Re: [PATCH 1/1] nfs42: client needs to update file mode after ALLOCATE op

[Trond Myklebust <trondmy@hammerspace.com>]

On Thu, 2023-08-24 at 08:18 -0700, Dai Ngo wrote:
> The Linux NFS server strips the SUID and SGID from the file mode
> on ALLOCATE op. The GETATTR op in the ALLOCATE compound needs to
> request the file mode from the server to update its file mode in
> case the SUID/SGUI bit were stripped.
> 
> Signed-off-by: Dai Ngo <dai.ngo@oracle.com>
> ---
>  fs/nfs/nfs42proc.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/nfs/nfs42proc.c b/fs/nfs/nfs42proc.c
> index 63802d195556..ba2b83bfb37c 100644
> --- a/fs/nfs/nfs42proc.c
> +++ b/fs/nfs/nfs42proc.c
> @@ -57,6 +57,9 @@ static int _nfs42_proc_fallocate(struct rpc_message
> *msg, struct file *filep,
>                 .falloc_server  = server,
>         };
>         int status;
> +       struct super_block *sb = inode->i_sb;
> +       u64 fattr_supported = NFS_SB(sb)->fattr_valid;
> +       unsigned long mask = NFS_INO_INVALID_BLOCKS;
>  
>         msg->rpc_argp = &args;
>         msg->rpc_resp = &res;
> @@ -69,8 +72,10 @@ static int _nfs42_proc_fallocate(struct
> rpc_message *msg, struct file *filep,
>                 return status;
>         }
>  
> +       if (fattr_supported & NFS_ATTR_FATTR_MODE)
> +               mask |= NFS_INO_INVALID_MODE;

nfs4_bitmask_set() will take care of masking out bits that are not
supported, so the condition above is redundant.

>         nfs4_bitmask_set(bitmask, server->cache_consistency_bitmask,
> inode,
> -                        NFS_INO_INVALID_BLOCKS);
> +                               mask);
>  
>         res.falloc_fattr = nfs_alloc_fattr();
>         if (!res.falloc_fattr)

-- 
Trond Myklebust
Linux NFS client maintainer, Hammerspace
trond.myklebust@hammerspace.com