[CRIU] [PATCH 5/7] sk-unix: Don't fail if socket path lays on btrfs volume

[Pavel Emelyanov]

On 12/04/2013 05:23 PM, Cyrill Gorcunov wrote:
> 
> Because socket migh lay on btrfs volume (thus the device
> number reported by diag module won't be the same as obtained
> from stat(2)) we need to do an additional test and try
> to resolve device number with help of btrfs engine.
> 
> Signed-off-by: Cyrill Gorcunov <gorcunov at openvz.org>
> ---
>  sk-unix.c | 22 +++++++++++++++++++++-
>  1 file changed, 21 insertions(+), 1 deletion(-)
> 


> +			udiag_vfs_dev = kdev_to_odev(uv->udiag_vfs_dev);
> +			if (st.st_dev != udiag_vfs_dev) {
> +				if (btrfs_lazy_parse_subvol(udiag_vfs_dev))
> +					goto err;
> +				if (is_btrfs_subvol(udiag_vfs_dev, st.st_dev)) {
> +					pr_info("unix: liftup btrfs dev %lx -> %lx\n",
> +						(long)st.st_dev, (long)udiag_vfs_dev);
> +					st.st_dev = udiag_vfs_dev;
> +				}
> +			}

This is way too complex. Why no

- if (st.st_dev != dev)
+ if (phys_stat_dev_match(st.st_dev, dev)

in all the places including ghosts and inotifies, where

bool phys_stat_dev_match(dev_t st_dev, dev_t phys_dev)
{
	struct mount_info *mi;

	if (st_dev == phys_dev)
		return true;

	/*
	 * comment here about btrfs devs screwup
	 */

	mi = lookup_mnt_sdev(phys_dev);
	if (!mi)
		/* no mountpoint with this device */
		return false;

	if (strcmp(mi->kfstype, "btrfs"))
		/* not btrfs -- no match */
		return false;

	if (!mi->private)
		btrfs_lazy_parse_subvol(mi);

	return is_btrfs_subvol(mi->private, st_dev);
}