From: Xiubo Li <xiubli@xxxxxxxxxx> 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. URL: https://tracker.ceph.com/issues/58126 Signed-off-by: Xiubo Li <xiubli@xxxxxxxxxx> --- fs/ceph/addr.c | 10 ++++++++++ fs/ceph/super.c | 11 +++++++++++ fs/ceph/super.h | 2 ++ 3 files changed, 23 insertions(+) diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c index 78ad45567dbb..de9b82905f18 100644 --- a/fs/ceph/addr.c +++ b/fs/ceph/addr.c @@ -284,6 +284,7 @@ static void finish_netfs_read(struct ceph_osd_request *req) } netfs_subreq_terminated(subreq, err, false); iput(req->r_inode); + ceph_dec_osd_stopping_blocker(fsc->mdsc); } static bool ceph_netfs_issue_op_inline(struct netfs_io_subrequest *subreq) @@ -411,6 +412,10 @@ static void ceph_netfs_issue_read(struct netfs_io_subrequest *subreq) } else { osd_req_op_extent_osd_iter(req, 0, &iter); } + if (!ceph_inc_osd_stopping_blocker(fsc->mdsc)) { + err = -EIO; + goto out; + } req->r_callback = finish_netfs_read; req->r_priv = subreq; req->r_inode = inode; @@ -906,6 +911,7 @@ static void writepages_finish(struct ceph_osd_request *req) else kfree(osd_data->pages); ceph_osdc_put_request(req); + ceph_dec_osd_stopping_blocker(fsc->mdsc); } /* @@ -1214,6 +1220,10 @@ static int ceph_writepages_start(struct address_space *mapping, BUG_ON(len < ceph_fscrypt_page_offset(pages[locked_pages - 1]) + thp_size(pages[locked_pages - 1]) - offset); + if (!ceph_inc_osd_stopping_blocker(fsc->mdsc)) { + rc = -EIO; + goto release_folios; + } req->r_callback = writepages_finish; req->r_inode = inode; diff --git a/fs/ceph/super.c b/fs/ceph/super.c index d3f54f3d7b17..401fe61ea53a 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -1524,6 +1524,17 @@ void ceph_dec_mds_stopping_blocker(struct ceph_mds_client *mdsc) __dec_stopping_blocker(mdsc); } +/* For data IO requests */ +bool ceph_inc_osd_stopping_blocker(struct ceph_mds_client *mdsc) +{ + return __inc_stopping_blocker(mdsc); +} + +void ceph_dec_osd_stopping_blocker(struct ceph_mds_client *mdsc) +{ + __dec_stopping_blocker(mdsc); +} + static void ceph_kill_sb(struct super_block *s) { struct ceph_fs_client *fsc = ceph_sb_to_client(s); diff --git a/fs/ceph/super.h b/fs/ceph/super.h index cd5b88d819ca..2f9b6fc667b8 100644 --- a/fs/ceph/super.h +++ b/fs/ceph/super.h @@ -1418,4 +1418,6 @@ extern void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc); bool ceph_inc_mds_stopping_blocker(struct ceph_mds_client *mdsc, struct ceph_mds_session *session); void ceph_dec_mds_stopping_blocker(struct ceph_mds_client *mdsc); +bool ceph_inc_osd_stopping_blocker(struct ceph_mds_client *mdsc); +void ceph_dec_osd_stopping_blocker(struct ceph_mds_client *mdsc); #endif /* _FS_CEPH_SUPER_H */ -- 2.40.1