[CRIU] [PATCH] mount: close mnt.ns_fd only for sub-namespaces

[Pavel Emelyanov]

On 11/26/2015 04:31 PM, Andrew Vagin wrote:
> On Thu, Nov 26, 2015 at 04:07:46PM +0300, Pavel Emelyanov wrote:
>> On 11/26/2015 01:49 PM, Andrey Vagin wrote:
>>> From: Andrew Vagin <avagin at virtuozzo.com>
>>>
>>> nsid->mnt.ns_fd is initialized into 0, so currently
>>> fini_restore_mntns() closes the 0 descriptor if processes
>>> lives in a current mount namespace (NS_CRIU).
>>>
>>> Without this patch I get the following error:
>>> (00.166444)   4109: Inherit fd tty:[8800:d] -> 0 has been closed
>>>
>>> Signed-off-by: Andrew Vagin <avagin at virtuozzo.com>
>>> ---
>>>  mount.c | 2 +-
>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/mount.c b/mount.c
>>> index 20ae79b..d3628c2 100644
>>> --- a/mount.c
>>> +++ b/mount.c
>>> @@ -2653,7 +2653,7 @@ void fini_restore_mntns(void)
>>>  	for (nsid = ns_ids; nsid != NULL; nsid = nsid->next) {
>>>  		if (nsid->nd != &mnt_ns_desc)
>>>  			continue;
>>> -		if (root_item->ids->mnt_ns_id == nsid->id)
>>> +		if (nsid->type == NS_OTHER)
>>
>> But all the nsid's with type OTHER have ns_fd set to some descriptor,
>> why do you skip them?
> 
> I want to say != ;)

But != also looks weird. Look at prepare_mnt_ns() it does this:

       for (nsid = ns_ids; nsid != NULL; nsid = nsid->next) {
                char path[PATH_MAX];

                if (nsid->nd != &mnt_ns_desc)
                        continue;
                if (root_item->ids->mnt_ns_id == nsid->id)
                        continue;

		...

		nsid->mnt.ns_fd = open_proc(PROC_SELF, "ns/mnt");
                if (nsid->mnt.ns_fd < 0)
                        goto err;

There's no single check for nsid->type, why do we have one on
the closing part?