Re: [PATCH] virtiofsd: Show submounts

[Max Reitz <mreitz@redhat.com>]

[-- Attachment #1.1.1: Type: text/plain, Size: 3312 bytes --]

On 30.04.20 11:21, Max Reitz wrote:
> On 30.04.20 10:58, Dr. David Alan Gilbert wrote:
>> * Max Reitz (mreitz@redhat.com) wrote:
>>> On 29.04.20 16:57, Dr. David Alan Gilbert wrote:
>>>> * Dr. David Alan Gilbert (dgilbert@redhat.com) wrote:
>>>>> * Max Reitz (mreitz@redhat.com) wrote:
>>>>>> Currently, setup_mounts() bind-mounts the shared directory without
>>>>>> MS_REC.  This makes all submounts disappear.
>>>>>>
>>>>>> Pass MS_REC so that the guest can see submounts again.
>>>>>
>>>>> Thanks!
>>>>>
>>>>>> Fixes: 3ca8a2b1c83eb185c232a4e87abbb65495263756
>>>>>
>>>>> Should this actually be 5baa3b8e95064c2434bd9e2f312edd5e9ae275dc ?
>>>>>
>>>>>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>>>>>> ---
>>>>>>  tools/virtiofsd/passthrough_ll.c | 2 +-
>>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/tools/virtiofsd/passthrough_ll.c b/tools/virtiofsd/passthrough_ll.c
>>>>>> index 4c35c95b25..9d7f863e66 100644
>>>>>> --- a/tools/virtiofsd/passthrough_ll.c
>>>>>> +++ b/tools/virtiofsd/passthrough_ll.c
>>>>>> @@ -2643,7 +2643,7 @@ static void setup_mounts(const char *source)
>>>>>>      int oldroot;
>>>>>>      int newroot;
>>>>>>  
>>>>>> -    if (mount(source, source, NULL, MS_BIND, NULL) < 0) {
>>>>>> +    if (mount(source, source, NULL, MS_BIND | MS_REC, NULL) < 0) {
>>>>>>          fuse_log(FUSE_LOG_ERR, "mount(%s, %s, MS_BIND): %m\n", source, source);
>>>>>>          exit(1);
>>>>>>      }
>>>>>
>>>>> Do we want MS_SLAVE to pick up future mounts that might happenf rom the
>>>>> host?
>>>>> What's the interaction between this and the MS_REC|MS_SLAVE that we have
>>>>> a few lines above for / ?
>>>>
>>>> Just to confirm something from vgoyal, and what had confused me about
>>>> why we hadn't spotted this earlier.
>>>>
>>>> Even without this patch, the SLAVE stuff worked so if you start the
>>>> daemon and *then* mount under the shared directory, the guest sees it
>>>> with or without this patch.
>>>
>>> Hm, I don’t.  Do you really?
>>
>> Yes! With your patch reverted:
>>
>> Start virtiofsd, mount in the guest:
>>
>> host:
>> # ./virtiofsd --socket-path=/tmp/vhostqemu -o source=/home/dgilbert/virtio-fs/fs  -o log_level=warn -o no_writeback
>>
>> guest:
>> # mount -t virtiofs myfs /sysroot
> 
> OK, for some reason I didn’t try to mount in the guest first, but did
> the host mount after starting virtiofsd.
> 
> My mount test uses multiple mounts, and two of them I see mounted, but
> three I still don’t see mounted.

And if I do a find /sysroot (or find /mnt in my case) in the guest
before mounting anything on the host, the mounts don’t appear in the
guest at all, but that’s kind of a virtiofsd bug (or feature?).  That
is, if it has opened some directory before it became a mountpoint, then
it stays not-a-mountpoint.

> Let me see whether I can come up with something reproducible that isn’t
> a script.

Unfortunately, I can’t.  I’ve attached the script.

It launches qemu, then in the guest you have to mount “host” somewhere
(before the script emits “mounting”, ideally).

I can see that in the guest, mnt0-0 is mounted, and mnt0-0/sub is, too,
but mnt0-1 and mnt1 aren’t.

With this patch here, everything is mounted.

Max

[-- Attachment #1.1.2: test.sh --]
[-- Type: application/x-shellscript, Size: 2214 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]