ceph-devel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Xiubo Li <xiubli@redhat.com>
To: Eric Biggers <ebiggers@kernel.org>, Jeff Layton <jlayton@kernel.org>
Cc: ceph-devel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-fscrypt@vger.kernel.org, dhowells@redhat.com,
	lhenriques@suse.de, khiremat@redhat.com
Subject: Re: [RFC PATCH v8 09/24] ceph: add ability to set fscrypt_auth via setattr
Date: Wed, 1 Sep 2021 08:53:51 +0800	[thread overview]
Message-ID: <fca1da91-07aa-61ab-8d43-17a15a72691b@redhat.com> (raw)
In-Reply-To: <YS5s5mYZtc3r+K/E@sol.localdomain>


On 9/1/21 1:54 AM, Eric Biggers wrote:
> On Tue, Aug 31, 2021 at 09:50:32AM -0400, Jeff Layton wrote:
>>>>>> +		/* It should never be re-set once set */
>>>>>> +		WARN_ON_ONCE(ci->fscrypt_auth);
>>>>>> +
>>>>> Maybe this should return -EEXIST if already set ?
>>>>>
>>>> I don't know. In general, once the context is set on an inode, we
>>>> shouldn't ever reset it. That said, I think we might need to allow
>>>> admins to override an existing context if it's corrupted.
>>>>
>>>> So, that's the rationale for the WARN_ON_ONCE. The admins should never
>>>> do this under normal circumstances but they do have the ability to
>>>> change it if needed (and we'll see a warning in the logs in that case).
>>> I may miss some code in the fs/crypto/ layer.
>>>
>>> I readed that once the directory/file has set the policy context, it
>>> will just return 0 if the new one matches the existence, if not match it
>>> will return -EEXIST, or will try to call ceph layer to set it.
>>>
>>> So once this is set, my understanding is that it shouldn't be here ?
>>>
>> Where did you read that? If we have documented semantics we need to
>> follow here, then we should change it to comply with them.
>>
> That is how FS_IOC_SET_ENCRYPTION_POLICY behaves, but the check for an existing
> policy already happens in fscrypt_ioctl_set_policy(), so ->set_context doesn't
> need to worry about it.

Yeah, the FS_IOC_SET_ENCRYPTION_POLICY section in 
"Documentation/filesystems/fscrypt.rst".


> - Eric
>


  reply	other threads:[~2021-09-01  0:54 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-26 16:19 [RFC PATCH v8 00/24] ceph+fscrypt: context, filename and symlink support Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 01/24] vfs: export new_inode_pseudo Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 02/24] fscrypt: export fscrypt_base64url_encode and fscrypt_base64url_decode Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 03/24] fscrypt: export fscrypt_fname_encrypt and fscrypt_fname_encrypted_size Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 04/24] fscrypt: add fscrypt_context_for_new_inode Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 05/24] ceph: preallocate inode for ops that may create one Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 06/24] ceph: parse new fscrypt_auth and fscrypt_file fields in inode traces Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 07/24] ceph: add fscrypt_* handling to caps.c Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 08/24] ceph: crypto context handling for ceph Jeff Layton
2021-08-26 16:19 ` [RFC PATCH v8 09/24] ceph: add ability to set fscrypt_auth via setattr Jeff Layton
2021-08-31  5:06   ` Xiubo Li
2021-08-31 12:43     ` Jeff Layton
2021-08-31 13:22       ` Xiubo Li
2021-08-31 13:50         ` Jeff Layton
2021-08-31 17:54           ` Eric Biggers
2021-09-01  0:53             ` Xiubo Li [this message]
2021-09-01  1:13           ` Xiubo Li
2021-09-01 12:02             ` Jeff Layton
2021-09-02  1:54               ` Xiubo Li
2021-08-26 16:20 ` [RFC PATCH v8 10/24] ceph: implement -o test_dummy_encryption mount option Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 11/24] ceph: add fscrypt ioctls Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 12/24] ceph: decode alternate_name in lease info Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 13/24] ceph: make ceph_msdc_build_path use ref-walk Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 14/24] ceph: add encrypted fname handling to ceph_mdsc_build_path Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 15/24] ceph: send altname in MClientRequest Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 16/24] ceph: encode encrypted name in dentry release Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 17/24] ceph: properly set DCACHE_NOKEY_NAME flag in lookup Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 18/24] ceph: make d_revalidate call fscrypt revalidator for encrypted dentries Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 19/24] ceph: add helpers for converting names for userland presentation Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 20/24] ceph: add fscrypt support to ceph_fill_trace Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 21/24] ceph: add support to readdir for encrypted filenames Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 22/24] ceph: create symlinks with encrypted and base64-encoded targets Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 23/24] ceph: make ceph_get_name decrypt filenames Jeff Layton
2021-08-26 16:20 ` [RFC PATCH v8 24/24] ceph: add a new ceph.fscrypt.auth vxattr Jeff Layton

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=fca1da91-07aa-61ab-8d43-17a15a72691b@redhat.com \
    --to=xiubli@redhat.com \
    --cc=ceph-devel@vger.kernel.org \
    --cc=dhowells@redhat.com \
    --cc=ebiggers@kernel.org \
    --cc=jlayton@kernel.org \
    --cc=khiremat@redhat.com \
    --cc=lhenriques@suse.de \
    --cc=linux-fscrypt@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --subject='Re: [RFC PATCH v8 09/24] ceph: add ability to set fscrypt_auth via setattr' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).