{"payload":{"feedbackUrl":"https://github.com/orgs/community/discussions/53140","repo":{"id":2003466,"defaultBranch":"master","name":"zfs","ownerLogin":"behlendorf","currentUserCanPush":false,"isFork":true,"isEmpty":false,"createdAt":"2011-07-05T22:51:43.000Z","ownerAvatar":"https://avatars.githubusercontent.com/u/148917?v=4","public":true,"private":false,"isOrgOwned":false},"refInfo":{"name":"","listCacheKey":"v0:1713979385.0","currentOid":""},"activityList":{"items":[{"before":"e15393c5f2bf815a502146b4daf39ff59f71fda4","after":"d18b42d0e228f97e1c7d336b9119ff177d21d4ab","ref":"refs/heads/issue-16126","pushedAt":"2024-04-24T17:48:09.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Python 3.12 deprecated python3-distutils\n\nAs for python-3.12 the distutils package has been deprecated.\nThe latest ax_python_devel.m4 macro from the autoconf archive\nhas been updated accordingly so let's pull in the new version.\n\nWe can also drop the changes made to our customized version\nto continue if the development version is not installed since\nthis functionality has been included upstream.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>\nIssue #16126","shortMessageHtmlLink":"Python 3.12 deprecated python3-distutils"}},{"before":"9bfcd487166d1f67d2adbd328115b3d34815d905","after":"e15393c5f2bf815a502146b4daf39ff59f71fda4","ref":"refs/heads/issue-16126","pushedAt":"2024-04-24T17:33:18.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Python 3.12 deprecated python3-distutils\n\nAs for python-3.12 the distutils package has been deprecated.\nThe latest ax_python_devel.m4 macro from the autoconf archive\nhas been updated accordingly so let's pull in the new version.\n\nWe can also drop the changes made to our customized version\nto continue if the development version is not installed since\nthis functionality has been included upstream.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>\nIssue #16126","shortMessageHtmlLink":"Python 3.12 deprecated python3-distutils"}},{"before":null,"after":"9bfcd487166d1f67d2adbd328115b3d34815d905","ref":"refs/heads/issue-16126","pushedAt":"2024-04-24T17:23:05.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Python 3.12 deprecated python3-distutils\n\nAs for python-3.12 the distutils package has been deprecated.\nThe latest ax_python_devel.m4 macro from the autoconf archive\nhas been updated accordingly so let's pull in the new version.\n\nWe can also drop the changes made to our customized version\nto continue if the development version is not installed since\nthis functionality has been included upstream.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>\nIssue #16126","shortMessageHtmlLink":"Python 3.12 deprecated python3-distutils"}},{"before":null,"after":"b45960d5b59a141c6ce7fb2cfacadfd9841d8660","ref":"refs/heads/pr-15839-backport","pushedAt":"2024-03-01T01:18:49.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"vdev probe to slow disk can stall mmp write checker\n\nSimplify vdev probes in the zio_vdev_io_done context to\navoid holding the spa config lock for long durations.\n\nSponsored-by: Klara, Inc.\nSponsored-by: Wasabi Technology, Inc.\n\nSigned-off-by: Don Brady <don.brady@klarasystems.com>","shortMessageHtmlLink":"vdev probe to slow disk can stall mmp write checker"}},{"before":"0964e82f27e32e0a0810832f584d71a4a743baa4","after":"c4d3d7a030076da8aebf5c0a3c98a2a426b0a92a","ref":"refs/heads/enforce-minimum-partition-size","pushedAt":"2024-02-15T20:05:44.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Check for minimum partition size\n\nOn Linux block devices used for vdevs will by partitioned.  The block\ndevice must be large enough for an 64M partition starting at offset\nof 2048 sectors (part1), and a second 64M reserved partition at the\nend of the device (part9).\n\nThis commit adds a capacity check when creating the GPT label to\nimmediately detect a device which is too small.  With the existing\ncode this would be caught slightly latter when attempting to use\nthe partition.  Catching it sooner let's us print a more useful error.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Check for minimum partition size"}},{"before":"3b99a5911ae740723655fbdc81f9829ff8e167e5","after":"0964e82f27e32e0a0810832f584d71a4a743baa4","ref":"refs/heads/enforce-minimum-partition-size","pushedAt":"2024-02-14T23:26:18.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Check for minimum partition size\n\nOn Linux block devices used for vdevs will by partitioned.  The block\ndevice must be large enough for an 64M partition starting at offset\nof 2048 sectors (part1), and a second 64M reserved partition at the\nend of the device (part9).\n\nThis commit adds a capacity check when creating the GPT label to\nimmediately detect a device which is too small.  With the existing\ncode this would be caught slightly latter when attempting to use\nthe partition.  Catching it sooner let's us print a more useful error.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Check for minimum partition size"}},{"before":null,"after":"3b99a5911ae740723655fbdc81f9829ff8e167e5","ref":"refs/heads/enforce-minimum-partition-size","pushedAt":"2024-02-14T23:25:21.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Check for minimum partition size\n\nOn Linux block devices used for vdevs will by paritioned.  The block\ndevice must be large enough for an 64M partition starting at offset\nof 2048 sectors (part1), and a second 64M reserved partition at the\nend of the device (part9).\n\nThis commit adds a capacity check when creaating the GPT label to\nimmediately detect a device which is too small.  With the existing\ncode this would be caught slightly latter when attempting to use\nthe partition.  Catching it sooner let's us print a more useful error.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Check for minimum partition size"}},{"before":"c6b83949eb70140a3965732d7cb099b540934e01","after":"64111843a676d6882c629e43dc568e680e665338","ref":"refs/heads/zpstat_unclean_2","pushedAt":"2024-02-06T21:13:11.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Updates to test case\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Updates to test case"}},{"before":"26128c6a7f69dd50701baa795ec3f25419ec072d","after":"4d5e9ea42584f2553d94e3e1b3627dc8505f7b0e","ref":"refs/heads/issue-15728","pushedAt":"2024-02-05T19:04:42.000Z","pushType":"push","commitsCount":1,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Move zfs_bclone_enabled and zfs_bclone_wait_dirty tunables\n\nRelocate declaration of zfs_bclone_enabled and zfs_bclone_wait_dirty\nto the platform independant code.  Add some additional documention\nto these tunables at the same time.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"BRT: Move zfs_bclone_enabled and zfs_bclone_wait_dirty tunables"}},{"before":"c5a762f5deaf62d030106cb417aa5c43b65c797e","after":"26128c6a7f69dd50701baa795ec3f25419ec072d","ref":"refs/heads/issue-15728","pushedAt":"2024-02-05T17:52:03.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifying that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":"4cc5f5779777ef440c73db57c7eb73f7467b2260","after":"c5a762f5deaf62d030106cb417aa5c43b65c797e","ref":"refs/heads/issue-15728","pushedAt":"2024-02-02T19:13:55.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifying that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":"46e82deae82d45cfdd55b2b539a96c90965517ea","after":"4cc5f5779777ef440c73db57c7eb73f7467b2260","ref":"refs/heads/issue-15728","pushedAt":"2024-02-02T18:36:01.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifying that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":"45d016aa63bbced900a7a231f78191591774608f","after":"46e82deae82d45cfdd55b2b539a96c90965517ea","ref":"refs/heads/issue-15728","pushedAt":"2024-02-02T00:24:21.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifying that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":"40667cdb8cdfb9293487b2c7b6d9393ab5ec1855","after":"45d016aa63bbced900a7a231f78191591774608f","ref":"refs/heads/issue-15728","pushedAt":"2024-02-01T01:06:10.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifying that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":null,"after":"40667cdb8cdfb9293487b2c7b6d9393ab5ec1855","ref":"refs/heads/issue-15728","pushedAt":"2024-02-01T00:11:59.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"BRT: Fix FICLONE/FICLONERANGE shortened copy\n\nOn Linux the ioctl_ficlonerange() and ioctl_ficlone() system calls\nare expected to either fully clone the specified range or return an\nerror.  The range may be for an entire file.  While internally ZFS\nsupports cloning partial ranges there's no way to return the length\ncloned to the caller so we need to make this all or nothing.\n\nAs part of this change support for the REMAP_FILE_CAN_SHORTEN flag\nhas been added.  When REMAP_FILE_CAN_SHORTEN is set zfs_clone_range()\nwill return a shortened range when encountering pending dirty records.\nWhen it's clear zfs_clone_range() will block and wait for the records\nto be written out allowing the blocks to be cloned.\n\nFurthermore, the file rangelock is held over the region being cloned\nto prevent it from being modified while cloning.  This doesn't quite\nprovide an atomic semantics since if an error is encountered only a\nportion of the range may be cloned.  This will be converted to an\nerror if REMAP_FILE_CAN_SHORTEN was not provided and returned to the\ncaller.  However, the destination file range is left in an undefined\nstate.\n\nA test case has been added which exercises this functionality by\nverifing that `cp --reflink=never|auto|always` works correctly.\n\nSigned-off-by: Brian D Behlendorf <behlendo@slag12.llnl.gov>\nIssue #15728","shortMessageHtmlLink":"BRT: Fix FICLONE/FICLONERANGE shortened copy"}},{"before":"89765895d88aa41719d3c510a1e3c8d8ac887862","after":"217fabc9d496dceef8914ed0a77ac70e4073a53c","ref":"refs/heads/issue-15510-backport","pushedAt":"2024-01-29T23:09:06.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Workaround UBSAN errors for variable arrays\n\nThis gets around UBSAN errors when using arrays at the end of\nstructs.  It converts some zero-length arrays to variable length\narrays and disables UBSAN checking on certain modules.\n\nIt is based off of the patch from #15460.\n\nReviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>\nTested-by: Thomas Lamprecht <t.lamprecht@proxmox.com>\nCo-authored-by: Thomas Lamprecht <t.lamprecht@proxmox.com>\nSigned-off-by: Tony Hutter <hutter2@llnl.gov>\nIssue #15145\nCloses #15510","shortMessageHtmlLink":"Workaround UBSAN errors for variable arrays"}},{"before":null,"after":"7abc719db56e321e00f7953868fd810fb6c84b4c","ref":"refs/heads/zfs-2.2.3-backports","pushedAt":"2024-01-29T21:51:49.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Linux 6.8 compat: handle mnt_idmap user_namespace change\n\nstruct mnt_idmap no longer has a struct user_namespace within it. Work\naround this by creating a temporary with the copy of the map we need\ntaken from the idmap.\n\nReviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>\nCo-authored-by: Youzhong Yang <yyang@mathworks.com>\nSigned-off-by: Rob Norris <robn@despairlabs.com>\nSponsored-by: https://despairlabs.com/sponsor/\nCloses #15805","shortMessageHtmlLink":"Linux 6.8 compat: handle mnt_idmap user_namespace change"}},{"before":null,"after":"01eaaa18c4cd25ffdc94175f970c77d1d5c90b27","ref":"refs/heads/linux-6.7-META","pushedAt":"2024-01-29T19:08:03.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Linux 6.7 compat: META\n\nUpdate the META file to reflect compatibility with the 6.7 kernel.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Linux 6.7 compat: META"}},{"before":null,"after":"89765895d88aa41719d3c510a1e3c8d8ac887862","ref":"refs/heads/issue-15510-backport","pushedAt":"2024-01-26T23:00:47.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Workaround UBSAN errors for variable arrays\n\nThis gets around UBSAN errors when using arrays at the end of\nstructs.  It converts some zero-length arrays to variable length\narrays and disables UBSAN checking on certain modules.\n\nIt is based off of the patch from #15460.\n\nReviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>\nTested-by: Thomas Lamprecht <t.lamprecht@proxmox.com>\nCo-authored-by: Thomas Lamprecht <t.lamprecht@proxmox.com>\nSigned-off-by: Tony Hutter <hutter2@llnl.gov>\nIssue #15145\nCloses #15510","shortMessageHtmlLink":"Workaround UBSAN errors for variable arrays"}},{"before":null,"after":"7bf8ef4108cceaf4851497d88dad9e4fe0738661","ref":"refs/heads/zts-zdb_args_pos","pushedAt":"2024-01-26T21:47:05.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"ZTS: Allow longer run time for zdb_args_pos\n\nThe zdb_args_pos may take slightly longer than 600 seconds to run on\nsome of the CI builders.  To prevent this from causing failures allow\nup to 1200 seconds for tests in this group.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"ZTS: Allow longer run time for zdb_args_pos"}},{"before":null,"after":"24d338ec7ae8aa4991d4b45c62ab0a3da405e30e","ref":"refs/heads/zfs-bclone-enable-backport","pushedAt":"2024-01-19T20:45:55.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"ZTS: Apply zfs_bclone_enabled to bclone tests\n\nIf block cloning is disabled by default then enable it when running\nthe bclone tests.  Follow up to #15529.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"ZTS: Apply zfs_bclone_enabled to bclone tests"}},{"before":null,"after":"41fe04167b5369d306eae37f4d2ad7b68a5f5c81","ref":"refs/heads/zts-bclone-enable","pushedAt":"2024-01-19T20:39:37.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"ZTS: Apply zfs_bclone_enabled to bclone tests\n\nIf block cloning is disabled by default then enable it when running\nthe bclone tests.  Follow up to #15529.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"ZTS: Apply zfs_bclone_enabled to bclone tests"}},{"before":"65099d2801316516e9643bb73c6fb8556d272a09","after":"c6b83949eb70140a3965732d7cb099b540934e01","ref":"refs/heads/zpstat_unclean_2","pushedAt":"2024-01-19T00:02:36.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Add 'zpool status -e' flag to see unhealthy vdevs\n\nWhen very large pools are present, it can be laborious to find\nreasons for why a pool is degraded and/or where an unhealthy vdev\nis. This option filters out vdevs that are ONLINE and with no errors\nto make it easier to see where the issues are. Root and parents of\nunhealthy vdevs will always be printed.\n\nTesting:\nZFS errors and drive failures for multiple vdevs were simulated with\nzinject.\n\nSample vdev listings with '-e' option\n- All vdevs healthy\n    NAME        STATE     READ WRITE CKSUM\n    iron5       ONLINE       0     0     0\n\n- ZFS errors\n    NAME        STATE     READ WRITE CKSUM\n    iron5       ONLINE       0     0     0\n      raidz2-5  ONLINE       1     0     0\n        L23     ONLINE       1     0     0\n        L24     ONLINE       1     0     0\n        L37     ONLINE       1     0     0\n\n- Vdev faulted\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     FAULTED      0     0     0  too many errors\n\n- Vdev faults and data errors\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-1  DEGRADED     0     0     0\n        L2      FAULTED      0     0     0  too many errors\n      raidz2-5  ONLINE       1     0     0\n        L23     ONLINE       1     0     0\n        L24     ONLINE       1     0     0\n        L37     ONLINE       1     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     FAULTED      0     0     0  too many errors\n\n- Vdev missing\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     UNAVAIL      3     1     0\n\nSigned-off-by: Cameron Harr <harr1@llnl.gov>\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Add 'zpool status -e' flag to see unhealthy vdevs"}},{"before":null,"after":"65099d2801316516e9643bb73c6fb8556d272a09","ref":"refs/heads/zpstat_unclean_2","pushedAt":"2024-01-19T00:00:03.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Add 'zpool status -e' flag to see unhealthy vdevs\n\nWhen very large pools are present, it can be laborious to find\nreasons for why a pool is degraded and/or where an unhealthy vdev\nis. This option filters out vdevs that are ONLINE and with no errors\nto make it easier to see where the issues are. Root and parents of\nunhealthy vdevs will always be printed.\n\nTesting:\nZFS errors and drive failures for multiple vdevs were simulated with\nzinject.\n\nSample vdev listings with '-e' option\n- All vdevs healthy\n    NAME        STATE     READ WRITE CKSUM\n    iron5       ONLINE       0     0     0\n\n- ZFS errors\n    NAME        STATE     READ WRITE CKSUM\n    iron5       ONLINE       0     0     0\n      raidz2-5  ONLINE       1     0     0\n        L23     ONLINE       1     0     0\n        L24     ONLINE       1     0     0\n        L37     ONLINE       1     0     0\n\n- Vdev faulted\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     FAULTED      0     0     0  too many errors\n\n- Vdev faults and data errors\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-1  DEGRADED     0     0     0\n        L2      FAULTED      0     0     0  too many errors\n      raidz2-5  ONLINE       1     0     0\n        L23     ONLINE       1     0     0\n        L24     ONLINE       1     0     0\n        L37     ONLINE       1     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     FAULTED      0     0     0  too many errors\n\n- Vdev missing\n    NAME        STATE     READ WRITE CKSUM\n    iron5       DEGRADED     0     0     0\n      raidz2-6  DEGRADED     0     0     0\n        L67     UNAVAIL      3     1     0\n\nSigned-off-by: Cameron Harr <harr1@llnl.gov>\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>","shortMessageHtmlLink":"Add 'zpool status -e' flag to see unhealthy vdevs"}},{"before":null,"after":"918533eddff9853f00b23ac94ff0786166d39827","ref":"refs/heads/issue-1567-backport","pushedAt":"2024-01-12T20:42:09.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"compact: workaround for GPL-only symbols on riscv from Linux 6.2\n\nSince Linux 6.2, the implementation of flush_dcache_page on riscv\r\nreferences GPL-only symbol `PageHuge`, breaking the build of zfs.\r\n\r\nThis patch uses existing mechanism to override flush_dcache_page,\r\nremoving the call to `PageHuge`. According to comments in kernel,\r\nit is only used to do some check against HugeTLB pages, which only\r\nexist in userspace. ZFS uses flush_dcache_page only on kernel pages,\r\nthus this patch will not introduce any behaviour change.\r\n\r\nSee also: torvalds/linux@d33deda, openzfs/zfs@589f59b\r\n\r\nReviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>\r\nSigned-off-by: Shengqi Chen <harry-chen@outlook.com>\r\nCloses #14974 \r\nCloses #15627","shortMessageHtmlLink":"compact: workaround for GPL-only symbols on riscv from Linux 6.2"}},{"before":"f90e6e0db396240a5222a70a4042c9d0c6482a68","after":"0a496185ce902ed435e611724a55d52afa4a516c","ref":"refs/heads/block_cloning-tests-freebsd","pushedAt":"2024-01-09T17:43:19.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Enable block_cloning tests on FreeBSD.\n\nSigned-off-by: Pawel Jakub Dawidek <pawel@dawidek.net>","shortMessageHtmlLink":"Enable block_cloning tests on FreeBSD."}},{"before":null,"after":"f90e6e0db396240a5222a70a4042c9d0c6482a68","ref":"refs/heads/block_cloning-tests-freebsd","pushedAt":"2024-01-09T00:57:08.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Enable block_cloning tests on FreeBSD.\n\nSigned-off-by: Pawel Jakub Dawidek <pawel@dawidek.net>","shortMessageHtmlLink":"Enable block_cloning tests on FreeBSD."}},{"before":null,"after":"41ba00cda8b7a847a8b3128b368c0ee99d3edbd6","ref":"refs/heads/issue-15711-backport","pushedAt":"2024-01-09T00:13:20.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Linux 6.2 compat: add check for kernel_neon_* availability\n\nThis patch adds check for `kernel_neon_*` symbols on arm and arm64\r\nplatforms to address the following issues:\r\n\r\n1. Linux 6.2+ on arm64 has exported them with `EXPORT_SYMBOL_GPL`, so\r\n   license compatibility must be checked before use.\r\n2. On both arm and arm64, the definitions of these symbols are guarded\r\n   by `CONFIG_KERNEL_MODE_NEON`, but their declarations are still\r\n   present. Checking in configuration phase only leads to MODPOST\r\n   errors (undefined references).\r\n\r\nReviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>\r\nSigned-off-by: Shengqi Chen <harry-chen@outlook.com>\r\nCloses #15711 \r\nCloses #14555 \r\nCloses: #15401","shortMessageHtmlLink":"Linux 6.2 compat: add check for kernel_neon_* availability"}},{"before":"23ee97e32e8cdd847ed3fb8463e800f5869787a4","after":"b2a8bae621dd67ab0ac134359806e2959c9e02f1","ref":"refs/heads/issue-13538","pushedAt":"2024-01-08T23:58:20.000Z","pushType":"force_push","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Fix \"out of memory\" error\n\nDrop the no_memory() call from zpool_in_use() when reading the\nlabel fails and instead return the error to the caller.  This\nprevents a misleading \"internal error: out of memory\" error\nwhen the label can't be read.  This will result in is_spare()\nreturning B_FALSE instead of aborting, which is already safely\nhandled.\n\nFurthermore, on Linux it's possible for EREMOTEIO to returned\nby an NVMe device if the device has been low-level formatted\nand not rescanned.  In this case we want to fallback to the\nlegacy scanning method and read any of the labels we can.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>\nIssue #13538","shortMessageHtmlLink":"Fix \"out of memory\" error"}},{"before":null,"after":"23ee97e32e8cdd847ed3fb8463e800f5869787a4","ref":"refs/heads/issue-13538","pushedAt":"2024-01-08T23:48:39.000Z","pushType":"branch_creation","commitsCount":0,"pusher":{"login":"behlendorf","name":"Brian Behlendorf","path":"/behlendorf","primaryAvatarUrl":"https://avatars.githubusercontent.com/u/148917?s=80&v=4"},"commit":{"message":"Fix \"out of memory\" error\n\nDrop the no_memory() call from zpool_in_use() when reading the\nlabel fails and instead return the error to the caller.  This\nprevents a misleading \"internal error: out of memory\" error\nwhen the label can't be read.  This will result in is_spare()\nreturning B_FALSE instead of aborting, which is already safely\nhandled.\n\nFurthermore, on Linux it's possible for EREMOTEIO to returned\nby an NVMVe device if the device has been low-level formatted\nand not rescanned.  In this case we want to fallback to the\nlegacy scanning method and read any of the labels we can.\n\nSigned-off-by: Brian Behlendorf <behlendorf1@llnl.gov>\nIssue #13538","shortMessageHtmlLink":"Fix \"out of memory\" error"}}],"hasNextPage":true,"hasPreviousPage":false,"activityType":"all","actor":null,"timePeriod":"all","sort":"DESC","perPage":30,"cursor":"djE6ks8AAAAEOWi66AA","startCursor":null,"endCursor":null}},"title":"Activity · behlendorf/zfs"}