Date: Tue, 03 Sep 2019 14:06:41 -0000 From: Alan Somers <asomers@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r346078 - head/sys/kern Message-ID: <201904101302.x3AD2XKg012359@repo.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: asomers Date: Wed Apr 10 13:02:33 2019 New Revision: 346078 URL: https://svnweb.freebsd.org/changeset/base/346078 Log: fix cache_lookup's documentation cache_lookup's documentation got dislocated by r324378. Relocate and expand it. Reviewed by: jhb, kib MFC after: 2 weeks Sponsored by: The FreeBSD Foundation Modified: head/sys/kern/vfs_cache.c Modified: head/sys/kern/vfs_cache.c ============================================================================== --- head/sys/kern/vfs_cache.c Wed Apr 10 10:50:48 2019 (r346077) +++ head/sys/kern/vfs_cache.c Wed Apr 10 13:02:33 2019 (r346078) @@ -150,7 +150,7 @@ struct namecache_ts { * Names found by directory scans are retained in a cache * for future reference. It is managed LRU, so frequently * used names will hang around. Cache is indexed by hash value - * obtained from (vp, name) where vp refers to the directory + * obtained from (dvp, name) where dvp refers to the directory * containing name. * * If it is a "negative" entry, (i.e. for a name that is known NOT to @@ -1129,23 +1129,6 @@ cache_lookup_dot(struct vnode *dvp, struct vnode **vpp return (-1); } -/* - * Lookup an entry in the cache - * - * Lookup is called with dvp pointing to the directory to search, - * cnp pointing to the name of the entry being sought. If the lookup - * succeeds, the vnode is returned in *vpp, and a status of -1 is - * returned. If the lookup determines that the name does not exist - * (negative caching), a status of ENOENT is returned. If the lookup - * fails, a status of zero is returned. If the directory vnode is - * recycled out from under us due to a forced unmount, a status of - * ENOENT is returned. - * - * vpp is locked and ref'd on return. If we're looking up DOTDOT, dvp is - * unlocked. If we're looking up . an extra ref is taken, but the lock is - * not recursively acquired. - */ - static __noinline int cache_lookup_nomakeentry(struct vnode *dvp, struct vnode **vpp, struct componentname *cnp, struct timespec *tsp, int *ticksp) @@ -1229,6 +1212,42 @@ out_no_entry: return (0); } +/** + * Lookup a name in the name cache + * + * # Arguments + * + * - dvp: Parent directory in which to search. + * - vpp: Return argument. Will contain desired vnode on cache hit. + * - cnp: Parameters of the name search. The most interesting bits of + * the cn_flags field have the following meanings: + * - MAKEENTRY: If clear, free an entry from the cache rather than look + * it up. + * - ISDOTDOT: Must be set if and only if cn_nameptr == ".." + * - tsp: Return storage for cache timestamp. On a successful (positive + * or negative) lookup, tsp will be filled with any timespec that + * was stored when this cache entry was created. However, it will + * be clear for "." entries. + * - ticks: Return storage for alternate cache timestamp. On a successful + * (positive or negative) lookup, it will contain the ticks value + * that was current when the cache entry was created, unless cnp + * was ".". + * + * # Returns + * + * - -1: A positive cache hit. vpp will contain the desired vnode. + * - ENOENT: A negative cache hit, or dvp was recycled out from under us due + * to a forced unmount. vpp will not be modified. If the entry + * is a whiteout, then the ISWHITEOUT flag will be set in + * cnp->cn_flags. + * - 0: A cache miss. vpp will not be modified. + * + * # Locking + * + * On a cache hit, vpp will be returned locked and ref'd. If we're looking up + * .., dvp is unlocked. If we're looking up . an extra ref is taken, but the + * lock is not recursively acquired. + */ int cache_lookup(struct vnode *dvp, struct vnode **vpp, struct componentname *cnp, struct timespec *tsp, int *ticksp)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201904101302.x3AD2XKg012359>