[src/trunk]: src/share/man/man9 Explain large write transaction bug in WAPBL.

[riastradh <riastradh%NetBSD.org@localhost>]

details:   https://anonhg.NetBSD.org/src/rev/b2ccdb3046cb
branches:  trunk
changeset: 807176:b2ccdb3046cb
user:      riastradh <riastradh%NetBSD.org@localhost>
date:      Tue Mar 31 00:03:58 2015 +0000

description:
Explain large write transaction bug in WAPBL.

diffstat:

 share/man/man9/wapbl.9 |  14 +++++++++++++-
 1 files changed, 13 insertions(+), 1 deletions(-)

diffs (28 lines):

diff -r 74719faa1245 -r b2ccdb3046cb share/man/man9/wapbl.9
--- a/share/man/man9/wapbl.9    Mon Mar 30 16:57:40 2015 +0000
+++ b/share/man/man9/wapbl.9    Tue Mar 31 00:03:58 2015 +0000
@@ -1,4 +1,4 @@
-.\"    $NetBSD: wapbl.9,v 1.8 2015/03/27 21:53:50 riastradh Exp $
+.\"    $NetBSD: wapbl.9,v 1.9 2015/03/31 00:03:58 riastradh Exp $
 .\"
 .\" Copyright (c) 2015 The NetBSD Foundation, Inc.
 .\" All rights reserved.
@@ -453,6 +453,18 @@
 .Nm
 in the transaction.
 .Pp
+.Nm
+mainly supports write-ahead, and has only limited support for rolling
+back operations, in the form of
+.Fn wapbl_register_inode
+and
+.Fn wapbl_unregister_inode .
+Consequently, for example, large writes appending to a file, which
+requires multiple disk block allocations and an inode update, must
+occur in a single transaction -- there is no way to roll back the disk
+block allocations if the write fails in the middle, e.g. because of a
+fault in the middle of the user buffer.
+.Pp
 There is only one
 .Nm
 transaction for each file system at any given time, and only one