FreeBSD source tree https://cgit-dev.freebsd.org/src-test/atom?h=stable%2F7 2016-02-02T19:29:29Z 2016-02-02T19:29:29Z John Baldwin jhb@FreeBSD.org 2016-02-02T19:29:29Z urn:sha1:546d648266fdaaa364294be44359fd13ff6f9f7f namecache hit. This was ommitted due to a merging error in r238913. The effect was to usually break caching of negative name lookups in the NFS client. Submitted by: bde 2012-07-30T19:05:41Z John Baldwin jhb@FreeBSD.org 2012-07-30T19:05:41Z urn:sha1:2e89c94eaa744b214d8f3a7180c3825152dfafe2 - For dotdot lookup in nfs_lookup, inline the vn_vget_ino() to prevent operating on the unmounted mount point and freed mount data in case of forced unmount performed while dvp is unlocked to nget the target vnode. - Close a race in NFS lookup processing that could result in stale name cache entries on one client when a directory was renamed on another client. The root cause for the stale entry being trusted is that each per-vnode nfsnode structure has a single 'n_ctime' timestamp used to validate positive name cache entries. However, if there are multiple entries for a single vnode, they all share a single timestamp. To fix this, extend the name cache to allow filesystems to optionally store a timestamp value in each name cache entry. The NFS clients now fetch the timestamp associated with each name cache entry and use that to validate cache hits instead of the timestamps previously stored in the nfsnode. Another part of the fix is that the NFS clients now use timestamps from the post-op attributes of RPCs when adding name cache entries rather than pulling the timestamps out of the file's attribute cache. The latter is subject to races with other lookups updating the attribute cache concurrently. - Adjust the nfs_skip_wcc_data_onerr setting so that it does not block post-op attributes for ENOENT errors now that the name caching logic depends on working post-op attributes. Tested by: Mark Saad nonesuch longcount org 2010-04-18T04:49:07Z Konstantin Belousov kib@FreeBSD.org 2010-04-18T04:49:07Z urn:sha1:f4e8785eb4952e053ab3964ad294c2197e25be4e Fix typo. 2009-05-15T20:36:54Z John Baldwin jhb@FreeBSD.org 2009-05-15T20:36:54Z urn:sha1:a43e11c99c1375e73805e23c14866f5c8179b977 not noticing new files created by another client for a "long" time, reduce the default timeout for caching attributes of directories from 30 seconds to 3 seconds. 2009-05-04T21:58:32Z John Baldwin jhb@FreeBSD.org 2009-05-04T21:58:32Z urn:sha1:cdc332da614786472df7d803990e077967fcae05 forced unmount, consistently return ENOENT rather than EBADF. 2009-04-17T21:10:53Z Alexander Kabaev kan@FreeBSD.org 2009-04-17T21:10:53Z urn:sha1:099cfe7ae9f581565a6c5e872b425129fc5235f9 More fallout from negative dotdot caching. Negative entries should be removed from and reinserted to proper ncneg list. Reported by: pho Submitted by: kib Approved by: re (kib) 2009-04-15T16:05:25Z Alexander Kabaev kan@FreeBSD.org 2009-04-15T16:05:25Z urn:sha1:dc34b7e1da15f05b04720b182d5d0a23ae0ff48a Properly handle negative cache enries found through dotdot lookups. Submitted by: tor Approved by: re(kib) 2009-04-12T19:41:16Z Alexander Kabaev kan@FreeBSD.org 2009-04-12T19:41:16Z urn:sha1:99e72b7f3b0cb01b7f3d3b4cf8970bebaf7abf60 Replace v_dd vnode pointer with v_cache_dd pointer to struct namecache in directory vnodes. Allow namecache dotdot entry to be created pointing from child vnode to parent vnode if no existing links in opposite direction exist. Use direct link from parent to child for dotdot lookups otherwise. This restores more efficient dotdot caching in NFS filesystems which was lost when vnodes stoppped being type stable. Majority of backporting work for this was done by jhb. Reviewed by: jhb, kib Approved by: re (kib) 2009-04-08T04:30:16Z Stephen McKay mckay@FreeBSD.org 2009-04-08T04:30:16Z urn:sha1:07a63855b62c985a19348981ca07d681f88e095b Approved by: re (kib) 2009-03-27T13:05:29Z Konstantin Belousov kib@FreeBSD.org 2009-03-27T13:05:29Z urn:sha1:17cbd3654e58330f4fdd61d34298ba275bf37e23 When ktracing namei operations, log a result of the __getcwd(). Approved by: re (kensmith)