The reason is that tsc stops and it causes the inaccuracy.
And later we can write some scripts based on this patch.
Signed-off-by: Guanqun Lu <guanqun.lu@intel.com>
Remove some unused variables and replaces read and write
to the pipe with read_exact and write_exact (these two functions are
implemented in libxc, that we have to link anyway).
This allows fs-backed to be compiled with -D_FORTIFY_SOURCE=2, hence
should fix the problems reported by Boris.
Signed-off-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
When XenAPI is available, xm scsi-attach fails. Also xm new and
xm create fail if a vscsi parameter is defined into domain
configuration files.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
If domU frame buffer configuration has no vnclisten attribute set, use
the global vnc-listen address specified in xend configuration file
when updating vfb location. Currently the hardcoded value 'localhost'
is used.
Signed-off-by: Jim Fehlig <jfehlig@novell.com>
When XenAPI is available, xm new and xm create fail.
# xm new vm1
Using config file "/etc/xen/vm1".
Attribute 's3_integrity' not declared
# xm create vm1
Using config file "/etc/xen/vm1".
Attribute 's3_integrity' not declared
This patch adds an s3_integrity into create.dtd.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Rather than use "0x0" to signify that the hotplug slot should
be automatically selected by qemu-dm, define AUTO_PHP_SLOT_STR.
At the same time, change the magic value form 0x0 to 0xff,
in line with changes made to qemu-dm to allow any unused PCI
device to be used for hotplug.
Signed-off-by: Simon Horman <horms@verge.net.au>
When a udev event is received, udevevent.py parses the udev data and
tells XendNode.py to update the physical resource information.
This patch also add a boolean parameter 'xend-udev-event-server', to
let users indicate whether we should enable this function or not.
Signed-off-by: Yosuke Iwamatsu <y-iwamatsu@ab.jp.nec.com>
Add __classes_lock to protect __classes, since it can be modified by
the udev listener thread.
Signed-off-by: Yosuke Iwamatsu <y-iwamatsu@ab.jp.nec.com>
Introduce SocketDgramListener and UnixDgramListener classes.
We already have STREAM (connection) type socket listener classes in
the source tree, but we need DGRAM (connectionless) type listeners to
receive udev events.
Signed-off-by: Yosuke Iwamatsu <y-iwamatsu@ab.jp.nec.com>
CPU topology info is necessary for power management analysis. For
example, to analysis the effect of Px state coordination, Cx
package/core coordination, the thread/core/socket topology
information is needed.
This patch add new command "get-cpu-topology" in xenpm to print the
CPU topology info:
Signed-off-by: Yu Ke <ke.yu@intel.com>
In the case of a host OS without lsscsi command, xm scsi-attach
command fails even if you specify existing SCSI device.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
This patch moves a call processing of a domain reset processing
from xm to xend. Also, it adds a reset argument to do_dump() of
SrvDomain.py.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Until now the udev script for blktap devices needs to decide if to
signal success or failure to xend. As this script runs completely
independent of blktapctrl and tapdisk/ioemu which do the real work,
the udev script can't even theoretically know if tapdisk is happy.
This patch removes the udev script and replaces its checks by new
ones in libblktap.
Signed-off-by: Kevin Wolf <kwolf@suse.de>
Connections to ioemu have single_handler set, so they are closed as
soon as all images of a certain type are closed. This is wrong with
ioemu: All images that belong to the same domain are handled by the
same backend process (usually qemu-dm, but also tapdisk-ioemu for
domains without device model), regardless of the image type.
This patch checks for the same-domain condition for ioemu connections.
Signed-off-by: Kevin Wolf <kwolf@suse.de>
Currently all disk types that are supported are defined in a header
file private to blktapctrl and tapdisk. When restoring ioemu as a
backend for blktap these constants are needed by ioemu, so move them
to a more public header file.
Signed-off-by: Kevin Wolf <kwolf@suse.de>
This patch adds support for specifying the backend (tapdisk or ioemu)
to blktapctrl. Images can be specified e.g. as tap:tapdisk:aio,
tap:ioemu:qcow2 or tap:vmdk. When omitting the backend, a default is
chosen based on the image type (currently always tapdisk because ioemu
as a backend is broken until a follow-up patch series against qemu-xen
is applied)
Signed-off-by: Kevin Wolf <kwolf@suse.de>
attached patch fixes a regression in c/s 19330 which
prevents to start guests on a Linux Dom0.
Signed-off-by: Christoph Egger <Christoph.Egger@amd.com>
This is a corrected version of a patch commited as c/s 19250 and
reverted by c/s 19314. Thanks to "trap sigerr ERR" in
xen-hotplug-common.sh the xen-hotplug-cleanup would exit when reading
/local/domain/ID/vm fails thus skipping all the xenstore-rm lines in
the rest of the script.
Change deviceDestroy behavior to remove /vm/UUID/device/...
path only when force was used (as it already does so for both frontend
and backend) and do the removing from xen-hotplug-cleanup script when
we are sure the device is really not attached to the guest any more.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
This patch prevents xend from creating a XendNetwork instance of
tmpbridge (a temporary bridge used in network-bridge script).
Signed-off-by: Yosuke Iwamatsu <y-iwamatsu@ab.jp.nec.com>
The code was attempting to use the domain's current number of pages
(info.nr_pages) as a maximum index. We then walk the memory map and
can easily over-write past the end of the nr_pages-sized array, if the
domain has more pages mapped in than earlier (live dump). Restrict
ourselves to the current number of pages.
Also fix the dump core method in xend to actually implement the crash
and live options. In particular this means that xend clients other
than xm now get non-live dumps by default.
Signed-off-by: John Levon <john.levon@sun.com>
The comment about the lack of restart prevention is not accurate.
Bump the minimum restart time to 60 seconds.
Signed-off-by: John Levon <john.levon@sun.com>
On Solaris, xend runs in a 'process contract' such that all children
are killed when the service is restarted. Spawn qemu processes in a
new contract to avoid this.
Signed-off-by: John Levon <john.levon@sun.com>
If a client asks about a non-existent domain, report the error back
properly instead of just backtracing within xend.
Signed-off-by: John Levon <john.levon@sun.com>
The Solaris curses library has a broken timeout() function: after a
first timeout() call with a positive value for an argument, subsequent
calls will fail to reset it. So, getch() always times out, confusing
the pygrub timer in the main loop. Add an extra check to avoid exiting
prematurely.
Signed-off-by: Frank van der Linden <frank.vanderlinden@sun.com>
Make sure that tty semantics are active for Solaris ptys, or if they
aren't (and not needed) to not do tcget/setattr on the filedescriptor
in Python code.
Also work around a bug in the Solaris ptm streams driver, which will
cause a write error on the master side of a pty (because of e.g. a
missing slave) to persist forever.
Signed-off-by: Frank van der Linden <frank.vanderlinden@sun.com>
Keir Fraser