Illumos 5027 - zfs large block support

5027 zfs large block support
Reviewed by: Alek Pinchuk <[email protected]>
Reviewed by: George Wilson <[email protected]>
Reviewed by: Josef 'Jeff' Sipek <[email protected]>
Reviewed by: Richard Elling <[email protected]>
Reviewed by: Saso Kiselkov <[email protected]>
Reviewed by: Brian Behlendorf <[email protected]>
Approved by: Dan McDonald <[email protected]>

References:
  https://www.illumos.org/issues/5027
  https://github.com/illumos/illumos-gate/commit/b515258

Porting Notes:

* Included in this patch is a tiny ISP2() cleanup in zio_init() from
Illumos 5255.

* Unlike the upstream Illumos commit this patch does not impose an
arbitrary 128K block size limit on volumes.  Volumes, like filesystems,
are limited by the zfs_max_recordsize=1M module option.

* By default the maximum record size is limited to 1M by the module
option zfs_max_recordsize.  This value may be safely increased up to
16M which is the largest block size supported by the on-disk format.
At the moment, 1M blocks clearly offer a significant performance
improvement but the benefits of going beyond this for the majority
of workloads are less clear.

* The illumos version of this patch increased DMU_MAX_ACCESS to 32M.
This was determined not to be large enough when using 16M blocks
because the zfs_make_xattrdir() function will fail (EFBIG) when
assigning a TX.  This was immediately observed under Linux because
all newly created files must have a security xattr created and
that was failing.  Therefore, we've set DMU_MAX_ACCESS to 64M.

* On 32-bit platforms a hard limit of 1M is set for blocks due
to the limited virtual address space.  We should be able to relax
this one the ABD patches are merged.

Ported-by: Brian Behlendorf <[email protected]>
Closes #354

2 files changed, 27 insertions, 18 deletions

diff --git a/lib/libzfs/libzfs_dataset.c b/lib/libzfs/libzfs_dataset.c
index 4087baf34..9f23d1801 100644
--- a/lib/libzfs/libzfs_dataset.c
+++ b/lib/libzfs/libzfs_dataset.c
@@ -1055,21 +1055,28 @@ zfs_valid_proplist(libzfs_handle_t *hdl, zfs_type_t type, nvlist_t *nvl,
 			break;
 		}
 
-		case ZFS_PROP_RECORDSIZE:
 		case ZFS_PROP_VOLBLOCKSIZE:
-			/* must be power of two within SPA_{MIN,MAX}BLOCKSIZE */
+		case ZFS_PROP_RECORDSIZE:
+		{
+			int maxbs = SPA_MAXBLOCKSIZE;
+			if (zhp != NULL) {
+				maxbs = zpool_get_prop_int(zhp->zpool_hdl,
+				    ZPOOL_PROP_MAXBLOCKSIZE, NULL);
+			}
+			/*
+			 * The value must be a power of two between
+			 * SPA_MINBLOCKSIZE and maxbs.
+			 */
 			if (intval < SPA_MINBLOCKSIZE ||
-			    intval > SPA_MAXBLOCKSIZE || !ISP2(intval)) {
+			    intval > maxbs || !ISP2(intval)) {
 				zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
-				    "'%s' must be power of 2 from %u "
-				    "to %uk"), propname,
-				    (uint_t)SPA_MINBLOCKSIZE,
-				    (uint_t)SPA_MAXBLOCKSIZE >> 10);
+				    "'%s' must be power of 2 from 512B "
+				    "to %uKB"), propname, maxbs >> 10);
 				(void) zfs_error(hdl, EZFS_BADPROP, errbuf);
 				goto error;
 			}
 			break;
-
+		}
 		case ZFS_PROP_MLSLABEL:
 		{
 #ifdef HAVE_MLSLABEL
@@ -1446,7 +1453,8 @@ zfs_setprop_error(libzfs_handle_t *hdl, zfs_prop_t prop, int err,
 		break;
 
 	case ERANGE:
-		if (prop == ZFS_PROP_COMPRESSION) {
+		if (prop == ZFS_PROP_COMPRESSION ||
+		    prop == ZFS_PROP_RECORDSIZE) {
 			(void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 			    "property setting is not allowed on "
 			    "bootable datasets"));
@@ -3212,9 +3220,7 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type,
 		case EDOM:
 			zfs_error_aux(hdl, dgettext(TEXT_DOMAIN,
 			    "volume block size must be power of 2 from "
-			    "%u to %uk"),
-			    (uint_t)SPA_MINBLOCKSIZE,
-			    (uint_t)SPA_MAXBLOCKSIZE >> 10);
+			    "512B to %uKB"), zfs_max_recordsize >> 10);
 
 			return (zfs_error(hdl, EZFS_BADPROP, errbuf));
 
diff --git a/lib/libzfs/libzfs_sendrecv.c b/lib/libzfs/libzfs_sendrecv.c
index 8911e0ed2..cdc872caa 100644
--- a/lib/libzfs/libzfs_sendrecv.c
+++ b/lib/libzfs/libzfs_sendrecv.c
@@ -214,7 +214,7 @@ static void *
 cksummer(void *arg)
 {
 	dedup_arg_t *dda = arg;
-	char *buf = malloc(1<<20);
+	char *buf = zfs_alloc(dda->dedup_hdl, SPA_MAXBLOCKSIZE);
 	dmu_replay_record_t thedrr;
 	dmu_replay_record_t *drr = &thedrr;
 	struct drr_begin *drrb = &thedrr.drr_u.drr_begin;
@@ -279,9 +279,9 @@ cksummer(void *arg)
 			    DMU_COMPOUNDSTREAM && drr->drr_payloadlen != 0) {
 				int sz = drr->drr_payloadlen;
 
-				if (sz > 1<<20) {
-					free(buf);
-					buf = malloc(sz);
+				if (sz > SPA_MAXBLOCKSIZE) {
+					buf = zfs_realloc(dda->dedup_hdl, buf,
+					    SPA_MAXBLOCKSIZE, sz);
 				}
 				(void) ssread(buf, sz, ofp);
 				if (ferror(stdin))
@@ -834,7 +834,7 @@ typedef struct send_dump_data {
 	char prevsnap[ZFS_MAXNAMELEN];
 	uint64_t prevsnap_obj;
 	boolean_t seenfrom, seento, replicate, doall, fromorigin;
-	boolean_t verbose, dryrun, parsable, progress, embed_data;
+	boolean_t verbose, dryrun, parsable, progress, embed_data, large_block;
 	int outfd;
 	boolean_t err;
 	nvlist_t *fss;
@@ -1181,6 +1181,8 @@ dump_snapshot(zfs_handle_t *zhp, void *arg)
 		}
 
 		enum lzc_send_flags flags = 0;
+		if (sdd->large_block)
+			flags |= LZC_SEND_FLAG_LARGE_BLOCK;
 		if (sdd->embed_data)
 			flags |= LZC_SEND_FLAG_EMBED_DATA;
 
@@ -1529,6 +1531,7 @@ zfs_send(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap,
 	sdd.parsable = flags->parsable;
 	sdd.progress = flags->progress;
 	sdd.dryrun = flags->dryrun;
+	sdd.large_block = flags->largeblock;
 	sdd.embed_data = flags->embed_data;
 	sdd.filter_cb = filter_func;
 	sdd.filter_cb_arg = cb_arg;
@@ -2564,7 +2567,7 @@ static int
 recv_skip(libzfs_handle_t *hdl, int fd, boolean_t byteswap)
 {
 	dmu_replay_record_t *drr;
-	void *buf = malloc(1<<20);
+	void *buf = zfs_alloc(hdl, SPA_MAXBLOCKSIZE);
 	char errbuf[1024];
 
 	(void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN,