- 24 Aug, 2023 32 commits
-
-
Xiubo Li authored
When fscrypt is enabled we will align the truncate size up to the CEPH_FSCRYPT_BLOCK_SIZE always, so if we truncate the size in the same block more than once, the latter ones will be skipped being invalidated from the page caches. This will force invalidating the page caches by using the smaller size than the real file size. At the same time add more debug log and fix the debug log for truncate code. Link: https://tracker.ceph.com/issues/58834Signed-off-by:
Xiubo Li <xiubli@redhat.com> Reviewed-and-tested-by:
Luís Henriques <lhenriques@suse.de> Reviewed-by:
Milind Changire <mchangir@redhat.com> Signed-off-by:
Ilya Dryomov <idryomov@gmail.com>
-
Xiubo Li authored
The sync_filesystem() will flush all the dirty buffer and submit the osd reqs to the osdc and then is blocked to wait for all the reqs to finish. But the when the reqs' replies come, the reqs will be removed from osdc just before the req->r_callback()s are called. Which means the sync_filesystem() will be woke up by leaving the req->r_callback()s are still running. This will be buggy when the waiter require the req->r_callback()s to release some resources before continuing. So we need to make sure the req->r_callback()s are called before removing the reqs from the osdc. WARNING: CPU: 4 PID: 168846 at fs/crypto/keyring.c:242 fscrypt_destroy_keyring+0x7e/0xd0 CPU: 4 PID: 168846 Comm: umount Tainted: G S 6.1.0-rc5-ceph-g72ead199864c #1 Hardware name: Supermicro SYS-5018R-WR/X10SRW-F, BIOS 2.0 12/17/2015 RIP: 0010:fscrypt_destroy_keyring+0x7e/0xd0 RSP: 0018:ffffc9000b277e28 EFLAGS: 00010202 RAX: 0000000000000002 RBX: ffff88810d52ac00 RCX: ffff88810b56aa00 RDX: 0000000080000000 RSI: ffffffff822f3a09 RDI: ffff888108f59000 RBP: ffff8881d394fb88 R08: 0000000000000028 R09: 0000000000000000 R10: 0000000000000001 R11: 11ff4fe6834fcd91 R12: ffff8881d394fc40 R13: ffff888108f59000 R14: ffff8881d394f800 R15: 0000000000000000 FS: 00007fd83f6f1080(0000) GS:ffff88885fd00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f918d417000 CR3: 000000017f89a005 CR4: 00000000003706e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <TASK> generic_shutdown_super+0x47/0x120 kill_anon_super+0x14/0x30 ceph_kill_sb+0x36/0x90 [ceph] deactivate_locked_super+0x29/0x60 cleanup_mnt+0xb8/0x140 task_work_run+0x67/0xb0 exit_to_user_mode_prepare+0x23d/0x240 syscall_exit_to_user_mode+0x25/0x60 do_syscall_64+0x40/0x80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7fd83dc39e9b We need to increase the blocker counter to make sure all the osd requests' callbacks have been finished just before calling the kill_anon_super() when unmounting. Link: https://tracker.ceph.com/issues/58126Signed-off-by:
-
Xiubo Li authored
When unmounting all the dirty buffers will be flushed and after the last osd request is finished the last reference of the i_count will be released. Then it will flush the dirty cap/snap to MDSs, and the unmounting won't wait the possible acks, which will ihold the inodes when updating the metadata locally but makes no sense any more, of this. This will make the evict_inodes() to skip these inodes. If encrypt is enabled the kernel generate a warning when removing the encrypt keys when the skipped inodes still hold the keyring: WARNING: CPU: 4 PID: 168846 at fs/crypto/keyring.c:242 fscrypt_destroy_keyring+0x7e/0xd0 CPU: 4 PID: 168846 Comm: umount Tainted: G S 6.1.0-rc5-ceph-g72ead199864c #1 Hardware name: Supermicro SYS-5018R-WR/X10SRW-F, BIOS 2.0 12/17/2015 RIP: 0010:fscrypt_destroy_keyring+0x7e/0xd0 RSP: 0018:ffffc9000b277e28 EFLAGS: 00010202 RAX: 0000000000000002 RBX: ffff88810d52ac00 RCX: ffff88810b56aa00 RDX: 0000000080000000 RSI: ffffffff822f3a09 RDI: ffff888108f59000 RBP: ffff8881d394fb88 R08: 0000000000000028 R09: 0000000000000000 R10: 0000000000000001 R11: 11ff4fe6834fcd91 R12: ffff8881d394fc40 R13: ffff888108f59000 R14: ffff8881d394f800 R15: 0000000000000000 FS: 00007fd83f6f1080(0000) GS:ffff88885fd00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f918d417000 CR3: 000000017f89a005 CR4: 00000000003706e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <TASK> generic_shutdown_super+0x47/0x120 kill_anon_super+0x14/0x30 ceph_kill_sb+0x36/0x90 [ceph] deactivate_locked_super+0x29/0x60 cleanup_mnt+0xb8/0x140 task_work_run+0x67/0xb0 exit_to_user_mode_prepare+0x23d/0x240 syscall_exit_to_user_mode+0x25/0x60 do_syscall_64+0x40/0x80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7fd83dc39e9b Later the kernel will crash when iput() the inodes and dereferencing the "sb->s_master_keys", which has been released by the generic_shutdown_super(). Link: https://tracker.ceph.com/issues/59162Signed-off-by:
-
Luís Henriques authored
Signed-off-by:
Jeff Layton <jlayton@kernel.org> Reviewed-by:
-
Luís Henriques authored
With snapshot names encryption we can not allow snapshots to be created in locked directories because the names wouldn't be encrypted. This patch forces the directory to be unlocked to allow a snapshot to be created. Signed-off-by:
-
Luís Henriques authored
Since filenames in encrypted directories are encrypted and shown as a base64-encoded string when the directory is locked, make snapshot names show a similar behaviour. When creating a snapshot, .snap directories for every subdirectory will show the snapshot name in the "long format": # mkdir .snap/my-snap # ls my-dir/.snap/ _my-snap_1099511627782 Encrypted snapshots will need to be able to handle these by encrypting/decrypting only the snapshot part of the string ('my-snap'). Also, since the MDS prevents snapshot names to be bigger than 240 characters it is necessary to adapt CEPH_NOHASH_NAME_MAX to accommodate this extra limitation. [ idryomov: drop const on !CONFIG_FS_ENCRYPTION branch too ] Signed-off-by: -
Luís Henriques authored
When doing a direct/sync write, we need to invalidate the page cache in the range being written to. If we don't do this, the cache will include invalid data as we just did a write that avoided the page cache. In the event that invalidation fails, just ignore the error. That likely just means that we raced with another task doing a buffered write, in which case we want to leave the page intact anyway. [ jlayton: minor comment update ] Signed-off-by:
-
Jeff Layton authored
Force the use of sparse reads when the inode is encrypted, and add the appropriate code to decrypt the extent map after receiving. Note that the crypto block may be smaller than a page, but the reverse cannot be true. Signed-off-by:
-
Jeff Layton authored
Allow writepage to issue encrypted writes. Extend out the requested size and offset to cover complete blocks, and then encrypt and write them to the OSDs. Add the appropriate machinery to write back dirty data with encryption. Signed-off-by:
-
Jeff Layton authored
When doing a synchronous write on an encrypted inode, we have no guarantee that the caller is writing crypto block-aligned data. When that happens, we must do a read/modify/write cycle. First, expand the range to cover complete blocks. If we had to change the original pos or length, issue a read to fill the first and/or last pages, and fetch the version of the object from the result. We then copy data into the pages as usual, encrypt the result and issue a write prefixed by an assertion that the version hasn't changed. If it has changed then we restart the whole thing again. If there is no object at that position in the file (-ENOENT), we prefix the write on an exclusive create of the object instead. Signed-off-by:
-
Jeff Layton authored
Encrypted files will need to be dealt with in block-sized chunks and once we do that, the way that ceph_sync_write aligns the data in the bounce buffer won't be acceptable. Change it to align the data the same way it would be aligned in the pagecache. Signed-off-by:
-
Jeff Layton authored
Eventually I want to merge the synchronous and direct read codepaths, possibly via new netfs infrastructure. For now, the direct path is not crypto-enabled, so use the sync read/write paths instead. Signed-off-by:
-
Xiubo Li authored
This will transfer the encrypted last block contents to the MDS along with the truncate request only when the new size is smaller and not aligned to the fscrypt BLOCK size. When the last block is located in the file hole, the truncate request will only contain the header. The MDS could fail to do the truncate if there has another client or process has already updated the RADOS object which contains the last block, and will return -EAGAIN, then the kclient needs to retry it. The RMW will take around 50ms, and will let it retry 20 times for now. Signed-off-by:
-
Xiubo Li authored
Turn the guts of ceph_sync_read into a new helper that takes an inode and an offset instead of a kiocb struct, and make ceph_sync_read call the helper as a wrapper. Make the new helper always return the last object's version. Signed-off-by:
-
Jeff Layton authored
Currently we have some special-casing for multi-op writes, but in the case of a read, we can't really handle it. All of the current multi-op callers call it with CEPH_OSD_FLAG_WRITE set. Have ceph_osdc_new_request check for CEPH_OSD_FLAG_READ and if it's set, allocate multiple reply ops instead of multiple request ops. If neither flag is set, return -EINVAL. Signed-off-by:
-
Jeff Layton authored
...and record the user_version in the reply in a new field in ceph_osd_request, so we can populate the assert_ver appropriately. Shuffle the fields a bit too so that the new field fits in an existing hole on x86_64. Signed-off-by:
-
Jeff Layton authored
...and allow test_dummy_encryption to bypass content encryption if mounted with test_dummy_encryption=clear. [ xiubli: remove test_dummy_encryption=clear support per Ilya ] Signed-off-by:
-
Jeff Layton authored
Handle the new fscrypt_file and fscrypt_auth fields in cap messages. Use them to populate new fields in cap_extra_info and update the inode with those values. Signed-off-by:
-
Jeff Layton authored
For encrypted inodes, transmit a rounded-up size to the MDS as the normal file size and send the real inode size in fscrypt_file field. Also, fix up creates and truncates to also transmit fscrypt_file. When we get an inode trace from the MDS, grab the fscrypt_file field if the inode is encrypted, and use it to populate the i_size field instead of the regular inode size field. Signed-off-by:
-
Luís Henriques authored
When setting a directory's crypt context, ceph_dir_clear_complete() needs to be called otherwise if it was complete before, any existing (old) dentry will still be valid. This patch adds a wrapper around __fscrypt_prepare_readdir() which will ensure a directory is marked as non-complete if key status changes. [ xiubli: revise commit title per Milind ] Signed-off-by:
-
Luís Henriques authored
If a client doesn't have Fx caps on a directory, it will get errors while trying encrypt it: ceph: handle_cap_grant: cap grant attempt to change fscrypt_auth on non-I_NEW inode (old len 0 new len 48) fscrypt (ceph, inode 1099511627812): Error -105 getting encryption context A simple way to reproduce this is to use two clients: client1 # mkdir /mnt/mydir client2 # ls /mnt/mydir client1 # fscrypt encrypt /mnt/mydir client1 # echo hello > /mnt/mydir/world This happens because, in __ceph_setattr(), we only initialize ci->fscrypt_auth if we have Ax and ceph_fill_inode() won't use the fscrypt_auth received if the inode state isn't I_NEW. Fix it by allowing ceph_fill_inode() to also set ci->fscrypt_auth if the inode doesn't have it set already. Signed-off-by: -
Jeff Layton authored
Add the appropriate calls into fscrypt for various actions, including link, rename, setattr, and the open codepaths. Disable fallocate for encrypted inodes -- hopefully, just for now. If we have an encrypted inode, then the client will need to re-encrypt the contents of the new object. Disable copy offload to or from encrypted inodes. Set i_blkbits to crypto block size for encrypted inodes -- some of the underlying infrastructure for fscrypt relies on i_blkbits being aligned to crypto blocksize. Report STATX_ATTR_ENCRYPTED on encrypted inodes. [ lhenriques: forbid encryption with striped layouts ] Signed-off-by:
-
Jeff Layton authored
When creating symlinks in encrypted directories, encrypt and base64-encode the target with the new inode's key before sending to the MDS. When filling a symlinked inode, base64-decode it into a buffer that we'll keep in ci->i_symlink. When get_link is called, decrypt the buffer into a new one that will hang off i_link. Signed-off-by:
-
Xiubo Li authored
To make it simpler to decrypt names in a readdir reply (i.e. before we have a dentry), add a new ceph_encode_encrypted_fname()-like helper that takes a qstr pointer instead of a dentry pointer. Once we've decrypted the names in a readdir reply, we no longer need the crypttext, so overwrite them in ceph_mds_reply_dir_entry with the unencrypted names. Then in both ceph_readdir_prepopulate() and ceph_readdir() we will use the dencrypted name directly. [ jlayton: convert some BUG_ONs into error returns ] Signed-off-by:
-
Xiubo Li authored
Instead of passing just the r_reply_info to the readdir reply parser, pass the request pointer directly instead. This will facilitate implementing readdir on fscrypted directories. Signed-off-by:
-
Jeff Layton authored
When we get a dentry in a trace, decrypt the name so we can properly instantiate the dentry or fill out ceph_get_name() buffer. Signed-off-by:
-
Jeff Layton authored
Define a new ceph_fname struct that we can use to carry information about encrypted dentry names. Add helpers for working with these objects, including ceph_fname_to_usr which formats an encrypted filename for userland presentation. [ xiubli: fix resulting name length check -- neither name_len nor ctext_len should exceed NAME_MAX ] Signed-off-by:
-
Jeff Layton authored
If we have a dentry which represents a no-key name, then we need to test whether the parent directory's encryption key has since been added. Do that before we test anything else about the dentry. Signed-off-by:
-
Jeff Layton authored
This is required so that we know to invalidate these dentries when the directory is unlocked. Atomic open can act as a lookup if handed a dentry that is negative on the MDS. Ensure that we set DCACHE_NOKEY_NAME on the dentry in atomic_open, if we don't have the key for the parent. Otherwise, we can end up validating the dentry inappropriately if someone later adds a key. Signed-off-by:
-
Jeff Layton authored
Ceph is a bit different from local filesystems, in that we don't want to store filenames as raw binary data, since we may also be dealing with clients that don't support fscrypt. We could just base64-encode the encrypted filenames, but that could leave us with filenames longer than NAME_MAX. It turns out that the MDS doesn't care much about filename length, but the clients do. To manage this, we've added a new "alternate name" field that can be optionally added to any dentry that we'll use to store the binary crypttext of the filename if its base64-encoded value will be longer than NAME_MAX. When a dentry has one of these names attached, the MDS will send it along in the lease info, which we can then store for later usage. Signed-off-by:
-
Jeff Layton authored
In the event that we have a filename longer than CEPH_NOHASH_NAME_MAX, we'll need to hash the tail of the filename. The client however will still need to know the full name of the file if it has a key. To support this, the MClientRequest field has grown a new alternate_name field that we populate with the full (binary) crypttext of the filename. This is then transmitted to the clients in readdir or traces as part of the dentry lease. Add support for populating this field when the filenames are very long. Signed-off-by:
-
Jeff Layton authored
Allow ceph_mdsc_build_path to encrypt and base64 encode the filename when the parent is encrypted and we're sending the path to the MDS. In a similar fashion, encode encrypted dentry names if including a dentry release in a request. In most cases, we just encrypt the filenames and base64 encode them, but when the name is longer than CEPH_NOHASH_NAME_MAX, we use a similar scheme to fscrypt proper, and hash the remaning bits with sha256. When doing this, we then send along the full crypttext of the name in the new alternate_name field of the MClientRequest. The MDS can then send that along in readdir responses and traces. [ idryomov: drop duplicate include reported by Abaci Robot ] Signed-off-by:
-
- 22 Aug, 2023 8 commits
-
-
Luís Henriques authored
The base64url encoding used by fscrypt includes the '_' character, which may cause problems in snapshot names (if the name starts with '_'). Thus, use the base64 encoding defined for IMAP mailbox names (RFC 3501), which uses '+' and ',' instead of '-' and '_'. Signed-off-by:
-
Xiubo Li authored
ioctl file 0000000004e6b054 cmd 2148296211 arg 824635143532 The numerical cmd value in the ioctl debug log message is too hard to understand even when you look at it in the code. Make it more readable. [ idryomov: add missing _ in ceph_ioctl_cmd_name() ] Signed-off-by: -
Jeff Layton authored
We gate most of the ioctls on MDS feature support. The exception is the key removal and status functions that we still want to work if the MDS's were to (inexplicably) lose the feature. For the set_policy ioctl, we take Fs caps to ensure that nothing can create files in the directory while the ioctl is running. That should be enough to ensure that the "empty_dir" check is reliable. The vxattr is read-only, added mostly for future debugging purposes. Signed-off-by:
-
Jeff Layton authored
Add support for the test_dummy_encryption mount option. This allows us to test the encrypted codepaths in ceph without having to manually set keys, etc. [ lhenriques: fix potential fsc->fsc_dummy_enc_policy memory leak in ceph_real_mount() ] Signed-off-by:
-
Jeff Layton authored
Most fscrypt-enabled filesystems store the crypto context in an xattr, but that's problematic for ceph as xatts are governed by the XATTR cap, but we really want the crypto context as part of the AUTH cap. Because of this, the MDS has added two new inode metadata fields: fscrypt_auth and fscrypt_file. The former is used to hold the crypto context, and the latter is used to track the real file size. Parse new fscrypt_auth and fscrypt_file fields in inode traces. For now, we don't use fscrypt_file, but fscrypt_auth is used to hold the fscrypt context. Allow the client to use a setattr request for setting the fscrypt_auth field. Since this is not a standard setattr request from the VFS, we add a new field to __ceph_setattr that carries ceph-specific inode attrs. Have the set_context op do a setattr that sets the fscrypt_auth value, and get_context just return the contents of that field (since it should always be available). Signed-off-by:
-
Jeff Layton authored
The netfs layer has already pinned the pages involved before calling issue_op, so we can just pass down the iter directly instead of calling iov_iter_get_pages_alloc. Instead of having to allocate a page array, use CEPH_MSG_DATA_ITER and pass it the iov_iter directly to clone. Signed-off-by:
-
Jeff Layton authored
Add an iov_iter to the unions in ceph_msg_data and ceph_msg_data_cursor. Instead of requiring a list of pages or bvecs, we can just use an iov_iter directly, and avoid extra allocations. We assume that the pages represented by the iter are pinned such that they shouldn't incur page faults, which is the case for the iov_iters created by netfs. While working on this, Al Viro informed me that he was going to change iov_iter_get_pages to auto-advance the iterator as that pattern is more or less required for ITER_PIPE anyway. We emulate that here for now by advancing in the _next op and tracking that amount in the "lastlen" field. In the event that _next is called twice without an intervening _advance, we revert the iov_iter by the remaining lastlen before calling iov_iter_get_pages. Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: David Howells <dhowells@redhat.com> Signed-off-by:
-
Jeff Layton authored
Encryption potentially requires allocation, at which point we'll need to be in a non-atomic context. Convert ceph_msdc_build_path to take dentry spinlocks and references instead of using rcu_read_lock to walk the path. This is slightly less efficient, and we may want to eventually allow using RCU when the leaf dentry isn't encrypted. Signed-off-by:
-
