[LU-14432] update e2fsprogs to 1.46.2 Created: 14/Feb/21  Updated: 28/Mar/22  Resolved: 09/Jun/21

Status: Resolved
Project: Lustre
Component/s: None
Affects Version/s: None
Fix Version/s: Lustre 2.15.0

Type: Task Priority: Minor
Reporter: Dongyang Li Assignee: Dongyang Li
Resolution: Fixed Votes: 0
Labels: None

Issue Links:
Related
is related to LU-15607 update e2fsprogs to 1.46.5 Resolved
Rank (Obsolete): 9223372036854775807

 Description   

this tracks the changes needed to rebase e2fsprogs to 1.46.1



 Comments   
Comment by Gerrit Updater [ 14/Feb/21 ]

Li Dongyang (dongyangli@ddn.com) uploaded a new patch: https://review.whamcloud.com/41665
Subject: LU-14432 configure.ac: fix AC_PREREQ
Project: tools/e2fsprogs
Branch: master-lustre-test
Current Patch Set: 1
Commit: 361f67dabd88dcd2e59db929d17a80579c693b6b

Comment by Gerrit Updater [ 02/Mar/21 ]

Andreas Dilger (adilger@whamcloud.com) merged in patch https://review.whamcloud.com/41665/
Subject: LU-14432 configure.ac: fix AC_PREREQ
Project: tools/e2fsprogs
Branch: master-lustre-test
Current Patch Set:
Commit: 93498f3b052680ef5d6e2ce9f53c7596cf596268

Comment by Andreas Dilger [ 09/Apr/21 ]

I was looking through the patches, and a few notes on what could be cleaned up if the patches need to be rebased again:

  • bugfixes like "e2fsck: fix memory leaks with pfsck enabled" and "e2fsck: misc cleanups for pfsck" and "e2fsck: propagate number of threads" should be merged into the respective original patches instead of being added at the end
  • in the dirdata patch it probably makes sense to add a check to mke2fs/tune2fs/e2fsck to refuse checking/formatting filesystems with both dirdata and casefold in the same filesystem until that feature combination is properly handled.
  • in the dirdata patch, I wonder if there is something better we can name EXT4_DIR_NAME_LEN() and avoid changing all of the current uses of EXT2_DIR_REC_LEN() in the code? that could avoid some potential bugs in the future.
  • in the "debugfs: dump "fid" and "lma" xattrs on inode stat" patch, it would be useful to increase the maximum size of xattrs that are hex-dumped by the debugfs command "stat". That is currently limited to 40-byte xattrs, but increasing this to 80 or 120 bytes would be more useful for debugging
  • ideally, patch https://review.whamcloud.com/35203 "LU-11604 debugfs: add support for printing linkEA" could be updated and included in this release. There have been a number of cases where this would have been useful recently to include printing LinkEA.
  • the patch "LU-11915 deubugfs: add support for xattrs in external inodes" looks like it could be submitted upstream
  • patch "filefrag: support PFL and FLR file" should be merged into the earlier Lustre filefrag patch
  • the patch "LU-14432 configure.ac: fix AC_PREREQ" could be merged earlier into the series so that the earlier builds are not broken during the rest of the patches
  • the patches to add .spec support could be brought together in the patch series

In general I think the patches look good (though I haven't really looked at each one in detail except the dirdata patch).

Comment by Dongyang Li [ 21/Apr/21 ]
  • bugfixes like "e2fsck: fix memory leaks with pfsck enabled" and "e2fsck: misc cleanups for pfsck" and "e2fsck: propagate number of threads" should be merged into the respective original patches instead of being added at the end

          We are now using the commits from upstream pfsck branch, maybe better to keep them like this.

  • in the dirdata patch it probably makes sense to add a check to mke2fs/tune2fs/e2fsck to refuse checking/formatting filesystems with both dirdata and casefold in the same filesystem until that feature combination is properly handled.

          Done

  • in the dirdata patch, I wonder if there is something better we can name EXT4_DIR_NAME_LEN() and avoid changing all of the current uses of EXT2_DIR_REC_LEN() in the code? that could avoid some potential bugs in the future.

          Unfortunately  EXT2_DIR_REC_LEN needs to look inside the dentry to determine rec len, it used to just use name_len and add it up, changing the param to dentry like now will make compiler tell us if new code trying to use EXT2_DIR_REC_LEN with name_len

  • in the "debugfs: dump "fid" and "lma" xattrs on inode stat" patch, it would be useful to increase the maximum size of xattrs that are hex-dumped by the debugfs command "stat". That is currently limited to 40-byte xattrs, but increasing this to 80 or 120 bytes would be more useful for debugging

          Done

  • ideally, patch https://review.whamcloud.com/35203 "LU-11604 debugfs: add support for printing linkEA" could be updated and included in this release. There have been a number of cases where this would have been useful recently to include printing LinkEA.

          Done updated patch at https://review.whamcloud.com/#/c/43383/ additional patch to print ACL https://review.whamcloud.com/#/c/43384/

  • the patch "LU-11915 deubugfs: add support for xattrs in external inodes" looks like it could be submitted upstream
  • patch "filefrag: support PFL and FLR file" should be merged into the earlier Lustre filefrag patch

          Done

  • the patch "LU-14432 configure.ac: fix AC_PREREQ" could be merged earlier into the series so that the earlier builds are not broken during the rest of the patches

          Done

  • the patches to add .spec support could be brought together in the patch series

          Done

Comment by Andreas Dilger [ 22/Apr/21 ]

Ok, now that EXA5.2.2 is released and e2fsprogs-1.45.6-wc7 is on the download site with the (hopefully) last nlinks fix, I think we can update to 1.46.2 if you are ready. What needs to be done for this?

Comment by Dongyang Li [ 22/Apr/21 ]

Let me refresh the patches for LU-11604 and after we land those to master-lustre-test we can tag 1.46.2.wc0, I think that's all.

Comment by Andreas Dilger [ 23/Apr/21 ]

I would prefer .wc1 for the first version, but I will leave it to you to manage this as I'm busy with several other areas right now.

Please let Minh know in advance of the change, so he can be aware if it is causing problems during testing and revert if necessary. That shouldn't happen, because we test with Lustre in advance, but we don't run all of the tests every time.

Comment by Dongyang Li [ 23/Apr/21 ]

my bad, it's already wc1 in the spec file . I will land those 2 patches from LU-11604 and tag 1.46.2.wc1. will give Minh a buzz before I change the branch over

Comment by Gerrit Updater [ 27/Apr/21 ]

Li Dongyang (dongyangli@ddn.com) uploaded a new patch: https://review.whamcloud.com/43469
Subject: LU-14432 misc: update e2fsprogs to 1.46.2.wc1
Project: fs/lustre-release
Branch: master
Current Patch Set: 1
Commit: c0d1ba54c06504049408ed7e319cb47362e7a64a

Comment by Gerrit Updater [ 31/May/21 ]

Andreas Dilger (adilger@whamcloud.com) uploaded a new patch: https://review.whamcloud.com/43881
Subject: LU-14432 ext2fs: fix ext2fs_get_avg_group() warning
Project: tools/e2fsprogs
Branch: master-lustre
Current Patch Set: 1
Commit: 2b15cc71186c9df7cbc584a78d20ac7b8ad3b32b

Comment by Gerrit Updater [ 01/Jun/21 ]

Andreas Dilger (adilger@whamcloud.com) merged in patch https://review.whamcloud.com/43881/
Subject: LU-14432 ext2fs: fix ext2fs_get_avg_group() warning
Project: tools/e2fsprogs
Branch: master-lustre
Current Patch Set:
Commit: 8495892a6a059be295210ea8e40006e1fa222429

Comment by Gerrit Updater [ 08/Jun/21 ]

Oleg Drokin (green@whamcloud.com) merged in patch https://review.whamcloud.com/43469/
Subject: LU-14432 misc: update e2fsprogs to 1.46.2.wc1
Project: fs/lustre-release
Branch: master
Current Patch Set:
Commit: 80f352fe90cad09cbdf7b61f74cc6ce4cd999bbf

Comment by Peter Jones [ 09/Jun/21 ]

Landed for 2.15

Generated at Sat Feb 10 03:09:41 UTC 2024 using Jira 9.4.14#940014-sha1:734e6822bbf0d45eff9af51f82432957f73aa32c.