BMC/Intel-BMC/linux.git/fs/ceph/caps.c, branch dev-4.17

ceph: quota: cache inode pointer in ceph_snap_realm

2018-04-02T09:17:53+00:00

Keep a pointer to the inode in struct ceph_snap_realm. This allows to optimize functions that walk the realms hierarchy (e.g. in quotas). Signed-off-by: Luis Henriques Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: optimizing cap reservation

2018-04-02T08:12:47+00:00

When caps_avail_count is in a low level, most newly trimmed caps will probably go into ->caps_list and caps_avail_count will be increased. Hence after trimming, should recheck caps_avail_count to effectly reuse newly trimmed caps. Also, when releasing unnecessary caps follow the same rule of ceph_put_cap. Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: release unreserved caps if having enough available caps

2018-04-02T08:12:47+00:00

When unreserving caps check if there is too mamy available caps in the ->caps_list, if so release unreserved caps. Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: optimizing cap allocation

2018-04-02T08:12:46+00:00

When setting high volume of caps_min_count or having many unreserved caps, unused caps may always keep in the ->caps_list even can't get new cap from kmem_cache_alloc because lack of maximum limitation of caps_avail_count. Hence reuse caps in ->caps_list if available, it's maybe better than setting max limitation of caps_avail_count and releasing unused caps when reaching the limit. Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: adding protection for showing cap reservation info

2018-04-02T08:12:46+00:00

Adding spinlock protection during getting cap reservation ralated fields so that the numbers match below BUG_ON condition in the code. BUG_ON(mdsc->caps_total_count != mdsc->caps_use_count + mdsc->caps_reserve_count + mdsc->caps_avail_count); Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: flush dirty caps of unlinked inode ASAP

2018-02-26T15:19:16+00:00

Client should release unlinked inode from its cache ASAP. But client can't release inode with dirty caps. Link: http://tracker.ceph.com/issues/22886 Signed-off-by: Zhi Zhang Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: improving efficiency of syncfs

2018-01-30T20:10:15+00:00

write_inode() could be called variety of reasons, in the case of syncfs(2) there is no need to wait for flush getting completed in write_inode(), ->sync_fs is for guaranteeing flush completion for all inodes at that point. Signed-off-by: Chengguang Xu Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: try to allocate enough memory for reserved caps

2018-01-29T17:36:12+00:00

ceph_reserve_caps() may not reserve enough caps under high memory pressure, but it saved the needed caps number that expected to be reserved. When getting caps, crash would happen due to number mismatch. Now we will try to trim more caps when failing to allocate memory for caps need to be reserved, then try again. If still failing to allocate memory, return -ENOMEM. Signed-off-by: Zhi Zhang Reviewed-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: fix race of queuing delayed caps

2018-01-29T17:36:11+00:00

When called with CHECK_CAPS_AUTHONLY flag, ceph_check_caps() only processes auth caps. In that case, it's unsafe to remove inode from mdsc->cap_delay_list, because there can be delayed non-auth caps. Besides, ceph_check_caps() may lock/unlock i_ceph_lock several times, when multiple threads call ceph_check_caps() at the same time. It's possible that one thread calls __cap_delay_requeue(), another thread calls __cap_delay_cancel(). __cap_delay_cancel() should be called at very beginning of ceph_check_caps(), so that it does not race with __cap_delay_requeue(). Signed-off-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov

ceph: delete unreachable code in ceph_check_caps()

2018-01-29T17:36:11+00:00

"revoking & (CEPH_CAP_FILE_CACHE|CEPH_CAP_FILE_LAZYIO)" has already been tested before calling try_nonblocking_invalidate() Signed-off-by: "Yan, Zheng" Signed-off-by: Ilya Dryomov