[PATCHSET,0/3] fstests: add inode btree blocks counters to the AGI header

Message

Darrick J. Wong March 23, 2021, 4:20 a.m. UTC

Hi all,

Years ago, Christoph diagnosed a problem where freeing an inode on a
totally full filesystem could fail due to finobt expansion not being
able to allocate enough blocks.  He solved the problem by using the
per-AG block reservation system to ensure that there are always enough
blocks for finobt expansion, but that came at the cost of having to walk
the entire finobt at mount time.  This new feature solves that
performance regression by adding inode btree block counts to the AGI
header.  The patches in this series amend fstests to handle the new
metadata fields and to test that upgrades work properly.

If you're going to start using this mess, you probably ought to just
pull from my git trees, which are linked below.

This is an extraordinary way to destroy everything.  Enjoy!
Comments and questions are, as always, welcome.

--D

kernel git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=inobt-counters

xfsprogs git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=inobt-counters

fstests git tree:
https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=inobt-counters
---
 common/xfs        |   20 ++++++
 tests/xfs/010     |    3 +
 tests/xfs/030     |    2 +
 tests/xfs/122.out |    2 -
 tests/xfs/764     |  190 +++++++++++++++++++++++++++++++++++++++++++++++++++++
 tests/xfs/764.out |   27 ++++++++
 tests/xfs/910     |   84 +++++++++++++++++++++++
 tests/xfs/910.out |   12 +++
 tests/xfs/group   |    2 +
 9 files changed, 340 insertions(+), 2 deletions(-)
 create mode 100755 tests/xfs/764
 create mode 100644 tests/xfs/764.out
 create mode 100755 tests/xfs/910
 create mode 100644 tests/xfs/910.out

Comments

Eryu Guan March 28, 2021, 3:14 p.m. UTC | #1

On Mon, Mar 22, 2021 at 09:20:44PM -0700, Darrick J. Wong wrote:
> Hi all,
> 
> Years ago, Christoph diagnosed a problem where freeing an inode on a
> totally full filesystem could fail due to finobt expansion not being
> able to allocate enough blocks.  He solved the problem by using the
> per-AG block reservation system to ensure that there are always enough
> blocks for finobt expansion, but that came at the cost of having to walk
> the entire finobt at mount time.  This new feature solves that
> performance regression by adding inode btree block counts to the AGI
> header.  The patches in this series amend fstests to handle the new
> metadata fields and to test that upgrades work properly.

I've applied the first two patches, the third one seems need to be
revised.

Thanks,
Eryu

> 
> If you're going to start using this mess, you probably ought to just
> pull from my git trees, which are linked below.
> 
> This is an extraordinary way to destroy everything.  Enjoy!
> Comments and questions are, as always, welcome.
> 
> --D
> 
> kernel git tree:
> https://git.kernel.org/cgit/linux/kernel/git/djwong/xfs-linux.git/log/?h=inobt-counters
> 
> xfsprogs git tree:
> https://git.kernel.org/cgit/linux/kernel/git/djwong/xfsprogs-dev.git/log/?h=inobt-counters
> 
> fstests git tree:
> https://git.kernel.org/cgit/linux/kernel/git/djwong/xfstests-dev.git/log/?h=inobt-counters
> ---
>  common/xfs        |   20 ++++++
>  tests/xfs/010     |    3 +
>  tests/xfs/030     |    2 +
>  tests/xfs/122.out |    2 -
>  tests/xfs/764     |  190 +++++++++++++++++++++++++++++++++++++++++++++++++++++
>  tests/xfs/764.out |   27 ++++++++
>  tests/xfs/910     |   84 +++++++++++++++++++++++
>  tests/xfs/910.out |   12 +++
>  tests/xfs/group   |    2 +
>  9 files changed, 340 insertions(+), 2 deletions(-)
>  create mode 100755 tests/xfs/764
>  create mode 100644 tests/xfs/764.out
>  create mode 100755 tests/xfs/910
>  create mode 100644 tests/xfs/910.out