On 12/23/2011 12:57 PM, Osier Yang wrote:
> IMHO "device=lun" is the right way to go here, per we want the
> the device is exposed to guest as a LUN but not a normal disk.
> But it seems for Laine's patch, it's also right to use "type=lun",
> as it tries to disable/enable SG_IO for normal disk?
Th
On 12/23/2011 12:15 AM, Daniel P. Berrange wrote:
The 'type' attribute refers to how the host emulator deals with the
disk.
The 'device' attribute refers to what type of device hardware is exposed
to the guest.
What we're doing here is controlling whether the host emulator allows
SG_IO. The gue
Here is a revised version of the virtio-scsi proposal. There's actually
not too much left intact from v1. :)
The main simplification is in how SCSI hosts can be addressed in a stable
manner.
SCSI controller models
==
Existing controller models are "auto", "buslogic", "lsilo
On 12/22/2011 03:20 PM, Eric Blake wrote:
> Usage:
>
>
>
>
I still don't like the proposed XML. We need to get consensus on where
to put it, with my proposal being:
Right, other types of drivers might include a userspace NFS client, or
some kind of "artificial"
On 11/28/2011 03:01 PM, Dave Allan wrote:
We'd then extend the domain disk XML to take a pool/volume
definition.
I like the idea. It could be used for other things:
- QEMU usermode iSCSI backend, including storing iSCSI secrets outside
the domain XML. In this case, multiple domains will usu
On 11/26/2011 02:14 PM, Bastien ROUCARIES wrote:
Malloc page than realloc to smaller does not work ?
There is no constraint that realloc(X, N) returns X, even if N is
smaller than the previous allocation size.
Paolo
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/m
> Indeed; Linux has posix_memalign, and mingw never runs the io helper
> (although it does compile it, hence the #if). If gnulib would give
> us posix_memalign on mingw, we could nuke this #if altogether.
That's pretty difficult (unless you also add a posix_memalign_free)
because at the time posi
When aligning you need to clear the bits in the mask and leave the
others aside. Likely this code has never run, and will never run.
Signed-off-by: Paolo Bonzini
---
src/util/iohelper.c |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/src/util/iohelper.c b/src/util
On 11/11/2011 03:32 PM, Daniel P. Berrange wrote:
+%makeinstall SYSTEMD_UNIT_DIR=%{?buildroot:%{buildroot}}%{_unitdir}
Is this correct since DESTDIR is prepended to it? Shouldn't it just be
just %{_unitdir}? Otherwise looks good; it would be nicer to make it a
configure option, but that's n
Hi all,
QEMU 1.0 test day will be held on November 7th. Any kind of testing
that you can contribute will be appreciated---including testing your
favorite hypervisor, desktop virtualization or management stack together
with the new release.
You are also welcome to add testing hints at
http:
On 10/27/2011 11:19 AM, Daniel P. Berrange wrote:
where host selects the qdev parent device, while channel/target/lun
are passed as qdev properties (the QEMU names are respectively
channel, scsi-id, lun).
The problem here is that we already have an address type that is used
for SCSI - typ
On 10/26/2011 06:15 PM, Dave Allan wrote:
> In NPIV, a virtual HBA is created using "virsh nodedev-create" and
> passed to the guest. Such virtual adapter does have a stable
> address, namely its WWN. As such, it can be addressed simply by
> generalizing the kind of source address that can
Hi all,
let's kick off the discussion on what changes are needed in domain XML
for more complete SCSI support.
There are three relevant topics:
1) providing channel/target/lun addresses for SCSI disks;
2) supporting LUN passthrough;
3) supporting SCSI host passthrough.
A fourth topic is s
On 10/11/2011 06:02 PM, Rick Vernam wrote:
On Tuesday 11 October 2011 10:45:14 Paolo Bonzini wrote:
Hi all, please go here to cast your vote on when QEMU 1.0 test day will
be help.
http://www.doodle.com/35ae4zk7e4ndb5qq
Having an "official" date will make it easier to put
On 10/04/2011 08:35 PM, Laine Stump wrote:
This resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=730909
When support for setting the qemu disk error policy to 'enospc' was
added, it was inadvertantly as "enospace". This patch corrects that on
the qemu commandline (while retaining the
On 08/17/2011 10:21 AM, Stefan Hajnoczi wrote:
On Wed, Aug 17, 2011 at 5:51 PM, Paolo Bonzini wrote:
following discussions yesterday with Juan Quintela and Marcelo Tosatti, here
is my humble proposal: remove block migration from qemu master. It seems to
me that keeping block migration is
On 07/25/2011 10:22 PM, Eric Blake wrote:
As long as QEMU supports -M ? it should not be a problem. The -M option
is not scraped from -help output.
If -M is going away, then libvirt _should_ scrape -help output to see
whether to use -M or -machine, rather than blindly using -M.
-M going aw
On 07/25/2011 11:12 AM, Richard W.M. Jones wrote:
I guess this may affect the qemu driver in future ... Thanks to
Markus Armbruster for bringing this to my/our attention.
As long as QEMU supports -M ? it should not be a problem. The -M option
is not scraped from -help output.
Paolo
--
lib
On 07/07/2011 04:17 PM, Daniel P. Berrange wrote:
if [ $1 -eq 1 ] ; then
# Initial installation
+/bin/systemctl enable virtlockd.service>/dev/null 2>&1 || :
/bin/systemctl enable libvirtd.service>/dev/null 2>&1 || :
/bin/systemctl enable cgconfig.service>/dev/null 2>&1 ||
On 07/07/2011 11:31 AM, Daniel P. Berrange wrote:
On Wed, Jul 06, 2011 at 07:36:26PM -0600, Jim Fehlig wrote:
Kernel cmdline args can be passed to xen pv domains even when a
bootloader is specified. The current config-to-sxpr mapping
ignores cmdline when bootloader is present.
Since the xend s
On 06/30/2011 07:56 PM, Eric Blake wrote:
> Yes, Windows pipes are that broken.:(
>
> Using socketpair is a possibly good idea, but I would do it on
> libvirtd only. I don't know exactly how libvirtd uses this pipe, but
> perhaps it can be changed to an eventfd-like abstraction that can be
>
On 05/25/2011 05:53 PM, Eric Blake wrote:
> -localtime_r(&creation,&time_info);
> +creation_time_t = creation_longlong;
> +if (creation_time_t != creation_longlong) {
> +vshError(ctl, "%s", _("time_t overflow"));
We shouldn't hit this erro
On 05/21/2011 12:11 AM, Eric Blake wrote:
I think we're safe. It shouldn't matter if thread 2 reads thread 3's
answer (because they are both read-only queries, and should be getting
the same answer; or even if the answers differ, reading the newer answer
is not horrible because it is still an ac
On 05/16/2011 07:30 PM, Eric Blake wrote:
> +while (priv->jobSignals& QEMU_JOB_SIGNAL_BLKSTAT)
> +ignore_value(virCondWait(&priv->signalCond,&vm->lock));
> +
> +priv->jobSignalsData.statDevName = disk->info.alias;
> +priv->jobSignalsData.blockStat = stats;
> +
On 05/20/2011 11:06 AM, Daniel P. Berrange wrote:
So if I'm understanding correctly, the only difference we'll see with
assigning devices to functions, instead of slots, is that hotplug of
individual devices is not possible. From a guest (driver) POV everything
else is 100% functionally unchanged
On 05/11/2011 09:55 AM, Daniel P. Berrange wrote:
Hmm, that's kinda wierd& i'm suprised it works, particularly for LSI
since I thought guest drivers would need support for multifunction too.
For well-behaved {kernel,device,driver}s multifunction should be totally
transparent.
Example from
libxl accepts hpet configuration in its domain info struct. Parse the
domain definition's element in order to set the value.
Signed-off-by: Paolo Bonzini
---
src/libxl/libxl_conf.c |8
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/src/libxl/libxl_conf.c
This patch series add support for enabling or disabling the HPET
in Xen domains. It's split in two because I didn't have means to test
the libxl part.
Paolo Bonzini (2):
xen: parse and generate hpet item in sxpr
libxl: support enabling the HPET
src/libxl/li
leaving the default to "enabled" for compatibility
reasons.
This patch provides support for the hpet item in both sexpr and xm
formats, and translates it to a element.
Signed-off-by: Paolo Bonzini
---
src/xenxs/xen_sxpr.c | 26 +++
src/xenx
On 03/31/2011 01:18 PM, Michal Novotny wrote:
You mean to compare command-line from the XML, i.e. to generate the
command-line from XML and put the tests whether the generated
command-line is the same as mentioned in the *.argv file containing
whole command-line? Like for QEMU?
Yes.
Paolo
--
On 03/31/2011 01:06 PM, Michal Novotny wrote:
> - this would be the third patch in the series, and it would add
> dnsmasqAddHost calls in networkSaveDnsmasqHostsfile
So this patch will be just adding the networkSaveDnsmasqHostsfile() call
to the dnsmasqAddHost() and nothing else ?
And all th
+
+
+
+
+
+
oneOrMore hostname elements seems preferrable.
+if ((fd = open(hostsFile,
You can use fopen and fprintf.
This function also should be in src/util/dnsmasq.c (e.g. with addnhost*
APIs in that file and dnsmasqAddHost as a public API). See the
hostsfile* APIs there.
Then yo
On 03/30/2011 02:57 PM, Michal Novotny wrote:
Thename
+attribute doesn't support commas in it's value and therefore you should
+avoid using them since they are automatically replaced by spaces
No, names with commas should be rejected.
Paolo
--
libvir-list mailing list
libvir-list@redh
On 03/30/2011 01:00 PM, Michal Novotny wrote:
I think you should triage it a bit more, e.g. with strace -ff. Anyway,
there is no hurry of doing this I think.
Well, you mean to use strace on the daemonized process?
Wherever it helps understanding what's happening. :)
Also, I've been testing
On 03/30/2011 12:41 PM, Michal Novotny wrote:
1) /usr/sbin/dnsmasq --strict-order --bind-interfaces
--pid-file=/var/run/libvirt/network/default.pid --conf-file=
--except-interface lo --txt-record="txt-record","some value, which is
something" --addn-hosts=/var/run/libvirt/network/default.hosts
--l
On 03/29/2011 05:08 PM, Michal Novotny wrote:
Ok, however vendor-class and user-class should be used together and
therefore it could be committed into one commit, right? The next commit
could be the dhcp-mac one... right ?
No, they can be one patch series, but they should be separate commits.
On 03/29/2011 01:37 PM, Michal Novotny wrote:
On 03/29/2011 01:16 PM, Paolo Bonzini wrote:
On 03/29/2011 12:52 PM, Michal Novotny wrote:
[snip]
It would be great to:
1) add and tags inside that allow
filtering according to user/vendor classes
Well, I didn't know this is support
On 03/29/2011 12:52 PM, Michal Novotny wrote:
[snip]
It would be great to:
1) add and tags inside that allow
filtering according to user/vendor classes
Well, I didn't know this is supported by DNSMasq but it seems to be
Yes, I am using it. :)
-4, --dhcp-mac=,
Map
On 03/29/2011 12:52 PM, Michal Novotny wrote:
[snip]
It would be great to:
1) add and tags inside that allow
filtering according to user/vendor classes
Well, I didn't know this is supported by DNSMasq but it seems to be
Yes, I am using it. :)
-4, --dhcp-mac=,
Map
On 03/28/2011 06:02 PM, Eric Blake wrote:
On 03/27/2011 02:59 AM, Laine Stump wrote:
On 03/26/2011 08:12 AM, Eric Blake wrote:
Diego reported a bug where virsh tries to initialize a readline
history directory during 'make check' run as root, but fails
because /root was read-only.
It turns out
On 03/29/2011 01:27 AM, Eric Blake wrote:
Paolo, any thoughts on the best approach to take? (I know which way I'm
leaning, but want some feedback before I give away my bias).
Without guessing what your bias is, I also :) prefer to implement
{g,s}et_nonblock_flag functions. It would use eithe
On 03/28/2011 01:34 PM, Michal Novotny wrote:
> Hi Laine,
> thanks for your reply. Comments inline...
>
> On 03/25/2011 10:02 PM, Laine Stump wrote:
>> I haven't had time yet to look at the code in detail, but thought I
>> should send this preliminary commentary.
>>
>> On 03/24/2011 09:58 AM, Mich
On 03/11/2011 05:30 AM, Wen Congyang wrote:
> Yep, makes sense. I'm double checking whether I reliably closed the
> write end of the socket in all paths, and it may boil down to changing
> the child to_always_ sendmsg even if there is no fd to send. I'm still
> thinking about it.
Using S
On 03/10/2011 03:18 AM, Eric Blake wrote:
+if (qemuCaps&& qemuCapsGet(qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD)&&
+priv->monConfig->type == VIR_DOMAIN_CHR_TYPE_UNIX) {
+cmd = virCommandNewArgs(args);
+virCommandSetInputFD(cmd, pipeFD[0]);
+virCo
On 03/09/2011 04:59 PM, Daniel P. Berrange wrote:
> +if (!flags) {
> + if (virDomainSetMemory(dom, kilobytes) != 0) {
> +ret = FALSE;
> + }
Indentation went a little bit wrong here. Also '!flags' could
in fact be '!flags || (flags == VIR_DOMAIN_MEM_LIVE)' s
On 03/04/2011 01:37 AM, Eric Blake wrote:
(virFileOperation): Rename...
(virFileOpenAs): ...and reduce parameters.
As an "external observer" of libvirt development, I must admit
virFileOperation always seemed an extremely heavyweight and clunky
interface. Thanks for cleaning it up!
Paolo
On 03/04/2011 01:37 AM, Eric Blake wrote:
* src/storage/storage_backend.c (createRawFileOpHook): Change
signature.
(struct createRawFileOpHookData): Delete unused struct.
(virStorageBackendCreateRaw): Adjust caller.
* src/qemu/qemu_driver.c (struct fileOpHookData): Delete unused
struct.
(qemudDom
On 03/04/2011 01:37 AM, Eric Blake wrote:
- unsigned int flags ATTRIBUTE_UNUSED)
+ unsigned int flags)
Uhm, you're still not using the flags, you're just using the argument as
a place to write the flags for virFileOperation. Perhaps the "inn
On 03/03/2011 03:09 PM, Daniel Veillard wrote:
ACK, removing should be safe then, but adding will still be
a big problem due to virHashGrow().
Removing _the current_ element is safe, on the other hand removing any
other element used to be safe and now it may cause a dangling pointer
access
On 03/02/2011 06:16 AM, Eric Blake wrote:
On 01/03/2011 06:39 AM, Paolo Bonzini wrote:
On 12/23/2010 05:51 PM, Eric Blake wrote:
In the outgoing direction, we still have to use a unix socket (outgoing
migration is started via a monitor command, and I don't know how to pass
a new fd into
On 02/23/2011 06:18 PM, Ian Jackson wrote:
Because for now you want to be able to have both drivers. Unless,
perhaps, it's somehow possible to select between them at runtime for
the same URI scheme ?
The Xen driver is already using a multitude of "subdrivers" (query
xenstore, query the hyperv
On 02/23/2011 03:56 AM, Jim Fehlig wrote:
Add a new xen driver based on libxenlight [1], which is the primary
toolstack starting with Xen 4.1.0. The driver is stateful, runs
privileged only, and is accessed with libxl:/// URI.
Why not let the user keep xen:/// ?
Paolo
--
libvir-list mailing
On 01/20/2011 02:39 PM, Michal Novotny wrote:
Well, I'm working on support for serial port configuration as serial = [
"/dev/null", "/dev/ttyS0", "/dev/ttyS1" ] to:
1) pass /dev/null to the guest as /dev/ttyS0
2) pass /dev/ttyS0 to the guest as /dev/ttyS1
3) pass /dev/ttyS1 to the guest as /dev/
On 01/20/2011 12:58 PM, root wrote:
+if ((list->type != VIR_CONF_STRING) || (list->str == NULL))
+goto skipport;
+
I believe this should do
xenXMError(VIR_ERR_INTERNAL_ERROR,
_("config value %s was malformed"), name);
goto
(qemudDomainMigratePrepareTunnel):
Replace Unix socket with simpler pipe.
Suggested by Paolo Bonzini.
---
I've since been corrected; qemu_monitor.c has qemuMonitorIOWriteWithFD
that looks like it can be used to avoid exec: on outgoing migrations
by instead passing an open fd over a unix socket, although I hav
On 12/23/2010 05:51 PM, Eric Blake wrote:
In the outgoing direction, we still have to use a unix socket (outgoing
migration is started via a monitor command, and I don't know how to pass
a new fd into qemu for using with an fd:n outgoing migration via the
monitor - it seems like fd is only useful
On 12/23/2010 05:51 PM, Eric Blake wrote:
Since you are at it, you may also want to replace the usage of netcat
>> and Unix sockets with "-incoming fd:" and a one-way pipe.
>
> Not sure it's worth it. Is unix:path any less efficient than creating a
> one-way pipe for use by fd:n? Also, we c
On 12/22/2010 11:27 PM, Eric Blake wrote:
https://bugzilla.redhat.com/show_bug.cgi?id=620363
When using -incoming stdio or -incoming exec:cat, qemu keeps the
stdin fd open long after the migration is complete. Not to
mention that exec:cat is horribly inefficient, by doubling the
I/O and going t
On 12/18/2010 01:10 AM, Eric Blake wrote:
case $ac_cv_search_dlopen:$host in
no:*-*-mingw* | no:*-*-msvc*) ;;
no:*) AC_MSG_ERROR([Unable to find dlopen()]) ;;
'none required:*') DLOPEN_LIBS= ;;
*) DLOPEN_LIBS=$ac_cv_search_dlopen ;;
esac
?
Paolo
--
libvir-list mailing list
l
custom_gnutls_push);
+gnutls_transport_set_pull_function(session, custom_gnutls_pull);
+#endif
+
/* Perform the TLS handshake. */
again:
err = gnutls_handshake (session);
Acked-By: Paolo Bonzini
Paolo
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On 11/12/2010 07:45 PM, Daniel P. Berrange wrote:
On Fri, Nov 12, 2010 at 07:32:05PM +0100, Paolo Bonzini wrote:
On 11/12/2010 05:22 PM, Daniel P. Berrange wrote:
+{
+const char *cmdargv[] = {
+"/lib/udev/scsi_id",
+"--replace-whitespace",
+
On 11/12/2010 05:22 PM, Daniel P. Berrange wrote:
+{
+const char *cmdargv[] = {
+"/lib/udev/scsi_id",
+"--replace-whitespace",
+"--whitelisted",
+"--device", dev,
+NULL
+};
Does this needs to be "/lib/udev/scsi_id -gud " for RHEL5
compatibility?
On 11/12/2010 05:22 PM, Daniel P. Berrange wrote:
src/storage/storage_backend_mpath.c | 20 +++
src/storage/storage_backend_scsi.c| 25
Not related to your patch, but these two drivers are undocumented.
Paolo
--
libvir-list mailing list
lib
On 11/02/2010 07:22 PM, Laine Stump wrote:
1) The element of the network xml will be expanded in the following
ways:
a) can now appear multiple times (although I think only one of
these can/should be allowed to have a element). The bridge
interface (eg virbr0) will be configured with all give
On 10/04/2010 06:31 PM, Justin Clift wrote:
Heh. Two periods in a name could lead to fun. Stuff like:
name="../../../etc/shadow"
Well, you know what I mean. :)
Not if you prohibit / though...
Paolo
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listi
On 10/03/2010 11:33 AM, Richard W.M. Jones wrote:
Indeed. I'm sure we need a whitelist, not a blacklist as suggested by
the other comment. All domains I'd ever want to create would match
the regexp
^[[:alpha:]][-_[:alnum:]]*$
This might break existing users however.
A period in a name is de
On 10/01/2010 04:18 PM, Daniel Veillard wrote:
I would think augmenting the sexpr should be sufficient for this
but the problem is really to find out when the feature is available,
and I don't know how to do this reliably either (except trying and if
there is an identifiable error keep it disa
On 09/10/2010 06:43 PM, Eric Blake wrote:
> * src/Makefile.am (libvirt.def, libvirt_qemu.def): '\}' and '\t'
> are not required by POSIX. Use '}' and literal tab instead.
> (install-data-local): Avoid sed -i.
> * tests/read-bufsiz: Likewise.
> Reported by Mitchell Hashimoto.
> ---
>
> This should
On 07/23/2010 09:34 PM, Eric Blake wrote:
+if (virAsprintf(&path, PCI_SYSFS "devices/%s/physfn", dev->name)< 0) {
Weird spacing around <
No, instead you are suffering from
https://bugzilla.mozilla.org/show_bug.cgi?id=571502
:(
Paolo
--
libvir-list mailing list
libvir-list@redhat.co
On 06/23/2010 01:13 PM, Daniel P. Berrange wrote:
On Wed, Jun 23, 2010 at 01:08:40PM +0200, Paolo Bonzini wrote:
On 06/23/2010 12:50 PM, Daniel P. Berrange wrote:
libvirt_driver_test_la_LDFLAGS = -module -avoid-version
+libvirt_driver_test_la_LIBADD = libvirt.la ../gnulib/lib/libgnu.la
On 06/23/2010 12:50 PM, Daniel P. Berrange wrote:
>libvirt_driver_test_la_LDFLAGS = -module -avoid-version
> +libvirt_driver_test_la_LIBADD = libvirt.la ../gnulib/lib/libgnu.la
I don't think it is right to link to libvirt.la here or with other
driver modules.
The libvirtd daemon links to l
On 04/29/2010 05:41 AM, Eric Blake wrote:
Gnulib guarantees that pthread.h exists, but for now, it is a dummy
header with no support for most pthread_* functions. Modify our
use of pthread to use function checks, rather than header checks,
to determine how much pthread support is present.
Shou
On 04/09/2010 11:30 PM, Eric Blake wrote:
Yeah, having the ability to specify an optional wrapper script, that
receives the name of the normal interpreter and all the arguments the
normal interpreter would have been given, sounds like the ultimate in
flexibility at a minimum of xml.
You can als
ACK
Thanks for the ACK.
However, when I saw the inconsistent spacing above,
(the "break" I added is indented one space less than the "else" above it)
I worried that my patch had a problem, so I went looking for the cause.
There is no problem on my end.
This seems to be coming from your mail cl
On 03/22/2010 02:25 AM, Matthias Bolte wrote:
Correctly disable pthread related code if pthread is not avialable,
in order to get it compile with MinGW on Windows.
What about, instead
#ifndef HAVE_PTHREAD_H
#define pthread_sigmask sigprocmask
#endif
(or actually the equivalent if...AC_DEFINE.
Adding to this C wrappers for QMP commands threatens to make QMP command
arguments part of the library ABI. Compatible QMP evolution (like
adding an optional argument) turns into a libqmp soname bump.
Counter-productive. How do you plan to avoid that?
.so versioning. Ugly as hell to do manu
On 03/24/2010 12:02 PM, Daniel P. Berrange wrote:
I don't see any need to special case the base image here. The main
virDOmainSnapshotDelete() API already lets you discard all snapshots
until you get to the base image - VMWare/VirtualBox already demonstrate
this is sufficient.
Is this as effici
On 03/24/2010 04:24 PM, Darryl L. Pierce wrote:
On Wed, Mar 24, 2010 at 03:34:04PM +0100, Paolo Bonzini wrote:
I'm still in the same state here. Shall I open a bug against libvirt
regarding it not recognizing the --tftp-root argument?
That would be against dnsmasq?
I'm not sure i
On 03/24/2010 02:09 PM, Chris Lalancette wrote:
The problem, though, is what Mattias points out; there is no (easy) way
that, given state C, I can get back to state A to make a new snapshot.
I actually have to be at state A to take a new snapshot with a parent of
A. I think this is a place where
On 03/24/2010 01:08 PM, Darryl L. Pierce wrote:
On Thu, Mar 11, 2010 at 04:10:24PM -0500, Darryl L. Pierce wrote:
Here is what's output by libvirtd:
11:36:57.608: debug : virRun:809 : /usr/sbin/dnsmasq --strict-order
--bind-interfaces --pid-file=/var/run/libvirt/network/testbr6762.pid
--conf-
> > /* Return the number of snapshots for this domain */
> > int virDomainSnapshotNum(virDomainPtr domain, unsigned int flags);
>
> Shouldn't this be called virDomainNumOfSnapshots to be consistent with
> the naming or other num-of functions?
Gah. I struggled with this; I like the fact that all
On 03/18/2010 06:04 PM, Eric Blake wrote:
>>
>> ACK. I had to double-check MSDN to make sure that threads-win32 already
>> creates recursive mutex by default.
>
> I don't supposed you read if its possible to make it non-recursive, since
> the orginal impl is supposed to be non-recursive !
On 03/01/2010 06:24 PM, Daniel P. Berrange wrote:
mingw == gcc built as a cross-compiler, so there's no issue there.
Unfortunately, _native_ mingw (i.e. building under MSYS) uses a
hideously old version of GCC (3.4.x), so it's not always a non-issue.
But it is fine in this case:
http://gcc.
On the other hand, perhaps a more invasive rewrite would also work while
also avoiding assert(), by hoisting the worker->quitRequest into the while
loop, something like:
while ((client = qemudPendingJob(server)) == NULL) {
if (worker->quitRequest
|| virCondWait(&se
On 02/24/2010 05:09 PM, Richard W.M. Jones wrote:
In bug 567931 we found that virt-top would exit occasionally
when the terminal window was resized. Tracking this down it
turned out that SIGWINCH was being delivered to the process at
exactly the point where the libvirt remote driver was calling
On 02/24/2010 04:09 PM, Richard W.M. Jones wrote:
No signal handler will necessarily do things like
> longjmp-ing out of the remote driver, so every signal could give rise to
> a similar bug.
Did you mean "signals handlers might longjmp out ..."?
No, I meant what I wrote, as longjmp-ing out
On 02/24/2010 04:22 PM, Richard W.M. Jones wrote:
On Wed, Feb 24, 2010 at 08:11:45AM -0700, Eric Blake wrote:
According to Diego Elio Pettenò on 2/24/2010 8:04 AM:
Check the host string before deciding whether to look for windres or not.
dnl Look for windres to build a Windows icon resource.
On 02/24/2010 04:21 PM, Eric Blake wrote:
> + AC_SEARCH_LIBS([dlopen], [dl],
> +[DLOPEN_LIBS=${LIBS#${old_libs}}
The shell expansion ${a#b} is not portable to Solaris /bin/sh. Is that a
problem for libvirt, or are we Linux-centric enough to not care?
It should definitely configure rig
On 02/24/2010 02:06 PM, Richard W.M. Jones wrote:
The correct solution is to mask out SIGWINCH for the duration
of the poll(2) system call. The per-thread mask is changed and
restored immediately after the call. Since we are using
pthread_sigmask, this should not affect other threads, and
since
On 02/24/2010 06:20 AM, Laine Stump wrote:
...since I see nothing in POSIX that guarantees that pipe() will preserve
your initialized values across failure. I think it's better to
explicitly
set things to -1 on the failure path, rather than initializing that way.
Interesting point. If we can't
On 02/16/2010 05:31 PM, Dave Allan wrote:
ANother option would be to add an explicit virStorageVolZero() API to
allow a volume to be wiped independently of create/delete operations.
I like this approach. Since the code is already written for pretty much
every permutation: regular non-sparse f
On 02/16/2010 12:33 PM, Daniel P. Berrange wrote:
The idea of zeroing upon delete, is that we want the currently allocated
extents to be overwritten with zeros. If we truncate to 0 size, then
extend to original size I imagine that would easily allow the filesystem
to give you a new set of extents
On 02/16/2010 05:08 AM, Dave Allan wrote:
What about for a sparse file doing just
ftruncate (fd, 0);
ftruncate (fd, st.st_size);
That's a good point. The only thing that makes me hesitate is that
I'm not certain that the file is guaranteed to contain zeros
following the second call to ftrunc
On 02/15/2010 10:29 PM, David Allan wrote:
* If the volume is a sparse file and the fiemap ioctl is available,
use fiemap to locate the volume's extents.
What about for a sparse file doing just
ftruncate (fd, 0);
ftruncate (fd, st.st_size);
?
It's already sparse, it doesn't hurt to make
On 02/10/2010 07:11 PM, Daniel P. Berrange wrote:
On Fri, Feb 05, 2010 at 04:56:56PM +0100, Jiri Denemark wrote:
Signed-off-by: Jiri Denemark
---
daemon/remote.c | 25 +
daemon/remote_dispatch_args.h |1 +
daemon/remote_dispatch_prototy
testbr31018
Are there any relative pathing values that should be used by the tftp
and bootp tags? Should the tftp->root attribute point to the work
directory and then bootp->file refer to tftpboot/pxelinux.0?
No, the above seems fine (I should have pointed you to
tests/networkxml2
On 01/08/2010 06:23 PM, Daniel P. Berrange wrote:
diff --git a/src/qemu/qemu_conf.c b/src/qemu/qemu_conf.c
index 5dcd50f..eded887 100644
--- a/src/qemu/qemu_conf.c
+++ b/src/qemu/qemu_conf.c
@@ -1592,6 +1592,9 @@ static char *qemuDiskDriveName(const virDomainDiskDefPtr
disk)
case VIR_D
On 01/08/2010 06:23 PM, Daniel P. Berrange wrote:
@@ -1206,9 +1206,7 @@ int qemuMonitorJSONAddUSBDeviceMatch(qemuMonitorPtr mon,
/* XXX qemu also returns a 'function' number now */
Wrong comment now.
Paolo
--
Libvir-list mailing list
Libvir-list@redhat.com
https://www.redhat.com/mailman/lis
You are not giving BOOTP info (you are missing the tag,
which goes inside).
Okay, here's the XML that's now being generated by my test script:
testbr6253
However, I still get the message "No filename or root path specified"
within the VM and it does not boot from the above direct
(mcpie...@mcpierce-desktop:node-image)$ sudo virsh net-dumpxml testbr541
testbr541
f389317f-8420-7516-df9d-756b7deb3d37
When I start up the VM, I see it get an IP address within the range
specified, but it never pxeboots the ISO image.
To what
601 - 700 of 790 matches
Mail list logo
