[SVM] Make sure that the WP bit is always set in CR0 when running SVM guests.
This is necessary to be sure that the shadow mode code can always detect
writes to guest page tables.

This should fix Windows restore from hibernation on SVM platforms.

Signed-off-by: Tom Woller <thomas.woller@amd.com>

[HVM] Update ACPI tables to reflect new PIRQ routing:
 1. Override ISA IRQs 5,6,10,11 as active-low level-triggered.
    These are 'wired' to PCI devices.
 2. Limit PCI-ISA link routers to route to 5,6,10,11. Others
    are active-high edge-triggered and so invalid to use for
    PCI devices. In any case, four ISA IRQs is sufficient to
    avoid unnecessary aliasing among four links.
Also remove PCI device 0 IRQ info. It's an embedded host bridge with
no interrupt sources.

Signed-off-by: Keir Fraser <keir@xensource.com>

[HVM] Move PCI and PCI-ISA bridge setup to hvmloader.
Signed-off-by: Keir Fraser <keir@xensource.com>

[QEMU] Add TCP_NODELAY to tcp connections exporting serial ports.

Signed-off-by: Steven Smith <sos22@cam.ac.uk>

Add -p/--paused flag to xm restore.

When --paused is specified, the domain is paused after it is loaded.

Signed-off-by: Brendan Cully <brendan@cs.ubc.ca>

The legacy apis return a value of None on success, which was being converted to a value of 'nil' in the xmlrpc response.
The Xen IPI spec states that destroy, hard_shutdown, pause, resume, start, suspend, and unpause return void.  This patch makes it so and permits proper parsing of the response by clients (libxen) that follow the spec.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>

Stop the kernel and ramdisk from being nuked if the kernel_external bootloader
setting is being used.

From Jim Fehlig <jfehlig@novell.com>.
Signed-off-by: Ewan Mellor <ewan@xensource.com>

[XEN] Fix the build.
Signed-off-by: Keir Fraser <keir@xensource.com>

[HVM] Set VIOAPIC ID to a value that does not conflict with LAPIC IDs.
Signed-off-by: Keir Fraser <keir@xensource.com>

[HVM] Fix instruction linear address computation.
This helps newer isolinux' graphical boot code (which crashes without
this).

Signed-off-by: Jan Beulich <jbeulich@novell.com>

[TOOLS] Check for python-devel.

Thanks to Ewan Mellor, Anthony Liguori, and Daniel P. Berrange for
pointing out problems with and solutions for the problems with the
first version.

Signed-off-by: Jan Beulich <jbeulich@novell.com>

[IA64] Change callback_irq spec of PV-on-HVM for IPF

Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
Signed-off-by: Alex Williamson <alex.williamson@hp.com>

[LINUX] Make xen /dev/mem aware of IA64 to use X.

The original /dev/mem (linux/drivers/char/mem.c) is aware of IA64
specific issues. On the other hand paravirtualized /dev/mem
(linux-2.6-xen-sparse/drivers/xen/char/mem.c) is simplifed not to be
aware of IA64. This patch makes it IA64-aware to use X on domain0/IA64.
With this patch and ia64-specific patches, X server can boot on domain0/IA64.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[MINIOS] Refactored mm.c and sched.c. x86 arch specific code got moved to
arch/x86/mm.c and arch/x86/sched.c. Header files were also refactored:
arch specific code got moved to include/x86/arch_mm.h and
include/x86/sched_mm.h.

Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>
Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>

[POWERPC][LIBFS] Fix build breakage in log2 assembly.
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>

[XENOPROFILE] Make xenoprof arch-generic with dynamic mapping/unmapping xenoprof
buffer support and auto translated mode support.
renamed xenoprof_get_buffer::buf_maddr, xenoprof_passive::buf_maddr to
xenoprof_get_buffer::buf_gmaddr, xenoprof_passive::buf_gmaddr
to support auto translated mode. With auto translated mode enabled,
it is gmaddr, not maddr.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
Simplify the share function.

Signed-off-by: Keir Fraser <keir@xensource.com>

[XENOPROFILE] Update xenoprof linux side patch following Renato's comment.
 - compilation fix when CONFIG_OPROFILE=m with minor clean up
 - don't export xenoprof_is_primary.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[XENOPROF] Make xenoprof of linux side arch generic with some bug fixes.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[XENOPROFILE] Split linux-2.6-xen-sparse/arch/i386/oprofile/xenoprof.c into
linux-2.6-xen-sparse/drivers/xen/xenoprof/xenoprofile.c
without code changes except slight adjustment to compile.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[XENOPROFILE] Replace x86 specific code in xen/common/xenoprof.c
 - replace nmi_ prefix with xenoprof_arch_ prefix
 - move config_counter to xen/arch/x86/oprofile/xenoprof.c

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[XENOPROFILE] Move code under xen/arch/x86/oprofile to xen/common.
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>

[HVM] Reworked interrupt distribution logic.

TODO:
 1. Fix IO-APIC ID to not conflict with LAPIC IDS.
 2. Fix i8259 device model (seems to work already though!).
 3. Add INTSRC overrides in MPBIOS and ACPI tables so
    that PCI legacy IRQ routing always ends up at an
    IO-APIC input with level trigger. Restricting link
    routing to {5,6,10,11} and setting overrides for all
    four of those would work.

Signed-off-by: Keir Fraser <keir@xensource.com>

The legacy HTTP entry point for Xend remote access didn't hook the
dump operation for a domain, this patch adds it.

Signed-off-by: Daniel Veillard <veillard@redhat.com>

[XEN] Fix potential crash in grant-table code if
guest misbehaves and implicitly unmaps a PTE.
Signed-off-by: Keir Fraser <keir@xensource.com>

Fix iptable handling when multiple addresses have been specified.

From Adrian Chadd <adrian@creative.net.au>.
Signed-off-by: Ewan Mellor <ewan@xensource.com>

Added an xm console -q flag.

Signed-off-by: Ewan Mellor <ewan@xensource.com>

This is a refactored version of a previous patch that destroys external
devices' state when a VM configuration file is destroyed. Currently only
the vTPM device's state needs to be explicitly destroyed.
I am also surrounding the saving of the managed domain's configuration
with a try-catch.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>

A small fix to the documentation to reflect recent changes to the
'cfgbootpolicy' command.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>

There is a somewhat trivial issue with XendCheckpoint.py right now in
that it logs everything written to stderr by xc_save and xc_restore as
errors whereas in fact the vast majority of this output is
information/debug (and all actual errors are marked by the string ERROR:
at the start of the message) -- this is confusing to folks looking at
the logs and makes automated log analysis tricky.

Fix is to scan for the ERROR: string and log anything without it using
log.info instead.

Signed-off by: Simon Graham <Simon.Graham@stratus.com>

[BLK] Handle block devices with more than 2^32 sectors.
Signed-off-by: Keir Fraser <keir@xensource.com>

Add example of keyboard layout

Signed-off-by: Takanori Kasai <Kasai.Takanori@jp.fujitsu.com>

Cleanup after the bootloader even if the domain building fails.

Signed-off-by: Ewan Mellor <ewan@xensource.com>

[XEN] Update shadow audit code.
Make the flag checks line up with the current flag propagation strategy.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

Merge

[XEN] Remove write access to new PT before discarding old shadow.
This allows us to use the old pagetables's linear maps in our
remove-writeable-mappings heuristics, fixing the same crash that
cset 12339 did, but still letting us do fast revoke of writeable
mappings of toplevel pagetables.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[MINIOS] Delay enabling event delivery at start of day.
Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>

[BLKTAP] fix race between ufe_ring and blk_ring in blktap.c

Signed-off-by: Jake Wires <jwires@xensource.com>

[XEND] Accomodate extra "devices" in SXP that do not have DevControllers

Signed-off-by: Alastair Tse <atse@xensource.com>

[XEND] Revert changes in changeset 12328

Mistakenly reverted some fixes for rebooting, un-reverting.

Signed-off-by: Alastair Tse <atse@xensource.com>

[XEND] Fix missing domains on domU reboot.

Fixed regression with rebooting domU caused by the change in handling
syncing state from Xen and Xend.

Added missing implementation for XendDomainInfo.unwatchVm() which
disappeared during a merge.

Added protection for console/vnc-port reading for managed domains.

Signed-off-by: Alastair Tse <atse@xensource.com>

[LIBXC] Replace API uses of size_t with ulong.

Avoids possibility of different FILE_OFFSET_BITS defn.
across the interface.

Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Fix build. Mercurial ate my changeset fixups. :-(
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Restrict access to grant-mapping operations.
TLB flushing is not done strictly before notifying
the mappee -- this creates scope for multi-processor
mapping guests to attempt to abuse a stale mapping
on another VCPU.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

[QEMU] rtl8139: Disallow chaining above 64K

As it stands the 8139C+ TX chaining is only bounded by realloc failure.
This is contrary to how the real hardware operates.  It also has DoS
potential when ioemu runs in dom0.

This patch makes any attempt to chain a frame beyond 64K fail
immediately.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

[LINUX] Add process profiling hook to timer ISR.
Signed-off-by: Rik van Riel <riel@redhat.com>

HVM cleanups:
1) make vmx/svm time functions hvm common, since they are actually the same.
2) move hvm_send_assist_req from platform.c to hvm.c.
2) rename VMX MSR context switch functions to make them more readable.
3) misc coding style clean ups.

Signed-off-by: Xin Li <xin.b.li@intel.com>

[HVMLOADER][VTPM] Add support for a TCG ACPI table.

The ACPI table is generated if the probing for the TIS interface
succeeds. I cleaned up the ACPI signatures in acpi2_0.h to be more
readable.

The specification for the layout of these ACPI tables can be found here:
https://www.trustedcomputinggroup.org/groups/server/TCG_ACPIGeneralSpecification_1-00_1-00_FINAL.pdf

Extensions to the Bochs BIOS that use these ACPI extensions will follow.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>

[OPROFILE] Add support for core and core2 chips.

Tested on Woodcrest  processors.  Requires Oprofile 0.9.2.

Signed-off-by: Andrew Theurer <habanero@us.ibm.com>

[HVMLOADEr] Reserve MMIO 0xa0000 to 0xc0000 in ACPI dsdt.

Avoids possible vga driver loading problem in HVM Windows guest.
Also fixes a Makefile bug in hvmloader directory.

Signed-off-by: Qing He <qing.he@intel.com>

[TOOLS] Provide common LFS CFLAGS/LDFLAGS additions in Rules.mk.
Signed-off-by: Keir Fraser <keir@xensource.com>

[MINIOS] Refactor spinlock header for multi-arch support.

I separated the spinlock parts special to the x86 architecture and
moved these to include/x86/arch_spinlock.h. The common code is now in
include/spinlock.h.

Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>

[MINIOS] Event channel hypercall update for 64-bit guests.
Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>

[MINIOS] Move initialisation of events (masking event channels)
earlier during the boot process. Otherwise 64bit guests would
sometimes crash.

Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>

[XEN] Add missing #include.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Scheduling hypercalls defer entry to the scheduler to softirq
context.

This avoids entering the scheduler with outstanding multicall state
and also happens to simplify the x86 entry protocol to the scheduler
hypercall (since we do not need to preset the return code).

Signed-off-by: Keir Fraser <keir@xensource.com>

[HVM] Decouple the RTC from the PIT periodic timer
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[XEN] Remove redundant calls to check_for_early_unshadow()
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[XEND] Fix PCI configuration parsing and SXP output

PCI configuration is formatted differently from other device
configurations because of the requirement that all pci devices need to
be present when it is initialised.

Signed-off-by: Alastair Tse <atse@xensource.com>

Add to select keyboard layout

Signed-off-by: Takanori Kasai <Kasai.Takanori@jp.fujitsu.com>

Fix clobbering bug when hde is specified; QEMU only supports up to hdd, but
hde is accepted and writes off the end of some tables.

From Hiromichi Itou <ito@begi.net>.
Signed-off-by: Ewan Mellor <ewan@xensource.com>

[HVM] Disallow PTE updates and MMIO accesses from hypervisor mode
(prevents copy_to/from_guest from causing problems, for example).
Signed-off-by: Keir Fraser <keir@xensource.com>

Fix of floppy disk configuration check

Signed-off-by: Takanori Kasai <Kasai.Takanori@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>

This patch improves an usage message about xm dump-core command.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

Fix cpu_cap option.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

Improve error message when trying to pause / unpause domain 0.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

Fix xm migrate error message when trying to migrate domain 0.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

[NET] back: Fix bug in 12262:ac2097d71e06dbbf77279af10d6ae7359d921ab0.
Pointed out by Jan Beulich.
Signed-off-by: Keir Fraser <keir@xensource.com>

[IA64] Fix build breakage.
Signed-off-by: Alex Williamson <alex.williamson@hp.com>

Xen trivial build fix. Oops.
Signed-off-by: Keir Fraser <keir@xensource.com>

[BLKTAP] Little error handling fix to blktapctrl.c.

Signed-off-by: Andrew Warfield <andy@xensource.com>

[XEN] Various selector and callback cleanups to simplify the tools
and assumptions about callback selector values on x86/32.
Original patch from Jan Beulich <jbeulich@novell.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[BUILDER] Simplify vcpu context initialisation by pre-zeroing
the context structure.
Signed-off-by: Jan Beulich <jbeulich@novell.com>

[XEN] Define remaining x86 public pointer fields as guest handles.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[TOOLS] Rename chekc_hotplug to check_udev and fix permissions.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Track high-water-mark of p2m map
and so avoid some unnecessary __copy_from_user faults.
Also tidy the p2m functions generally.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

Add the VNC port to the output of xm list --long, if the value has been written
to xenstore by QEMU.

Signed-off-by: Ewan Mellor <ewan@xensource.com>

[MINIOS] Add timer support.

Based on an original patch by Robert Kaiser.

Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>

[LIBXC] Clarify error messages to talk about allocation 'extents'
rather than 'pages'.
Signed-off-by: Jimi Xenidis <jimix@watson.ibm.com>

[PV-ON-HVM] Correct some out of date documentation.

Noticed-by: Nowatzki, Thomas L <Thomas.Nowatzki@UNISYS.com>
Signed-off-by: Steven Smith <sos22@cam.ac.uk>

merge with xen-ia64-unstable.hg

merge with xen-unstable.hg

[IA64] remove ASSERT in vmx_init.c

Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>

[IA64] Add CCISS driver to build config and change nbd to a module

LVM complains a lot when nbd is built into the kernel

Signed-off-by: Alex Williamson <alex.williamson@hp.com>

Fix breakage of blktap device strings from cset 63d1b02dad347e8feb845d8a8b482e251a478164.

Signed-off-by: Andrew Warfield <andy@xensource.com>

Fix xm block-configure, by allowing the device ID to be unspecified inside
device_configure (in which case it will be specified by the incoming config).

Signed-off-by: Ewan Mellor <ewan@xensource.com>

Speed up shadow of x86_64 Linux guests using 3 paging levels.
Linux kernel 2.6.9 or older uses 3 paging levels and its direct map is
from 0x10000000000.

Signed-off-by: Xin Li <xin.b.li@intel.com>

[XEN] Extend fix from 12403:680b37789200 to other emulation paths.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[LINUX] x86/64: Fix backward compatibility to Xen 3.0.2.
This requires us to explicitly specify _PAGE_USER for kernel mappings.
Original patch by Jan Beulich <jbeulich@novell.com> and Gerd Hoffmann.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Fix the "early unshadow" optimization to look at the entire pte.

The optimization is designed to detect pagetable destruction, but re-uses the
pagetable entries to store data under certain conditions (paging, zeroing, &c);
sometimes the lower 8 bits are 0.  These "false positives" cause unnecessary
shadowing and unshadowing.

Signed-off-by: George Dunlap <gdunlap@xensource.com>

[LINUX] Preserve flags when converting PTEs from machine to phys.
Signed-off-by: John Byrne <john.l.byrne@hp.com>

[XENPERF] Pretty-print hypercall stats.
From: Yoshihiko Yaegashi <y.yaegashi@jp.fujitsu.com>
From: Ken Hironaka <kenny@logos.ic.i.u-tokyo.ac.jp>
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Don't pass spurious reserved-bit faults to the guest
since x86_64 linux can't handle them.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[XEN] Don't use shadow linear maps in sh_set_toplevel_shadow()
because they will not be there.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[XEN][POWERPC] Change license on public/ headers from GPL to BSD.
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>

[VMXASSIST] Remove unused variable declarations to avoid compiler warning.
Signed-off-by: Ben Thomas <ben@virtualiron.com>

[IA64] Fix weird behaviour when configuring >3G memory for a VTI domain.
Signed-off-by, Anthony Xu < anthony.xu@intel.com >

[HVM] vlapic: More cleanups, simplifications and fixes.
Signed-off-by: Keir Fraser <keir@xensource.com>

Add uuid to host record returned by host_get_record.  libxen fails to parse the host record with this field missing.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>

[IA64] Fix xencomm for xm mem-set command

This patch saves and restores the hypercall parameter within xencomm.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

[IA64] Fix HVM interrupts on IPF

Xen has changed to set-irq-level hypercall from shared-memory PIC stat.
This patch makes IPF accomodate this change

Signed-off-by: Anthony Xu <anthony.xu@intel.com>