Re: [PATCH] ceph: make sure directories aren't complete after setting crypt context

From: Luís Henriques
Date: Mon Nov 21 2022 - 08:51:41 EST


Xiubo Li <xiubli@xxxxxxxxxx> writes:

> On 18/11/2022 19:12, Luís Henriques wrote:
>> Xiubo Li <xiubli@xxxxxxxxxx> writes:
>>
>>> On 18/11/2022 02:02, Luís Henriques wrote:
>>>> Xiubo Li <xiubli@xxxxxxxxxx> writes:
>>>>
>>>>> On 17/11/2022 16:03, Xiubo Li wrote:
>>>>>> On 16/11/2022 23:37, Luís Henriques wrote:
>>>>>>> When setting a directory's crypt context, __ceph_dir_clear_complete() needs
>>>>>>> to be used otherwise, if it was complete before, any old dentry that's still
>>>>>>> around will be valid.
>>>>>>>
>>>>>>> Signed-off-by: Luís Henriques <lhenriques@xxxxxxx>
>>>>>>> ---
>>>>>>> Hi!
>>>>>>>
>>>>>>> Here's a simple way to trigger the bug this patch is fixing:
>>>>>>>
>>>>>>> # cd /cephfs
>>>>>>> # ls mydir
>>>>>>> nKRhofOAVNsAwVLvDw7a0c9ypsjbZfK3n0Npnmni6j0
>>>>>>> # ls mydir/nKRhofOAVNsAwVLvDw7a0c9ypsjbZfK3n0Npnmni6j0/
>>>>>>> Cyuer5xT+kBlEPgtwAqSj0WK2taEljP5vHZ,D8VXCJ8
>>>>>>> u+46b2XVCt7Obpz0gznZyNLRj79Q2l4KmkwbKOzdQKw
>>>>>>> # fscrypt unlock mydir
>>>>>>> # touch /mnt/test/mydir/mysubdir/file
>>>>>>> touch: cannot touch '/mnt/test/mydir/mysubdir/file': No such file or
>>>>>>> directory
>>>>>>>
>>>>>>>   fs/ceph/crypto.c | 4 ++++
>>>>>>>   1 file changed, 4 insertions(+)
>>>>>>>
>>>>>>> diff --git a/fs/ceph/crypto.c b/fs/ceph/crypto.c
>>>>>>> index 35a2ccfe6899..dc1557967032 100644
>>>>>>> --- a/fs/ceph/crypto.c
>>>>>>> +++ b/fs/ceph/crypto.c
>>>>>>> @@ -87,6 +87,10 @@ static int ceph_crypt_get_context(struct inode *inode,
>>>>>>> void *ctx, size_t len)
>>>>>>>           return -ERANGE;
>>>>>>>         memcpy(ctx, cfa->cfa_blob, ctxlen);
>>>>>>> +
>>>>>>> +    /* Directory isn't complete anymore */
>>>>>>> +    if (S_ISDIR(inode->i_mode) && __ceph_dir_is_complete(ci))
>>>>>>> +        __ceph_dir_clear_complete(ci);
>>>>>> Hi Luis,
>>>>>>
>>>>>> Good catch!
>>>>>>
>>>>>> BTW, why do this in the ceph_crypt_get_context() ? As my understanding is that
>>>>>> we should mark 'mydir' as incomplete when unlocking it. While as I remembered
>>>>>> the unlock operation will do:
>>>>>>
>>>>>>
>>>>>> Step1: get_encpolicy via 'mydir' as ctx
>>>>>> Step2: rm_enckey of ctx from the superblock
>>>>>>
>>>>> Sorry, it should be add_enckey.
>>>>>> Since I am still running the test cases for the file lock patches, so I didn't
>>>>>> catch logs to confirm the above steps yet.
>>>>>>
>>>>>> If I am right IMO then we should mark the dir as incomplete in the Step2
>>>>>> instead, because for non-unlock operations they may also do the Step1.
>>>>>>
>>>>> Your patch will work. But probably we should do this just around
>>>>> __fscrypt_prepare_readdir() or fscrypt_prepare_readdir() instead ? We need to
>>>>> detect that once the 'inode->i_crypt_info' changed then mark the dir as
>>>>> incomplete.
>>>>>
>>>>> For now for the lock operation it will evict the inode, which will help do this
>>>>> for us already. But for unlock case, we need to handle it by ourself.
>>>> OK, that makes sense and to be honest I thought that there should be
>>>> another place for doing this. Unfortunately, I didn't found it: in the
>>>> test case I have the fscrypt_prepare_readdir() isn't called:
>>>>
>>>> # cd /cephfs
>>>> # ls mydir
>>>> nKRhofOAVNsAwVLvDw7a0c9ypsjbZfK3n0Npnmni6j0
>>>> # ls mydir/nKRhofOAVNsAwVLvDw7a0c9ypsjbZfK3n0Npnmni6j0/
>>>> Cyuer5xT+kBlEPgtwAqSj0WK2taEljP5vHZ,D8VXCJ8 u+46b2XVCt7Obpz0gznZyNLRj79Q2l4KmkwbKOzdQKw
>>>>
>>>> At this point readdir was executed, of course. And
>>>> __ceph_dir_set_complete() is also used to indicate that we have the full
>>>> contents. However, executing the following commands won't result in any
>>>> new readdir():
>>>>
>>>> # fscrypt unlock mydir
>>>> # touch /mnt/test/mydir/mysubdir/file
>>>>
>>>> and since the encryption key is set at the sb level, I couldn't find a way
>>>> to detect changes in inode->i_crypt_info. ceph_d_revalidate() is invoked
>>>> but at that point I don't thing we have a way to know what is changing.
>>>>
>>>> Any ideas?
>>> # ls mydir/
>>> zy94Zt01M90xwYq+nxJsEvea+HYq49mqVgrUBkYrJAU
>>> # ls mydir/zy94Zt01M90xwYq+nxJsEvea+HYq49mqVgrUBkYrJAU/
>>> 24iNa8ICYc6nZZIkL,4n1sOtp9KiaTGtpq8PCZDQ6LU
>>> ivG4rQYzg6YENagbTZaV2kh2sVouEdlkmBZPfgVDl48
>>>
>>> I have added the debug logs in all the places just like:
>>>
>>> @@ -784,7 +786,9 @@ static struct dentry *ceph_lookup(struct inode *dir, struct
>>> dentry *dentry,
>>>                 return ERR_PTR(-ENAMETOOLONG);
>>>
>>>         if (IS_ENCRYPTED(dir)) {
>>> +       printk("%s fscrypt_has_encryption_key(dir): %d ======\n", __func__,
>>> fscrypt_has_encryption_key(dir));
>>>                 err = __fscrypt_prepare_readdir(dir);
>>> +       printk("%s fscrypt_has_encryption_key(dir): %d ======\n", __func__,
>>> fscrypt_has_encryption_key(dir));
>>>                 if (err)
>>>                         return ERR_PTR(err);
>>>                 if (!fscrypt_has_encryption_key(dir)) {
>>>
>>>
>>> # fscrypt unlock mydir
>>> Enter custom passphrase for protector "l":
>>> "mydir" is now unlocked and ready for use.
>>>
>>> And after "mydir/" was unlocked and then when doing:
>>>
>>> # touch mydir/dir/file2
>>>
>>> And when lookup the "dir/" dentry I can see the "mydir/" encryption key changed:
>>>
>>> 709 <7>[79125.023676] ceph:  __ceph_caps_issued_mask ino 0x10000000004 cap
>>> 000000004dc11892 issued pAsLsXsFs (mask As)
>>> 710 <7>[79125.023687] ceph:  __touch_cap 000000007071b095 cap 000000004dc11892
>>> mds0
>>> 711 <7>[79125.023823] ceph:  lookup 000000007071b095 dentry 00000000f97501b2
>>> 'dir'
>>> 712 <4>[79125.023838] ceph_lookup fscrypt_has_encryption_key(dir): 0 ======
>>> 713 <4>[79125.024186] ceph_lookup fscrypt_has_encryption_key(dir): 1 ======
>>> 714 <7>[79125.024194] ceph:   dir 000000007071b095 flags are 0x0
>>> 715 <7>[79125.024269] ceph:  do_request on 00000000a93fafef
>>>
>>> I am thinking could we just make __fscrypt_prepare_readdir(), which will return
>>> 0 when the key is already set or successfully set, to return 1 instead of 0 to
>>> mark that the key changed ?
>> OK, I see what you mean. Thanks. What about simply detect this change
>> here instead of changing __fscrypt_prepare_readdir() semantics? Because I
>> think that would require changes in several other places, including other
>> filesystems.
>>
>> What about something like this?
>>
>> diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
>> index edc2bf0aab83..499ec75d2496 100644
>> --- a/fs/ceph/dir.c
>> +++ b/fs/ceph/dir.c
>> @@ -784,6 +784,8 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
>> return ERR_PTR(-ENAMETOOLONG);
>> if (IS_ENCRYPTED(dir)) {
>> + bool had_key = fscrypt_has_encryption_key(dir);
>> +
>> err = __fscrypt_prepare_readdir(dir);
>> if (err)
>> return ERR_PTR(err);
>> @@ -791,6 +793,8 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
>> spin_lock(&dentry->d_lock);
>> dentry->d_flags |= DCACHE_NOKEY_NAME;
>> spin_unlock(&dentry->d_lock);
>> + } else if (!had_key) {
>> + __ceph_dir_clear_complete(ceph_inode(dir));
>> }
>> }
>>
>
> Yeah. Could we just add one helper, such as:

Sure, that makes sense. I'll send out a new version soon. Thanks!

Cheers,
--
Luís


> bool ceph_fscrypt_prepare_readdir(dir)
>
> {
>
>     bool had_key = fscrypt_has_encryption_key(dir);
>
>     err = __fscrypt_prepare_readdir(dir);
>
>     if (!err && !had_key || err = -ENOKEY && had_key) {
>
>         return true;   // key changed
>
>     }
>
>     return false;  // key not changed
>
> }
>
> Thanks!
>
> - Xiubo
>
>> Cheers,
>