Re: [GIT PULL] bcachefs fixes for 6.14-rc2

[Akira Yokosawa <akiyks@gmail.com>]

Hello Kent,

Kent Overstreet wrote:
> Nothing major, things continue to be fairly quiet over here.
> 
> Tracking some bugs that show up on nix build servers; would like someone
> to confirm or deny that these are or are not still happening on 6.14,
> and if you've got an environment where this is reproducing and can work
> with me to debug that would save me some trouble...
> 
> Also, another bug report is implicating hibernate, please let me know if
> you have any data on that, still trying to piece together what's going
> on there.
> 
> And as usual, be noisy if you're seeing a serious bug that's not getting
> resolved.
> 

I'm not Linus, and not a serious bug, but let me be noisy.

See below.

> The following changes since commit 2014c95afecee3e76ca4a56956a936e23283f05b:
> 
>   Linux 6.14-rc1 (2025-02-02 15:39:26 -0800)
> 
> are available in the Git repository at:
> 
>   git://evilpiepirate.org/bcachefs.git tags/bcachefs-2025-02-06
> 
> for you to fetch changes up to 44a7bfed6f352b4bae4fd244d0fcd32aa25d0deb:
> 
>   bcachefs: bch2_bkey_sectors_need_rebalance() now only depends on bch_extent_rebalance (2025-02-05 19:56:24 -0500)
> 
> ----------------------------------------------------------------
> bcachefs fixes for 6.14-rc2
> 
> - add a SubmittingPatches to clarify that patches submitted for bcachefs
>   do, in fact, need to be tested
> - discard path now correctly issues journal flushes when needed, this
>   fixes performance issues when the filesystem is nearly full and we're
>   bottlenecked on copygc
> - fix a bug that could cause the pending rebalance work accounting to be
>   off when devices are being onlined/offlined; users should report if
>   they are still seeing this
> - and a few more trivial ones
> 
> ----------------------------------------------------------------
> Jeongjun Park (2):
>       bcachefs: fix incorrect pointer check in __bch2_subvolume_delete()
>       bcachefs: fix deadlock in journal_entry_open()
> 
> Kent Overstreet (4):
>       bcachefs docs: SubmittingPatches.rst

This caused a new warning in "make htmldocs", which was reported at:

    https://lore.kernel.org/20250204141216.4a2635ee@canb.auug.org.au/

, and has not been taken care of.

I'd like you not to ignore such a friendly report!

Regards, Akira

>       bcachefs: Fix discard path journal flushing
>       bcachefs: Fix rcu imbalance in bch2_fs_btree_key_cache_exit()
>       bcachefs: bch2_bkey_sectors_need_rebalance() now only depends on bch_extent_rebalance
> 
>  .../filesystems/bcachefs/SubmittingPatches.rst     | 98 ++++++++++++++++++++++
>  MAINTAINERS                                        |  1 +
>  fs/bcachefs/alloc_background.c                     | 47 ++++++-----
>  fs/bcachefs/alloc_foreground.c                     | 10 ++-
>  fs/bcachefs/alloc_types.h                          |  1 +
>  fs/bcachefs/btree_key_cache.c                      |  1 -
>  fs/bcachefs/buckets_waiting_for_journal.c          | 12 ++-
>  fs/bcachefs/buckets_waiting_for_journal.h          |  4 +-
>  fs/bcachefs/inode.h                                |  4 +-
>  fs/bcachefs/journal.c                              | 18 +++-
>  fs/bcachefs/journal.h                              |  1 +
>  fs/bcachefs/journal_types.h                        |  1 +
>  fs/bcachefs/opts.h                                 | 14 ----
>  fs/bcachefs/rebalance.c                            |  8 +-
>  fs/bcachefs/rebalance.h                            | 20 +++++
>  fs/bcachefs/subvolume.c                            |  7 +-
>  fs/bcachefs/super.c                                | 11 +++
>  fs/bcachefs/super.h                                |  1 +
>  fs/bcachefs/trace.h                                | 14 +++-
>  19 files changed, 214 insertions(+), 59 deletions(-)
>  create mode 100644 Documentation/filesystems/bcachefs/SubmittingPatches.rst