> On 30. Mar 2017, at 17:20, Christos Zoulas <chris...@astron.com> wrote: > > In article <534552fb-af29-4219-8390-7514a2665...@eis.cs.tu-bs.de>, > J. Hannken-Illjes <hann...@eis.cs.tu-bs.de> wrote: >> Currently vfs_busy() / vfs_unbusy() get used to >> >> - Enter/leave a critical section against unmounting >> >> - Add a reference to the mount >> >> - return the next mount from the mountlist on error >> >> Plan is to untangle the mountlist processing from vfs_busy() / vfs_unbusy() >> and add an iterator for the mountlist: >> >> - void mountlist_iterator_init(mount_iterator_t **) >> to initialize an iterator in mounted order. >> >> - void mountlist_iterator_init_reverse(mount_iterator_t **) >> to initialize an iterator in reverse mounted order. >> >> - void mountlist_iterator_destroy(mount_iterator_t *) >> to destroy an iterator aquired with one of the above. >> >> - struct mount *mountlist_iterator_next(mount_iterator_t *, int) >> to retrieve the next mount from an iterator or NULL if done. >> The mount is returned vfs_busy'd, caller has to vfs_unbusy. >> Accepts a flag MNT_NOWAIT to skip mounts whose unmount is in progress. >> >> - struct mount *_mountlist_next(struct mount *) >> a lockless variant to be used from DDB only. >> >> Diffs are here: https://www.netbsd.org/~hannken/ml_iterator/ >> >> Comments or objections anyone? > > LGTM, but can't all the copies of the dostatvfs iteration be merged > together and utilize a converter function to handle the data massaging?
Looks like vfs_syscalls.c::do_sys_getvfsstat() - iterate over the mountlist and call a copyfn() function on the stat buffer - already does it. -- J. Hannken-Illjes - hann...@eis.cs.tu-bs.de - TU Braunschweig (Germany)
