This fixes the #include's for the macvtap build detection.
Signed-off-by: Stefan Berger stef...@us.ibm.com
Index: libvirt-acl/configure.ac
===
--- libvirt-acl.orig/configure.ac
+++ libvirt-acl/configure.ac
@@ -1849,7 +1849,8 @@
On Mon, Mar 08, 2010 at 09:46:14PM +0100, Soren Hansen wrote:
If a special cache strategy for a disk has been specified in a domain
definition, but no driverName has been set, virDomainGetXMLDesc will not
include the driver tag at all.
This patch makes sure that the driver tag is included if
libvir-list-boun...@redhat.com wrote on 03/08/2010 07:29:42 PM:
Using a bridge to connect a qemu NIC to a host interface offers a fair
amount of flexibility to reconfigure the host without restarting the VM.
For example, if the bridge connects host interface eth0 to the qemu tap0
interface,
On Tue, Mar 09, 2010 at 07:30:56AM -0500, Stefan Berger wrote:
This fixes the #include's for the macvtap build detection.
Signed-off-by: Stefan Berger stef...@us.ibm.com
Index: libvirt-acl/configure.ac
===
---
libvir-list-boun...@redhat.com
Hi,
Debian's (and therefore likely the derivatives') kernel headers are
lacking the typdef for sa_family_t in linux/socket.h so the
compilation of the macvtap check fails. Additionally including
sockaddr.h fixes this and doesn't hurt the other cases where
On Mon, Mar 08, 2010 at 10:01:25PM -0500, Laine Stump wrote:
A few more non-literal format strings in error log messages have crept
in. Fix them in the standard way - turn the format string into %s
with the original string as the arg.
---
src/qemu/qemu_conf.c|2 +-
Hi,
Debian's (and therefore likely the derivatives') kernel headers are
lacking the typdef for sa_family_t in linux/socket.h so the
compilation of the macvtap check fails. Additionally including
sockaddr.h fixes this and doesn't hurt the other cases where there's
also a typedef in socket.h since
Signed-off-by: Jiri Denemark jdene...@redhat.com
---
src/util/processinfo.c |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/src/util/processinfo.c b/src/util/processinfo.c
index 5ae19f8..97ee779 100644
--- a/src/util/processinfo.c
+++ b/src/util/processinfo.c
@@
On 03/09/2010 08:54 AM, Jiri Denemark wrote:
The nodeinfo structure wasn't initialized in qemu driver and with the
recent change in CPU topology parsing, old value of nodeinfo-sockets
could be used and incremented giving totally bogus results.
Let's just wipe the structure completely.
On 03/09/2010 06:59 AM, Jiri Denemark wrote:
+++ b/src/util/processinfo.c
@@ -130,31 +130,31 @@ realloc:
Is there a policy on indentation of labels? By default, emacs starts
labels in column 2, so that utilities that expect only function names to
appear in
On 03/08/2010 07:19 PM, Eric Blake wrote:
On 03/05/2010 12:06 PM, Chris Lalancette wrote:
+#define CPU_SYS_PATH /sys/devices/system/cpu
+if (virAsprintf(path, %s/cpu%d/topology/thread_siblings,
CPU_SYS_PATH,
+cpu) 0) {
Where is the documentation about what this
Hi Matthias,
On Mon, Mar 08, 2010 at 08:28:07PM +0100, Matthias Bolte wrote:
2010/3/8 Dimitris Kalogeras d.kaloge...@noc.ntua.gr:
Hi *,
Apologies for cross posting.
I have installed the the libvirt and virt-toolss in an ubuntu karmic
9.10. I am trying to access-manage an ESX 3.5i via
Eric Blake wrote:
...
+
+static int parse_socket(int cpu)
+{
Several tools (such as ctag, or even more simply, 'git grep ^func')
work better if all function implementations are listed with split lines,
such that the function name starts at the first column:
static int
parse_socket(int
The nodeinfo structure wasn't initialized in qemu driver and with the
recent change in CPU topology parsing, old value of nodeinfo-sockets
could be used and incremented giving totally bogus results.
Let's just wipe the structure completely.
Signed-off-by: Jiri Denemark jdene...@redhat.com
---
Eric Blake wrote:
* docs/hacking.html.in (committers): Fix spelling and grammar.
---
Per the very text I'm editing, this counts as trivial :)
docs/hacking.html.in | 26 +++---
1 files changed, 15 insertions(+), 11 deletions(-)
...
Applied, with these additional
On 02/26/2010 11:49 AM, Chris Lalancette wrote:
On 02/26/2010 01:29 PM, Eric Blake wrote:
Ever since I first started playing with libvirt last month, 'make check'
has been failing for me on the same test:
$ VIR_TEST_DEBUG=2 make -C tests check TESTS=daemon-conf
...
1) corrupted config
Eric Blake wrote:
AUTHORS |5 +
Thanks. Pushed.
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
Eric Blake wrote:
* doc/hacking.html.in (preprocessor): New section to document
recently-discussed style issues.
This patch assumes that we are willing to mass-update preprocessor
indentation via Jim's cppi tool.
Thanks for writing that.
Obviously I would not object.
Daniel Veillard ok'd it,
On 03/08/2010 05:29 PM, Eric Blake wrote:
On 03/05/2010 12:06 PM, Chris Lalancette wrote:
The current code for nodeinfo is pretty naive
about socket and thread information. To determine the
sockets, it just takes the number of cpus and divides
by the number of cores. For the thread count,
On Tue, Mar 09, 2010 at 08:35:09AM -0700, Eric Blake wrote:
On 03/09/2010 06:59 AM, Jiri Denemark wrote:
+++ b/src/util/processinfo.c
@@ -130,31 +130,31 @@ realloc:
Is there a policy on indentation of labels? By default, emacs starts
start on first
On 03/09/2010 09:03 AM, Jim Meyering wrote:
The AUTHORS files indicates the list of people with commit acces right
who can actually merge the patches.
-The general rule for commiting patches is to make sure it has been reviewed
-properly in the mailing-list first, usually if a couple of
Eric Blake wrote:
On 03/09/2010 09:03 AM, Jim Meyering wrote:
The AUTHORS files indicates the list of people with commit acces right
who can actually merge the patches.
-The general rule for commiting patches is to make sure it has been reviewed
-properly in the mailing-list first, usually
On 03/09/2010 09:22 AM, Jim Meyering wrote:
Eric Blake wrote:
On 03/09/2010 09:03 AM, Jim Meyering wrote:
The AUTHORS files indicates the list of people with commit acces right
who can actually merge the patches.
-The general rule for commiting patches is to make sure it has been reviewed
On 03/09/2010 08:47 AM, Jim Meyering wrote:
Eric Blake wrote:
...
+
+static int parse_socket(int cpu)
+{
Several tools (such as ctag, or even more simply, 'git grep ^func')
work better if all function implementations are listed with split lines,
such that the function name starts at the
The nodeinfo structure wasn't initialized in qemu driver and with the
recent change in CPU topology parsing, old value of nodeinfo-sockets
could be used and incremented giving totally bogus results.
Let's just wipe the structure completely.
Yeah, this makes sense.
ACK
Thanks,
Resending patch 2 as attachment.
Sharadha
-Original Message-
From: Matthias Bolte [mailto:matthias.bo...@googlemail.com]
Sent: 05 March 2010 21:42
To: Sharadha Prabhakar (3P)
Cc: libvir-list@redhat.com; Ewan Mellor
Subject: Re: [libvirt] [PATCH 2/2] Addition of XenAPI support to libvirt
+++ b/src/util/processinfo.c
@@ -130,31 +130,31 @@ realloc:
Is there a policy on indentation of labels? By default, emacs starts
labels in column 2, so that utilities that expect only function names to
appear in column 1 behave in a more predictable
Eric Blake wrote:
On 03/09/2010 09:22 AM, Jim Meyering wrote:
Eric Blake wrote:
On 03/09/2010 09:03 AM, Jim Meyering wrote:
The AUTHORS files indicates the list of people with commit acces right
who can actually merge the patches.
-The general rule for commiting patches is to make sure it
It is a bad idea to call gettext on an already-translated
string. In cases where a string must be translated separately
from where it is exposed to xgettext, the gettext manual
recommends the idiom of N_() wrapping gettext_noop for
marking the string.
* src/internal.h (N_): Fix definition to
* cfg.mk (sc_preprocessor_indentation): New syntax-check rule.
---
This has no mercy if cppi is not installed. Should I rework
it to give a nicer warning in that case, especially since
cppi is not in FC12?
cfg.mk |7 +++
1 files changed, 7 insertions(+), 0 deletions(-)
diff --git
Eric Blake wrote:
It is a bad idea to call gettext on an already-translated
string. In cases where a string must be translated separately
from where it is exposed to xgettext, the gettext manual
recommends the idiom of N_() wrapping gettext_noop for
marking the string.
* src/internal.h
On Tue, Mar 9, 2010 at 5:27 AM, Stefan Berger stef...@us.ibm.com wrote:
The problem I guess is that the underlying interface can disappear at any
time due to hotplug leaving you with a race condition at other times as well
until a watcher thread detects the change and can act, no?
Right. I
When starting a VM 3 times, on the 3rd time the number of sockets is
completely out of control. Initialize the socket numbers to zero so the
cpu affinity does not occur due to max cpus being completely out of
range and the VM can start.
Signed-off-by: Stefan Berger stef...@us.ibm.com
diff --git
Eric Blake wrote:
* cfg.mk (sc_preprocessor_indentation): New syntax-check rule.
---
This has no mercy if cppi is not installed. Should I rework
it to give a nicer warning in that case, especially since
cppi is not in FC12?
Yes, please.
I doubt that requiring people to build from
source
Hi there,
I trying to enable domain events callbacks under a cross compiled (mingwin
under fedora12 env) libvirt 0.7.4 (I haven't been able to compile more recent
versions under mingwin) dll. Here a little sample of code I'm trying to write :
#include stdafx.h
#include windows.h
#include
On 03/09/2010 10:34 AM, Stefan Berger wrote:
When starting a VM 3 times, on the 3rd time the number of sockets is
completely out of control. Initialize the socket numbers to zero so the
cpu affinity does not occur due to max cpus being completely out of
range and the VM can start.
ACK.
Eric Blake wrote:
On 03/09/2010 10:02 AM, Eric Blake wrote:
* cfg.mk (sc_preprocessor_indentation): New syntax-check rule.
I don't see 1/2 in the archives yet; perhaps it was held up by
moderation, because it is big? (400k)
That was it. I approved it.
At any rate, I've pushed it to my
Eric Blake wrote:
This has no mercy if cppi is not installed. Should I rework
it to give a nicer warning in that case, especially since
cppi is not in FC12?
Please make it print a warning when cppi is not available,
so that people who run make syntax-check know when this check
is being
On 03/09/2010 01:14 PM, Eric Blake wrote:
On 03/09/2010 10:34 AM, Stefan Berger wrote:
When starting a VM 3 times, on the 3rd time the number of sockets is
completely out of control. Initialize the socket numbers to zero so the
cpu affinity does not occur due to max cpus being completely out
Eric Blake wrote:
On 03/09/2010 10:02 AM, Eric Blake wrote:
* cfg.mk (sc_preprocessor_indentation): New syntax-check rule.
I don't see 1/2 in the archives yet; perhaps it was held up by
moderation, because it is big? (400k)
At any rate, I've pushed it to my cloned repo, if that makes life
On 03/03/2010 01:47 PM, Eric Blake wrote:
Based on a warning from coverity. The safe* functions
guarantee complete transactions on success, but don't guarantee
freedom from failure.
snip
diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c
index b7b2e09..d9cae8b 100644
---
On 03/03/2010 03:58 PM, Ed Swierk wrote:
The udev node device driver relies on the device symlink in sysfs to
populate the node device's parent property. Loopback, vlan and bridge
network interfaces don't have a real hardware parent but could be useful
to include as node devices anyway (and
On 03/09/2010 11:53 AM, Chris Lalancette wrote:
if (priv-errfd != -1) {
-saferead(priv-errfd, errout, sizeof(errout));
+if (saferead(priv-errfd, errout, sizeof(errout)) 0) {
+virReportSystemError(errno, %s,
+
On 03/03/2010 11:52 AM, Eric Blake wrote:
Otherwise, a malicious packet could cause a DoS via spurious
out-of-memory failure.
* src/uml/uml_driver.c (umlMonitorCommand): Validate that incoming
data is reliable before using it to allocate/dereference memory.
Don't report bogus errno on short
On 03/09/2010 02:09 PM, Eric Blake wrote:
remoteDriverLock(priv);
if (fds[1].revents) {
DEBUG0(Woken up from poll by other thread);
-saferead(priv-wakeupReadFD, ignore, sizeof(ignore));
+ignore_value (saferead(priv-wakeupReadFD, ignore,
As pointed out by eblake, I made a real hash of the
nodeinfo code with commit
aa2f6f96ddd7a57011c3d25586d588100651feb2. This patch
cleans it up:
1) Do more work at compile time instead of runtime (minor)
2) Properly handle the hex digits that come from
On 03/09/2010 01:48 PM, Jim Meyering wrote:
Eric Blake wrote:
On 03/09/2010 10:02 AM, Eric Blake wrote:
* cfg.mk (sc_preprocessor_indentation): New syntax-check rule.
I don't see 1/2 in the archives yet; perhaps it was held up by
moderation, because it is big? (400k)
At any rate, I've
Chris Lalancette wrote:
...
commit 4d3b7e06bc8e2e76584da369406df0909c63f2d4
Author: Eric Blake ebl...@redhat.com
Date: Tue Mar 9 09:33:48 2010 -0700
build: consistently indent preprocessor directives
* global: patch created by running:
for f in $(git ls-files '*.[ch]') ; do
On 03/09/2010 12:30 PM, Chris Lalancette wrote:
1) Do more work at compile time instead of runtime (minor)
2) Properly handle the hex digits that come from
/sys/devices/system/cpu/cpu*/topology/thread_siblings
3) Fix up some error paths that could cause SEGV
4) Used unsigned's for the cpu
On 03/09/2010 12:43 PM, Jim Meyering wrote:
I'll go ahead and push this if no one speaks up soon.
Rather than downloading and applying the patch, I ran the command
and committed as you.
By this, you mean patch 1/2, correct? I just want to make sure
Yes. The one induced by the for loop
libvir-list-boun...@redhat.com wrote on 03/08/2010 07:29:59 PM:
[image removed]
[libvirt] [RFC][PATCH v2 2/2] Automatically recreate macvtap
interface and reconnect to qemu
Ed Swierk
to:
libvir-list
03/08/2010 07:45 PM
Sent by:
libvir-list-boun...@redhat.com
(See the
Since cppi is not part of Fedora Core 12, the check is conditional:
without cppi, running 'make syntax-check' merely warns:
$ make sc_preprocessor_indentation
preprocessor_indentation
maint.mk: skipping test sc_preprocessor_indentation: cppi not installed
* cfg.mk (sc_preprocessor_indentation):
On Tue, Mar 9, 2010 at 10:55 AM, Eric Blake ebl...@redhat.com wrote:
Where is udev_device_get_syspath declared? I'm assuming it's generated,
since 'git grep' didn't find it. The only concern I have is whether it
sanely reacts to a NULL argument, but since I couldn't find it, I'm not
sure.
On 03/09/2010 12:17 PM, Chris Lalancette wrote:
if (nbytes sizeof res) {
-virReportSystemError(errno,
- _(incomplete reply %s),
- cmd);
+virReportSystemError(0, _(incomplete reply %s), cmd);
+
On 03/09/2010 02:31 PM, Ed Swierk wrote:
On Tue, Mar 9, 2010 at 10:55 AM, Eric Blake ebl...@redhat.com wrote:
Where is udev_device_get_syspath declared? I'm assuming it's generated,
since 'git grep' didn't find it. The only concern I have is whether it
sanely reacts to a NULL argument, but
On 03/09/2010 04:31 PM, Eric Blake wrote:
On 03/09/2010 12:17 PM, Chris Lalancette wrote:
if (nbytes sizeof res) {
-virReportSystemError(errno,
- _(incomplete reply %s),
- cmd);
+
Hi there,
I am using libvirt+qemu/kvm.
For a guest VM that is running, I could change its cdrom media by running the
following virsh command:
virsh attach-disk vm-id path-to-iso-file hdc --driver file --type cdrom
--mode readonly
This will change the media in the cdrom for the guest VM.
Once I
* src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Report
failures on pipe used for wakeup.
Reported by Chris Lalancette.
---
I think this is what you had in mind for additional error
checking. However, in practice, will one-byte reads and
writes on fd's created by pipe() ever encounter
Based on a warning from coverity. The safe* functions
guarantee complete transactions on success, but don't guarantee
freedom from failure.
* src/util/util.h (saferead, safewrite, safezero): Add
ATTRIBUTE_RETURN_CHECK.
* src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Ignore
some
On 03/09/2010 03:25 PM, Eric Blake wrote:
* src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Report
failures on pipe used for wakeup.
Reported by Chris Lalancette.
---
I think this is what you had in mind for additional error
checking. However, in practice, will one-byte reads
* src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Report
failures on pipe used for wakeup.
Reported by Chris Lalancette.
---
src/remote/remote_driver.c | 28
1 files changed, 24 insertions(+), 4 deletions(-)
diff --git a/src/remote/remote_driver.c
Based on a warning from coverity. The safe* functions
guarantee complete transactions on success, but don't guarantee
freedom from failure.
* src/util/util.h (saferead, safewrite, safezero): Add
ATTRIBUTE_RETURN_CHECK.
* src/remote/remote_driver.c (remoteIO, remoteIOEventLoop): Ignore
some
On 03/09/2010 01:55 PM, Eric Blake wrote:
On 03/03/2010 03:58 PM, Ed Swierk wrote:
The udev node device driver relies on the device symlink in sysfs to
populate the node device's parent property. Loopback, vlan and bridge
network interfaces don't have a real hardware parent but could be useful
On 03/05/2010 10:41 AM, Dave Allan wrote:
On 03/04/2010 07:49 PM, Ed Swierk wrote:
On Thu, Mar 4, 2010 at 10:31 AM, Dave Allandal...@redhat.com wrote:
I reworked udevSetupSystemDev into this style (which also fixes the
bug you
pointed out that it didn't properly free resources on error).
64 matches
Mail list logo
