fs: avoid softlockups in s_inodes iterators
authorEric Sandeen <sandeen@redhat.com>
Fri, 6 Dec 2019 16:54:23 +0000 (10:54 -0600)
committerAl Viro <viro@zeniv.linux.org.uk>
Wed, 18 Dec 2019 05:03:01 +0000 (00:03 -0500)
commit04646aebd30b99f2cfa0182435a2ec252fcb16d0
tree5b820d440100bf1c9ce8db4be85d73ceb503cdca
parente0ff126ee7ad405c1ef531f9f3db92929de4f20f
fs: avoid softlockups in s_inodes iterators

Anything that walks all inodes on sb->s_inodes list without rescheduling
risks softlockups.

Previous efforts were made in 2 functions, see:

c27d82f fs/drop_caches.c: avoid softlockups in drop_pagecache_sb()
ac05fbb inode: don't softlockup when evicting inodes

but there hasn't been an audit of all walkers, so do that now.  This
also consistently moves the cond_resched() calls to the bottom of each
loop in cases where it already exists.

One loop remains: remove_dquot_ref(), because I'm not quite sure how
to deal with that one w/o taking the i_lock.

Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
fs/drop_caches.c
fs/inode.c
fs/notify/fsnotify.c
fs/quota/dquot.c