local network migration without the 'live' parameter 3 times. The test
'04' that does live migration now tests local live migration 3 times.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
all the necessary fields initialised. In particular, before we were
not copying ip_summed and that screws up checksum offload.
Signed-off-by: Keir Fraser <keir@xensource.com>
XendCheckpoint class to differentiate between network suspend/resume
(network=True) and local suspend/resume (network=False).
Instead of passing the 'live' parameter to the migration methods, this
'network' parameter is passed now. This ends up being merely a renaming
of the parameter.
A check with the xm-test suite against this change has resulted in no
additional errors.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
configuration of an existing domain. This is needed for a domain to be
able to reboot and retain its PCI device configuration. This patch adds
such support.
Thanks to Mike Wright for reporting this problem and working with me to
fix it.
Signed-off-by: Ryan Wilson <hap9@epoch.ncsc.mil>
1) Move the tablet to the second serial port. This way the tablet will
not interfere with people who want to use a serial console on the guest.
2) Enhance the Summagraphics emulation so that the Windows XP driver
works. Turns out the Windows driver was using capabilities the X driver
didn't care about and it wouldn't recognize the tablet without these
capabilities.
Signed-off-by: donald.d.dugger@intel.com
If you change the video resolution on a Windows XP guest such that it uses
fewer bytes of VRAM (either by using fewer bytes per pixel or by lowering the
resolution) then some window backgrounds will become corrupted. This happens
because the Windows XP Cirrus Logic driver assumes that VRAM is initialized
to 0xff whenever the video mode switches between VGA and SVGA.
Signed-off-by: Donald.D.Dugger@intel.com
In the external device migration script I am introducing a subtype
(-subtype <subtype>) from which the script's name is formed. Also all
unparsed parameters are passed to the called script.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
sysfs (and hence trigger hotplug callbacks) until the devices really
are dead. This fixes a bug where the deferred code to free a blk
device was running concurrently with a hotplug-remove callback which
would try to reclaim the underlying storage. In some cases the race
would be lost and the hotplug script would fail.
Thanks to the Zhu Han at Intel for finding the root cause of this
long-term and annoying bug!
Signed-off-by: Keir Fraser <keir@xensource.com>
the EXCEPTION_BITMAP vmcs control. There is no need for that as VMM is
just plainly re-injecting back to the guest. The attached is a simple
patch to set the vmcs control properly.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
The current code will break if there are multiple scope global routes
since the "ip route list default" command doesn't use "default" as a
filter, and will output multiple devices which breaks the rest of the
script.
Also make the vif/xenbr/eth naming consistent, i.e. if your default
network interface is eth1, you get vif1 and xenbr1.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Made it a minor release bump for xm-test so old ramdisk will work. The
ramdisk is functionally the same.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
merge with given values. This will allow a bootloader to not have to
specify everything which is useful for bootloaders other than pygrub.
Signed-off-by: Jeremy Katz <katzj@redhat.com>
previous event. We enter an infinite loop if the original failed
injection cannot be fixed up by Xen (e.g., because it's not a shadow
pagetable issue).
The RHEL4 HVM guest hang issue was actually a side effect of
change-set 9699. In the rhel4 guest hang rc.sysinit init-script was
calls kmodule program to probe the hardware. The kmodule uses the kudzu
library call probeDevices(). For probing the graphics hardware in the
vbe_get_mode_info() function, sets up the environment and goes into the
vm86 mode to do the int x10 call. For returning back to protected mode
it sets up a int 0xff call. At the time of calling the int 0xff the
guest process pages were not filled up. And it was causing an infinite
loop of vmexits with the IDT_VECTORING_INFO on the int 0xff instruction.
The reason for the infinite loop is changeset 9699. With that
the guest page fault was always getting overridden by the int 0xff gp
fault coming from the IDT_VECTORING_INFO. With the attached patch if VMM
is injecting exceptions like page faults or gp faults then
IDT_VECTORING_INFO field does not override it, and that breaks the
vmexit infinite loop for the rhel4.
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
In the Xen 2.0.x compatibility section of xend (where we try to parse
the s-expressions if they came from an SXP configuration file for Xen
2.0.x), the wrong variable is referenced. This fix corrects the python
script to use the correct variable.
Thanks to Mike Wright for reporting this.
Signed-off-by: Ryan Wilson <hap9@epoch.ncsc.mil>
substitutions. This allows a Xen-patched Linux to build when /bin/sh
is a POSIX shell other than bash.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
When you generate non-linear skb's as netback does, it is up to you
to ensure that truesize reflects the size of paged part of the skb.
Otherwise socket memory accounting in Linux falls apart.
Thanks to the new check that David Miller added to TCP we were able
to catch this bug.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Changeset 9889:42a8e3101c6c reorganized the code on this file, and
missed this spin_unlock_irq(). Without this patch, my machine hangs
completely during boot. With this, it works.
Signed-off-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
during a context switch to the SVM domain's vcpu. This patch also
initializes the tlb_control to 1 for the initial do_launch().
Signed-off-by: Tom Woller <thomas.woller@amd.com>
creating domains with networking very easy. This patch:
1) Adds new XenDevice class, with the XenNetDevice subclass. These
classes represent devices for xm-test and are tied to XenDomains.
This can eventually be used for block devices as well. Currently,
devices must be added to domains prior to starting the domain. The
attach and detach needs to be handled.
2) Adds a new NetConfig class to handle configuring the network
environment in which the tests run. This patch only handles ranges
of IPs in a bridged environment. DHCP needs to be added as well as
handling NAT and routed environments.
3) Modifies XenDomain class to handle XenDevices.
4) Adds new configuration options for defining a range of IPs, their
network address, and their netmask.
5) Removes the old Network.py and Network class.
6) Modifies the existing tests to use the new infrastructure.
7) Adds some documentation to help creating domains.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
in XenDomain. The devices we will add, like network, will need to run
console commands when a domain is started. So I have made console be
a part of domain object, which it's tied to anyway.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
it as HVM or PV rather than using the config time option. The forthcoming
XenNetDevice will need to check. This is also the first step toward running
PV and HVM domains at the same time.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
NULL. This problem fixes errors that occur when linking Mini-OS with
ANSI standard code that uses stddef.h.
Signed-off-by: Keir Fraser <keir@xensource.com>
