From owner-svn-src-stable@freebsd.org Wed Oct 3 14:46:26 2018 Return-Path: Delivered-To: svn-src-stable@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 3D25510C5443; Wed, 3 Oct 2018 14:46:26 +0000 (UTC) (envelope-from mav@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "mxrelay.nyi.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id E67BB7DE8A; Wed, 3 Oct 2018 14:46:25 +0000 (UTC) (envelope-from mav@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id E146318531; Wed, 3 Oct 2018 14:46:25 +0000 (UTC) (envelope-from mav@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id w93EkPs9030129; Wed, 3 Oct 2018 14:46:25 GMT (envelope-from mav@FreeBSD.org) Received: (from mav@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id w93EkP7h030126; Wed, 3 Oct 2018 14:46:25 GMT (envelope-from mav@FreeBSD.org) Message-Id: <201810031446.w93EkP7h030126@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: mav set sender to mav@FreeBSD.org using -f From: Alexander Motin Date: Wed, 3 Oct 2018 14:46:25 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-11@freebsd.org Subject: svn commit: r339128 - in stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs: . sys X-SVN-Group: stable-11 X-SVN-Commit-Author: mav X-SVN-Commit-Paths: in stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs: . sys X-SVN-Commit-Revision: 339128 X-SVN-Commit-Repository: base MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-stable@freebsd.org X-Mailman-Version: 2.1.27 Precedence: list List-Id: SVN commit messages for all the -stable branches of the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 03 Oct 2018 14:46:26 -0000 Author: mav Date: Wed Oct 3 14:46:25 2018 New Revision: 339128 URL: https://svnweb.freebsd.org/changeset/base/339128 Log: MFC r337181: 9539 Make zvol operations use _by_dnode routines Continues what was started in 7801 add more by-dnode routines by fully converting zvols to avoid unnecessary dnode_hold() calls. This saves a small amount of CPU time and slightly improves latencies of operations on zvols. illumos/illumos-gate@8dfe5547fbf0979fc1065a8b6fddc1e940a7cf4f Reviewed by: Matthew Ahrens Reviewed by: Brian Behlendorf Reviewed by: Rick McNeal Approved by: Dan McDonald Author: Richard Yao Modified: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zvol.c Directory Properties: stable/11/ (props changed) Modified: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c ============================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Wed Oct 3 14:45:48 2018 (r339127) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu.c Wed Oct 3 14:46:25 2018 (r339128) @@ -449,7 +449,7 @@ dmu_spill_hold_by_bonus(dmu_buf_t *bonus, void *tag, d * and can induce severe lock contention when writing to several files * whose dnodes are in the same block. */ -static int +int dmu_buf_hold_array_by_dnode(dnode_t *dn, uint64_t offset, uint64_t length, boolean_t read, void *tag, int *numbufsp, dmu_buf_t ***dbpp, uint32_t flags) { @@ -1321,7 +1321,7 @@ xuio_stat_wbuf_nocopy(void) } #ifdef _KERNEL -static int +int dmu_read_uio_dnode(dnode_t *dn, uio_t *uio, uint64_t size) { dmu_buf_t **dbp; @@ -1437,7 +1437,7 @@ dmu_read_uio(objset_t *os, uint64_t object, uio_t *uio return (err); } -static int +int dmu_write_uio_dnode(dnode_t *dn, uio_t *uio, uint64_t size, dmu_tx_t *tx) { dmu_buf_t **dbp; @@ -1873,22 +1873,17 @@ dmu_return_arcbuf(arc_buf_t *buf) * dmu_write(). */ void -dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, arc_buf_t *buf, +dmu_assign_arcbuf_dnode(dnode_t *dn, uint64_t offset, arc_buf_t *buf, dmu_tx_t *tx) { - dmu_buf_impl_t *dbuf = (dmu_buf_impl_t *)handle; - dnode_t *dn; dmu_buf_impl_t *db; uint32_t blksz = (uint32_t)arc_buf_lsize(buf); uint64_t blkid; - DB_DNODE_ENTER(dbuf); - dn = DB_DNODE(dbuf); rw_enter(&dn->dn_struct_rwlock, RW_READER); blkid = dbuf_whichblock(dn, 0, offset); VERIFY((db = dbuf_hold(dn, blkid, FTAG)) != NULL); rw_exit(&dn->dn_struct_rwlock); - DB_DNODE_EXIT(dbuf); /* * We can only assign if the offset is aligned, the arc buf is the @@ -1916,17 +1911,25 @@ dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, ASSERT3U(arc_get_compression(buf), ==, ZIO_COMPRESS_OFF); ASSERT(!(buf->b_flags & ARC_BUF_FLAG_COMPRESSED)); - DB_DNODE_ENTER(dbuf); - dn = DB_DNODE(dbuf); os = dn->dn_objset; object = dn->dn_object; - DB_DNODE_EXIT(dbuf); dbuf_rele(db, FTAG); dmu_write(os, object, offset, blksz, buf->b_data, tx); dmu_return_arcbuf(buf); XUIOSTAT_BUMP(xuiostat_wbuf_copied); } +} + +void +dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, arc_buf_t *buf, + dmu_tx_t *tx) +{ + dmu_buf_impl_t *dbuf = (dmu_buf_impl_t *)handle; + + DB_DNODE_ENTER(dbuf); + dmu_assign_arcbuf_dnode(DB_DNODE(dbuf), offset, buf, tx); + DB_DNODE_EXIT(dbuf); } typedef struct { Modified: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h ============================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h Wed Oct 3 14:45:48 2018 (r339127) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu.h Wed Oct 3 14:46:25 2018 (r339128) @@ -519,6 +519,9 @@ uint64_t dmu_buf_refcount(dmu_buf_t *db); int dmu_buf_hold_array_by_bonus(dmu_buf_t *db, uint64_t offset, uint64_t length, boolean_t read, void *tag, int *numbufsp, dmu_buf_t ***dbpp); +int dmu_buf_hold_array_by_dnode(dnode_t *dn, uint64_t offset, uint64_t length, + boolean_t read, void *tag, int *numbufsp, dmu_buf_t ***dbpp, + uint32_t flags); void dmu_buf_rele_array(dmu_buf_t **, int numbufs, void *tag); typedef void dmu_buf_evict_func_t(void *user_ptr); @@ -757,10 +760,13 @@ void dmu_prealloc(objset_t *os, uint64_t object, uint6 dmu_tx_t *tx); int dmu_read_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size); int dmu_read_uio_dbuf(dmu_buf_t *zdb, struct uio *uio, uint64_t size); +int dmu_read_uio_dnode(dnode_t *dn, struct uio *uio, uint64_t size); int dmu_write_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size, dmu_tx_t *tx); int dmu_write_uio_dbuf(dmu_buf_t *zdb, struct uio *uio, uint64_t size, dmu_tx_t *tx); +int dmu_write_uio_dnode(dnode_t *dn, struct uio *uio, uint64_t size, + dmu_tx_t *tx); #ifdef _KERNEL #ifdef illumos int dmu_write_pages(objset_t *os, uint64_t object, uint64_t offset, @@ -774,6 +780,8 @@ int dmu_read_pages(objset_t *os, uint64_t object, vm_p #endif struct arc_buf *dmu_request_arcbuf(dmu_buf_t *handle, int size); void dmu_return_arcbuf(struct arc_buf *buf); +void dmu_assign_arcbuf_dnode(dnode_t *handle, uint64_t offset, + struct arc_buf *buf, dmu_tx_t *tx); void dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, struct arc_buf *buf, dmu_tx_t *tx); int dmu_xuio_init(struct xuio *uio, int niov); Modified: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zvol.c ============================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zvol.c Wed Oct 3 14:45:48 2018 (r339127) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zvol.c Wed Oct 3 14:46:25 2018 (r339128) @@ -174,7 +174,7 @@ typedef struct zvol_state { zilog_t *zv_zilog; /* ZIL handle */ list_t zv_extents; /* List of extents for dump */ znode_t zv_znode; /* for range locking */ - dmu_buf_t *zv_dbuf; /* bonus handle */ + dnode_t *zv_dn; /* dnode hold */ #ifndef illumos int zv_state; int zv_volmode; /* Provide GEOM or cdev */ @@ -868,7 +868,7 @@ zvol_first_open(zvol_state_t *zv) } zv->zv_volblocksize = doi.doi_data_block_size; - error = dmu_bonus_hold(os, ZVOL_OBJ, zvol_tag, &zv->zv_dbuf); + error = dnode_hold(os, ZVOL_OBJ, zvol_tag, &zv->zv_dn); if (error) { dmu_objset_disown(os, zvol_tag); return (error); @@ -893,8 +893,8 @@ zvol_last_close(zvol_state_t *zv) zil_close(zv->zv_zilog); zv->zv_zilog = NULL; - dmu_buf_rele(zv->zv_dbuf, zvol_tag); - zv->zv_dbuf = NULL; + dnode_rele(zv->zv_dn, zvol_tag); + zv->zv_dn = NULL; /* * Evict cached data @@ -1342,8 +1342,6 @@ static int zvol_get_data(void *arg, lr_write_t *lr, char *buf, struct lwb *lwb, zio_t *zio) { zvol_state_t *zv = arg; - objset_t *os = zv->zv_objset; - uint64_t object = ZVOL_OBJ; uint64_t offset = lr->lr_offset; uint64_t size = lr->lr_length; /* length of user data */ dmu_buf_t *db; @@ -1367,7 +1365,7 @@ zvol_get_data(void *arg, lr_write_t *lr, char *buf, st if (buf != NULL) { /* immediate write */ zgd->zgd_rl = zfs_range_lock(&zv->zv_znode, offset, size, RL_READER); - error = dmu_read(os, object, offset, size, buf, + error = dmu_read_by_dnode(zv->zv_dn, offset, size, buf, DMU_READ_NO_PREFETCH); } else { /* indirect write */ /* @@ -1380,7 +1378,7 @@ zvol_get_data(void *arg, lr_write_t *lr, char *buf, st offset = P2ALIGN(offset, size); zgd->zgd_rl = zfs_range_lock(&zv->zv_znode, offset, size, RL_READER); - error = dmu_buf_hold(os, object, offset, zgd, &db, + error = dmu_buf_hold_by_dnode(zv->zv_dn, offset, zgd, &db, DMU_READ_NO_PREFETCH); if (error == 0) { blkptr_t *bp = &lr->lr_blkptr; @@ -1451,8 +1449,8 @@ zvol_log_write(zvol_state_t *zv, dmu_tx_t *tx, offset_ itx = zil_itx_create(TX_WRITE, sizeof (*lr) + (wr_state == WR_COPIED ? len : 0)); lr = (lr_write_t *)&itx->itx_lr; - if (wr_state == WR_COPIED && dmu_read(zv->zv_objset, - ZVOL_OBJ, off, len, lr + 1, DMU_READ_NO_PREFETCH) != 0) { + if (wr_state == WR_COPIED && dmu_read_by_dnode(zv->zv_dn, + off, len, lr + 1, DMU_READ_NO_PREFETCH) != 0) { zil_itx_destroy(itx); itx = zil_itx_create(TX_WRITE, sizeof (*lr)); lr = (lr_write_t *)&itx->itx_lr; @@ -1874,7 +1872,7 @@ zvol_read(struct cdev *dev, struct uio *uio, int iofla if (bytes > volsize - uio->uio_loffset) bytes = volsize - uio->uio_loffset; - error = dmu_read_uio_dbuf(zv->zv_dbuf, uio, bytes); + error = dmu_read_uio_dnode(zv->zv_dn, uio, bytes); if (error) { /* convert checksum errors into IO errors */ if (error == ECKSUM) @@ -1946,7 +1944,7 @@ zvol_write(struct cdev *dev, struct uio *uio, int iofl dmu_tx_abort(tx); break; } - error = dmu_write_uio_dbuf(zv->zv_dbuf, uio, bytes, tx); + error = dmu_write_uio_dnode(zv->zv_dn, uio, bytes, tx); if (error == 0) zvol_log_write(zv, tx, off, bytes, sync); dmu_tx_commit(tx); @@ -2028,7 +2026,7 @@ zvol_getefi(void *arg, int flag, uint64_t vs, uint8_t int zvol_get_volume_params(minor_t minor, uint64_t *blksize, uint64_t *max_xfer_len, void **minor_hdl, void **objset_hdl, void **zil_hdl, - void **rl_hdl, void **bonus_hdl) + void **rl_hdl, void **dnode_hdl) { zvol_state_t *zv; @@ -2039,7 +2037,7 @@ zvol_get_volume_params(minor_t minor, uint64_t *blksiz return (SET_ERROR(ENXIO)); ASSERT(blksize && max_xfer_len && minor_hdl && - objset_hdl && zil_hdl && rl_hdl && bonus_hdl); + objset_hdl && zil_hdl && rl_hdl && dnode_hdl); *blksize = zv->zv_volblocksize; *max_xfer_len = (uint64_t)zvol_maxphys; @@ -2047,7 +2045,7 @@ zvol_get_volume_params(minor_t minor, uint64_t *blksiz *objset_hdl = zv->zv_objset; *zil_hdl = zv->zv_zilog; *rl_hdl = &zv->zv_znode; - *bonus_hdl = zv->zv_dbuf; + *dnode_hdl = zv->zv_dn; return (0); }