Tag zfs-2.2.3

META file and changelog updated.

Signed-off-by: Tony Hutter <hutter2@llnl.gov>

.github/codeql-cpp.yml

@@ -0,0 +1,4 @@
+name: "Custom CodeQL Analysis"
+
+queries:
+  - uses: ./.github/codeql/custom-queries/cpp/deprecatedFunctionUsage.ql

.github/codeql-python.yml

@@ -0,0 +1,4 @@
+name: "Custom CodeQL Analysis"
+
+paths-ignore:
+  - tests

.github/codeql/custom-queries/cpp/deprecatedFunctionUsage.ql

@@ -0,0 +1,59 @@
+/**
+ * @name Deprecated function usage detection
+ * @description Detects functions whose usage is banned from the OpenZFS
+ *              codebase due to QA concerns.
+ * @kind problem
+ * @severity error
+ * @id cpp/deprecated-function-usage
+*/
+
+import cpp
+
+predicate isDeprecatedFunction(Function f) {
+  f.getName() = "strtok" or
+  f.getName() = "__xpg_basename" or
+  f.getName() = "basename" or
+  f.getName() = "dirname" or
+  f.getName() = "bcopy" or
+  f.getName() = "bcmp" or
+  f.getName() = "bzero" or
+  f.getName() = "asctime" or
+  f.getName() = "asctime_r" or
+  f.getName() = "gmtime" or
+  f.getName() = "localtime" or
+  f.getName() = "strncpy"
+
+}
+
+string getReplacementMessage(Function f) {
+  if f.getName() = "strtok" then
+    result = "Use strtok_r(3) instead!"
+  else if f.getName() = "__xpg_basename" then
+    result = "basename(3) is underspecified. Use zfs_basename() instead!"
+  else if f.getName() = "basename" then
+    result = "basename(3) is underspecified. Use zfs_basename() instead!"
+  else if f.getName() = "dirname" then
+    result = "dirname(3) is underspecified. Use zfs_dirnamelen() instead!"
+  else if f.getName() = "bcopy" then
+    result = "bcopy(3) is deprecated. Use memcpy(3)/memmove(3) instead!"
+  else if f.getName() = "bcmp" then
+    result = "bcmp(3) is deprecated. Use memcmp(3) instead!"
+  else if f.getName() = "bzero" then
+    result = "bzero(3) is deprecated. Use memset(3) instead!"
+  else if f.getName() = "asctime" then
+    result = "Use strftime(3) instead!"
+  else if f.getName() = "asctime_r" then
+    result = "Use strftime(3) instead!"
+  else if f.getName() = "gmtime" then
+    result = "gmtime(3) isn't thread-safe. Use gmtime_r(3) instead!"
+  else if f.getName() = "localtime" then
+    result = "localtime(3) isn't thread-safe. Use localtime_r(3) instead!"
+  else
+    result = "strncpy(3) is deprecated. Use strlcpy(3) instead!"
+}
+
+from FunctionCall fc, Function f
+where
+  fc.getTarget() = f and
+  isDeprecatedFunction(f)
+select fc, getReplacementMessage(f)

.github/codeql/custom-queries/cpp/qlpack.yml

@@ -0,0 +1,4 @@
+name: openzfs-cpp-queries
+version: 0.0.0
+libraryPathDependencies: codeql-cpp
+suites: openzfs-cpp-suite

.github/workflows/README.md

@@ -4,44 +4,54 @@
 ```mermaid
 flowchart TB
 subgraph CleanUp and Summary
-  Part1-20.04-->CleanUp+nice+Summary
-  Part2-20.04-->CleanUp+nice+Summary
-  PartN-20.04-->CleanUp+nice+Summary
-  Part1-22.04-->CleanUp+nice+Summary
-  Part2-22.04-->CleanUp+nice+Summary
-  PartN-22.04-->CleanUp+nice+Summary
+  CleanUp+Summary
 end
 
 subgraph Functional Testings
+  sanity-checks-20.04
+  zloop-checks-20.04
   functional-testing-20.04-->Part1-20.04
   functional-testing-20.04-->Part2-20.04
-  functional-testing-20.04-->PartN-20.04
+  functional-testing-20.04-->Part3-20.04
+  functional-testing-20.04-->Part4-20.04
   functional-testing-22.04-->Part1-22.04
   functional-testing-22.04-->Part2-22.04
-  functional-testing-22.04-->PartN-22.04
-end
-
-subgraph Sanity and zloop Testings
-  sanity-checks-20.04-->functional-testing-20.04
-  sanity-checks-22.04-->functional-testing-22.04
-  zloop-checks-20.04-->functional
-  zloop-checks-22.04-->functional
+  functional-testing-22.04-->Part3-22.04
+  functional-testing-22.04-->Part4-22.04
+  sanity-checks-22.04
+  zloop-checks-22.04
 end
 
 subgraph Code Checking + Building
+  Build-Ubuntu-20.04
   codeql.yml
   checkstyle.yml
-  Build-Ubuntu-20.04-->sanity-checks-20.04
-  Build-Ubuntu-22.04-->sanity-checks-22.04
-  Build-Ubuntu-20.04-->zloop-checks-20.04
-  Build-Ubuntu-22.04-->zloop-checks-22.04
+  Build-Ubuntu-22.04
 end
+
+  Build-Ubuntu-20.04-->sanity-checks-20.04
+  Build-Ubuntu-20.04-->zloop-checks-20.04
+  Build-Ubuntu-20.04-->functional-testing-20.04
+  Build-Ubuntu-22.04-->sanity-checks-22.04
+  Build-Ubuntu-22.04-->zloop-checks-22.04
+  Build-Ubuntu-22.04-->functional-testing-22.04
+
+  sanity-checks-20.04-->CleanUp+Summary
+  Part1-20.04-->CleanUp+Summary
+  Part2-20.04-->CleanUp+Summary
+  Part3-20.04-->CleanUp+Summary
+  Part4-20.04-->CleanUp+Summary
+  Part1-22.04-->CleanUp+Summary
+  Part2-22.04-->CleanUp+Summary
+  Part3-22.04-->CleanUp+Summary
+  Part4-22.04-->CleanUp+Summary
+  sanity-checks-22.04-->CleanUp+Summary
 ```
 
 
 1) build zfs modules for Ubuntu 20.04 and 22.04 (~15m)
 2) 2x zloop test (~10m) + 2x sanity test (~25m)
-3) functional testings in parts 1..5 (each ~1h)
+3) 4x functional testings in parts 1..4 (each ~1h)
 4) cleanup and create summary
    - content of summary depends on the results of the steps
 

.github/workflows/checkstyle.yaml

@@ -8,7 +8,7 @@ jobs:
   checkstyle:
     runs-on: ubuntu-22.04
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Install dependencies
@@ -52,7 +52,7 @@ jobs:
       if: failure() && steps.CheckABI.outcome == 'failure'
       run: |
         find -name *.abi | tar -cf abi_files.tar -T -
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       if: failure() && steps.CheckABI.outcome == 'failure'
       with:
         name: New ABI files (use only if you're sure about interface changes)

.github/workflows/codeql.yml

@@ -24,11 +24,12 @@ jobs:
         echo "MAKEFLAGS=-j$(nproc)" >> $GITHUB_ENV
 
     - name: Checkout repository
-      uses: actions/checkout@v3
+      uses: actions/checkout@v4
 
     - name: Initialize CodeQL
       uses: github/codeql-action/init@v2
       with:
+        config-file: .github/codeql-${{ matrix.language }}.yml
         languages: ${{ matrix.language }}
 
     - name: Autobuild

.github/workflows/scripts/generate-summary.sh

@@ -87,7 +87,7 @@ function summarize_f() {
   output "\n## $headline\n"
   rm -rf testfiles
   for i in $(seq 1 $FUNCTIONAL_PARTS); do
-    tarfile="$2/part$i.tar"
+    tarfile="$2-part$i/part$i.tar"
     check_tarfile "$tarfile"
     check_logfile "testfiles/log"
   done

.github/workflows/scripts/setup-dependencies.sh

@@ -55,29 +55,24 @@ function mod_install() {
   cat /proc/spl/kstat/zfs/chksum_bench
   echo "::endgroup::"
 
-  echo "::group::Reclaim and report disk space"
-  # remove 4GiB of images
-  sudo systemd-run docker system prune --force --all --volumes
+  echo "::group::Optimize storage for ZFS testings"
+  # remove swap and umount fast storage
+  # 89GiB -> rootfs + bootfs with ~80MB/s -> don't care
+  # 64GiB -> /mnt with 420MB/s -> new testing ssd
+  sudo swapoff -a
 
-  # remove unused software
-  sudo systemd-run --wait rm -rf \
-    "$AGENT_TOOLSDIRECTORY" \
-    /opt/* \
-    /usr/local/* \
-    /usr/share/az* \
-    /usr/share/dotnet \
-    /usr/share/gradle* \
-    /usr/share/miniconda \
-    /usr/share/swift \
-    /var/lib/gems \
-    /var/lib/mysql \
-    /var/lib/snapd
-
-  # trim the cleaned space
-  sudo fstrim /
+  # this one is fast and mounted @ /mnt
+  # -> we reformat with ext4 + move it to /var/tmp
+  DEV="/dev/disk/azure/resource-part1"
+  sudo umount /mnt
+  sudo mkfs.ext4 -O ^has_journal -F $DEV
+  sudo mount -o noatime,barrier=0 $DEV /var/tmp
+  sudo chmod 1777 /var/tmp
 
   # disk usage afterwards
-  df -h /
+  sudo df -h /
+  sudo df -h /var/tmp
+  sudo fstrim -a
   echo "::endgroup::"
 }
 

.github/workflows/zfs-linux-tests.yml

@@ -13,10 +13,10 @@ jobs:
   zloop:
     runs-on: ubuntu-${{ inputs.os }}
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         ref: ${{ github.event.pull_request.head.sha }}
-    - uses: actions/download-artifact@v3
+    - uses: actions/download-artifact@v4
       with:
         name: modules-${{ inputs.os }}
     - name: Install modules
@@ -34,7 +34,7 @@ jobs:
       if: failure()
       run: |
         sudo chmod +r -R /var/tmp/zloop/
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       if: failure()
       with:
         name: Zpool-logs-${{ inputs.os }}
@@ -43,7 +43,7 @@ jobs:
           !/var/tmp/zloop/*/vdev/
         retention-days: 14
         if-no-files-found: ignore
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       if: failure()
       with:
         name: Zpool-files-${{ inputs.os }}
@@ -55,10 +55,10 @@ jobs:
   sanity:
     runs-on: ubuntu-${{ inputs.os }}
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         ref: ${{ github.event.pull_request.head.sha }}
-    - uses: actions/download-artifact@v3
+    - uses: actions/download-artifact@v4
       with:
         name: modules-${{ inputs.os }}
     - name: Install modules
@@ -77,7 +77,7 @@ jobs:
         RESPATH="/var/tmp/test_results"
         mv -f $RESPATH/current $RESPATH/testfiles
         tar cf $RESPATH/sanity.tar -h -C $RESPATH testfiles
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       if: success() || failure()
       with:
         name: Logs-${{ inputs.os }}-sanity
@@ -91,10 +91,10 @@ jobs:
       matrix:
         tests: [ part1, part2, part3, part4 ]
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         ref: ${{ github.event.pull_request.head.sha }}
-    - uses: actions/download-artifact@v3
+    - uses: actions/download-artifact@v4
       with:
         name: modules-${{ inputs.os }}
     - name: Install modules
@@ -116,9 +116,9 @@ jobs:
         RESPATH="/var/tmp/test_results"
         mv -f $RESPATH/current $RESPATH/testfiles
         tar cf $RESPATH/${{ matrix.tests }}.tar -h -C $RESPATH testfiles
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       if: success() || failure()
       with:
-        name: Logs-${{ inputs.os }}-functional
+        name: Logs-${{ inputs.os }}-functional-${{ matrix.tests }}
         path: /var/tmp/test_results/${{ matrix.tests }}.tar
         if-no-files-found: ignore

.github/workflows/zfs-linux.yml

@@ -14,14 +14,14 @@ jobs:
         os: [20.04, 22.04]
     runs-on: ubuntu-${{ matrix.os }}
     steps:
-    - uses: actions/checkout@v3
+    - uses: actions/checkout@v4
       with:
         ref: ${{ github.event.pull_request.head.sha }}
     - name: Build modules
       run: .github/workflows/scripts/setup-dependencies.sh build
     - name: Prepare modules upload
       run: tar czf modules-${{ matrix.os }}.tgz *.deb .github tests/test-runner tests/ImageOS.txt
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       with:
         name: modules-${{ matrix.os }}
         path: modules-${{ matrix.os }}.tgz
@@ -44,7 +44,7 @@ jobs:
     runs-on: ubuntu-22.04
     needs: testings
     steps:
-    - uses: actions/download-artifact@v3
+    - uses: actions/download-artifact@v4
     - name: Generating summary
       run: |
         tar xzf modules-22.04/modules-22.04.tgz .github tests
@@ -58,7 +58,7 @@ jobs:
       run: .github/workflows/scripts/generate-summary.sh 3
     - name: Summary for errors #4
       run: .github/workflows/scripts/generate-summary.sh 4
-    - uses: actions/upload-artifact@v3
+    - uses: actions/upload-artifact@v4
       with:
         name: Summary Files
         path: Summary/

META

@@ -1,10 +1,10 @@
 Meta:          1
 Name:          zfs
 Branch:        1.0
-Version:       2.2.1
+Version:       2.2.3
 Release:       1
 Release-Tags:  relext
 License:       CDDL
 Author:        OpenZFS
-Linux-Maximum: 6.6
+Linux-Maximum: 6.7
 Linux-Minimum: 3.10

README.md

@@ -32,4 +32,4 @@ For more details see the NOTICE, LICENSE and COPYRIGHT files; `UCRL-CODE-235197`
 
 # Supported Kernels
   * The `META` file contains the officially recognized supported Linux kernel versions.
-  * Supported FreeBSD versions are any supported branches and releases starting from 12.2-RELEASE.
+  * Supported FreeBSD versions are any supported branches and releases starting from 12.4-RELEASE.

cmd/zdb/zdb.c

@@ -34,6 +34,7 @@
  * Copyright (c) 2021 Allan Jude
  * Copyright (c) 2021 Toomas Soome <tsoome@me.com>
  * Copyright (c) 2023, Klara Inc.
+ * Copyright (c) 2023, Rob Norris <robn@despairlabs.com>
  */
 
 #include <stdio.h>
@@ -80,6 +81,7 @@
 #include <sys/dsl_scan.h>
 #include <sys/btree.h>
 #include <sys/brt.h>
+#include <sys/brt_impl.h>
 #include <zfs_comutil.h>
 #include <sys/zstd/zstd.h>
 
@@ -899,6 +901,8 @@ usage(void)
 	    "don't print label contents\n");
 	(void) fprintf(stderr, "        -t --txg=INTEGER             "
 	    "highest txg to use when searching for uberblocks\n");
+	(void) fprintf(stderr, "        -T --brt-stats               "
+	    "BRT statistics\n");
 	(void) fprintf(stderr, "        -u --uberblock               "
 	    "uberblock\n");
 	(void) fprintf(stderr, "        -U --cachefile=PATH          "
@@ -999,6 +1003,15 @@ zdb_nicenum(uint64_t num, char *buf, size_t buflen)
 		nicenum(num, buf, buflen);
 }
 
+static void
+zdb_nicebytes(uint64_t bytes, char *buf, size_t buflen)
+{
+	if (dump_opt['P'])
+		(void) snprintf(buf, buflen, "%llu", (longlong_t)bytes);
+	else
+		zfs_nicebytes(bytes, buf, buflen);
+}
+
 static const char histo_stars[] = "****************************************";
 static const uint64_t histo_width = sizeof (histo_stars) - 1;
 
@@ -2081,6 +2094,76 @@ dump_all_ddts(spa_t *spa)
 	dump_dedup_ratio(&dds_total);
 }
 
+static void
+dump_brt(spa_t *spa)
+{
+	if (!spa_feature_is_enabled(spa, SPA_FEATURE_BLOCK_CLONING)) {
+		printf("BRT: unsupported on this pool\n");
+		return;
+	}
+
+	if (!spa_feature_is_active(spa, SPA_FEATURE_BLOCK_CLONING)) {
+		printf("BRT: empty\n");
+		return;
+	}
+
+	brt_t *brt = spa->spa_brt;
+	VERIFY(brt);
+
+	char count[32], used[32], saved[32];
+	zdb_nicebytes(brt_get_used(spa), used, sizeof (used));
+	zdb_nicebytes(brt_get_saved(spa), saved, sizeof (saved));
+	uint64_t ratio = brt_get_ratio(spa);
+	printf("BRT: used %s; saved %s; ratio %llu.%02llux\n", used, saved,
+	    (u_longlong_t)(ratio / 100), (u_longlong_t)(ratio % 100));
+
+	if (dump_opt['T'] < 2)
+		return;
+
+	for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
+		brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
+		if (brtvd == NULL)
+			continue;
+
+		if (!brtvd->bv_initiated) {
+			printf("BRT: vdev %" PRIu64 ": empty\n", vdevid);
+			continue;
+		}
+
+		zdb_nicenum(brtvd->bv_totalcount, count, sizeof (count));
+		zdb_nicebytes(brtvd->bv_usedspace, used, sizeof (used));
+		zdb_nicebytes(brtvd->bv_savedspace, saved, sizeof (saved));
+		printf("BRT: vdev %" PRIu64 ": refcnt %s; used %s; saved %s\n",
+		    vdevid, count, used, saved);
+	}
+
+	if (dump_opt['T'] < 3)
+		return;
+
+	char dva[64];
+	printf("\n%-16s %-10s\n", "DVA", "REFCNT");
+
+	for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
+		brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
+		if (brtvd == NULL || !brtvd->bv_initiated)
+			continue;
+
+		zap_cursor_t zc;
+		zap_attribute_t za;
+		for (zap_cursor_init(&zc, brt->brt_mos, brtvd->bv_mos_entries);
+		    zap_cursor_retrieve(&zc, &za) == 0;
+		    zap_cursor_advance(&zc)) {
+			uint64_t offset = *(uint64_t *)za.za_name;
+			uint64_t refcnt = za.za_first_integer;
+
+			snprintf(dva, sizeof (dva), "%" PRIu64 ":%llx", vdevid,
+			    (u_longlong_t)offset);
+			printf("%-16s %-10llu\n", dva, (u_longlong_t)refcnt);
+		}
+		zap_cursor_fini(&zc);
+	}
+}
+
 static void
 dump_dtl_seg(void *arg, uint64_t start, uint64_t size)
 {
@@ -2277,7 +2360,7 @@ static void
 snprintf_zstd_header(spa_t *spa, char *blkbuf, size_t buflen,
     const blkptr_t *bp)
 {
-	abd_t *pabd;
+	static abd_t *pabd = NULL;
 	void *buf;
 	zio_t *zio;
 	zfs_zstdhdr_t zstd_hdr;
@@ -2308,7 +2391,8 @@ snprintf_zstd_header(spa_t *spa, char *blkbuf, size_t buflen,
 		return;
 	}
 
-	pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
+	if (!pabd)
+		pabd = abd_alloc_for_io(SPA_MAXBLOCKSIZE, B_FALSE);
 	zio = zio_root(spa, NULL, NULL, 0);
 
 	/* Decrypt but don't decompress so we can read the compression header */
@@ -7957,6 +8041,17 @@ dump_mos_leaks(spa_t *spa)
 		}
 	}
 
+	if (spa->spa_brt != NULL) {
+		brt_t *brt = spa->spa_brt;
+		for (uint64_t vdevid = 0; vdevid < brt->brt_nvdevs; vdevid++) {
+			brt_vdev_t *brtvd = &brt->brt_vdevs[vdevid];
+			if (brtvd != NULL && brtvd->bv_initiated) {
+				mos_obj_refd(brtvd->bv_mos_brtvdev);
+				mos_obj_refd(brtvd->bv_mos_entries);
+			}
+		}
+	}
+
 	/*
 	 * Visit all allocated objects and make sure they are referenced.
 	 */
@@ -8093,6 +8188,9 @@ dump_zpool(spa_t *spa)
 	if (dump_opt['D'])
 		dump_all_ddts(spa);
 
+	if (dump_opt['T'])
+		dump_brt(spa);
+
 	if (dump_opt['d'] > 2 || dump_opt['m'])
 		dump_metaslabs(spa);
 	if (dump_opt['M'])
@@ -8404,6 +8502,14 @@ zdb_decompress_block(abd_t *pabd, void *buf, void *lbuf, uint64_t lsize,
 	*cfuncp++ = ZIO_COMPRESS_LZ4;
 	*cfuncp++ = ZIO_COMPRESS_LZJB;
 	mask |= ZIO_COMPRESS_MASK(LZ4) | ZIO_COMPRESS_MASK(LZJB);
+	/*
+	 * Every gzip level has the same decompressor, no need to
+	 * run it 9 times per bruteforce attempt.
+	 */
+	mask |= ZIO_COMPRESS_MASK(GZIP_2) | ZIO_COMPRESS_MASK(GZIP_3);
+	mask |= ZIO_COMPRESS_MASK(GZIP_4) | ZIO_COMPRESS_MASK(GZIP_5);
+	mask |= ZIO_COMPRESS_MASK(GZIP_6) | ZIO_COMPRESS_MASK(GZIP_7);
+	mask |= ZIO_COMPRESS_MASK(GZIP_8) | ZIO_COMPRESS_MASK(GZIP_9);
 	for (int c = 0; c < ZIO_COMPRESS_FUNCTIONS; c++)
 		if (((1ULL << c) & mask) == 0)
 			*cfuncp++ = c;
@@ -8879,6 +8985,7 @@ main(int argc, char **argv)
 		{"io-stats",		no_argument,		NULL, 's'},
 		{"simulate-dedup",	no_argument,		NULL, 'S'},
 		{"txg",			required_argument,	NULL, 't'},
+		{"brt-stats",		no_argument,		NULL, 'T'},
 		{"uberblock",		no_argument,		NULL, 'u'},
 		{"cachefile",		required_argument,	NULL, 'U'},
 		{"verbose",		no_argument,		NULL, 'v'},
@@ -8892,7 +8999,7 @@ main(int argc, char **argv)
 	};
 
 	while ((c = getopt_long(argc, argv,
-	    "AbBcCdDeEFGhiI:kK:lLmMNo:Op:PqrRsSt:uU:vVx:XYyZ",
+	    "AbBcCdDeEFGhiI:kK:lLmMNo:Op:PqrRsSt:TuU:vVx:XYyZ",
 	    long_options, NULL)) != -1) {
 		switch (c) {
 		case 'b':
@@ -8914,6 +9021,7 @@ main(int argc, char **argv)
 		case 'R':
 		case 's':
 		case 'S':
+		case 'T':
 		case 'u':
 		case 'y':
 		case 'Z':
@@ -9076,22 +9184,6 @@ main(int argc, char **argv)
 	if (dump_opt['l'])
 		return (dump_label(argv[0]));
 
-	if (dump_opt['O']) {
-		if (argc != 2)
-			usage();
-		dump_opt['v'] = verbose + 3;
-		return (dump_path(argv[0], argv[1], NULL));
-	}
-	if (dump_opt['r']) {
-		target_is_spa = B_FALSE;
-		if (argc != 3)
-			usage();
-		dump_opt['v'] = verbose;
-		error = dump_path(argv[0], argv[1], &object);
-		if (error != 0)
-			fatal("internal error: %s", strerror(error));
-	}
-
 	if (dump_opt['X'] || dump_opt['F'])
 		rewind = ZPOOL_DO_REWIND |
 		    (dump_opt['X'] ? ZPOOL_EXTREME_REWIND : 0);
@@ -9192,6 +9284,29 @@ main(int argc, char **argv)
 		searchdirs = NULL;
 	}
 
+	/*
+	 * We need to make sure to process -O option or call
+	 * dump_path after the -e option has been processed,
+	 * which imports the pool to the namespace if it's
+	 * not in the cachefile.
+	 */
+	if (dump_opt['O']) {
+		if (argc != 2)
+			usage();
+		dump_opt['v'] = verbose + 3;
+		return (dump_path(argv[0], argv[1], NULL));
+	}
+
+	if (dump_opt['r']) {
+		target_is_spa = B_FALSE;
+		if (argc != 3)
+			usage();
+		dump_opt['v'] = verbose;
+		error = dump_path(argv[0], argv[1], &object);
+		if (error != 0)
+			fatal("internal error: %s", strerror(error));
+	}
+
 	/*
 	 * import_checkpointed_state makes the assumption that the
 	 * target pool that we pass it is already part of the spa

cmd/zdb/zdb_il.c

@@ -168,7 +168,7 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, const void *arg)
 	    (u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_offset,
 	    (u_longlong_t)lr->lr_length);
 
-	if (txtype == TX_WRITE2 || verbose < 5)
+	if (txtype == TX_WRITE2 || verbose < 4)
 		return;
 
 	if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
@@ -178,6 +178,8 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, const void *arg)
 		    "will claim" : "won't claim");
 		print_log_bp(bp, tab_prefix);
 
+		if (verbose < 5)
+			return;
 		if (BP_IS_HOLE(bp)) {
 			(void) printf("\t\t\tLSIZE 0x%llx\n",
 			    (u_longlong_t)BP_GET_LSIZE(bp));
@@ -202,6 +204,9 @@ zil_prt_rec_write(zilog_t *zilog, int txtype, const void *arg)
 		if (error)
 			goto out;
 	} else {
+		if (verbose < 5)
+			return;
+
 		/* data is stored after the end of the lr_write record */
 		data = abd_alloc(lr->lr_length, B_FALSE);
 		abd_copy_from_buf(data, lr + 1, lr->lr_length);
@@ -217,6 +222,28 @@ out:
 	abd_free(data);
 }
 
+static void
+zil_prt_rec_write_enc(zilog_t *zilog, int txtype, const void *arg)
+{
+	(void) txtype;
+	const lr_write_t *lr = arg;
+	const blkptr_t *bp = &lr->lr_blkptr;
+	int verbose = MAX(dump_opt['d'], dump_opt['i']);
+
+	(void) printf("%s(encrypted)\n", tab_prefix);
+
+	if (verbose < 4)
+		return;
+
+	if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
+		(void) printf("%shas blkptr, %s\n", tab_prefix,
+		    !BP_IS_HOLE(bp) &&
+		    bp->blk_birth >= spa_min_claim_txg(zilog->zl_spa) ?
+		    "will claim" : "won't claim");
+		print_log_bp(bp, tab_prefix);
+	}
+}
+
 static void
 zil_prt_rec_truncate(zilog_t *zilog, int txtype, const void *arg)
 {
@@ -312,11 +339,34 @@ zil_prt_rec_clone_range(zilog_t *zilog, int txtype, const void *arg)
 {
 	(void) zilog, (void) txtype;
 	const lr_clone_range_t *lr = arg;
+	int verbose = MAX(dump_opt['d'], dump_opt['i']);
 
 	(void) printf("%sfoid %llu, offset %llx, length %llx, blksize %llx\n",
 	    tab_prefix, (u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_offset,
 	    (u_longlong_t)lr->lr_length, (u_longlong_t)lr->lr_blksz);
 
+	if (verbose < 4)
+		return;
+
+	for (unsigned int i = 0; i < lr->lr_nbps; i++) {
+		(void) printf("%s[%u/%llu] ", tab_prefix, i + 1,
+		    (u_longlong_t)lr->lr_nbps);
+		print_log_bp(&lr->lr_bps[i], "");
+	}
+}
+
+static void
+zil_prt_rec_clone_range_enc(zilog_t *zilog, int txtype, const void *arg)
+{
+	(void) zilog, (void) txtype;
+	const lr_clone_range_t *lr = arg;
+	int verbose = MAX(dump_opt['d'], dump_opt['i']);
+
+	(void) printf("%s(encrypted)\n", tab_prefix);
+
+	if (verbose < 4)
+		return;
+
 	for (unsigned int i = 0; i < lr->lr_nbps; i++) {
 		(void) printf("%s[%u/%llu] ", tab_prefix, i + 1,
 		    (u_longlong_t)lr->lr_nbps);
@@ -327,6 +377,7 @@ zil_prt_rec_clone_range(zilog_t *zilog, int txtype, const void *arg)
 typedef void (*zil_prt_rec_func_t)(zilog_t *, int, const void *);
 typedef struct zil_rec_info {
 	zil_prt_rec_func_t	zri_print;
+	zil_prt_rec_func_t	zri_print_enc;
 	const char		*zri_name;
 	uint64_t		zri_count;
 } zil_rec_info_t;
@@ -341,7 +392,9 @@ static zil_rec_info_t zil_rec_info[TX_MAX_TYPE] = {
 	{.zri_print = zil_prt_rec_remove,   .zri_name = "TX_RMDIR           "},
 	{.zri_print = zil_prt_rec_link,	    .zri_name = "TX_LINK            "},
 	{.zri_print = zil_prt_rec_rename,   .zri_name = "TX_RENAME          "},
-	{.zri_print = zil_prt_rec_write,    .zri_name = "TX_WRITE           "},
+	{.zri_print = zil_prt_rec_write,
+	    .zri_print_enc = zil_prt_rec_write_enc,
+	    .zri_name = "TX_WRITE           "},
 	{.zri_print = zil_prt_rec_truncate, .zri_name = "TX_TRUNCATE        "},
 	{.zri_print = zil_prt_rec_setattr,  .zri_name = "TX_SETATTR         "},
 	{.zri_print = zil_prt_rec_acl,	    .zri_name = "TX_ACL_V0          "},
@@ -358,6 +411,7 @@ static zil_rec_info_t zil_rec_info[TX_MAX_TYPE] = {
 	{.zri_print = zil_prt_rec_rename,   .zri_name = "TX_RENAME_EXCHANGE "},
 	{.zri_print = zil_prt_rec_rename,   .zri_name = "TX_RENAME_WHITEOUT "},
 	{.zri_print = zil_prt_rec_clone_range,
+	    .zri_print_enc = zil_prt_rec_clone_range_enc,
 	    .zri_name = "TX_CLONE_RANGE     "},
 };
 
@@ -384,6 +438,8 @@ print_log_record(zilog_t *zilog, const lr_t *lr, void *arg, uint64_t claim_txg)
 	if (txtype && verbose >= 3) {
 		if (!zilog->zl_os->os_encrypted) {
 			zil_rec_info[txtype].zri_print(zilog, txtype, lr);
+		} else if (zil_rec_info[txtype].zri_print_enc) {
+			zil_rec_info[txtype].zri_print_enc(zilog, txtype, lr);
 		} else {
 			(void) printf("%s(encrypted)\n", tab_prefix);
 		}

cmd/zed/agents/zfs_mod.c

@@ -233,8 +233,12 @@ zfs_process_add(zpool_handle_t *zhp, nvlist_t *vdev, boolean_t labeled)
 	}
 
 	(void) nvlist_lookup_string(vdev, ZPOOL_CONFIG_PHYS_PATH, &physpath);
+
+	update_vdev_config_dev_sysfs_path(vdev, path,
+	    ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);
 	(void) nvlist_lookup_string(vdev, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
 	    &enc_sysfs_path);
+
 	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk);
 	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_OFFLINE, &offline);
 	(void) nvlist_lookup_uint64(vdev, ZPOOL_CONFIG_FAULTED, &faulted);

cmd/zed/zed.d/statechange-slot_off.sh

@@ -5,7 +5,7 @@
 #
 # Bad SCSI disks can often "disappear and reappear" causing all sorts of chaos
 # as they flip between FAULTED and ONLINE.  If
-# ZED_POWER_OFF_ENCLOUSRE_SLOT_ON_FAULT is set in zed.rc, and the disk gets
+# ZED_POWER_OFF_ENCLOSURE_SLOT_ON_FAULT is set in zed.rc, and the disk gets
 # FAULTED, then power down the slot via sysfs:
 #
 # /sys/class/enclosure/<enclosure>/<slot>/power_status
@@ -19,7 +19,7 @@
 # Exit codes:
 #   0: slot successfully powered off
 #   1: enclosure not available
-#   2: ZED_POWER_OFF_ENCLOUSRE_SLOT_ON_FAULT disabled
+#   2: ZED_POWER_OFF_ENCLOSURE_SLOT_ON_FAULT disabled
 #   3: vdev was not FAULTED
 #   4: The enclosure sysfs path passed from ZFS does not exist
 #   5: Enclosure slot didn't actually turn off after we told it to
@@ -32,7 +32,7 @@ if [ ! -d /sys/class/enclosure ] ; then
 	exit 1
 fi
 
-if [ "${ZED_POWER_OFF_ENCLOUSRE_SLOT_ON_FAULT}" != "1" ] ; then
+if [ "${ZED_POWER_OFF_ENCLOSURE_SLOT_ON_FAULT}" != "1" ] ; then
 	exit 2
 fi
 

cmd/zed/zed.d/zed-functions.sh

@@ -205,6 +205,10 @@ zed_notify()
     [ "${rv}" -eq 0 ] && num_success=$((num_success + 1))
     [ "${rv}" -eq 1 ] && num_failure=$((num_failure + 1))
 
+    zed_notify_ntfy "${subject}" "${pathname}"; rv=$?
+    [ "${rv}" -eq 0 ] && num_success=$((num_success + 1))
+    [ "${rv}" -eq 1 ] && num_failure=$((num_failure + 1))
+
     [ "${num_success}" -gt 0 ] && return 0
     [ "${num_failure}" -gt 0 ] && return 1
     return 2
@@ -527,6 +531,100 @@ zed_notify_pushover()
 }
 
 
+# zed_notify_ntfy (subject, pathname)
+#
+# Send a notification via Ntfy.sh <https://ntfy.sh/>.
+# The ntfy topic (ZED_NTFY_TOPIC) identifies the topic that the notification
+# will be sent to Ntfy.sh server. The ntfy url (ZED_NTFY_URL) defines the
+# self-hosted or provided hosted ntfy service location. The ntfy access token
+# <https://docs.ntfy.sh/publish/#access-tokens> (ZED_NTFY_ACCESS_TOKEN) reprsents an
+# access token that could be used if a topic is read/write protected. If a
+# topic can be written to publicaly, a ZED_NTFY_ACCESS_TOKEN is not required.
+#
+# Requires curl and sed executables to be installed in the standard PATH.
+#
+# References
+#   https://docs.ntfy.sh
+#
+# Arguments
+#   subject: notification subject
+#   pathname: pathname containing the notification message (OPTIONAL)
+#
+# Globals
+#   ZED_NTFY_TOPIC
+#   ZED_NTFY_ACCESS_TOKEN (OPTIONAL)
+#   ZED_NTFY_URL
+#
+# Return
+#   0: notification sent
+#   1: notification failed
+#   2: not configured
+#
+zed_notify_ntfy()
+{
+    local subject="$1"
+    local pathname="${2:-"/dev/null"}"
+    local msg_body
+    local msg_out
+    local msg_err
+
+    [ -n "${ZED_NTFY_TOPIC}" ] || return 2
+    local url="${ZED_NTFY_URL:-"https://ntfy.sh"}/${ZED_NTFY_TOPIC}"
+
+    if [ ! -r "${pathname}" ]; then
+        zed_log_err "ntfy cannot read \"${pathname}\""
+        return 1
+    fi
+
+    zed_check_cmd "curl" "sed" || return 1
+
+    # Read the message body in.
+    #
+    msg_body="$(cat "${pathname}")"
+
+    if [ -z "${msg_body}" ]
+    then
+        msg_body=$subject
+        subject=""
+    fi
+
+    # Send the POST request and check for errors.
+    #
+    if [ -n "${ZED_NTFY_ACCESS_TOKEN}" ]; then
+        msg_out="$( \
+        curl \
+        -u ":${ZED_NTFY_ACCESS_TOKEN}" \
+        -H "Title: ${subject}" \
+        -d "${msg_body}" \
+        -H "Priority: high" \
+        "${url}" \
+        2>/dev/null \
+        )"; rv=$?
+    else
+        msg_out="$( \
+        curl \
+        -H "Title: ${subject}" \
+        -d "${msg_body}" \
+        -H "Priority: high" \
+        "${url}" \
+        2>/dev/null \
+        )"; rv=$?
+    fi
+    if [ "${rv}" -ne 0 ]; then
+        zed_log_err "curl exit=${rv}"
+        return 1
+    fi
+    msg_err="$(echo "${msg_out}" \
+        | sed -n -e 's/.*"errors" *:.*\[\(.*\)\].*/\1/p')"
+    if [ -n "${msg_err}" ]; then
+        zed_log_err "ntfy \"${msg_err}"\"
+        return 1
+    fi
+    return 0
+}
+
+
+
 # zed_rate_limit (tag, [interval])
 #
 # Check whether an event of a given type [tag] has already occurred within the

cmd/zed/zed.d/zed.rc

@@ -146,4 +146,26 @@ ZED_SYSLOG_SUBCLASS_EXCLUDE="history_event"
 # Power off the drive's slot in the enclosure if it becomes FAULTED.  This can
 # help silence misbehaving drives.  This assumes your drive enclosure fully
 # supports slot power control via sysfs.
-#ZED_POWER_OFF_ENCLOUSRE_SLOT_ON_FAULT=1
+#ZED_POWER_OFF_ENCLOSURE_SLOT_ON_FAULT=1
+
+##
+# Ntfy topic
+# This defines which topic will receive the ntfy notification.
+#  <https://docs.ntfy.sh/publish/>
+# Disabled by default; uncomment to enable.
+#ZED_NTFY_TOPIC=""
+
+##
+# Ntfy access token (optional for public topics)
+# This defines an access token which can be used
+# to allow you to authenticate when sending to topics
+# <https://docs.ntfy.sh/publish/#access-tokens>
+# Disabled by default; uncomment to enable.
+#ZED_NTFY_ACCESS_TOKEN=""
+
+##
+# Ntfy Service URL
+# This defines which service the ntfy call will be directed toward
+#  <https://docs.ntfy.sh/install/>
+# https://ntfy.sh by default; uncomment to enable an alternative service url.
+#ZED_NTFY_URL="https://ntfy.sh"

cmd/zed/zed_event.c

@@ -35,6 +35,7 @@
 #include "zed_strings.h"
 
 #include "agents/zfs_agents.h"
+#include <libzutil.h>
 
 #define	MAXBUF	4096
 
@@ -922,6 +923,25 @@ _zed_event_add_time_strings(uint64_t eid, zed_strings_t *zsp, int64_t etime[])
 	}
 }
 
+
+static void
+_zed_event_update_enc_sysfs_path(nvlist_t *nvl)
+{
+	const char *vdev_path;
+
+	if (nvlist_lookup_string(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_PATH,
+	    &vdev_path) != 0) {
+		return; /* some other kind of event, ignore it */
+	}
+
+	if (vdev_path == NULL) {
+		return;
+	}
+
+	update_vdev_config_dev_sysfs_path(nvl, vdev_path,
+	    FM_EREPORT_PAYLOAD_ZFS_VDEV_ENC_SYSFS_PATH);
+}
+
 /*
  * Service the next zevent, blocking until one is available.
  */
@@ -969,6 +989,17 @@ zed_event_service(struct zed_conf *zcp)
 		zed_log_msg(LOG_WARNING,
 		    "Failed to lookup zevent class (eid=%llu)", eid);
 	} else {
+		/*
+		 * Special case: If we can dynamically detect an enclosure sysfs
+		 * path, then use that value rather than the one stored in the
+		 * vd->vdev_enc_sysfs_path.  There have been rare cases where
+		 * vd->vdev_enc_sysfs_path becomes outdated.  However, there
+		 * will be other times when we can not dynamically detect the
+		 * sysfs path (like if a disk disappears) and have to rely on
+		 * the old value for things like turning on the fault LED.
+		 */
+		_zed_event_update_enc_sysfs_path(nvl);
+
 		/* let internal modules see this event first */
 		zfs_agent_post_event(class, NULL, nvl);
 

cmd/zfs/zfs_main.c

@@ -3672,15 +3672,25 @@ zfs_do_list(int argc, char **argv)
 
 			for (char *tok; (tok = strsep(&optarg, ",")); ) {
 				static const char *const type_subopts[] = {
-					"filesystem", "volume",
-					"snapshot", "snap",
+					"filesystem",
+					"fs",
+					"volume",
+					"vol",
+					"snapshot",
+					"snap",
 					"bookmark",
-					"all" };
+					"all"
+				};
 				static const int type_types[] = {
-					ZFS_TYPE_FILESYSTEM, ZFS_TYPE_VOLUME,
-					ZFS_TYPE_SNAPSHOT, ZFS_TYPE_SNAPSHOT,
+					ZFS_TYPE_FILESYSTEM,
+					ZFS_TYPE_FILESYSTEM,
+					ZFS_TYPE_VOLUME,
+					ZFS_TYPE_VOLUME,
+					ZFS_TYPE_SNAPSHOT,
+					ZFS_TYPE_SNAPSHOT,
 					ZFS_TYPE_BOOKMARK,
-					ZFS_TYPE_DATASET | ZFS_TYPE_BOOKMARK };
+					ZFS_TYPE_DATASET | ZFS_TYPE_BOOKMARK
+				};
 
 				for (c = 0; c < ARRAY_SIZE(type_subopts); ++c)
 					if (strcmp(tok, type_subopts[c]) == 0) {
@@ -7230,7 +7240,8 @@ share_mount(int op, int argc, char **argv)
 		pthread_mutex_init(&share_mount_state.sm_lock, NULL);
 
 		/* For a 'zfs share -a' operation start with a clean slate. */
-		zfs_truncate_shares(NULL);
+		if (op == OP_SHARE)
+			zfs_truncate_shares(NULL);
 
 		/*
 		 * libshare isn't mt-safe, so only do the operation in parallel

cmd/zpool/os/freebsd/zpool_vdev_os.c

@@ -124,3 +124,24 @@ check_file(const char *file, boolean_t force, boolean_t isspare)
 {
 	return (check_file_generic(file, force, isspare));
 }
+
+int
+zpool_power_current_state(zpool_handle_t *zhp, char *vdev)
+{
+
+	(void) zhp;
+	(void) vdev;
+	/* Enclosure slot power not supported on FreeBSD yet */
+	return (-1);
+}
+
+int
+zpool_power(zpool_handle_t *zhp, char *vdev, boolean_t turn_on)
+{
+
+	(void) zhp;
+	(void) vdev;
+	(void) turn_on;
+	/* Enclosure slot power not supported on FreeBSD yet */
+	return (ENOTSUP);
+}

cmd/zpool/os/linux/zpool_vdev_os.c

@@ -416,3 +416,258 @@ check_file(const char *file, boolean_t force, boolean_t isspare)
 {
 	return (check_file_generic(file, force, isspare));
 }
+
+/*
+ * Read from a sysfs file and return an allocated string.  Removes
+ * the newline from the end of the string if there is one.
+ *
+ * Returns a string on success (which must be freed), or NULL on error.
+ */
+static char *zpool_sysfs_gets(char *path)
+{
+	int fd;
+	struct stat statbuf;
+	char *buf = NULL;
+	ssize_t count = 0;
+	fd = open(path, O_RDONLY);
+	if (fd < 0)
+		return (NULL);
+
+	if (fstat(fd, &statbuf) != 0) {
+		close(fd);
+		return (NULL);
+	}
+
+	buf = calloc(sizeof (*buf), statbuf.st_size + 1);
+	if (buf == NULL) {
+		close(fd);
+		return (NULL);
+	}
+
+	/*
+	 * Note, we can read less bytes than st_size, and that's ok.  Sysfs
+	 * files will report their size is 4k even if they only return a small
+	 * string.
+	 */
+	count = read(fd, buf, statbuf.st_size);
+	if (count < 0) {
+		/* Error doing read() or we overran the buffer */
+		close(fd);
+		free(buf);
+		return (NULL);
+	}
+
+	/* Remove trailing newline */
+	if (buf[count - 1] == '\n')
+		buf[count - 1] = 0;
+
+	close(fd);
+
+	return (buf);
+}
+
+/*
+ * Write a string to a sysfs file.
+ *
+ * Returns 0 on success, non-zero otherwise.
+ */
+static int zpool_sysfs_puts(char *path, char *str)
+{
+	FILE *file;
+
+	file = fopen(path, "w");
+	if (!file) {
+		return (-1);
+	}
+
+	if (fputs(str, file) < 0) {
+		fclose(file);
+		return (-2);
+	}
+	fclose(file);
+	return (0);
+}
+
+/* Given a vdev nvlist_t, rescan its enclosure sysfs path */
+static void
+rescan_vdev_config_dev_sysfs_path(nvlist_t *vdev_nv)
+{
+	update_vdev_config_dev_sysfs_path(vdev_nv,
+	    fnvlist_lookup_string(vdev_nv, ZPOOL_CONFIG_PATH),
+	    ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);
+}
+
+/*
+ * Given a power string: "on", "off", "1", or "0", return 0 if it's an
+ * off value, 1 if it's an on value, and -1 if the value is unrecognized.
+ */
+static int zpool_power_parse_value(char *str)
+{
+	if ((strcmp(str, "off") == 0) || (strcmp(str, "0") == 0))
+		return (0);
+
+	if ((strcmp(str, "on") == 0) || (strcmp(str, "1") == 0))
+		return (1);
+
+	return (-1);
+}
+
+/*
+ * Given a vdev string return an allocated string containing the sysfs path to
+ * its power control file.  Also do a check if the power control file really
+ * exists and has correct permissions.
+ *
+ * Example returned strings:
+ *
+ * /sys/class/enclosure/0:0:122:0/10/power_status
+ * /sys/bus/pci/slots/10/power
+ *
+ * Returns allocated string on success (which must be freed), NULL on failure.
+ */
+static char *
+zpool_power_sysfs_path(zpool_handle_t *zhp, char *vdev)
+{
+	const char *enc_sysfs_dir = NULL;
+	char *path = NULL;
+	nvlist_t *vdev_nv = zpool_find_vdev(zhp, vdev, NULL, NULL, NULL);
+
+	if (vdev_nv == NULL) {
+		return (NULL);
+	}
+
+	/* Make sure we're getting the updated enclosure sysfs path */
+	rescan_vdev_config_dev_sysfs_path(vdev_nv);
+
+	if (nvlist_lookup_string(vdev_nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
+	    &enc_sysfs_dir) != 0) {
+		return (NULL);
+	}
+
+	if (asprintf(&path, "%s/power_status", enc_sysfs_dir) == -1)
+		return (NULL);
+
+	if (access(path, W_OK) != 0) {
+		free(path);
+		path = NULL;
+		/* No HDD 'power_control' file, maybe it's NVMe? */
+		if (asprintf(&path, "%s/power", enc_sysfs_dir) == -1) {
+			return (NULL);
+		}
+
+		if (access(path, R_OK | W_OK) != 0) {
+			/* Not NVMe either */
+			free(path);
+			return (NULL);
+		}
+	}
+
+	return (path);
+}
+
+/*
+ * Given a path to a sysfs power control file, return B_TRUE if you should use
+ * "on/off" words to control it, or B_FALSE otherwise ("0/1" to control).
+ */
+static boolean_t
+zpool_power_use_word(char *sysfs_path)
+{
+	if (strcmp(&sysfs_path[strlen(sysfs_path) - strlen("power_status")],
+	    "power_status") == 0) {
+		return (B_TRUE);
+	}
+	return (B_FALSE);
+}
+
+/*
+ * Check the sysfs power control value for a vdev.
+ *
+ * Returns:
+ *  0 - Power is off
+ *  1 - Power is on
+ * -1 - Error or unsupported
+ */
+int
+zpool_power_current_state(zpool_handle_t *zhp, char *vdev)
+{
+	char *val;
+	int rc;
+
+	char *path = zpool_power_sysfs_path(zhp, vdev);
+	if (path == NULL)
+		return (-1);
+
+	val = zpool_sysfs_gets(path);
+	if (val == NULL) {
+		free(path);
+		return (-1);
+	}
+
+	rc = zpool_power_parse_value(val);
+	free(val);
+	free(path);
+	return (rc);
+}
+
+/*
+ * Turn on or off the slot to a device
+ *
+ * Device path is the full path to the device (like /dev/sda or /dev/sda1).
+ *
+ * Return code:
+ * 0:		Success
+ * ENOTSUP:	Power control not supported for OS
+ * EBADSLT:	Couldn't read current power state
+ * ENOENT:	No sysfs path to power control
+ * EIO:	Couldn't write sysfs power value
+ * EBADE:	Sysfs power value didn't change
+ */
+int
+zpool_power(zpool_handle_t *zhp, char *vdev, boolean_t turn_on)
+{
+	char *sysfs_path;
+	const char *val;
+	int rc;
+	int timeout_ms;
+
+	rc = zpool_power_current_state(zhp, vdev);
+	if (rc == -1) {
+		return (EBADSLT);
+	}
+
+	/* Already correct value? */
+	if (rc == (int)turn_on)
+		return (0);
+
+	sysfs_path = zpool_power_sysfs_path(zhp, vdev);
+	if (sysfs_path == NULL)
+		return (ENOENT);
+
+	if (zpool_power_use_word(sysfs_path)) {
+		val = turn_on ? "on" : "off";
+	} else {
+		val = turn_on ? "1" : "0";
+	}
+
+	rc = zpool_sysfs_puts(sysfs_path, (char *)val);
+
+	free(sysfs_path);
+	if (rc != 0) {
+		return (EIO);
+	}
+
+	/*
+	 * Wait up to 30 seconds for sysfs power value to change after
+	 * writing it.
+	 */
+	timeout_ms = zpool_getenv_int("ZPOOL_POWER_ON_SLOT_TIMEOUT_MS", 30000);
+	for (int i = 0; i < MAX(1, timeout_ms / 200); i++) {
+		rc = zpool_power_current_state(zhp, vdev);
+		if (rc == (int)turn_on)
+			return (0);	/* success */
+
+		fsleep(0.200);	/* 200ms */
+	}
+
+	/* sysfs value never changed */
+	return (EBADE);
+}

cmd/zpool/zpool.d/ses

@@ -33,10 +33,18 @@ for i in $scripts ; do
 	val=""
 	case $i in
 	enc)
-		val=$(ls "$VDEV_ENC_SYSFS_PATH/../../" 2>/dev/null)
+		if echo "$VDEV_ENC_SYSFS_PATH" | grep -q '/sys/bus/pci/slots' ; then
+			val="$VDEV_ENC_SYSFS_PATH"
+		else
+			val="$(ls """$VDEV_ENC_SYSFS_PATH/../../""" 2>/dev/null)"
+		fi
 		;;
 	slot)
-		val=$(cat "$VDEV_ENC_SYSFS_PATH/slot" 2>/dev/null)
+		if echo "$VDEV_ENC_SYSFS_PATH" | grep -q '/sys/bus/pci/slots' ; then
+			val="$(basename """$VDEV_ENC_SYSFS_PATH""")"
+		else
+			val="$(cat """$VDEV_ENC_SYSFS_PATH/slot""" 2>/dev/null)"
+		fi
 		;;
 	encdev)
 		val=$(ls "$VDEV_ENC_SYSFS_PATH/../device/scsi_generic" 2>/dev/null)

cmd/zpool/zpool_iter.c

@@ -554,6 +554,10 @@ for_each_vdev_run_cb(void *zhp_data, nvlist_t *nv, void *cb_vcdl)
 	if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) != 0)
 		return (1);
 
+	/* Make sure we're getting the updated enclosure sysfs path */
+	update_vdev_config_dev_sysfs_path(nv, path,
+	    ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);
+
 	nvlist_lookup_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH,
 	    &vdev_enc_sysfs_path);
 

cmd/zpool/zpool_main.c

@@ -353,7 +353,7 @@ get_usage(zpool_help_t idx)
 		return (gettext("\tattach [-fsw] [-o property=value] "
 		    "<pool> <device> <new-device>\n"));
 	case HELP_CLEAR:
-		return (gettext("\tclear [-nF] <pool> [device]\n"));
+		return (gettext("\tclear [[--power]|[-nF]] <pool> [device]\n"));
 	case HELP_CREATE:
 		return (gettext("\tcreate [-fnd] [-o property=value] ... \n"
 		    "\t    [-O file-system-property=value] ... \n"
@@ -389,9 +389,11 @@ get_usage(zpool_help_t idx)
 		    "[-T d|u] [pool] ... \n"
 		    "\t    [interval [count]]\n"));
 	case HELP_OFFLINE:
-		return (gettext("\toffline [-f] [-t] <pool> <device> ...\n"));
+		return (gettext("\toffline [--power]|[[-f][-t]] <pool> "
+		    "<device> ...\n"));
 	case HELP_ONLINE:
-		return (gettext("\tonline [-e] <pool> <device> ...\n"));
+		return (gettext("\tonline [--power][-e] <pool> <device> "
+		    "...\n"));
 	case HELP_REPLACE:
 		return (gettext("\treplace [-fsw] [-o property=value] "
 		    "<pool> <device> [new-device]\n"));
@@ -410,7 +412,7 @@ get_usage(zpool_help_t idx)
 		return (gettext("\ttrim [-dw] [-r <rate>] [-c | -s] <pool> "
 		    "[<device> ...]\n"));
 	case HELP_STATUS:
-		return (gettext("\tstatus [-c [script1,script2,...]] "
+		return (gettext("\tstatus [--power] [-c [script1,script2,...]] "
 		    "[-igLpPstvxD]  [-T d|u] [pool] ... \n"
 		    "\t    [interval [count]]\n"));
 	case HELP_UPGRADE:
@@ -516,6 +518,77 @@ print_vdev_prop_cb(int prop, void *cb)
 	return (ZPROP_CONT);
 }
 
+/*
+ * Given a leaf vdev name like 'L5' return its VDEV_CONFIG_PATH like
+ * '/dev/disk/by-vdev/L5'.
+ */
+static const char *
+vdev_name_to_path(zpool_handle_t *zhp, char *vdev)
+{
+	nvlist_t *vdev_nv = zpool_find_vdev(zhp, vdev, NULL, NULL, NULL);
+	if (vdev_nv == NULL) {
+		return (NULL);
+	}
+	return (fnvlist_lookup_string(vdev_nv, ZPOOL_CONFIG_PATH));
+}
+
+static int
+zpool_power_on(zpool_handle_t *zhp, char *vdev)
+{
+	return (zpool_power(zhp, vdev, B_TRUE));
+}
+
+static int
+zpool_power_on_and_disk_wait(zpool_handle_t *zhp, char *vdev)
+{
+	int rc;
+
+	rc = zpool_power_on(zhp, vdev);
+	if (rc != 0)
+		return (rc);
+
+	zpool_disk_wait(vdev_name_to_path(zhp, vdev));
+
+	return (0);
+}
+
+static int
+zpool_power_on_pool_and_wait_for_devices(zpool_handle_t *zhp)
+{
+	nvlist_t *nv;
+	const char *path = NULL;
+	int rc;
+
+	/* Power up all the devices first */
+	FOR_EACH_REAL_LEAF_VDEV(zhp, nv) {
+		path = fnvlist_lookup_string(nv, ZPOOL_CONFIG_PATH);
+		if (path != NULL) {
+			rc = zpool_power_on(zhp, (char *)path);
+			if (rc != 0) {
+				return (rc);
+			}
+		}
+	}
+
+	/*
+	 * Wait for their devices to show up.  Since we powered them on
+	 * at roughly the same time, they should all come online around
+	 * the same time.
+	 */
+	FOR_EACH_REAL_LEAF_VDEV(zhp, nv) {
+		path = fnvlist_lookup_string(nv, ZPOOL_CONFIG_PATH);
+		zpool_disk_wait(path);
+	}
+
+	return (0);
+}
+
+static int
+zpool_power_off(zpool_handle_t *zhp, char *vdev)
+{
+	return (zpool_power(zhp, vdev, B_FALSE));
+}
+
 /*
  * Display usage message.  If we're inside a command, display only the usage for
  * that command.  Otherwise, iterate over the entire command table and display
@@ -2088,11 +2161,13 @@ typedef struct status_cbdata {
 	boolean_t	cb_explain;
 	boolean_t	cb_first;
 	boolean_t	cb_dedup_stats;
+	boolean_t	cb_print_unhealthy;
 	boolean_t	cb_print_status;
 	boolean_t	cb_print_slow_ios;
 	boolean_t	cb_print_vdev_init;
 	boolean_t	cb_print_vdev_trim;
 	vdev_cmd_data_list_t	*vcdl;
+	boolean_t	cb_print_power;
 } status_cbdata_t;
 
 /* Return 1 if string is NULL, empty, or whitespace; return 0 otherwise. */
@@ -2283,6 +2358,35 @@ health_str_to_color(const char *health)
 	return (NULL);
 }
 
+/*
+ * Called for each leaf vdev.  Returns 0 if the vdev is healthy.
+ * A vdev is unhealthy if any of the following are true:
+ * 1) there are read, write, or checksum errors,
+ * 2) its state is not ONLINE, or
+ * 3) slow IO reporting was requested (-s) and there are slow IOs.
+ */
+static int
+vdev_health_check_cb(void *hdl_data, nvlist_t *nv, void *data)
+{
+	status_cbdata_t *cb = data;
+	vdev_stat_t *vs;
+	uint_t vsc;
+	(void) hdl_data;
+
+	if (nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS,
+	    (uint64_t **)&vs, &vsc) != 0)
+		return (1);
+
+	if (vs->vs_checksum_errors || vs->vs_read_errors ||
+	    vs->vs_write_errors || vs->vs_state != VDEV_STATE_HEALTHY)
+		return (1);
+
+	if (cb->cb_print_slow_ios && vs->vs_slow_ios)
+		return (1);
+
+	return (0);
+}
+
 /*
  * Print out configuration state as requested by status_callback.
  */
@@ -2301,7 +2405,8 @@ print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
 	const char *state;
 	const char *type;
 	const char *path = NULL;
-	const char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL;
+	const char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL,
+	    *scolor = NULL;
 
 	if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN,
 	    &child, &children) != 0)
@@ -2328,6 +2433,15 @@ print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
 			state = gettext("AVAIL");
 	}
 
+	/*
+	 * If '-e' is specified then top-level vdevs and their children
+	 * can be pruned if all of their leaves are healthy.
+	 */
+	if (cb->cb_print_unhealthy && depth > 0 &&
+	    for_each_vdev_in_nvlist(nv, vdev_health_check_cb, cb) == 0) {
+		return;
+	}
+
 	printf_color(health_str_to_color(state),
 	    "\t%*s%-*s  %-8s", depth, "", cb->cb_namewidth - depth,
 	    name, state);
@@ -2342,6 +2456,9 @@ print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
 		if (vs->vs_checksum_errors)
 			ccolor = ANSI_RED;
 
+		if (vs->vs_slow_ios)
+			scolor = ANSI_BLUE;
+
 		if (cb->cb_literal) {
 			fputc(' ', stdout);
 			printf_color(rcolor, "%5llu",
@@ -2374,9 +2491,30 @@ print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name,
 			}
 
 			if (cb->cb_literal)
-				printf(" %5llu", (u_longlong_t)vs->vs_slow_ios);
+				printf_color(scolor, " %5llu",
+				    (u_longlong_t)vs->vs_slow_ios);
 			else
-				printf(" %5s", rbuf);
+				printf_color(scolor, " %5s", rbuf);
+		}
+		if (cb->cb_print_power) {
+			if (children == 0)  {
+				/* Only leaf vdevs have physical slots */
+				switch (zpool_power_current_state(zhp, (char *)
+				    fnvlist_lookup_string(nv,
+				    ZPOOL_CONFIG_PATH))) {
+				case 0:
+					printf_color(ANSI_RED, " %5s",
+					    gettext("off"));
+					break;
+				case 1:
+					printf(" %5s", gettext("on"));
+					break;
+				default:
+					printf(" %5s", "-");
+				}
+			} else {
+				printf(" %5s", "-");
+			}
 		}
 	}
 
@@ -5428,19 +5566,6 @@ get_interval_count_filter_guids(int *argc, char **argv, float *interval,
 	    interval, count);
 }
 
-/*
- * Floating point sleep().  Allows you to pass in a floating point value for
- * seconds.
- */
-static void
-fsleep(float sec)
-{
-	struct timespec req;
-	req.tv_sec = floor(sec);
-	req.tv_nsec = (sec - (float)req.tv_sec) * NANOSEC;
-	nanosleep(&req, NULL);
-}
-
 /*
  * Terminal height, in rows. Returns -1 if stdout is not connected to a TTY or
  * if we were unable to determine its size.
@@ -6939,10 +7064,12 @@ zpool_do_split(int argc, char **argv)
 	return (ret);
 }
 
-
+#define	POWER_OPT 1024
 
 /*
- * zpool online <pool> <device> ...
+ * zpool online [--power] <pool> <device> ...
+ *
+ * --power: Power on the enclosure slot to the drive (if possible)
  */
 int
 zpool_do_online(int argc, char **argv)
@@ -6953,13 +7080,21 @@ zpool_do_online(int argc, char **argv)
 	int ret = 0;
 	vdev_state_t newstate;
 	int flags = 0;
+	boolean_t is_power_on = B_FALSE;
+	struct option long_options[] = {
+		{"power", no_argument, NULL, POWER_OPT},
+		{0, 0, 0, 0}
+	};
 
 	/* check options */
-	while ((c = getopt(argc, argv, "e")) != -1) {
+	while ((c = getopt_long(argc, argv, "e", long_options, NULL)) != -1) {
 		switch (c) {
 		case 'e':
 			flags |= ZFS_ONLINE_EXPAND;
 			break;
+		case POWER_OPT:
+			is_power_on = B_TRUE;
+			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -6967,6 +7102,9 @@ zpool_do_online(int argc, char **argv)
 		}
 	}
 
+	if (libzfs_envvar_is_set("ZPOOL_AUTO_POWER_ON_SLOT"))
+		is_power_on = B_TRUE;
+
 	argc -= optind;
 	argv += optind;
 
@@ -6988,6 +7126,18 @@ zpool_do_online(int argc, char **argv)
 	for (i = 1; i < argc; i++) {
 		vdev_state_t oldstate;
 		boolean_t avail_spare, l2cache;
+		int rc;
+
+		if (is_power_on) {
+			rc = zpool_power_on_and_disk_wait(zhp, argv[i]);
+			if (rc == ENOTSUP) {
+				(void) fprintf(stderr,
+				    gettext("Power control not supported\n"));
+			}
+			if (rc != 0)
+				return (rc);
+		}
+
 		nvlist_t *tgt = zpool_find_vdev(zhp, argv[i], &avail_spare,
 		    &l2cache, NULL);
 		if (tgt == NULL) {
@@ -7033,12 +7183,15 @@ zpool_do_online(int argc, char **argv)
 }
 
 /*
- * zpool offline [-ft] <pool> <device> ...
+ * zpool offline [-ft]|[--power] <pool> <device> ...
+ *
  *
  *	-f	Force the device into a faulted state.
  *
  *	-t	Only take the device off-line temporarily.  The offline/faulted
  *		state will not be persistent across reboots.
+ *
+ *	--power Power off the enclosure slot to the drive (if possible)
  */
 int
 zpool_do_offline(int argc, char **argv)
@@ -7049,9 +7202,15 @@ zpool_do_offline(int argc, char **argv)
 	int ret = 0;
 	boolean_t istmp = B_FALSE;
 	boolean_t fault = B_FALSE;
+	boolean_t is_power_off = B_FALSE;
+
+	struct option long_options[] = {
+		{"power", no_argument, NULL, POWER_OPT},
+		{0, 0, 0, 0}
+	};
 
 	/* check options */
-	while ((c = getopt(argc, argv, "ft")) != -1) {
+	while ((c = getopt_long(argc, argv, "ft", long_options, NULL)) != -1) {
 		switch (c) {
 		case 'f':
 			fault = B_TRUE;
@@ -7059,6 +7218,9 @@ zpool_do_offline(int argc, char **argv)
 		case 't':
 			istmp = B_TRUE;
 			break;
+		case POWER_OPT:
+			is_power_off = B_TRUE;
+			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -7066,6 +7228,20 @@ zpool_do_offline(int argc, char **argv)
 		}
 	}
 
+	if (is_power_off && fault) {
+		(void) fprintf(stderr,
+		    gettext("-0 and -f cannot be used together\n"));
+		usage(B_FALSE);
+		return (1);
+	}
+
+	if (is_power_off && istmp) {
+		(void) fprintf(stderr,
+		    gettext("-0 and -t cannot be used together\n"));
+		usage(B_FALSE);
+		return (1);
+	}
+
 	argc -= optind;
 	argv += optind;
 
@@ -7085,8 +7261,22 @@ zpool_do_offline(int argc, char **argv)
 		return (1);
 
 	for (i = 1; i < argc; i++) {
-		if (fault) {
-			uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]);
+		uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]);
+		if (is_power_off) {
+			/*
+			 * Note: we have to power off first, then set REMOVED,
+			 * or else zpool_vdev_set_removed_state() returns
+			 * EAGAIN.
+			 */
+			ret = zpool_power_off(zhp, argv[i]);
+			if (ret != 0) {
+				(void) fprintf(stderr, "%s %s %d\n",
+				    gettext("unable to power off slot for"),
+				    argv[i], ret);
+			}
+			zpool_vdev_set_removed_state(zhp, guid, VDEV_AUX_NONE);
+
+		} else if (fault) {
 			vdev_aux_t aux;
 			if (istmp == B_FALSE) {
 				/* Force the fault to persist across imports */
@@ -7109,7 +7299,7 @@ zpool_do_offline(int argc, char **argv)
 }
 
 /*
- * zpool clear <pool> [device]
+ * zpool clear [-nF]|[--power] <pool> [device]
  *
  * Clear all errors associated with a pool or a particular device.
  */
@@ -7121,13 +7311,20 @@ zpool_do_clear(int argc, char **argv)
 	boolean_t dryrun = B_FALSE;
 	boolean_t do_rewind = B_FALSE;
 	boolean_t xtreme_rewind = B_FALSE;
+	boolean_t is_power_on = B_FALSE;
 	uint32_t rewind_policy = ZPOOL_NO_REWIND;
 	nvlist_t *policy = NULL;
 	zpool_handle_t *zhp;
 	char *pool, *device;
 
+	struct option long_options[] = {
+		{"power", no_argument, NULL, POWER_OPT},
+		{0, 0, 0, 0}
+	};
+
 	/* check options */
-	while ((c = getopt(argc, argv, "FnX")) != -1) {
+	while ((c = getopt_long(argc, argv, "FnX", long_options,
+	    NULL)) != -1) {
 		switch (c) {
 		case 'F':
 			do_rewind = B_TRUE;
@@ -7138,6 +7335,9 @@ zpool_do_clear(int argc, char **argv)
 		case 'X':
 			xtreme_rewind = B_TRUE;
 			break;
+		case POWER_OPT:
+			is_power_on = B_TRUE;
+			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -7145,6 +7345,9 @@ zpool_do_clear(int argc, char **argv)
 		}
 	}
 
+	if (libzfs_envvar_is_set("ZPOOL_AUTO_POWER_ON_SLOT"))
+		is_power_on = B_TRUE;
+
 	argc -= optind;
 	argv += optind;
 
@@ -7185,6 +7388,14 @@ zpool_do_clear(int argc, char **argv)
 		return (1);
 	}
 
+	if (is_power_on) {
+		if (device == NULL) {
+			zpool_power_on_pool_and_wait_for_devices(zhp);
+		} else {
+			zpool_power_on_and_disk_wait(zhp, device);
+		}
+	}
+
 	if (zpool_clear(zhp, device, policy) != 0)
 		ret = 1;
 
@@ -8801,6 +9012,10 @@ status_callback(zpool_handle_t *zhp, void *data)
 			printf_color(ANSI_BOLD, " %5s", gettext("SLOW"));
 		}
 
+		if (cbp->cb_print_power) {
+			printf_color(ANSI_BOLD, " %5s", gettext("POWER"));
+		}
+
 		if (cbp->vcdl != NULL)
 			print_cmd_columns(cbp->vcdl, 0);
 
@@ -8828,9 +9043,11 @@ status_callback(zpool_handle_t *zhp, void *data)
 				(void) printf(gettext(
 				    "errors: No known data errors\n"));
 			} else if (!cbp->cb_verbose) {
+				color_start(ANSI_RED);
 				(void) printf(gettext("errors: %llu data "
 				    "errors, use '-v' for a list\n"),
 				    (u_longlong_t)nerr);
+				color_end();
 			} else {
 				print_error_log(zhp);
 			}
@@ -8847,10 +9064,11 @@ status_callback(zpool_handle_t *zhp, void *data)
 }
 
 /*
- * zpool status [-c [script1,script2,...]] [-igLpPstvx] [-T d|u] [pool] ...
- *              [interval [count]]
+ * zpool status [-c [script1,script2,...]] [-igLpPstvx] [--power] [-T d|u] ...
+ *              [pool] [interval [count]]
  *
  *	-c CMD	For each vdev, run command CMD
+ *	-e	Display only unhealthy vdevs
  *	-i	Display vdev initialization status.
  *	-g	Display guid for individual vdev name.
  *	-L	Follow links when resolving vdev path name.
@@ -8862,6 +9080,7 @@ status_callback(zpool_handle_t *zhp, void *data)
  *	-D	Display dedup status (undocumented)
  *	-t	Display vdev TRIM status.
  *	-T	Display a timestamp in date(1) or Unix format
+ *	--power	Display vdev enclosure slot power status
  *
  * Describes the health status of all pools or some subset.
  */
@@ -8875,8 +9094,14 @@ zpool_do_status(int argc, char **argv)
 	status_cbdata_t cb = { 0 };
 	char *cmd = NULL;
 
+	struct option long_options[] = {
+		{"power", no_argument, NULL, POWER_OPT},
+		{0, 0, 0, 0}
+	};
+
 	/* check options */
-	while ((c = getopt(argc, argv, "c:igLpPsvxDtT:")) != -1) {
+	while ((c = getopt_long(argc, argv, "c:eigLpPsvxDtT:", long_options,
+	    NULL)) != -1) {
 		switch (c) {
 		case 'c':
 			if (cmd != NULL) {
@@ -8902,6 +9127,9 @@ zpool_do_status(int argc, char **argv)
 			}
 			cmd = optarg;
 			break;
+		case 'e':
+			cb.cb_print_unhealthy = B_TRUE;
+			break;
 		case 'i':
 			cb.cb_print_vdev_init = B_TRUE;
 			break;
@@ -8935,6 +9163,9 @@ zpool_do_status(int argc, char **argv)
 		case 'T':
 			get_timestamp_arg(*optarg);
 			break;
+		case POWER_OPT:
+			cb.cb_print_power = B_TRUE;
+			break;
 		case '?':
 			if (optopt == 'c') {
 				print_zpool_script_list("status");
@@ -10752,6 +10983,9 @@ print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row)
 		col_widths[i] = MAX(strlen(headers[i]), 6) + 2;
 	}
 
+	if (timestamp_fmt != NODATE)
+		print_timestamp(timestamp_fmt);
+
 	/* Print header if appropriate */
 	int term_height = terminal_height();
 	boolean_t reprint_header = (!wd->wd_headers_once && term_height > 0 &&
@@ -10819,9 +11053,6 @@ print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row)
 	if (vdev_any_spare_replacing(nvroot))
 		bytes_rem[ZPOOL_WAIT_REPLACE] =  bytes_rem[ZPOOL_WAIT_RESILVER];
 
-	if (timestamp_fmt != NODATE)
-		print_timestamp(timestamp_fmt);
-
 	for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) {
 		char buf[64];
 		if (!wd->wd_enabled[i])

cmd/zpool/zpool_util.h

@@ -138,6 +138,9 @@ int check_file(const char *file, boolean_t force, boolean_t isspare);
 void after_zpool_upgrade(zpool_handle_t *zhp);
 int check_file_generic(const char *file, boolean_t force, boolean_t isspare);
 
+int zpool_power(zpool_handle_t *zhp, char *vdev, boolean_t turn_on);
+int zpool_power_current_state(zpool_handle_t *zhp, char *vdev);
+
 #ifdef	__cplusplus
 }
 #endif

cmd/zpool/zpool_vdev.c

@@ -372,6 +372,10 @@ make_leaf_vdev(nvlist_t *props, const char *arg, boolean_t is_primary)
 	verify(nvlist_add_string(vdev, ZPOOL_CONFIG_PATH, path) == 0);
 	verify(nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE, type) == 0);
 
+	/* Lookup and add the enclosure sysfs path (if exists) */
+	update_vdev_config_dev_sysfs_path(vdev, path,
+	    ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH);
+
 	if (strcmp(type, VDEV_TYPE_DISK) == 0)
 		verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_WHOLE_DISK,
 		    (uint64_t)wholedisk) == 0);

config/Rules.am

@@ -42,21 +42,6 @@ AM_CPPFLAGS += $(DEBUG_CPPFLAGS)
 AM_CPPFLAGS += $(CODE_COVERAGE_CPPFLAGS)
 AM_CPPFLAGS += -DTEXT_DOMAIN=\"zfs-@ac_system_l@-user\"
 
-AM_CPPFLAGS_NOCHECK  = -D"strtok(...)=strtok(__VA_ARGS__) __attribute__((deprecated(\"Use strtok_r(3) instead!\")))"
-AM_CPPFLAGS_NOCHECK += -D"__xpg_basename(...)=__xpg_basename(__VA_ARGS__) __attribute__((deprecated(\"basename(3) is underspecified. Use zfs_basename() instead!\")))"
-AM_CPPFLAGS_NOCHECK += -D"basename(...)=basename(__VA_ARGS__) __attribute__((deprecated(\"basename(3) is underspecified. Use zfs_basename() instead!\")))"
-AM_CPPFLAGS_NOCHECK += -D"dirname(...)=dirname(__VA_ARGS__) __attribute__((deprecated(\"dirname(3) is underspecified. Use zfs_dirnamelen() instead!\")))"
-AM_CPPFLAGS_NOCHECK += -D"bcopy(...)=__attribute__((deprecated(\"bcopy(3) is deprecated. Use memcpy(3)/memmove(3) instead!\"))) bcopy(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"bcmp(...)=__attribute__((deprecated(\"bcmp(3) is deprecated. Use memcmp(3) instead!\"))) bcmp(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"bzero(...)=__attribute__((deprecated(\"bzero(3) is deprecated. Use memset(3) instead!\"))) bzero(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"asctime(...)=__attribute__((deprecated(\"Use strftime(3) instead!\"))) asctime(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"asctime_r(...)=__attribute__((deprecated(\"Use strftime(3) instead!\"))) asctime_r(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"gmtime(...)=__attribute__((deprecated(\"gmtime(3) isn't thread-safe. Use gmtime_r(3) instead!\"))) gmtime(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"localtime(...)=__attribute__((deprecated(\"localtime(3) isn't thread-safe. Use localtime_r(3) instead!\"))) localtime(__VA_ARGS__)"
-AM_CPPFLAGS_NOCHECK += -D"strncpy(...)=__attribute__((deprecated(\"strncpy(3) is deprecated. Use strlcpy(3) instead!\"))) strncpy(__VA_ARGS__)"
-
-AM_CPPFLAGS += $(AM_CPPFLAGS_NOCHECK)
-
 if ASAN_ENABLED
 AM_CPPFLAGS += -DZFS_ASAN_ENABLED
 endif

config/kernel-acl.m4

@@ -172,7 +172,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_get_acl], [
 		#include <linux/fs.h>
 
-		struct posix_acl *get_acl_fn(struct inode *inode, int type)
+		static struct posix_acl *get_acl_fn(struct inode *inode, int type)
 		    { return NULL; }
 
 		static const struct inode_operations
@@ -184,7 +184,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_get_acl_rcu], [
 		#include <linux/fs.h>
 
-		struct posix_acl *get_acl_fn(struct inode *inode, int type,
+		static struct posix_acl *get_acl_fn(struct inode *inode, int type,
 		    bool rcu) { return NULL; }
 
 		static const struct inode_operations
@@ -196,7 +196,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_GET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_get_inode_acl], [
 		#include <linux/fs.h>
 
-		struct posix_acl *get_inode_acl_fn(struct inode *inode, int type,
+		static struct posix_acl *get_inode_acl_fn(struct inode *inode, int type,
 		    bool rcu) { return NULL; }
 
 		static const struct inode_operations
@@ -243,7 +243,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_set_acl_mnt_idmap_dentry], [
 		#include <linux/fs.h>
 
-		int set_acl_fn(struct mnt_idmap *idmap,
+		static int set_acl_fn(struct mnt_idmap *idmap,
 		    struct dentry *dent, struct posix_acl *acl,
 		    int type) { return 0; }
 
@@ -255,7 +255,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns_dentry], [
 		#include <linux/fs.h>
 
-		int set_acl_fn(struct user_namespace *userns,
+		static int set_acl_fn(struct user_namespace *userns,
 		    struct dentry *dent, struct posix_acl *acl,
 		    int type) { return 0; }
 
@@ -267,7 +267,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_set_acl_userns], [
 		#include <linux/fs.h>
 
-		int set_acl_fn(struct user_namespace *userns,
+		static int set_acl_fn(struct user_namespace *userns,
 		    struct inode *inode, struct posix_acl *acl,
 		    int type) { return 0; }
 
@@ -279,7 +279,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_OPERATIONS_SET_ACL], [
 	ZFS_LINUX_TEST_SRC([inode_operations_set_acl], [
 		#include <linux/fs.h>
 
-		int set_acl_fn(struct inode *inode, struct posix_acl *acl,
+		static int set_acl_fn(struct inode *inode, struct posix_acl *acl,
 		    int type) { return 0; }
 
 		static const struct inode_operations

config/kernel-automount.m4

@@ -8,7 +8,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_AUTOMOUNT], [
 	ZFS_LINUX_TEST_SRC([dentry_operations_d_automount], [
 		#include <linux/dcache.h>
-		struct vfsmount *d_automount(struct path *p) { return NULL; }
+		static struct vfsmount *d_automount(struct path *p) { return NULL; }
 		struct dentry_operations dops __attribute__ ((unused)) = {
 			.d_automount = d_automount,
 		};

config/kernel-bio.m4

@@ -247,7 +247,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_BIO_END_IO_T_ARGS], [
 	ZFS_LINUX_TEST_SRC([bio_end_io_t_args], [
 		#include <linux/bio.h>
-		void wanted_end_io(struct bio *bio) { return; }
+		static void wanted_end_io(struct bio *bio) { return; }
 		bio_end_io_t *end_io __attribute__ ((unused)) = wanted_end_io;
 	], [])
 ])

config/kernel-blkdev.m4

@@ -35,6 +35,25 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_GET_BY_PATH_4ARG], [
 	])
 ])
 
+dnl #
+dnl # 6.8.x API change
+dnl # bdev_open_by_path() replaces blkdev_get_by_path()
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_OPEN_BY_PATH], [
+	ZFS_LINUX_TEST_SRC([bdev_open_by_path], [
+		#include <linux/fs.h>
+		#include <linux/blkdev.h>
+	], [
+		struct bdev_handle *bdh __attribute__ ((unused)) = NULL;
+		const char *path = "path";
+		fmode_t mode = 0;
+		void *holder = NULL;
+		struct blk_holder_ops h;
+
+		bdh = bdev_open_by_path(path, mode, holder, &h);
+	])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_GET_BY_PATH], [
 	AC_MSG_CHECKING([whether blkdev_get_by_path() exists and takes 3 args])
 	ZFS_LINUX_TEST_RESULT([blkdev_get_by_path], [
@@ -47,7 +66,15 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_GET_BY_PATH], [
 				[blkdev_get_by_path() exists and takes 4 args])
 			AC_MSG_RESULT(yes)
 		], [
-			ZFS_LINUX_TEST_ERROR([blkdev_get_by_path()])
+			AC_MSG_RESULT(no)
+			AC_MSG_CHECKING([whether bdev_open_by_path() exists])
+			ZFS_LINUX_TEST_RESULT([bdev_open_by_path], [
+				AC_DEFINE(HAVE_BDEV_OPEN_BY_PATH, 1,
+					[bdev_open_by_path() exists])
+				AC_MSG_RESULT(yes)
+			], [
+				ZFS_LINUX_TEST_ERROR([blkdev_get_by_path()])
+			])
 		])
 	])
 ])
@@ -108,18 +135,41 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_PUT_HOLDER], [
 	])
 ])
 
+dnl #
+dnl # 6.8.x API change
+dnl # bdev_release() replaces blkdev_put()
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_RELEASE], [
+	ZFS_LINUX_TEST_SRC([bdev_release], [
+		#include <linux/fs.h>
+		#include <linux/blkdev.h>
+	], [
+		struct bdev_handle *bdh = NULL;
+		bdev_release(bdh);
+	])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_PUT], [
 	AC_MSG_CHECKING([whether blkdev_put() exists])
 	ZFS_LINUX_TEST_RESULT([blkdev_put], [
 		AC_MSG_RESULT(yes)
 	], [
+		AC_MSG_RESULT(no)
 		AC_MSG_CHECKING([whether blkdev_put() accepts void* as arg 2])
 		ZFS_LINUX_TEST_RESULT([blkdev_put_holder], [
 			AC_MSG_RESULT(yes)
 			AC_DEFINE(HAVE_BLKDEV_PUT_HOLDER, 1,
 				[blkdev_put() accepts void* as arg 2])
 		], [
-			ZFS_LINUX_TEST_ERROR([blkdev_put()])
+			AC_MSG_RESULT(no)
+			AC_MSG_CHECKING([whether bdev_release() exists])
+			ZFS_LINUX_TEST_RESULT([bdev_release], [
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_BDEV_RELEASE, 1,
+					[bdev_release() exists])
+			], [
+				ZFS_LINUX_TEST_ERROR([blkdev_put()])
+			])
 		])
 	])
 ])
@@ -474,6 +524,7 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_BDEVNAME], [
 
 dnl #
 dnl # 5.19 API: blkdev_issue_secure_erase()
+dnl # 4.7  API: __blkdev_issue_discard(..., BLKDEV_DISCARD_SECURE)
 dnl # 3.10 API: blkdev_issue_discard(..., BLKDEV_DISCARD_SECURE)
 dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_ISSUE_SECURE_ERASE], [
@@ -489,6 +540,20 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV_ISSUE_SECURE_ERASE], [
 		    sector, nr_sects, GFP_KERNEL);
 	])
 
+	ZFS_LINUX_TEST_SRC([blkdev_issue_discard_async_flags], [
+		#include <linux/blkdev.h>
+	],[
+		struct block_device *bdev = NULL;
+		sector_t sector = 0;
+		sector_t nr_sects = 0;
+		unsigned long flags = 0;
+		struct bio *biop = NULL;
+		int error __attribute__ ((unused));
+
+		error = __blkdev_issue_discard(bdev,
+		    sector, nr_sects, GFP_KERNEL, flags, &biop);
+	])
+
 	ZFS_LINUX_TEST_SRC([blkdev_issue_discard_flags], [
 		#include <linux/blkdev.h>
 	],[
@@ -512,13 +577,22 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_ISSUE_SECURE_ERASE], [
 	],[
 		AC_MSG_RESULT(no)
 
-		AC_MSG_CHECKING([whether blkdev_issue_discard() is available])
-		ZFS_LINUX_TEST_RESULT([blkdev_issue_discard_flags], [
+		AC_MSG_CHECKING([whether __blkdev_issue_discard() is available])
+		ZFS_LINUX_TEST_RESULT([blkdev_issue_discard_async_flags], [
 			AC_MSG_RESULT(yes)
-			AC_DEFINE(HAVE_BLKDEV_ISSUE_DISCARD, 1,
-			    [blkdev_issue_discard() is available])
+			AC_DEFINE(HAVE_BLKDEV_ISSUE_DISCARD_ASYNC, 1,
+			    [__blkdev_issue_discard() is available])
 		],[
-			ZFS_LINUX_TEST_ERROR([blkdev_issue_discard()])
+			AC_MSG_RESULT(no)
+
+			AC_MSG_CHECKING([whether blkdev_issue_discard() is available])
+			ZFS_LINUX_TEST_RESULT([blkdev_issue_discard_flags], [
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_BLKDEV_ISSUE_DISCARD, 1,
+					[blkdev_issue_discard() is available])
+			],[
+				ZFS_LINUX_TEST_ERROR([blkdev_issue_discard()])
+			])
 		])
 	])
 ])
@@ -570,8 +644,10 @@ AC_DEFUN([ZFS_AC_KERNEL_BLKDEV_BLK_STS_RESV_CONFLICT], [
 AC_DEFUN([ZFS_AC_KERNEL_SRC_BLKDEV], [
 	ZFS_AC_KERNEL_SRC_BLKDEV_GET_BY_PATH
 	ZFS_AC_KERNEL_SRC_BLKDEV_GET_BY_PATH_4ARG
+	ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_OPEN_BY_PATH
 	ZFS_AC_KERNEL_SRC_BLKDEV_PUT
 	ZFS_AC_KERNEL_SRC_BLKDEV_PUT_HOLDER
+	ZFS_AC_KERNEL_SRC_BLKDEV_BDEV_RELEASE
 	ZFS_AC_KERNEL_SRC_BLKDEV_REREAD_PART
 	ZFS_AC_KERNEL_SRC_BLKDEV_INVALIDATE_BDEV
 	ZFS_AC_KERNEL_SRC_BLKDEV_LOOKUP_BDEV

config/kernel-block-device-operations.m4

@@ -5,7 +5,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLOCK_DEVICE_OPERATIONS_CHECK_EVENTS], [
 	ZFS_LINUX_TEST_SRC([block_device_operations_check_events], [
 		#include <linux/blkdev.h>
 
-		unsigned int blk_check_events(struct gendisk *disk,
+		static unsigned int blk_check_events(struct gendisk *disk,
 		    unsigned int clearing) {
 			(void) disk, (void) clearing;
 			return (0);
@@ -34,7 +34,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLOCK_DEVICE_OPERATIONS_RELEASE_VOID], [
 	ZFS_LINUX_TEST_SRC([block_device_operations_release_void], [
 		#include <linux/blkdev.h>
 
-		void blk_release(struct gendisk *g, fmode_t mode) {
+		static void blk_release(struct gendisk *g, fmode_t mode) {
 			(void) g, (void) mode;
 			return;
 		}
@@ -56,7 +56,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLOCK_DEVICE_OPERATIONS_RELEASE_1ARG], [
 	ZFS_LINUX_TEST_SRC([block_device_operations_release_void_1arg], [
 		#include <linux/blkdev.h>
 
-		void blk_release(struct gendisk *g) {
+		static void blk_release(struct gendisk *g) {
 			(void) g;
 			return;
 		}
@@ -96,7 +96,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK], [
 	ZFS_LINUX_TEST_SRC([block_device_operations_revalidate_disk], [
 		#include <linux/blkdev.h>
 
-		int blk_revalidate_disk(struct gendisk *disk) {
+		static int blk_revalidate_disk(struct gendisk *disk) {
 			(void) disk;
 			return(0);
 		}

config/kernel-commit-metadata.m4

@@ -7,7 +7,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_COMMIT_METADATA], [
 	ZFS_LINUX_TEST_SRC([export_operations_commit_metadata], [
 		#include <linux/exportfs.h>
-		int commit_metadata(struct inode *inode) { return 0; }
+		static int commit_metadata(struct inode *inode) { return 0; }
 		static struct export_operations eops __attribute__ ((unused))={
 			.commit_metadata = commit_metadata,
 		};

config/kernel-current-time.m4

@@ -2,12 +2,15 @@ dnl #
 dnl # 4.9, current_time() added
 dnl # 4.18, return type changed from timespec to timespec64
 dnl #
+dnl # Note that we don't care about the return type in this check. If we have
+dnl # to implement a fallback, we'll know we're <4.9, which was timespec.
+dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_CURRENT_TIME], [
 	ZFS_LINUX_TEST_SRC([current_time], [
 		#include <linux/fs.h>
 	], [
 		struct inode ip __attribute__ ((unused));
-		ip.i_atime = current_time(&ip);
+		(void) current_time(&ip);
 	])
 ])
 

config/kernel-dentry-operations.m4

@@ -98,7 +98,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_D_REVALIDATE_NAMEIDATA], [
 		#include <linux/dcache.h>
 		#include <linux/sched.h>
 
-		int revalidate (struct dentry *dentry,
+		static int revalidate (struct dentry *dentry,
 		    struct nameidata *nidata) { return 0; }
 
 		static const struct dentry_operations

config/kernel-dirty-inode.m4

@@ -8,7 +8,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_DIRTY_INODE], [
 	ZFS_LINUX_TEST_SRC([dirty_inode_with_flags], [
 		#include <linux/fs.h>
 
-		void dirty_inode(struct inode *a, int b) { return; }
+		static void dirty_inode(struct inode *a, int b) { return; }
 
 		static const struct super_operations
 		    sops __attribute__ ((unused)) = {

config/kernel-encode-fh-inode.m4

@@ -7,7 +7,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_ENCODE_FH_WITH_INODE], [
 	ZFS_LINUX_TEST_SRC([export_operations_encode_fh], [
 		#include <linux/exportfs.h>
-		int encode_fh(struct inode *inode, __u32 *fh, int *max_len,
+		static int encode_fh(struct inode *inode, __u32 *fh, int *max_len,
 		              struct inode *parent) { return 0; }
 		static struct export_operations eops __attribute__ ((unused))={
 			.encode_fh = encode_fh,

config/kernel-evict-inode.m4

@@ -6,7 +6,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_EVICT_INODE], [
 	ZFS_LINUX_TEST_SRC([evict_inode], [
 		#include <linux/fs.h>
-		void evict_inode (struct inode * t) { return; }
+		static void evict_inode (struct inode * t) { return; }
 		static struct super_operations sops __attribute__ ((unused)) = {
 			.evict_inode = evict_inode,
 		};

config/kernel-fallocate.m4

@@ -11,7 +11,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_FALLOCATE], [
 	ZFS_LINUX_TEST_SRC([file_fallocate], [
 		#include <linux/fs.h>
 
-		long test_fallocate(struct file *file, int mode,
+		static long test_fallocate(struct file *file, int mode,
 		    loff_t offset, loff_t len) { return 0; }
 
 		static const struct file_operations

config/kernel-flush_dcache_page.m4

@@ -1,7 +1,8 @@
 dnl #
 dnl # Starting from Linux 5.13, flush_dcache_page() becomes an inline
-dnl # function and may indirectly referencing GPL-only cpu_feature_keys on
-dnl # powerpc
+dnl # function and may indirectly referencing GPL-only symbols:
+dnl # on powerpc: cpu_feature_keys
+dnl # on riscv: PageHuge (added from 6.2)
 dnl #
 
 dnl #

config/kernel-fpu.m4

@@ -79,6 +79,12 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_FPU], [
 		__kernel_fpu_end();
 	], [], [ZFS_META_LICENSE])
 
+	ZFS_LINUX_TEST_SRC([kernel_neon], [
+		#include <asm/neon.h>
+	], [
+		kernel_neon_begin();
+		kernel_neon_end();
+	], [], [ZFS_META_LICENSE])
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_FPU], [
@@ -105,9 +111,20 @@ AC_DEFUN([ZFS_AC_KERNEL_FPU], [
 			AC_DEFINE(KERNEL_EXPORTS_X86_FPU, 1,
 			    [kernel exports FPU functions])
 		],[
-			AC_MSG_RESULT(internal)
-			AC_DEFINE(HAVE_KERNEL_FPU_INTERNAL, 1,
-			    [kernel fpu internal])
+			dnl #
+			dnl # ARM neon symbols (only on arm and arm64)
+			dnl # could be GPL-only on arm64 after Linux 6.2
+			dnl #
+			ZFS_LINUX_TEST_RESULT([kernel_neon_license],[
+				AC_MSG_RESULT(kernel_neon_*)
+				AC_DEFINE(HAVE_KERNEL_NEON, 1,
+				    [kernel has kernel_neon_* functions])
+			],[
+				# catch-all
+				AC_MSG_RESULT(internal)
+				AC_DEFINE(HAVE_KERNEL_FPU_INTERNAL, 1,
+				    [kernel fpu internal])
+			])
 		])
 	])
 ])

config/kernel-fsync.m4

@@ -5,7 +5,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_FSYNC], [
 	ZFS_LINUX_TEST_SRC([fsync_without_dentry], [
 		#include <linux/fs.h>
 
-		int test_fsync(struct file *f, int x) { return 0; }
+		static int test_fsync(struct file *f, int x) { return 0; }
 
 		static const struct file_operations
 		    fops __attribute__ ((unused)) = {
@@ -16,7 +16,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_FSYNC], [
 	ZFS_LINUX_TEST_SRC([fsync_range], [
 		#include <linux/fs.h>
 
-		int test_fsync(struct file *f, loff_t a, loff_t b, int c)
+		static int test_fsync(struct file *f, loff_t a, loff_t b, int c)
 		    { return 0; }
 
 		static const struct file_operations

config/kernel-get-link.m4

@@ -5,7 +5,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_GET_LINK], [
 	ZFS_LINUX_TEST_SRC([inode_operations_get_link], [
 		#include <linux/fs.h>
-		const char *get_link(struct dentry *de, struct inode *ip,
+		static const char *get_link(struct dentry *de, struct inode *ip,
 		    struct delayed_call *done) { return "symlink"; }
 		static struct inode_operations
 		     iops __attribute__ ((unused)) = {
@@ -15,7 +15,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_GET_LINK], [
 
 	ZFS_LINUX_TEST_SRC([inode_operations_get_link_cookie], [
 		#include <linux/fs.h>
-		const char *get_link(struct dentry *de, struct
+		static const char *get_link(struct dentry *de, struct
 		    inode *ip, void **cookie) { return "symlink"; }
 		static struct inode_operations
 		     iops __attribute__ ((unused)) = {
@@ -25,7 +25,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_GET_LINK], [
 
 	ZFS_LINUX_TEST_SRC([inode_operations_follow_link], [
 		#include <linux/fs.h>
-		const char *follow_link(struct dentry *de,
+		static const char *follow_link(struct dentry *de,
 		    void **cookie) { return "symlink"; }
 		static struct inode_operations
 		    iops __attribute__ ((unused)) = {
@@ -35,7 +35,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_GET_LINK], [
 
 	ZFS_LINUX_TEST_SRC([inode_operations_follow_link_nameidata], [
 		#include <linux/fs.h>
-		void *follow_link(struct dentry *de, struct
+		static void *follow_link(struct dentry *de, struct
 		    nameidata *nd) { return (void *)NULL; }
 		static struct inode_operations
 		    iops __attribute__ ((unused)) = {

config/kernel-idmap_mnt_api.m4

@@ -23,3 +23,28 @@ AC_DEFUN([ZFS_AC_KERNEL_IDMAP_MNT_API], [
         ])
 ])
 
+dnl #
+dnl # 6.8 decouples mnt_idmap from user_namespace. This is all internal
+dnl # to mnt_idmap so we can't detect it directly, but we detect a related
+dnl # change as use that as a signal.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_IDMAP_NO_USERNS], [
+	ZFS_LINUX_TEST_SRC([idmap_no_userns], [
+		#include <linux/uidgid.h>
+	], [
+		struct uid_gid_map *map = NULL;
+		map_id_down(map, 0);
+	])
+])
+
+
+AC_DEFUN([ZFS_AC_KERNEL_IDMAP_NO_USERNS], [
+	AC_MSG_CHECKING([whether idmapped mounts have a user namespace])
+	ZFS_LINUX_TEST_RESULT([idmap_no_userns], [
+		AC_MSG_RESULT([yes])
+		AC_DEFINE(HAVE_IDMAP_NO_USERNS, 1,
+			[mnt_idmap does not have user_namespace])
+	], [
+		AC_MSG_RESULT([no])
+	])
+])

config/kernel-inode-create.m4

@@ -7,7 +7,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_CREATE], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int inode_create(struct mnt_idmap *idmap,
+		static int inode_create(struct mnt_idmap *idmap,
 		    struct inode *inode ,struct dentry *dentry,
 		    umode_t umode, bool flag) { return 0; }
 
@@ -25,7 +25,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_CREATE], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int inode_create(struct user_namespace *userns,
+		static int inode_create(struct user_namespace *userns,
 		    struct inode *inode ,struct dentry *dentry,
 		    umode_t umode, bool flag) { return 0; }
 
@@ -42,7 +42,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_CREATE], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int inode_create(struct inode *inode ,struct dentry *dentry,
+		static int inode_create(struct inode *inode ,struct dentry *dentry,
 		    umode_t umode, bool flag) { return 0; }
 
 		static const struct inode_operations

config/kernel-inode-getattr.m4

@@ -7,7 +7,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_GETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_getattr_mnt_idmap], [
 		#include <linux/fs.h>
 
-		int test_getattr(
+		static int test_getattr(
 		    struct mnt_idmap *idmap,
 		    const struct path *p, struct kstat *k,
 		    u32 request_mask, unsigned int query_flags)
@@ -28,7 +28,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_GETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_getattr_userns], [
 		#include <linux/fs.h>
 
-		int test_getattr(
+		static int test_getattr(
 			struct user_namespace *userns,
 		    const struct path *p, struct kstat *k,
 		    u32 request_mask, unsigned int query_flags)
@@ -47,7 +47,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_GETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_getattr_path], [
 		#include <linux/fs.h>
 
-		int test_getattr(
+		static int test_getattr(
 		    const struct path *p, struct kstat *k,
 		    u32 request_mask, unsigned int query_flags)
 		    { return 0; }
@@ -61,7 +61,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_GETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_getattr_vfsmount], [
 		#include <linux/fs.h>
 
-		int test_getattr(
+		static int test_getattr(
 		    struct vfsmount *mnt, struct dentry *d,
 		    struct kstat *k)
 		    { return 0; }

config/kernel-inode-lookup.m4

@@ -6,7 +6,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_LOOKUP_FLAGS], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		struct dentry *inode_lookup(struct inode *inode,
+		static struct dentry *inode_lookup(struct inode *inode,
 		    struct dentry *dentry, unsigned int flags) { return NULL; }
 
 		static const struct inode_operations iops

config/kernel-inode-permission.m4

@@ -8,12 +8,12 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_PERMISSION], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int inode_permission(struct mnt_idmap *idmap,
+		static int test_permission(struct mnt_idmap *idmap,
 		    struct inode *inode, int mask) { return 0; }
 
 		static const struct inode_operations
 			iops __attribute__ ((unused)) = {
-			.permission             = inode_permission,
+			.permission             = test_permission,
 		};
 	],[])
 
@@ -25,12 +25,12 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_PERMISSION], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int inode_permission(struct user_namespace *userns,
+		static int test_permission(struct user_namespace *userns,
 		    struct inode *inode, int mask) { return 0; }
 
 		static const struct inode_operations
 			iops __attribute__ ((unused)) = {
-			.permission		= inode_permission,
+			.permission		= test_permission,
 		};
 	],[])
 ])

config/kernel-inode-setattr.m4

@@ -7,7 +7,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_SETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_setattr_mnt_idmap], [
 		#include <linux/fs.h>
 
-		int test_setattr(
+		static int test_setattr(
 		    struct mnt_idmap *idmap,
 		    struct dentry *de, struct iattr *ia)
 		    { return 0; }
@@ -27,7 +27,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_SETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_setattr_userns], [
 		#include <linux/fs.h>
 
-		int test_setattr(
+		static int test_setattr(
 		    struct user_namespace *userns,
 		    struct dentry *de, struct iattr *ia)
 		    { return 0; }
@@ -41,7 +41,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_SETATTR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_setattr], [
 		#include <linux/fs.h>
 
-		int test_setattr(
+		static int test_setattr(
 		    struct dentry *de, struct iattr *ia)
 		    { return 0; }
 

config/kernel-inode-times.m4

@@ -47,11 +47,53 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_INODE_TIMES], [
 		#include <linux/fs.h>
 	],[
 		struct inode ip;
-		struct timespec64 ts;
+		struct timespec64 ts = {0};
 
 		memset(&ip, 0, sizeof(ip));
 		inode_set_ctime_to_ts(&ip, ts);
 	])
+
+	dnl #
+	dnl # 6.7 API change
+	dnl # i_atime/i_mtime no longer directly accessible, must use
+	dnl # inode_get_mtime(ip), inode_set_mtime*(ip) to
+	dnl # read/write.
+	dnl #
+	ZFS_LINUX_TEST_SRC([inode_get_atime], [
+		#include <linux/fs.h>
+	],[
+		struct inode ip;
+
+		memset(&ip, 0, sizeof(ip));
+		inode_get_atime(&ip);
+	])
+	ZFS_LINUX_TEST_SRC([inode_get_mtime], [
+		#include <linux/fs.h>
+	],[
+		struct inode ip;
+
+		memset(&ip, 0, sizeof(ip));
+		inode_get_mtime(&ip);
+	])
+
+	ZFS_LINUX_TEST_SRC([inode_set_atime_to_ts], [
+		#include <linux/fs.h>
+	],[
+		struct inode ip;
+		struct timespec64 ts = {0};
+
+		memset(&ip, 0, sizeof(ip));
+		inode_set_atime_to_ts(&ip, ts);
+	])
+	ZFS_LINUX_TEST_SRC([inode_set_mtime_to_ts], [
+		#include <linux/fs.h>
+	],[
+		struct inode ip;
+		struct timespec64 ts = {0};
+
+		memset(&ip, 0, sizeof(ip));
+		inode_set_mtime_to_ts(&ip, ts);
+	])
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_INODE_TIMES], [
@@ -90,4 +132,40 @@ AC_DEFUN([ZFS_AC_KERNEL_INODE_TIMES], [
 	],[
 		AC_MSG_RESULT(no)
 	])
+
+	AC_MSG_CHECKING([whether inode_get_atime() exists])
+	ZFS_LINUX_TEST_RESULT([inode_get_atime], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_INODE_GET_ATIME, 1,
+		    [inode_get_atime() exists in linux/fs.h])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
+	AC_MSG_CHECKING([whether inode_set_atime_to_ts() exists])
+	ZFS_LINUX_TEST_RESULT([inode_set_atime_to_ts], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_INODE_SET_ATIME_TO_TS, 1,
+		    [inode_set_atime_to_ts() exists in linux/fs.h])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
+	AC_MSG_CHECKING([whether inode_get_mtime() exists])
+	ZFS_LINUX_TEST_RESULT([inode_get_mtime], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_INODE_GET_MTIME, 1,
+		    [inode_get_mtime() exists in linux/fs.h])
+	],[
+		AC_MSG_RESULT(no)
+	])
+
+	AC_MSG_CHECKING([whether inode_set_mtime_to_ts() exists])
+	ZFS_LINUX_TEST_RESULT([inode_set_mtime_to_ts], [
+		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_INODE_SET_MTIME_TO_TS, 1,
+		    [inode_set_mtime_to_ts() exists in linux/fs.h])
+	],[
+		AC_MSG_RESULT(no)
+	])
 ])

config/kernel-make-request-fn.m4

@@ -4,7 +4,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN], [
 	ZFS_LINUX_TEST_SRC([make_request_fn_void], [
 		#include <linux/blkdev.h>
-		void make_request(struct request_queue *q,
+		static void make_request(struct request_queue *q,
 		    struct bio *bio) { return; }
 	],[
 		blk_queue_make_request(NULL, &make_request);
@@ -12,7 +12,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN], [
 
 	ZFS_LINUX_TEST_SRC([make_request_fn_blk_qc_t], [
 		#include <linux/blkdev.h>
-		blk_qc_t make_request(struct request_queue *q,
+		static blk_qc_t make_request(struct request_queue *q,
 		    struct bio *bio) { return (BLK_QC_T_NONE); }
 	],[
 		blk_queue_make_request(NULL, &make_request);
@@ -20,7 +20,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN], [
 
 	ZFS_LINUX_TEST_SRC([blk_alloc_queue_request_fn], [
 		#include <linux/blkdev.h>
-		blk_qc_t make_request(struct request_queue *q,
+		static blk_qc_t make_request(struct request_queue *q,
 		    struct bio *bio) { return (BLK_QC_T_NONE); }
 	],[
 		struct request_queue *q __attribute__ ((unused));
@@ -29,7 +29,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN], [
 
 	ZFS_LINUX_TEST_SRC([blk_alloc_queue_request_fn_rh], [
 		#include <linux/blkdev.h>
-		blk_qc_t make_request(struct request_queue *q,
+		static blk_qc_t make_request(struct request_queue *q,
 		    struct bio *bio) { return (BLK_QC_T_NONE); }
 	],[
 		struct request_queue *q __attribute__ ((unused));

config/kernel-mkdir.m4

@@ -9,7 +9,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MKDIR], [
 	ZFS_LINUX_TEST_SRC([mkdir_mnt_idmap], [
 		#include <linux/fs.h>
 
-		int mkdir(struct mnt_idmap *idmap,
+		static int mkdir(struct mnt_idmap *idmap,
 			struct inode *inode, struct dentry *dentry,
 			umode_t umode) { return 0; }
 		static const struct inode_operations
@@ -26,7 +26,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MKDIR], [
 	ZFS_LINUX_TEST_SRC([mkdir_user_namespace], [
 		#include <linux/fs.h>
 
-		int mkdir(struct user_namespace *userns,
+		static int mkdir(struct user_namespace *userns,
 			struct inode *inode, struct dentry *dentry,
 		    umode_t umode) { return 0; }
 
@@ -47,7 +47,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MKDIR], [
 	ZFS_LINUX_TEST_SRC([inode_operations_mkdir], [
 		#include <linux/fs.h>
 
-		int mkdir(struct inode *inode, struct dentry *dentry,
+		static int mkdir(struct inode *inode, struct dentry *dentry,
 		    umode_t umode) { return 0; }
 
 		static const struct inode_operations

config/kernel-mknod.m4

@@ -7,7 +7,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MKNOD], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int tmp_mknod(struct mnt_idmap *idmap,
+		static int tmp_mknod(struct mnt_idmap *idmap,
 		    struct inode *inode ,struct dentry *dentry,
 		    umode_t u, dev_t d) { return 0; }
 
@@ -25,7 +25,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_MKNOD], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int tmp_mknod(struct user_namespace *userns,
+		static int tmp_mknod(struct user_namespace *userns,
 		    struct inode *inode ,struct dentry *dentry,
 		    umode_t u, dev_t d) { return 0; }
 

config/kernel-proc-operations.m4

@@ -7,14 +7,14 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_PROC_OPERATIONS], [
 	ZFS_LINUX_TEST_SRC([proc_ops_struct], [
 		#include <linux/proc_fs.h>
 
-		int test_open(struct inode *ip, struct file *fp) { return 0; }
-		ssize_t test_read(struct file *fp, char __user *ptr,
+		static int test_open(struct inode *ip, struct file *fp) { return 0; }
+		static ssize_t test_read(struct file *fp, char __user *ptr,
 		    size_t size, loff_t *offp) { return 0; }
-		ssize_t test_write(struct file *fp, const char __user *ptr,
+		static ssize_t test_write(struct file *fp, const char __user *ptr,
 		    size_t size, loff_t *offp) { return 0; }
-		loff_t test_lseek(struct file *fp, loff_t off, int flag)
+		static loff_t test_lseek(struct file *fp, loff_t off, int flag)
 		    { return 0; }
-		int test_release(struct inode *ip, struct file *fp)
+		static int test_release(struct inode *ip, struct file *fp)
 		    { return 0; }
 
 		const struct proc_ops test_ops __attribute__ ((unused)) = {

config/kernel-put-link.m4

@@ -4,7 +4,7 @@ dnl #
 AC_DEFUN([ZFS_AC_KERNEL_SRC_PUT_LINK], [
 	ZFS_LINUX_TEST_SRC([put_link_cookie], [
 		#include <linux/fs.h>
-		void put_link(struct inode *ip, void *cookie)
+		static void put_link(struct inode *ip, void *cookie)
 		    { return; }
 		static struct inode_operations
 		    iops __attribute__ ((unused)) = {
@@ -14,7 +14,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_PUT_LINK], [
 
 	ZFS_LINUX_TEST_SRC([put_link_nameidata], [
 		#include <linux/fs.h>
-		void put_link(struct dentry *de, struct
+		static void put_link(struct dentry *de, struct
 		    nameidata *nd, void *ptr) { return; }
 		static struct inode_operations
 		    iops __attribute__ ((unused)) = {

config/kernel-rename.m4

@@ -8,7 +8,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_rename2], [
 		#include <linux/fs.h>
-		int rename2_fn(struct inode *sip, struct dentry *sdp,
+		static int rename2_fn(struct inode *sip, struct dentry *sdp,
 			struct inode *tip, struct dentry *tdp,
 			unsigned int flags) { return 0; }
 
@@ -26,7 +26,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_rename_flags], [
 		#include <linux/fs.h>
-		int rename_fn(struct inode *sip, struct dentry *sdp,
+		static int rename_fn(struct inode *sip, struct dentry *sdp,
 			struct inode *tip, struct dentry *tdp,
 			unsigned int flags) { return 0; }
 
@@ -44,7 +44,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([dir_inode_operations_wrapper_rename2], [
 		#include <linux/fs.h>
-		int rename2_fn(struct inode *sip, struct dentry *sdp,
+		static int rename2_fn(struct inode *sip, struct dentry *sdp,
 			struct inode *tip, struct dentry *tdp,
 			unsigned int flags) { return 0; }
 
@@ -62,7 +62,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_rename_userns], [
 		#include <linux/fs.h>
-		int rename_fn(struct user_namespace *user_ns, struct inode *sip,
+		static int rename_fn(struct user_namespace *user_ns, struct inode *sip,
 			struct dentry *sdp, struct inode *tip, struct dentry *tdp,
 			unsigned int flags) { return 0; }
 
@@ -77,7 +77,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_RENAME], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_rename_mnt_idmap], [
 		#include <linux/fs.h>
-		int rename_fn(struct mnt_idmap *idmap, struct inode *sip,
+		static int rename_fn(struct mnt_idmap *idmap, struct inode *sip,
 			struct dentry *sdp, struct inode *tip, struct dentry *tdp,
 			unsigned int flags) { return 0; }
 

config/kernel-show-options.m4

@@ -5,7 +5,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SHOW_OPTIONS], [
 	ZFS_LINUX_TEST_SRC([super_operations_show_options], [
 		#include <linux/fs.h>
 
-		int show_options(struct seq_file * x, struct dentry * y) {
+		static int show_options(struct seq_file * x, struct dentry * y) {
 			return 0;
 		};
 

config/kernel-shrink.m4

@@ -8,9 +8,6 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SUPER_BLOCK_S_SHRINK], [
 	ZFS_LINUX_TEST_SRC([super_block_s_shrink], [
 		#include <linux/fs.h>
 
-		int shrink(struct shrinker *s, struct shrink_control *sc)
-		    { return 0; }
-
 		static const struct super_block
 		    sb __attribute__ ((unused)) = {
 			.s_shrink.seeks = DEFAULT_SEEKS,
@@ -19,12 +16,44 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SUPER_BLOCK_S_SHRINK], [
 	],[])
 ])
 
+dnl #
+dnl # 6.7 API change
+dnl # s_shrink is now a pointer.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_SUPER_BLOCK_S_SHRINK_PTR], [
+	ZFS_LINUX_TEST_SRC([super_block_s_shrink_ptr], [
+		#include <linux/fs.h>
+		static unsigned long shrinker_cb(struct shrinker *shrink,
+		    struct shrink_control *sc) { return 0; }
+		static struct shrinker shrinker = {
+			.count_objects = shrinker_cb,
+			.scan_objects = shrinker_cb,
+			.seeks = DEFAULT_SEEKS,
+		};
+		static const struct super_block
+		    sb __attribute__ ((unused)) = {
+			.s_shrink = &shrinker,
+		};
+	],[])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_SUPER_BLOCK_S_SHRINK], [
 	AC_MSG_CHECKING([whether super_block has s_shrink])
 	ZFS_LINUX_TEST_RESULT([super_block_s_shrink], [
 		AC_MSG_RESULT(yes)
+		AC_DEFINE(HAVE_SUPER_BLOCK_S_SHRINK, 1,
+			[have super_block s_shrink])
 	],[
-		ZFS_LINUX_TEST_ERROR([sb->s_shrink()])
+		AC_MSG_RESULT(no)
+		AC_MSG_CHECKING([whether super_block has s_shrink pointer])
+		ZFS_LINUX_TEST_RESULT([super_block_s_shrink_ptr], [
+			AC_MSG_RESULT(yes)
+			AC_DEFINE(HAVE_SUPER_BLOCK_S_SHRINK_PTR, 1,
+				[have super_block s_shrink pointer])
+		],[
+			AC_MSG_RESULT(no)
+			ZFS_LINUX_TEST_ERROR([sb->s_shrink()])
+		])
 	])
 ])
 
@@ -57,7 +86,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SHRINK_CONTROL_HAS_NID], [
 AC_DEFUN([ZFS_AC_KERNEL_SRC_REGISTER_SHRINKER_VARARG], [
 	ZFS_LINUX_TEST_SRC([register_shrinker_vararg], [
 		#include <linux/mm.h>
-		unsigned long shrinker_cb(struct shrinker *shrink,
+		static unsigned long shrinker_cb(struct shrinker *shrink,
 		    struct shrink_control *sc) { return 0; }
 	],[
 		struct shrinker cache_shrinker = {
@@ -72,7 +101,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_REGISTER_SHRINKER_VARARG], [
 AC_DEFUN([ZFS_AC_KERNEL_SRC_SHRINKER_CALLBACK], [
 	ZFS_LINUX_TEST_SRC([shrinker_cb_shrink_control], [
 		#include <linux/mm.h>
-		int shrinker_cb(struct shrinker *shrink,
+		static int shrinker_cb(struct shrinker *shrink,
 		    struct shrink_control *sc) { return 0; }
 	],[
 		struct shrinker cache_shrinker = {
@@ -84,7 +113,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SHRINKER_CALLBACK], [
 
 	ZFS_LINUX_TEST_SRC([shrinker_cb_shrink_control_split], [
 		#include <linux/mm.h>
-		unsigned long shrinker_cb(struct shrinker *shrink,
+		static unsigned long shrinker_cb(struct shrinker *shrink,
 		    struct shrink_control *sc) { return 0; }
 	],[
 		struct shrinker cache_shrinker = {
@@ -96,6 +125,25 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SHRINKER_CALLBACK], [
 	])
 ])
 
+dnl #
+dnl # 6.7 API change
+dnl # register_shrinker has been replaced by shrinker_register.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_SHRINKER_REGISTER], [
+	ZFS_LINUX_TEST_SRC([shrinker_register], [
+		#include <linux/shrinker.h>
+		static unsigned long shrinker_cb(struct shrinker *shrink,
+		    struct shrink_control *sc) { return 0; }
+	],[
+		struct shrinker cache_shrinker = {
+			.count_objects = shrinker_cb,
+			.scan_objects = shrinker_cb,
+			.seeks = DEFAULT_SEEKS,
+		};
+		shrinker_register(&cache_shrinker);
+	])
+])
+
 AC_DEFUN([ZFS_AC_KERNEL_SHRINKER_CALLBACK],[
 	dnl #
 	dnl # 6.0 API change
@@ -133,14 +181,36 @@ AC_DEFUN([ZFS_AC_KERNEL_SHRINKER_CALLBACK],[
 			dnl # cs->shrink() is logically split in to
 			dnl # cs->count_objects() and cs->scan_objects()
 			dnl #
-			AC_MSG_CHECKING([if cs->count_objects callback exists])
+			AC_MSG_CHECKING(
+			    [whether cs->count_objects callback exists])
 			ZFS_LINUX_TEST_RESULT(
-				[shrinker_cb_shrink_control_split],[
-					AC_MSG_RESULT(yes)
-					AC_DEFINE(HAVE_SPLIT_SHRINKER_CALLBACK, 1,
-						[cs->count_objects exists])
+			    [shrinker_cb_shrink_control_split],[
+				AC_MSG_RESULT(yes)
+				AC_DEFINE(HAVE_SPLIT_SHRINKER_CALLBACK, 1,
+				    [cs->count_objects exists])
 			],[
+				AC_MSG_RESULT(no)
+
+				AC_MSG_CHECKING(
+				    [whether shrinker_register exists])
+				ZFS_LINUX_TEST_RESULT([shrinker_register], [
+					AC_MSG_RESULT(yes)
+					AC_DEFINE(HAVE_SHRINKER_REGISTER, 1,
+					    [shrinker_register exists])
+
+					dnl # We assume that the split shrinker
+					dnl # callback exists if
+					dnl # shrinker_register() exists,
+					dnl # because the latter is a much more
+					dnl # recent addition, and the macro
+					dnl # test for shrinker_register() only
+					dnl # works if the callback is split
+					AC_DEFINE(HAVE_SPLIT_SHRINKER_CALLBACK,
+					    1, [cs->count_objects exists])
+				],[
+					AC_MSG_RESULT(no)
 					ZFS_LINUX_TEST_ERROR([shrinker])
+				])
 			])
 		])
 	])
@@ -174,10 +244,12 @@ AC_DEFUN([ZFS_AC_KERNEL_SHRINK_CONTROL_STRUCT], [
 
 AC_DEFUN([ZFS_AC_KERNEL_SRC_SHRINKER], [
 	ZFS_AC_KERNEL_SRC_SUPER_BLOCK_S_SHRINK
+	ZFS_AC_KERNEL_SRC_SUPER_BLOCK_S_SHRINK_PTR
 	ZFS_AC_KERNEL_SRC_SHRINK_CONTROL_HAS_NID
 	ZFS_AC_KERNEL_SRC_SHRINKER_CALLBACK
 	ZFS_AC_KERNEL_SRC_SHRINK_CONTROL_STRUCT
 	ZFS_AC_KERNEL_SRC_REGISTER_SHRINKER_VARARG
+	ZFS_AC_KERNEL_SRC_SHRINKER_REGISTER
 ])
 
 AC_DEFUN([ZFS_AC_KERNEL_SHRINKER], [

config/kernel-strlcpy.m4

@@ -0,0 +1,47 @@
+dnl #
+dnl # 6.8.x replaced strlcpy with strscpy. Check for both so we can provide
+dnl # appropriate fallbacks.
+dnl #
+AC_DEFUN([ZFS_AC_KERNEL_SRC_STRLCPY], [
+	ZFS_LINUX_TEST_SRC([kernel_has_strlcpy], [
+		#include <linux/string.h>
+	], [
+		const char *src = "goodbye";
+		char dst[32];
+		size_t len;
+		len = strlcpy(dst, src, sizeof (dst));
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_SRC_STRSCPY], [
+	ZFS_LINUX_TEST_SRC([kernel_has_strscpy], [
+		#include <linux/string.h>
+	], [
+		const char *src = "goodbye";
+		char dst[32];
+		ssize_t len;
+		len = strscpy(dst, src, sizeof (dst));
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_STRLCPY], [
+	AC_MSG_CHECKING([whether strlcpy() exists])
+	ZFS_LINUX_TEST_RESULT([kernel_has_strlcpy], [
+		AC_MSG_RESULT([yes])
+		AC_DEFINE(HAVE_KERNEL_STRLCPY, 1,
+			[strlcpy() exists])
+	], [
+		AC_MSG_RESULT([no])
+	])
+])
+
+AC_DEFUN([ZFS_AC_KERNEL_STRSCPY], [
+	AC_MSG_CHECKING([whether strscpy() exists])
+	ZFS_LINUX_TEST_RESULT([kernel_has_strscpy], [
+		AC_MSG_RESULT([yes])
+		AC_DEFINE(HAVE_KERNEL_STRSCPY, 1,
+			[strscpy() exists])
+	], [
+		AC_MSG_RESULT([no])
+	])
+])

config/kernel-symlink.m4

@@ -6,7 +6,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SYMLINK], [
 	ZFS_LINUX_TEST_SRC([symlink_mnt_idmap], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
-		int tmp_symlink(struct mnt_idmap *idmap,
+		static int tmp_symlink(struct mnt_idmap *idmap,
 		    struct inode *inode ,struct dentry *dentry,
 		    const char *path) { return 0; }
 
@@ -23,7 +23,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_SYMLINK], [
 		#include <linux/fs.h>
 		#include <linux/sched.h>
 
-		int tmp_symlink(struct user_namespace *userns,
+		static int tmp_symlink(struct user_namespace *userns,
 		    struct inode *inode ,struct dentry *dentry,
 		    const char *path) { return 0; }
 

config/kernel-timer.m4

@@ -18,7 +18,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TIMER_SETUP], [
 			int data;
 		};
 
-		void task_expire(struct timer_list *tl)
+		static void task_expire(struct timer_list *tl)
 		{
 			struct my_task_timer *task_timer =
 			    from_timer(task_timer, tl, timer);
@@ -31,7 +31,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TIMER_SETUP], [
 
 	ZFS_LINUX_TEST_SRC([timer_list_function], [
 		#include <linux/timer.h>
-		void task_expire(struct timer_list *tl) {}
+		static void task_expire(struct timer_list *tl) {}
 	],[
 		struct timer_list tl;
 		tl.function = task_expire;

config/kernel-tmpfile.m4

@@ -9,7 +9,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_mnt_idmap], [
 		#include <linux/fs.h>
-		int tmpfile(struct mnt_idmap *idmap,
+		static int tmpfile(struct mnt_idmap *idmap,
 		    struct inode *inode, struct file *file,
 		    umode_t mode) { return 0; }
 		static struct inode_operations
@@ -22,7 +22,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile], [
 		#include <linux/fs.h>
-		int tmpfile(struct user_namespace *userns,
+		static int tmpfile(struct user_namespace *userns,
 		    struct inode *inode, struct file *file,
 		    umode_t mode) { return 0; }
 		static struct inode_operations
@@ -36,7 +36,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry_userns], [
 		#include <linux/fs.h>
-		int tmpfile(struct user_namespace *userns,
+		static int tmpfile(struct user_namespace *userns,
 		    struct inode *inode, struct dentry *dentry,
 		    umode_t mode) { return 0; }
 		static struct inode_operations
@@ -46,7 +46,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_TMPFILE], [
 	],[])
 	ZFS_LINUX_TEST_SRC([inode_operations_tmpfile_dentry], [
 			#include <linux/fs.h>
-			int tmpfile(struct inode *inode, struct dentry *dentry,
+			static int tmpfile(struct inode *inode, struct dentry *dentry,
 			    umode_t mode) { return 0; }
 			static struct inode_operations
 			    iops __attribute__ ((unused)) = {

config/kernel-vfs-direct_IO.m4

@@ -5,7 +5,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO], [
 	ZFS_LINUX_TEST_SRC([direct_io_iter], [
 		#include <linux/fs.h>
 
-		ssize_t test_direct_IO(struct kiocb *kiocb,
+		static ssize_t test_direct_IO(struct kiocb *kiocb,
 		    struct iov_iter *iter) { return 0; }
 
 		static const struct address_space_operations
@@ -17,7 +17,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO], [
 	ZFS_LINUX_TEST_SRC([direct_io_iter_offset], [
 		#include <linux/fs.h>
 
-		ssize_t test_direct_IO(struct kiocb *kiocb,
+		static ssize_t test_direct_IO(struct kiocb *kiocb,
 		    struct iov_iter *iter, loff_t offset) { return 0; }
 
 		static const struct address_space_operations
@@ -29,7 +29,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO], [
 	ZFS_LINUX_TEST_SRC([direct_io_iter_rw_offset], [
 		#include <linux/fs.h>
 
-		ssize_t test_direct_IO(int rw, struct kiocb *kiocb,
+		static ssize_t test_direct_IO(int rw, struct kiocb *kiocb,
 		    struct iov_iter *iter, loff_t offset) { return 0; }
 
 		static const struct address_space_operations
@@ -41,7 +41,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DIRECT_IO], [
 	ZFS_LINUX_TEST_SRC([direct_io_iovec], [
 		#include <linux/fs.h>
 
-		ssize_t test_direct_IO(int rw, struct kiocb *kiocb,
+		static ssize_t test_direct_IO(int rw, struct kiocb *kiocb,
 		    const struct iovec *iov, loff_t offset,
 		    unsigned long nr_segs) { return 0; }
 

config/kernel-vfs-iterate.m4

@@ -1,7 +1,7 @@
 AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_ITERATE], [
 	ZFS_LINUX_TEST_SRC([file_operations_iterate_shared], [
 		#include <linux/fs.h>
-		int iterate(struct file *filp, struct dir_context * context)
+		static int iterate(struct file *filp, struct dir_context * context)
 		    { return 0; }
 
 		static const struct file_operations fops
@@ -12,7 +12,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_ITERATE], [
 
 	ZFS_LINUX_TEST_SRC([file_operations_iterate], [
 		#include <linux/fs.h>
-		int iterate(struct file *filp,
+		static int iterate(struct file *filp,
 		    struct dir_context *context) { return 0; }
 
 		static const struct file_operations fops
@@ -27,7 +27,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_ITERATE], [
 
 	ZFS_LINUX_TEST_SRC([file_operations_readdir], [
 		#include <linux/fs.h>
-		int readdir(struct file *filp, void *entry,
+		static int readdir(struct file *filp, void *entry,
 		    filldir_t func) { return 0; }
 
 		static const struct file_operations fops

config/kernel-vfs-rw-iterate.m4

@@ -5,9 +5,9 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_RW_ITERATE], [
 	ZFS_LINUX_TEST_SRC([file_operations_rw], [
 		#include <linux/fs.h>
 
-		ssize_t test_read(struct kiocb *kiocb, struct iov_iter *to)
+		static ssize_t test_read(struct kiocb *kiocb, struct iov_iter *to)
 		    { return 0; }
-		ssize_t test_write(struct kiocb *kiocb, struct iov_iter *from)
+		static ssize_t test_write(struct kiocb *kiocb, struct iov_iter *from)
 		    { return 0; }
 
 		static const struct file_operations

config/kernel-writepage_t.m4

@@ -6,7 +6,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_WRITEPAGE_T], [
 	dnl #
 	ZFS_LINUX_TEST_SRC([writepage_t_folio], [
 		#include <linux/writeback.h>
-		int putpage(struct folio *folio,
+		static int putpage(struct folio *folio,
 		    struct writeback_control *wbc, void *data)
 		{ return 0; }
 		writepage_t func = putpage;

config/kernel-xattr-handler.m4

@@ -68,7 +68,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry_inode], [
 		#include <linux/xattr.h>
 
-		int get(const struct xattr_handler *handler,
+		static int get(const struct xattr_handler *handler,
 		    struct dentry *dentry, struct inode *inode,
 		    const char *name, void *buffer, size_t size) { return 0; }
 		static const struct xattr_handler
@@ -80,7 +80,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_get_xattr_handler], [
 		#include <linux/xattr.h>
 
-		int get(const struct xattr_handler *handler,
+		static int get(const struct xattr_handler *handler,
 		    struct dentry *dentry, const char *name,
 		    void *buffer, size_t size) { return 0; }
 		static const struct xattr_handler
@@ -92,7 +92,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry], [
 		#include <linux/xattr.h>
 
-		int get(struct dentry *dentry, const char *name,
+		static int get(struct dentry *dentry, const char *name,
 		    void *buffer, size_t size, int handler_flags)
 		    { return 0; }
 		static const struct xattr_handler
@@ -104,7 +104,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_GET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_get_dentry_inode_flags], [
 		#include <linux/xattr.h>
 
-		int get(const struct xattr_handler *handler,
+		static int get(const struct xattr_handler *handler,
 		    struct dentry *dentry, struct inode *inode,
 		    const char *name, void *buffer,
 		    size_t size, int flags) { return 0; }
@@ -182,7 +182,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_set_mnt_idmap], [
 		#include <linux/xattr.h>
 
-		int set(const struct xattr_handler *handler,
+		static int set(const struct xattr_handler *handler,
 			struct mnt_idmap *idmap,
 			struct dentry *dentry, struct inode *inode,
 			const char *name, const void *buffer,
@@ -197,7 +197,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_set_userns], [
 		#include <linux/xattr.h>
 
-		int set(const struct xattr_handler *handler,
+		static int set(const struct xattr_handler *handler,
 			struct user_namespace *mnt_userns,
 			struct dentry *dentry, struct inode *inode,
 			const char *name, const void *buffer,
@@ -212,7 +212,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_set_dentry_inode], [
 		#include <linux/xattr.h>
 
-		int set(const struct xattr_handler *handler,
+		static int set(const struct xattr_handler *handler,
 		    struct dentry *dentry, struct inode *inode,
 		    const char *name, const void *buffer,
 		    size_t size, int flags)
@@ -226,7 +226,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_set_xattr_handler], [
 		#include <linux/xattr.h>
 
-		int set(const struct xattr_handler *handler,
+		static int set(const struct xattr_handler *handler,
 		    struct dentry *dentry, const char *name,
 		    const void *buffer, size_t size, int flags)
 		    { return 0; }
@@ -239,7 +239,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_SET], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_set_dentry], [
 		#include <linux/xattr.h>
 
-		int set(struct dentry *dentry, const char *name,
+		static int set(struct dentry *dentry, const char *name,
 		    const void *buffer, size_t size, int flags,
 		    int handler_flags) { return 0; }
 		static const struct xattr_handler
@@ -325,7 +325,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_LIST], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_list_simple], [
 		#include <linux/xattr.h>
 
-		bool list(struct dentry *dentry) { return 0; }
+		static bool list(struct dentry *dentry) { return 0; }
 		static const struct xattr_handler
 		    xops __attribute__ ((unused)) = {
 			.list = list,
@@ -335,7 +335,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_LIST], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_list_xattr_handler], [
 		#include <linux/xattr.h>
 
-		size_t list(const struct xattr_handler *handler,
+		static size_t list(const struct xattr_handler *handler,
 		    struct dentry *dentry, char *list, size_t list_size,
 		    const char *name, size_t name_len) { return 0; }
 		static const struct xattr_handler
@@ -347,7 +347,7 @@ AC_DEFUN([ZFS_AC_KERNEL_SRC_XATTR_HANDLER_LIST], [
 	ZFS_LINUX_TEST_SRC([xattr_handler_list_dentry], [
 		#include <linux/xattr.h>
 
-		size_t list(struct dentry *dentry,
+		static size_t list(struct dentry *dentry,
 		    char *list, size_t list_size,
 		    const char *name, size_t name_len,
 		    int handler_flags) { return 0; }

config/kernel.m4

@@ -149,6 +149,8 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
 	ZFS_AC_KERNEL_SRC_SYSFS
 	ZFS_AC_KERNEL_SRC_SET_SPECIAL_STATE
 	ZFS_AC_KERNEL_SRC_STANDALONE_LINUX_STDARG
+	ZFS_AC_KERNEL_SRC_STRLCPY
+	ZFS_AC_KERNEL_SRC_STRSCPY
 	ZFS_AC_KERNEL_SRC_PAGEMAP_FOLIO_WAIT_BIT
 	ZFS_AC_KERNEL_SRC_ADD_DISK
 	ZFS_AC_KERNEL_SRC_KTHREAD
@@ -156,6 +158,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
 	ZFS_AC_KERNEL_SRC___COPY_FROM_USER_INATOMIC
 	ZFS_AC_KERNEL_SRC_USER_NS_COMMON_INUM
 	ZFS_AC_KERNEL_SRC_IDMAP_MNT_API
+	ZFS_AC_KERNEL_SRC_IDMAP_NO_USERNS
 	ZFS_AC_KERNEL_SRC_IATTR_VFSID
 	ZFS_AC_KERNEL_SRC_FILEMAP
 	ZFS_AC_KERNEL_SRC_WRITEPAGE_T
@@ -168,6 +171,9 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [
 			ZFS_AC_KERNEL_SRC_CPU_HAS_FEATURE
 			ZFS_AC_KERNEL_SRC_FLUSH_DCACHE_PAGE
 			;;
+		riscv*)
+			ZFS_AC_KERNEL_SRC_FLUSH_DCACHE_PAGE
+			;;
 	esac
 
 	AC_MSG_CHECKING([for available kernel interfaces])
@@ -291,6 +297,8 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 	ZFS_AC_KERNEL_SYSFS
 	ZFS_AC_KERNEL_SET_SPECIAL_STATE
 	ZFS_AC_KERNEL_STANDALONE_LINUX_STDARG
+	ZFS_AC_KERNEL_STRLCPY
+	ZFS_AC_KERNEL_STRSCPY
 	ZFS_AC_KERNEL_PAGEMAP_FOLIO_WAIT_BIT
 	ZFS_AC_KERNEL_ADD_DISK
 	ZFS_AC_KERNEL_KTHREAD
@@ -298,6 +306,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 	ZFS_AC_KERNEL___COPY_FROM_USER_INATOMIC
 	ZFS_AC_KERNEL_USER_NS_COMMON_INUM
 	ZFS_AC_KERNEL_IDMAP_MNT_API
+	ZFS_AC_KERNEL_IDMAP_NO_USERNS
 	ZFS_AC_KERNEL_IATTR_VFSID
 	ZFS_AC_KERNEL_FILEMAP
 	ZFS_AC_KERNEL_WRITEPAGE_T
@@ -310,6 +319,9 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [
 			ZFS_AC_KERNEL_CPU_HAS_FEATURE
 			ZFS_AC_KERNEL_FLUSH_DCACHE_PAGE
 			;;
+		riscv*)
+			ZFS_AC_KERNEL_FLUSH_DCACHE_PAGE
+			;;
 	esac
 ])
 

config/rpm.am

@@ -83,6 +83,11 @@ srpm-common:
 		rpm-local || exit 1; \
 	LANG=C $(RPMBUILD) \
 		--define "_tmppath $$rpmbuild/TMP" \
+		--define "_builddir $$rpmbuild/BUILD" \
+		--define "_rpmdir $$rpmbuild/RPMS" \
+		--define "_srcrpmdir $$rpmbuild/SRPMS" \
+		--define "_specdir $$rpmbuild/SPECS" \
+		--define "_sourcedir $$rpmbuild/SOURCES" \
 		--define "_topdir $$rpmbuild" \
 		$(def) -bs $$rpmbuild/SPECS/$$rpmspec || exit 1; \
 	cp $$rpmbuild/SRPMS/$$rpmpkg . || exit 1; \
@@ -99,6 +104,11 @@ rpm-common:
 		rpm-local || exit 1; \
 	LANG=C ${RPMBUILD} \
 		--define "_tmppath $$rpmbuild/TMP" \
+		--define "_builddir $$rpmbuild/BUILD" \
+		--define "_rpmdir $$rpmbuild/RPMS" \
+		--define "_srcrpmdir $$rpmbuild/SRPMS" \
+		--define "_specdir $$rpmbuild/SPECS" \
+		--define "_sourcedir $$rpmbuild/SOURCES" \
 		--define "_topdir $$rpmbuild" \
 		$(def) --rebuild $$rpmpkg || exit 1; \
 	cp $$rpmbuild/RPMS/*/* . || exit 1; \

include/Makefile.am

@@ -33,6 +33,7 @@ COMMON_H = \
 	sys/bqueue.h \
 	sys/btree.h \
 	sys/brt.h \
+	sys/brt_impl.h \
 	sys/dataset_kstats.h \
 	sys/dbuf.h \
 	sys/ddt.h \

include/libzfs.h

@@ -318,6 +318,9 @@ _LIBZFS_H int zpool_vdev_remove_wanted(zpool_handle_t *, const char *);
 
 _LIBZFS_H int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t);
 _LIBZFS_H int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t);
+_LIBZFS_H int zpool_vdev_set_removed_state(zpool_handle_t *, uint64_t,
+    vdev_aux_t);
+
 _LIBZFS_H int zpool_vdev_clear(zpool_handle_t *, uint64_t);
 
 _LIBZFS_H nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *,

include/libzutil.h

@@ -97,6 +97,7 @@ _LIBZUTIL_H int zpool_find_config(libpc_handle_t *, const char *, nvlist_t **,
 _LIBZUTIL_H const char * const * zpool_default_search_paths(size_t *count);
 _LIBZUTIL_H int zpool_read_label(int, nvlist_t **, int *);
 _LIBZUTIL_H int zpool_label_disk_wait(const char *, int);
+_LIBZUTIL_H int zpool_disk_wait(const char *);
 
 struct udev_device;
 
@@ -163,6 +164,8 @@ _LIBZUTIL_H void zfs_niceraw(uint64_t, char *, size_t);
 _LIBZUTIL_H void zpool_dump_ddt(const ddt_stat_t *, const ddt_histogram_t *);
 _LIBZUTIL_H int zpool_history_unpack(char *, uint64_t, uint64_t *, nvlist_t ***,
     uint_t *);
+_LIBZUTIL_H void fsleep(float sec);
+_LIBZUTIL_H int zpool_getenv_int(const char *env, int default_val);
 
 struct zfs_cmd;
 
@@ -205,9 +208,65 @@ _LIBZUTIL_H void zfs_setproctitle(const char *fmt, ...);
 typedef int (*pool_vdev_iter_f)(void *, nvlist_t *, void *);
 int for_each_vdev_cb(void *zhp, nvlist_t *nv, pool_vdev_iter_f func,
     void *data);
+int for_each_vdev_macro_helper_func(void *zhp_data, nvlist_t *nv, void *data);
+int for_each_real_leaf_vdev_macro_helper_func(void *zhp_data, nvlist_t *nv,
+    void *data);
+/*
+ * Often you'll want to iterate over all the vdevs in the pool, but don't want
+ * to use for_each_vdev() since it requires a callback function.
+ *
+ * Instead you can use FOR_EACH_VDEV():
+ *
+ *     zpool_handle_t *zhp      // Assume this is initialized
+ *     nvlist_t *nv
+ *     ...
+ *     FOR_EACH_VDEV(zhp, nv) {
+ *	 const char *path = NULL;
+ *	 nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path);
+ *	 printf("Looking at vdev %s\n", path);
+ *     }
+ *
+ * Note: FOR_EACH_VDEV runs in O(n^2) time where n = number of vdevs.  However,
+ * there's an upper limit of 256 vdevs per dRAID top-level vdevs (TLDs), 255 for
+ * raidz2 TLDs, a real world limit of ~500 vdevs for mirrors, so this shouldn't
+ * really be an issue.
+ *
+ * Here are some micro-benchmarks of a complete FOR_EACH_VDEV loop on a RAID0
+ * pool:
+ *
+ * 100  vdevs = 0.7ms
+ * 500  vdevs = 17ms
+ * 750  vdevs = 40ms
+ * 1000 vdevs = 82ms
+ *
+ * The '__nv += 0' at the end of the for() loop gets around a "comma or
+ * semicolon followed by non-blank" checkstyle error.  Note on most compliers
+ * the '__nv += 0' can just be replaced with 'NULL', but gcc on Centos 7
+ * will give a 'warning: statement with no effect' error if you do that.
+ */
+#define	__FOR_EACH_VDEV(__zhp, __nv, __func) { \
+	__nv = zpool_get_config(__zhp, NULL); \
+	VERIFY0(nvlist_lookup_nvlist(__nv, ZPOOL_CONFIG_VDEV_TREE, &__nv)); \
+	} \
+	for (nvlist_t *__root_nv = __nv, *__state = (nvlist_t *)0; \
+	    for_each_vdev_cb(&__state, __root_nv, __func, &__nv) == 1; \
+	    __nv += 0)
+
+#define	FOR_EACH_VDEV(__zhp, __nv) \
+	__FOR_EACH_VDEV(__zhp, __nv, for_each_vdev_macro_helper_func)
+
+/*
+ * "real leaf" vdevs are leaf vdevs that are real devices (disks or files).
+ * This excludes leaf vdevs like like draid spares.
+ */
+#define	FOR_EACH_REAL_LEAF_VDEV(__zhp, __nv) \
+	__FOR_EACH_VDEV(__zhp, __nv, for_each_real_leaf_vdev_macro_helper_func)
+
 int for_each_vdev_in_nvlist(nvlist_t *nvroot, pool_vdev_iter_f func,
     void *data);
 void update_vdevs_config_dev_sysfs_path(nvlist_t *config);
+_LIBZUTIL_H void update_vdev_config_dev_sysfs_path(nvlist_t *nv,
+    const char *path, const char *key);
 #ifdef	__cplusplus
 }
 #endif

include/os/freebsd/spl/sys/mod_os.h

@@ -91,6 +91,12 @@
 #define	param_set_max_auto_ashift_args(var) \
     CTLTYPE_UINT, NULL, 0, param_set_max_auto_ashift, "IU"
 
+#define	spa_taskq_read_param_set_args(var) \
+    CTLTYPE_STRING, NULL, 0, spa_taskq_read_param, "A"
+
+#define	spa_taskq_write_param_set_args(var) \
+    CTLTYPE_STRING, NULL, 0, spa_taskq_write_param, "A"
+
 #define	fletcher_4_param_set_args(var) \
     CTLTYPE_STRING, NULL, 0, fletcher_4_param, "A"
 

include/os/freebsd/spl/sys/uio.h

@@ -62,7 +62,7 @@ zfs_uio_setoffset(zfs_uio_t *uio, offset_t off)
 }
 
 static inline void
-zfs_uio_advance(zfs_uio_t *uio, size_t size)
+zfs_uio_advance(zfs_uio_t *uio, ssize_t size)
 {
 	zfs_uio_resid(uio) -= size;
 	zfs_uio_offset(uio) += size;

include/os/freebsd/spl/sys/vfs.h

@@ -101,7 +101,7 @@ void vfs_setmntopt(vfs_t *vfsp, const char *name, const char *arg,
 void vfs_clearmntopt(vfs_t *vfsp, const char *name);
 int vfs_optionisset(const vfs_t *vfsp, const char *opt, char **argp);
 int mount_snapshot(kthread_t *td, vnode_t **vpp, const char *fstype,
-    char *fspath, char *fspec, int fsflags);
+    char *fspath, char *fspec, int fsflags, vfs_t *parent_vfsp);
 
 typedef	uint64_t	vfs_feature_t;
 

include/os/freebsd/spl/sys/vnode.h

@@ -56,6 +56,7 @@ enum symfollow { NO_FOLLOW = NOFOLLOW };
 #ifndef IN_BASE
 #include_next <sys/vnode.h>
 #endif
+#include <sys/ccompat.h>
 #include <sys/mount.h>
 #include <sys/cred.h>
 #include <sys/fcntl.h>
@@ -104,7 +105,7 @@ vn_flush_cached_data(vnode_t *vp, boolean_t sync)
 		zfs_vmobject_wlock(vp->v_object);
 		vm_object_page_clean(vp->v_object, 0, 0, flags);
 		zfs_vmobject_wunlock(vp->v_object);
-		VOP_UNLOCK(vp);
+		VOP_UNLOCK1(vp);
 	}
 }
 #endif

include/os/freebsd/zfs/sys/zfs_vfsops_os.h

@@ -286,7 +286,6 @@ typedef struct zfid_long {
 
 extern uint_t zfs_fsyncer_key;
 extern int zfs_super_owner;
-extern int zfs_bclone_enabled;
 
 extern void zfs_init(void);
 extern void zfs_fini(void);

include/os/linux/kernel/linux/dcache_compat.h

@@ -42,8 +42,8 @@
 /*
  * Starting from Linux 5.13, flush_dcache_page() becomes an inline function
  * and under some configurations, may indirectly referencing GPL-only
- * cpu_feature_keys on powerpc. Override this function when it is detected
- * being GPL-only.
+ * symbols, e.g., cpu_feature_keys on powerpc and PageHuge on riscv.
+ * Override this function when it is detected being GPL-only.
  */
 #if defined __powerpc__ && defined HAVE_FLUSH_DCACHE_PAGE_GPL_ONLY
 #include <linux/simd_powerpc.h>
@@ -53,6 +53,17 @@
 			clear_bit(PG_dcache_clean, &(page)->flags);	\
 	} while (0)
 #endif
+/*
+ * For riscv implementation, the use of PageHuge can be safely removed.
+ * Because it handles pages allocated by HugeTLB, while flush_dcache_page
+ * in zfs module is only called on kernel pages.
+ */
+#if defined __riscv && defined HAVE_FLUSH_DCACHE_PAGE_GPL_ONLY
+#define	flush_dcache_page(page)	do {					\
+		if (test_bit(PG_dcache_clean, &(page)->flags))		\
+			clear_bit(PG_dcache_clean, &(page)->flags);	\
+	} while (0)
+#endif
 
 /*
  * 2.6.30 API change,

include/os/linux/kernel/linux/simd_aarch64.h

@@ -71,9 +71,15 @@
 #define	ID_AA64PFR0_EL1		sys_reg(3, 0, 0, 1, 0)
 #define	ID_AA64ISAR0_EL1	sys_reg(3, 0, 0, 6, 0)
 
+#if (defined(HAVE_KERNEL_NEON) && defined(CONFIG_KERNEL_MODE_NEON))
 #define	kfpu_allowed()		1
 #define	kfpu_begin()		kernel_neon_begin()
 #define	kfpu_end()		kernel_neon_end()
+#else
+#define	kfpu_allowed()		0
+#define	kfpu_begin()		do {} while (0)
+#define	kfpu_end()		do {} while (0)
+#endif
 #define	kfpu_init()		(0)
 #define	kfpu_fini()		do {} while (0)
 

include/os/linux/kernel/linux/simd_arm.h

@@ -53,9 +53,15 @@
 #include <asm/elf.h>
 #include <asm/hwcap.h>
 
+#if (defined(HAVE_KERNEL_NEON) && defined(CONFIG_KERNEL_MODE_NEON))
 #define	kfpu_allowed()		1
 #define	kfpu_begin()		kernel_neon_begin()
 #define	kfpu_end()		kernel_neon_end()
+#else
+#define	kfpu_allowed()		0
+#define	kfpu_begin()		do {} while (0)
+#define	kfpu_end()		do {} while (0)
+#endif
 #define	kfpu_init()		(0)
 #define	kfpu_fini()		do {} while (0)
 

include/os/linux/spl/sys/cred.h

@@ -73,13 +73,25 @@ static inline struct user_namespace *zfs_i_user_ns(struct inode *inode)
 static inline boolean_t zfs_no_idmapping(struct user_namespace *mnt_userns,
     struct user_namespace *fs_userns)
 {
-	return (zfs_is_init_userns(mnt_userns) || mnt_userns == fs_userns);
+	return (zfs_is_init_userns(mnt_userns) ||
+	    mnt_userns == fs_userns);
 }
 
 static inline uid_t zfs_uid_to_vfsuid(zidmap_t *mnt_userns,
     struct user_namespace *fs_userns, uid_t uid)
 {
-	struct user_namespace *owner = idmap_owner(mnt_userns);
+	struct user_namespace *owner;
+#ifdef HAVE_IOPS_CREATE_IDMAP
+	if (mnt_userns == zfs_init_idmap)
+		return (uid);
+#endif
+#ifdef HAVE_IDMAP_NO_USERNS
+	struct user_namespace ns;
+	ns.uid_map = mnt_userns->uid_map;
+	owner = &ns;
+#else
+	owner = idmap_owner(mnt_userns);
+#endif
 	if (zfs_no_idmapping(owner, fs_userns))
 		return (uid);
 	if (!zfs_is_init_userns(fs_userns))
@@ -92,7 +104,18 @@ static inline uid_t zfs_uid_to_vfsuid(zidmap_t *mnt_userns,
 static inline gid_t zfs_gid_to_vfsgid(zidmap_t *mnt_userns,
     struct user_namespace *fs_userns, gid_t gid)
 {
-	struct user_namespace *owner = idmap_owner(mnt_userns);
+	struct user_namespace *owner;
+#ifdef HAVE_IOPS_CREATE_IDMAP
+	if (mnt_userns == zfs_init_idmap)
+		return (gid);
+#endif
+#ifdef HAVE_IDMAP_NO_USERNS
+	struct user_namespace ns;
+	ns.gid_map = mnt_userns->gid_map;
+	owner = &ns;
+#else
+	owner = idmap_owner(mnt_userns);
+#endif
 	if (zfs_no_idmapping(owner, fs_userns))
 		return (gid);
 	if (!zfs_is_init_userns(fs_userns))
@@ -105,7 +128,18 @@ static inline gid_t zfs_gid_to_vfsgid(zidmap_t *mnt_userns,
 static inline uid_t zfs_vfsuid_to_uid(zidmap_t *mnt_userns,
     struct user_namespace *fs_userns, uid_t uid)
 {
-	struct user_namespace *owner = idmap_owner(mnt_userns);
+	struct user_namespace *owner;
+#ifdef HAVE_IOPS_CREATE_IDMAP
+	if (mnt_userns == zfs_init_idmap)
+		return (uid);
+#endif
+#ifdef HAVE_IDMAP_NO_USERNS
+	struct user_namespace ns;
+	ns.uid_map = mnt_userns->uid_map;
+	owner = &ns;
+#else
+	owner = idmap_owner(mnt_userns);
+#endif
 	if (zfs_no_idmapping(owner, fs_userns))
 		return (uid);
 	uid = from_kuid(owner, KUIDT_INIT(uid));
@@ -119,7 +153,18 @@ static inline uid_t zfs_vfsuid_to_uid(zidmap_t *mnt_userns,
 static inline gid_t zfs_vfsgid_to_gid(zidmap_t *mnt_userns,
     struct user_namespace *fs_userns, gid_t gid)
 {
-	struct user_namespace *owner = idmap_owner(mnt_userns);
+	struct user_namespace *owner;
+#ifdef HAVE_IOPS_CREATE_IDMAP
+	if (mnt_userns == zfs_init_idmap)
+		return (gid);
+#endif
+#ifdef HAVE_IDMAP_NO_USERNS
+	struct user_namespace ns;
+	ns.gid_map = mnt_userns->gid_map;
+	owner = &ns;
+#else
+	owner = idmap_owner(mnt_userns);
+#endif
 	if (zfs_no_idmapping(owner, fs_userns))
 		return (gid);
 	gid = from_kgid(owner, KGIDT_INIT(gid));

include/os/linux/spl/sys/kmem_cache.h

@@ -70,8 +70,6 @@ typedef enum kmem_cbrc {
 #define	KMC_REAP_CHUNK		INT_MAX
 #define	KMC_DEFAULT_SEEKS	1
 
-#define	KMC_RECLAIM_ONCE	0x1	/* Force a single shrinker pass */
-
 extern struct list_head spl_kmem_cache_list;
 extern struct rw_semaphore spl_kmem_cache_sem;
 

include/os/linux/spl/sys/shrinker.h

@@ -29,12 +29,13 @@
 
 /*
  * Due to frequent changes in the shrinker API the following
- * compatibility wrappers should be used.  They are as follows:
+ * compatibility wrapper should be used.
  *
- *   SPL_SHRINKER_DECLARE(varname, countfunc, scanfunc, seek_cost);
+ *   shrinker = spl_register_shrinker(name, countfunc, scanfunc, seek_cost);
+ *   spl_unregister_shrinker(shrinker);
  *
- * SPL_SHRINKER_DECLARE is used to declare a shrinker with the name varname,
- * which is passed to spl_register_shrinker()/spl_unregister_shrinker().
+ * spl_register_shrinker is used to create and register a shrinker with the
+ * given name.
  * The countfunc returns the number of free-able objects.
  * The scanfunc returns the number of objects that were freed.
  * The callbacks can return SHRINK_STOP if further calls can't make any more
@@ -57,57 +58,28 @@
  *	...scan objects in the cache and reclaim them...
  * }
  *
- * SPL_SHRINKER_DECLARE(my_shrinker, my_count, my_scan, DEFAULT_SEEKS);
+ * static struct shrinker *my_shrinker;
  *
  * void my_init_func(void) {
- *	spl_register_shrinker(&my_shrinker);
+ *	my_shrinker = spl_register_shrinker("my-shrinker",
+ *	    my_count, my_scan, DEFAULT_SEEKS);
+ * }
+ *
+ * void my_fini_func(void) {
+ *	spl_unregister_shrinker(my_shrinker);
  * }
  */
 
-#ifdef HAVE_REGISTER_SHRINKER_VARARG
-#define	spl_register_shrinker(x)	register_shrinker(x, "zfs-arc-shrinker")
-#else
-#define	spl_register_shrinker(x)	register_shrinker(x)
-#endif
-#define	spl_unregister_shrinker(x)	unregister_shrinker(x)
+typedef unsigned long (*spl_shrinker_cb)
+	(struct shrinker *, struct shrink_control *);
 
-/*
- * Linux 3.0 to 3.11 Shrinker API Compatibility.
- */
-#if defined(HAVE_SINGLE_SHRINKER_CALLBACK)
-#define	SPL_SHRINKER_DECLARE(varname, countfunc, scanfunc, seek_cost)	\
-static int								\
-__ ## varname ## _wrapper(struct shrinker *shrink, struct shrink_control *sc)\
-{									\
-	if (sc->nr_to_scan != 0) {					\
-		(void) scanfunc(shrink, sc);				\
-	}								\
-	return (countfunc(shrink, sc));					\
-}									\
-									\
-static struct shrinker varname = {					\
-	.shrink = __ ## varname ## _wrapper,				\
-	.seeks = seek_cost,						\
-}
+struct shrinker *spl_register_shrinker(const char *name,
+    spl_shrinker_cb countfunc, spl_shrinker_cb scanfunc, int seek_cost);
+void spl_unregister_shrinker(struct shrinker *);
 
+#ifndef SHRINK_STOP
+/* 3.0-3.11 compatibility */
 #define	SHRINK_STOP	(-1)
-
-/*
- * Linux 3.12 and later Shrinker API Compatibility.
- */
-#elif defined(HAVE_SPLIT_SHRINKER_CALLBACK)
-#define	SPL_SHRINKER_DECLARE(varname, countfunc, scanfunc, seek_cost)	\
-static struct shrinker varname = {					\
-	.count_objects = countfunc,					\
-	.scan_objects = scanfunc,					\
-	.seeks = seek_cost,						\
-}
-
-#else
-/*
- * Linux 2.x to 2.6.22, or a newer shrinker API has been introduced.
- */
-#error "Unknown shrinker callback"
 #endif
 
 #endif /* SPL_SHRINKER_H */

include/os/linux/spl/sys/string.h

@@ -1 +1,50 @@
+/*
+ *  Copyright (C) 2007-2010 Lawrence Livermore National Security, LLC.
+ *  Produced at Lawrence Livermore National Laboratory (cf, DISCLAIMER).
+ *  Written by Brian Behlendorf <behlendorf1@llnl.gov>.
+ *  UCRL-CODE-235197
+ *
+ *  This file is part of the SPL, Solaris Porting Layer.
+ *
+ *  The SPL is free software; you can redistribute it and/or modify it
+ *  under the terms of the GNU General Public License as published by the
+ *  Free Software Foundation; either version 2 of the License, or (at your
+ *  option) any later version.
+ *
+ *  The SPL is distributed in the hope that it will be useful, but WITHOUT
+ *  ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ *  FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
+ *  for more details.
+ *
+ *  You should have received a copy of the GNU General Public License along
+ *  with the SPL.  If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef _SPL_STRING_H
+#define	_SPL_STRING_H
+
 #include <linux/string.h>
+
+/* Fallbacks for kernel missing strlcpy */
+#ifndef HAVE_KERNEL_STRLCPY
+
+#if defined(HAVE_KERNEL_STRSCPY)
+/*
+ * strscpy is strlcpy, but returns an error on truncation. strlcpy is defined
+ * to return strlen(src), so detect error and override it.
+ */
+static inline size_t
+strlcpy(char *dest, const char *src, size_t size)
+{
+	ssize_t ret = strscpy(dest, src, size);
+	if (likely(ret > 0))
+		return ((size_t)ret);
+	return (strlen(src));
+}
+#else
+#error "no strlcpy fallback available"
+#endif
+
+#endif /* HAVE_KERNEL_STRLCPY */
+
+#endif /* _SPL_STRING_H */

include/os/linux/spl/sys/types.h

@@ -57,12 +57,23 @@ typedef int			minor_t;
 struct user_namespace;
 #ifdef HAVE_IOPS_CREATE_IDMAP
 #include <linux/refcount.h>
+#ifdef HAVE_IDMAP_NO_USERNS
+#include <linux/user_namespace.h>
+struct mnt_idmap {
+	struct uid_gid_map uid_map;
+	struct uid_gid_map gid_map;
+	refcount_t count;
+};
+typedef struct mnt_idmap	zidmap_t;
+#define	idmap_owner(p)	(NULL)
+#else
 struct mnt_idmap {
 	struct user_namespace *owner;
 	refcount_t count;
 };
 typedef struct mnt_idmap	zidmap_t;
 #define	idmap_owner(p)	(((struct mnt_idmap *)p)->owner)
+#endif
 #else
 typedef struct user_namespace	zidmap_t;
 #define	idmap_owner(p)	((struct user_namespace *)p)

include/os/linux/spl/sys/uio.h

@@ -95,7 +95,7 @@ zfs_uio_setoffset(zfs_uio_t *uio, offset_t off)
 }
 
 static inline void
-zfs_uio_advance(zfs_uio_t *uio, size_t size)
+zfs_uio_advance(zfs_uio_t *uio, ssize_t size)
 {
 	uio->uio_resid -= size;
 	uio->uio_loffset += size;

include/os/linux/zfs/sys/zfs_vfsops_os.h

@@ -45,8 +45,6 @@ extern "C" {
 typedef struct zfsvfs zfsvfs_t;
 struct znode;
 
-extern int zfs_bclone_enabled;
-
 /*
  * This structure emulates the vfs_t from other platforms.  It's purpose
  * is to facilitate the handling of mount options and minimize structural

include/os/linux/zfs/sys/zpl.h

@@ -273,5 +273,25 @@ extern long zpl_ioctl_fideduperange(struct file *filp, void *arg);
 #else
 #define	zpl_inode_set_ctime_to_ts(ip, ts)	(ip->i_ctime = ts)
 #endif
+#ifdef HAVE_INODE_GET_ATIME
+#define	zpl_inode_get_atime(ip)	inode_get_atime(ip)
+#else
+#define	zpl_inode_get_atime(ip)	(ip->i_atime)
+#endif
+#ifdef HAVE_INODE_SET_ATIME_TO_TS
+#define	zpl_inode_set_atime_to_ts(ip, ts)	inode_set_atime_to_ts(ip, ts)
+#else
+#define	zpl_inode_set_atime_to_ts(ip, ts)	(ip->i_atime = ts)
+#endif
+#ifdef HAVE_INODE_GET_MTIME
+#define	zpl_inode_get_mtime(ip)	inode_get_mtime(ip)
+#else
+#define	zpl_inode_get_mtime(ip)	(ip->i_mtime)
+#endif
+#ifdef HAVE_INODE_SET_MTIME_TO_TS
+#define	zpl_inode_set_mtime_to_ts(ip, ts)	inode_set_mtime_to_ts(ip, ts)
+#else
+#define	zpl_inode_set_mtime_to_ts(ip, ts)	(ip->i_mtime = ts)
+#endif
 
 #endif	/* _SYS_ZPL_H */

include/sys/brt_impl.h

@@ -0,0 +1,199 @@
+/*
+ * CDDL HEADER START
+ *
+ * The contents of this file are subject to the terms of the
+ * Common Development and Distribution License (the "License").
+ * You may not use this file except in compliance with the License.
+ *
+ * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
+ * or https://opensource.org/licenses/CDDL-1.0.
+ * See the License for the specific language governing permissions
+ * and limitations under the License.
+ *
+ * When distributing Covered Code, include this CDDL HEADER in each
+ * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
+ * If applicable, add the following below this CDDL HEADER, with the
+ * fields enclosed by brackets "[]" replaced with your own identifying
+ * information: Portions Copyright [yyyy] [name of copyright owner]
+ *
+ * CDDL HEADER END
+ */
+/*
+ * Copyright (c) 2020, 2021, 2022 by Pawel Jakub Dawidek
+ */
+
+#ifndef _SYS_BRT_IMPL_H
+#define	_SYS_BRT_IMPL_H
+
+#ifdef	__cplusplus
+extern "C" {
+#endif
+
+/*
+ * BRT - Block Reference Table.
+ */
+#define	BRT_OBJECT_VDEV_PREFIX	"com.fudosecurity:brt:vdev:"
+
+/*
+ * We divide each VDEV into 16MB chunks. Each chunk is represented in memory
+ * by a 16bit counter, thus 1TB VDEV requires 128kB of memory: (1TB / 16MB) * 2B
+ * Each element in this array represents how many BRT entries do we have in this
+ * chunk of storage. We always load this entire array into memory and update as
+ * needed. By having it in memory we can quickly tell (during zio_free()) if
+ * there are any BRT entries that we might need to update.
+ *
+ * This value cannot be larger than 16MB, at least as long as we support
+ * 512 byte block sizes. With 512 byte block size we can have exactly
+ * 32768 blocks in 16MB. In 32MB we could have 65536 blocks, which is one too
+ * many for a 16bit counter.
+ */
+#define	BRT_RANGESIZE	(16 * 1024 * 1024)
+_Static_assert(BRT_RANGESIZE / SPA_MINBLOCKSIZE <= UINT16_MAX,
+	"BRT_RANGESIZE is too large.");
+/*
+ * We don't want to update the whole structure every time. Maintain bitmap
+ * of dirty blocks within the regions, so that a single bit represents a
+ * block size of entcounts. For example if we have a 1PB vdev then all
+ * entcounts take 128MB of memory ((64TB / 16MB) * 2B). We can divide this
+ * 128MB array of entcounts into 32kB disk blocks, as we don't want to update
+ * the whole 128MB on disk when we have updated only a single entcount.
+ * We maintain a bitmap where each 32kB disk block within 128MB entcounts array
+ * is represented by a single bit. This gives us 4096 bits. A set bit in the
+ * bitmap means that we had a change in at least one of the 16384 entcounts
+ * that reside on a 32kB disk block (32kB / sizeof (uint16_t)).
+ */
+#define	BRT_BLOCKSIZE	(32 * 1024)
+#define	BRT_RANGESIZE_TO_NBLOCKS(size)					\
+	(((size) - 1) / BRT_BLOCKSIZE / sizeof (uint16_t) + 1)
+
+#define	BRT_LITTLE_ENDIAN	0
+#define	BRT_BIG_ENDIAN		1
+#ifdef _ZFS_LITTLE_ENDIAN
+#define	BRT_NATIVE_BYTEORDER		BRT_LITTLE_ENDIAN
+#define	BRT_NON_NATIVE_BYTEORDER	BRT_BIG_ENDIAN
+#else
+#define	BRT_NATIVE_BYTEORDER		BRT_BIG_ENDIAN
+#define	BRT_NON_NATIVE_BYTEORDER	BRT_LITTLE_ENDIAN
+#endif
+
+typedef struct brt_vdev_phys {
+	uint64_t	bvp_mos_entries;
+	uint64_t	bvp_size;
+	uint64_t	bvp_byteorder;
+	uint64_t	bvp_totalcount;
+	uint64_t	bvp_rangesize;
+	uint64_t	bvp_usedspace;
+	uint64_t	bvp_savedspace;
+} brt_vdev_phys_t;
+
+typedef struct brt_vdev {
+	/*
+	 * VDEV id.
+	 */
+	uint64_t	bv_vdevid;
+	/*
+	 * Is the structure initiated?
+	 * (bv_entcount and bv_bitmap are allocated?)
+	 */
+	boolean_t	bv_initiated;
+	/*
+	 * Object number in the MOS for the entcount array and brt_vdev_phys.
+	 */
+	uint64_t	bv_mos_brtvdev;
+	/*
+	 * Object number in the MOS for the entries table.
+	 */
+	uint64_t	bv_mos_entries;
+	/*
+	 * Entries to sync.
+	 */
+	avl_tree_t	bv_tree;
+	/*
+	 * Does the bv_entcount[] array needs byte swapping?
+	 */
+	boolean_t	bv_need_byteswap;
+	/*
+	 * Number of entries in the bv_entcount[] array.
+	 */
+	uint64_t	bv_size;
+	/*
+	 * This is the array with BRT entry count per BRT_RANGESIZE.
+	 */
+	uint16_t	*bv_entcount;
+	/*
+	 * Sum of all bv_entcount[]s.
+	 */
+	uint64_t	bv_totalcount;
+	/*
+	 * Space on disk occupied by cloned blocks (without compression).
+	 */
+	uint64_t	bv_usedspace;
+	/*
+	 * How much additional space would be occupied without block cloning.
+	 */
+	uint64_t	bv_savedspace;
+	/*
+	 * brt_vdev_phys needs updating on disk.
+	 */
+	boolean_t	bv_meta_dirty;
+	/*
+	 * bv_entcount[] needs updating on disk.
+	 */
+	boolean_t	bv_entcount_dirty;
+	/*
+	 * bv_entcount[] potentially can be a bit too big to sychronize it all
+	 * when we just changed few entcounts. The fields below allow us to
+	 * track updates to bv_entcount[] array since the last sync.
+	 * A single bit in the bv_bitmap represents as many entcounts as can
+	 * fit into a single BRT_BLOCKSIZE.
+	 * For example we have 65536 entcounts in the bv_entcount array
+	 * (so the whole array is 128kB). We updated bv_entcount[2] and
+	 * bv_entcount[5]. In that case only first bit in the bv_bitmap will
+	 * be set and we will write only first BRT_BLOCKSIZE out of 128kB.
+	 */
+	ulong_t		*bv_bitmap;
+	uint64_t	bv_nblocks;
+} brt_vdev_t;
+
+/*
+ * In-core brt
+ */
+typedef struct brt {
+	krwlock_t	brt_lock;
+	spa_t		*brt_spa;
+#define	brt_mos		brt_spa->spa_meta_objset
+	uint64_t	brt_rangesize;
+	uint64_t	brt_usedspace;
+	uint64_t	brt_savedspace;
+	avl_tree_t	brt_pending_tree[TXG_SIZE];
+	kmutex_t	brt_pending_lock[TXG_SIZE];
+	/* Sum of all entries across all bv_trees. */
+	uint64_t	brt_nentries;
+	brt_vdev_t	*brt_vdevs;
+	uint64_t	brt_nvdevs;
+} brt_t;
+
+/* Size of bre_offset / sizeof (uint64_t). */
+#define	BRT_KEY_WORDS	(1)
+
+/*
+ * In-core brt entry.
+ * On-disk we use bre_offset as the key and bre_refcount as the value.
+ */
+typedef struct brt_entry {
+	uint64_t	bre_offset;
+	uint64_t	bre_refcount;
+	avl_node_t	bre_node;
+} brt_entry_t;
+
+typedef struct brt_pending_entry {
+	blkptr_t	bpe_bp;
+	int		bpe_count;
+	avl_node_t	bpe_node;
+} brt_pending_entry_t;
+
+#ifdef	__cplusplus
+}
+#endif
+
+#endif	/* _SYS_BRT_IMPL_H */

include/sys/dataset_kstats.h

@@ -71,6 +71,7 @@ typedef struct dataset_kstats {
 
 int dataset_kstats_create(dataset_kstats_t *, objset_t *);
 void dataset_kstats_destroy(dataset_kstats_t *);
+void dataset_kstats_rename(dataset_kstats_t *dk, const char *);
 
 void dataset_kstats_update_write_kstats(dataset_kstats_t *, int64_t);
 void dataset_kstats_update_read_kstats(dataset_kstats_t *, int64_t);

include/sys/dbuf.h

@@ -379,8 +379,8 @@ dmu_buf_impl_t *dbuf_find(struct objset *os, uint64_t object, uint8_t level,
 int dbuf_read(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags);
 void dmu_buf_will_clone(dmu_buf_t *db, dmu_tx_t *tx);
 void dmu_buf_will_not_fill(dmu_buf_t *db, dmu_tx_t *tx);
-void dmu_buf_will_fill(dmu_buf_t *db, dmu_tx_t *tx);
-void dmu_buf_fill_done(dmu_buf_t *db, dmu_tx_t *tx);
+void dmu_buf_will_fill(dmu_buf_t *db, dmu_tx_t *tx, boolean_t canfail);
+boolean_t dmu_buf_fill_done(dmu_buf_t *db, dmu_tx_t *tx, boolean_t failed);
 void dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx);
 dbuf_dirty_record_t *dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx);
 dbuf_dirty_record_t *dbuf_dirty_lightweight(dnode_t *dn, uint64_t blkid,

include/sys/dmu.h

@@ -1072,8 +1072,7 @@ int dmu_offset_next(objset_t *os, uint64_t object, boolean_t hole,
 int dmu_read_l0_bps(objset_t *os, uint64_t object, uint64_t offset,
     uint64_t length, struct blkptr *bps, size_t *nbpsp);
 int dmu_brt_clone(objset_t *os, uint64_t object, uint64_t offset,
-    uint64_t length, dmu_tx_t *tx, const struct blkptr *bps, size_t nbps,
-    boolean_t replay);
+    uint64_t length, dmu_tx_t *tx, const struct blkptr *bps, size_t nbps);
 
 /*
  * Initial setup and final teardown.

include/sys/dsl_crypt.h

@@ -206,6 +206,7 @@ void dsl_dataset_promote_crypt_sync(dsl_dir_t *target, dsl_dir_t *origin,
     dmu_tx_t *tx);
 int dmu_objset_create_crypt_check(dsl_dir_t *parentdd,
     dsl_crypto_params_t *dcp, boolean_t *will_encrypt);
+boolean_t dmu_objset_crypto_key_equal(objset_t *osa, objset_t *osb);
 void dsl_dataset_create_crypt_sync(uint64_t dsobj, dsl_dir_t *dd,
     struct dsl_dataset *origin, dsl_crypto_params_t *dcp, dmu_tx_t *tx);
 uint64_t dsl_crypto_key_create_sync(uint64_t crypt, dsl_wrapping_key_t *wkey,

include/sys/spa_impl.h

@@ -266,6 +266,7 @@ struct spa {
 
 	spa_aux_vdev_t	spa_spares;		/* hot spares */
 	spa_aux_vdev_t	spa_l2cache;		/* L2ARC cache devices */
+	boolean_t	spa_aux_sync_uber;	/* need to sync aux uber */
 	nvlist_t	*spa_label_features;	/* Features for reading MOS */
 	uint64_t	spa_config_object;	/* MOS object for pool config */
 	uint64_t	spa_config_generation;	/* config generation number */

include/sys/zfs_vnops.h

@@ -24,8 +24,11 @@
 
 #ifndef	_SYS_FS_ZFS_VNOPS_H
 #define	_SYS_FS_ZFS_VNOPS_H
+
 #include <sys/zfs_vnops_os.h>
 
+extern int zfs_bclone_enabled;
+
 extern int zfs_fsync(znode_t *, int, cred_t *);
 extern int zfs_read(znode_t *, zfs_uio_t *, int, cred_t *);
 extern int zfs_write(znode_t *, zfs_uio_t *, int, cred_t *);

lib/libspl/include/assert.h

@@ -64,6 +64,9 @@ libspl_assert(const char *buf, const char *file, const char *func, int line)
 #undef verify
 #endif
 
+#define	PANIC(fmt, a...)						\
+	libspl_assertf(__FILE__, __FUNCTION__, __LINE__, fmt, ## a)
+
 #define	VERIFY(cond)							\
 	(void) ((!(cond)) &&						\
 	    libspl_assert(#cond, __FILE__, __FUNCTION__, __LINE__))

lib/libspl/include/os/freebsd/sys/stat.h

@@ -76,8 +76,12 @@ fstat64_blk(int fd, struct stat64 *st)
 /*
  * Only Intel-based Macs have a separate stat64; Arm-based Macs are like
  * FreeBSD and have a full 64-bit stat from the start.
+ *
+ * On Linux, musl libc is full 64-bit too and has deprecated its own version
+ * of these defines since version 1.2.4.
  */
-#if defined(__APPLE__) && !(defined(__i386__) || defined(__x86_64__))
+#if (defined(__APPLE__) && !(defined(__i386__) || defined(__x86_64__))) || \
+	(defined(__linux__) && !defined(__GLIBC__))
 #define	stat64	stat
 #define	fstat64	fstat
 #endif

lib/libspl/include/sys/uio.h

@@ -90,7 +90,7 @@ zfs_uio_iov_at_index(zfs_uio_t *uio, uint_t idx, void **base, uint64_t *len)
 }
 
 static inline void
-zfs_uio_advance(zfs_uio_t *uio, size_t size)
+zfs_uio_advance(zfs_uio_t *uio, ssize_t size)
 {
 	uio->uio_resid -= size;
 	uio->uio_loffset += size;