summaryrefslogtreecommitdiffstats
path: root/fs/ceph
Commit message (Collapse)AuthorAgeFilesLines
* ceph: fix __choose_mds() for LSSNAP requestYan, Zheng2017-10-021-5/+6
| | | | | | | | | previous commit 5d37ca14 "ceph: send LSSNAP request to auth mds of directory inode" is buggy. It makes __choose_mds() choose mds base on hash of '.snap' dentry. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: properly queue cap snap for newly created snap realmYan, Zheng2017-10-021-5/+3
| | | | | | | | | | | commit 3ae0bebc "ceph: queue cap snap only when snap realm's context changes" introduced a regression: we may not call queue_realm_cap_snaps() for newly created snap realm. This regression allows unflushed snapshot data to be overwritten. Link: http://tracker.ceph.com/issues/21483 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: avoid panic in create_session_open_msg() if utsname() returns NULLYan, Zheng2017-09-192-3/+7
| | | | | | | | | | | | | utsname() can return NULL while process is exiting. Kernel releases file locks during process exits. We send request to mds when releasing file lock. So it's possible that we open mds session while process is exiting. utsname() is called in create_session_open_msg(). Link: http://tracker.ceph.com/issues/21275 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> [idryomov@gmail.com: drop utsname.h include from mds_client.c] Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* Merge tag 'ceph-for-4.14-rc1' of git://github.com/ceph/ceph-clientLinus Torvalds2017-09-1213-330/+408
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Pull ceph updates from Ilya Dryomov: "The highlights include: - a large series of fixes and improvements to the snapshot-handling code (Zheng Yan) - individual read/write OSD requests passed down to libceph are now limited to 16M in size to avoid hitting OSD-side limits (Zheng Yan) - encode MStatfs v2 message to allow for more accurate space usage reporting (Douglas Fuller) - switch to the new writeback error tracking infrastructure (Jeff Layton)" * tag 'ceph-for-4.14-rc1' of git://github.com/ceph/ceph-client: (35 commits) ceph: stop on-going cached readdir if mds revokes FILE_SHARED cap ceph: wait on writeback after writing snapshot data ceph: fix capsnap dirty pages accounting ceph: ignore wbc->range_{start,end} when write back snapshot data ceph: fix "range cyclic" mode writepages ceph: cleanup local variables in ceph_writepages_start() ceph: optimize pagevec iterating in ceph_writepages_start() ceph: make writepage_nounlock() invalidate page that beyonds EOF ceph: properly get capsnap's size in get_oldest_context() ceph: remove stale check in ceph_invalidatepage() ceph: queue cap snap only when snap realm's context changes ceph: handle race between vmtruncate and queuing cap snap ceph: fix message order check in handle_cap_export() ceph: fix NULL pointer dereference in ceph_flush_snaps() ceph: adjust 36 checks for NULL pointers ceph: delete an unnecessary return statement in update_dentry_lease() ceph: ENOMEM pr_err in __get_or_create_frag() is redundant ceph: check negative offsets in ceph_llseek() ceph: more accurate statfs ceph: properly set snap follows for cap reconnect ...
| * ceph: stop on-going cached readdir if mds revokes FILE_SHARED capYan, Zheng2017-09-061-6/+7
| | | | | | | | | | | | | | | | | | | | | | | | If directory's FILE_SHARED cap get revoked, dentry in the directory can get spliced into other directory (Eg, other client move the dentry into directory B, then we do readdir on directory B). So we should stop on-going cached readdir. this can be achieved by marking dir not complete, because __dcache_readdir() checks dir completeness before emitting each dentry. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: wait on writeback after writing snapshot dataYan, Zheng2017-09-061-0/+24
| | | | | | | | | | | | | | | | | | | | | | | | | | In sync mode, writepages() needs to write all dirty pages. But it can only write dirty pages associated with the oldest snapc. To write dirty pages associated with next snapc, it needs to wait until current writes complete. Without this wait, writepages() keeps looking up dirty pages, but the found dirty pages are not writeable. It wastes CPU time. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: fix capsnap dirty pages accountingYan, Zheng2017-09-061-2/+2
| | | | | | | | | | | | | | | | | | | | writepages_finish() calls ceph_put_wrbuffer_cap_refs() once for all pages, parameter snapc is set to req->r_snapc. So writepages() shouldn't write dirty pages associated with different snapc in one OSD request. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: ignore wbc->range_{start,end} when write back snapshot dataYan, Zheng2017-09-061-34/+46
| | | | | | | | | | | | | | | | | | | | | | | | writepages() needs to write dirty pages to OSD in strict order of snapshot context. It must first write dirty pages associated with the oldest snapshot context. In the write range case, dirty pages in the specified range can be associated with newer snapc. They are not writeable until we write all dirty pages associated with the oldest snapc. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: fix "range cyclic" mode writepagesYan, Zheng2017-09-061-17/+24
| | | | | | | | | | | | | | | | | | | | In range cyclic mode, writepages() should first write dirty pages in range [writeback_index, (pgoff_t)-1], then write pages in range [0, writeback_index -1]. Besides, if writepages() encounters a page that beyond EOF, it should restart from the beginning. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: cleanup local variables in ceph_writepages_start()Yan, Zheng2017-09-061-12/+9
| | | | | | | | | | | | | | Remove two variables and define variables of same type together. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: optimize pagevec iterating in ceph_writepages_start()Yan, Zheng2017-09-061-29/+25
| | | | | | | | | | | | | | | | | | ceph_writepages_start() supports writing non-continuous pages. If it encounters a non-dirty or non-writeable page in pagevec, it can continue to check the rest pages in pagevec. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: make writepage_nounlock() invalidate page that beyonds EOFYan, Zheng2017-09-061-18/+32
| | | | | | | | | | | | | | | | Otherwise, the page left in state that page is associated with a snapc, but (PageDirty(page) || PageWriteback(page)) is false. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: properly get capsnap's size in get_oldest_context()Yan, Zheng2017-09-061-57/+80
| | | | | | | | | | | | | | | | | | | | capsnap's size is set by __ceph_finish_cap_snap(). If capsnap is under writing, its size is zero. In this case, get_oldest_context() should read i_size. Besides, ceph_writepages_start() should re-check capsnap's size after dirty pages get locked. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: remove stale check in ceph_invalidatepage()Yan, Zheng2017-09-061-8/+1
| | | | | | | | | | | | | | | | Both set_page_dirty and truncate_complete_page should be called for locked page, they can't race with each other. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: queue cap snap only when snap realm's context changesYan, Zheng2017-09-061-21/+16
| | | | | | | | | | | | | | | | | | | | If we create capsnap when snap realm's context does not change, the new capsnap's snapc is equal to ci->i_head_snapc. Page writeback code can't differentiates dirty pages associated with the new capsnap from dirty pages associated with i_head_snapc. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: handle race between vmtruncate and queuing cap snapYan, Zheng2017-09-061-1/+12
| | | | | | | | | | | | | | | | | | It's possible that we create a cap snap while there is pending vmtruncate (truncate hasn't been processed by worker thread). We should truncate dirty pages beyond capsnap->size in that case. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: fix message order check in handle_cap_export()Yan, Zheng2017-09-061-1/+1
| | | | | | | | | | | | | | | | | | If caps for importer mds exists, but cap id mismatch, client should have received corresponding import message. Because cap ID does not change as long as client holds the caps. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: fix NULL pointer dereference in ceph_flush_snaps()Yan, Zheng2017-09-061-1/+1
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: adjust 36 checks for NULL pointersMarkus Elfring2017-09-0610-36/+36
| | | | | | | | | | | | | | | | | | | | | | | | The script “checkpatch.pl” pointed information out like the following. Comparison to NULL could be written ... Thus fix the affected source code places. Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Reviewed-by: Yan, Zheng <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: delete an unnecessary return statement in update_dentry_lease()Markus Elfring2017-09-061-1/+0
| | | | | | | | | | | | | | | | | | | | | | | | The script "checkpatch.pl" pointed information out like the following. WARNING: void function return statements are not generally useful Thus remove such a statement in the affected function. Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Reviewed-by: Yan, Zheng <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: ENOMEM pr_err in __get_or_create_frag() is redundantMarkus Elfring2017-09-061-5/+2
| | | | | | | | | | | | | | | | | | | | Omit an extra message for a memory allocation failure in this function. This issue was detected by using the Coccinelle software. Signed-off-by: Markus Elfring <elfring@users.sourceforge.net> Reviewed-by: Yan, Zheng <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: check negative offsets in ceph_llseek()Luis Henriques2017-09-061-2/+2
| | | | | | | | | | | | | | | | | | | | When a user requests SEEK_HOLE or SEEK_DATA with a negative offset ceph_llseek should return -ENXIO. Currently -EINVAL is being returned for SEEK_DATA and 0 for SEEK_HOLE. Signed-off-by: Luis Henriques <lhenriques@suse.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: more accurate statfsDouglas Fuller2017-09-061-1/+8
| | | | | | | | | | | | | | | | | | | | | | | | | | Improve accuracy of statfs reporting for Ceph filesystems comprising exactly one data pool. In this case, the Ceph monitor can now report the space usage for the single data pool instead of the global data for the entire Ceph cluster. Include support for this message in mon_client and leverage it in ceph/super. Signed-off-by: Douglas Fuller <dfuller@redhat.com> Reviewed-by: Yan, Zheng <zyan@redhat.com> Reviewed-by: Ilya Dryomov <idryomov@gmail.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: properly set snap follows for cap reconnectYan, Zheng2017-09-061-1/+1
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: don't use CEPH_OSD_FLAG_ORDERSNAPYan, Zheng2017-09-061-3/+3
| | | | | | | | | | | | | | | | | | | | Inode can be moved between snap realms. It's possible inode is moved into a snap realm whose seq number is smaller than old snap realm's. So there is no guarantee that seq number inode's snap context always increases. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: include snapc in debug message of writeYan, Zheng2017-09-062-7/+8
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: make sure flushsnap messages are sent in proper orderYan, Zheng2017-09-061-5/+7
| | | | | | | | | | | | | | | | | | Before sending new flushsnap message, check if there are old flushsnap messages that need to be re-sent. If there are, re-send old messages first. This guarantees ordering of flushsnap messages. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: fix -EOLDSNAPC handlingYan, Zheng2017-09-061-10/+7
| | | | | | | | | | | | | | | | Need to drop cap reference before retry. Besides, it's better to redo file write checks for each retry because we re-lock inode. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: send LSSNAP request to auth mds of directory inodeYan, Zheng2017-09-062-5/+14
| | | | | | | | | | | | | | | | Snapdir inode has no capability. __choose_mds() should choose mds base on capabilities of snapdir's parent inode. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: don't fill readdir cache for LSSNAP replyYan, Zheng2017-09-061-8/+11
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: cleanup ceph_readdir_prepopulate()Yan, Zheng2017-09-061-7/+0
| | | | | | | | | | | | | | In LSSNAP case, req->r_dentry is already set to snapdir dentry. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: use errseq_t for writeback error reportingJeff Layton2017-09-061-1/+1
| | | | | | | | | | | | | | | | | | Ensure that when writeback errors are marked that we report those to all file descriptions that were open at the time of the error. Signed-off-by: Jeff Layton <jlayton@redhat.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: new cap message flags indicate if there is pending capsnapYan, Zheng2017-09-061-1/+4
| | | | | | | | | | | | | | | | | | | | These flags tell mds if there is pending capsnap explicitly. Without this explicit notification, mds can only conclude if client has pending capsnap. The method mds use is inefficient and error-prone. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: nuke startsync opYanhu Cao2017-09-062-22/+4
| | | | | | | | | | | | | | | | | | startsync is a no-op, has been for years. Remove it. Link: http://tracker.ceph.com/issues/20604 Signed-off-by: Yanhu Cao <gmayyyha@gmail.com> Reviewed-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: validate correctness of some mount optionsYan, Zheng2017-09-062-7/+23
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: limit osd write sizeYan, Zheng2017-09-064-5/+11
| | | | | | | | | | | | | | | | | | | | OSD has a configurable limitation of max write size. OSD return error if write request size is larger than the limitation. For now, set max write size to CEPH_MSG_MAX_DATA_LEN. It should be small enough. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: limit osd read size to CEPH_MSG_MAX_DATA_LENYan, Zheng2017-09-064-18/+15
| | | | | | | | | | | | | | | | libceph returns -EIO when read size > CEPH_MSG_MAX_DATA_LEN. Link: http://tracker.ceph.com/issues/20528 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
| * ceph: remove unused cap_release_safety mount optionYan, Zheng2017-09-062-7/+0
| | | | | | | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* | Merge branch 'akpm' (patches from Andrew)Linus Torvalds2017-09-061-31/+0
|\ \ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Merge updates from Andrew Morton: - various misc bits - DAX updates - OCFS2 - most of MM * emailed patches from Andrew Morton <akpm@linux-foundation.org>: (119 commits) mm,fork: introduce MADV_WIPEONFORK x86,mpx: make mpx depend on x86-64 to free up VMA flag mm: add /proc/pid/smaps_rollup mm: hugetlb: clear target sub-page last when clearing huge page mm: oom: let oom_reap_task and exit_mmap run concurrently swap: choose swap device according to numa node mm: replace TIF_MEMDIE checks by tsk_is_oom_victim mm, oom: do not rely on TIF_MEMDIE for memory reserves access z3fold: use per-cpu unbuddied lists mm, swap: don't use VMA based swap readahead if HDD is used as swap mm, swap: add sysfs interface for VMA based swap readahead mm, swap: VMA based swap readahead mm, swap: fix swap readahead marking mm, swap: add swap readahead hit statistics mm/vmalloc.c: don't reinvent the wheel but use existing llist API mm/vmstat.c: fix wrong comment selftests/memfd: add memfd_create hugetlbfs selftest mm/shmem: add hugetlbfs support to memfd_create() mm, devm_memremap_pages: use multi-order radix for ZONE_DEVICE lookups mm/vmalloc.c: halve the number of comparisons performed in pcpu_get_vm_areas() ...
| * | fscache: remove unused ->now_uncached callbackJan Kara2017-09-061-31/+0
| |/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Patch series "Ranged pagevec lookup", v2. In this series I make pagevec_lookup() update the index (to be consistent with pagevec_lookup_tag() and also as a preparation for ranged lookups), provide ranged variant of pagevec_lookup() and use it in places where it makes sense. This not only removes some common code but is also a measurable performance win for some use cases (see patch 4/10) where radix tree is sparse and searching & grabing of a page after the end of the range has measurable overhead. This patch (of 10): The callback doesn't ever get called. Remove it. Link: http://lkml.kernel.org/r/20170726114704.7626-2-jack@suse.cz Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* | Merge tag 'locks-v4.14-1' of ↵Linus Torvalds2017-09-061-1/+1
|\ \ | |/ |/| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | git://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux Pull file locking updates from Jeff Layton: "This pile just has a few file locking fixes from Ben Coddington. There are a couple of cleanup patches + an attempt to bring sanity to the l_pid value that is reported back to userland on an F_GETLK request. After a few gyrations, he came up with a way for filesystems to communicate to the VFS layer code whether the pid should be translated according to the namespace or presented as-is to userland" * tag 'locks-v4.14-1' of git://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux: locks: restore a warn for leaked locks on close fs/locks: Remove fl_nspid and use fs-specific l_pid for remote locks fs/locks: Use allocation rather than the stack in fcntl_getlk()
| * fs/locks: Remove fl_nspid and use fs-specific l_pid for remote locksBenjamin Coddington2017-07-161-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Since commit c69899a17ca4 "NFSv4: Update of VFS byte range lock must be atomic with the stateid update", NFSv4 has been inserting locks in rpciod worker context. The result is that the file_lock's fl_nspid is the kworker's pid instead of the original userspace pid. The fl_nspid is only used to represent the namespaced virtual pid number when displaying locks or returning from F_GETLK. There's no reason to set it for every inserted lock, since we can usually just look it up from fl_pid. So, instead of looking up and holding struct pid for every lock, let's just look up the virtual pid number from fl_pid when it is needed. That means we can remove fl_nspid entirely. The translaton and presentation of fl_pid should handle the following four cases: 1 - F_GETLK on a remote file with a remote lock: In this case, the filesystem should determine the l_pid to return here. Filesystems should indicate that the fl_pid represents a non-local pid value that should not be translated by returning an fl_pid <= 0. 2 - F_GETLK on a local file with a remote lock: This should be the l_pid of the lock manager process, and translated. 3 - F_GETLK on a remote file with a local lock, and 4 - F_GETLK on a local file with a local lock: These should be the translated l_pid of the local locking process. Fuse was already doing the correct thing by translating the pid into the caller's namespace. With this change we must update fuse to translate to init's pid namespace, so that the locks API can then translate from init's pid namespace into the pid namespace of the caller. With this change, the locks API will expect that if a filesystem returns a remote pid as opposed to a local pid for F_GETLK, that remote pid will be <= 0. This signifies that the pid is remote, and the locks API will forego translating that pid into the pid namespace of the local calling process. Finally, we convert remote filesystems to present remote pids using negative numbers. Have lustre, 9p, ceph, cifs, and dlm negate the remote pid returned for F_GETLK lock requests. Since local pids will never be larger than PID_MAX_LIMIT (which is currently defined as <= 4 million), but pid_t is an unsigned int, we should have plenty of room to represent remote pids with negative numbers if we assume that remote pid numbers are similarly limited. If this is not the case, then we run the risk of having a remote pid returned for which there is also a corresponding local pid. This is a problem we have now, but this patch should reduce the chances of that occurring, while also returning those remote pid numbers, for whatever that may be worth. Signed-off-by: Benjamin Coddington <bcodding@redhat.com> Signed-off-by: Jeff Layton <jlayton@redhat.com>
* | ceph: fix readpage from fscacheYan, Zheng2017-09-012-18/+18
| | | | | | | | | | | | | | | | | | | | | | | | ceph_readpage() unlocks page prematurely prematurely in the case that page is reading from fscache. Caller of readpage expects that page is uptodate when it get unlocked. So page shoule get locked by completion callback of fscache_read_or_alloc_pages() Cc: stable@vger.kernel.org # 4.1+, needs backporting for < 4.7 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* | ceph: fix race in concurrent readdirYan, Zheng2017-07-171-0/+5
|/ | | | | | | | | | | | | | | | | | | | | | | | | For a large directory, program needs to issue multiple readdir syscalls to get all dentries. When there are multiple programs read the directory concurrently. Following sequence of events can happen. - program calls readdir with pos = 2. ceph sends readdir request to mds. The reply contains N1 entries. ceph adds these N1 entries to readdir cache. - program calls readdir with pos = N1+2. The readdir is satisfied by the readdir cache, N2 entries are returned. (Other program calls readdir in the middle, which fills the cache) - program calls readdir with pos = N1+N2+2. ceph sends readdir request to mds. The reply contains N3 entries and it reaches directory end. ceph adds these N3 entries to the readdir cache and marks directory complete. The second readdir call does not update fi->readdir_cache_idx. ceph add the last N3 entries to wrong places. Cc: stable@vger.kernel.org # v4.3+ Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: update ceph_dentry_info::lease_session when necessaryYan, Zheng2017-07-071-2/+7
| | | | | | | | | | Current code does not update ceph_dentry_info::lease_session once it is set. If auth mds of corresponding dentry changes, dentry lease keeps in an invalid state. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Reviewed-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: new mount option that specifies fscache uniquifierYan, Zheng2017-07-073-21/+113
| | | | | | | | | | | | | | | Current ceph uses FSID as primary index key of fscache data. This allows ceph to retain cached data across remount. But this causes problem (kernel opps, fscache does not support sharing data) when a filesystem get mounted several times (with fscache enabled, with different mount options). The fix is adding a new mount option, which specifies uniquifier for fscache. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Acked-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: avoid accessing freeing inode in ceph_check_delayed_caps()Yan, Zheng2017-07-071-2/+9
| | | | | Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: avoid invalid memory dereference in the middle of umountYan, Zheng2017-07-072-4/+6
| | | | | | | | | extra_mon_dispatch() and debugfs' foo_show functions dereference fsc->mdsc. we should clean up fsc->client->extra_mon_dispatch and debugfs before destroying fsc->mds. Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: getattr before read on ceph.* xattrsYan, Zheng2017-07-071-0/+3
| | | | | | | | | | | | | | Previously we were returning values for quota, layout xattrs without any kind of update -- the user just got whatever happened to be in our cache. Clearly this extra round trip has a cost, but reads of these xattrs are fairly rare, happening on admin intervention rather than in normal operation. Link: http://tracker.ceph.com/issues/17939 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
* ceph: don't re-send interrupted flock requestYan, Zheng2017-07-071-1/+24
| | | | | | | | | | Don't re-send interrupted flock request in cases of mds failover and receiving request forward. Because corresponding 'lock intr' request may have been finished, it won't get re-sent. Link: http://tracker.ceph.com/issues/20170 Signed-off-by: "Yan, Zheng" <zyan@redhat.com> Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
OpenPOWER on IntegriCloud