Implement Redacted Send/Receive

Redacted send/receive allows users to send subsets of their data to 
a target system. One possible use case for this feature is to not 
transmit sensitive information to a data warehousing, test/dev, or 
analytics environment. Another is to save space by not replicating 
unimportant data within a given dataset, for example in backup tools 
like zrepl.

Redacted send/receive is a three-stage process. First, a clone (or 
clones) is made of the snapshot to be sent to the target. In this 
clone (or clones), all unnecessary or unwanted data is removed or
modified. This clone is then snapshotted to create the "redaction 
snapshot" (or snapshots). Second, the new zfs redact command is used 
to create a redaction bookmark. The redaction bookmark stores the 
list of blocks in a snapshot that were modified by the redaction 
snapshot(s). Finally, the redaction bookmark is passed as a parameter 
to zfs send. When sending to the snapshot that was redacted, the
redaction bookmark is used to filter out blocks that contain sensitive 
or unwanted information, and those blocks are not included in the send 
stream.  When sending from the redaction bookmark, the blocks it 
contains are considered as candidate blocks in addition to those 
blocks in the destination snapshot that were modified since the 
creation_txg of the redaction bookmark.  This step is necessary to 
allow the target to rehydrate data in the case where some blocks are 
accidentally or unnecessarily modified in the redaction snapshot.

The changes to bookmarks to enable fast space estimation involve 
adding deadlists to bookmarks. There is also logic to manage the 
life cycles of these deadlists.

The new size estimation process operates in cases where previously 
an accurate estimate could not be provided. In those cases, a send 
is performed where no data blocks are read, reducing the runtime 
significantly and providing a byte-accurate size estimate.

Reviewed-by: Dan Kimmel <[email protected]>
Reviewed-by: Matt Ahrens <[email protected]>
Reviewed-by: Prashanth Sreenivasa <[email protected]>
Reviewed-by: John Kennedy <[email protected]>
Reviewed-by: George Wilson <[email protected]>
Reviewed-by: Chris Williamson <[email protected]>
Reviewed-by: Pavel Zhakarov <[email protected]>
Reviewed-by: Sebastien Roy <[email protected]>
Reviewed-by: Prakash Surya <[email protected]>
Reviewed-by: Brian Behlendorf <[email protected]>
Signed-off-by: Paul Dagnelie <[email protected]>
Closes #7958 

7 files changed, 376 insertions, 1 deletions

diff --git a/tests/zfs-tests/cmd/Makefile.am b/tests/zfs-tests/cmd/Makefile.am
index 39a538d2d..09c59f591 100644
--- a/tests/zfs-tests/cmd/Makefile.am
+++ b/tests/zfs-tests/cmd/Makefile.am
@@ -8,6 +8,7 @@ SUBDIRS = \
 	file_check \
 	file_trunc \
 	file_write \
+	get_diff \
 	largest_file \
 	libzfs_input_check \
 	mkbusy \
@@ -24,4 +25,5 @@ SUBDIRS = \
 	rename_dir \
 	rm_lnkcnt_zero_file \
 	threadsappend \
-	xattrtest
+	xattrtest \
+	stride_dd
diff --git a/tests/zfs-tests/cmd/get_diff/Makefile.am b/tests/zfs-tests/cmd/get_diff/Makefile.am
new file mode 100644
index 000000000..06c39ddd8
--- /dev/null
+++ b/tests/zfs-tests/cmd/get_diff/Makefile.am
@@ -0,0 +1,6 @@
+include $(top_srcdir)/config/Rules.am
+
+pkgexecdir = $(datadir)/@PACKAGE@/zfs-tests/bin
+
+pkgexec_PROGRAMS = get_diff
+get_diff_SOURCES = get_diff.c
diff --git a/tests/zfs-tests/cmd/get_diff/get_diff.c b/tests/zfs-tests/cmd/get_diff/get_diff.c
new file mode 100644
index 000000000..2799f46b0
--- /dev/null
+++ b/tests/zfs-tests/cmd/get_diff/get_diff.c
@@ -0,0 +1,109 @@
+/*
+ * This file and its contents are supplied under the terms of the
+ * Common Development and Distribution License ("CDDL"), version 1.0.
+ * You may only use this file in accordance with the terms of version
+ * 1.0 of the CDDL.
+ *
+ * A full copy of the text of the CDDL should have accompanied this
+ * source.  A copy of the CDDL is also available via the Internet at
+ * http://www.illumos.org/license/CDDL.
+ */
+
+/*
+ * Copyright (c) 2018 by Delphix. All rights reserved.
+ */
+
+#include <stdio.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include <sys/param.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <errno.h>
+
+static void
+usage(char *msg, int exit_value)
+{
+	(void) fprintf(stderr, "get_diff file redacted_file\n");
+	(void) fprintf(stderr, "%s\n", msg);
+	exit(exit_value);
+}
+
+/*
+ * This utility compares two files, an original and its redacted counterpart
+ * (in that order). It compares the files 512 bytes at a time, printing out
+ * any ranges (as offset and length) where the redacted file does not match
+ * the original. This output is used to verify that the expected ranges of
+ * a redacted file do not contain the original data.
+ */
+int
+main(int argc, char *argv[])
+{
+	off_t		diff_off = 0, diff_len = 0, off = 0;
+	int		fd1, fd2;
+	char		*fname1, *fname2;
+	char		buf1[DEV_BSIZE], buf2[DEV_BSIZE];
+	ssize_t		bytes;
+
+	if (argc != 3)
+		usage("Incorrect number of arguments.", 1);
+
+	if ((fname1 = argv[1]) == NULL)
+		usage("Filename missing.", 1);
+	if ((fd1 = open(fname1, O_LARGEFILE | O_RDONLY)) < 0) {
+		perror("open1 failed");
+		exit(1);
+	}
+
+	if ((fname2 = argv[2]) == NULL)
+		usage("Redacted filename missing.", 1);
+	if ((fd2 = open(fname2, O_LARGEFILE | O_RDONLY)) < 0) {
+		perror("open2 failed");
+		exit(1);
+	}
+
+	while ((bytes = pread(fd1, buf1, DEV_BSIZE, off)) > 0) {
+		if (pread(fd2, buf2, DEV_BSIZE, off) < 0) {
+			if (errno == EIO) {
+				/*
+				 * A read in a redacted section of a file will
+				 * fail with EIO. If we get EIO, continue on
+				 * but ensure that a comparison of buf1 and
+				 * buf2 will fail, indicating a redacted block.
+				 */
+				buf2[0] = ~buf1[0];
+			} else {
+				perror("pread failed");
+				exit(1);
+			}
+		}
+		if (memcmp(buf1, buf2, bytes) == 0) {
+			if (diff_len != 0) {
+				(void) fprintf(stdout, "%lld,%lld\n",
+				    (long long)diff_off, (long long)diff_len);
+				assert(off == diff_off + diff_len);
+				diff_len = 0;
+			}
+			diff_off = 0;
+		} else {
+			if (diff_len == 0)
+				diff_off = off;
+			assert(off == diff_off + diff_len);
+			diff_len += bytes;
+		}
+		off += bytes;
+	}
+
+	if (diff_len != 0 && diff_len != 0) {
+		(void) fprintf(stdout, "%lld,%lld\n", (long long)diff_off,
+		    (long long)diff_len);
+	}
+
+	(void) close(fd1);
+	(void) close(fd2);
+
+	return (0);
+}
diff --git a/tests/zfs-tests/cmd/libzfs_input_check/libzfs_input_check.c b/tests/zfs-tests/cmd/libzfs_input_check/libzfs_input_check.c
index 977b9e2f3..2de1ba20c 100644
--- a/tests/zfs-tests/cmd/libzfs_input_check/libzfs_input_check.c
+++ b/tests/zfs-tests/cmd/libzfs_input_check/libzfs_input_check.c
@@ -691,6 +691,34 @@ zfs_destroy(const char *dataset)
 }
 
 static void
+test_redact(const char *snapshot1, const char *snapshot2)
+{
+	nvlist_t *required = fnvlist_alloc();
+	nvlist_t *snapnv = fnvlist_alloc();
+	char bookmark[MAXNAMELEN + 32];
+
+	fnvlist_add_string(required, "bookname", "testbookmark");
+	fnvlist_add_boolean(snapnv, snapshot2);
+	fnvlist_add_nvlist(required, "snapnv", snapnv);
+
+	IOC_INPUT_TEST(ZFS_IOC_REDACT, snapshot1, required, NULL, 0);
+
+	nvlist_free(snapnv);
+	nvlist_free(required);
+
+	strncpy(bookmark, snapshot1, sizeof (bookmark) - 1);
+	*strchr(bookmark, '@') = '\0';
+	strncat(bookmark, "#testbookmark", sizeof (bookmark));
+	zfs_destroy(bookmark);
+}
+
+static void
+test_get_bookmark_props(const char *bookmark)
+{
+	IOC_INPUT_TEST(ZFS_IOC_GET_BOOKMARK_PROPS, bookmark, NULL, NULL, 0);
+}
+
+static void
 zfs_ioc_input_tests(const char *pool)
 {
 	char filepath[] = "/tmp/ioc_test_file_XXXXXX";
@@ -700,6 +728,7 @@ zfs_ioc_input_tests(const char *pool)
 	char bookmark[ZFS_MAX_DATASET_NAME_LEN + 32];
 	char backup[ZFS_MAX_DATASET_NAME_LEN];
 	char clone[ZFS_MAX_DATASET_NAME_LEN];
+	char clonesnap[ZFS_MAX_DATASET_NAME_LEN + 32];
 	int tmpfd, err;
 
 	/*
@@ -710,6 +739,7 @@ zfs_ioc_input_tests(const char *pool)
 	(void) snprintf(snapshot, sizeof (snapshot), "%s@snapshot", dataset);
 	(void) snprintf(bookmark, sizeof (bookmark), "%s#bookmark", dataset);
 	(void) snprintf(clone, sizeof (clone), "%s/test-fs-clone", pool);
+	(void) snprintf(clonesnap, sizeof (clonesnap), "%s@snap", clone);
 	(void) snprintf(backup, sizeof (backup), "%s/backup", pool);
 
 	err = lzc_create(dataset, DMU_OST_ZFS, NULL, NULL, 0);
@@ -747,6 +777,7 @@ zfs_ioc_input_tests(const char *pool)
 
 	test_bookmark(pool, snapshot, bookmark);
 	test_get_bookmarks(dataset);
+	test_get_bookmark_props(bookmark);
 	test_destroy_bookmarks(pool, bookmark);
 
 	test_hold(pool, snapshot);
@@ -754,6 +785,9 @@ zfs_ioc_input_tests(const char *pool)
 	test_release(pool, snapshot);
 
 	test_clone(snapshot, clone);
+	test_snapshot(pool, clonesnap);
+	test_redact(snapshot, clonesnap);
+	zfs_destroy(clonesnap);
 	zfs_destroy(clone);
 
 	test_rollback(dataset, snapshot);
@@ -909,6 +943,8 @@ validate_ioc_values(void)
 	    ZFS_IOC_BASE + 78 == ZFS_IOC_POOL_DISCARD_CHECKPOINT &&
 	    ZFS_IOC_BASE + 79 == ZFS_IOC_POOL_INITIALIZE &&
 	    ZFS_IOC_BASE + 80 == ZFS_IOC_POOL_TRIM &&
+	    ZFS_IOC_BASE + 81 == ZFS_IOC_REDACT &&
+	    ZFS_IOC_BASE + 82 == ZFS_IOC_GET_BOOKMARK_PROPS &&
 	    LINUX_IOC_BASE + 1 == ZFS_IOC_EVENTS_NEXT &&
 	    LINUX_IOC_BASE + 2 == ZFS_IOC_EVENTS_CLEAR &&
 	    LINUX_IOC_BASE + 3 == ZFS_IOC_EVENTS_SEEK);
diff --git a/tests/zfs-tests/cmd/stride_dd/.gitignore b/tests/zfs-tests/cmd/stride_dd/.gitignore
new file mode 100644
index 000000000..7c072ee0d
--- /dev/null
+++ b/tests/zfs-tests/cmd/stride_dd/.gitignore
@@ -0,0 +1 @@
+/stride_dd
diff --git a/tests/zfs-tests/cmd/stride_dd/Makefile.am b/tests/zfs-tests/cmd/stride_dd/Makefile.am
new file mode 100644
index 000000000..d6f1adbac
--- /dev/null
+++ b/tests/zfs-tests/cmd/stride_dd/Makefile.am
@@ -0,0 +1,7 @@
+include $(top_srcdir)/config/Rules.am
+
+pkgexecdir = $(datadir)/@PACKAGE@/zfs-tests/bin
+
+pkgexec_PROGRAMS = stride_dd
+stride_dd_SOURCES = stride_dd.c
+stride_dd_LDADD = -lrt
diff --git a/tests/zfs-tests/cmd/stride_dd/stride_dd.c b/tests/zfs-tests/cmd/stride_dd/stride_dd.c
new file mode 100644
index 000000000..88bd53292
--- /dev/null
+++ b/tests/zfs-tests/cmd/stride_dd/stride_dd.c
@@ -0,0 +1,214 @@
+/*
+ * This file and its contents are supplied under the terms of the
+ * Common Development and Distribution License ("CDDL"), version 1.0.
+ * You may only use this file in accordance with the terms of version
+ * 1.0 of the CDDL.
+ *
+ * A full copy of the text of the CDDL should have accompanied this
+ * source.  A copy of the CDDL is also available via the Internet at
+ * http://www.illumos.org/license/CDDL.
+ */
+
+/*
+ * Copyright (c) 2018 by Delphix. All rights reserved.
+ */
+
+#include <sys/types.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+
+static int bsize = 0;
+static int count = 0;
+static char *ifile = NULL;
+static char *ofile = NULL;
+static int stride = 0;
+static int seek = 0;
+static char *execname = "stride_dd";
+
+static void usage(void);
+static void parse_options(int argc, char *argv[]);
+
+static void
+usage(void)
+{
+	(void) fprintf(stderr,
+	    "usage: %s -i inputfile -o outputfile -b blocksize -c count \n"
+	    "           -s stride [ -k seekblocks]\n"
+	    "\n"
+	    "Simplified version of dd that supports the stride option.\n"
+	    "A stride of n means that for each block written, n - 1 blocks\n"
+	    "are skipped in both the input and output file. A stride of 1\n"
+	    "means that blocks are read and written consecutively.\n"
+	    "All numeric parameters must be integers.\n"
+	    "\n"
+	    "    inputfile:  File to read from\n"
+	    "    outputfile: File to write to\n"
+	    "    blocksize:  Size of each block to read/write\n"
+	    "    count:      Number of blocks to read/write\n"
+	    "    stride:     Read/write a block then skip (stride - 1) blocks\n"
+	    "    seekblocks: Number of blocks to skip at start of output\n",
+	    execname);
+	(void) exit(1);
+}
+
+static void
+parse_options(int argc, char *argv[])
+{
+	int c;
+	int errflag = 0;
+
+	execname = argv[0];
+
+	extern char *optarg;
+	extern int optind, optopt;
+
+	while ((c = getopt(argc, argv, ":b:c:i:o:s:k:")) != -1) {
+		switch (c) {
+			case 'b':
+				bsize = atoi(optarg);
+				break;
+
+			case 'c':
+				count = atoi(optarg);
+				break;
+
+			case 'i':
+				ifile = optarg;
+				break;
+
+			case 'o':
+				ofile = optarg;
+				break;
+
+			case 's':
+				stride = atoi(optarg);
+				break;
+
+			case 'k':
+				seek = atoi(optarg);
+				break;
+
+			case ':':
+				(void) fprintf(stderr,
+				    "Option -%c requires an operand\n", optopt);
+				errflag++;
+				break;
+
+			case '?':
+			default:
+				(void) fprintf(stderr,
+				    "Unrecognized option: -%c\n", optopt);
+				errflag++;
+				break;
+		}
+
+		if (errflag) {
+			(void) usage();
+		}
+	}
+
+	if (bsize <= 0 || count <= 0 || stride <= 0 || ifile == NULL ||
+	    ofile == NULL || seek < 0) {
+		(void) fprintf(stderr,
+		    "Required parameter(s) missing or invalid.\n");
+		(void) usage();
+	}
+}
+
+int
+main(int argc, char *argv[])
+{
+	int i;
+	int ifd;
+	int ofd;
+	void *buf;
+	int c;
+
+	parse_options(argc, argv);
+
+	ifd = open(ifile, O_RDONLY);
+	if (ifd == -1) {
+		(void) fprintf(stderr, "%s: %s: ", execname, ifile);
+		perror("open");
+		exit(2);
+	}
+
+	ofd = open(ofile, O_WRONLY | O_CREAT, 0666);
+	if (ofd == -1) {
+		(void) fprintf(stderr, "%s: %s: ", execname, ofile);
+		perror("open");
+		exit(2);
+	}
+
+	/*
+	 * We use valloc because some character block devices expect a
+	 * page-aligned buffer.
+	 */
+	int err = posix_memalign(&buf, 4096, bsize);
+	if (err != 0) {
+		(void) fprintf(stderr,
+		    "%s: %s\n", execname, strerror(err));
+		exit(2);
+	}
+
+	if (seek > 0) {
+		if (lseek(ofd, seek * bsize, SEEK_CUR) == -1) {
+			perror("output lseek");
+			exit(2);
+		}
+	}
+
+	for (i = 0; i < count; i++) {
+		c = read(ifd, buf, bsize);
+		if (c != bsize) {
+
+			perror("read");
+			exit(2);
+		}
+		if (c != bsize) {
+			if (c < 0) {
+				perror("read");
+			} else {
+				(void) fprintf(stderr,
+				    "%s: unexpected short read, read %d "
+				    "bytes, expected %d\n", execname,
+				    c, bsize);
+			}
+			exit(2);
+		}
+
+		c = write(ofd, buf, bsize);
+		if (c != bsize) {
+			if (c < 0) {
+				perror("write");
+			} else {
+				(void) fprintf(stderr,
+				    "%s: unexpected short write, wrote %d "
+				    "bytes, expected %d\n", execname,
+				    c, bsize);
+			}
+			exit(2);
+		}
+
+		if (stride > 1) {
+			if (lseek(ifd, (stride - 1) * bsize, SEEK_CUR) == -1) {
+				perror("input lseek");
+				exit(2);
+			}
+			if (lseek(ofd, (stride - 1) * bsize, SEEK_CUR) == -1) {
+				perror("output lseek");
+				exit(2);
+			}
+		}
+	}
+	free(buf);
+
+	(void) close(ofd);
+	(void) close(ifd);
+
+	return (0);
+}