[v3,0/6] ext4: Fix stale buffer loading from last failed

  Patch 1 fixes reusing stale buffer heads from last failed mounting.
Patch 2~4 reconstructs 'j_format_version' initialization and checking
in loading process.

v1->v2:
  Adopt suggestions from Tudor, add fix tag and corrupt 'stable' field
  in patch 1.
  Reserve empty lines in patch 4.
v2->v3:
  Split block device checking cleanup into a new patch (2th).
  Add 'Reviewed-by' tag in patch 3-6.


Zhang Yi (4):
  jbd2: remove unused feature macros
  jbd2: switch to check format version in superblock directly
  jbd2: factor out journal initialization from journal_get_superblock()
  jbd2: remove j_format_version

Zhihao Cheng (2):
  ext4: Fix reusing stale buffer heads from last failed mounting
  ext4: ext4_put_super: Remove redundant checking for
    'sbi->s_journal_bdev'

 fs/ext4/super.c      | 15 +++++++------
 fs/jbd2/journal.c    | 53 +++++++++++++++++---------------------------
 include/linux/jbd2.h | 33 ++++++++++++---------------
 3 files changed, 42 insertions(+), 59 deletions(-)
  

在 2023/3/15 9:31, Zhihao Cheng 写道:
> Patch 1 fixes reusing stale buffer heads from last failed mounting.
> Patch 2~4 reconstructs 'j_format_version' initialization and checking
> in loading process.
> 
> v1->v2:
>    Adopt suggestions from Tudor, add fix tag and corrupt 'stable' field
>    in patch 1.
>    Reserve empty lines in patch 4.
> v2->v3:
>    Split block device checking cleanup into a new patch (2th).
>    Add 'Reviewed-by' tag in patch 3-6.
> 
> 
> Zhang Yi (4):
>    jbd2: remove unused feature macros
>    jbd2: switch to check format version in superblock directly
>    jbd2: factor out journal initialization from journal_get_superblock()
>    jbd2: remove j_format_version
> 
> Zhihao Cheng (2):
>    ext4: Fix reusing stale buffer heads from last failed mounting
>    ext4: ext4_put_super: Remove redundant checking for
>      'sbi->s_journal_bdev'
> 
>   fs/ext4/super.c      | 15 +++++++------
>   fs/jbd2/journal.c    | 53 +++++++++++++++++---------------------------
>   include/linux/jbd2.h | 33 ++++++++++++---------------
>   3 files changed, 42 insertions(+), 59 deletions(-)
> 

Hi Ted, will this patchset be merged in next window?
  

On Fri, Jun 09, 2023 at 04:04:47PM +0800, Zhihao Cheng wrote:
> 
> Hi Ted, will this patchset be merged in next window?

It's currently in the dev branch.  I haven't set the ack out for it
yet because I'm still seeing some test failures, including some test
hangs in my regression tests.  There are a number of patches series
submission that I'm currently working through, so we'll have to see
what the "guilty" patch set might be, and whether there is an obvious
fix for it or not.  (I've found one such problem that was missed by
code review[1], and unfortunately, there is at least one more issue
that I'm currently trying to pin down.)

[1] https://lore.kernel.org/r/20230610190319.GB1436857@mit.edu

Cheers,

					- Ted
  

在 2023/6/11 12:42, Theodore Ts'o 写道:
> On Fri, Jun 09, 2023 at 04:04:47PM +0800, Zhihao Cheng wrote:
>>
>> Hi Ted, will this patchset be merged in next window?
> 
> It's currently in the dev branch.  I haven't set the ack out for it
> yet because I'm still seeing some test failures, including some test
> hangs in my regression tests.  There are a number of patches series
> submission that I'm currently working through, so we'll have to see
> what the "guilty" patch set might be, and whether there is an obvious
> fix for it or not.  (I've found one such problem that was missed by
> code review[1], and unfortunately, there is at least one more issue
> that I'm currently trying to pin down.)
> 
> [1] https://lore.kernel.org/r/20230610190319.GB1436857@mit.edu
> 
> Cheers,
> 
> 					- Ted
> 
> .
> 

Thanks for reminding. now I'm sure this patchset is tracked. Please 
follow your process.
  

On Wed, 15 Mar 2023 09:31:22 +0800, Zhihao Cheng wrote:
> Patch 1 fixes reusing stale buffer heads from last failed mounting.
> Patch 2~4 reconstructs 'j_format_version' initialization and checking
> in loading process.
> 
> v1->v2:
>   Adopt suggestions from Tudor, add fix tag and corrupt 'stable' field
>   in patch 1.
>   Reserve empty lines in patch 4.
> v2->v3:
>   Split block device checking cleanup into a new patch (2th).
>   Add 'Reviewed-by' tag in patch 3-6.
> 
> [...]

Applied, thanks!

[1/6] ext4: Fix reusing stale buffer heads from last failed mounting
      commit: ffea255f4052e6416a4b5738925337afbccd795a
[2/6] ext4: ext4_put_super: Remove redundant checking for 'sbi->s_journal_bdev'
      commit: a8f17d78525adf325c80f9dd1db469d337a5ce49
[3/6] jbd2: remove unused feature macros
      commit: 870a42846c1055c4ff9dfd492a0929c52a367d63
[4/6] jbd2: switch to check format version in superblock directly
      commit: 6014c2204f10b1199e15ab61aa30274a14999b1d
[5/6] jbd2: factor out journal initialization from journal_get_superblock()
      commit: 51bacdba23d85af2a9a145d97bfb77e6e85c98ad
[6/6] jbd2: remove j_format_version
      commit: 1f15ee267c0498016cc4aee2cdcc18e56ff42bae

Best regards,