Re: [PATCH v5 00/18] nfsd: refactor nfs4_create_file()

On Fri, 2026-07-17 at 19:27 +1000, NeilBrown wrote:
> For for another version so soon - I hadn't noticed Jeff's LLM results.
> Most of the issue raised there were already addressed I think, but the
> using  nfserr_noent instead of NFSERR_NOENT, and adding a condition on
> the call to mnt_want_write() in the final patch were still needed.
> 
> Thanks,
> NeilBrown
> 
> 
>  [PATCH v5 01/18] nfsd: honour client-provided attributes for
>  [PATCH v5 02/18] nfsd: move check_nfsd_access() call into
>  [PATCH v5 03/18] nfsd: correctly handle CREATE of mounted-on files
>  [PATCH v5 04/18] nfsd: replace fh_fill_both_attrs() with
>  [PATCH v5 05/18] nfsd: move fh_want_write() after preamble in
>  [PATCH v5 06/18] nfsd: move more nfs-specific code into preamble of
>  [PATCH v5 07/18] nfsd: remove subtlety from nfsd4_create_file()
>  [PATCH v5 08/18] nfsd: in nfsd4_create_file() let VFS report if file
>  [PATCH v5 09/18] nfsd: nfsd4_create_file(): Move NFSD_MAY_CREATE
>  [PATCH v5 10/18] nfsd: fh_want_write) failure need not be immediately
>  [PATCH v5 11/18] nfsd: (almost) always open file in
>  [PATCH v5 12/18] nfsd: reduce range of directory lock in
>  [PATCH v5 13/18] nfsd: open-code nfsd4_vfs_create() into
>  [PATCH v5 14/18] nfsd: move some code out of the
>  [PATCH v5 15/18] nfsd: reduce want-write range in nfsd4_create_file()
>  [PATCH v5 16/18] nfsd: move v0 checking out of nfsd_check_obj_isreg()
>  [PATCH v5 17/18] nfsd: separate out VFS-specific code from
>  [PATCH v5 18/18] nfsd: use do_lookup_open() for non-creating open

Nice work, Neil. You can add this to any that don't already have it:

Reviewed-by: Jeff Layton <[email protected]>