)]}' { "log": [ { "commit": "9b66ea472260dac53bff6eb40d36ac27f922c260", "tree": "c58347e32247de52a5b3ab09d13fb4da11c98911", "parents": [ "a88448dd5a83590e9647c131337a9df909dcce1f" ], "author": { "name": "Jeff Layton", "email": "jlayton@kernel.org", "time": "Wed Feb 19 10:50:14 2025 -0500" }, "committer": { "name": "Jeff Layton", "email": "jlayton@kernel.org", "time": "Wed Feb 19 17:43:46 2025 -0500" }, "message": "generic/126: run it inside its own subdirectory\n\nI had been seeing some failures in generic/126 when running on NFS under\nkdevops. The problem turned out to be that kdevops makes the root of the\nNFS export mode 01777 by default (i.e. with the sticky bit set). This\ncauses some of these permission tests to fail because the open() is\ncalled with the O_CREAT bit set.\n\nMake the outcome of this test not depend on the permissions on\n$TEST_DIR. Create a new directory to do these tests in, and clean it up\nwhen done.\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Jeff Layton \u003cjlayton@kernel.org\u003e\n" }, { "commit": "a88448dd5a83590e9647c131337a9df909dcce1f", "tree": "db63bd486ae1f8c4c6a2d9b07d144aa5380d193d", "parents": [ "8467552f09e1672a02712653b532a84bd46ea10e" ], "author": { "name": "Jeff Layton", "email": "jlayton@kernel.org", "time": "Wed Feb 19 11:44:07 2025 -0500" }, "committer": { "name": "Jeff Layton", "email": "jlayton@kernel.org", "time": "Wed Feb 19 11:44:07 2025 -0500" }, "message": "generic/126: run it inside its own directory\n\nTo: fstests@vger.kernel.org\nCc: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nCc: Dave Chinner \u003cdavid@fromorbit.com\u003e\nSigned-off-by: Jeff Layton \u003cjlayton@kernel.org\u003e\n\n---\nChanges in v3:\n- Add comment explaining the weird setup\n- Move all of the cleanup into _cleanup()\n- rm -rf $testdir before doing mkdir\n- Link to v2: https://lore.kernel.org/r/20250219-generic126-v2-1-e7537f6c9607@kernel.org\n\nChanges in v2:\n- add commands to _cleanup() to clean up test directory (Thanks Darrick!)\n- Link to v1: https://lore.kernel.org/r/20250219-generic126-v1-1-018e4e60c811@kernel.org\n\n--- b4-submit-tracking ---\n# This section is used internally by b4 prep for tracking purposes.\n{\n \"series\": {\n \"revision\": 3,\n \"change-id\": \"20250219-generic126-f81b3022c4b7\",\n \"prefixes\": [\n \"fstests\"\n ],\n \"history\": {\n \"v1\": [\n \"20250219-generic126-v1-1-018e4e60c811@kernel.org\"\n ],\n \"v2\": [\n \"20250219-generic126-v2-1-e7537f6c9607@kernel.org\"\n ]\n }\n }\n}" }, { "commit": "8467552f09e1672a02712653b532a84bd46ea10e", "tree": "db63bd486ae1f8c4c6a2d9b07d144aa5380d193d", "parents": [ "4906a5ab94ee4a75c9bb013cfb8a832b887a4fd0" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Fri Nov 15 19:49:26 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 29 11:20:18 2024 +0800" }, "message": "btrfs/327: add a test case to verify inline extent data read\n\n[BUG]\nWhen developing sector size \u003c page size handling for btrfs, I\u0027m hitting\na data corruption, which is only possible with the following out-of-tree\npatches:\n\n btrfs: allow inline data extents creation if sector size \u003c page size\n btrfs: allow buffered write to skip full page if it\u0027s sector aligned\n\n[CAUSE]\nThankfully no upstream kernels are affected, even if some one is\nmounting a btrfs created by x86_64 with inlined data extents, they won\u0027t\nhit the corruption.\n\nThe root cause is that when reading inline extents, we zero out the\nwhole remaining range until folio end.\n\nThis means such zeroing out can cover ranges that is dirtied but not yet\nwritten back, thus lead to data corruption.\n\nThis needs all the following conditions to be met:\n\n- Sector size \u003c page size\n So no x86_64 is affected. The most common users should be Asahi Linux.\n But they are safe due to the next two conditions.\n\n- Inline data extents are present\n For sector size \u003c page size cases, we do not allow creating new inline\n data extents but only reading it.\n\n But even all above cases are met by using a x86_64 created btrfs with\n inlined data extents, the next point will still save us.\n\n- Partial uptodate folios are allowed\n This requires the out-of-tree patch \"btrfs: allow buffered write to skip\n full page if it\u0027s sector aligned\", or buffered write will read out the\n whole folio before dirting any range.\n\nSo end users are completely safe.\n\n[TEST CASE]\nThe test case itself is pretty straightforward:\n\n- Buffered write [0, 4k)\n- Drop all page cache\n- Buffered write [8k, 12k)\n- Verify the file content\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "4906a5ab94ee4a75c9bb013cfb8a832b887a4fd0", "tree": "f9c6cedd2a9844166262a24644393d511b3efb41", "parents": [ "0c01033c834fe8c13d635757b88793571e1a415e" ], "author": { "name": "Anand Jain", "email": "anand.jain@oracle.com", "time": "Wed Nov 20 19:40:41 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:42:34 2024 +0800" }, "message": "fstests: fix blksize_t printf format warnings across architectures\n\nFix format string warnings when printing blksize_t values that vary\nacross architectures. The warning occurs because blksize_t is defined\ndifferently between architectures: aarch64 architectures blksize_t is\nint, on x86-64 it\u0027s long-int. Cast the values to long. Fixes warnings\nas below.\n\n seek_sanity_test.c:110:45: warning: format \u0027%ld\u0027 expects argument of type\n \u0027long int\u0027, but argument 3 has type \u0027blksize_t\u0027 {aka \u0027int\u0027}\n\n attr_replace_test.c:70:22: warning: format \u0027%ld\u0027 expects argument of type\n \u0027long int\u0027, but argument 3 has type \u0027__blksize_t\u0027 {aka \u0027int\u0027}\n\nSigned-off-by: Anand Jain \u003canand.jain@oracle.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "0c01033c834fe8c13d635757b88793571e1a415e", "tree": "df734db07f172bd2d2cfd55f43bdbb3e01d24a95", "parents": [ "ab1ece9ee32dd45816374e41ffe06bdef0534728" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Nov 26 12:27:29 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/459: prevent collisions between test VMs backed by a shared disk pool\n\nIf you happen to be running fstests on a bunch of VMs and the VMs all\nhave access to a shared disk pool, then it\u0027s possible that two VMs could\nbe running generic/459 at exactly the same time. In that case, it\u0027s a\nVERY bad thing to have two nodes trying to create an LVM volume group\nnamed \"vg_459\" because one node will succeed, after which the other node\nwill see the vg_459 volume group that it didn\u0027t create:\n\n A volume group called vg_459 already exists.\n Logical volume pool_459 already exists in Volume group vg_459.\n Logical Volume \"lv_459\" already exists in volume group \"vg_459\"\n\nBut then, because this is bash, we don\u0027t abort the test script and\ncontinue executing. If we\u0027re lucky this fails when /dev/vg_459/lv_459\ndisappears before mkfs can run:\n\n Error accessing specified device /dev/mapper/vg_459-lv_459: No such file or directory\n Usage: mkfs.xfs\n\nBut in the bad case both nodes write filesystems to the same device and\nthen they trample all over each other. Fix this by adding the hostname\nand pid to all the LVM names so that they won\u0027t collide.\n\nFixes: 461dad511f6b91 (\"generic: Test filesystem lockup on full overprovisioned dm-thin\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "ab1ece9ee32dd45816374e41ffe06bdef0534728", "tree": "a35e3a4cba04b28e6a1d90ed2e82bc218b658c54", "parents": [ "d78afe6b6f0c9afd48a6323e4126a5c2aab491e5" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:24:43 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/122: add tests for commitrange structures\n\nUpdate this test to check the ioctl structure for XFS_IOC_COMMIT_RANGE,\nwhich was added in 6.12. This will be the last ever addition to\nxfs/122, because in 6.13 we moved the ondisk structure checks to libxfs\nafter which we\u0027ll be able to _notrun this test on newer codebases.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d78afe6b6f0c9afd48a6323e4126a5c2aab491e5", "tree": "00a5bd70c50dfa6e752b6bfa69e389d08d4b5693", "parents": [ "f4ac46a85121f09656e2863fb437057d08408e6a" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:24:27 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/454: actually set attr value for llamapirate subtest\n\nTed reported that this test fails on his setup, and I noticed that I\nforgot to actually set a value for the xattr. In theory filesystems\nsupport zero-byte xattrs, but we might as well set and check the values\nso that we can make sure nobody got confused.\n\nThe actual test failure comes from attr 2.4.47 refusing to set a\nzero-legnth xattr, whereas 2.5 and newer will. That was changed in the\nattr commit 0550d2bc989d39 (\"Properly set and report empty attribute\nvalues\") prior to 2.4.48:\n\nhttps://git.savannah.nongnu.org/cgit/attr.git/commit/?id\u003d0550d2bc989d390eb25f7004ee0fae2dbc693a0d\n\nCc: fstests@vger.kernel.org # v2024.10.28\nFixes: 9c3762ceafd430 (\"misc: amend unicode confusing name tests to check for hidden tag characters\")\nReported-and-tested-by: tytso@mit.edu\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "f4ac46a85121f09656e2863fb437057d08408e6a", "tree": "ac4b6f345322e1dacc1c052c4082262053e3be69", "parents": [ "d6596effc4dfa276e7a3389a07b6dbcb65f9f87a" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:24:11 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/366: fix directio requirements checking\n\nOn a system with 4k-sector storage devices, this test fails with:\n\n --- /tmp/fstests/tests/generic/366.out\t2024-11-17 09:04:53.161104479 -0800\n +++ /var/tmp/fstests/generic/366.out.bad\t2024-11-20 21:02:30.948000000 -0800\n @@ -1,2 +1,34 @@\n QA output created by 366\n +fio: io_u error on file /opt/file1: Invalid argument: read offset\u003d15360, buflen\u003d512\n +fio: io_u error on file /opt/file1: Invalid argument: read offset\u003d15360, buflen\u003d512\n\nThe cause of this failure is that we cannot do 512byte directios to a\ndevice with 4k LBAs. Update the precondition checking to exclude this\nscenario.\n\nCc: fstests@vger.kernel.org # v2024.11.17\nFixes: 4c1629ae3a3a56 (\"generic: new test case to verify if certain fio load will hang the filesystem\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d6596effc4dfa276e7a3389a07b6dbcb65f9f87a", "tree": "ed47e3e1308d03fa8cfd38f6f5f468a6bab887e0", "parents": [ "98a76c6c51da4b3048e491977a76fdfd27f01b6a" ], "author": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Mon Nov 25 17:23:56 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/157: do not drop necessary mkfs options\n\nTo give the test option \"-L oldlabel\" to _scratch_mkfs_sized, xfs/157\ndoes:\n\n MKFS_OPTIONS\u003d\"-L oldlabel $MKFS_OPTIONS\" _scratch_mkfs_sized $fs_size\n\nbut the _scratch_mkfs_sized trys to keep the $fs_size, when mkfs\nfails with incompatible $MKFS_OPTIONS options, likes this:\n\n ** mkfs failed with extra mkfs options added to \"-L oldlabel -m rmapbt\u003d1\" by test 157 **\n ** attempting to mkfs using only test 157 options: -d size\u003d524288000 -b size\u003d4096 **\n\nbut the \"-L oldlabel\" is necessary, we shouldn\u0027t drop it. To avoid\nthat, we give the \"-L oldlabel\" to _scratch_mkfs_sized through\nfunction parameters, not through global MKFS_OPTIONS.\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\n[djwong: fix more string quoting issues]\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "98a76c6c51da4b3048e491977a76fdfd27f01b6a", "tree": "3938e923e245ba27439986f9f62a95ff4120abb2", "parents": [ "f06e93a4b1afbf960c1bae00a8123062b169ab15" ], "author": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Mon Nov 25 17:23:40 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "common/rc: _scratch_mkfs_sized supports extra arguments\n\nTo give more arguments to _scratch_mkfs_sized, we generally do as:\n\n MKFS_OPTIONS\u003d\"-L oldlabel $MKFS_OPTIONS\" _scratch_mkfs_sized $fs_size\n\nto give \"-L oldlabel\" to it. But if _scratch_mkfs_sized fails, it\nwill get rid of the whole MKFS_OPTIONS and try to mkfs again.\nLikes:\n\n ** mkfs failed with extra mkfs options added to \"-L oldlabel -m rmapbt\u003d1\" by test 157 **\n ** attempting to mkfs using only test 157 options: -d size\u003d524288000 -b size\u003d4096 **\n\nBut that\u0027s not the fault of \"-L oldlabel\". So for keeping the mkfs\noptions (\"-L oldlabel\") we need, we\u0027d better to let the\nscratch_mkfs_sized to support extra arguments, rather than using\nglobal MKFS_OPTIONS.\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\n[djwong: fix string quoting issues]\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "f06e93a4b1afbf960c1bae00a8123062b169ab15", "tree": "f62e4e6603c50b0a0c26fb3b9300ff7a28cfff33", "parents": [ "dbdbe1ec8591d090127952ac91be40d48f4af9f7" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:23:24 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/251: don\u0027t copy the fsstress source code\n\nRun fsstress for a short time to generate test data to replicate on the\nscratch device so that we don\u0027t blow out the test runtimes on\nunintentionally copying .git directories or large corefiles from the\ndeveloper\u0027s systems, etc.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "dbdbe1ec8591d090127952ac91be40d48f4af9f7", "tree": "15d870a9d368f54a5629831818e3da5a93f4e194", "parents": [ "2d6e7681acff1ec6bfcb76525b648c99676f23f2" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:23:09 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/251: constrain runtime via time/load/soak factors\n\nOn my test fleet, this test can run for well in excess of 20 minutes:\n\n 613 generic/251\n 616 generic/251\n 624 generic/251\n 630 generic/251\n 634 generic/251\n 652 generic/251\n 675 generic/251\n 749 generic/251\n 777 generic/251\n 808 generic/251\n 832 generic/251\n 946 generic/251\n 1082 generic/251\n 1221 generic/251\n 1241 generic/251\n 1254 generic/251\n 1305 generic/251\n 1366 generic/251\n 1646 generic/251\n 1936 generic/251\n 1952 generic/251\n 2358 generic/251\n 4359 generic/251\n 5325 generic/251\n 34046 generic/251\n\nbecause it hardcodes 20 threads and 10 copies. It\u0027s not great to have a\ntest that results in a significant fraction of the total test runtime.\nFix the looping and load on this test to use LOAD and TIME_FACTOR to\nscale up its operations, along with the usual SOAK_DURATION override.\nThat brings the default runtime down to less than a minute.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "2d6e7681acff1ec6bfcb76525b648c99676f23f2", "tree": "b17481cb5a0b879b0a9147d9ef6d3e7f93274ead", "parents": [ "d9352ed5e24a51fead3c5760cddce28262bc9e07" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:22:52 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/251: use sentinel files to kill the fstrim loop\n\nApparently the subshell kill doesn\u0027t always take, and then the test runs\nfor hours and hours because nothing stops it. Instead, use a sentinel\nfile to detect when fstrim_loop should stop execing background fstrims.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d9352ed5e24a51fead3c5760cddce28262bc9e07", "tree": "9704b717b28ef4f93377df0454557d401043a853", "parents": [ "38257f559615e56c0f788ee01ce4e4b99d0c29a5" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:22:37 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/009: allow logically contiguous preallocations\n\nThe new rtgroups feature implements a simplistic rotor to pick the\nrtgroup for an initial allocation to a file. This causes test failures\nif the preallocations are spread across two rtgroups, which happens if\nthere are more subtests than rtgroups.\n\nOne way to fix this would be to reset the rotor then each subtest starts\nallocating from rtgroup 0, but the only way to do that is to cycle the\nscratch mount, which is a bit gross.\n\nInstead, report logically contiguous mappings as a single mapping even\nif the physical space is not contiguous. Unfortunately, there\u0027s not\nenough context in the comments to know if the test actually was checking\nfor physical contiguity? Or if this is just an exerciser of the old\npreallocation calls, and it\u0027s fine as long as the file ranges are mapped\n(or unmapped) as desired.\n\nMessing with some awk is a lot cheaper than umount/mount cycling.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "38257f559615e56c0f788ee01ce4e4b99d0c29a5", "tree": "5addd30953b8a7ed96a3537052f978bf74f1de1d", "parents": [ "93a8e88d0f6ed506dfed281d6914c59a5b85b115" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:22:21 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/163: skip test if we can\u0027t shrink due to enospc issues\n\nIf this test fails due to insufficient space, skip this test. This can\nhappen if a realtime volume is enabled on the filesystem and we cannot\nshrink due to the rtbitmap.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "93a8e88d0f6ed506dfed281d6914c59a5b85b115", "tree": "13a95c82be2eef83b160f37f63a7e679ccdf29bc", "parents": [ "67e01fc52bbc2101f8ce2dcc7298716ffbc4e990" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:22:05 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/562: handle ENOSPC while cloning gracefully\n\nThis test creates a couple of patterned files on a tiny filesystem,\nfragments the free space, clones one patterned file to the other, and\nchecks that the entire file was cloned.\n\nHowever, this test doesn\u0027t work on a 64k fsblock filesystem because\nwe\u0027ve used up all the free space reservation for the rmapbt, and that\ncauses the FICLONE to error out with ENOSPC partway through. Hence we\nneed to detect the ENOSPC and _notrun the test.\n\nThat said, it turns out that XFS has been silently dropping error codes\nif we managed to make some progress cloning extents. That\u0027s ok if the\noperation has REMAP_FILE_CAN_SHORTEN like copy_file_range does, but\nFICLONE/FICLONERANGE do not permit partial results, so the dropped error\ncodes is actually an error.\n\nTherefore, this testcase now becomes a regression test for the patch to\nfix that.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "67e01fc52bbc2101f8ce2dcc7298716ffbc4e990", "tree": "254dca0cf604ec2df7b8da1006e49f9cb96ae0d0", "parents": [ "27ce2a47a7735b56c09e1250f7ded12ccf8fcbef" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:21:50 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "common/rc: capture dmesg when oom kills happen\n\nCapture the dmesg output if the OOM killer is invoked during fstests.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "27ce2a47a7735b56c09e1250f7ded12ccf8fcbef", "tree": "70722635e8f49d6940473784e82967ce29b748f9", "parents": [ "fa95cde52dcd7dcc30b9968da68741f8b16a6e43" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:21:34 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/508: fix test for 64k blocksize\n\nIt turns out that icreate transactions will try to reserve quite a bit\nof space on a 64k fsblock filesystem -- enough to handle the worst case\nparent directory expansion, a new inode chunk, and these days a parent\npointer as well. This can work out to quite a bit of space:\n\nfsblock\t\treservation\n1k\t\t172K\n4k\t\t368K\n16k\t\t1136K\n64k\t\t3650K\n\nUnfortunately, this test sets its block quota limits at 1-2MB, so we\ncan\u0027t even create a child file. Bump the limits up by 10x so that this\ntest will pass even if there\u0027s more metadata size creep in the future.\n\nFixes: f769a923f576df (\"xfs: project quota ineritance flag test\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "fa95cde52dcd7dcc30b9968da68741f8b16a6e43", "tree": "3624c8c257b12120687623c83ae90235f80b4727", "parents": [ "e21e278065bf3e8ab168c5d901a07dfa4f2705fc" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:21:18 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "xfs/113: fix failure to corrupt the entire directory\n\nThis test tries to corrupt the data blocks of a directory, but it\ndoesn\u0027t take into account the fact that __populate_check_xfs_dir can\nremove enough entries to cause sparse holes in the directory. If that\nhappens, this \"file data block is unmapped\" logic will cause the\ncorruption loop to exit early. Then we can add to the directory, which\ncauses the test to fail.\n\nInstead, create a list of mappable dir block offsets, and run 100\ncorruptions at a time to reduce the amount of time we spend initializing\nxfs_db. This fixes the regressions that I see with 32k/64k block sizes.\n\nCc: fstests@vger.kernel.org # v2022.05.01\nFixes: c8e6dbc8812653 (\"xfs: test directory metadata corruption checking and repair\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "e21e278065bf3e8ab168c5d901a07dfa4f2705fc", "tree": "174d73aca21f81fb68017c269a942a3e5c75ebc4", "parents": [ "92aad87ec067eee1fc6800f86cd02f1c57841a9a" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:21:03 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:49 2024 +0800" }, "message": "generic/757: convert to thinp\n\nConvert this test to use dm-thinp so that discards always zero the data.\nThis prevents weird replay problems if the scratch device doesn\u0027t\nguarantee that read after discard returns zeroes.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "92aad87ec067eee1fc6800f86cd02f1c57841a9a", "tree": "a7ae4c0902b7707b75a416595aadcc69577f3edd", "parents": [ "09d86a113f0a46639f3cffe98e4e9c22c085f08f" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Mon Nov 25 17:20:47 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:39:48 2024 +0800" }, "message": "generic/757: fix various bugs in this test\n\nFix this test so the check doesn\u0027t fail on XFS, and restrict runtime to\n100 loops because otherwise this test takes many hours.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "09d86a113f0a46639f3cffe98e4e9c22c085f08f", "tree": "71fef229ebec59f3f26e843d678817c81c8f121d", "parents": [ "a9c233162ef3055691ccbef685c6a245f2b9681c" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Thu Nov 28 12:14:56 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 21:32:56 2024 +0800" }, "message": "btrfs/028: kill lingering processes when test is interrupted\n\nIf we interrupt the test after it spawned the fsstress and balance\nprocesses (while it\u0027s sleeping for 30 seconds * $TIME_FACTOR), we don\u0027t\nkill them and they stay around for a long time, making it impossible to\nunmount the scratch filesystem (failing with -EBUSY).\n\nFix this by adding a _cleanup function that kills the processes and\nwaits for them to exit.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: David Sterba \u003cdsterba@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "a9c233162ef3055691ccbef685c6a245f2b9681c", "tree": "e4dec04193328274b5a6a310899eab29ba75b38d", "parents": [ "5c42ec5b9fd3491b540c425318a8e03a11bd398d" ], "author": { "name": "Nirjhar Roy", "email": "nirjhar@linux.ibm.com", "time": "Wed Nov 27 09:58:01 2024 +0530" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 10:25:36 2024 +0800" }, "message": "generic: Addition of new tests for extsize hints\n\nThis commit adds new tests that checks the behaviour of xfs/ext4\nfilesystems when extsize hint is set on file with inode size as 0,\nnon-empty files with allocated and delalloc extents and so on.\nAlthough currently this test is placed under tests/generic, it\nonly runs on xfs and there is an ongoing patch series[1] to\nenable extsize hints for ext4 as well.\n\n[1] https://lore.kernel.org/linux-ext4/cover.1726034272.git.ojaswin@linux.ibm.com/\n\nReviewed-by Ritesh Harjani (IBM) \u003critesh.list@gmail.com\u003e\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Ojaswin Mujoo \u003cojaswin@linux.ibm.com\u003e\nSuggested-by: Ojaswin Mujoo \u003cojaswin@linux.ibm.com\u003e\nSigned-off-by: Nirjhar Roy \u003cnirjhar@linux.ibm.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "5c42ec5b9fd3491b540c425318a8e03a11bd398d", "tree": "e081999deabef2ce0a0cc00838d4653fbd1c3c62", "parents": [ "e715441790b8cd2223f9759d423f54a273752466" ], "author": { "name": "Nirjhar Roy", "email": "nirjhar@linux.ibm.com", "time": "Wed Nov 27 09:58:00 2024 +0530" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 10:25:36 2024 +0800" }, "message": "common/rc: Add a new _require_scratch_extsize helper function\n\n_require_scratch_extsize helper function will be used in the\nthe next patch to make the test run only on filesystems with\nextsize support.\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Ojaswin Mujoo \u003cojaswin@linux.ibm.com\u003e\nSigned-off-by: Nirjhar Roy \u003cnirjhar@linux.ibm.com\u003e\nReviewed-by: Ritesh Harjani (IBM) \u003critesh.list@gmail.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "e715441790b8cd2223f9759d423f54a273752466", "tree": "2cd2ede106d2d5fcdaa177735e1b8f4b60bc2602", "parents": [ "c8114d98186a7c03d3c2b5e895d06b4da010cc02" ], "author": { "name": "Nirjhar Roy", "email": "nirjhar@linux.ibm.com", "time": "Wed Nov 27 09:57:59 2024 +0530" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 10:25:36 2024 +0800" }, "message": "common/rc,xfs/207: Add a common helper function to check xflag bits\n\nThis patch defines a common helper function to test whether any of\nfsxattr xflags field is set or not. We will use this helper in\nan upcoming patch for checking extsize (e) flag.\n\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Ritesh Harjani (IBM) \u003critesh.list@gmail.com\u003e\nReviewed-by: Ojaswin Mujoo \u003cojaswin@linux.ibm.com\u003e\nSigned-off-by: Nirjhar Roy \u003cnirjhar@linux.ibm.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c8114d98186a7c03d3c2b5e895d06b4da010cc02", "tree": "a2f9423d08b9bfd8f5c0af87f8ee46e52d1be9c8", "parents": [ "88be6071da67f94af9721e0f5f0afa159b2605f8" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Tue Nov 19 15:55:07 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Nov 28 10:19:26 2024 +0800" }, "message": "xfs/229: call on the test directory\n\nxfs/229 operates on a directory that is forced to the data volume, but\nit calls _require_fs_space on $TEST_DIR which might point to the RT\ndevice when -d rtinherit is set.\n\nCall _require_fs_space on $TDIR after it is created to check for the\nspace actually used by the test.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: \"Hans Holmberg\" \u003chans.holmberg@wdc.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "88be6071da67f94af9721e0f5f0afa159b2605f8", "tree": "d5016abf2cbdf70f1ee197a942f43a10671ac13e", "parents": [ "47aa187fb226189a95d26c55c6fc7328c7ebbec2" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Nov 12 17:37:14 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 22:07:41 2024 +0800" }, "message": "xfs/185: don\u0027t fail when rtfile is larger than rblocks\n\nThis test creates a 200MB rt volume on a file-backed loopdev. However,\nif the size of the loop file is not congruent with the rt extent size\n(e.g. 28k) then the rt volume will not use all 200MB because we cannot\nhave partial rt extents. Because of this rounding, we can end up with\nan fsmap listing that covers fewer sectors than the bmap of the loop\nfile.\n\nFix the test to allow this case.\n\nCc: fstests@vger.kernel.org # v2022.05.01\nFixes: 410a2e3186a1e8 (\"xfs: regresion test for fsmap problems with realtime\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "47aa187fb226189a95d26c55c6fc7328c7ebbec2", "tree": "ef4d9f880faf2540f5c5c120f5eb4b373f4a6d7c", "parents": [ "e2affc453cf5a5c172565a6e12be6a45d34b0114" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Nov 12 17:36:58 2024 -0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 22:07:41 2024 +0800" }, "message": "xfs/273: check thoroughness of the mappings\n\nEnhance this test to make sure that there are no gaps in the fsmap\nrecords, and (especially) that they we report all the way to the end of\nthe device.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "e2affc453cf5a5c172565a6e12be6a45d34b0114", "tree": "18e6e932063f3823a116c5e087c9a007c4eab268", "parents": [ "c3a28053dc09c31b031a733e243e35d78841f801" ], "author": { "name": "Johannes Thumshirn", "email": "jth@kernel.org", "time": "Mon Nov 11 13:24:53 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 14:51:04 2024 +0800" }, "message": "btrfs/136: check for ext3 support\n\nTest-case btrfs/136 requires ext3 support, so check for ext3 using\n_require_extra_fs.\n\nSigned-off-by: Johannes Thumshirn \u003cjohannes.thumshirn@wdc.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c3a28053dc09c31b031a733e243e35d78841f801", "tree": "f2e9d1efdd4b5e0ead43000025ef29426c21b6ce", "parents": [ "3fc00cd4ab986a76ddf1ab355713744749866c86" ], "author": { "name": "André Almeida", "email": "andrealmeid@igalia.com", "time": "Sat Nov 09 20:46:18 2024 -0300" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 14:50:56 2024 +0800" }, "message": "common/casefold: Support for tmpfs casefold test\n\nTest casefold support for tmpfs.\n\nSigned-off-by: André Almeida \u003candrealmeid@igalia.com\u003e\nReviewed-by: Gabriel Krisman Bertazi \u003cgabriel@krisman.be\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "3fc00cd4ab986a76ddf1ab355713744749866c86", "tree": "d6170a4d3e037f8058dba792ac8364d31b7ca6b2", "parents": [ "cdbcb808ecc1f7b0f13943e12997292d50ed324b" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Wed Nov 13 19:58:38 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 14:50:56 2024 +0800" }, "message": "btrfs/321: make the filter to handle older btrfs-progs\n\n[FALSE ALERT]\nWith much older distros like SLE12SP5, which is using btrfs-progs 4.5.3,\ntest case btrfs/321 fails like this:\n\n btrfs/321 QA output created by 321\n unable to locate the last csum tree leaf\n (see /opt/xfstests/results//btrfs/321.full for details)\n [failed, exit status 1]- output mismatch (see /opt/xfstests/results//btrfs/321.out.bad)\n --- tests/btrfs/321.out\t2024-10-28 07:03:54.000000000 -0400\n +++ /opt/xfstests/results//btrfs/321.out.bad\t2024-11-07 09:33:58.238442033 -0500\n @@ -1,2 +1,3 @@\n QA output created by 321\n -Silence is golden\n +unable to locate the last csum tree leaf\n +(see /opt/xfstests/results//btrfs/321.full for details)\n ...\n (Run diff -u /opt/xfstests/tests/btrfs/321.out /opt/xfstests/results//btrfs/321.out.bad to see the entire diff)\n\n[CAUSE]\nThe full output shows the regular csum tree as usual:\n\n btrfs-progs v4.5.3+20160729\n checksum tree key (CSUM_TREE ROOT_ITEM 0)\n node 4247552 level 1 items 9 free 112 generation 7 owner 7\n fs uuid 5623d533-ff79-4ddf-b9a1-7d359fa97c48\n chunk uuid 0af5a7bd-d2d8-4146-ada8-444f2a2f5351\n \tkey (EXTENT_CSUM EXTENT_CSUM 20971520) block 4243456 (1036) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 25006080) block 4251648 (1038) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 29040640) block 4255744 (1039) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 33075200) block 4259840 (1040) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 37109760) block 4263936 (1041) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 41144320) block 4268032 (1042) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 45178880) block 4272128 (1043) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 49213440) block 4276224 (1044) gen 7\n \tkey (EXTENT_CSUM EXTENT_CSUM 53248000) block 4280320 (1045) gen 7\n leaf 4243456 items 1 free space 30 generation 7 owner 7\n fs uuid 5623d533-ff79-4ddf-b9a1-7d359fa97c48\n chunk uuid 0af5a7bd-d2d8-4146-ada8-444f2a2f5351\n \titem 0 key (EXTENT_CSUM EXTENT_CSUM 20971520) itemoff 55 itemsize 3940\n \t\textent csum item\n [...]\n leaf 4280320 items 1 free space 2722 generation 7 owner 7\n fs uuid 5623d533-ff79-4ddf-b9a1-7d359fa97c48\n chunk uuid 0af5a7bd-d2d8-4146-ada8-444f2a2f5351\n \titem 0 key (EXTENT_CSUM EXTENT_CSUM 53248000) itemoff 2747 itemsize 1248\n \t\textent csum item\n total bytes 25768755200\n bytes used 34213888\n uuid 5623d533-ff79-4ddf-b9a1-7d359fa97c48\n\nBut notice the header for each leaf, there is no flags for the leaf.\nOn newer btrfs-progs, the leaf header lines looks like this:\n\n leaf 5423104 items 1 free space 2918 generation 7 owner CSUM_TREE\n leaf 5423104 flags 0x1(WRITTEN) backref revision 1\n\nIt\u0027s two lines, not the old one line output.\nThe new behavior is introduced in btrfs-progs commit 9cc9c9ab3220\n(\"btrfs-progs: print the eb flags for nodes as well\"), included by v5.10\nrelease.\n\nSo the test case doesn\u0027t handle older output format and failed to locate\nthe target leaf.\n\n[FIX]\nInstead of relying on the leaf flags line, use the much older\n\"leaf \u003cbytenr\u003e items\" line as the filter target, so we can support much\nolder distros.\n\nReported-by: Long An \u003clan@suse.com\u003e\nLink: https://bugzilla.suse.com/show_bug.cgi?id\u003d1233303\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "cdbcb808ecc1f7b0f13943e12997292d50ed324b", "tree": "51e5141bc7f6e507c6b358ba05629a8d2b12a69c", "parents": [ "81477600d36391dcbdfefd0b258fa55c17ae6932" ], "author": { "name": "Alexander Mikhalitsyn", "email": "aleksandr.mikhalitsyn@canonical.com", "time": "Mon Nov 11 16:21:00 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 15 14:50:56 2024 +0800" }, "message": "ext4/032: add a new testcase in online resize tests\n\nAdd a new testcase for [1] commit in ext4 online resize testsuite.\n\nLink: https://lore.kernel.org/linux-ext4/20240927133329.1015041-1-libaokun@huaweicloud.com [1]\nSigned-off-by: Alexander Mikhalitsyn \u003caleksandr.mikhalitsyn@canonical.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nReviewed-by: Baokun Li \u003clibaokun1@huawei.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "81477600d36391dcbdfefd0b258fa55c17ae6932", "tree": "1f0b2ecf0ea6df38d803ff2354eba76161de6710", "parents": [ "1e990ee9aaa05a4248a243be7385a6d7de6db724" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Wed Nov 06 16:13:28 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Nov 09 22:34:05 2024 +0800" }, "message": "btrfs: a new test case to verify mount behavior with background remounting\n\n[BUG]\nWhen there is a process in the background remounting a btrfs, switching\nbetween RO/RW, then another process try to mount another subvolume of\nthe same btrfs read-only, we can hit a race causing the RW mount to fail\nwith -EBUSY:\n\n[CAUSE]\nDuring the btrfs mount, to support mounting different subvolumes with\ndifferent RO/RW flags, we have a small hack during the mount:\n\n Retry with matching RO flags if the initial mount fail with -EBUSY.\n\nThe problem is, during that retry we do not hold any super block lock\n(s_umount), this meanings there can be a remount process changing the RO\nflags of the original fs super block.\n\nIf so, we can have an EBUSY error during retry.\nAnd this time we treat any failure as an error, without any retry and\ncause the above EBUSY mount failure.\n\n[FIX]\nThe fix is already sent to the mailing list.\nThe fix is to allow btrfs to have different RO flag between super block\nand mount point during mount, and if the RO flag mismatch, reconfigure\nthe fs to RW with s_umount hold, so that there will be no race.\n\n[TEST CASE]\nThe test case will create two processes:\n\n- Remounting an existing subvolume mount point\n Switching between RO and RW\n\n- Mounting another subvolume RW\n After a successful mount, unmount and retry.\n\nThis is enough to trigger the -EBUSY error in less than 5 seconds.\nTo be extra safe, the test case will run for 10 seconds at least, and\nfollow TIME_FACTOR for extra loads.\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "1e990ee9aaa05a4248a243be7385a6d7de6db724", "tree": "31fbaf5bcf9950caff53c836ca6a066caeba9423", "parents": [ "083bb0ea9473e709cc11ba0314fc4ab2ad8710a7" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Mon Nov 04 11:41:23 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Nov 06 22:10:41 2024 +0800" }, "message": "btrfs: add a test for defrag of contiguous file extents\n\nTest that defrag merges adjacent extents that are contiguous.\nThis exercises a regression fixed by a patchset for the kernel that is\ncomprissed of the following patches:\n\n btrfs: fix extent map merging not happening for adjacent extents\n btrfs: fix defrag not merging contiguous extents due to merged extent maps\n\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "083bb0ea9473e709cc11ba0314fc4ab2ad8710a7", "tree": "5f020a5a9c02d6518c78c6903d5069ce1ef831e5", "parents": [ "2fe95b92d6da3669ceaa8d01590448cc9ccac1c6" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Tue Oct 29 13:21:35 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 01 04:01:53 2024 +0800" }, "message": "xfs: online grow vs. log recovery stress test (realtime version)\n\nThis is fundamentally the same as the previous growfs vs. log\nrecovery test, with tweaks to support growing the XFS realtime\nvolume on such configurations. Changes include using the appropriate\nmkfs params, growfs params, and enabling realtime inheritance on the\nscratch fs.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redaht.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "2fe95b92d6da3669ceaa8d01590448cc9ccac1c6", "tree": "71c96c03024c47dce183b9f01f7bcd07ac7573f3", "parents": [ "d9228b41b30fa00973153d688638a336081b2627" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Tue Oct 29 13:21:34 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Nov 01 04:00:41 2024 +0800" }, "message": "xfs: online grow vs. log recovery stress test\n\nfstests includes decent functional tests for online growfs and\nshrink, and decent stress tests for crash and log recovery, but no\ncombination of the two. This test combines bits from a typical\ngrowfs stress test like xfs/104 with crash recovery cycles from a\ntest like generic/388. As a result, this reproduces at least a\ncouple recently fixed issues related to log recovery of online\ngrowfs operations.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d9228b41b30fa00973153d688638a336081b2627", "tree": "00be6456debd1f5171f929a06d0d63b5d4816a95", "parents": [ "7838e3082b15354323f29efbcf6380dddb11a751" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Tue Oct 29 17:21:28 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 30 16:21:26 2024 +0800" }, "message": "btrfs/287: make the test work when compression is enabled\n\nWhen running btrfs/287 with compression enabled (mount options), the test\nfails because it expects to find 4M extents, however compression limits\nthe maximum size of extents to 128K, breaking the tests\u0027 expectations.\n\nExample:\n\n $ MOUNT_OPTIONS\u003d\"-o compress\" ./check btrfs/287\n FSTYP -- btrfs\n PLATFORM -- Linux/x86_64 debian0 6.12.0-rc4-btrfs-next-177+ #1 SMP PREEMPT_DYNAMIC Thu Oct 24 17:14:37 WEST 2024\n MKFS_OPTIONS -- /dev/sdc\n MOUNT_OPTIONS -- -o compress /dev/sdc /home/fdmanana/btrfs-tests/scratch_1\n\n btrfs/287 2s ... - output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/287.out.bad)\n --- tests/btrfs/287.out\t2024-10-19 18:21:30.451644840 +0100\n +++ /home/fdmanana/git/hub/xfstests/results//btrfs/287.out.bad\t2024-10-29 16:31:20.926612583 +0000\n @@ -25,22 +25,14 @@\n resolve first extent with ignore offset option:\n inode 257 offset 16777216 root 5\n inode 257 offset 8388608 root 5\n -inode 257 offset 2097152 root 5\n resolve first extent +1M offset:\n -inode 257 offset 17825792 root 5\n -inode 257 offset 9437184 root 5\n ...\n (Run \u0027diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/287.out /home/fdmanana/git/hub/xfstests/results//btrfs/287.out.bad\u0027 to see the entire diff)\n\n HINT: You _MAY_ be missing kernel fix:\n 0cad8f14d70c btrfs: fix backref walking not returning all inode refs\n\n Ran: btrfs/287\n Failures: btrfs/287\n Failed 1 of 1 tests\n\nFix this by creating the two 4M extents with fallocate, so that the test\nworks regardless of compression being enabled or not.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "7838e3082b15354323f29efbcf6380dddb11a751", "tree": "fbe2dbbe601f4324fd4796b0ef25199bca8a7e0c", "parents": [ "23ebca50151d3e12f6a4675e4269c10774f2a772" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Tue Oct 29 18:26:44 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 30 16:19:43 2024 +0800" }, "message": "f2fs/007: add testcase to check consistency of compressed inode metadata\n\nmetadata of compressed inode should always be consistent after file\ncompression, reservation, releasement and decompression, let\u0027s add\na testcase to check it.\n\nCc: Jaegeuk Kim \u003cjaegeuk@kernel.org\u003e\nCc: Qi Han \u003chanqi@vivo.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "23ebca50151d3e12f6a4675e4269c10774f2a772", "tree": "f625c8a0613286550ad49dbf701d87084e6499cb", "parents": [ "5089b050a6724c8226f28ab3e7c70912bc610299" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Tue Oct 29 18:26:43 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 30 16:19:43 2024 +0800" }, "message": "f2fs/006: add testcase to check out-of-space case\n\nThis is a regression test to check whether f2fs handles dirty\ndata correctly when checkpoint is disabled, if lfs mode is on,\nit will trigger OPU for all overwritten data, this will cost\nfree segments, so f2fs must account overwritten data as OPU\ndata when calculating free space, otherwise, it may run out\nof free segments in f2fs\u0027 allocation function. If kernel config\nCONFIG_F2FS_CHECK_FS is on, it will cause system panic, otherwise,\ndd may encounter I/O error.\n\nCc: Jaegeuk Kim \u003cjaegeuk@kernel.org\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "5089b050a6724c8226f28ab3e7c70912bc610299", "tree": "84f69acc42507031cc71676bf9cf0bfda680021a", "parents": [ "4c1629ae3a3a56d467a010f48f83e3511aa63943" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Mon Oct 28 22:17:00 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 30 16:19:43 2024 +0800" }, "message": "f2fs/005: add testcase to check checkpoint disabling functionality\n\nThis patch introduce a regression test to check whether f2fs handles\ndirty inode correctly when checkpoint is disabled in a corner case,\nit may hang umount before the bug is fixed.\n\nCc: Qi Han \u003chanqi@vivo.com\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "4c1629ae3a3a56d467a010f48f83e3511aa63943", "tree": "d5faa4bab874dacf320e91243178a2a5369c4f99", "parents": [ "66434a1cb73131f45457208ca0a0382ff8708067" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Tue Oct 29 10:05:25 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 30 16:19:43 2024 +0800" }, "message": "generic: new test case to verify if certain fio load will hang the filesystem\n\n[BUG]\nDuring the development to make btrfs pass generic/563 (which needs to\nmake btrfs to support partial folios), generic/095 causes hangs\nduring tests.\n\nThe call trace for the hanging process looks like this:\n\n __switch_to+0xf8/0x168\n __schedule+0x328/0x8a8\n schedule+0x54/0x140\n io_schedule+0x44/0x68\n folio_wait_bit_common+0x198/0x3f8\n __folio_lock+0x24/0x40\n extent_write_cache_pages+0x2e0/0x4c0 [btrfs]\n btrfs_writepages+0x94/0x158 [btrfs]\n do_writepages+0x74/0x190\n filemap_fdatawrite_wbc+0x88/0xc8\n __filemap_fdatawrite_range+0x6c/0xa8\n filemap_fdatawrite_range+0x1c/0x30\n btrfs_start_ordered_extent+0x264/0x2e0 [btrfs]\n btrfs_lock_and_flush_ordered_range+0x8c/0x160 [btrfs]\n __get_extent_map+0xa0/0x220 [btrfs]\n btrfs_do_readpage+0x1bc/0x5d8 [btrfs]\n btrfs_read_folio+0x50/0xa0 [btrfs]\n filemap_read_folio+0x54/0x110\n filemap_update_page+0x2e0/0x3b8\n filemap_get_pages+0x228/0x4d8\n filemap_read+0x11c/0x3b8\n btrfs_file_read_iter+0x74/0x90 [btrfs]\n new_sync_read+0xd0/0x1d0\n vfs_read+0x1a0/0x1f0\n\n[CAUSE]\nThe root cause is a btrfs specific behavior that during a folio read, we\ncan trigger writeback of the same folio, which will try to lock the same\nfolio already locked by the read process.\n\nThe fix is already sent to the mailing list:\nhttps://lore.kernel.org/linux-btrfs/62bf73ada7be2888d45a787c2b6fd252103a5d25.1729725088.git.wqu@suse.com/\n\nThis problem can only happen if all the following conditions are met:\n\n- The sector size of btrfs is smaller than page size\n To have partial uptodate folios.\n\n- Btrfs won\u0027t read the full folio if buffered write is block aligned\n This is done by the not yet merged patch:\n https://lore.kernel.org/linux-btrfs/ac2639ec4e9ac176d33e95ef7ecf008fa6be5461.1727833878.git.wqu@suse.com/\n\n[TEST CASE]\nDuring the debugging of that generic/095 hang, I extracted a minimal\nreproducer which is much smaller and faster, although it still requires\nseveral runs to trigger a hang.\n\nThe test case will run the fio workload 32 times by default, which is\nmore than enough to trigger the hang.\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "66434a1cb73131f45457208ca0a0382ff8708067", "tree": "082ab12db14e5b2b716ad484c2d1244c70eca9b3", "parents": [ "f056971a4d2f2314b162e8e2114514dc86cd8515" ], "author": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Oct 27 04:12:34 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Mon Oct 28 21:36:16 2024 +0800" }, "message": "xfs: notrun if kernel xfs not supports ascii-ci feature\n\nAs the ascii-ci feature is deprecated, if linux build without the\nCONFIG_XFS_SUPPORT_ASCII_CI, mount xfs with \"-n version\u003dci\" will\nget EINVAL. So let\u0027s notrun if it\u0027s not supported by kernel.\n\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "f056971a4d2f2314b162e8e2114514dc86cd8515", "tree": "ab81734a735fd21e367b94c73c76b2a78c0ab8a4", "parents": [ "6d97ae8ddebf56b029a256a2292036995ec900fb" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Fri Oct 25 18:24:21 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 26 22:28:38 2024 +0800" }, "message": "btrfs: test remount with \"compress\" clears \"compress-force\"\n\nTest that remounting with the \"compress\" mount option clears the\n\"compress-force\" mount option previously specified.\n\nThis tests a regression introduced with kernel 6.8 and recently fixed by\nthe following kernel commit:\n\n 3510e684b8f6 (\"btrfs: clear force-compress on remount when compress mount option is given\")\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "6d97ae8ddebf56b029a256a2292036995ec900fb", "tree": "f9a5d39c91d2061abcff95beb23e273ce1f16018", "parents": [ "b71de9073a655a01941beb9e40a45c56cfd3bb38" ], "author": { "name": "Pankaj Raghav", "email": "p.raghav@samsung.com", "time": "Thu Oct 24 13:23:11 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 26 00:59:36 2024 +0800" }, "message": "generic: increase file size to match CoW delayed allocation for XFS 64k bs\n\ngeneric/305,326,328 have been failing for 32k and 64k blocksizes.\n\nWe do the following in the test 305 and 326 (highlighting only the part\nthat is related to failure):\n\n- create a 1M test-1/file1\n- reflink test-1/file2 and test-1/file3 based on test-1/file1\n- Overwrite first half of test-1/file2 to do a CoW operation\n- Expect the size of the test-1 dir to be 3M\n\nThe test is failing for 32k and 64k blocksizes as the number of blocks\n(direct + delayed) is higher than number of blocks allocated for\nblocksizes \u003c 32k in XFS, resulting in size of test-1 to be more than 3M.\nThough generic/328 has a different IO pattern, the reason for failure is\nthe same.\n\nThis is the failure output :\n --- tests/generic/305.out 2024-06-05 11:52:27.430262812 +0000\n +++ /root/results//64k_4ks/generic/305.out.bad 2024-10-23 10:56:57.643986870 +0000\n @@ -11,7 +11,7 @@\n CoW one of the files\n root 0 0 0\n nobody 0 0 0\n -fsgqa 3072 0 0\n +fsgqa 4608 0 0\n Remount the FS to see if accounting changes\n root 0 0 0\n\nIn these tests, XFS is doing a delayed allocation of\nXFS_DEFAULT_COWEXTSIZE_HINT(32). Increase the size of the file so that\nthe CoW write(sz/2) matches the maximum size of the delayed allocation\nfor the max blocksize of 64k. This will ensure that all parts of the\ndelayed extents are converted to real extents for all blocksizes.\n\nEven though this is not the most complete solution to fix these tests,\nthe objective of these tests are to test quota and not the effect of delayed\nallocations.\n\nSigned-off-by: Pankaj Raghav \u003cp.raghav@samsung.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "b71de9073a655a01941beb9e40a45c56cfd3bb38", "tree": "f86f66e8589cc5d0461c0a56f883f854756eb3de", "parents": [ "18795160b0f324401a62b4e8e2b3a1f8753a896b" ], "author": { "name": "Pankaj Raghav", "email": "p.raghav@samsung.com", "time": "Thu Oct 24 13:23:10 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 26 00:58:59 2024 +0800" }, "message": "generic/219: use filesystem blocksize while calculating the file size\n\ngeneric/219 was failing for XFS with 32k and 64k blocksize. Even though\nwe do only 48k IO, XFS will allocate blocks rounded to the nearest\nblocksize.\n\nSigned-off-by: Pankaj Raghav \u003cp.raghav@samsung.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "18795160b0f324401a62b4e8e2b3a1f8753a896b", "tree": "3a54fd11d493c452829e1b3a832368a35fb7b999", "parents": [ "6f4a6b5789d7584e76964bb3ae366df35dce2138" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Sat Oct 12 09:14:19 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Oct 25 12:03:08 2024 +0800" }, "message": "f2fs/004: add missing _fixed_by_kernel_commit line\n\nThe bug related to this regression testcase has been fixed by commit\nb2c160f4f3cf (\"f2fs: atomic: fix to forbid dio in atomic_file\"), let\u0027s\nadd missing _fixed_by_kernel_commit line for this testcase.\n\nCc: Jaegeuk Kim \u003cjaegeuk@kernel.org\u003e\nCc: Daeho Jeong \u003cdaehojeong@google.com\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "6f4a6b5789d7584e76964bb3ae366df35dce2138", "tree": "efed6d0642a1720e0cfb02cf99ea7e24ee854f9c", "parents": [ "9c3762ceafd430bb4442d07d7bd89c55987729ed" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Sat Oct 12 17:48:24 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Oct 25 11:32:33 2024 +0800" }, "message": "fstests: btrfs/002: fix the OOM caused by too large block size\n\n[BUG]\nWhen running the test case btrfs/002, with 64K page size and 64K sector\nsize, and the VM doesn\u0027t have much memory (in my case 4G Vram), the test\ncase will trigger OOM and fail:\n\nbtrfs/002 4s ... [failed, exit status 1]- output mismatch (see /home/adam/xfstests-dev/results//btrfs/002.out.bad)\n --- tests/btrfs/002.out\t2024-04-25 18:13:45.035555469 +0930\n +++ /home/adam/xfstests-dev/results//btrfs/002.out.bad\t2024-10-12 17:19:48.785156223 +1030\n @@ -1,2 +1 @@\n QA output created by 002\n -Silence is golden\n ...\n\nThe OOM is triggered by the dd process, and a lot of dd processes are\nusing too much memory:\n\n dd invoked oom-killer: gfp_mask\u003d0x140dca(GFP_HIGHUSER_MOVABLE|__GFP_COMP|__GFP_ZERO), order\u003d0, oom_score_adj\u003d250\n CPU: 0 UID: 0 PID: 185764 Comm: dd Not tainted 6.12.0-rc2-custom+ #76\n Hardware name: QEMU KVM Virtual Machine, BIOS unknown 2/2/2022\n Tasks state (memory values in pages):\n [ pid ] uid tgid total_vm rss rss_anon rss_file rss_shmem pgtables_bytes swapents oom_score_adj name\n [ 185665] 0 185665 8688 3840 3840 0 0 458752 4832 250 dd\n [ 185672] 0 185672 8688 2432 2432 0 0 393216 5312 250 dd\n [ 185680] 0 185680 8688 2016 2016 0 0 458752 4960 250 dd\n [ 185686] 0 185686 8688 2080 2080 0 0 458752 3584 250 dd\n [ 185693] 0 185693 8688 2144 2144 0 0 458752 4384 250 dd\n [ 185700] 0 185700 8688 2176 2176 0 0 458752 3584 250 dd\n [ 185707] 0 185707 8688 1792 1792 0 0 524288 3616 250 dd\n [ 185714] 0 185714 8688 2304 2304 0 0 458752 3488 250 dd\n [ 185721] 0 185721 8688 1920 1920 0 0 458752 2624 250 dd\n [ 185728] 0 185728 8688 2272 2272 0 0 393216 2528 250 dd\n [ 185735] 0 185735 8688 2048 2048 0 0 393216 3552 250 dd\n [ 185742] 0 185742 8688 1984 1984 0 0 458752 2816 250 dd\n [ 185751] 0 185751 8688 1600 1600 0 0 458752 2784 250 dd\n [ 185756] 0 185756 8688 1120 1120 0 0 458752 2400 250 dd\n [ 185764] 0 185764 8688 1504 1504 0 0 393216 2240 250 dd\n [ 185772] 0 185772 8688 1504 1504 0 0 458752 1984 250 dd\n [ 185777] 0 185777 8688 1280 1280 0 0 393216 2336 250 dd\n [ 185784] 0 185784 8688 2144 2144 0 0 393216 2272 250 dd\n [ 185791] 0 185791 8688 2176 2176 0 0 458752 576 250 dd\n [ 185798] 0 185798 8688 1696 1696 0 0 458752 1536 250 dd\n [ 185806] 0 185806 8688 1728 1728 0 0 393216 544 250 dd\n [ 185815] 0 185815 8688 2240 2240 0 0 458752 0 250 dd\n [ 185819] 0 185819 8688 1504 1504 0 0 458752 384 250 dd\n [ 185826] 0 185826 8688 1536 1536 0 0 458752 160 250 dd\n [ 185833] 0 185833 8688 2944 2944 0 0 458752 64 250 dd\n [ 185838] 0 185838 8688 2400 2400 0 0 458752 0 250 dd\n [ 185847] 0 185847 8688 864 864 0 0 458752 0 250 dd\n [ 185853] 0 185853 8688 1088 1088 0 0 393216 0 250 dd\n [ 185860] 0 185860 8688 416 416 0 0 393216 0 250 dd\n [ 185867] 0 185867 8688 352 352 0 0 458752 0 250 dd\n\n[CAUSE]\nThe test case workload _fill_blk() is going to fill the file to its block\nboundary.\n\nBut the implementation is not taking larger blocks into consideration.\n\n\tFSIZE\u003d`stat -t $i | cut -d\" \" -f2`\n\tBLKS\u003d`stat -c \"%B\" $i`\n\tNBLK\u003d`stat -c \"%b\" $i`\n\tFALLOC\u003d$(($BLKS * $NBLK))\n\tWS\u003d$(($FALLOC - $FSIZE))\n\n$FSIZE is the file size, $BLKS is the size of each reported block,\n$NBLK is the number of blocks the file takes, thus $FALLOC is the\nrounded up block size.\n\nFor 64K sector size, the BLKS is 512, and NBLKS is 128 (one 64K sector).\n$FALLOC is the correct value of 64K (10K rounded up to 64K).\n\nThen the problem comes to how the write is done:\n\n\t_ddt of\u003d$i oseek\u003d$FSIZE obs\u003d$WS count\u003d1 status\u003dnoxfer 2\u003e/dev/null \u0026\n\nUnfrotunately the above command is using output block size of 54K, and\nneed to skip 10K * 54K bytes, resulting a file size of 540M.\n\nSo far although it\u0027s not the correct intention, it\u0027s not yet causing\nproblem.\n\nBut during _append_file(), we further enlarge the file by:\n\n\t\tFSIZE\u003d`stat -t $i | cut -d\" \" -f2`\n\t\tdd if\u003d$X of\u003d$i seek\u003d1 bs\u003d$FSIZE obs\u003d$FSIZE count\u003d1 status\u003dnoxfer 2\u003e/dev/null \u0026\n\nIn above case, since the previous file is 540M size, the output block\nsize will also be 540M, taking a lot of memory.\n\nFurthermore since the workload is run in background, we can have many dd\nprocesses taking up at least 540M, causing huge memory usage and trigger\nOOM.\n\n[FIX]\nThe original code is already not doing what it should do, just get rid of\nthe cursed dd command usage inside _fill_blk(), and use pwrite from\nxfs_io instead.\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "9c3762ceafd430bb4442d07d7bd89c55987729ed", "tree": "2e95d7d5f69344a511017739d85cdad1480bb242", "parents": [ "0ee667ece05c790f2d91422ff7c54eeee49d57fb" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Wed Oct 16 16:15:48 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Oct 25 11:24:11 2024 +0800" }, "message": "misc: amend unicode confusing name tests to check for hidden tag characters\n\nThe Unicode consortium has twice defined (and later deprecated) special\n\"tag\" codepoints. These tag codepoints are not supposed to be rendered\n(i.e. they\u0027re invisible) but you can certainly encode them in\ndirectories and labels to try to confuse users.\n\nxfs_scrub already knows how complain about these tag characters because\nlibicu can detect both their presence and their use in confusing name\nattacks, so add this as an explicit regression test.\n\nLink: https://embracethered.com/blog/posts/2024/hiding-and-finding-text-with-unicode-tags/\nLink: https://arstechnica.com/security/2024/10/ai-chatbots-can-read-and-write-invisible-text-creating-an-ideal-covert-channel/\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "0ee667ece05c790f2d91422ff7c54eeee49d57fb", "tree": "2dd0f6a731c4432b8165a71633f4e9f0d47f1cbb", "parents": [ "0d203737049ddafb9bbec4ad404234e8bb006748" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Thu Oct 24 20:35:03 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Oct 25 11:00:47 2024 +0800" }, "message": "fstests: btrfs/330: enable the test case for both new and old APIs\n\n[BUG]\nIf the mount tool is utilizing the new fs-based API\n(e.g. util-linux 2.40.2 from Archlinux), btrfs\u0027 per-subvolume RO/RW mount\nis broken again:\n\n # mount -o subvol\u003dsubv1,ro /dev/test/scratch1 /mnt/test\n # mount -o rw,subvol\u003dsubv2 /dev/test/scratch1 /mnt/scratch\n # mount | grep mnt\n /dev/mapper/test-scratch1 on /mnt/test type btrfs (ro,relatime,discard\u003dasync,space_cache\u003dv2,subvolid\u003d256,subvol\u003d/subv1)\n /dev/mapper/test-scratch1 on /mnt/scratch type btrfs (ro,relatime,discard\u003dasync,space_cache\u003dv2,subvolid\u003d257,subvol\u003d/subv2)\n # touch /mnt/scratch/foobar\n touch: cannot touch \u0027/mnt/scratch/foobar\u0027: Read-only file system\n\n[CAUSE]\nBtrfs has an extra remount hack to handle above case, which will\nre-configure the super block to be RW on the first RW mount.\n\nThe initial promise is, the new fd-based API will not set ro FLAG, but\nonly MOUNT_ATTR_RDONLY, so that btrfs will skip the remount hack for new\nAPI based mount request.\n\nHowever it\u0027s not the case, the first RO subvolume mount will set ro flag\nat fsconfig(), and also set MOUNT_ATTR_RDONLY attribute for the mount\npoint:\n\n # strace mount -o subvol\u003dsubv1,ro /dev/test/scratch1 /mnt/test/\n ...\n fsconfig(3, FSCONFIG_SET_STRING, \"source\", \"/dev/mapper/test-scratch1\", 0) \u003d 0\n fsconfig(3, FSCONFIG_SET_STRING, \"subvol\", \"subv1\", 0) \u003d 0\n fsconfig(3, FSCONFIG_SET_FLAG, \"ro\", NULL, 0) \u003d 0\n fsconfig(3, FSCONFIG_CMD_CREATE, NULL, NULL, 0) \u003d 0\n fsmount(3, FSMOUNT_CLOEXEC, 0) \u003d 4\n mount_setattr(4, \"\", AT_EMPTY_PATH, {attr_set\u003dMOUNT_ATTR_RDONLY, attr_clr\u003d0, propagation\u003d0 /* MS_??? */, userns_fd\u003d0}, 32) \u003d 0\n move_mount(4, \"\", AT_FDCWD, \"/mnt/test\", MOVE_MOUNT_F_EMPTY_PATH) \u003d 0\n\nThis will result exactly the same behavior, no matter if it\u0027s the new\nAPI or not.\n\nFurthermore we can even have corner cases like mounting the initial RO\nsubvolume using the old API, then mount the RW subvolume using the new\nAPI.\n\nSo even using the new API, there is no guarantee to keep the\nper-subvolume RO/RW mount feature.\nWe have to do the reconfigure anyway.\n\n[FIX]\nThe kernel fix is already submitted, but for the test case part, we\nshould enable btrfs/330 for all mount tools, no matter the API it\nutilizes.\n\nThe only difference for the new API based mount is the new\n_fixed_by_kernel_commit call, to show the proper fix.\n\nNow it can properly detects the broken feature.\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "0d203737049ddafb9bbec4ad404234e8bb006748", "tree": "70b3a47ca48c6a35dd65333bb28daf77b78cee2d", "parents": [ "bd0d0179baf1f88b333c2240f4f181a59dc381e5" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Fri Oct 18 10:34:53 2024 +1030" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 24 20:48:38 2024 +0800" }, "message": "btrfs/012: fix false alerts when SELinux is enabled\n\n[FALSE FAILURE]\nIf SELinux is enabled, the test btrfs/012 will fail due to metadata\nmismatch:\n\nFSTYP -- btrfs\nPLATFORM -- Linux/x86_64 localhost 6.4.0-150600.23.25-default #1 SMP PREEMPT_DYNAMIC Tue Oct 1 10:54:01 UTC 2024 (ea7c56d)\nMKFS_OPTIONS -- /dev/loop1\nMOUNT_OPTIONS -- -o context\u003dsystem_u:object_r:root_t:s0 /dev/loop1 /mnt/scratch\n\nbtrfs/012 - output mismatch (see /home/adam/xfstests-dev/results//btrfs/012.out.bad)\n --- tests/btrfs/012.out\t2024-10-18 10:15:29.132894338 +1030\n +++ /home/adam/xfstests-dev/results//btrfs/012.out.bad\t2024-10-18 10:25:51.834819708 +1030\n @@ -1,6 +1,1390 @@\n QA output created by 012\n Checking converted btrfs against the original one:\n -OK\n +metadata mismatch in /p0/d2/f4\n +metadata mismatch in /p0/d2/f5\n +metadata and data mismatch in /p0/d2/\n +metadata and data mismatch in /p0/\n ...\n\n[CAUSE]\nAll the mismatch happens in the metadata, to be more especific, it\u0027s the\nsecurity xattrs.\n\nAlthough btrfs-convert properly convert all xattrs including the\nsecurity ones, at mount time we will get new SELinux labels, causing the\nmismatch between the converted and original fs.\n\n[FIX]\nOverride SELINUX_MOUNT_OPTIONS so that we will not touch the security\nxattrs, and that should fix the false alert.\n\nReported-by: Long An \u003clan@suse.com\u003e\nLink: https://bugzilla.suse.com/show_bug.cgi?id\u003d1231524\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "bd0d0179baf1f88b333c2240f4f181a59dc381e5", "tree": "a52f7520972ad69f3da87fc81f74be965732d871", "parents": [ "1f8fedd08f0683f395a7fc645851517b2c5711e1" ], "author": { "name": "Pankaj Raghav", "email": "p.raghav@samsung.com", "time": "Wed Oct 16 16:15:32 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 23 12:23:57 2024 +0800" }, "message": "xfs/161: adapt the test case for LBS filesystem\n\nThis test fails for \u003e\u003d 64k filesystem block size on a 4k PAGE_SIZE\nsystem(see LBS efforts[1]). Adapt the blksz so that we create more than\none block for the testcase.\n\nCap the blksz to be at least 64k to retain the same behaviour as before\nfor smaller filesystem blocksizes.\n\n[1] LBS effort: https://lore.kernel.org/lkml/20230915183848.1018717-1-kernel@pankajraghav.com/\n\nSigned-off-by: Pankaj Raghav \u003cp.raghav@samsung.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "1f8fedd08f0683f395a7fc645851517b2c5711e1", "tree": "70ddb14a2962f7e0d53a7e6fd6b443576ba5391e", "parents": [ "cf97fa373d61e14d3d661fe367a3449996893867" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Wed Oct 16 16:15:16 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 23 12:23:46 2024 +0800" }, "message": "common/xfs: _notrun tests that fail due to block size \u003c sector size\n\nIt makes no sense to fail a test that failed to format a filesystem with\na block size smaller than the sector size since the test preconditions\nare not valid.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "cf97fa373d61e14d3d661fe367a3449996893867", "tree": "d54df8adf3ab7a468899bfe6e000b28d6d6b5564", "parents": [ "9eda73ff46959ab6f0504c46062f9071a8cedc3b" ], "author": { "name": "Mark Harmstone", "email": "maharmstone@fb.com", "time": "Tue Oct 15 16:39:34 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 23 11:56:04 2024 +0800" }, "message": "generic: add test for missing btrfs csums in log when doing async on subpage vol\n\nAdds a test for a bug we encountered on Linux 6.4 on aarch64, where a\nrace could mean that csums weren\u0027t getting written to the log tree,\nleading to corruption when it was replayed.\n\nThe patches to detect log this tree corruption are in btrfs-progs 6.11.\n\nSigned-off-by: Mark Harmstone \u003cmaharmstone@fb.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "9eda73ff46959ab6f0504c46062f9071a8cedc3b", "tree": "42fc17683638ca07529ff9c23d828d31b347b50e", "parents": [ "891f4995ab07ee0a07eca156915ed87ab5f479f6" ], "author": { "name": "Boris Burkov", "email": "boris@bur.io", "time": "Wed Oct 16 10:54:33 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Wed Oct 23 11:14:23 2024 +0800" }, "message": "btrfs: add test for cleaner thread under seed-sprout\n\nWe have a longstanding bug that creating a seed sprout fs with the\nro-\u003erw transition done with\n\nmount -o remount,rw $mnt\n\ninstead of\n\numount $mnt\nmount $sprout_dev $mnt\n\nresults in an fs without BTRFS_FS_OPEN set, which fails to ever run the\ncritical btrfs cleaner thread.\n\nThis test reproduces that bug and detects it by creating and deleting a\nsubvolume, then triggering the cleaner thread. The expected behavior is\nfor the cleaner thread to delete the stale subvolume and for the list to\nshow no entries. Without the fix, we see a DELETED entry for the subvol.\n\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Boris Burkov \u003cboris@bur.io\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "891f4995ab07ee0a07eca156915ed87ab5f479f6", "tree": "e45ac8d8ba6975709e7b0b0c5747f073891fa601", "parents": [ "ff07e0a329361c752d77d0f446d247f0ebdb5418" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Oct 01 09:49:11 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:46 2024 +0800" }, "message": "src/fiexchange.h: add the start-commit/commit-range ioctls\n\nAdd these two ioctls as well, since they\u0027re a part of the file content\nexchange functionality.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "ff07e0a329361c752d77d0f446d247f0ebdb5418", "tree": "56b7283a4d3a66ed341c677996f3914a56e0a3bd", "parents": [ "dd0eeae9ead0706fa1102d87ad02e822f2f02133" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Thu Oct 10 18:07:38 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "fsstress: add support for FALLOC_FL_UNSHARE_RANGE\n\nTeach fsstress to try to unshare file blocks on filesystems, seeing how\nthe recent addition to fsx has uncovered a lot of bugs.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nReviewed-by: Brian Foster \u003cbfoster@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "dd0eeae9ead0706fa1102d87ad02e822f2f02133", "tree": "38c87e6a2d50cd8307fd07bb937c991c3bbc2395", "parents": [ "cf838edbaec793c0172686f942b8025418ab33cc" ], "author": { "name": "Daniel Gomez", "email": "da.gomez@samsung.com", "time": "Sun Jun 30 23:52:40 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "common/config: fix RECREATE_TEST_DEV initialization\n\nDo not allow the overwriting of the RECREATE_TEST_DEV variable. When\nthis variable is enabled, common/rc -\u003e common/config will reset it\nto false after the test device recreation process. This allows for\ndifferentiation in mount options for SCRATCH and TEST.\n\nSigned-off-by: Daniel Gomez \u003cda.gomez@samsung.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "cf838edbaec793c0172686f942b8025418ab33cc", "tree": "18c70ee2bbd8ebdd32139e51a507a2cb5ff73982", "parents": [ "37f61a8fd323669fb39c4191ebd78d6d91e6826f" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Oct 01 09:48:56 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "common/populate: fix bash syntax error in _fill_fs\n\nIn bash, one does not set a variable by prepending the dollar sign to\nthe variable name. Amazingly, this was copied verbatim from generic/256\nin 2016 and hasn\u0027t been caught since its introduction in 2011. :(\n\nCc: allison.henderson@oracle.com\nFixes: 815015e9ee (\"generic: make 17[1-4] work well when btrfs compression is enabled\")\nFixes: b55fb0807c (\"xfstests: Add ENOSPC Hole Punch Test\")\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "37f61a8fd323669fb39c4191ebd78d6d91e6826f", "tree": "0fa7eb5a9e989e698ed672cd8089f9e300e2c133", "parents": [ "8babf87596c84f3c342e3f0d47c7d6a31e3e0602" ], "author": { "name": "An Long", "email": "lan@suse.com", "time": "Fri Oct 11 14:12:16 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "btrfs/315: update filter to match mount cmd\n\nMount error info changed since util-linux v2.40\n(91ea38e libmount: report failed syscall name).\nSo update _filter_mount_error() to match it.\n\nSigned-off-by: An Long \u003clan@suse.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "8babf87596c84f3c342e3f0d47c7d6a31e3e0602", "tree": "b8fac1eb2298013ad04ca68713bb3e3f089c6187", "parents": [ "b37aa9c8b0f9a648b2b1d508a952c1c4047683a8" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Mon Oct 07 13:02:16 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "btrfs/322: add git commit ID\n\nThe corresponding btrfs kernel patch was merged into Linus\u0027 tree and\nincluded in kernel 6.12-rc2, so update the test with the commit ID.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: David Sterba \u003cdsterba@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "b37aa9c8b0f9a648b2b1d508a952c1c4047683a8", "tree": "12f4afa00fceaf1dbb7cb871324e0a93df5eb773", "parents": [ "f4b00fafa4ef091c9d73818f7843e0881ca773ca" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Mon Oct 07 12:32:50 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "btrfs: update some tests to be able to run with btrfs-progs v6.11\n\nIn btrfs-progs v6.11 the output of the \"filesystem show\" command changed\nso that it no longers prints blank lines. This happened with commit\n4331bfb011bd (\"btrfs-progs: fi show: remove stray newline in filesystem\nshow\").\n\nWe have some tests that expect the blank lines in their golden output,\nand therefore they fail with btrfs-progs v6.11.\n\nSo update the filter _filter_btrfs_filesystem_show to remove blank lines\nand change the golden output of the tests to not expect the blank lines,\nmaking the tests work with btrfs-progs v6.11 and older versions.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Boris Burkov \u003cboris@bur.io\u003e\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: David Sterba \u003cdsterba@suse.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "f4b00fafa4ef091c9d73818f7843e0881ca773ca", "tree": "3e2e5cdfe3f878f8d82ef47d5bd99e7c8772f56b", "parents": [ "c9a2881aa6f890211cb0b816cd8f0f202a00cde5" ], "author": { "name": "Mark Harmstone", "email": "maharmstone@fb.com", "time": "Thu Oct 10 16:36:25 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "btrfs/318: add _require_loop\n\nbtrfs/318 uses loopback devices, but was missing a call to _require_loop\nto print the correct message if CONFIG_LOOP is not set.\n\nSigned-off-by: Mark Harmstone \u003cmaharmstone@fb.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c9a2881aa6f890211cb0b816cd8f0f202a00cde5", "tree": "57f0efd042814324ccdf7256fe334c91f207b1d3", "parents": [ "74777e97c56b516e8dd8b0b40d64ffd00d9216fc" ], "author": { "name": "Mark Harmstone", "email": "maharmstone@fb.com", "time": "Wed Jul 31 12:41:33 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Oct 12 22:06:30 2024 +0800" }, "message": "common/fail_make_request: fix error message\n\nfail_make_request depends on the kernel option CONFIG_FAIL_MAKE_REQUEST\nto function, not CONFIG_FAULT_INJECTION_DEBUG_FS.\n\nSigned-off-by: Mark Harmstone \u003cmaharmstone@fb.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "74777e97c56b516e8dd8b0b40d64ffd00d9216fc", "tree": "b8c9ddc275aa175d91bcc1c5922911a51d0ef46a", "parents": [ "2f7e1b8a6f09b6e9f7fcfc69618e05bb76ac10c0" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Tue Oct 08 09:12:09 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 10 17:46:52 2024 +0800" }, "message": "generic/694: sync before sampling i_blocks\n\nWithout a sync there might still be temporary blocks in i_blocks like\nindirect block reservations or additional blocks reserved for out of\nplace writes.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "2f7e1b8a6f09b6e9f7fcfc69618e05bb76ac10c0", "tree": "a9454dcf4cae69250a873bc2c899e3abb60fd5cf", "parents": [ "0ffcda29e8a05a139593a46dea6ef75efdc6e805" ], "author": { "name": "Hans Holmberg", "email": "Hans.Holmberg@wdc.com", "time": "Tue Oct 08 10:52:04 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 10 16:03:40 2024 +0800" }, "message": "xfs/157,xfs/547,xfs/548: switch to using _scratch_mkfs_sized\n\nThese test cases specify small -d sizes which combined with a rt dev of\nunrestricted size and the rtrmap feature can cause mkfs to fail with\nerror:\n\nmkfs.xfs: cannot handle expansion of realtime rmap btree; need \u003cx\u003e free\nblocks, have \u003cy\u003e\n\nThis is due to that the -d size is not big enough to support the\nmetadata space allocation required for the rt groups.\n\nSwitch to use _scratch_mkfs_sized that sets up the -r size parameter\nto avoid this. If -r size\u003dx and -d size\u003dx we will not risk running\nout of space on the ddev as the metadata size is just a fraction of\nthe rt data size.\n\nSigned-off-by: Hans Holmberg \u003chans.holmberg@wdc.com\u003e\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "0ffcda29e8a05a139593a46dea6ef75efdc6e805", "tree": "6ca2f913f9db9269cfcb7893b93bf856587925f6", "parents": [ "034e5655933729ef334b6355c94607d115414282" ], "author": { "name": "Hans Holmberg", "email": "Hans.Holmberg@wdc.com", "time": "Tue Oct 08 10:52:04 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 10 16:03:40 2024 +0800" }, "message": "common: make rt_ops local in _try_scratch_mkfs_sized\n\nIf we call _try_scratch_mkfs_size with $SCRATCH_RTDEV set followed by\na call with $SCRATCH_RTDEV cleared, rt_ops will have stale size\nparameters that will cause mkfs.xfs to fail with:\n\"size specified for non-existent rt subvolume\"\n\nMake rt_ops local to fix this.\n\nSigned-off-by: Hans Holmberg \u003chans.holmberg@wdc.com\u003e\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "034e5655933729ef334b6355c94607d115414282", "tree": "37b1328f9f19ae9d7e6f4466e1cd2f76ed36128f", "parents": [ "e0c84186bfdaa7c48f144d60ff1c8d03a416b113" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Mon Sep 30 09:20:38 2024 +0930" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 10 14:46:21 2024 +0800" }, "message": "fstests: generic/563: use fs blocksize to do the writes\n\n[FALSE ALERTS]\nIf the system has a page size larger than 4K, and the fs block size\nmatches the page size, test case generic/563 will fail:\n\n --- tests/generic/563.out\t2024-04-25 18:13:45.178550333 +0930\n +++ /home/adam/xfstests-dev/results//generic/563.out.bad\t2024-09-30 09:09:16.155312379 +0930\n @@ -3,7 +3,8 @@\n read is in range\n write is in range\n write -\u003e read/write\n -read is in range\n +read has value of 8388608\n +read is NOT in range -33792 .. 33792\n write is in range\n ...\n\nBoth Ext4 and btrfs fail with 64K block size and 64K page size\n\n[CAUSE]\nThe test case writes the 8MiB file using the default block size xfs_io\npwrite, which is 4KiB.\n\nSince the fs block size is 64K, such 4KiB write is unaligned inside a\nblock, causing the fs to read out the full page.\n\nThus the pwrite will cause the fs to read out every page, resulting the\nabove 8MiB+ read value.\n\n[FIX]\nFix the test case by using the fs block size to avoid such unaligned\nbuffered write.\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Boris Burkov \u003cboris@bur.io\u003e\nReviewed-by: Mark Harmstone \u003cmaharmstone@fb.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "e0c84186bfdaa7c48f144d60ff1c8d03a416b113", "tree": "97538abafe3243d01397ee698b4fde61d0a4e911", "parents": [ "4c9c08ec05c66766fbbc1e352d1dc62933fbc85c" ], "author": { "name": "An Long", "email": "lan@suse.com", "time": "Thu Oct 10 14:23:14 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Oct 10 14:24:16 2024 +0800" }, "message": "src/Makefile: install two necessary files\n\nparse-dev-tree.awk and parse-extent-tree.awk are used by generic/746.\nWe need to make sure them are installed, otherwise generic/746 will\nhave problems if fstests is installed via \"make install\".\n\nSigned-off-by: An Long \u003clan@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "4c9c08ec05c66766fbbc1e352d1dc62933fbc85c", "tree": "3192265aa4b4ea16783acd8b55b4eb5e3a8924cc", "parents": [ "790f4d8444fa4bed79d2b441212eec2167414c65" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Wed Sep 11 11:53:33 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Oct 08 11:20:59 2024 +0800" }, "message": "src/min_dio_alignment: don\u0027t return invalid dio_offset_align\n\nIf returned parameters of statx() are: a)STATX_DIOALIGN is set in\nstx_mask, b)stx.stx_dio_offset_align is zero, it indicates filesystem\nsupports DIO, but the file doesn\u0027t.\n\nIt needs to avoid returning zeroed stx.stx_dio_offset_align value,\ninstead, we can fallthrough to get alignment size of block device or\npage size, otherwise, it may cause potential deadloop, e.g.\ngeneric/465:\n\nalign\u003dstx_dio_offset_align(it equals to zero)\npage_size\u003d4096\nwhile [ $align -le $page_size ]; do\n echo \"$AIO_TEST -a $align -d $testfile.$align\" \u003e\u003e $seqres.full\n $AIO_TEST -a $align -d $testfile.$align 2\u003e\u00261 | tee -a $seqres.full\n align\u003d$((align * 2))\ndone\n\nCc: Christoph Hellwig \u003chch@lst.de\u003e\nCc: Eric Biggers \u003cebiggers@kernel.org\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "790f4d8444fa4bed79d2b441212eec2167414c65", "tree": "66b7699aa776b03b67b01af3682248aa39057189", "parents": [ "d26d32028a78aaa972364c6d57894ed04bbcf9ba" ], "author": { "name": "Dave Chinner", "email": "dchinner@redhat.com", "time": "Tue Sep 24 10:45:48 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 29 21:08:52 2024 +0800" }, "message": "xfs: new EOF fragmentation tests\n\nThese tests create substantial file fragmentation as a result of\napplication actions that defeat post-EOF preallocation\noptimisations. They are intended to replicate known vectors for\nthese problems, and provide a check that the fragmentation levels\nhave been controlled. The mitigations we make may not completely\nremove fragmentation (e.g. they may demonstrate speculative delalloc\nrelated extent size growth) so the checks don\u0027t assume we\u0027ll end up\nwith perfect layouts and hence check for an exceptable level of\nfragmentation rather than none.\n\nSigned-off-by: Dave Chinner \u003cdchinner@redhat.com\u003e\n[move to different test number, update to current xfstest APIs]\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d26d32028a78aaa972364c6d57894ed04bbcf9ba", "tree": "0f710f440bf5fcda133ddd163a553087d602d95c", "parents": [ "34ce77ce867d27ad7ad9a180489e4311c3cd15e4" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Fri Sep 27 11:28:07 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 29 21:06:46 2024 +0800" }, "message": "btrfs: test an incremental send scenario with cloning of unaligned extent\n\nTest that doing an incremental send with a file that had its size\ndecreased and became the destination for a clone operation of an extent\nwith an unaligned end offset that matches the new file size, works\ncorrectly.\n\nThis tests a bug fixed by the following kernel patch:\n\n \"btrfs: send: fix invalid clone operation for file that got its size decreased\"\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "34ce77ce867d27ad7ad9a180489e4311c3cd15e4", "tree": "1b30eca2616199c42d392214fd68e47f6f6e771f", "parents": [ "10e6c6772b3d21a233dda763674c297337e88294" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Thu Sep 26 10:41:46 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 29 21:06:46 2024 +0800" }, "message": "fsx: support unshare range fallocate mode\n\nThe fallocate unshare mode flag modifies traditional preallocate\nmode to unshare any shared extents backing the target range. Without\nthe unshare flag, preallocate mode simply assures that blocks are\nphysically allocated, regardless of whether they might be shared.\nUnshare mode behaves the same as preallocate mode outside of the\nshared extent case.\n\nSince unshare is fundamentally a modifier to preallocate mode,\nenable it via an operation flag. Similar to keep size mode, select\nit randomly for fallocate operations and track it via a flag and\nstring combination for operation logging and replay.\n\nUnshare is mainly used for filesystems that support reflink, but the\noperation is equivalent to preallocate mode for non-shared ranges,\nso enable it by default. Filesystems that do not support the\nfallocate flag (such as those that might not support reflink) will\nfail the test operation and disable unshare calls at runtime. Also\nprovide a new command line option to explicitly disable unshare\ncalls.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "10e6c6772b3d21a233dda763674c297337e88294", "tree": "ade7da9d0be7c47cb977c86a9d00cbfd40e9fb65", "parents": [ "c3140a6b87943a2e32c9c77ed4fe1fd2894c779b" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Thu Sep 12 13:45:00 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Sep 20 22:38:58 2024 +0800" }, "message": "btrfs/321: make the test work when compression is enabled\n\nWhen running btrfs/321 with compression enabled it fails like this:\n\n $ MOUNT_OPTIONS\u003d\"-o compress\" ./check btrfs/321\n FSTYP -- btrfs\n PLATFORM -- Linux/x86_64 debian0 6.11.0-rc7-btrfs-next-174+ #1 SMP PREEMPT_DYNAMIC Tue Sep 10 17:11:38 WEST 2024\n MKFS_OPTIONS -- /dev/sdc\n MOUNT_OPTIONS -- -o compress /dev/sdc /home/fdmanana/btrfs-tests/scratch_1\n\n btrfs/321 2s ... [failed, exit status 1]- output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/321.out.bad)\n --- tests/btrfs/321.out\t2024-09-12 12:12:11.259272125 +0100\n +++ /home/fdmanana/git/hub/xfstests/results//btrfs/321.out.bad\t2024-09-12 13:18:40.231120012 +0100\n @@ -1,2 +1,5 @@\n QA output created by 321\n -Silence is golden\n +mount: /home/fdmanana/btrfs-tests/scratch_1: can\u0027t read superblock on /dev/sdc.\n + dmesg(1) may have more information after failed mount system call.\n +mount -o compress -o ro /dev/sdc /home/fdmanana/btrfs-tests/scratch_1 failed\n +(see /home/fdmanana/git/hub/xfstests/results//btrfs/321.full for details)\n ...\n (Run \u0027diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/321.out /home/fdmanana/git/hub/xfstests/results//btrfs/321.out.bad\u0027 to see the entire diff)\n\n HINT: You _MAY_ be missing kernel fix:\n 10d9d8c3512f btrfs: fix a use-after-free bug when hitting errors inside btrfs_submit_chunk()\n\n Ran: btrfs/321\n Failures: btrfs/321\n Failed 1 of 1 tests\n\nThis is because with compression enabled we get a csum tree that has only\none leaf, and that leaf is the root of the csum tree. That means that\nafter the test corrupts the leaf, the next mount will fail since an error\nloading the root is critical and makes the mount operation fail.\n\nFix this by creating a file with 128M of data instead of 32M, as this\nguarantees that even if compression is enabled, and even with the maximum\nallowed leaf size (64K), we still get a csum tree with multiple leaves,\nmaking the test work.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c3140a6b87943a2e32c9c77ed4fe1fd2894c779b", "tree": "491aa7af73eb286de477a8355ab18d0d0e61891a", "parents": [ "185d797c73e1d9313d70e362ea7fcd2d1daa7832" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Wed Sep 18 13:57:19 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Sep 20 22:38:55 2024 +0800" }, "message": "generic: add a regression test for sub-block fsmap queries\n\nZizhi Wo found some bugs in the GETFSMAP implementation if it is fed\nsub-fsblock ranges. Add a regression test for this.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "185d797c73e1d9313d70e362ea7fcd2d1daa7832", "tree": "1764eda1d5c052f52799ccfe765126d4d6a822ad", "parents": [ "9b41da945b200087ca9bb0dc59bff62505ddcd74" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Mon Sep 16 13:03:12 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Sep 20 04:08:48 2024 +0800" }, "message": "fstests: fix min_dio_alignment logic for getting device block size\n\nIf we failed to get the dio alignment from statx we try to get the\ndevice\u0027s block size using the BLKSSZGET ioctl, however we failed to\nreturn it because we don\u0027t check if the ioctl succeeded (returned 0).\nFurthermore in case the ioctl returned an error, we end up returning an\nundefined value since the \u0027logical_block_size\u0027 variable ends up not\nbeing initialized.\n\nThis was causing some tests to be skipped on btrfs after commit\nee799a0cf1d4 (\"replace _min_dio_alignment with calls to\nsrc/min_dio_alignment\"), like generic/240 for example:\n\n $ ./check generic/240\n FSTYP -- btrfs\n PLATFORM -- Linux/x86_64 debian0 6.11.0-rc7-btrfs-next-174+ #1 SMP PREEMPT_DYNAMIC Tue Sep 10 17:11:38 WEST 2024\n MKFS_OPTIONS -- /dev/sdc\n MOUNT_OPTIONS -- /dev/sdc /home/fdmanana/btrfs-tests/scratch_1\n\n generic/240 1s ... [not run] fs block size must be larger than the device block size. fs block size: 4096, device block size: 4096\n Ran: generic/240\n Not run: generic/240\n Passed all 1 tests\n\nWhere before that commit the test ran.\n\nFix this by checking that the ioctl succeeded.\n\nFixes: 0e5f196d0a6a (\"add a new min_dio_alignment helper\")\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "9b41da945b200087ca9bb0dc59bff62505ddcd74", "tree": "ecfbe88b6e0c025e233d5a71d4747bb96517743b", "parents": [ "d6433d8c8402891ea2025b20ff6039a48de68693" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Thu Sep 12 12:26:38 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Sep 20 04:07:44 2024 +0800" }, "message": "fstests: add missing kernel git commit IDs to some tests\n\nThree tests (btrfs/321, generic/364 and xfs/608) refer to kernel patches\nthat are now in Linus\u0027 git kernel tree, so update the tests to include\nthe commit IDs.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "d6433d8c8402891ea2025b20ff6039a48de68693", "tree": "12a78cab99b844c91518b004efe435d174a3d3c4", "parents": [ "7ba74c6a3c6956715b00fce46e5547a488f65e89" ], "author": { "name": "Amir Goldstein", "email": "amir73il@gmail.com", "time": "Fri Aug 30 20:08:44 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Fri Sep 20 03:37:12 2024 +0800" }, "message": "overlay: create a variant to syncfs error test xfs/546\n\nTest overlayfs over xfs with and without \"volatile\" mount option.\n\nSigned-off-by: Amir Goldstein \u003camir73il@gmail.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "7ba74c6a3c6956715b00fce46e5547a488f65e89", "tree": "d37c61a75f16beaf3fb662b3375099bfd97a7194", "parents": [ "f71228e304b262162a5bec6d5529a89a6884d8c4" ], "author": { "name": "Srivathsa Dara", "email": "srivathsa.d.dara@oracle.com", "time": "Tue Sep 17 13:58:13 2024 +0000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Thu Sep 19 21:53:09 2024 +0800" }, "message": "ext4: Regression test for corruption during on-line resize\n\nRegression test for:\n\ta6b3bfe176e8 ext4: fix corruption during on-line resize\n\nSigned-off-by: Srivathsa Dara \u003csrivathsa.d.dara@oracle.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "f71228e304b262162a5bec6d5529a89a6884d8c4", "tree": "c281f5d0fa80fccf4f20c1e57b42f315b026bca5", "parents": [ "6171b03c92a67b5f4ff1ae730cf33e2c3a42866f" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Thu Sep 05 16:38:18 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:19:28 2024 +0800" }, "message": "btrfs/319: make the test work when compression is used\n\nCurrently btrfs/319 assumes there is no compression and that the files\nget a single extent (1 fiemap line) with a size of 1048581 bytes. However\nwhen testing with compression, for example by passing \"-o compress\" to\nMOUNT_OPTIONS environment variable, we get several extents and two lines\nof fiemap output, which makes the test fail since it hardcodes the fiemap\noutput:\n\n $ MOUNT_OPTIONS\u003d\"-o compress\" ./check btrfs/319\n FSTYP -- btrfs\n PLATFORM -- Linux/x86_64 debian0 6.11.0-rc6-btrfs-next-173+ #1 SMP PREEMPT_DYNAMIC Tue Sep 3 17:40:24 WEST 2024\n MKFS_OPTIONS -- /dev/sdc\n MOUNT_OPTIONS -- -o compress /dev/sdc /home/fdmanana/btrfs-tests/scratch_1\n\n btrfs/319 1s ... - output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/319.out.bad)\n --- tests/btrfs/319.out\t2024-08-12 14:16:55.653383284 +0100\n +++ /home/fdmanana/git/hub/xfstests/results//btrfs/319.out.bad\t2024-09-05 15:24:53.323076548 +0100\n @@ -6,11 +6,13 @@\n e61178ee0288ebe3fa36a3c975b02c94 SCRATCH_MNT/snap/foo\n e61178ee0288ebe3fa36a3c975b02c94 SCRATCH_MNT/snap/bar\n File bar fiemap in the original filesystem:\n -0: [0..2055]: shared|last\n +0: [0..2047]: shared\n +1: [2048..2055]: shared|last\n Creating a new filesystem to receive the send stream...\n ...\n (Run \u0027diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/319.out /home/fdmanana/git/hub/xfstests/results//btrfs/319.out.bad\u0027 to see the entire diff)\n\n HINT: You _MAY_ be missing kernel fix:\n 46a6e10a1ab1 btrfs: send: allow cloning non-aligned extent if it ends at i_size\n\n Ran: btrfs/319\n Failures: btrfs/319\n Failed 1 of 1 tests\n\nSo change the test to not rely on the fiemap output in its golden output\nand instead just check if all the extents reported by fiemap have the\nshared flag set (failing if there are any without the shared flag).\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "6171b03c92a67b5f4ff1ae730cf33e2c3a42866f", "tree": "054c6b4793822c62d8d52806389d29903960d934", "parents": [ "1e207662b3ce3ed712a827f4fedc580d0dc8c02f" ], "author": { "name": "Aleksa Sarai", "email": "cyphar@cyphar.com", "time": "Thu Sep 05 05:48:23 2024 +1000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:19:28 2024 +0800" }, "message": "generic/756: test name_to_handle_at(AT_HANDLE_MNT_ID_UNIQUE) explicitly\n\nIn order to make sure we are actually testing AT_HANDLE_MNT_ID_UNIQUE,\nadd a test (based on generic/426) which runs the open_by_handle in a\nmode where it will error out if there is a problem with getting mount\nIDs. The test is skipped if the kernel doesn\u0027t support the necessary\nfeatures.\n\nSuggested-by: Amir Goldstein \u003camir73il@gmail.com\u003e\nReviewed-by: Amir Goldstein \u003camir73il@gmail.com\u003e\nSigned-off-by: Aleksa Sarai \u003ccyphar@cyphar.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "1e207662b3ce3ed712a827f4fedc580d0dc8c02f", "tree": "de9c141d62530bed0a0808e49022836f21598c71", "parents": [ "0afb586a7e7790b2ee3dbc4f4b4742c6f9305b48" ], "author": { "name": "Aleksa Sarai", "email": "cyphar@cyphar.com", "time": "Thu Sep 05 05:48:22 2024 +1000" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:19:28 2024 +0800" }, "message": "open_by_handle: verify u32 and u64 mount IDs\n\nNow that open_by_handle_at(2) can return u64 mount IDs, do some tests to\nmake sure they match properly as part of the regular open_by_handle\ntests. Also, add automatic tests for the old u32 mount IDs as well.\n\nBy default, we do mount ID checks but silently skip the tests if the\nsyscalls are not supported by the running kernel (to ensure the tests\ncontinue to work for old kernels). We will add some tests explicitly\nchecking the new features (with no silent skipping) in a future patch.\n\nThe u32 mount ID tests require STATX_MNT_ID (Linux 5.8), while the u64\nmount ID tests require STATX_MNT_ID_UNIQUE (Linux 6.9) and\nAT_HANDLE_MNT_ID_UNIQUE (linux-next).\n\nLink: https://lore.kernel.org/all/20240828-exportfs-u64-mount-id-v3-0-10c2c4c16708@cyphar.com/\nReviewed-by: Amir Goldstein \u003camir73il@gmail.com\u003e\nSigned-off-by: Aleksa Sarai \u003ccyphar@cyphar.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "0afb586a7e7790b2ee3dbc4f4b4742c6f9305b48", "tree": "f005430a840278fee884fba2b2359fd3e882df85", "parents": [ "136c5099d07eae5cd02ae9bfc992e44f3b21ffe7" ], "author": { "name": "Chen Hanxiao", "email": "chenhx.fnst@fujitsu.com", "time": "Tue Sep 03 10:18:09 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:19:28 2024 +0800" }, "message": "generic/362: skip test on NFS mount\n\nxfstests complains:\n\nFSTYP -- nfs\nPLATFORM -- Linux/x86_64 r95b-1 5.14.0-496.el9.x86_64 #1 SMP PREEMPT_DYNAMIC Mon Aug 12 18:50:44 EDT 2024\nMKFS_OPTIONS -- 192.168.122.42:/nfsscratch\nMOUNT_OPTIONS -- -o vers\u003d4.2 192.168.122.42:/nfsscratch /mnt/scratch\n\ngeneric/362 QA output created by 362\nFailed to open/create file: Invalid argument\nSilence is golden\n- output mismatch (see /var/lib/xfstests/results//generic/362.out.bad)\n --- tests/generic/362.out 2024-09-02 14:27:09.162636093 -0400\n +++ /var/lib/xfstests/results//generic/362.out.bad 2024-09-02 14:33:36.167636093 -0400\n @@ -1,2 +1,3 @@\n QA output created by 362\n +Failed to open/create file: Invalid argument\n Silence is golden\n ...\n (Run \u0027diff -u /var/lib/xfstests/tests/generic/362.out /var/lib/xfstests/results//generic/362.out.bad\u0027 to see the entire diff)\nRan: generic/362\nFailures: generic/362\nFailed 1 of 1 tests\n\nNFS commit 9597c13b forbade open with O_APPEND|O_DIRECT\n\nstrace show that dio-append-buf-fault use (O_APPEND|O_DIRECT):\n\n mount -o vers\u003d4.2 192.168.122.42:/nfstest /mnt/scratch/\n strace ./src/dio-append-buf-fault /mnt/scratch/111\n..\n openat(AT_FDCWD, \"/mnt/scratch/111\", O_WRONLY|O_CREAT|O_TRUNC|O_APPEND|O_DIRECT, 0666) \u003d 3\n\nSo skip generic/362 on NFS\n\nSigned-off-by: Chen Hanxiao \u003cchenhx.fnst@fujitsu.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "136c5099d07eae5cd02ae9bfc992e44f3b21ffe7", "tree": "83b119b9fc487a4e91b71d5c5a7983f601528818", "parents": [ "87f14ee1f79f5f222472d86ba0f7e800944ce221" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Fri Aug 30 00:10:21 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:19:27 2024 +0800" }, "message": "generic: test concurrent direct IO writes and fsync using same fd\n\nTest that a program that has 2 threads using the same file descriptor and\nconcurrently doing direct IO writes and fsync doesn\u0027t trigger any crash\nor deadlock.\n\nThis is motivated by a bug found in btrfs fixed by the following patch:\n\n \"btrfs: fix race between direct IO write and fsync when using same fd\"\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "87f14ee1f79f5f222472d86ba0f7e800944ce221", "tree": "b07f0677ad4c2b6548ec12c59dd5d75af1fd1a91", "parents": [ "c9b3c1c392fd1a7e752920f539d936809ff2faa5" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:47:26 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sun Sep 08 13:17:19 2024 +0800" }, "message": "xfs: functional testing for filesystem properties\n\nMake sure that fs property storage and retrieval actually work.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c9b3c1c392fd1a7e752920f539d936809ff2faa5", "tree": "ea43bd3de931401d74b9ff11196fc91a4dde3e04", "parents": [ "fff4042b60bec943274258a7cbce1917ae512eec" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:47:10 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:50:44 2024 +0800" }, "message": "common/xfs: FITRIM now supports realtime volumes\n\nXFS now supports FITRIM to the realtime volume. Detect this support and\nenable it.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "fff4042b60bec943274258a7cbce1917ae512eec", "tree": "dd22a7d5b6b9f68667165868dc0d256a049e03b6", "parents": [ "18c331d98c29c3c09927eb4a2c2be53655c7e478" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:46:54 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:50:28 2024 +0800" }, "message": "xfs: refactor statfs field extraction\n\nPrepare for the next patch by refactoring the open-coded bits that call\nstatfs on a mounted xfs filesystem to extract a status field.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "18c331d98c29c3c09927eb4a2c2be53655c7e478", "tree": "603b2c610101f184aebbf4d0af9725b2b01c5e2a", "parents": [ "652692a8e4754a6996d8252b9a73bea1cb39511b" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:46:39 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:50:17 2024 +0800" }, "message": "xfs/004: fix column extraction code\n\nNow that the xfs_db freesp command prints a CDF of the free space\nhistograms, fix the pct column extraction code to handle the two\nnew columns by \u003ccough\u003e using awk.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "652692a8e4754a6996d8252b9a73bea1cb39511b", "tree": "7cb79de624276144d0f1328a30b9826afdb35784", "parents": [ "c63a6fe023531f7fb246c363d9e39e0ca1d6158a" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:46:23 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:50:00 2024 +0800" }, "message": "xfs: test xfs_scrub services\n\nCreate a pair of new tests that check that xfs_scrub and xfs_scrub_all\nwill find and test mounted filesystems.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c63a6fe023531f7fb246c363d9e39e0ca1d6158a", "tree": "f433bd70c5c54ae04ff56dac91935130a96cddd4", "parents": [ "1dbd10d1bbca6d67fca02754ee8ffab531892763" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:46:07 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:45:49 2024 +0800" }, "message": "generic/453: check xfs_scrub detection of confusing job offers\n\nEarlier this year, ESET revealed that Linux users had been tricked into\nopening executables containing malware payloads. The trickery came in\nthe form of a malicious zip file containing a filename with the string\n\"job offer․pdf\". Note that the filename does *not* denote a real pdf\nfile, since the last four codepoints in the file name are \"ONE DOT\nLEADER\", p, d, and f. Not period (ok, FULL STOP), p, d, f like you\u0027d\nnormally expect.\n\nNow that xfs_scrub can look for codepoints that could be confused with a\nperiod followed by alphanumerics, let\u0027s make sure it actually works.\n\nLink: https://www.welivesecurity.com/2023/04/20/linux-malware-strengthens-links-lazarus-3cx-supply-chain-attack/\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "1dbd10d1bbca6d67fca02754ee8ffab531892763", "tree": "69398d11d1f478eb7e0d6e781ecf65f25011260f", "parents": [ "a8d61229a949ab5372bde61d43207c47f25afd06" ], "author": { "name": "Darrick J. Wong", "email": "djwong@kernel.org", "time": "Tue Aug 27 11:45:52 2024 -0700" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:45:04 2024 +0800" }, "message": "generic/453: test confusable name detection with 32-bit unicode codepoints\n\nTest the confusable name detection when there are 32-bit unicode\nsequences in use. In other words, emoji. Change the xfs_scrub test to\ndump the output to a file instead of passing huge echo commands around.\n\nSigned-off-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nReviewed-by: Christoph Hellwig \u003chch@lst.de\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "a8d61229a949ab5372bde61d43207c47f25afd06", "tree": "fc96a7d840fe0ab76cf0541fc71856504adaf1ea", "parents": [ "66a2179efbde1fc2639419edac23494a8e8f75e9" ], "author": { "name": "Filipe Manana", "email": "fdmanana@suse.com", "time": "Mon Aug 26 13:22:57 2024 +0100" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:33:54 2024 +0800" }, "message": "btrfs/319: add git commit ID\n\nThe kernel patch was merged into Linus\u0027 tree, so update the \u0027xxxxxxxxxxxx\u0027\nstub with the respective git commit ID.\n\nSigned-off-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nReviewed-by: Josef Bacik \u003cjosef@toxicpanda.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "66a2179efbde1fc2639419edac23494a8e8f75e9", "tree": "d8afe3810f0b9184c4d67c1eae13c97ab5109d2f", "parents": [ "869685f2f308a3b79cd97ad90923d2b595b916d4" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Tue Aug 27 09:43:54 2024 +0930" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:32:31 2024 +0800" }, "message": "fstests: btrfs: test reading data with a corrupted checksum tree leaf\n\n[BUG]\nThere is a bug report that, KASAN get triggered when:\n\n- A read bio needs to be split\n This can happen for profiles with stripes, including\n RAID0/RAID10/RAID5/RAID6.\n\n- An error happens before submitting the new split bio\n This includes:\n * chunk map lookup failure\n * data csum lookup failure\n\nThen during the error path of btrfs_submit_chunk(), the original bio is\nfully freed before submitted range has a chance to call its endio\nfunction, resulting a use-after-free bug.\n\n[NEW TEST CASE]\nIntroduce a new test case to verify the specific behavior by:\n\n- Create a btrfs with enough csum leaves with data RAID0 profile\n To bump the csum tree level, use the minimal nodesize possible (4K).\n Writing 32M data which needs at least 8 leaves for data checksum\n\n RAID0 profile ensures the data read bios will get split.\n\n- Find the last csum tree leave and corrupt it\n\n- Read the data many times until we trigger the bug or exit gracefully\n With an x86_64 VM with KASAN enabled, it can trigger the KASAN report in\n just 4 iterations (the default iteration number is 32).\n\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nReviewed-by: Anand Jain \u003canand.jain@oracle.com\u003e\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "869685f2f308a3b79cd97ad90923d2b595b916d4", "tree": "5e21f2e3ed43c531bca7aa86d3c4b13b32c64af8", "parents": [ "fea26871799f2326626c7a78ba5548853c4b1159" ], "author": { "name": "Qu Wenruo", "email": "wqu@suse.com", "time": "Tue Aug 27 08:17:08 2024 +0930" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:31:44 2024 +0800" }, "message": "fstests: btrfs/125: do not use raid5 for metadata\n\n[BUG]\nThere are several bug reports of btrfs/125 failure recently, either\ncausing balance failure (-EIO), or even kernel crash.\n\nThe balance failure looks like this:\n\n Mount normal and balance\n +ERROR: error during balancing \u0027/mnt/scratch\u0027: Input/output error\n +There may be more info in syslog - try dmesg | tail\n +md5sum: /mnt/scratch/tf2: Input/output error\n\nThe test case btrfs/125 is not reliable in the past, and has been\ndiscussed several times:\n\nhttps://lore.kernel.org/linux-btrfs/CAL3q7H4oa70DUhOFE7kot62KjxcbvvZKxu62VfLpAcmgsinBFw@mail.gmail.com/\nhttps://lore.kernel.org/linux-btrfs/53f7bace2ac75d88ace42dd811d48b7912647301.1654672140.git.wqu@suse.com/#t\n\n[CAUSE]\nThere are several different factors involved.\n\n1. RMW mix the old and new metadata, causing unrepairable corruption\n E.g. with the following layout:\n\n data 1 |\u003c- Stale metadata -\u003e| (from the out-of-date device)\n data 2 | Unused |\n parity |PPPPPPPPPPPPPPPPPPPP|\n\n In above case, although metadata on data 1 is out-of-date, we can\n still rebuild the correct data from parity and data 2.\n\n But if we have new metadata writes into the data 2 stripe, an RMW\n will screw up the whole situation:\n\n data 1 |\u003c- Stale metadata -\u003e| (from the out-of-date device)\n data 2 |\u003c- New metadata -\u003e|\n parity |XXXXXXXXXXXXXXXXXXXX|\n\n The RMW will use the stale metadata and new metadata to calculate new\n parity.\n The resulted new parity will no longer be able to recover the old\n data 1.\n\n This is a known bug, thus our documentation is already recommending\n to avoid RAID56 for metadata usage.\n\n \u003e Metadata\n \u003e Do not use raid5 nor raid6 for metadata. Use raid1 or raid1c3\n \u003e respectively.\n\n And this is very hard to fix, unlike data we can fetch the\n data csum and verify during RMW, we can not do that during RMW.\n\n At the timing of RMW, we\u0027re holding the rbio lock for the full\n stripe.\n If the extent tree search requires a read-recover, it will generate\n another rbio, which may cover the same full stripe we\u0027re working on,\n leading to a deadlock.\n\n Furthermore the current RAID56 repair code is all based on veritical\n sectors, but metadata can cross several horizontal sectors.\n This will require multiple combinations to repair a metadata.\n\n2. Crash caused by double freeing a bio\n By chance if the above RMW corrupted csum tree, then during\n btrfs_submit_chunk() we will hit an error path that leads to double\n freeing of a bio, resulting crash or a KASAN report.\n\n Thankfully the patch fixing the use-after-free is already sent to the\n mailing list:\n https://lore.kernel.org/linux-btrfs/f4f916352ddf3f80048567ec7d8cc64cb388dc09.1724493430.git.wqu@suse.com/\n\n[WORKAROUND]\nSince it\u0027s very hard to fix the RAID56 metadata problem without a\ndeadlock or a huge code rework, for now just use RAID1 for the metadata\nof this particular test case.\n\nThere may be a chance to fix the situation by properly marking the\nmissing-then-reappear device as out-of-date, so no direct read from that\ndevice.\n\nBut that will also be a huge new feature, not something can be done\nimmediately.\n\nReviewed-by: Filipe Manana \u003cfdmanana@suse.com\u003e\nSigned-off-by: Qu Wenruo \u003cwqu@suse.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "fea26871799f2326626c7a78ba5548853c4b1159", "tree": "5bbf319ca97d2ef7c8f3b1fabc94f701eac2a120", "parents": [ "9e2b0f4c9579a74c940b4132d77ac6f59d3d8c4f" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Wed Aug 28 14:15:34 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:12:14 2024 +0800" }, "message": "generic: test to run fsx eof pollution\n\nFilesystem regressions related to partial page zeroing can go\nunnoticed for a decent amount of time. A recent example is the issue\nof iomap zero range not handling dirty pagecache over unwritten\nextents, which leads to wrong behavior on certain file extending\noperations (i.e. truncate, write extension, etc.).\n\nfsx does occasionally uncover these sorts of problems, but failures\ncan be rare and/or require longer running tests outside what is\ntypically run via full fstests regression runs. fsx now supports a\nmode that injects post-eof data in order to explicitly test partial\neof zeroing behavior. This uncovers certain problems more quickly\nand applies coverage more broadly across size changing operations.\n\nAdd a new test that runs an fsx instance (modeled after generic/127)\nwith eof pollution mode enabled. While the test is generic, it is\ncurrently limited to XFS as that is currently the only known major\nfs that does enough zeroing to satisfy the strict semantics expected\nby fsx. The long term goal is to uncover and fix issues so more\nfilesystems can enable this test.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "9e2b0f4c9579a74c940b4132d77ac6f59d3d8c4f", "tree": "4a4e8b29056a86360b69955e54a1d6db80146ca8", "parents": [ "11dbbdff173734b9f01c9ed46f9e52c008774fa7" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Wed Aug 28 14:15:33 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:12:14 2024 +0800" }, "message": "fsx: support eof page pollution for eof zeroing test coverage\n\nFile ranges that are newly exposed via size changing operations are\nexpected to return zeroes until written to. This behavior tends to\nbe difficult to regression test as failures can be racy and\ntransient. fsx is probably the best tool for this type of test\ncoverage, but uncovering issues can require running for a\nsignificantly longer period of time than is typically invoked\nthrough fstests tests. As a result, these types of regressions tend\nto go unnoticed for an unfortunate amount of time.\n\nTo facilitate uncovering these problems more quickly, implement an\neof pollution mode in fsx that opportunistically injects post-eof\ndata prior to operations that change file size. Since data injection\noccurs immediately before the size changing operation, it can be\nused to detect problems in partial eof page/block zeroing associated\nwith each relevant operation.\n\nThe implementation takes advantage of the fact that mapped writes\ncan place data beyond eof so long as the page starts within eof. The\nmain reason for the isolated per-operation approach (vs. something\nlike allowing mapped writes to write beyond eof, for example) is to\naccommodate the fact that writeback zeroes post-eof data on the eof\npage. The current approach is therefore not necessarily guaranteed\nto detect all problems, but provides more generic and broad test\ncoverage than the alternative of testing explicit command sequences\nand doesn\u0027t require significant changes to how fsx works. If this\nproves useful long term, further enhancements can be considered that\nmight facilitate the presence of post-eof data across operations.\n\nEnable the feature with the -e command line option. It is disabled\nby default because zeroing behavior is inconsistent across\nfilesystems. This can also be revisited in the future if zeroing\nbehavior is refined for the major filesystems that rely on fstests\nfor regression testing.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "11dbbdff173734b9f01c9ed46f9e52c008774fa7", "tree": "de123b00563abbc0e4c12682036ce471bde9cd90", "parents": [ "bc0c52655c5a80fcfa4a9b44e6661e5aa14cb67f" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Wed Aug 28 14:15:32 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:12:14 2024 +0800" }, "message": "fsx: factor out a file size update helper\n\nIn preparation for support for eof page pollution, factor out a file\nsize update helper. This updates the internally tracked file size\nbased on the upcoming operation and zeroes the appropriate range in\nthe good buffer for extending operations.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "bc0c52655c5a80fcfa4a9b44e6661e5aa14cb67f", "tree": "b8b6b64a18b2f322de28c78df46586a3c8437f25", "parents": [ "c88734099723030956e8fd69587812ef34392195" ], "author": { "name": "Brian Foster", "email": "bfoster@redhat.com", "time": "Wed Aug 28 14:15:31 2024 -0400" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:12:14 2024 +0800" }, "message": "fsx: don\u0027t skip file size and buf updates on simulated ops\n\nfsx supports the ability to skip through a certain number of\noperations of a given command sequence before beginning full\noperation. The way this works is by tracking the operation count,\nsimulating minimal side effects of skipped operations in-memory, and\nthen finally writing out the in-memory state to the target file when\nfull operation begins.\n\nSeveral fallocate() related operations don\u0027t correctly track\nin-memory state when simulated, however. For example, consider an\nops file with the following two operations:\n\n zero_range 0x0 0x1000 0x0\n read 0x0 0x1000 0x0\n\n... and an fsx run like so:\n\n fsx -d -b 2 --replay-ops\u003d\u003copsfile\u003e \u003cfile\u003e\n\nThis simulates the zero_range operation, but fails to track the file\nextension that occurs as a side effect such that the subsequent read\ndoesn\u0027t occur as expected:\n\n Will begin at operation 2\n skipping zero size read\n\nThe read is skipped in this case because the file size is zero. The\nproper behavior, and what is consistent with other size changing\noperations, is to make the appropriate in-core changes before\nchecking whether an operation is simulated so the end result of\nthose changes can be reflected on-disk for eventual non-simulated\noperations. This results in expected behavior with the same ops file\nand test command:\n\n Will begin at operation 2\n 2 read 0x0 thru 0xfff (0x1000 bytes)\n\nUpdate zero, copy and clone range to do the file size and EOF change\nrelated zeroing before checking against the simulated ops count.\n\nSigned-off-by: Brian Foster \u003cbfoster@redhat.com\u003e\nReviewed-by: Darrick J. Wong \u003cdjwong@kernel.org\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "c88734099723030956e8fd69587812ef34392195", "tree": "72351ed21077d043f8acb2a407d85831886a6764", "parents": [ "55158bfc994632f1ef81e2e1c1838bc8fd9bd5af" ], "author": { "name": "Chao Yu", "email": "chao@kernel.org", "time": "Mon Sep 02 18:18:44 2024 +0800" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:06:13 2024 +0800" }, "message": "f2fs/003: add missing _fixed_by_kernel_commit line\n\nThe bug related to this regression testcase has been fixed by commit\nb40a2b003709 (\"f2fs: use meta inode for GC of atomic file\"), let\u0027s\nadd missing _fixed_by_kernel_commit line for this testcase.\n\nCc: Jaegeuk Kim \u003cjaegeuk@kernel.org\u003e\nCc: Daeho Jeong \u003cdaehojeong@google.com\u003e\nSigned-off-by: Chao Yu \u003cchao@kernel.org\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "55158bfc994632f1ef81e2e1c1838bc8fd9bd5af", "tree": "833cab5159879a11e59c59483a39899f18034b5a", "parents": [ "b14652801e2e0db73b66e217f243dde122a459c5" ], "author": { "name": "Amir Goldstein", "email": "amir73il@gmail.com", "time": "Fri Aug 30 21:45:46 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Tue Sep 03 04:02:55 2024 +0800" }, "message": "overlay: deprecate test t_truncate_self\n\nSince kernel commit 2a010c412853 (\"fs: don\u0027t block i_writecount during\nexec\"), truncating an executable file while it is being executed is\nallowed. Therefore, the test t_truncate_self now fails, so remove it.\n\nSigned-off-by: Amir Goldstein \u003camir73il@gmail.com\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" }, { "commit": "b14652801e2e0db73b66e217f243dde122a459c5", "tree": "6706e59016ce21b60687de54091e450df422b6f3", "parents": [ "d77a998175749f189385a2e71d05ef3bafb3fb49" ], "author": { "name": "Christoph Hellwig", "email": "hch@lst.de", "time": "Wed Aug 21 06:47:43 2024 +0200" }, "committer": { "name": "Zorro Lang", "email": "zlang@kernel.org", "time": "Sat Aug 24 04:32:02 2024 +0800" }, "message": "generic/615: add to the quick group\n\nThis is a simple and very quick test, add it to the quick group.\n\nSigned-off-by: Christoph Hellwig \u003chch@lst.de\u003e\nReviewed-by: Zorro Lang \u003czlang@redhat.com\u003e\nSigned-off-by: Zorro Lang \u003czlang@kernel.org\u003e\n" } ], "next": "d77a998175749f189385a2e71d05ef3bafb3fb49" }