From c79d1bae75376e2d66051a3cfbeae53e3df5775d Mon Sep 17 00:00:00 2001 From: Umer Saleem Date: Tue, 25 Jul 2023 21:01:27 +0500 Subject: [PATCH 01/13] Update changelog for OpenZFS 2.2.0 release This commit updates changelog for native Debian packages for OpenZFS 2.2.0 release. Reviewed-by: Brian Behlendorf Signed-off-by: Umer Saleem Closes #15104 --- contrib/debian/changelog | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/contrib/debian/changelog b/contrib/debian/changelog index 6273d60383..ba42ea59fa 100644 --- a/contrib/debian/changelog +++ b/contrib/debian/changelog @@ -1,3 +1,9 @@ +openzfs-linux (2.2.0-0) unstable; urgency=low + + * OpenZFS 2.2.0 is tagged. + + -- Umer Saleem Tue, 25 Jul 2023 15:00:00 +0500 + openzfs-linux (2.1.99-1) unstable; urgency=low * Integrate minimally modified Debian packaging from ZFS on Linux From 41a0f66279b19a144ec6a9615c514674e208e60b Mon Sep 17 00:00:00 2001 From: Alexander Motin Date: Mon, 24 Jul 2023 16:41:11 -0400 Subject: [PATCH 02/13] ZIL: Fix config lock deadlock. When we have some LWBs closed and their ZIOs ready to be issued, we can not afford sleeping on config lock if somebody else try to lock it as writer, or it will cause a deadlock. To solve it, move spa_config_enter() from zil_lwb_write_issue() to zil_lwb_write_close() under zl_issuer_lock to enforce lock ordering with other threads. Now if we can't immediately lock config, issue all previously closed LWBs so that they could drop their config locks after completion, and only then allow sleeping on our lock. Reviewed-by: Mark Maybee Reviewed-by: Prakash Surya Reviewed-by: George Wilson Signed-off-by: Alexander Motin Sponsored by: iXsystems, Inc. Closes #15078 Closes #15080 --- module/zfs/zil.c | 34 +++++++++++++++++++++++++++------- 1 file changed, 27 insertions(+), 7 deletions(-) diff --git a/module/zfs/zil.c b/module/zfs/zil.c index 00d66a2481..af7137faac 100644 --- a/module/zfs/zil.c +++ b/module/zfs/zil.c @@ -151,6 +151,7 @@ static kmem_cache_t *zil_lwb_cache; static kmem_cache_t *zil_zcw_cache; static void zil_lwb_commit(zilog_t *zilog, lwb_t *lwb, itx_t *itx); +static void zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb); static itx_t *zil_itx_clone(itx_t *oitx); static int @@ -1768,7 +1769,7 @@ static uint_t zil_maxblocksize = SPA_OLD_MAXBLOCKSIZE; * Has to be called under zl_issuer_lock to chain more lwbs. */ static lwb_t * -zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb) +zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb, list_t *ilwbs) { lwb_t *nlwb = NULL; zil_chain_t *zilc; @@ -1870,6 +1871,27 @@ zil_lwb_write_close(zilog_t *zilog, lwb_t *lwb) dmu_tx_commit(tx); + /* + * We need to acquire the config lock for the lwb to issue it later. + * However, if we already have a queue of closed parent lwbs already + * holding the config lock (but not yet issued), we can't block here + * waiting on the lock or we will deadlock. In that case we must + * first issue to parent IOs before waiting on the lock. + */ + if (ilwbs && !list_is_empty(ilwbs)) { + if (!spa_config_tryenter(spa, SCL_STATE, lwb, RW_READER)) { + lwb_t *tlwb; + while ((tlwb = list_remove_head(ilwbs)) != NULL) + zil_lwb_write_issue(zilog, tlwb); + spa_config_enter(spa, SCL_STATE, lwb, RW_READER); + } + } else { + spa_config_enter(spa, SCL_STATE, lwb, RW_READER); + } + + if (ilwbs) + list_insert_tail(ilwbs, lwb); + /* * If there was an allocation failure then nlwb will be null which * forces a txg_wait_synced(). @@ -1933,7 +1955,7 @@ zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb) ZIL_STAT_INCR(zilog, zil_itx_metaslab_normal_alloc, BP_GET_LSIZE(&lwb->lwb_blk)); } - spa_config_enter(zilog->zl_spa, SCL_STATE, lwb, RW_READER); + ASSERT(spa_config_held(zilog->zl_spa, SCL_STATE, RW_READER)); zil_lwb_add_block(lwb, &lwb->lwb_blk); lwb->lwb_issued_timestamp = gethrtime(); zio_nowait(lwb->lwb_root_zio); @@ -2037,8 +2059,7 @@ cont: lwb_sp < zil_max_waste_space(zilog) && (dlen % max_log_data == 0 || lwb_sp < reclen + dlen % max_log_data))) { - list_insert_tail(ilwbs, lwb); - lwb = zil_lwb_write_close(zilog, lwb); + lwb = zil_lwb_write_close(zilog, lwb, ilwbs); if (lwb == NULL) return (NULL); zil_lwb_write_open(zilog, lwb); @@ -2937,8 +2958,7 @@ zil_process_commit_list(zilog_t *zilog, zil_commit_waiter_t *zcw, list_t *ilwbs) zfs_commit_timeout_pct / 100; if (sleep < zil_min_commit_timeout || lwb->lwb_sz - lwb->lwb_nused < lwb->lwb_sz / 8) { - list_insert_tail(ilwbs, lwb); - lwb = zil_lwb_write_close(zilog, lwb); + lwb = zil_lwb_write_close(zilog, lwb, ilwbs); zilog->zl_cur_used = 0; if (lwb == NULL) { while ((lwb = list_remove_head(ilwbs)) @@ -3096,7 +3116,7 @@ zil_commit_waiter_timeout(zilog_t *zilog, zil_commit_waiter_t *zcw) * since we've reached the commit waiter's timeout and it still * hasn't been issued. */ - lwb_t *nlwb = zil_lwb_write_close(zilog, lwb); + lwb_t *nlwb = zil_lwb_write_close(zilog, lwb, NULL); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); From 991834f5dcff9d515fce3b74a83a40ffc626e012 Mon Sep 17 00:00:00 2001 From: Alexander Motin Date: Tue, 25 Jul 2023 12:08:36 -0400 Subject: [PATCH 03/13] Remove zl_issuer_lock from zil_suspend(). This locking was recently added as part of #14979. But appears it is illegal to take zl_issuer_lock while holding dp_config_rwlock, taken by dsl_pool_hold(). It causes deadlock with sync thread in spa_sync_upgrades(). On a second thought, we should not need this locking, since zil_commit_impl() we call below takes zl_issuer_lock, that should sufficiently protect zl_suspend reads, combined with other logic from #14979. Reviewed-by: Brian Behlendorf Signed-off-by: Alexander Motin Sponsored by: iXsystems, Inc. Closes #15103 --- module/zfs/zil.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/module/zfs/zil.c b/module/zfs/zil.c index af7137faac..be5b9edf6e 100644 --- a/module/zfs/zil.c +++ b/module/zfs/zil.c @@ -3941,13 +3941,11 @@ zil_suspend(const char *osname, void **cookiep) return (error); zilog = dmu_objset_zil(os); - mutex_enter(&zilog->zl_issuer_lock); mutex_enter(&zilog->zl_lock); zh = zilog->zl_header; if (zh->zh_flags & ZIL_REPLAY_NEEDED) { /* unplayed log */ mutex_exit(&zilog->zl_lock); - mutex_exit(&zilog->zl_issuer_lock); dmu_objset_rele(os, suspend_tag); return (SET_ERROR(EBUSY)); } @@ -3961,7 +3959,6 @@ zil_suspend(const char *osname, void **cookiep) if (cookiep == NULL && !zilog->zl_suspending && (zilog->zl_suspend > 0 || BP_IS_HOLE(&zh->zh_log))) { mutex_exit(&zilog->zl_lock); - mutex_exit(&zilog->zl_issuer_lock); dmu_objset_rele(os, suspend_tag); return (0); } @@ -3970,7 +3967,6 @@ zil_suspend(const char *osname, void **cookiep) dsl_pool_rele(dmu_objset_pool(os), suspend_tag); zilog->zl_suspend++; - mutex_exit(&zilog->zl_issuer_lock); if (zilog->zl_suspend > 1) { /* From 571762b29008f6bd30ac3ea2484752ebc8eeb327 Mon Sep 17 00:00:00 2001 From: Brian Behlendorf Date: Mon, 24 Jul 2023 11:20:42 -0700 Subject: [PATCH 04/13] Linux 6.4 compat: META Update the META file to reflect compatibility with the 6.4 kernel. Reviewed-by: George Melikov Reviewed-by: Rob Norris Signed-off-by: Brian Behlendorf Closes #15095 --- META | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/META b/META index f17be8f36e..456ee81229 100644 --- a/META +++ b/META @@ -6,5 +6,5 @@ Release: rc2 Release-Tags: relext License: CDDL Author: OpenZFS -Linux-Maximum: 6.3 +Linux-Maximum: 6.4 Linux-Minimum: 3.10 From b9aa32ff3950b38e86e369bed59b9819433ee804 Mon Sep 17 00:00:00 2001 From: Brian Behlendorf Date: Tue, 25 Jul 2023 13:55:29 -0700 Subject: [PATCH 05/13] zed: Reduce log noise for large JBODs For large JBODs the log message "zfs_iter_vdev: no match" can account for the bulk of the log messages (over 70%). Since this message is purely informational and not that useful we remove it. Reviewed-by: Olaf Faaland Reviewed-by: Brian Atkinson Signed-off-by: Brian Behlendorf Closes #15086 Closes #15094 --- cmd/zed/agents/zfs_mod.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/cmd/zed/agents/zfs_mod.c b/cmd/zed/agents/zfs_mod.c index b07a027122..a8d084bb4b 100644 --- a/cmd/zed/agents/zfs_mod.c +++ b/cmd/zed/agents/zfs_mod.c @@ -607,8 +607,6 @@ zfs_iter_vdev(zpool_handle_t *zhp, nvlist_t *nvl, void *data) */ if (nvlist_lookup_string(nvl, dp->dd_prop, &path) != 0 || strcmp(dp->dd_compare, path) != 0) { - zed_log_msg(LOG_INFO, " %s: no match (%s != vdev %s)", - __func__, dp->dd_compare, path); return; } if (dp->dd_new_vdev_guid != 0 && dp->dd_new_vdev_guid != guid) { From 7698503dcae0acf1e2c74351b258d2ed5c231eaa Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Thu, 22 Jun 2023 13:44:00 +1000 Subject: [PATCH 06/13] zfs_clone_range: use vmem_malloc for large allocation Just silencing the warning about large allocations. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- module/zfs/zfs_vnops.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/module/zfs/zfs_vnops.c b/module/zfs/zfs_vnops.c index 7bdcc16393..3ebd2d0ff7 100644 --- a/module/zfs/zfs_vnops.c +++ b/module/zfs/zfs_vnops.c @@ -1212,7 +1212,7 @@ zfs_clone_range(znode_t *inzp, uint64_t *inoffp, znode_t *outzp, gid = KGID_TO_SGID(ZTOGID(outzp)); projid = outzp->z_projid; - bps = kmem_alloc(sizeof (bps[0]) * maxblocks, KM_SLEEP); + bps = vmem_alloc(sizeof (bps[0]) * maxblocks, KM_SLEEP); /* * Clone the file in reasonable size chunks. Each chunk is cloned @@ -1330,7 +1330,7 @@ zfs_clone_range(znode_t *inzp, uint64_t *inoffp, znode_t *outzp, done += size; } - kmem_free(bps, sizeof (bps[0]) * maxblocks); + vmem_free(bps, sizeof (bps[0]) * maxblocks); zfs_znode_update_vfs(outzp); unlock: From 8aa4f0f0fc16bfecf47cc36fd88a7c360ffe0953 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Thu, 22 Jun 2023 13:44:00 +1000 Subject: [PATCH 07/13] brt_vdev_realloc: use vmem_alloc for large allocation bv_entcount can be a relatively large allocation (see comment for BRT_RANGESIZE), so get it from the big allocator. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- module/zfs/brt.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/module/zfs/brt.c b/module/zfs/brt.c index 99bd472d6f..877b503a1b 100644 --- a/module/zfs/brt.c +++ b/module/zfs/brt.c @@ -680,7 +680,7 @@ brt_vdev_realloc(brt_t *brt, brt_vdev_t *brtvd) size = (vdev_get_min_asize(vd) - 1) / brt->brt_rangesize + 1; spa_config_exit(brt->brt_spa, SCL_VDEV, FTAG); - entcount = kmem_zalloc(sizeof (entcount[0]) * size, KM_SLEEP); + entcount = vmem_zalloc(sizeof (entcount[0]) * size, KM_SLEEP); nblocks = BRT_RANGESIZE_TO_NBLOCKS(size); bitmap = kmem_zalloc(BT_SIZEOFMAP(nblocks), KM_SLEEP); @@ -709,7 +709,7 @@ brt_vdev_realloc(brt_t *brt, brt_vdev_t *brtvd) sizeof (entcount[0]) * MIN(size, brtvd->bv_size)); memcpy(bitmap, brtvd->bv_bitmap, MIN(BT_SIZEOFMAP(nblocks), BT_SIZEOFMAP(brtvd->bv_nblocks))); - kmem_free(brtvd->bv_entcount, + vmem_free(brtvd->bv_entcount, sizeof (entcount[0]) * brtvd->bv_size); kmem_free(brtvd->bv_bitmap, BT_SIZEOFMAP(brtvd->bv_nblocks)); } @@ -792,7 +792,7 @@ brt_vdev_dealloc(brt_t *brt, brt_vdev_t *brtvd) ASSERT(RW_WRITE_HELD(&brt->brt_lock)); ASSERT(brtvd->bv_initiated); - kmem_free(brtvd->bv_entcount, sizeof (uint16_t) * brtvd->bv_size); + vmem_free(brtvd->bv_entcount, sizeof (uint16_t) * brtvd->bv_size); brtvd->bv_entcount = NULL; kmem_free(brtvd->bv_bitmap, BT_SIZEOFMAP(brtvd->bv_nblocks)); brtvd->bv_bitmap = NULL; From 0426e13271306f3c776eea0a6c938d23f3d65a75 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Mon, 24 Jul 2023 17:54:05 +1000 Subject: [PATCH 08/13] dmu_buf_will_clone: only check that current txg is clean dbuf_undirty() will (correctly) only removed dirty records for the given (open) txg. If there is a dirty record for an earlier closed txg that has not been synced out yet, then db_dirty_records will still have entries on it, tripping the assertion. Instead, change the assertion to only consider the current txg. To some extent this is redundant, as its really just saying "did dbuf_undirty() work?", but it it doesn't hurt and accurately expresses our expectations. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Original-patch-by: Kay Pedersen Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- module/zfs/dbuf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c index 1ea075217f..fbeac866ae 100644 --- a/module/zfs/dbuf.c +++ b/module/zfs/dbuf.c @@ -2701,7 +2701,7 @@ dmu_buf_will_clone(dmu_buf_t *db_fake, dmu_tx_t *tx) */ mutex_enter(&db->db_mtx); VERIFY(!dbuf_undirty(db, tx)); - ASSERT(list_head(&db->db_dirty_records) == NULL); + ASSERT0(dbuf_find_dirty_eq(db, tx->tx_txg)); if (db->db_buf != NULL) { arc_buf_destroy(db->db_buf, db); db->db_buf = NULL; From a3ea8c8ee6d01aba0d15b4619b8641376fef57e2 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Mon, 24 Jul 2023 18:02:21 +1000 Subject: [PATCH 09/13] dbuf_sync_leaf: check DB_READ in state assertions Block cloning introduced a new state transition from DB_NOFILL to DB_READ. This occurs when a block is cloned and then read on the current txg. In this case, the clone will move the dbuf to DB_NOFILL, and then the read will be issued for the overidden block pointer. If that read is still outstanding when it comes time to write, the dbuf will be in DB_READ, which is not handled by the checks in dbuf_sync_leaf, thus tripping the assertions. This updates those checks to allow DB_READ as a valid state iff the dirty record is for a BRT write and there is a override block pointer. This is a safe situation because the block already exists, so there's nothing that could change from underneath the read. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Original-patch-by: Kay Pedersen Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- module/zfs/dbuf.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c index fbeac866ae..b7453578a7 100644 --- a/module/zfs/dbuf.c +++ b/module/zfs/dbuf.c @@ -4457,6 +4457,15 @@ dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx) } else if (db->db_state == DB_FILL) { /* This buffer was freed and is now being re-filled */ ASSERT(db->db.db_data != dr->dt.dl.dr_data); + } else if (db->db_state == DB_READ) { + /* + * This buffer has a clone we need to write, and an in-flight + * read on the BP we're about to clone. Its safe to issue the + * write here because the read has already been issued and the + * contents won't change. + */ + ASSERT(dr->dt.dl.dr_brtwrite && + dr->dt.dl.dr_override_state == DR_OVERRIDDEN); } else { ASSERT(db->db_state == DB_CACHED || db->db_state == DB_NOFILL); } From 5d12545da8c112aa813560950f39315956338963 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Tue, 27 Jun 2023 23:44:53 +1000 Subject: [PATCH 10/13] linux: implement filesystem-side copy/clone functions This implements the Linux VFS ops required to service the file copy/clone APIs: .copy_file_range (4.5+) .clone_file_range (4.5-4.19) .dedupe_file_range (4.5-4.19) .remap_file_range (4.20+) Note that dedupe_file_range() and remap_file_range(REMAP_FILE_DEDUP) are hooked up here, but are not implemented yet. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- config/kernel-vfs-file_range.m4 | 164 ++++++++++++++++++++++++ config/kernel.m4 | 10 ++ include/os/linux/zfs/sys/zpl.h | 14 ++ module/Kbuild.in | 1 + module/os/linux/zfs/zpl_file.c | 13 +- module/os/linux/zfs/zpl_file_range.c | 183 +++++++++++++++++++++++++++ 6 files changed, 384 insertions(+), 1 deletion(-) create mode 100644 config/kernel-vfs-file_range.m4 create mode 100644 module/os/linux/zfs/zpl_file_range.c diff --git a/config/kernel-vfs-file_range.m4 b/config/kernel-vfs-file_range.m4 new file mode 100644 index 0000000000..cc96404d8b --- /dev/null +++ b/config/kernel-vfs-file_range.m4 @@ -0,0 +1,164 @@ +dnl # +dnl # The *_file_range APIs have a long history: +dnl # +dnl # 2.6.29: BTRFS_IOC_CLONE and BTRFS_IOC_CLONE_RANGE ioctl introduced +dnl # 3.12: BTRFS_IOC_FILE_EXTENT_SAME ioctl introduced +dnl # +dnl # 4.5: copy_file_range() syscall introduced, added to VFS +dnl # 4.5: BTRFS_IOC_CLONE and BTRFS_IOC_CLONE_RANGE renamed to FICLONE ands +dnl # FICLONERANGE, added to VFS as clone_file_range() +dnl # 4.5: BTRFS_IOC_FILE_EXTENT_SAME renamed to FIDEDUPERANGE, added to VFS +dnl # as dedupe_file_range() +dnl # +dnl # 4.20: VFS clone_file_range() and dedupe_file_range() replaced by +dnl # remap_file_range() +dnl # +dnl # 5.3: VFS copy_file_range() expected to do its own fallback, +dnl # generic_copy_file_range() added to support it +dnl # +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_COPY_FILE_RANGE], [ + ZFS_LINUX_TEST_SRC([vfs_copy_file_range], [ + #include + + static ssize_t test_copy_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + size_t len, unsigned int flags) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; (void) flags; + return (0); + } + + static const struct file_operations + fops __attribute__ ((unused)) = { + .copy_file_range = test_copy_file_range, + }; + ],[]) +]) +AC_DEFUN([ZFS_AC_KERNEL_VFS_COPY_FILE_RANGE], [ + AC_MSG_CHECKING([whether fops->copy_file_range() is available]) + ZFS_LINUX_TEST_RESULT([vfs_copy_file_range], [ + AC_MSG_RESULT([yes]) + AC_DEFINE(HAVE_VFS_COPY_FILE_RANGE, 1, + [fops->copy_file_range() is available]) + ],[ + AC_MSG_RESULT([no]) + ]) +]) + +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_GENERIC_COPY_FILE_RANGE], [ + ZFS_LINUX_TEST_SRC([generic_copy_file_range], [ + #include + ], [ + struct file *src_file __attribute__ ((unused)) = NULL; + loff_t src_off __attribute__ ((unused)) = 0; + struct file *dst_file __attribute__ ((unused)) = NULL; + loff_t dst_off __attribute__ ((unused)) = 0; + size_t len __attribute__ ((unused)) = 0; + unsigned int flags __attribute__ ((unused)) = 0; + generic_copy_file_range(src_file, src_off, dst_file, dst_off, + len, flags); + ]) +]) +AC_DEFUN([ZFS_AC_KERNEL_VFS_GENERIC_COPY_FILE_RANGE], [ + AC_MSG_CHECKING([whether generic_copy_file_range() is available]) + ZFS_LINUX_TEST_RESULT_SYMBOL([generic_copy_file_range], + [generic_copy_file_range], [fs/read_write.c], [ + AC_MSG_RESULT(yes) + AC_DEFINE(HAVE_VFS_GENERIC_COPY_FILE_RANGE, 1, + [generic_copy_file_range() is available]) + ],[ + AC_MSG_RESULT(no) + ]) +]) + +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_CLONE_FILE_RANGE], [ + ZFS_LINUX_TEST_SRC([vfs_clone_file_range], [ + #include + + static int test_clone_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + u64 len) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; + return (0); + } + + static const struct file_operations + fops __attribute__ ((unused)) = { + .clone_file_range = test_clone_file_range, + }; + ],[]) +]) +AC_DEFUN([ZFS_AC_KERNEL_VFS_CLONE_FILE_RANGE], [ + AC_MSG_CHECKING([whether fops->clone_file_range() is available]) + ZFS_LINUX_TEST_RESULT([vfs_clone_file_range], [ + AC_MSG_RESULT([yes]) + AC_DEFINE(HAVE_VFS_CLONE_FILE_RANGE, 1, + [fops->clone_file_range() is available]) + ],[ + AC_MSG_RESULT([no]) + ]) +]) + +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_DEDUPE_FILE_RANGE], [ + ZFS_LINUX_TEST_SRC([vfs_dedupe_file_range], [ + #include + + static int test_dedupe_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + u64 len) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; + return (0); + } + + static const struct file_operations + fops __attribute__ ((unused)) = { + .dedupe_file_range = test_dedupe_file_range, + }; + ],[]) +]) +AC_DEFUN([ZFS_AC_KERNEL_VFS_DEDUPE_FILE_RANGE], [ + AC_MSG_CHECKING([whether fops->dedupe_file_range() is available]) + ZFS_LINUX_TEST_RESULT([vfs_dedupe_file_range], [ + AC_MSG_RESULT([yes]) + AC_DEFINE(HAVE_VFS_DEDUPE_FILE_RANGE, 1, + [fops->dedupe_file_range() is available]) + ],[ + AC_MSG_RESULT([no]) + ]) +]) + +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_REMAP_FILE_RANGE], [ + ZFS_LINUX_TEST_SRC([vfs_remap_file_range], [ + #include + + static loff_t test_remap_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + loff_t len, unsigned int flags) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; (void) flags; + return (0); + } + + static const struct file_operations + fops __attribute__ ((unused)) = { + .remap_file_range = test_remap_file_range, + }; + ],[]) +]) + +AC_DEFUN([ZFS_AC_KERNEL_VFS_REMAP_FILE_RANGE], [ + AC_MSG_CHECKING([whether fops->remap_file_range() is available]) + ZFS_LINUX_TEST_RESULT([vfs_remap_file_range], [ + AC_MSG_RESULT([yes]) + AC_DEFINE(HAVE_VFS_REMAP_FILE_RANGE, 1, + [fops->remap_file_range() is available]) + ],[ + AC_MSG_RESULT([no]) + ]) +]) diff --git a/config/kernel.m4 b/config/kernel.m4 index cb7e736c9a..b17ccfdeec 100644 --- a/config/kernel.m4 +++ b/config/kernel.m4 @@ -116,6 +116,11 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [ ZFS_AC_KERNEL_SRC_VFS_RW_ITERATE ZFS_AC_KERNEL_SRC_VFS_GENERIC_WRITE_CHECKS ZFS_AC_KERNEL_SRC_VFS_IOV_ITER + ZFS_AC_KERNEL_SRC_VFS_COPY_FILE_RANGE + ZFS_AC_KERNEL_SRC_VFS_GENERIC_COPY_FILE_RANGE + ZFS_AC_KERNEL_SRC_VFS_REMAP_FILE_RANGE + ZFS_AC_KERNEL_SRC_VFS_CLONE_FILE_RANGE + ZFS_AC_KERNEL_SRC_VFS_DEDUPE_FILE_RANGE ZFS_AC_KERNEL_SRC_KMAP_ATOMIC_ARGS ZFS_AC_KERNEL_SRC_FOLLOW_DOWN_ONE ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN @@ -249,6 +254,11 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [ ZFS_AC_KERNEL_VFS_RW_ITERATE ZFS_AC_KERNEL_VFS_GENERIC_WRITE_CHECKS ZFS_AC_KERNEL_VFS_IOV_ITER + ZFS_AC_KERNEL_VFS_COPY_FILE_RANGE + ZFS_AC_KERNEL_VFS_GENERIC_COPY_FILE_RANGE + ZFS_AC_KERNEL_VFS_REMAP_FILE_RANGE + ZFS_AC_KERNEL_VFS_CLONE_FILE_RANGE + ZFS_AC_KERNEL_VFS_DEDUPE_FILE_RANGE ZFS_AC_KERNEL_KMAP_ATOMIC_ARGS ZFS_AC_KERNEL_FOLLOW_DOWN_ONE ZFS_AC_KERNEL_MAKE_REQUEST_FN diff --git a/include/os/linux/zfs/sys/zpl.h b/include/os/linux/zfs/sys/zpl.h index 2b302e9dab..8b0e79afb0 100644 --- a/include/os/linux/zfs/sys/zpl.h +++ b/include/os/linux/zfs/sys/zpl.h @@ -180,6 +180,20 @@ zpl_dir_emit_dots(struct file *file, zpl_dir_context_t *ctx) } #endif /* HAVE_VFS_ITERATE */ + +/* zpl_file_range.c */ + +/* handlers for file_operations of the same name */ +extern ssize_t zpl_copy_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, size_t len, unsigned int flags); +extern loff_t zpl_remap_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, loff_t len, unsigned int flags); +extern int zpl_clone_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, uint64_t len); +extern int zpl_dedupe_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, uint64_t len); + + #if defined(HAVE_INODE_TIMESTAMP_TRUNCATE) #define zpl_inode_timestamp_truncate(ts, ip) timestamp_truncate(ts, ip) #elif defined(HAVE_INODE_TIMESPEC64_TIMES) diff --git a/module/Kbuild.in b/module/Kbuild.in index 485331ac65..c132171592 100644 --- a/module/Kbuild.in +++ b/module/Kbuild.in @@ -461,6 +461,7 @@ ZFS_OBJS_OS := \ zpl_ctldir.o \ zpl_export.o \ zpl_file.o \ + zpl_file_range.o \ zpl_inode.o \ zpl_super.o \ zpl_xattr.o \ diff --git a/module/os/linux/zfs/zpl_file.c b/module/os/linux/zfs/zpl_file.c index e690525d3c..92b603e98a 100644 --- a/module/os/linux/zfs/zpl_file.c +++ b/module/os/linux/zfs/zpl_file.c @@ -1283,7 +1283,6 @@ zpl_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) } #endif /* CONFIG_COMPAT */ - const struct address_space_operations zpl_address_space_operations = { #ifdef HAVE_VFS_READPAGES .readpages = zpl_readpages, @@ -1333,6 +1332,18 @@ const struct file_operations zpl_file_operations = { .aio_fsync = zpl_aio_fsync, #endif .fallocate = zpl_fallocate, +#ifdef HAVE_VFS_COPY_FILE_RANGE + .copy_file_range = zpl_copy_file_range, +#endif +#ifdef HAVE_VFS_REMAP_FILE_RANGE + .remap_file_range = zpl_remap_file_range, +#endif +#ifdef HAVE_VFS_CLONE_FILE_RANGE + .clone_file_range = zpl_clone_file_range, +#endif +#ifdef HAVE_VFS_DEDUPE_FILE_RANGE + .dedupe_file_range = zpl_dedupe_file_range, +#endif #ifdef HAVE_FILE_FADVISE .fadvise = zpl_fadvise, #endif diff --git a/module/os/linux/zfs/zpl_file_range.c b/module/os/linux/zfs/zpl_file_range.c new file mode 100644 index 0000000000..db387a7481 --- /dev/null +++ b/module/os/linux/zfs/zpl_file_range.c @@ -0,0 +1,183 @@ +/* + * 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) 2023, Klara Inc. + */ + +#ifdef CONFIG_COMPAT +#include +#endif +#include +#include +#include +#include +#include + +/* + * Clone part of a file via block cloning. + * + * Note that we are not required to update file offsets; the kernel will take + * care of that depending on how it was called. + */ +static ssize_t +__zpl_clone_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, size_t len) +{ + struct inode *src_i = file_inode(src_file); + struct inode *dst_i = file_inode(dst_file); + uint64_t src_off_o = (uint64_t)src_off; + uint64_t dst_off_o = (uint64_t)dst_off; + uint64_t len_o = (uint64_t)len; + cred_t *cr = CRED(); + fstrans_cookie_t cookie; + int err; + + if (!spa_feature_is_enabled( + dmu_objset_spa(ITOZSB(dst_i)->z_os), SPA_FEATURE_BLOCK_CLONING)) + return (-EOPNOTSUPP); + + if (src_i != dst_i) + spl_inode_lock_shared(src_i); + spl_inode_lock(dst_i); + + crhold(cr); + cookie = spl_fstrans_mark(); + + err = -zfs_clone_range(ITOZ(src_i), &src_off_o, ITOZ(dst_i), + &dst_off_o, &len_o, cr); + + spl_fstrans_unmark(cookie); + crfree(cr); + + spl_inode_unlock(dst_i); + if (src_i != dst_i) + spl_inode_unlock_shared(src_i); + + if (err < 0) + return (err); + + return ((ssize_t)len_o); +} + +#ifdef HAVE_VFS_COPY_FILE_RANGE +/* + * Entry point for copy_file_range(). Copy len bytes from src_off in src_file + * to dst_off in dst_file. We are permitted to do this however we like, so we + * try to just clone the blocks, and if we can't support it, fall back to the + * kernel's generic byte copy function. + */ +ssize_t +zpl_copy_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, size_t len, unsigned int flags) +{ + ssize_t ret; + + if (flags != 0) + return (-EINVAL); + + /* Try to do it via zfs_clone_range() */ + ret =__zpl_clone_file_range(src_file, src_off, + dst_file, dst_off, len); + +#ifdef HAVE_VFS_GENERIC_COPY_FILE_RANGE + /* + * Since Linux 5.3 the filesystem driver is responsible for executing + * an appropriate fallback, and a generic fallback function is provided. + */ + if (ret == -EOPNOTSUPP || ret == -EXDEV) + ret = generic_copy_file_range(src_file, src_off, dst_file, + dst_off, len, flags); +#endif /* HAVE_VFS_GENERIC_COPY_FILE_RANGE */ + + return (ret); +} +#endif /* HAVE_VFS_COPY_FILE_RANGE */ + +#ifdef HAVE_VFS_REMAP_FILE_RANGE +/* + * Entry point for FICLONE/FICLONERANGE/FIDEDUPERANGE. + * + * FICLONE and FICLONERANGE are basically the same as copy_file_range(), except + * that they must clone - they cannot fall back to copying. FICLONE is exactly + * FICLONERANGE, for the entire file. We don't need to try to tell them apart; + * the kernel will sort that out for us. + * + * FIDEDUPERANGE is for turning a non-clone into a clone, that is, compare the + * range in both files and if they're the same, arrange for them to be backed + * by the same storage. + */ +loff_t +zpl_remap_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, loff_t len, unsigned int flags) +{ + if (flags & ~(REMAP_FILE_DEDUP | REMAP_FILE_CAN_SHORTEN)) + return (-EINVAL); + + /* + * REMAP_FILE_CAN_SHORTEN lets us know we can clone less than the given + * range if we want. Its designed for filesystems that make data past + * EOF available, and don't want it to be visible in both files. ZFS + * doesn't do that, so we just turn the flag off. + */ + flags &= ~REMAP_FILE_CAN_SHORTEN; + + if (flags & REMAP_FILE_DEDUP) + /* No support for dedup yet */ + return (-EOPNOTSUPP); + + /* Zero length means to clone everything to the end of the file */ + if (len == 0) + len = i_size_read(file_inode(src_file)) - src_off; + + return (__zpl_clone_file_range(src_file, src_off, + dst_file, dst_off, len)); +} +#endif /* HAVE_VFS_REMAP_FILE_RANGE */ + +#ifdef HAVE_VFS_CLONE_FILE_RANGE +/* + * Entry point for FICLONE and FICLONERANGE, before Linux 4.20. + */ +int +zpl_clone_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, uint64_t len) +{ + /* Zero length means to clone everything to the end of the file */ + if (len == 0) + len = i_size_read(file_inode(src_file)) - src_off; + + return (__zpl_clone_file_range(src_file, src_off, + dst_file, dst_off, len)); +} +#endif /* HAVE_VFS_CLONE_FILE_RANGE */ + +#ifdef HAVE_VFS_DEDUPE_FILE_RANGE +/* + * Entry point for FIDEDUPERANGE, before Linux 4.20. + */ +int +zpl_dedupe_file_range(struct file *src_file, loff_t src_off, + struct file *dst_file, loff_t dst_off, uint64_t len) +{ + /* No support for dedup yet */ + return (-EOPNOTSUPP); +} +#endif /* HAVE_VFS_DEDUPE_FILE_RANGE */ From 3366ceaf3ab7f9f5f96313a2421c7c3418e541dd Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Tue, 27 Jun 2023 23:45:00 +1000 Subject: [PATCH 11/13] linux: implement filesystem-side clone ioctls Prior to Linux 4.5, the FICLONE etc ioctls were specific to BTRFS, and were implemented as regular filesystem-specific ioctls. This implements those ioctls directly in OpenZFS, allowing cloning to work on older kernels. There's no need to gate these behind version checks; on later kernels Linux will simply never deliver these ioctls, instead calling the approprate VFS op. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- include/os/linux/zfs/sys/zpl.h | 35 ++++++++++++ module/os/linux/zfs/zpl_file.c | 6 +++ module/os/linux/zfs/zpl_file_range.c | 79 ++++++++++++++++++++++++++++ 3 files changed, 120 insertions(+) diff --git a/include/os/linux/zfs/sys/zpl.h b/include/os/linux/zfs/sys/zpl.h index 8b0e79afb0..b62ab5eec8 100644 --- a/include/os/linux/zfs/sys/zpl.h +++ b/include/os/linux/zfs/sys/zpl.h @@ -193,6 +193,41 @@ extern int zpl_clone_file_range(struct file *src_file, loff_t src_off, extern int zpl_dedupe_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, uint64_t len); +/* compat for FICLONE/FICLONERANGE/FIDEDUPERANGE ioctls */ +typedef struct { + int64_t fcr_src_fd; + uint64_t fcr_src_offset; + uint64_t fcr_src_length; + uint64_t fcr_dest_offset; +} zfs_ioc_compat_file_clone_range_t; + +typedef struct { + int64_t fdri_dest_fd; + uint64_t fdri_dest_offset; + uint64_t fdri_bytes_deduped; + int32_t fdri_status; + uint32_t fdri_reserved; +} zfs_ioc_compat_dedupe_range_info_t; + +typedef struct { + uint64_t fdr_src_offset; + uint64_t fdr_src_length; + uint16_t fdr_dest_count; + uint16_t fdr_reserved1; + uint32_t fdr_reserved2; + zfs_ioc_compat_dedupe_range_info_t fdr_info[]; +} zfs_ioc_compat_dedupe_range_t; + +#define ZFS_IOC_COMPAT_FICLONE _IOW(0x94, 9, int) +#define ZFS_IOC_COMPAT_FICLONERANGE \ + _IOW(0x94, 13, zfs_ioc_compat_file_clone_range_t) +#define ZFS_IOC_COMPAT_FIDEDUPERANGE \ + _IOWR(0x94, 54, zfs_ioc_compat_dedupe_range_t) + +extern long zpl_ioctl_ficlone(struct file *filp, void *arg); +extern long zpl_ioctl_ficlonerange(struct file *filp, void *arg); +extern long zpl_ioctl_fideduperange(struct file *filp, void *arg); + #if defined(HAVE_INODE_TIMESTAMP_TRUNCATE) #define zpl_inode_timestamp_truncate(ts, ip) timestamp_truncate(ts, ip) diff --git a/module/os/linux/zfs/zpl_file.c b/module/os/linux/zfs/zpl_file.c index 92b603e98a..87a248af83 100644 --- a/module/os/linux/zfs/zpl_file.c +++ b/module/os/linux/zfs/zpl_file.c @@ -1257,6 +1257,12 @@ zpl_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) return (zpl_ioctl_getdosflags(filp, (void *)arg)); case ZFS_IOC_SETDOSFLAGS: return (zpl_ioctl_setdosflags(filp, (void *)arg)); + case ZFS_IOC_COMPAT_FICLONE: + return (zpl_ioctl_ficlone(filp, (void *)arg)); + case ZFS_IOC_COMPAT_FICLONERANGE: + return (zpl_ioctl_ficlonerange(filp, (void *)arg)); + case ZFS_IOC_COMPAT_FIDEDUPERANGE: + return (zpl_ioctl_fideduperange(filp, (void *)arg)); default: return (-ENOTTY); } diff --git a/module/os/linux/zfs/zpl_file_range.c b/module/os/linux/zfs/zpl_file_range.c index db387a7481..aad502a809 100644 --- a/module/os/linux/zfs/zpl_file_range.c +++ b/module/os/linux/zfs/zpl_file_range.c @@ -181,3 +181,82 @@ zpl_dedupe_file_range(struct file *src_file, loff_t src_off, return (-EOPNOTSUPP); } #endif /* HAVE_VFS_DEDUPE_FILE_RANGE */ + +/* Entry point for FICLONE, before Linux 4.5. */ +long +zpl_ioctl_ficlone(struct file *dst_file, void *arg) +{ + unsigned long sfd = (unsigned long)arg; + + struct file *src_file = fget(sfd); + if (src_file == NULL) + return (-EBADF); + + if (dst_file->f_op != src_file->f_op) + return (-EXDEV); + + size_t len = i_size_read(file_inode(src_file)); + + ssize_t ret = + __zpl_clone_file_range(src_file, 0, dst_file, 0, len); + + fput(src_file); + + if (ret < 0) { + if (ret == -EOPNOTSUPP) + return (-ENOTTY); + return (ret); + } + + if (ret != len) + return (-EINVAL); + + return (0); +} + +/* Entry point for FICLONERANGE, before Linux 4.5. */ +long +zpl_ioctl_ficlonerange(struct file *dst_file, void __user *arg) +{ + zfs_ioc_compat_file_clone_range_t fcr; + + if (copy_from_user(&fcr, arg, sizeof (fcr))) + return (-EFAULT); + + struct file *src_file = fget(fcr.fcr_src_fd); + if (src_file == NULL) + return (-EBADF); + + if (dst_file->f_op != src_file->f_op) + return (-EXDEV); + + size_t len = fcr.fcr_src_length; + if (len == 0) + len = i_size_read(file_inode(src_file)) - fcr.fcr_src_offset; + + ssize_t ret = __zpl_clone_file_range(src_file, fcr.fcr_src_offset, + dst_file, fcr.fcr_dest_offset, len); + + fput(src_file); + + if (ret < 0) { + if (ret == -EOPNOTSUPP) + return (-ENOTTY); + return (ret); + } + + if (ret != len) + return (-EINVAL); + + return (0); +} + +/* Entry point for FIDEDUPERANGE, before Linux 4.5. */ +long +zpl_ioctl_fideduperange(struct file *filp, void *arg) +{ + (void) arg; + + /* No support for dedup yet */ + return (-ENOTTY); +} From 2768dc04cc45b2506a5f2f3b7c5da17eb42ec3c3 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Sun, 25 Jun 2023 20:50:19 +1000 Subject: [PATCH 12/13] linux: implement filesystem-side copy/clone functions for EL7 Redhat have backported copy_file_range and clone_file_range to the EL7 kernel using an "extended file operations" wrapper structure. This connects all that up to let cloning work there too. Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 --- config/kernel-vfs-extended-file_range.m4 | 50 ++++++++++++++++++++++++ config/kernel.m4 | 2 + include/os/linux/zfs/sys/zpl.h | 4 ++ module/os/linux/zfs/zfs_vfsops.c | 6 +++ module/os/linux/zfs/zfs_znode.c | 8 ++++ module/os/linux/zfs/zpl_file.c | 16 ++++++-- module/os/linux/zfs/zpl_file_range.c | 12 +++--- 7 files changed, 90 insertions(+), 8 deletions(-) create mode 100644 config/kernel-vfs-extended-file_range.m4 diff --git a/config/kernel-vfs-extended-file_range.m4 b/config/kernel-vfs-extended-file_range.m4 new file mode 100644 index 0000000000..a262231312 --- /dev/null +++ b/config/kernel-vfs-extended-file_range.m4 @@ -0,0 +1,50 @@ +dnl # +dnl # EL7 have backported copy_file_range and clone_file_range and +dnl # added them to an "extended" file_operations struct. +dnl # +dnl # We're testing for both functions in one here, because they will only +dnl # ever appear together and we don't want to match a similar method in +dnl # some future vendor kernel. +dnl # +AC_DEFUN([ZFS_AC_KERNEL_SRC_VFS_FILE_OPERATIONS_EXTEND], [ + ZFS_LINUX_TEST_SRC([vfs_file_operations_extend], [ + #include + + static ssize_t test_copy_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + size_t len, unsigned int flags) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; (void) flags; + return (0); + } + + static int test_clone_file_range(struct file *src_file, + loff_t src_off, struct file *dst_file, loff_t dst_off, + u64 len) { + (void) src_file; (void) src_off; + (void) dst_file; (void) dst_off; + (void) len; + return (0); + } + + static const struct file_operations_extend + fops __attribute__ ((unused)) = { + .kabi_fops = {}, + .copy_file_range = test_copy_file_range, + .clone_file_range = test_clone_file_range, + }; + ],[]) +]) +AC_DEFUN([ZFS_AC_KERNEL_VFS_FILE_OPERATIONS_EXTEND], [ + AC_MSG_CHECKING([whether file_operations_extend takes \ +.copy_file_range() and .clone_file_range()]) + ZFS_LINUX_TEST_RESULT([vfs_file_operations_extend], [ + AC_MSG_RESULT([yes]) + AC_DEFINE(HAVE_VFS_FILE_OPERATIONS_EXTEND, 1, + [file_operations_extend takes .copy_file_range() + and .clone_file_range()]) + ],[ + AC_MSG_RESULT([no]) + ]) +]) diff --git a/config/kernel.m4 b/config/kernel.m4 index b17ccfdeec..1487fa2e77 100644 --- a/config/kernel.m4 +++ b/config/kernel.m4 @@ -121,6 +121,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_SRC], [ ZFS_AC_KERNEL_SRC_VFS_REMAP_FILE_RANGE ZFS_AC_KERNEL_SRC_VFS_CLONE_FILE_RANGE ZFS_AC_KERNEL_SRC_VFS_DEDUPE_FILE_RANGE + ZFS_AC_KERNEL_SRC_VFS_FILE_OPERATIONS_EXTEND ZFS_AC_KERNEL_SRC_KMAP_ATOMIC_ARGS ZFS_AC_KERNEL_SRC_FOLLOW_DOWN_ONE ZFS_AC_KERNEL_SRC_MAKE_REQUEST_FN @@ -259,6 +260,7 @@ AC_DEFUN([ZFS_AC_KERNEL_TEST_RESULT], [ ZFS_AC_KERNEL_VFS_REMAP_FILE_RANGE ZFS_AC_KERNEL_VFS_CLONE_FILE_RANGE ZFS_AC_KERNEL_VFS_DEDUPE_FILE_RANGE + ZFS_AC_KERNEL_VFS_FILE_OPERATIONS_EXTEND ZFS_AC_KERNEL_KMAP_ATOMIC_ARGS ZFS_AC_KERNEL_FOLLOW_DOWN_ONE ZFS_AC_KERNEL_MAKE_REQUEST_FN diff --git a/include/os/linux/zfs/sys/zpl.h b/include/os/linux/zfs/sys/zpl.h index b62ab5eec8..0bd20f6489 100644 --- a/include/os/linux/zfs/sys/zpl.h +++ b/include/os/linux/zfs/sys/zpl.h @@ -52,7 +52,11 @@ extern const struct inode_operations zpl_special_inode_operations; /* zpl_file.c */ extern const struct address_space_operations zpl_address_space_operations; +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND +extern const struct file_operations_extend zpl_file_operations; +#else extern const struct file_operations zpl_file_operations; +#endif extern const struct file_operations zpl_dir_file_operations; /* zpl_super.c */ diff --git a/module/os/linux/zfs/zfs_vfsops.c b/module/os/linux/zfs/zfs_vfsops.c index 87c4e6dcaf..464c12e110 100644 --- a/module/os/linux/zfs/zfs_vfsops.c +++ b/module/os/linux/zfs/zfs_vfsops.c @@ -2092,6 +2092,9 @@ zfs_init(void) zfs_znode_init(); dmu_objset_register_type(DMU_OST_ZFS, zpl_get_file_info); register_filesystem(&zpl_fs_type); +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND + register_fo_extend(&zpl_file_operations); +#endif } void @@ -2102,6 +2105,9 @@ zfs_fini(void) */ taskq_wait(system_delay_taskq); taskq_wait(system_taskq); +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND + unregister_fo_extend(&zpl_file_operations); +#endif unregister_filesystem(&zpl_fs_type); zfs_znode_fini(); zfsctl_fini(); diff --git a/module/os/linux/zfs/zfs_znode.c b/module/os/linux/zfs/zfs_znode.c index 02b1af3edc..335ae3460c 100644 --- a/module/os/linux/zfs/zfs_znode.c +++ b/module/os/linux/zfs/zfs_znode.c @@ -415,7 +415,11 @@ zfs_inode_set_ops(zfsvfs_t *zfsvfs, struct inode *ip) switch (ip->i_mode & S_IFMT) { case S_IFREG: ip->i_op = &zpl_inode_operations; +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND + ip->i_fop = &zpl_file_operations.kabi_fops; +#else ip->i_fop = &zpl_file_operations; +#endif ip->i_mapping->a_ops = &zpl_address_space_operations; break; @@ -455,7 +459,11 @@ zfs_inode_set_ops(zfsvfs_t *zfsvfs, struct inode *ip) /* Assume the inode is a file and attempt to continue */ ip->i_mode = S_IFREG | 0644; ip->i_op = &zpl_inode_operations; +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND + ip->i_fop = &zpl_file_operations.kabi_fops; +#else ip->i_fop = &zpl_file_operations; +#endif ip->i_mapping->a_ops = &zpl_address_space_operations; break; } diff --git a/module/os/linux/zfs/zpl_file.c b/module/os/linux/zfs/zpl_file.c index 87a248af83..73526db731 100644 --- a/module/os/linux/zfs/zpl_file.c +++ b/module/os/linux/zfs/zpl_file.c @@ -1311,7 +1311,12 @@ const struct address_space_operations zpl_address_space_operations = { #endif }; +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND +const struct file_operations_extend zpl_file_operations = { + .kabi_fops = { +#else const struct file_operations zpl_file_operations = { +#endif .open = zpl_open, .release = zpl_release, .llseek = zpl_llseek, @@ -1341,12 +1346,12 @@ const struct file_operations zpl_file_operations = { #ifdef HAVE_VFS_COPY_FILE_RANGE .copy_file_range = zpl_copy_file_range, #endif -#ifdef HAVE_VFS_REMAP_FILE_RANGE - .remap_file_range = zpl_remap_file_range, -#endif #ifdef HAVE_VFS_CLONE_FILE_RANGE .clone_file_range = zpl_clone_file_range, #endif +#ifdef HAVE_VFS_REMAP_FILE_RANGE + .remap_file_range = zpl_remap_file_range, +#endif #ifdef HAVE_VFS_DEDUPE_FILE_RANGE .dedupe_file_range = zpl_dedupe_file_range, #endif @@ -1357,6 +1362,11 @@ const struct file_operations zpl_file_operations = { #ifdef CONFIG_COMPAT .compat_ioctl = zpl_compat_ioctl, #endif +#ifdef HAVE_VFS_FILE_OPERATIONS_EXTEND + }, /* kabi_fops */ + .copy_file_range = zpl_copy_file_range, + .clone_file_range = zpl_clone_file_range, +#endif }; const struct file_operations zpl_dir_file_operations = { diff --git a/module/os/linux/zfs/zpl_file_range.c b/module/os/linux/zfs/zpl_file_range.c index aad502a809..18efebfc1d 100644 --- a/module/os/linux/zfs/zpl_file_range.c +++ b/module/os/linux/zfs/zpl_file_range.c @@ -77,7 +77,8 @@ __zpl_clone_file_range(struct file *src_file, loff_t src_off, return ((ssize_t)len_o); } -#ifdef HAVE_VFS_COPY_FILE_RANGE +#if defined(HAVE_VFS_COPY_FILE_RANGE) || \ + defined(HAVE_VFS_FILE_OPERATIONS_EXTEND) /* * Entry point for copy_file_range(). Copy len bytes from src_off in src_file * to dst_off in dst_file. We are permitted to do this however we like, so we @@ -94,7 +95,7 @@ zpl_copy_file_range(struct file *src_file, loff_t src_off, return (-EINVAL); /* Try to do it via zfs_clone_range() */ - ret =__zpl_clone_file_range(src_file, src_off, + ret = __zpl_clone_file_range(src_file, src_off, dst_file, dst_off, len); #ifdef HAVE_VFS_GENERIC_COPY_FILE_RANGE @@ -109,7 +110,7 @@ zpl_copy_file_range(struct file *src_file, loff_t src_off, return (ret); } -#endif /* HAVE_VFS_COPY_FILE_RANGE */ +#endif /* HAVE_VFS_COPY_FILE_RANGE || HAVE_VFS_FILE_OPERATIONS_EXTEND */ #ifdef HAVE_VFS_REMAP_FILE_RANGE /* @@ -152,7 +153,8 @@ zpl_remap_file_range(struct file *src_file, loff_t src_off, } #endif /* HAVE_VFS_REMAP_FILE_RANGE */ -#ifdef HAVE_VFS_CLONE_FILE_RANGE +#if defined(HAVE_VFS_CLONE_FILE_RANGE) || \ + defined(HAVE_VFS_FILE_OPERATIONS_EXTEND) /* * Entry point for FICLONE and FICLONERANGE, before Linux 4.20. */ @@ -167,7 +169,7 @@ zpl_clone_file_range(struct file *src_file, loff_t src_off, return (__zpl_clone_file_range(src_file, src_off, dst_file, dst_off, len)); } -#endif /* HAVE_VFS_CLONE_FILE_RANGE */ +#endif /* HAVE_VFS_CLONE_FILE_RANGE || HAVE_VFS_FILE_OPERATIONS_EXTEND */ #ifdef HAVE_VFS_DEDUPE_FILE_RANGE /* From 36d1a3ef4ed923803a5b57c55cd46e071bdb4826 Mon Sep 17 00:00:00 2001 From: Rob Norris Date: Tue, 11 Jul 2023 20:46:33 +1000 Subject: [PATCH 13/13] zts: block cloning tests Reviewed-by: Brian Behlendorf Reviewed-by: Kay Pedersen Signed-off-by: Rob Norris Sponsored-By: OpenDrives Inc. Sponsored-By: Klara Inc. Closes #15050 Closes #405 Closes #13349 --- tests/runfiles/linux.run | 9 + tests/test-runner/bin/zts-report.py.in | 14 + tests/zfs-tests/cmd/.gitignore | 1 + tests/zfs-tests/cmd/Makefile.am | 1 + tests/zfs-tests/cmd/clonefile.c | 333 ++++++++++++++++++ tests/zfs-tests/include/commands.cfg | 1 + tests/zfs-tests/tests/Makefile.am | 12 + .../block_cloning/block_cloning.kshlib | 46 +++ .../block_cloning_copyfilerange.ksh | 60 ++++ ...ck_cloning_copyfilerange_cross_dataset.ksh | 65 ++++ .../block_cloning_copyfilerange_partial.ksh | 68 ++++ .../block_cloning_disabled_copyfilerange.ksh | 60 ++++ .../block_cloning_disabled_ficlone.ksh | 50 +++ .../block_cloning_disabled_ficlonerange.ksh | 50 +++ .../block_cloning/block_cloning_ficlone.ksh | 56 +++ .../block_cloning_ficlonerange.ksh | 56 +++ .../block_cloning_ficlonerange_partial.ksh | 64 ++++ .../functional/block_cloning/cleanup.ksh | 34 ++ .../tests/functional/block_cloning/setup.ksh | 36 ++ 19 files changed, 1016 insertions(+) create mode 100644 tests/zfs-tests/cmd/clonefile.c create mode 100644 tests/zfs-tests/tests/functional/block_cloning/block_cloning.kshlib create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_partial.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_copyfilerange.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlone.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlonerange.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlone.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange_partial.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh create mode 100755 tests/zfs-tests/tests/functional/block_cloning/setup.ksh diff --git a/tests/runfiles/linux.run b/tests/runfiles/linux.run index 618eeb9340..b68202d849 100644 --- a/tests/runfiles/linux.run +++ b/tests/runfiles/linux.run @@ -34,6 +34,15 @@ tags = ['functional', 'acl', 'posix-sa'] tests = ['atime_003_pos', 'root_relatime_on'] tags = ['functional', 'atime'] +[tests/functional/block_cloning:Linux] +tests = ['block_cloning_copyfilerange', 'block_cloning_copyfilerange_partial', + 'block_cloning_ficlone', 'block_cloning_ficlonerange', + 'block_cloning_ficlonerange_partial', + 'block_cloning_disabled_copyfilerange', 'block_cloning_disabled_ficlone', + 'block_cloning_disabled_ficlonerange', + 'block_cloning_copyfilerange_cross_dataset'] +tags = ['functional', 'block_cloning'] + [tests/functional/chattr:Linux] tests = ['chattr_001_pos', 'chattr_002_neg'] tags = ['functional', 'chattr'] diff --git a/tests/test-runner/bin/zts-report.py.in b/tests/test-runner/bin/zts-report.py.in index cf438e0e64..c9a2b4179a 100755 --- a/tests/test-runner/bin/zts-report.py.in +++ b/tests/test-runner/bin/zts-report.py.in @@ -134,6 +134,12 @@ ci_reason = 'CI runner doesn\'t have all requirements' # idmap_reason = 'Idmapped mount needs kernel 5.12+' +# +# copy_file_range() is not supported by all kernels +# +cfr_reason = 'Kernel copy_file_range support required' +cfr_cross_reason = 'copy_file_range(2) cross-filesystem needs kernel 5.3+' + # # These tests are known to fail, thus we use this list to prevent these # failures from failing the job as a whole; only unexpected failures @@ -288,6 +294,14 @@ elif sys.platform.startswith('linux'): 'idmap_mount/idmap_mount_003': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_004': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_005': ['SKIP', idmap_reason], + 'block_cloning/block_cloning_disabled_copyfilerange': + ['SKIP', cfr_reason], + 'block_cloning/block_cloning_copyfilerange': + ['SKIP', cfr_reason], + 'block_cloning/block_cloning_copyfilerange_partial': + ['SKIP', cfr_reason], + 'block_cloning/block_cloning_copyfilerange_cross_dataset': + ['SKIP', cfr_cross_reason], }) diff --git a/tests/zfs-tests/cmd/.gitignore b/tests/zfs-tests/cmd/.gitignore index f68f580728..5f53b68719 100644 --- a/tests/zfs-tests/cmd/.gitignore +++ b/tests/zfs-tests/cmd/.gitignore @@ -1,6 +1,7 @@ /badsend /btree_test /chg_usr_exec +/clonefile /devname2devid /dir_rd_update /draid diff --git a/tests/zfs-tests/cmd/Makefile.am b/tests/zfs-tests/cmd/Makefile.am index 066abb6ce3..9bdb3c2097 100644 --- a/tests/zfs-tests/cmd/Makefile.am +++ b/tests/zfs-tests/cmd/Makefile.am @@ -119,6 +119,7 @@ scripts_zfs_tests_bin_PROGRAMS += %D%/renameat2 scripts_zfs_tests_bin_PROGRAMS += %D%/xattrtest scripts_zfs_tests_bin_PROGRAMS += %D%/zed_fd_spill-zedlet scripts_zfs_tests_bin_PROGRAMS += %D%/idmap_util +scripts_zfs_tests_bin_PROGRAMS += %D%/clonefile %C%_idmap_util_LDADD = libspl.la diff --git a/tests/zfs-tests/cmd/clonefile.c b/tests/zfs-tests/cmd/clonefile.c new file mode 100644 index 0000000000..a7e7277ae4 --- /dev/null +++ b/tests/zfs-tests/cmd/clonefile.c @@ -0,0 +1,333 @@ +/* + * SPDX-License-Identifier: MIT + * + * Copyright (c) 2023, Rob Norris + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ + +/* + * This program is to test the availability and behaviour of copy_file_range, + * FICLONE, FICLONERANGE and FIDEDUPERANGE in the Linux kernel. It should + * compile and run even if these features aren't exposed through the libc. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifndef __NR_copy_file_range +#if defined(__x86_64__) +#define __NR_copy_file_range (326) +#elif defined(__i386__) +#define __NR_copy_file_range (377) +#elif defined(__s390__) +#define __NR_copy_file_range (375) +#elif defined(__arm__) +#define __NR_copy_file_range (391) +#elif defined(__aarch64__) +#define __NR_copy_file_range (285) +#elif defined(__powerpc__) +#define __NR_copy_file_range (379) +#else +#error "no definition of __NR_copy_file_range for this platform" +#endif +#endif /* __NR_copy_file_range */ + +ssize_t +copy_file_range(int, loff_t *, int, loff_t *, size_t, unsigned int) + __attribute__((weak)); + +static inline ssize_t +cf_copy_file_range(int sfd, loff_t *soff, int dfd, loff_t *doff, + size_t len, unsigned int flags) +{ + if (copy_file_range) + return (copy_file_range(sfd, soff, dfd, doff, len, flags)); + return ( + syscall(__NR_copy_file_range, sfd, soff, dfd, doff, len, flags)); +} + +/* Define missing FICLONE */ +#ifdef FICLONE +#define CF_FICLONE FICLONE +#else +#define CF_FICLONE _IOW(0x94, 9, int) +#endif + +/* Define missing FICLONERANGE and support structs */ +#ifdef FICLONERANGE +#define CF_FICLONERANGE FICLONERANGE +typedef struct file_clone_range cf_file_clone_range_t; +#else +typedef struct { + int64_t src_fd; + uint64_t src_offset; + uint64_t src_length; + uint64_t dest_offset; +} cf_file_clone_range_t; +#define CF_FICLONERANGE _IOW(0x94, 13, cf_file_clone_range_t) +#endif + +/* Define missing FIDEDUPERANGE and support structs */ +#ifdef FIDEDUPERANGE +#define CF_FIDEDUPERANGE FIDEDUPERANGE +#define CF_FILE_DEDUPE_RANGE_SAME FILE_DEDUPE_RANGE_SAME +#define CF_FILE_DEDUPE_RANGE_DIFFERS FILE_DEDUPE_RANGE_DIFFERS +typedef struct file_dedupe_range_info cf_file_dedupe_range_info_t; +typedef struct file_dedupe_range cf_file_dedupe_range_t; +#else +typedef struct { + int64_t dest_fd; + uint64_t dest_offset; + uint64_t bytes_deduped; + int32_t status; + uint32_t reserved; +} cf_file_dedupe_range_info_t; +typedef struct { + uint64_t src_offset; + uint64_t src_length; + uint16_t dest_count; + uint16_t reserved1; + uint32_t reserved2; + cf_file_dedupe_range_info_t info[0]; +} cf_file_dedupe_range_t; +#define CF_FIDEDUPERANGE _IOWR(0x94, 54, cf_file_dedupe_range_t) +#define CF_FILE_DEDUPE_RANGE_SAME (0) +#define CF_FILE_DEDUPE_RANGE_DIFFERS (1) +#endif + +typedef enum { + CF_MODE_NONE, + CF_MODE_CLONE, + CF_MODE_CLONERANGE, + CF_MODE_COPYFILERANGE, + CF_MODE_DEDUPERANGE, +} cf_mode_t; + +static int +usage(void) +{ + printf( + "usage:\n" + " FICLONE:\n" + " clonefile -c \n" + " FICLONERANGE:\n" + " clonefile -r \n" + " copy_file_range:\n" + " clonefile -f \n" + " FIDEDUPERANGE:\n" + " clonefile -d \n"); + return (1); +} + +int do_clone(int sfd, int dfd); +int do_clonerange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len); +int do_copyfilerange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len); +int do_deduperange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len); + +int quiet = 0; + +int +main(int argc, char **argv) +{ + cf_mode_t mode = CF_MODE_NONE; + + char c; + while ((c = getopt(argc, argv, "crfdq")) != -1) { + switch (c) { + case 'c': + mode = CF_MODE_CLONE; + break; + case 'r': + mode = CF_MODE_CLONERANGE; + break; + case 'f': + mode = CF_MODE_COPYFILERANGE; + break; + case 'd': + mode = CF_MODE_DEDUPERANGE; + break; + case 'q': + quiet = 1; + break; + } + } + + if (mode == CF_MODE_NONE || (argc-optind) < 2 || + (mode != CF_MODE_CLONE && (argc-optind) < 5)) + return (usage()); + + loff_t soff = 0, doff = 0; + size_t len = 0; + if (mode != CF_MODE_CLONE) { + soff = strtoull(argv[optind+2], NULL, 10); + if (soff == ULLONG_MAX) { + fprintf(stderr, "invalid source offset"); + return (1); + } + doff = strtoull(argv[optind+3], NULL, 10); + if (doff == ULLONG_MAX) { + fprintf(stderr, "invalid dest offset"); + return (1); + } + len = strtoull(argv[optind+4], NULL, 10); + if (len == ULLONG_MAX) { + fprintf(stderr, "invalid length"); + return (1); + } + } + + int sfd = open(argv[optind], O_RDONLY); + if (sfd < 0) { + fprintf(stderr, "open: %s: %s\n", + argv[optind], strerror(errno)); + return (1); + } + + int dfd = open(argv[optind+1], O_WRONLY|O_CREAT, + S_IRUSR|S_IWUSR|S_IRGRP|S_IROTH); + if (sfd < 0) { + fprintf(stderr, "open: %s: %s\n", + argv[optind+1], strerror(errno)); + close(sfd); + return (1); + } + + int err; + switch (mode) { + case CF_MODE_CLONE: + err = do_clone(sfd, dfd); + break; + case CF_MODE_CLONERANGE: + err = do_clonerange(sfd, dfd, soff, doff, len); + break; + case CF_MODE_COPYFILERANGE: + err = do_copyfilerange(sfd, dfd, soff, doff, len); + break; + case CF_MODE_DEDUPERANGE: + err = do_deduperange(sfd, dfd, soff, doff, len); + break; + default: + abort(); + } + + off_t spos = lseek(sfd, 0, SEEK_CUR); + off_t slen = lseek(sfd, 0, SEEK_END); + off_t dpos = lseek(dfd, 0, SEEK_CUR); + off_t dlen = lseek(dfd, 0, SEEK_END); + + fprintf(stderr, "file offsets: src=%lu/%lu; dst=%lu/%lu\n", spos, slen, + dpos, dlen); + + close(dfd); + close(sfd); + + return (err == 0 ? 0 : 1); +} + +int +do_clone(int sfd, int dfd) +{ + fprintf(stderr, "using FICLONE\n"); + int err = ioctl(dfd, CF_FICLONE, sfd); + if (err < 0) { + fprintf(stderr, "ioctl(FICLONE): %s\n", strerror(errno)); + return (err); + } + return (0); +} + +int +do_clonerange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len) +{ + fprintf(stderr, "using FICLONERANGE\n"); + cf_file_clone_range_t fcr = { + .src_fd = sfd, + .src_offset = soff, + .src_length = len, + .dest_offset = doff, + }; + int err = ioctl(dfd, CF_FICLONERANGE, &fcr); + if (err < 0) { + fprintf(stderr, "ioctl(FICLONERANGE): %s\n", strerror(errno)); + return (err); + } + return (0); +} + +int +do_copyfilerange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len) +{ + fprintf(stderr, "using copy_file_range\n"); + ssize_t copied = cf_copy_file_range(sfd, &soff, dfd, &doff, len, 0); + if (copied < 0) { + fprintf(stderr, "copy_file_range: %s\n", strerror(errno)); + return (1); + } + if (copied != len) { + fprintf(stderr, "copy_file_range: copied less than requested: " + "requested=%lu; copied=%lu\n", len, copied); + return (1); + } + return (0); +} + +int +do_deduperange(int sfd, int dfd, loff_t soff, loff_t doff, size_t len) +{ + fprintf(stderr, "using FIDEDUPERANGE\n"); + + char buf[sizeof (cf_file_dedupe_range_t)+ + sizeof (cf_file_dedupe_range_info_t)] = {0}; + cf_file_dedupe_range_t *fdr = (cf_file_dedupe_range_t *)&buf[0]; + cf_file_dedupe_range_info_t *fdri = + (cf_file_dedupe_range_info_t *) + &buf[sizeof (cf_file_dedupe_range_t)]; + + fdr->src_offset = soff; + fdr->src_length = len; + fdr->dest_count = 1; + + fdri->dest_fd = dfd; + fdri->dest_offset = doff; + + int err = ioctl(sfd, CF_FIDEDUPERANGE, fdr); + if (err != 0) + fprintf(stderr, "ioctl(FIDEDUPERANGE): %s\n", strerror(errno)); + + if (fdri->status < 0) { + fprintf(stderr, "dedup failed: %s\n", strerror(-fdri->status)); + err = -1; + } else if (fdri->status == CF_FILE_DEDUPE_RANGE_DIFFERS) { + fprintf(stderr, "dedup failed: range differs\n"); + err = -1; + } + + return (err); +} diff --git a/tests/zfs-tests/include/commands.cfg b/tests/zfs-tests/include/commands.cfg index b3cfe149ff..fa545e06bb 100644 --- a/tests/zfs-tests/include/commands.cfg +++ b/tests/zfs-tests/include/commands.cfg @@ -182,6 +182,7 @@ export ZFS_FILES='zdb export ZFSTEST_FILES='badsend btree_test chg_usr_exec + clonefile devname2devid dir_rd_update draid diff --git a/tests/zfs-tests/tests/Makefile.am b/tests/zfs-tests/tests/Makefile.am index ff65dc1ac2..0819cb6b57 100644 --- a/tests/zfs-tests/tests/Makefile.am +++ b/tests/zfs-tests/tests/Makefile.am @@ -90,6 +90,7 @@ nobase_dist_datadir_zfs_tests_tests_DATA += \ functional/alloc_class/alloc_class.kshlib \ functional/atime/atime.cfg \ functional/atime/atime_common.kshlib \ + functional/block_cloning/block_cloning.kshlib \ functional/cache/cache.cfg \ functional/cache/cache.kshlib \ functional/cachefile/cachefile.cfg \ @@ -437,6 +438,17 @@ nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \ functional/atime/root_atime_on.ksh \ functional/atime/root_relatime_on.ksh \ functional/atime/setup.ksh \ + functional/block_cloning/cleanup.ksh \ + functional/block_cloning/setup.ksh \ + functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh \ + functional/block_cloning/block_cloning_copyfilerange.ksh \ + functional/block_cloning/block_cloning_copyfilerange_partial.ksh \ + functional/block_cloning/block_cloning_disabled_copyfilerange.ksh \ + functional/block_cloning/block_cloning_disabled_ficlone.ksh \ + functional/block_cloning/block_cloning_disabled_ficlonerange.ksh \ + functional/block_cloning/block_cloning_ficlone.ksh \ + functional/block_cloning/block_cloning_ficlonerange.ksh \ + functional/block_cloning/block_cloning_ficlonerange_partial.ksh \ functional/bootfs/bootfs_001_pos.ksh \ functional/bootfs/bootfs_002_neg.ksh \ functional/bootfs/bootfs_003_pos.ksh \ diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning.kshlib b/tests/zfs-tests/tests/functional/block_cloning/block_cloning.kshlib new file mode 100644 index 0000000000..9998e5a87b --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning.kshlib @@ -0,0 +1,46 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib + +function have_same_content +{ + typeset hash1=$(cat $1 | md5sum) + typeset hash2=$(cat $2 | md5sum) + + log_must [ "$hash1" = "$hash2" ] +} + +function unique_blocks +{ + typeset zdbout=${TMPDIR:-$TEST_BASE_DIR}/zdbout.$$ + zdb -vvvvv $1 -O $2 | \ + awk '/ L0 / { print ++l " " $3 " " $7 }' > $zdbout.a + zdb -vvvvv $3 -O $4 | \ + awk '/ L0 / { print ++l " " $3 " " $7 }' > $zdbout.b + echo $(sort $zdbout.a $zdbout.b | uniq -d | cut -f1 -d' ') +} + diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange.ksh new file mode 100755 index 0000000000..9adcbfcd88 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange.ksh @@ -0,0 +1,60 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +if [[ $(linux_version) -lt $(linux_version "4.5") ]]; then + log_unsupported "copy_file_range not available before Linux 4.5" +fi + +claim="The copy_file_range syscall can clone whole files." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must clonefile -f /$TESTPOOL/file1 /$TESTPOOL/file2 0 0 524288 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "1 2 3 4" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh new file mode 100755 index 0000000000..07e089e89c --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_cross_dataset.ksh @@ -0,0 +1,65 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +if [[ $(linux_version) -lt $(linux_version "5.3") ]]; then + log_unsupported "copy_file_range can't copy cross-filesystem before Linux 5.3" +fi + +claim="The copy_file_range syscall can clone across datasets." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must zfs create $TESTPOOL/$TESTFS1 +log_must zfs create $TESTPOOL/$TESTFS2 + +log_must dd if=/dev/urandom of=/$TESTPOOL/$TESTFS1/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must \ + clonefile -f /$TESTPOOL/$TESTFS1/file1 /$TESTPOOL/$TESTFS2/file2 0 0 524288 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/$TESTFS1/file1 /$TESTPOOL/$TESTFS2/file2 + +typeset blocks=$(unique_blocks \ + $TESTPOOL/$TESTFS1 file1 $TESTPOOL/$TESTFS2 file2) +log_must [ "$blocks" = "1 2 3 4" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_partial.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_partial.ksh new file mode 100755 index 0000000000..ecac62b203 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_copyfilerange_partial.ksh @@ -0,0 +1,68 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +if [[ $(linux_version) -lt $(linux_version "4.5") ]]; then + log_unsupported "copy_file_range not available before Linux 4.5" +fi + +claim="The copy_file_range syscall can clone parts of a file." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must dd if=/$TESTPOOL/file1 of=/$TESTPOOL/file2 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "" ] + +log_must clonefile -f /$TESTPOOL/file1 /$TESTPOOL/file2 131072 131072 262144 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "2 3" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_copyfilerange.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_copyfilerange.ksh new file mode 100755 index 0000000000..30b155a140 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_copyfilerange.ksh @@ -0,0 +1,60 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +if [[ $(linux_version) -lt $(linux_version "4.5") ]]; then + log_unsupported "copy_file_range not available before Linux 4.5" +fi + +claim="The copy_file_range syscall copies files when block cloning is disabled." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=disabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must clonefile -f /$TESTPOOL/file1 /$TESTPOOL/file2 0 0 524288 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlone.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlone.ksh new file mode 100755 index 0000000000..10a2715ea2 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlone.ksh @@ -0,0 +1,50 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +claim="The FICLONE ioctl fails when block cloning is disabled." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=disabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_mustnot clonefile -c /$TESTPOOL/file1 /$TESTPOOL/file2 + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlonerange.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlonerange.ksh new file mode 100755 index 0000000000..e8461e6d3c --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_disabled_ficlonerange.ksh @@ -0,0 +1,50 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +claim="The FICLONERANGE ioctl fails when block cloning is disabled." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=disabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_mustnot clonefile -r /$TESTPOOL/file1 /$TESTPOOL/file2 0 0 524288 + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlone.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlone.ksh new file mode 100755 index 0000000000..d13a392298 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlone.ksh @@ -0,0 +1,56 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +claim="The FICLONE ioctl can clone files." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must clonefile -c /$TESTPOOL/file1 /$TESTPOOL/file2 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "1 2 3 4" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange.ksh new file mode 100755 index 0000000000..6556050c43 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange.ksh @@ -0,0 +1,56 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +claim="The FICLONERANGE ioctl can clone whole files." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must clonefile -r /$TESTPOOL/file1 /$TESTPOOL/file2 0 0 524288 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "1 2 3 4" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange_partial.ksh b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange_partial.ksh new file mode 100755 index 0000000000..37a3511a26 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/block_cloning_ficlonerange_partial.ksh @@ -0,0 +1,64 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +claim="The FICLONERANGE ioctl can clone parts of a file." + +log_assert $claim + +function cleanup +{ + datasetexists $TESTPOOL && destroy_pool $TESTPOOL +} + +log_onexit cleanup + +log_must zpool create -o feature@block_cloning=enabled $TESTPOOL $DISKS + +log_must dd if=/dev/urandom of=/$TESTPOOL/file1 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must dd if=/$TESTPOOL/file1 of=/$TESTPOOL/file2 bs=128K count=4 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "" ] + +log_must clonefile -r /$TESTPOOL/file1 /$TESTPOOL/file2 131072 131072 262144 +log_must sync_pool $TESTPOOL + +log_must have_same_content /$TESTPOOL/file1 /$TESTPOOL/file2 + +typeset blocks=$(unique_blocks $TESTPOOL file1 $TESTPOOL file2) +log_must [ "$blocks" = "2 3" ] + +log_pass $claim diff --git a/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh b/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh new file mode 100755 index 0000000000..7ac13adb63 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh @@ -0,0 +1,34 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +verify_runnable "global" + +default_cleanup_noexit + +log_pass diff --git a/tests/zfs-tests/tests/functional/block_cloning/setup.ksh b/tests/zfs-tests/tests/functional/block_cloning/setup.ksh new file mode 100755 index 0000000000..512f5a0644 --- /dev/null +++ b/tests/zfs-tests/tests/functional/block_cloning/setup.ksh @@ -0,0 +1,36 @@ +#!/bin/ksh -p +# +# 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) 2023, Klara Inc. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib + +if ! command -v clonefile > /dev/null ; then + log_unsupported "clonefile program required to test block cloning" +fi + +verify_runnable "global" + +log_pass