[CRIU] [PATCH 1/3] fsnotify: Add ability to force resolve paths for watchees

[Pavel Emelyanov]

On 10/30/2015 04:16 PM, Cyrill Gorcunov wrote:
> In case of migrating filesystem with nonpersistent inodes\devices
> we may hit the case where we successfully dump the container but
> on restore we will fail because block device get another id
> (while its contents might be fully sinc'ed). In this case we
> will try to lookup device from its number carried in image
> and fail.
> 
> Instead, when user know that he is migrating and prefer
> a flexible scheme -- we try to resolve file handle into
> a path and use it instead on the restore (just like
> for tmpfs/devtmpfs).
> 
> https://jira.sw.ru/browse/PSBM-40871
> 
> Reported-by: Nikita Spiridonov <nspiridonov at odin.com>
> Signed-off-by: Cyrill Gorcunov <gorcunov at openvz.org>
> ---
>  crtools.c            |  5 +++++
>  fsnotify.c           | 11 +++++++++--
>  include/cr_options.h |  1 +
>  3 files changed, 15 insertions(+), 2 deletions(-)
> 
> diff --git a/crtools.c b/crtools.c
> index d3812a18f4f0..243d90188ac8 100644
> --- a/crtools.c
> +++ b/crtools.c
> @@ -253,6 +253,7 @@ int main(int argc, char *argv[], char *envp[])
>  		{ "freeze-cgroup",		required_argument,	0, 1068 },
>  		{ "ghost-limit",		required_argument,	0, 1069 },
>  		{ "irmap-scan-path",		required_argument,	0, 1070 },
> +		{ "force-inotify-path",		required_argument,	0, 1071 },

One more option is excessive here. There are 3 big blocks that are in the
game here -- get inode handle, try to get the inode name, force FS scan to
get the inode name. And the 2nd block is must-have for TMPFS or (with this
patch) option. I propose to remove the option and make soft path lookup 
happen always, but only to fail the dump for TMPFS-s. The forced FS scan
should a) be performed iff the soft path resolve failed and b) still be
under option, which should be turned on by p.haul rsync driver.

-- Pavel

>  		{ },
>  	};
>  
> @@ -498,6 +499,9 @@ int main(int argc, char *argv[], char *envp[])
>  			if (irmap_scan_path_add(optarg))
>  				return -1;
>  			break;
> +		case 1071:
> +			opts.force_inotify_path = true;
> +			break;
>  		case 'M':
>  			{
>  				char *aux;
> @@ -727,6 +731,7 @@ usage:
>  "  -l|--" OPT_FILE_LOCKS "       handle file locks, for safety, only used for container\n"
>  "  -L|--libdir           path to a plugin directory (by default " CR_PLUGIN_DEFAULT ")\n"
>  "  --force-irmap         force resolving names for inotify/fsnotify watches\n"
> +"  --force-inotify-path  force inotify/fsnotify watches to carry paths (for migration)\n"
>  "  --irmap-scan-path FILE\n"
>  "                        add a path the irmap hints to scan\n"
>  "  -M|--ext-mount-map KEY:VALUE\n"
> diff --git a/fsnotify.c b/fsnotify.c
> index 9629977de8bc..57077a974562 100644
> --- a/fsnotify.c
> +++ b/fsnotify.c
> @@ -241,9 +241,16 @@ int check_open_handle(unsigned int s_dev, unsigned long i_ino,
>  		/*
>  		 * Inode numbers are not restored for tmpfs content, but we can
>  		 * get file names, becasue tmpfs cache is not pruned.
> +		 *
> +		 * Moreover, when we're migrating to another filesystem
> +		 * previously say sync'ed its contents, the device may
> +		 * be changed so on restore we won't be able to lookup
> +		 * for same hardcoded @s_dev. So carry a path if user
> +		 * requested.
>  		 */
> -		if ((mi->fstype->code == FSTYPE__TMPFS) ||
> -				(mi->fstype->code == FSTYPE__DEVTMPFS)) {
> +		if ((mi->fstype->code == FSTYPE__TMPFS)		||
> +		    (mi->fstype->code == FSTYPE__DEVTMPFS)	||
> +		    opts.force_inotify_path) {
>  			path = alloc_openable(s_dev, i_ino, f_handle);
>  			if (IS_ERR_OR_NULL(path)) {
>  				pr_err("Can't find suitable path for handle (%d)\n",
> diff --git a/include/cr_options.h b/include/cr_options.h
> index eac7283a4821..f655ad09e195 100644
> --- a/include/cr_options.h
> +++ b/include/cr_options.h
> @@ -84,6 +84,7 @@ struct cr_options {
>  	bool			auto_dedup;
>  	unsigned int		cpu_cap;
>  	bool			force_irmap;
> +	bool			force_inotify_path;
>  	char			**exec_cmd;
>  	unsigned int		manage_cgroups;
>  	char			*new_global_cg_root;
>