summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Implement bookmark copyingChristian Schwarz2020-02-1119-141/+671
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This feature allows copying existing bookmarks using zfs bookmark fs#target fs#newbookmark There are some niche use cases for such functionality, e.g. when using bookmarks as markers for replication progress. Copying redaction bookmarks produces a normal bookmark that cannot be used for redacted send (we are not duplicating the redaction object). ZCP support for bookmarking (both creation and copying) will be implemented in a separate patch based on this work. Overview: - Terminology: - source = existing snapshot or bookmark - new/bmark = new bookmark - Implement bookmark copying in `dsl_bookmark.c` - create new bookmark node - copy source's `zbn_phys` to new's `zbn_phys` - zero-out redaction object id in copy - Extend existing bookmark ioctl nvlist schema to accept bookmarks as sources - => `dsl_bookmark_create_nvl_validate` is authoritative - use `dsl_dataset_is_before` check for both snapshot and bookmark sources - Adjust CLI - refactor shortname expansion logic in `zfs_do_bookmark` - Update man pages - warn about redaction bookmark handling - Add test cases - CLI - pyyzfs libzfs_core bindings Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9571
* Address Coverity warnings in #9902Matthew Macy2020-02-111-4/+5
| | | | | | | | | | Coverity reports the variable may be NULL, but due to the way the dirty records are handled this cannot be the case. Add a comment and VERIFY to make this clear and silence the warning. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9962
* Add missing dmu_buf_unlock_parent() calls to dbuf_read_impl()Brian Behlendorf2020-02-101-1/+3
| | | | | | | | | | | | As explained by the comment in dbuf_read() and above dbuf_read_impl(). Under all circumstances the parent lock specified by dblt should be dropped when existing dbuf_read_impl(). This was not being done for two exist paths. Additionally, ensure the mutex is unlocked before dropping the parent lock. Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Igor Kozhukhov <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9968
* Fix zdb -R with 'b' flagPaul Zuchowski2020-02-107-117/+376
| | | | | | | | | | | | | zdb -R :b fails due to the indirect block being compressed, and the 'b' and 'd' flag not working in tandem when specified. Fix the flag parsing code and create a zfs test for zdb -R block display. Also fix the zio flags where the dotted notation for the vdev portion of DVA (i.e. 0.0:offset:length) fails. Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Paul Zuchowski <[email protected]> Closes #9640 Closes #9729
* bash scripts: use /usr/bin/env for bash shebangsGraham Christensen2020-02-1010-9/+10
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Not all systems / distros have a `/bin/bash`, and these scripts are more difficult to run at development time. For example, my system is NixOS which doesn't have a /bin/bash. This is not a problem for NixOS building ZFS as a package: the build environment automatically replaces these shebangs with corrected paths. The problem is much more annoying at development time: either the scripts don't run, or I correct them for my local machine and deal with a perpetually dirty work tree. Before committing this patch I confirmed there are existing scripts which use `/usr/bin/env` to locate bash, so I am thinking this is a safe transformation. There are a handful of other shebangs in this repository which don't work on my system. This patch is useful on its own specifically for `commitcheck.sh`, otherwise I can't validate my commits before submission. Here are the remaining shebangs which NixOS systems won't have: 1274 #!/bin/ksh -p 91 #!/bin/ksh 89 #! /bin/ksh -p 2 #!/bin/sed -f 1 #!/usr/bin/perl -w 1 #!/usr/bin/ksh 1 #!/bin/nawk -f plus this which will create an invalid shebang in `tests/zfs-tests/tests/functional/mv_files/mv_files_common.kshlib`: echo "#!/bin/ksh" > $TEST_BASE_DIR/exitsZero.ksh I chose to leave those alone for now, and gauge the interest in this much smaller patch first. The fixes for these are easy enough by simply using `/usr/bin/env ksh`: 91 #!/bin/ksh 1 #!/usr/bin/ksh The fix for the other set is much trickier. Quoting the GNU coreutils manual: Most operating systems (e.g. GNU/Linux, BSDs) treat all text after the first space as a single argument. When using env in a script it is thus not possible to specify multiple arguments. and not all `env`'s support arguments. Mine (GNU Coreutils 8.31) does, though this feature is new since April 2018, GNU Coreutils 8.30: https://git.savannah.gnu.org/cgit/coreutils.git/commit/?id=668306ed86c8c79b0af0db8b9c882654ebb66db2 and worse, requires the -S argument: -S, --split-string=S process and split S into separate arguments; used to pass multiple arguments on shebang lines Example: $ seq 1 2 | $(nix-build '<nixpkgs>' -A coreutils)/bin/env "sort -nr" /nix/[...]-coreutils-8.31/bin/env: ‘sort -nr’: No such file or directory /nix/[...]-coreutils-8.31/bin/env: use -[v]S to pass options in shebang lines $ seq 1 2 | $(nix-build '<nixpkgs>' -A coreutils)/bin/env "-S sort -nr" 2 1 GNU Coreutils says FreeBSD's `env` does, though I wonder if FreeBSD's would be unhappy with the `-S`: https://www.gnu.org/software/coreutils/manual/html_node/env-invocation.html#env-invocation BusyBox v1.30.1 does not, and does not have a `-S`-like option: $ seq 1 2 | $(nix-build '<nixpkgs>' -A busybox)/bin/env "sort -nr" env: can't execute 'sort -nr': No such file or directory Toybox 0.8.1 also does not, and also does not have a `-S` option: $ seq 1 2 | $(nix-build '<nixpkgs>' -A toybox)/bin/env "sort -nr" env: exec sort -nr: No such file or directory --- At any rate, if this patch merges and the remaining ~1,500 are updated, the much larger patch should probably include a checkstyle-like test asserting all new shebangs use `/usr/bin/env`. I also don't mind dealing with NixOS weirdness if the project would prefer that. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Graham Christensen <[email protected]> Closes #9893
* Share some code for spa deadman tunablesRyan Moeller2020-02-103-19/+33
| | | | | | | | | | | | | We need to do the same thing to update all spas on any OS for these tunables, so let's share the code. While here let's match the types of the literals initializing the variables with the type of the variable. Reviewed-by: Allan Jude <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Olaf Faaland <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9964
* ZTS: Test zvol I/O in different volmodesRyan Moeller2020-02-101-0/+15
| | | | | | | | | | | | We found that our zvol code had some issues with volmode=dev that were not revealed by ZTS. Add some basic I/O operations to exercise more code paths in the volmode test. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9953
* ICP: Improve AES-GCM performanceAttila Fülöp2020-02-1020-31/+2654
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Currently SIMD accelerated AES-GCM performance is limited by two factors: a. The need to disable preemption and interrupts and save the FPU state before using it and to do the reverse when done. Due to the way the code is organized (see (b) below) we have to pay this price twice for each 16 byte GCM block processed. b. Most processing is done in C, operating on single GCM blocks. The use of SIMD instructions is limited to the AES encryption of the counter block (AES-NI) and the Galois multiplication (PCLMULQDQ). This leads to the FPU not being fully utilized for crypto operations. To solve (a) we do crypto processing in larger chunks while owning the FPU. An `icp_gcm_avx_chunk_size` module parameter was introduced to make this chunk size tweakable. It defaults to 32 KiB. This step alone roughly doubles performance. (b) is tackled by porting and using the highly optimized openssl AES-GCM assembler routines, which do all the processing (CTR, AES, GMULT) in a single routine. Both steps together result in up to 32x reduction of the time spend in the en/decryption routines, leading up to approximately 12x throughput increase for large (128 KiB) blocks. Lastly, this commit changes the default encryption algorithm from AES-CCM to AES-GCM when setting the `encryption=on` property. Reviewed-By: Brian Behlendorf <[email protected]> Reviewed-By: Jason King <[email protected]> Reviewed-By: Tom Caputi <[email protected]> Reviewed-By: Richard Laager <[email protected]> Signed-off-by: Attila Fülöp <[email protected]> Closes #9749
* Factor out dbuf_sync_bonusMatthew Macy2020-02-071-31/+52
| | | | | | | | | Factor the portion of dbuf_sync_leaf() responsible for handling bonus buffers out in to its own dbuf_sync_bonus() helper function. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9909
* ZTS: Add an is_dilos function for future ZTS updatesIgor K2020-02-071-0/+15
| | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Igor Kozhukhov <[email protected]> Closes #9960
* ZTS: Use wc -c instead of --bytes for portabilityRyan Moeller2020-02-071-2/+2
| | | | | | | | FreeBSD does not have the long opts for wc. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Paul Dagnelie <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9963
* Linux 5.6 compat: timestamp_truncate()Brian Behlendorf2020-02-074-18/+40
| | | | | | | | | | | | The timestamp_truncate() function was added, it replaces the existing timespec64_trunc() function. This change renames our wrapper function to be consistent with the upstream name and updates the compatibility code for older kernels accordingly. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9956 Closes #9961
* Linux 5.6 compat: struct proc_opsBrian Behlendorf2020-02-076-12/+97
| | | | | | | | | | | | | | The proc_ops structure was introduced to replace the use of of the file_operations structure when registering proc handlers. This change creates a new kstat_proc_op_t typedef for compatibility which can be used to pass around the correct structure. This change additionally adds the 'const' keyword to all of the existing proc operations structures. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9961
* Reduce number of atomic_add() calls in aggsumAlexander Motin2020-02-062-33/+34
| | | | | | | | | | | | | | | | | | | | | | | Previous code used 4 atomics to do aggsum_flush_bucket() and 2 more to re-borrow after the flush. But since asc_borrowed and asc_delta are accessed only while holding asc_lock, it makes no any sense to modify as_lower_bound and as_upper_bound in multiple steps. Instead of that the new code uses only 2 atomics in all the cases, one per as_*_bound variable. I think even that is overkill, simple atomic store and load could be used here, since all modifications are done under the as_lock, but there are no such primitives in ZFS code now. While there, make borrow code consider previous borrow value, so that on mixed request patterns reduce chance of needing to borrow again if much larger request follows tiny one that needed borrow. Also reduce as_numbuckets from uint64_t to u_int. It makes no sense to use so large division operation on every aggsum_add(). Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Paul Dagnelie <[email protected]> Signed-off-by: Alexander Motin <[email protected]> Sponsored-By: iXsystems, Inc. Closes #9930
* Replace static per-cpu with dynamic per-cpu dataRomain Dolbeau2020-02-061-4/+14
| | | | | | | This solves the issue of loading the spl module on RISC-V. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Romain Dolbeau <[email protected]> Closes #9942
* Fix static data to link with -fno-commonRomain Dolbeau2020-02-065-3/+7
| | | | | | | | -fno-common is the new default in GCC 10, replacing -fcommon in GCC <= 9, so static data must only be allocated once. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Romain Dolbeau <[email protected]> Closes #9943
* Fix unknown cc flag -fno-ipa-sraRyan Moeller2020-02-062-2/+24
| | | | | | | | Clang does not recognize -fno-ipa-sra, so add a check for it. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Prakash Surya <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9946
* Suggest using visudo to editGerardwx2020-02-051-0/+1
| | | | | | | Suggest visudo which allows editing the sudoers file in a safe fashion. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Gerardwx <[email protected]> Closes #9918
* Few microoptimizations to dbuf layerAlexander Motin2020-02-052-28/+16
| | | | | | | | | | | | | | | | | | | Move db_link into the same cache line as db_blkid and db_level. It allows significantly reduce avl_add() time in dbuf_create() on systems with large RAM and huge number of dbufs per dnode. Avoid few accesses to dbuf_caches[].size, which is highly congested under high IOPS and never stays in cache for a long time. Use local value we are receiving from zfs_refcount_add_many() any way. Remove cache_size_bytes_max bump from dbuf_evict_one(). I don't see a point to do it on dbuf eviction after we done it on insertion in dbuf_rele_and_unlock(). Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Alexander Motin <[email protected]> Sponsored-By: iXsystems, Inc. Closes #9931
* Convert dbuf dirty record record list to a list_tMatthew Macy2020-02-056-91/+103
| | | | | | | | | Additionally pull in state machine comments about upcoming async cow work. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9902
* Prepare ks_data before calling kstat_install()Alexander Motin2020-02-042-7/+5
| | | | | | | | | | | | It violated sequence described in kstat.h, and at least on FreeBSD kstat_install() uses provided names to create the sysctls. If the names are not available at the time, it ends up bad. Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Alexander Motin <[email protected]> Sponsored-By: iXsystems, Inc. Closes #9933
* Use -Werror to check if the compiler supports specific options韩朴宇2020-02-041-6/+6
| | | | | | | | | Be default, clang treats unknown warning option as warning. We need to use -Werror to make it an error. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: 12101111 <[email protected]> Closes #9927
* Restore aclmode and remove acltype on FreeBSDRyan Moeller2020-02-043-5/+78
| | | | | | | | | | | | | | This replaces the placeholder ZFS_PROP_PRIVATE with ZFS_PROP_ACLMODE, matching what is done in the NFSv4 ACLs PR (#9709). On FreeBSD we hide ZFS_PROP_ACLTYPE, while on Linux we hide ZFS_PROP_ACLMODE. The tests already assume this arrangement. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9913
* Fix const-correctness in raidz mathRyan Moeller2020-02-031-8/+8
| | | | | | | | Clang warns (errors) that "cast from 'const void *' to 'struct v *' drops const qualifier." Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9917
* async zvol minor node creation interferes with receiveMatthew Ahrens2020-02-0313-68/+132
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When we finish a zfs receive, dmu_recv_end_sync() calls zvol_create_minors(async=TRUE). This kicks off some other threads that create the minor device nodes (in /dev/zvol/poolname/...). These async threads call zvol_prefetch_minors_impl() and zvol_create_minor(), which both call dmu_objset_own(), which puts a "long hold" on the dataset. Since the zvol minor node creation is asynchronous, this can happen after the `ZFS_IOC_RECV[_NEW]` ioctl and `zfs receive` process have completed. After the first receive ioctl has completed, userland may attempt to do another receive into the same dataset (e.g. the next incremental stream). This second receive and the asynchronous minor node creation can interfere with one another in several different ways, because they both require exclusive access to the dataset: 1. When the second receive is finishing up, dmu_recv_end_check() does dsl_dataset_handoff_check(), which can fail with EBUSY if the async minor node creation already has a "long hold" on this dataset. This causes the 2nd receive to fail. 2. The async udev rule can fail if zvol_id and/or systemd-udevd try to open the device while the the second receive's async attempt at minor node creation owns the dataset (via zvol_prefetch_minors_impl). This causes the minor node (/dev/zd*) to exist, but the udev-generated /dev/zvol/... to not exist. 3. The async minor node creation can silently fail with EBUSY if the first receive's zvol_create_minor() trys to own the dataset while the second receive's zvol_prefetch_minors_impl already owns the dataset. To address these problems, this change synchronously creates the minor node. To avoid the lock ordering problems that the asynchrony was introduced to fix (see #3681), we create the minor nodes from open context, with no locks held, rather than from syncing contex as was originally done. Implementation notes: We generally do not need to traverse children or prefetch anything (e.g. when running the recv, snapshot, create, or clone subcommands of zfs). We only need recursion when importing/opening a pool and when loading encryption keys. The existing recursive, asynchronous, prefetching code is preserved for use in these cases. Channel programs may need to create zvol minor nodes, when creating a snapshot of a zvol with the snapdev property set. We figure out what snapshots are created when running the LUA program in syncing context. In this case we need to remember what snapshots were created, and then try to create their minor nodes from open context, after the LUA code has completed. There are additional zvol use cases that asynchronously own the dataset, which can cause similar problems. E.g. changing the volmode or snapdev properties. These are less problematic because they are not recursive and don't touch datasets that are not involved in the operation, there is still potential for interference with subsequent operations. In the future, these cases should be similarly converted to create the zvol minor node synchronously from open context. The async tasks of removing and renaming minors do not own the objset, so they do not have this problem. However, it may make sense to also convert these operations to happen synchronously from open context, in the future. Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Prakash Surya <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matthew Ahrens <[email protected]> External-issue: DLPX-65948 Closes #7863 Closes #9885
* Disable "-fipa-sra" when using "--enable-debuginfo"Prakash Surya2020-02-031-2/+2
| | | | | | | | | | | | | | | To better enable dynamic tracing tools (e.g. "bpftrace") this change disables the "-fipa-sra" compilation optimization. This way, function signatures are not changed by the compiler, which allows us to better attach to kprobes and kretprobes with dynamic tracing tools. Otherwise, the compiler may append ".isra" to the function name, and possibly change the function arguments as well. Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: George Melikov <[email protected]> Signed-off-by: Prakash Surya <[email protected]> Closes #9921
* ZTS: Only use ext4 on LinuxRyan Moeller2020-01-311-2/+11
| | | | | | | | | | | And while here, add a workaround for FreeBSD to ensure dirty data is synced before taking a snapshot, by remounting read-only, then remount again read-write after the snapshot. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9914
* Left-align index propsRyan Moeller2020-01-311-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | Index type props display as strings, which should be aligned to the left not to the right. Before: ``` FreeBSD-13_0-CURRENT-r356528 ➜ ~ zfs list -ro name,aclmode,mountpoint NAME ACLMODE MOUNTPOINT p0 passthrough /p0 p0/foo discard /p0/foo ``` After: ``` FreeBSD-13_0-CURRENT-r356528 ➜ ~ zfs list -ro name,aclmode,mountpoint NAME ACLMODE MOUNTPOINT p0 passthrough /p0 p0/foo discard /p0/foo ``` Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: George Melikov <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9912
* Use the correct spelling of 'jailed' in testsRyan Moeller2020-01-3114-19/+72
| | | | | | | | FreeBSD has jails, not zones. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9899
* ZTS: Fix a few defaultsRyan Moeller2020-01-3114-35/+11
| | | | | | | | | | | Linux was missing a default value for DEV_DSKDIR. Set it to /dev. Fix resulting fallout. SLICE_PREFIX seems like a good candidate for including in the defaults. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9898
* ZTS: FreeBSD tunable for DISABLE_IVSET_GUID_CHECKRyan Moeller2020-01-291-1/+1
| | | | | | Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9907
* ZTS: Simplify zero_partitions on FreeBSDRyan Moeller2020-01-291-1/+3
| | | | | | | | | We can avoid a great deal of `sleep 3` by simply destroying the whole partition table in one shot with gpart. Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9908
* ZTS: Reverse constrained path lookup orderRyan Moeller2020-01-291-2/+5
| | | | | | | | | | | | FreeBSD base system zfs utils are in /sbin. ZoF utils install to /usr/local/sbin. Ensure we link to the ZoF utils not the base utils when searching for utils to constrain paths to for the tests. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9906
* ZTS: Don't use edonr on FreeBSDRyan Moeller2020-01-285-14/+19
| | | | | | | FreeBSD doesn't support feature@edonr. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9901
* ZTS: Eliminate random and shuf, consolidate codeRyan Moeller2020-01-282-7/+1
| | | | | | | | | Both GNU and FreeBSD sort have -R to randomize input. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9900
* Add an .editorconfig; document git whitespace settingsGraham Christensen2020-01-272-3/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | Most of the projects I work on don't use tabs, and while authoring my first patch I had to wrestle with my editor to not introduce whitespace editors. The `.editorconfig` file is supported by a large number of editors out of the box, and many more with plugins. As a first-time contributor, I can't say for certain these settings are totally correct, but thus far git and my editor are satisfied enough. I considered adding `git config --local format.signOff true` but wanted to respect the warning: format.signOff A boolean value which lets you enable the -s/--signoff option of format-patch by default. Note: Adding the Signed-off-by: line to a patch should be a conscious act and means that you certify you have the rights to submit this work under the same open source license. Please see the SubmittingPatches document for further discussion. Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Graham Christensen <[email protected]> Closes #9892
* ZTS: Move more tests out of common.runRyan Moeller2020-01-278-44/+39
| | | | | | | | | | | | | | | These tests won't run on all platforms as currently implemented: * add_nested_replacing_spare (needs zed) * fault (needs zed) * mmp (needs multihost_history) * umount_unlink_drained (needs procfs) * zpool_expand (needs udev events and zed) * zpool_reopen (needs scsi_debug) * zvol_swap_003_pos (needs to modify vfstab) * zvol_swap_00[56]_pos (needs swaphigh/swaplen) Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9871
* zdb: add support for object ranges for zdb -dNed Bass2020-01-246-50/+555
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Allow a range of object identifiers to dump with -d. This may be useful when dumping a large dataset and you want to break it up into multiple phases, or to resume where a previous scan left off. Object type selection flags are supported to reduce the performance overhead of verbosely dumping unwanted objects, and to reduce the amount of post-processing work needed to filter out unwanted objects from zdb output. This change extends existing syntax in a backward-compatible way. That is, the base case of a range is to specify a single object identifier to dump. Ranges and object identifiers can be intermixed as command line parameters. Usage synopsis: Object ranges take the form <start>:<end>[:<flags>] start Starting object number end Ending object number, or -1 for no upper bound flags Optional flags to select object types: A All objects (this is the default) d ZFS directories f ZFS files m SPA space maps z ZAPs - Negate effect of next flag Examples: # Dump all file objects zdb -dd tank/fish 0:-1:f # Dump all file and directory objects zdb -dd tank/fish 0:-1:fd # Dump all types except file and directory objects zdb -dd tank/fish 0:-1:A-f-d # Dump object IDs in a specific range zdb -dd tank/fish 1000:2000 Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Paul Zuchowski <[email protected]> Signed-off-by: Ned Bass <[email protected]> Closes #9832
* Performance tests, some variables missing PERF_ prefixTony Nguyen2020-01-2311-74/+75
| | | | | | | | | Adding the expected PERF_ prefix to RANDSEED, COMPPERCENT, and COMPCHUNK. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Tony Nguyen <[email protected]> Closes #9877
* ZTS: zpool offline requires whole disk nameJohn Wren Kennedy2020-01-231-1/+3
| | | | | | | | | | When used with non-loop devices, zdb_004_pos fails because the disk argument provided is the partition rather than the expected whole disk name. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: John Kennedy <[email protected]> Closes #9876
* dsl_bookmark_create_check: fix NULL pointer deref if dbca_errors == NULLChristian Schwarz2020-01-231-2/+6
| | | | | | | | Discovered in preparation of zcp support for creating bookmarks. Handle the case where dbca_errors is NULL. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9880
* entity_namecheck: doc comment: include space as allowed characterChristian Schwarz2020-01-231-1/+1
| | | | | | | | The helper function valid_char already allows it but the doc comment was out of date. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9879
* scripts/zfs-test.sh: example for -tChristian Schwarz2020-01-231-0/+3
| | | | | | | | Add an example for running a single test case. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9878
* ZTS: Get xattr tests running on FreeBSDRyan Moeller2020-01-2314-140/+166
| | | | | | | | | | This mostly involves reworking platform checks to make illumos the exception (thanks to their unusual way of exposing xattrs). Other platforms are able to take advantage of the recently added xattr wrappers in libtest. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9872
* Add AltiVec RAID-ZRomain Dolbeau2020-01-2312-1/+5200
| | | | | | | | | | | | | Implements the RAID-Z function using AltiVec SIMD. This is basically the NEON code translated to AltiVec. Note that the 'fletcher' algorithm requires 64-bits operations, and the initial implementations of AltiVec (PPC74xx a.k.a. G4, PPC970 a.k.a. G5) only has up to 32-bits operations, so no 'fletcher'. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Romain Dolbeau <[email protected]> Closes #9539
* ZTS: Add missing export for DEV_DSKDIR defaultRyan Moeller2020-01-231-2/+3
| | | | | | Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9873
* dmu_send: redacted: fix memory leak on invalid redaction/from bookmarkChristian Schwarz2020-01-232-6/+10
| | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9867
* cmd/zfs: redact: better error message for common usage errorsChristian Schwarz2020-01-232-2/+19
| | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9867
* cmd/zfs: send: meaningful error message for incorrect redaction bookmarkChristian Schwarz2020-01-231-0/+6
| | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Christian Schwarz <[email protected]> Closes #9867
* Simplify FreeBSD's locking requirements in zfs_replay.cMatthew Macy2020-01-221-24/+12
| | | | | | | | | | | Now that the FreeBSD zfs_vnops code avoids asserting that a vnode lock is held when z_replay is true we can limit the FreeBSD specific changes to the couple of changes where it is necessary to drop the vnode locks because a function returns with it held. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9865
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-19 03:26:58 +0000