJS := \
|
|
|
|
|
spl-atomic.o \
|
|
|
|
|
spl-condvar.o \
|
|
|
|
|
spl-cred.o \
|
|
|
|
|
spl-err.o \
|
|
|
|
|
spl-generic.o \
|
|
|
|
|
spl-kmem-cache.o \
|
|
|
|
|
spl-kmem.o \
|
|
|
|
|
spl-kstat.o \
|
|
|
|
|
spl-proc.o \
|
|
|
|
|
spl-procfs-list.o \
|
|
|
|
|
spl-taskq.o \
|
|
|
|
|
spl-thread.o \
|
|
|
|
|
spl-trace.o \
|
|
|
|
|
spl-tsd.o \
|
|
|
|
|
spl-vmem.o \
|
|
|
|
|
spl-xdr.o \
|
|
|
|
|
spl-zlib.o
|
|
|
|
|
|
|
|
|
|
spl-objs += $(addprefix os/linux/spl/,$(SPL_OBJS))
|
|
|
|
|
|
|
|
|
|
zfs-objs += avl/avl.o
|
|
|
|
|
|
|
|
|
|
ICP_OBJS := \
|
|
|
|
|
algs/aes/aes_impl.o \
|
|
|
|
|
algs/aes/aes_impl_generic.o \
|
|
|
|
|
algs/aes/aes_modes.o \
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
algs/blake3/blake3.o \
|
|
|
|
|
algs/blake3/blake3_generic.o \
|
|
|
|
|
algs/blake3/blake3_impl.o \
|
|
|
|
|
algs/blake3/blake3_x86-64.o \
|
2022-03-31 18:20:50 +03:00
|
|
|
algs/edonr/edonr.o \
|
|
|
|
|
algs/modes/cbc.o \
|
|
|
|
|
algs/modes/ccm.o \
|
|
|
|
|
algs/modes/ctr.o \
|
|
|
|
|
algs/modes/ecb.o \
|
|
|
|
|
algs/modes/gcm.o \
|
|
|
|
|
algs/modes/gcm_generic.o \
|
|
|
|
|
algs/modes/modes.o \
|
|
|
|
|
algs/sha2/sha2.o \
|
|
|
|
|
algs/skein/skein.o \
|
|
|
|
|
algs/skein/skein_block.o \
|
|
|
|
|
algs/skein/skein_iv.o \
|
|
|
|
|
api/kcf_cipher.o \
|
|
|
|
|
api/kcf_ctxops.o \
|
|
|
|
|
api/kcf_mac.o \
|
|
|
|
|
core/kcf_callprov.o \
|
|
|
|
|
core/kcf_mech_tabs.o \
|
|
|
|
|
core/kcf_prov_lib.o \
|
|
|
|
|
core/kcf_prov_tabs.o \
|
|
|
|
|
core/kcf_sched.o \
|
|
|
|
|
illumos-crypto.o \
|
|
|
|
|
io/aes.o \
|
|
|
|
|
io/sha2_mod.o \
|
|
|
|
|
io/skein_mod.o \
|
|
|
|
|
spi/kcf_spi.o
|
|
|
|
|
|
|
|
|
|
ICP_OBJS_X86_64 := \
|
|
|
|
|
asm-x86_64/aes/aes_aesni.o \
|
|
|
|
|
asm-x86_64/aes/aes_amd64.o \
|
|
|
|
|
asm-x86_64/aes/aeskey.o \
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
asm-x86_64/blake3/blake3_avx2.o \
|
|
|
|
|
asm-x86_64/blake3/blake3_avx512.o \
|
|
|
|
|
asm-x86_64/blake3/blake3_sse2.o \
|
|
|
|
|
asm-x86_64/blake3/blake3_sse41.o \
|
2022-03-31 18:20:50 +03:00
|
|
|
asm-x86_64/modes/aesni-gcm-x86_64.o \
|
|
|
|
|
asm-x86_64/modes/gcm_pclmulqdq.o \
|
|
|
|
|
asm-x86_64/modes/ghash-x86_64.o \
|
|
|
|
|
asm-x86_64/sha2/sha256_impl.o \
|
|
|
|
|
asm-x86_64/sha2/sha512_impl.o
|
|
|
|
|
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
|
2022-03-31 18:20:50 +03:00
|
|
|
ICP_OBJS_X86 := \
|
|
|
|
|
algs/aes/aes_impl_aesni.o \
|
|
|
|
|
algs/aes/aes_impl_x86-64.o \
|
|
|
|
|
algs/modes/gcm_pclmulqdq.o
|
|
|
|
|
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
|
|
|
|
|
ICP_OBJS_ARM64 := \
|
|
|
|
|
asm-aarch64/blake3/b3_aarch64_sse2.o \
|
|
|
|
|
asm-aarch64/blake3/b3_aarch64_sse41.o
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ICP_OBJS_PPC_PPC64 := \
|
|
|
|
|
asm-ppc64/blake3/b3_ppc64le_sse2.o \
|
|
|
|
|
asm-ppc64/blake3/b3_ppc64le_sse41.o
|
|
|
|
|
|
2022-03-31 18:20:50 +03:00
|
|
|
zfs-objs += $(addprefix icp/,$(ICP_OBJS))
|
|
|
|
|
zfs-$(CONFIG_X86) += $(addprefix icp/,$(ICP_OBJS_X86))
|
|
|
|
|
zfs-$(CONFIG_X86_64) += $(addprefix icp/,$(ICP_OBJS_X86_64))
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
zfs-$(CONFIG_ARM64) += $(addprefix icp/,$(ICP_OBJS_ARM64))
|
|
|
|
|
zfs-$(CONFIG_PPC) += $(addprefix icp/,$(ICP_OBJS_PPC_PPC64))
|
|
|
|
|
zfs-$(CONFIG_PPC64) += $(addprefix icp/,$(ICP_OBJS_PPC_PPC64))
|
|
|
|
|
|
|
|
|
|
$(addprefix $(obj)/icp/,$(ICP_OBJS) $(ICP_OBJS_X86) $(ICP_OBJS_X86_64) \
|
|
|
|
|
$(ICP_OBJS_ARM64) $(ICP_OBJS_PPC_PPC64)) : asflags-y += -I$(icp_include)
|
2022-03-31 18:20:50 +03:00
|
|
|
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
$(addprefix $(obj)/icp/,$(ICP_OBJS) $(ICP_OBJS_X86) $(ICP_OBJS_X86_64) \
|
|
|
|
|
$(ICP_OBJS_ARM64) $(ICP_OBJS_PPC_PPC64)) : ccflags-y += -I$(icp_include)
|
2022-03-31 18:20:50 +03:00
|
|
|
|
|
|
|
|
# Suppress objtool "can't find jump dest instruction at" warnings. They
|
|
|
|
|
# are caused by the constants which are defined in the text section of the
|
|
|
|
|
# assembly file using .byte instructions (e.g. bswap_mask). The objtool
|
|
|
|
|
# utility tries to interpret them as opcodes and obviously fails doing so.
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_aesni-gcm-x86_64.o := y
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_ghash-x86_64.o := y
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
|
2022-03-31 18:20:50 +03:00
|
|
|
# Suppress objtool "unsupported stack pointer realignment" warnings. We are
|
|
|
|
|
# not using a DRAP register while aligning the stack to a 64 byte boundary.
|
|
|
|
|
# See #6950 for the reasoning.
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_sha256_impl.o := y
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_sha512_impl.o := y
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
LUA_OBJS := \
|
|
|
|
|
lapi.o \
|
|
|
|
|
lauxlib.o \
|
|
|
|
|
lbaselib.o \
|
|
|
|
|
lcode.o \
|
|
|
|
|
lcompat.o \
|
|
|
|
|
lcorolib.o \
|
|
|
|
|
lctype.o \
|
|
|
|
|
ldebug.o \
|
|
|
|
|
ldo.o \
|
|
|
|
|
lfunc.o \
|
|
|
|
|
lgc.o \
|
|
|
|
|
llex.o \
|
|
|
|
|
lmem.o \
|
|
|
|
|
lobject.o \
|
|
|
|
|
lopcodes.o \
|
|
|
|
|
lparser.o \
|
|
|
|
|
lstate.o \
|
|
|
|
|
lstring.o \
|
|
|
|
|
lstrlib.o \
|
|
|
|
|
ltable.o \
|
|
|
|
|
ltablib.o \
|
|
|
|
|
ltm.o \
|
|
|
|
|
lvm.o \
|
|
|
|
|
lzio.o \
|
|
|
|
|
setjmp/setjmp.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix lua/,$(LUA_OBJS))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
NVPAIR_OBJS := \
|
|
|
|
|
fnvpair.o \
|
|
|
|
|
nvpair.o \
|
|
|
|
|
nvpair_alloc_fixed.o \
|
|
|
|
|
nvpair_alloc_spl.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix nvpair/,$(NVPAIR_OBJS))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
UNICODE_OBJS := \
|
|
|
|
|
u8_textprep.o \
|
|
|
|
|
uconv.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix unicode/,$(UNICODE_OBJS))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ZCOMMON_OBJS := \
|
|
|
|
|
cityhash.o \
|
|
|
|
|
zfeature_common.o \
|
|
|
|
|
zfs_comutil.o \
|
|
|
|
|
zfs_deleg.o \
|
|
|
|
|
zfs_fletcher.o \
|
|
|
|
|
zfs_fletcher_superscalar.o \
|
|
|
|
|
zfs_fletcher_superscalar4.o \
|
|
|
|
|
zfs_namecheck.o \
|
|
|
|
|
zfs_prop.o \
|
|
|
|
|
zpool_prop.o \
|
|
|
|
|
zprop_common.o
|
|
|
|
|
|
|
|
|
|
ZCOMMON_OBJS_X86 := \
|
|
|
|
|
zfs_fletcher_avx512.o \
|
|
|
|
|
zfs_fletcher_intel.o \
|
|
|
|
|
zfs_fletcher_sse.o
|
|
|
|
|
|
|
|
|
|
ZCOMMON_OBJS_ARM64 := \
|
|
|
|
|
zfs_fletcher_aarch64_neon.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix zcommon/,$(ZCOMMON_OBJS))
|
|
|
|
|
zfs-$(CONFIG_X86) += $(addprefix zcommon/,$(ZCOMMON_OBJS_X86))
|
|
|
|
|
zfs-$(CONFIG_ARM64) += $(addprefix zcommon/,$(ZCOMMON_OBJS_ARM64))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Zstd uses -O3 by default, so we should follow
|
|
|
|
|
ZFS_ZSTD_FLAGS := -O3
|
|
|
|
|
|
|
|
|
|
# -fno-tree-vectorize gets set for gcc in zstd/common/compiler.h
|
|
|
|
|
# Set it for other compilers, too.
|
|
|
|
|
ZFS_ZSTD_FLAGS += -fno-tree-vectorize
|
|
|
|
|
|
|
|
|
|
# SSE register return with SSE disabled if -march=znverX is passed
|
|
|
|
|
ZFS_ZSTD_FLAGS += -U__BMI__
|
|
|
|
|
|
|
|
|
|
# Quiet warnings about frame size due to unused code in unmodified zstd lib
|
|
|
|
|
ZFS_ZSTD_FLAGS += -Wframe-larger-than=20480
|
|
|
|
|
|
|
|
|
|
ZSTD_OBJS := \
|
|
|
|
|
zfs_zstd.o \
|
|
|
|
|
zstd_sparc.o
|
|
|
|
|
|
|
|
|
|
ZSTD_UPSTREAM_OBJS := \
|
|
|
|
|
lib/common/entropy_common.o \
|
|
|
|
|
lib/common/error_private.o \
|
|
|
|
|
lib/common/fse_decompress.o \
|
|
|
|
|
lib/common/pool.o \
|
|
|
|
|
lib/common/zstd_common.o \
|
|
|
|
|
lib/compress/fse_compress.o \
|
|
|
|
|
lib/compress/hist.o \
|
|
|
|
|
lib/compress/huf_compress.o \
|
|
|
|
|
lib/compress/zstd_compress.o \
|
|
|
|
|
lib/compress/zstd_compress_literals.o \
|
|
|
|
|
lib/compress/zstd_compress_sequences.o \
|
|
|
|
|
lib/compress/zstd_compress_superblock.o \
|
|
|
|
|
lib/compress/zstd_double_fast.o \
|
|
|
|
|
lib/compress/zstd_fast.o \
|
|
|
|
|
lib/compress/zstd_lazy.o \
|
|
|
|
|
lib/compress/zstd_ldm.o \
|
|
|
|
|
lib/compress/zstd_opt.o \
|
|
|
|
|
lib/decompress/huf_decompress.o \
|
|
|
|
|
lib/decompress/zstd_ddict.o \
|
|
|
|
|
lib/decompress/zstd_decompress.o \
|
|
|
|
|
lib/decompress/zstd_decompress_block.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS))
|
|
|
|
|
|
|
|
|
|
# Disable aarch64 neon SIMD instructions for kernel mode
|
|
|
|
|
$(addprefix $(obj)/zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS)) : ccflags-y += -I$(zstd_include) $(ZFS_ZSTD_FLAGS)
|
|
|
|
|
$(addprefix $(obj)/zstd/,$(ZSTD_OBJS) $(ZSTD_UPSTREAM_OBJS)) : asflags-y += -I$(zstd_include)
|
|
|
|
|
$(addprefix $(obj)/zstd/,$(ZSTD_UPSTREAM_OBJS)) : ccflags-y += -include $(zstd_include)/aarch64_compat.h -include $(zstd_include)/zstd_compat_wrapper.h -Wp,-w
|
|
|
|
|
$(obj)/zstd/zfs_zstd.o : ccflags-y += -include $(zstd_include)/zstd_compat_wrapper.h
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ZFS_OBJS := \
|
|
|
|
|
abd.o \
|
|
|
|
|
aggsum.o \
|
|
|
|
|
arc.o \
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
blake3_zfs.o \
|
2022-03-31 18:20:50 +03:00
|
|
|
blkptr.o \
|
|
|
|
|
bplist.o \
|
|
|
|
|
bpobj.o \
|
|
|
|
|
bptree.o \
|
|
|
|
|
bqueue.o \
|
|
|
|
|
btree.o \
|
|
|
|
|
dataset_kstats.o \
|
|
|
|
|
dbuf.o \
|
|
|
|
|
dbuf_stats.o \
|
|
|
|
|
ddt.o \
|
|
|
|
|
ddt_zap.o \
|
|
|
|
|
dmu.o \
|
|
|
|
|
dmu_diff.o \
|
|
|
|
|
dmu_object.o \
|
|
|
|
|
dmu_objset.o \
|
|
|
|
|
dmu_recv.o \
|
|
|
|
|
dmu_redact.o \
|
|
|
|
|
dmu_send.o \
|
|
|
|
|
dmu_traverse.o \
|
|
|
|
|
dmu_tx.o \
|
|
|
|
|
dmu_zfetch.o \
|
|
|
|
|
dnode.o \
|
|
|
|
|
dnode_sync.o \
|
|
|
|
|
dsl_bookmark.o \
|
|
|
|
|
dsl_crypt.o \
|
|
|
|
|
dsl_dataset.o \
|
|
|
|
|
dsl_deadlist.o \
|
|
|
|
|
dsl_deleg.o \
|
|
|
|
|
dsl_destroy.o \
|
|
|
|
|
dsl_dir.o \
|
|
|
|
|
dsl_pool.o \
|
|
|
|
|
dsl_prop.o \
|
|
|
|
|
dsl_scan.o \
|
|
|
|
|
dsl_synctask.o \
|
|
|
|
|
dsl_userhold.o \
|
|
|
|
|
edonr_zfs.o \
|
|
|
|
|
fm.o \
|
|
|
|
|
gzip.o \
|
|
|
|
|
hkdf.o \
|
|
|
|
|
lz4.o \
|
|
|
|
|
lz4_zfs.o \
|
|
|
|
|
lzjb.o \
|
|
|
|
|
metaslab.o \
|
|
|
|
|
mmp.o \
|
|
|
|
|
multilist.o \
|
|
|
|
|
objlist.o \
|
|
|
|
|
pathname.o \
|
|
|
|
|
range_tree.o \
|
|
|
|
|
refcount.o \
|
|
|
|
|
rrwlock.o \
|
|
|
|
|
sa.o \
|
|
|
|
|
sha256.o \
|
|
|
|
|
skein_zfs.o \
|
|
|
|
|
spa.o \
|
|
|
|
|
spa_boot.o \
|
|
|
|
|
spa_checkpoint.o \
|
|
|
|
|
spa_config.o \
|
|
|
|
|
spa_errlog.o \
|
|
|
|
|
spa_history.o \
|
|
|
|
|
spa_log_spacemap.o \
|
|
|
|
|
spa_misc.o \
|
|
|
|
|
spa_stats.o \
|
|
|
|
|
space_map.o \
|
|
|
|
|
space_reftree.o \
|
|
|
|
|
txg.o \
|
|
|
|
|
uberblock.o \
|
|
|
|
|
unique.o \
|
|
|
|
|
vdev.o \
|
|
|
|
|
vdev_cache.o \
|
|
|
|
|
vdev_draid.o \
|
|
|
|
|
vdev_draid_rand.o \
|
|
|
|
|
vdev_indirect.o \
|
|
|
|
|
vdev_indirect_births.o \
|
|
|
|
|
vdev_indirect_mapping.o \
|
|
|
|
|
vdev_initialize.o \
|
|
|
|
|
vdev_label.o \
|
|
|
|
|
vdev_mirror.o \
|
|
|
|
|
vdev_missing.o \
|
|
|
|
|
vdev_queue.o \
|
|
|
|
|
vdev_raidz.o \
|
|
|
|
|
vdev_raidz_math.o \
|
|
|
|
|
vdev_raidz_math_scalar.o \
|
|
|
|
|
vdev_rebuild.o \
|
|
|
|
|
vdev_removal.o \
|
|
|
|
|
vdev_root.o \
|
|
|
|
|
vdev_trim.o \
|
|
|
|
|
zap.o \
|
|
|
|
|
zap_leaf.o \
|
|
|
|
|
zap_micro.o \
|
|
|
|
|
zcp.o \
|
|
|
|
|
zcp_get.o \
|
|
|
|
|
zcp_global.o \
|
|
|
|
|
zcp_iter.o \
|
|
|
|
|
zcp_set.o \
|
|
|
|
|
zcp_synctask.o \
|
|
|
|
|
zfeature.o \
|
|
|
|
|
zfs_byteswap.o \
|
Introduce BLAKE3 checksums as an OpenZFS feature
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes #10058
Closes #12918
2022-06-09 01:55:57 +03:00
|
|
|
zfs_chksum.o \
|
2022-03-31 18:20:50 +03:00
|
|
|
zfs_fm.o \
|
|
|
|
|
zfs_fuid.o \
|
|
|
|
|
zfs_ioctl.o \
|
|
|
|
|
zfs_log.o \
|
|
|
|
|
zfs_onexit.o \
|
|
|
|
|
zfs_quota.o \
|
|
|
|
|
zfs_ratelimit.o \
|
|
|
|
|
zfs_replay.o \
|
|
|
|
|
zfs_rlock.o \
|
|
|
|
|
zfs_sa.o \
|
|
|
|
|
zfs_vnops.o \
|
|
|
|
|
zil.o \
|
|
|
|
|
zio.o \
|
|
|
|
|
zio_checksum.o \
|
|
|
|
|
zio_compress.o \
|
|
|
|
|
zio_inject.o \
|
|
|
|
|
zle.o \
|
|
|
|
|
zrlock.o \
|
|
|
|
|
zthr.o \
|
|
|
|
|
zvol.o
|
|
|
|
|
|
|
|
|
|
ZFS_OBJS_OS := \
|
|
|
|
|
abd_os.o \
|
|
|
|
|
arc_os.o \
|
|
|
|
|
mmp_os.o \
|
|
|
|
|
policy.o \
|
|
|
|
|
qat.o \
|
|
|
|
|
qat_compress.o \
|
|
|
|
|
qat_crypt.o \
|
|
|
|
|
spa_misc_os.o \
|
|
|
|
|
trace.o \
|
|
|
|
|
vdev_disk.o \
|
|
|
|
|
vdev_file.o \
|
|
|
|
|
zfs_acl.o \
|
|
|
|
|
zfs_ctldir.o \
|
|
|
|
|
zfs_debug.o \
|
|
|
|
|
zfs_dir.o \
|
|
|
|
|
zfs_file_os.o \
|
|
|
|
|
zfs_ioctl_os.o \
|
|
|
|
|
zfs_racct.o \
|
|
|
|
|
zfs_sysfs.o \
|
|
|
|
|
zfs_uio.o \
|
|
|
|
|
zfs_vfsops.o \
|
|
|
|
|
zfs_vnops_os.o \
|
|
|
|
|
zfs_znode.o \
|
|
|
|
|
zio_crypt.o \
|
|
|
|
|
zpl_ctldir.o \
|
|
|
|
|
zpl_export.o \
|
|
|
|
|
zpl_file.o \
|
|
|
|
|
zpl_inode.o \
|
|
|
|
|
zpl_super.o \
|
|
|
|
|
zpl_xattr.o \
|
|
|
|
|
zvol_os.o
|
|
|
|
|
|
|
|
|
|
ZFS_OBJS_X86 := \
|
|
|
|
|
vdev_raidz_math_avx2.o \
|
|
|
|
|
vdev_raidz_math_avx512bw.o \
|
|
|
|
|
vdev_raidz_math_avx512f.o \
|
|
|
|
|
vdev_raidz_math_sse2.o \
|
|
|
|
|
vdev_raidz_math_ssse3.o
|
|
|
|
|
|
|
|
|
|
ZFS_OBJS_ARM64 := \
|
|
|
|
|
vdev_raidz_math_aarch64_neon.o \
|
|
|
|
|
vdev_raidz_math_aarch64_neonx2.o
|
|
|
|
|
|
|
|
|
|
ZFS_OBJS_PPC_PPC64 := \
|
|
|
|
|
vdev_raidz_math_powerpc_altivec.o
|
|
|
|
|
|
|
|
|
|
zfs-objs += $(addprefix zfs/,$(ZFS_OBJS)) $(addprefix os/linux/zfs/,$(ZFS_OBJS_OS))
|
|
|
|
|
zfs-$(CONFIG_X86) += $(addprefix zfs/,$(ZFS_OBJS_X86))
|
|
|
|
|
zfs-$(CONFIG_ARM64) += $(addprefix zfs/,$(ZFS_OBJS_ARM64))
|
|
|
|
|
zfs-$(CONFIG_PPC) += $(addprefix zfs/,$(ZFS_OBJS_PPC_PPC64))
|
|
|
|
|
zfs-$(CONFIG_PPC64) += $(addprefix zfs/,$(ZFS_OBJS_PPC_PPC64))
|
|
|
|
|
|
|
|
|
|
# Suppress incorrect warnings from versions of objtool which are not
|
|
|
|
|
# aware of x86 EVEX prefix instructions used for AVX512.
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_vdev_raidz_math_avx512bw.o := y
|
|
|
|
|
OBJECT_FILES_NON_STANDARD_vdev_raidz_math_avx512f.o := y
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
|
2022-03-31 18:20:50 +03:00
|
|
|
ifeq ($(CONFIG_ALTIVEC),y)
|
|
|
|
|
$(obj)/zfs/vdev_raidz_math_powerpc_altivec.o : c_flags += -maltivec
|
Cleanup linux module kbuild files
The linux module can be built either as an external module, or compiled
into the kernel, using copy-builtin. The source and build directories
are slightly different between the two cases, and currently, compiling
into the kernel still refers to some files from the configured ZFS
source tree, instead of the copies inside the kernel source tree. There
is also duplication between copy-builtin, which creates a Kbuild file to
build ZFS inside the kernel tree, and the top-level module/Makefile.in.
Fix this by moving the list of modules and the CFLAGS settings into a
new module/Kbuild.in, which will be used by the kernel kbuild
infrastructure, and using KBUILD_EXTMOD to distinguish the two cases
within the Makefiles, in order to choose appropriate include
directories etc.
Module CFLAGS setting is simplified by using subdir-ccflags-y (available
since 2.6.30) to set them in the top-level Kbuild instead of each
individual module. The disabling of -Wunused-but-set-variable is removed
from the lua and zfs modules. The variable that the Makefile uses is
actually not defined, so this has no effect; and the warning has long
been disabled by the kernel Makefile itself.
The target_cpu definition in module/{zfs,zcommon} is removed as it was
replaced by use of CONFIG_SPARC64 in
commit 70835c5b755e ("Unify target_cpu handling")
os/linux/{spl,zfs} are removed from obj-m, as they are not modules in
themselves, but are included by the Makefile in the spl and zfs module
directories. The vestigial Makefiles in os and os/linux are removed.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes #10379
Closes #10421
2020-06-08 00:03:12 +03:00
|
|
|
endif
|