[Cluster-devel] [PATCH 0/8] Get rid of gfs2_write_calc_reserv in gfs2_iomap_begin_write

[Steven Whitehouse <swhiteho@redhat.com>]

Hi,


On 28/08/18 22:47, Andreas Gruenbacher wrote:
> Following several cleanups, this is a first step towards being more
> accurate in the number of blocks we reserve in the journal.  I hope this
> part of the code is easier to understand now as well.
>
> Thanks,
> Andreas
Looks good to me. Lets make sure it gets lots of testing,

Steve.

> Andreas Gruenbacher (8):
>    gfs2: Don't depend on mp_aheight in clone_metapath
>    gfs2: Split gfs2_indirect_init
>    gfs2: Move empty inode check out of gfs2_unstuff_dinode
>    gfs2: Don't create unnecessary indirect blocks
>    gfs2: Improve gfs2_alloc_size for stuffed files
>    gfs2: Rename size to len in gfs2_alloc_size
>    gfs2: Move gfs2_alloc_size out of gfs2_iomap_get
>    gfs2: Get rid of gfs2_write_calc_reserv in gfs2_iomap_begin_write
>
>   fs/gfs2/bmap.c | 336 ++++++++++++++++++++++++++++++-------------------
>   1 file changed, 205 insertions(+), 131 deletions(-)
>