Re: [PATCH 15/28] ovl: Open file with data except for the case of fsync

[Vivek Goyal]

On Tue, May 29, 2018 at 04:45:59PM +0200, Miklos Szeredi wrote:
> From: Vivek Goyal <vgoyal@xxxxxxxxxx>
> 
> ovl_open() should open file which contains data and not open metacopy
> inode.  With the introduction of metacopy inodes, with current
> implementaion we will end up opening metacopy inode as well.
> 
> But there can be certain circumstances like ovl_fsync() where we want to
> allow opening a metacopy inode instead.
> 
> Hence, change ovl_open_realfile() and add _ovl_open_real() and add extra
> parameter which specifies whether to allow opening metacopy inode or not.
> If this parameter is false, we look for data inode and open that.
> 
> This should allow covering both the cases.
> 
> Signed-off-by: Vivek Goyal <vgoyal@xxxxxxxxxx>
> Reviewed-by: Amir Goldstein <amir73il@xxxxxxxxx>
> Signed-off-by: Miklos Szeredi <mszeredi@xxxxxxxxxx>
> ---
>  fs/overlayfs/file.c | 39 ++++++++++++++++++++++++++++++---------
>  1 file changed, 30 insertions(+), 9 deletions(-)
> 
> diff --git a/fs/overlayfs/file.c b/fs/overlayfs/file.c
> index 266692ce9a9a..c7738ef492c8 100644
> --- a/fs/overlayfs/file.c
> +++ b/fs/overlayfs/file.c
> @@ -14,11 +14,20 @@
>  #include <linux/uio.h>
>  #include "overlayfs.h"
>  
> -static struct file *ovl_open_realfile(const struct file *file)
> +static char ovl_whatisit(struct inode *inode, struct inode *realinode)
> +{
> +	if (realinode != ovl_inode_upper(inode))
> +		return 'l';
> +	if (ovl_has_upperdata(inode))
> +		return 'u';
> +	else
> +		return 'm';
> +}
> +
> +static struct file *ovl_open_realfile(const struct file *file,
> +				      struct inode *realinode)
>  {
>  	struct inode *inode = file_inode(file);
> -	struct inode *upperinode = ovl_inode_upper(inode);
> -	struct inode *realinode = upperinode ?: ovl_inode_lower(inode);
>  	struct file *realfile;
>  	const struct cred *old_cred;
>  
> @@ -28,7 +37,7 @@ static struct file *ovl_open_realfile(const struct file *file)
>  	revert_creds(old_cred);
>  
>  	pr_debug("open(%p[%pD2/%c], 0%o) -> (%p, 0%o)\n",
> -		 file, file, upperinode ? 'u' : 'l', file->f_flags,
> +		 file, file, ovl_whatisit(inode, realinode), file->f_flags,
>  		 realfile, IS_ERR(realfile) ? 0 : realfile->f_flags);
>  
>  	return realfile;
> @@ -72,17 +81,24 @@ static int ovl_change_flags(struct file *file, unsigned int flags)
>  	return 0;
>  }
>  
> -static int ovl_real_fdget(const struct file *file, struct fd *real)
> +static int ovl_real_fdget_meta(const struct file *file, struct fd *real,
> +			       bool allow_meta)
>  {
>  	struct inode *inode = file_inode(file);
> +	struct inode *realinode;
>  
>  	real->flags = 0;
>  	real->file = file->private_data;
>  
> +	if (allow_meta)
> +		realinode = ovl_inode_real(inode);
> +	else
> +		realinode = ovl_inode_realdata(inode);
> +
>  	/* Has it been copied up since we'd opened it? */
> -	if (unlikely(file_inode(real->file) != ovl_inode_real(inode))) {
> +	if (unlikely(file_inode(real->file) != realinode)) {
>  		real->flags = FDPUT_FPUT;
> -		real->file = ovl_open_realfile(file);
> +		real->file = ovl_open_realfile(file, realinode);
>  
>  		return PTR_ERR_OR_ZERO(real->file);
>  	}
> @@ -94,6 +110,11 @@ static int ovl_real_fdget(const struct file *file, struct fd *real)
>  	return 0;
>  }
>  
> +static int ovl_real_fdget(const struct file *file, struct fd *real)
> +{
> +	return ovl_real_fdget_meta(file, real, false);
> +}
> +
>  static int ovl_open(struct inode *inode, struct file *file)
>  {
>  	struct dentry *dentry = file_dentry(file);
> @@ -107,7 +128,7 @@ static int ovl_open(struct inode *inode, struct file *file)
>  	/* No longer need these flags, so don't pass them on to underlying fs */
>  	file->f_flags &= ~(O_CREAT | O_EXCL | O_NOCTTY | O_TRUNC);
>  
> -	realfile = ovl_open_realfile(file);
> +	realfile = ovl_open_realfile(file, ovl_inode_real(file_inode(file)));

Hi Miklos,

Hmm...so there have been some changes in this patch. My original intention
was that to always open data inode (lower/upper) in ovl_open(). So if upper
inode is a metacopy only, I will open lower inode instead.

But new logic seems to be to always open real inode (that means upper
metacopy inode as well). And that means that later when read happens
on the file we will end up opening lower file, read data and close
lower file.

I am concerned a bit if there are performance implications to this.
This will be hot path for containers.

Thanks
Vivek