mir/criu
2
0
Fork 0
mirror of https://github.com/checkpoint-restore/criu synced 2025-08-31 06:15:24 +00:00
Code Issues Releases Wiki Activity

mount: fix regression where open_mountpoint failed on readonly fs

Browse Source 
If we fail to create temporary directory for doing a clean mount we can
make mount clean reusing the code which enters new mountns to umount
overmounts. As when last process exits mntns all mounts are implicitly
cleaned from children, see in kernel source - sys_exit->do_exit
->exit_task_namespaces->switch_task_namespaces->free_nsproxy
->put_mnt_ns->umount_tree->drop_collected_mounts->umount_tree:

    /* Hide the mounts from mnt_mounts */
    list_for_each_entry(p, &tmp_list, mnt_list) {
            list_del_init(&p->mnt_child);
    }

Fixes commit b6cfb1ce29 ("mount: make open_mountpoint handle overmouts
properly")

https://github.com/checkpoint-restore/criu/issues/520
Signed-off-by: Pavel Tikhomirov <ptikhomirov@virtuozzo.com>
Acked-by: Adrian Reber <areber@redhat.com>
Signed-off-by: Andrei Vagin <avagin@virtuozzo.com>
This commit is contained in:
Pavel Tikhomirov
2018-07-11 13:43:33 +03:00
committed by Andrei Vagin
parent 382756a4f5
commit e54d0d26ba
1 changed files with 20 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

28
criu/mount.c
View File

@@ -1325,10 +1325,18 @@ int ns_open_mountpoint(void *arg)
if (umount_overmounts(mi))
goto err;
/* Save fd which we opened for parent due to CLONE_FILES flag */
*fd = get_clean_fd(mi);
if (*fd < 0)
/*
* Save fd which we opened for parent due to CLONE_FILES flag
*
* Mount can still have children in it, but we don't need to clean it
* explicitly as when last process exits mntns all mounts in it are
* cleaned from their children, and we are exactly the last process.
*/
*fd = open(mi->mountpoint, O_DIRECTORY|O_RDONLY);
if (*fd < 0) {
pr_perror("Unable to open %s", mi->mountpoint);
goto err;
}
return 0;
err:
@@ -1367,18 +1375,22 @@ int open_mountpoint(struct mount_info *pm)
if (!mnt_is_overmounted(pm)) {
pr_info("\tmount has children %s\n", pm->mountpoint);
fd = get_clean_fd(pm);
if (fd < 0)
goto err;
} else {
}
/*
* Mount is overmounted or probably we can't create a temporary
* direcotry for a cleaned mount
*/
if (fd < 0) {
int pid, status;
struct clone_arg ca = {
.mi = pm,
.fd = &fd
};
pr_info("\tmount is overmounted %s\n", pm->mountpoint);
pr_info("\tmount is overmounted or has children %s\n",
pm->mountpoint);
/*
* We are overmounted - not accessible in a regular way. We