This patch implements TPM frontend driver resume after checkpointing a
domain. I also now let the front-end wait for the response from the
TPM until the response is there. A response must come due to the
strict request/response nature of the protocol, but the time until the
responses arrives depends on the type of operation that is executed by
the vTPM. Resending a request is not possible with this type of device.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
This changes netback to use struct page's mapping field instead of the
index one, and the PageForeign logic the other way around. That way,
CONFIG_SPLIT_PTLOCK_CPUS doesn't have to be suppressed anymore for x86.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Ensure the backend terminates when the frontend vanishes before the
connection is established. This happens when the guest doesn't have
the devices.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
While object files generated from C sources should generally not
depend on asm-offsets.h, those generated from assembly sources
should.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
This capability means that save/restore will remember contexts of all
online VCPUs at the time of final guest suspend. Any initialised but
offline VCPUs will not be saved!
Signed-off-by: Keir Fraser <keir@xensource.com>
If xc_map_foreign_range fails, the new domain object is freed,
but it is not removed from the connection list.
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
Each serial, parallel and monitor device in qemu that is connected to
a pty creates a xenstore node of the form:
<domain-path>/monitor/tty
<domain-path>/serial/<n>/tty
<domain-path>/parallel/<n>/tty
In addition, serial/0 (com1) also registers its information at:
<domain-path>/console/tty
Also fix a realloc() failure memory leak.
Signed-off-by: Ben Thomas <ben@virtualiron.com>
domain: raise the correct error code (XendInvalidDomain instead of
XendError). Also remove redundance exception code that is handled by
domain_lookup.
Signed-off-by: Yung Giang <ygiang@gmail.com>
grant table entries instead of the shared ones. I had to move some
functions from grant_table.c into grant_table.h to make them usable by
the ACM module.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
across it. No longer make calls to the function dependent on 'non-live
save/restore': it's not a good test of whether the fd is a socket.
Signed-off-by: Keir Fraser <keir@xensource.com>
will also work on redhat-based distributions before RHEL 5. See
discussion "xendomains init script" about a year ago on xen-devel.
Signed-off-by: Florian Kirstein <ray@ray.net>
workqueue_mutex. This is a new deadlock since the workqueue_mutex is
acquired in the workqueue_cpu_calbback() function across
CPU_UP_PREPARE->CPU_ONLINE.
The fix is for us not to rudely grab the cpu_hotplug_lock() during
save/restore -- it's really not necessary.
This patch is applicable to any of our 2.6 kernels, but is absolutely
required from 2.6.18 onwards.
Signed-off-by: Keir Fraser <keir@xensource.com>
The "fast" method fails a significant portion of the time.
unshadow will still use the "fast" method, but will fall
back to a brute-force method if it fails.
Signed-off-by: George Dunlap <gdunlap@xensource.com>
Make sure to reset the early-unshadow logic when we
do a fast-path propagate or when we emulate a non-zero
write.
Signed-off-by: George Dunlap <gdunlap@xensource.com>
Related changesets:
14148:b67c253d1cdb4f502dec2
13519:b4a8000e76db6b4b27341
These three changesets must be applied as a set!
Signed-off-by: Keir Fraser <keir@xensource.com>
