[ovs-dev] [PATCH] xenserver: Merge changes to scripts from XenServer 5.6.100.
andrew at fivevalve.com
Fri Feb 11 20:39:51 UTC 2011
On 2/11/11 12:17 PM, Ben Pfaff wrote:
> On Fri, Feb 11, 2011 at 11:44:11AM -0800, Andrew Evans wrote:
>> On 2/11/11 10:40 AM, Ben Pfaff wrote:
>>> This commit should update the documentation to remove references to
>>> XenServer 5.5 and 5.6 support. At least INSTALL.Linux and
>>> INSTALL.XenServer need to be updated.
>> Justin was doing that, but I'll roll his changes into my patch.
> Unless you've already done that, don't bother--we might as well use
> Justin's patch directly, ideally applying it before yours but really
> it doesn't matter much.
I've already incorporated this part of his patch:
INSTALL.Linux | 2 +-
INSTALL.XenServer | 5 ++---
vswitchd/vswitch.xml | 2 +-
xenserver/README | 6 +++---
4 files changed, 7 insertions(+), 8 deletions(-)
diff --git a/INSTALL.Linux b/INSTALL.Linux
index 9d69bd8..8ba73ed 100644
@@ -3,7 +3,7 @@
This document describes how to build and install Open vSwitch on a
generic Linux host. If you want to install Open vSwitch on a Citrix
-XenServer version 5.5.0, see INSTALL.XenServer instead.
+XenServer, see INSTALL.XenServer instead.
This version of Open vSwitch may be built manually with "configure"
and "make", as described below. You may also build Debian packages by
diff --git a/INSTALL.XenServer b/INSTALL.XenServer
index 33b7f8b..e7be494 100644
@@ -5,8 +5,7 @@ This document describes how to build and install Open
vSwitch on a
Citrix XenServer host. If you want to install Open vSwitch on a
generic Linux host, see INSTALL.Linux instead.
-These instructions have been tested with XenServer versions 5.5.0,
-5.5.0-24648p (Update 1), 5.5.0-25727p (Update 2), and 5.6.0.
+These instructions have been tested with XenServer 5.6 FP1.
Building Open vSwitch for XenServer
@@ -38,7 +37,7 @@ where:
name of the Open vSwitch tarball, e.g. 0.90.0.
<Xen kernel version> is the version number of the Xen kernel,
- e.g. 2.6.18-128.1.1.el5.xs18.104.22.1687.1006xen. This version number
+ e.g. 22.214.171.124-0.7.1.xs126.96.36.1997.170586xen. This version number
appears as the name of a directory in /lib/modules inside the VM.
It always ends in "xen".
diff --git a/vswitchd/vswitch.xml b/vswitchd/vswitch.xml
index f4515f4..c80e8cd 100644
@@ -244,7 +244,7 @@
The version of the system identified by <ref
- e.g. <code>5.5.0-24648p</code> on XenServer 5.5.0 build 24648.
+ e.g. <code>5.6.100-39265p</code> on XenServer 5.6.100 build
System integrators are responsible for choosing and setting an
diff --git a/xenserver/README b/xenserver/README
index 7da3ac2..3110aab 100644
@@ -79,8 +79,8 @@ files are:
This is uuid.py from Python 2.5. It is installed into the
- Open vSwitch RPM because XenServer 5.5 and 5.6 use Python 2.4,
- which do not have uuid.py.
+ Open vSwitch RPM because XenServer uses Python 2.4, which does
+ not have uuid.py.
To install, build the Open vSwitch RPM with a command like this:
@@ -94,7 +94,7 @@ question and reboot them. (The openvswitch-dbg
package that is also
produced need not be installed, but it is harmless to do so.)
-Copyright (C) 2009, 2010 Nicira Networks, Inc.
+Copyright (C) 2009, 2010, 2011 Nicira Networks, Inc.
Copying and distribution of this file, with or without modification,
are permitted in any medium without royalty provided the copyright
>>> * etc_xensource_scripts_vif has just whitespace differences.
>>> Can you fix those up?
>> Those are the only lines that use tabs for indentation. Justin cleaned
>> those up in a previous commit. Shouldn't we ask Citrix to merge those
> I guess I don't feel particularly eager about pushing whitespace
> changes upstream. It takes time, for questionable benefit. I'd
> rather just adopt their version, since the whitespace changes make no
> functional difference.
>>> * The following difference in InterfaceReconfigure.py appears
>>> to be for compatibility with pre-5.6.100 that wouldn't have
>>> a tunnel_access_PIF_of field, so I think that we could
>>> revert this difference now.
>>> @@ -826,7 +871,8 @@
>>> # Tunnel PIFs
>>> def pif_is_tunnel(pif):
>>> - return len(db().get_pif_record(pif)['tunnel_access_PIF_of']) > 0
>>> + rec = db().get_pif_record(pif)
>>> + return rec.has_key('tunnel_access_PIF_of') and len(rec['tunnel_access_PIF_of']) > 0
>> This breaks the unit test with a KeyError. I guess the test just needs
>> to be updated?
> Yes, I imagine that the PIF records in the test just need to have
> empty tunnel_access_PIF_of added.
Ok to push this?
More information about the dev