Add BLAKE3 hash support and chksum benchmarking to OpenZFS #12918

mcmilk · 2021-12-27T19:06:15Z

This patchset adds another checksum to OpenZFS that has similar performance
to Edon-R, but without the caveats around the latter.

It adds micro benchmarking for all advanced hashing functions like it's
done for fletcher-4.

The statistics are located here: /proc/spl/kstat/zfs/chksum_bench and look like
this example of my test system with debian 11 and this cpu: i3-1005G1:

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

The units are MiB/s for each method.

Signed-off-by: Tino Reichardt milky-zfs@mcmilk.de
Co-authored-by: Rich Ercolani rincebrain@gmail.com

Motivation and Context

Description

How Has This Been Tested?

Types of changes

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Performance enhancement (non-breaking change which improves efficiency)
I have run the ZFS Test Suite with this change applied.
I have added tests to cover my changes.
Code cleanup (non-breaking change which makes code smaller or more readable)
Breaking change (fix or feature that would cause existing functionality to change)
Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
Documentation (a change to man pages or other documentation)

Checklist:

My code follows the OpenZFS code style requirements.
I have updated the documentation accordingly.
I have read the contributing document.
All commit messages are properly formatted and contain Signed-off-by.

mcmilk · 2021-12-27T19:15:02Z

My local tests are building fine on fedore35 (x64) and debian 11 (x64) .

include/sys/Makefile.am

rincebrain · 2021-12-28T05:49:18Z

Tried to load it:

[15450.915597] Insufficient stack space to handle exception!
[15450.915604] ESR: 0x96000047 -- DABT (current EL)
[15450.920441] FAR: 0xffff00001413fff0
[15450.922471] Task stack:     [0xffff000014140000..0xffff000014150000]
[15450.925688] IRQ stack:      [0xffff000010040000..0xffff000010050000]
[15450.928787] Overflow stack: [0xffff80007fc002b0..0xffff80007fc012b0]
[15450.935391] CPU: 2 PID: 382855 Comm: modprobe Kdump: loaded Tainted: P           OE    ---------r-  - 4.18.0-348.2.1.el8_5.aarch64 #1
[15450.941302] Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015
[15450.944636] pstate: 60000005 (nZCv daif -PAN -UAO)
[15450.947204] pc : blake3_compress_in_place_generic+0x0/0x68 [icp]
[15450.950235] lr : blake3_compress_in_place_generic+0x54/0x68 [icp]
[15450.953199] sp : ffff000014140030
[15450.955187] x29: ffff000014140030 x28: 0000000000000040
[15450.957913] x27: 0000000000000400 x26: 0000000000000400
[15450.960575] x25: ffff800028240778 x24: ffff0000091c42b0
[15450.963292] x23: 0000000000000011 x22: 0000000000000000
[15450.966155] x21: 0000000000000040 x20: ffff00001a180000
[15450.968810] x19: ffff800028240798 x18: 00000000ffffffff
[15450.971624] x17: 0000000000000022 x16: 0000000000000000
[15450.974603] x15: 00000000000003bb x14: 0000000000000000
[15450.977314] x13: 0000000000000000 x12: 0000000000000000
[15450.980099] x11: 0000000000000000 x10: 0000000000000000
[15450.985656] x9 : 0000000000000000 x8 : 0000000000000000
[15450.988333] x7 : 0000000000000000 x6 : 0000000000000001
[15450.991563] x5 : ffff0000091bb290 x4 : 0000000000000011
[15450.994483] x3 : 0000000000000000 x2 : 0000000000000040
[15450.997198] x1 : ffff00001a180000 x0 : ffff800028240798
[15450.999983] Kernel panic - not syncing: kernel stack overflow
[15451.002893] CPU: 2 PID: 382855 Comm: modprobe Kdump: loaded Tainted: P           OE    ---------r-  - 4.18.0-348.2.1.el8_5.aarch64 #1
[15451.008834] Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015
[15451.012124] Call trace:
[15451.013879]  dump_backtrace+0x0/0x190
[15451.015981]  show_stack+0x28/0x38
[15451.018069]  dump_stack+0x9c/0xbc
[15451.020020]  panic+0x134/0x2fc
[15451.021967]  nmi_panic+0x84/0x88
[15451.023892]  handle_bad_stack+0xf8/0x110
[15451.026224]  __bad_stack+0x90/0x94
[15451.028301]  blake3_compress_in_place_generic+0x0/0x68 [icp]
[15451.031198]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.034141]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.037020]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.039938]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.042888]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.045863]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.049927]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.052988]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.056078]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.059321]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.062372]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.065253]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.068177]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15451.071105]  blake3_compress_in_place_generic+0x54/0x68 [icp]

[snip]

[15454.276911]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15454.281122]  blake3_compress_in_place_generic+0x54/0x68 [icp]
[15454.284164]  chunk_state_update+0x84/0x148 [icp]
[15454.286867]  Blake3_Update.part.2+0x21c/0x278 [icp]
[15454.289615]  Blake3_Update+0x3c/0x50 [icp]
[15454.292482]  blake3_incremental+0x38/0x50 [zfs]
[15454.295318]  abd_iterate_func.part.1+0x100/0x278 [zfs]
[15454.298501]  abd_iterate_func+0x54/0x68 [zfs]
[15454.301219]  abd_checksum_blake3_native+0x7c/0x140 [zfs]
[15454.304394]  chksum_run.isra.0+0x98/0x110 [zfs]
[15454.307242]  chksum_benchit+0x84/0x1b8 [zfs]
[15454.310978]  chksum_init+0x1ac/0x268 [zfs]
[15454.314257]  spa_init+0x128/0x1c8 [zfs]
[15454.316897]  zfs_kmod_init+0x4c/0x110 [zfs]
[15454.319625]  openzfs_init+0x2c/0x10000 [zfs]
[15454.321981]  do_one_initcall+0x58/0x1f0
[15454.324278]  do_init_module+0x64/0x238
[15454.326593]  load_module+0x1404/0x16a8
[15454.328865]  __se_sys_finit_module+0xa4/0x100
[15454.331437]  __arm64_sys_finit_module+0x28/0x38
[15454.334079]  el0_svc_handler+0xb4/0x188
[15454.336374]  el0_svc+0x8/0xc

So that's new and different.

e: Ah, I see the problem, I think. It might even be my fault, I don't feel like looking at the commit history. Fixing...

rincebrain · 2021-12-28T12:53:43Z

From this branch, running in kvm on a Pi 4:

21 0 0x01 -1 0 31088416609174 31097040395854
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
fletcher-4       4         1944      2352      2568      2734      2709      2700      2732      2771      2743      2743      2562
edonr-generic    256       509       521       608       628       633       637       640       643       636       639       608
skein-generic    256       39        293       309       316       318       315       318       319       318       319       314
sha256-generic   256       113       128       130       129       132       133       133       133       133       133       132
sha512-generic   512       147       177       184       188       189       190       191       191       191       191       189
blake3-generic   256       61        59        58        60        60        59        60        60        59        60        60
blake3-neon      256       62        58        173       180       187       189       191       190       190       188       166

mcmilk · 2021-12-28T13:30:28Z

From this branch, running in kvm on a Pi 4:
...

Thank you for the benchmarks on Pi 4 ... I will integrate your fixes for arm.

rincebrain · 2021-12-29T04:46:53Z

Just as a followup, on a baremetal Pi4 (all of these were/are with --enable-debug, to be clear):

21 0 0x01 -1 0 5697937181252 5698490351399
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
fletcher-4       4         1974      2325      2588      2731      2806      2731      2746      2611      2634      2627      2543
edonr-generic    256       510       571       592       626       637       637       639       642       643       644       639
skein-generic    256       274       291       295       303       305       307       308       309       309       309       308
sha256-generic   256       122       128       132       133       134       134       135       135       135       135       135
sha512-generic   512       163       178       182       189       190       191       192       192       192       192       192
blake3-generic   256       91        93        93        93        93        93        93        93        93        93        93
blake3-neon      256       93        90        182       184       187       189       191       191       191       192       187

For bonus fun, in a VM on my Ryzen 5900x:

21 0 0x01 -1 0 945696798991138 945699672232315
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
fletcher-4       4         3932      7112      11824     18053     24549     28667     32892     35115     36338     36156     36562
edonr-generic    256       1775      2017      2206      2317      2376      2379      2398      2429      2424      2418      2398
skein-generic    256       786       837       856       878       885       883       887       886       889       886       885
sha256-generic   256       368       384       387       399       398       397       400       401       400       400       401
sha512-generic   512       532       579       601       624       605       627       631       632       631       632       629
blake3-generic   256       271       277       283       279       281       283       283       283       283       283       283
blake3-sse2      256       314       1138      1660      1806      1872      1927      1897      1956      1954      1938      1935
blake3-sse41     256       324       1249      1830      1991      2116      2185      2168      2209      2220      2211      2214
blake3-avx2      256       327       1388      2231      3331      3782      4067      4290      4386      4406      4382      4386

And because I could, on my little sparc64 Netra T1:

21 0 0x01 -1 0 12665540404381 12682821164863
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
fletcher-4       4         336       430       490       525       545       247       260       267       271       233       205
edonr-generic    256       45        52        56        58        59        57        57        57        57        55        53
skein-generic    256       19        21        21        22        22        21        21        21        21        21        21
sha256-generic   256       9         10        10        11        11        11        11        11        11        11        11
sha512-generic   512       5         6         6         6         6         6         6         6         6         6         6
blake3-generic   256       1         2         3         6         5         4         7         7         7         7         7

e: As evidence that "premature optimization is evil, but occasionally useful", here's the numbers if I remake blake3_neon.S with -mcpu=cortex-a72 (which is what's in the Pi 4 specifically) (and the numbers from the original below for contrast):

blake3-neon      256       89        186       293       309       319       321       319       322       322       322       310

blake3-generic   256       91        93        93        93        93        93        93        93        93        93        93
blake3-neon      256       93        90        182       184       187       189       191       191       191       192       187

mcmilk · 2021-12-29T17:11:10Z

Benchmark with notebook cpu i3-1005G1:

20 0 0x01 -1 0 369640498761 381345412977                                                                                                                                                                 
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m                                                                        
fletcher-4       4         4809      8517      13614     19884     25669     29604     31464     33402     34112     33982     34060                                                                     
edonr-generic    256       1260      1488      1622      1714      1763      1781      1763      1762      1772      1768      1770                                                                      
skein-generic    256       545       582       602       609       613       618       623       620       620       620       621                                                                       
sha256-generic   256       285       302       310       316       318       320       318       319       319       319       319                                                                       
sha512-generic   512       393       439       462       471       479       481       483       484       484       485       484                                                                       
blake3-generic   256       300       298       303       304       296       298       299       300       300       301       300                                                                       
blake3-sse2      256       390       932       1308      1388      1429      1453      1464      1472      1467      1473      1471                                                                      
blake3-sse41     256       414       979       1488      1589      1648      1682      1700      1713      1715      1717      1717                                                                      
blake3-avx2      256       412       1304      1908      2871      3133      3278      3356      3398      3400      3416      3406                                                                      
blake3-avx512    256       451       1506      2667      3439      5043      5575      5860      6032      6118      5850      6079

rincebrain · 2021-12-29T17:34:35Z

The numbers from the SHA256/512 SIMD/SHA-NI PR are also interesting to compare against.

mcmilk · 2021-12-30T08:15:33Z

The numbers from the SHA256/512 SIMD/SHA-NI PR are also interesting to compare against.

Again with the i3-1005G1 - but sha2 with assembler:

implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
edonr-generic    256       1301      1504      1635      1730      1773      1783      1766      1759      1761      1766      1768
skein-generic    256       539       580       605       601       615       618       617       618       620       619       620
sha256-sha_ni    256       1056      1140      1188      1207      1220      1230      1238      1240      1239      1238      1236
sha512-avx2      512       459       504       522       537       544       543       547       547       550       550       550
blake3-generic   256       295       295       295       292       296       300       298       299       299       298       299
blake3-sse2      256       387       935       1308      1375      1421      1450      1461      1468      1471      1473      1472
blake3-sse41     256       416       976       1479      1578      1642      1676      1700      1709      1715      1721      1720
blake3-avx2      256       416       1318      1905      2843      3105      3258      3345      3390      3362      3408      3400
blake3-avx512    256       451       1505      2642      3403      4969      5519      5821      6014      6113      6135      6091

rincebrain · 2021-12-31T00:35:23Z

Isn't that cute...

From a ppc64le Debian box, running 077b0c6:

20 0 0x01 -1 0 1214651156219429 1214657651329681
implementation   digest    1k        2k        4k        8k        16k       32k       64k       128k      256k      512k      1m
edonr-generic    256       1361      1619      1702      1785      1734      1749      1754      1824      1977      1971      1983
skein-generic    256       455       485       511       512       520       527       530       530       529       529       530
sha256-generic   256       239       257       264       267       268       266       270       270       270       270       268
sha512-generic   512       340       372       386       400       404       406       408       408       409       409       407
blake3-generic   256       154       162       163       164       165       165       165       165       165       165       165
(null)-(null)    0         0         0         0         0         0         0         0         0         0         0         0

Guessing this is just a side effect of
chksum_stat_cnt = 5 + blake3_get_impl_count();
since it goes away when I uncomment the fletcher4 section and add the appropriate _init and _fini calls.

I find this patch useful for stopping it from panicking if you uncomment fletcher4 and add _init() and _fini() calls, it seems to happen because fletcher_4_init doesn't handle a second call of it well.

diff --git a/module/zcommon/zfs_fletcher.c b/module/zcommon/zfs_fletcher.c
index e7a8553f3..54d4cc3c2 100644
--- a/module/zcommon/zfs_fletcher.c
+++ b/module/zcommon/zfs_fletcher.c
@@ -771,26 +771,30 @@ fletcher_4_benchmark(void)
 void
 fletcher_4_init(void)
 {
-       /* Determine the fastest available implementation. */
-       fletcher_4_benchmark();
+       if (!fletcher_4_initialized) {
+               /* Determine the fastest available implementation. */
+               fletcher_4_benchmark();

 #if defined(_KERNEL)
-       /* Install kstats for all implementations */
-       fletcher_4_kstat = kstat_create("zfs", 0, "fletcher_4_bench", "misc",
-           KSTAT_TYPE_RAW, 0, KSTAT_FLAG_VIRTUAL);
-       if (fletcher_4_kstat != NULL) {
-               fletcher_4_kstat->ks_data = NULL;
-               fletcher_4_kstat->ks_ndata = UINT32_MAX;
-               kstat_set_raw_ops(fletcher_4_kstat,
-                   fletcher_4_kstat_headers,
-                   fletcher_4_kstat_data,
-                   fletcher_4_kstat_addr);
-               kstat_install(fletcher_4_kstat);
-       }
+               /* Install kstats for all implementations */
+               if (!fletcher_4_kstat) {
+                       fletcher_4_kstat = kstat_create("zfs", 0, "fletcher_4_bench", "misc",
+                           KSTAT_TYPE_RAW, 0, KSTAT_FLAG_VIRTUAL);
+                       if (fletcher_4_kstat != NULL) {
+                               fletcher_4_kstat->ks_data = NULL;
+                               fletcher_4_kstat->ks_ndata = UINT32_MAX;
+                               kstat_set_raw_ops(fletcher_4_kstat,
+                                   fletcher_4_kstat_headers,
+                                   fletcher_4_kstat_data,
+                                   fletcher_4_kstat_addr);
+                               kstat_install(fletcher_4_kstat);
+                       }
+               }
 #endif

-       /* Finish initialization */
-       fletcher_4_initialized = B_TRUE;
+               /* Finish initialization */
+               fletcher_4_initialized = B_TRUE;
+       }
 }

mcmilk · 2021-12-31T09:42:42Z

Yes, a second call of fletcher_4_init() will trigger an assert...

I am thinking about reworking all the hashing/chksumming stuff into one place.... currently FreeBSD does sth. wired with the HAVE_SSE4* HAVE_AVX* defines. They are not set correctly, but the toolchain finds them. Maybe it's the Makefile.bsd stuff problem again :-/ I will find it out and put together an even bigger patch set.

rincebrain · 2021-12-31T12:35:59Z

You could break the checksum/hash reworking into a separate patchset and then make the BLAKE3 patchset depend on it, perhaps, if that suits your fancy?

(I'm also just generally in favor of more, smaller PRs rather than larger ones doing multiple things at once, where possible.)

module/icp/algs/blake3/blake3.c

module/icp/algs/blake3/blake3_impl.c

module/icp/asm-aarch64/blake3/blake3_neon.S

tests/zfs-tests/tests/functional/checksum/blake3_test.c

mcmilk · 2022-01-06T05:24:21Z

I just rebased again. As mentioned by Ahelenia Ziemiańska, the blank first lines are fixed and the the blake3_impl[] array is const now.

From my side, I think the PR is ready for inclusion now.

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 openzfs#10058 Closes openzfs#12918

New features: - Fully adaptive ARC eviction (#14359) - Block cloning (#13392) - Scrub error log (#12812, #12355) - Linux container support (#14070, #14097, #12263) - BLAKE3 Checksums (#12918) - Corrective "zfs receive" (#9372) Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

New Features - Block cloning (#13392) - Linux container support (#14070, #14097, #12263) - Scrub error log (#12812, #12355) - BLAKE3 checksums (#12918) - Corrective "zfs receive" - Vdev and zpool user properties Performance - Fully adaptive ARC (#14359) - SHA2 checksums (#13741) - Edon-R checksums (#13618) - Zstd early abort (#13244) - Prefetch improvements (#14603, #14516, #14402, #14243, #13452) - General optimization (#14121, #14123, #14039, #13680, #13613, #13606, #13576, #13553, #12789, #14925, #14948) Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

New features: - Fully adaptive ARC eviction (openzfs#14359) - Block cloning (openzfs#13392) - Scrub error log (openzfs#12812, openzfs#12355) - Linux container support (openzfs#14070, openzfs#14097, openzfs#12263) - BLAKE3 Checksums (openzfs#12918) - Corrective "zfs receive" (openzfs#9372) Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>

rincebrain reviewed Dec 27, 2021

View reviewed changes

include/sys/Makefile.am Outdated Show resolved Hide resolved

mcmilk force-pushed the blake3 branch from 2e2fdfc to 6aa84ab Compare December 27, 2021 22:06

mcmilk force-pushed the blake3 branch 3 times, most recently from 7cde1ae to 077b0c6 Compare December 29, 2021 00:35

mcmilk changed the title ~~Add BLAKE3 support to OpenZFS~~ Rework checksum/hash support of OpenZFS Dec 31, 2021

mcmilk force-pushed the blake3 branch 2 times, most recently from 5114389 to 30feff8 Compare January 3, 2022 11:00