Source-Changes-HG archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
[src/netbsd-2-0]: src/doc Tickets 424 to 426.
details: https://anonhg.NetBSD.org/src/rev/4336baaacf98
branches: netbsd-2-0
changeset: 561246:4336baaacf98
user: tron <tron%NetBSD.org@localhost>
date: Sun May 30 15:18:55 2004 +0000
description:
Tickets 424 to 426.
diffstat:
doc/CHANGES-2.0 | 36 +++++++++++++++++++++++++++++++++++-
1 files changed, 35 insertions(+), 1 deletions(-)
diffs (47 lines):
diff -r faa769033af9 -r 4336baaacf98 doc/CHANGES-2.0
--- a/doc/CHANGES-2.0 Sun May 30 15:17:46 2004 +0000
+++ b/doc/CHANGES-2.0 Sun May 30 15:18:55 2004 +0000
@@ -1,4 +1,4 @@
-# $NetBSD: CHANGES-2.0,v 1.1.2.238 2004/05/30 15:02:18 tron Exp $
+# $NetBSD: CHANGES-2.0,v 1.1.2.239 2004/05/30 15:18:55 tron Exp $
A complete list of changes from the initial NetBSD 2.0 branch on 28 March 2004
until 2.0 release:
@@ -3380,3 +3380,37 @@
random warnings that zgrep isn't ready to handle.
[mrg, ticket #421]
+sys/miscfs/genfs/layer.h 1.6
+sys/miscfs/genfs/layer_extern.h 1.14
+sys/miscfs/genfs/layer_vnops.c 1.16
+sys/miscfs/nullfs/null_vnops.c 1.29
+sys/miscfs/overlay/overlay_vnops.c 1.12
+sys/miscfs/umapfs/umap_vnops.c 1.26
+
+ Since VOP_UPCALL() has been a long time in coming, add this partial
+ fix for layered-file-removal. It will work for the case of accessing
+ and deleting a file through the layered file system. Accessing via
+ the layer and deleting on the underlying still won't work, nor will
+ accessing via complicated structures (like two umap layers over a
+ given file systems).
+ We still need VOP_UPCALL(), but this is better than things were before.
+ This patch has been discussed off & on for a while. This incarnation
+ was tested by hannken at netbsd dot org.
+ [wrstuden, ticket #424]
+
+share/man/man9/vnode.9 1.21
+
+ Change comments around locking. While you _can_ grab vnode locks
+ as LK_SHARED, the VOP_ code (in all our existing file systems)
+ was implememted assuming exclusive locking. Use of LK_SHARED
+ is technically a bug and only works right for uni-processor and
+ big-lock SMP systems. Our current file systems will blow up
+ (probably quite subtlely) with LK_SHARED and fine-grained SMP
+ locking.
+ [wrstuden, ticket #425]
+
+share/man/man4/fss.4 1.6
+
+ The fss command appeared -> The fss device appeared
+ [bouyer, ticket #426]
+
Home |
Main Index |
Thread Index |
Old Index