[XEN] New event-channel reset operation.
Plumbed through libxenctrl to python.

From: Andrei Petrov <andrei.petrov@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

Add unsuspend hook for resuming devices in the checkpoint parent.

Nothing uses it at the moment.

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

Make xen_suspend handle resume.

Don't destroy xenstore watches on suspend, and only recreate them when
resuming in a new domain. Likewise, only invoke frontend device resume
code when in a new domain (the resume functions all tear down the
existing function and wait for the backend to negotiate a new one,
which does not happen in the source domain).

Signed-off-by: Brendan Cully <brendan@cs.ubc.ca>
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Add missing change for XS_RESUME.
Signed-off-by: Keir Fraser <keir@xensource.com>

xenstore: add XS_RESUME command; export it to xend.

This clears the shutdown flag for a domain in xenstore, allowing
subsequent shutdowns of the same domain to fire the appropriate
watches.

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

Add resumedomain domctl to resume a domain after checkpoint.
Export resumedomain domctl to libxc, xend.

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

[merge] with xen-unstable

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

[MINIOS] More cleanups for the ia64 port.
Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>

[MINIOS] Fix compile warnings with -O.
Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>

[TOOLS] Avoid void* arithmetic when parsing elf notes.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Avoid void* arithmetic in elfnote code.
From: Christoph Egger <Christoph.Egger@amd.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[merge] with xen-unstable

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

[LIBXC] Handle trace-buffer cpumap endianness.
Original patch from Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[LIBXC] Convert between byte-based and 64-bit bitmap arrays. Use this
for conversion of the domctl_cpumap.
Original patch from Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Convert between long-based and byte-based bitmap arrays.
Use this for conversion of the domctl_cpumap type on big-endian
systems.
Original patch from Jimi Xenidis <jimix@watson.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[HVM] Whitespace cleanup.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

[MINIOS] Added domctl and sysctl hypercalls to the
architecture-specific header files. Also increased the
__XEN_INTERFACE_VERSION__ to 0x00030205.

Signed-off-by: Derek Murray <Derek.Murray@cl.cam.ac.uk>

[XEN] Revert change to vcpu_guest_context which breaks guest ABI.

HVM save/restore needs to save/restore vcpu state via the new
stream-based load/save hypercalls.

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

[HVM] Only update guest time on timer interrupt delivery if the guest time is in the past.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

Tidy up the restore path following HVM save/restore patches.

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

Update pygrub for new Solaris directory names.

Signed-off-by: John Levon <john.levon@sun.com>

[HVM] Fix slow wallclock in x64 Vista. This is due to confusing a
timeout in the past vs. a timeout in the future when looking at a
32-bit HPET comparator.
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>

[linux networking] Remove the code to allocate page sized skbs.
This is no longer used much since the default is to always copy packets
to guests.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

Build PAE by default on x86_32.
Signed-off-by: Keir Fraser <keir@xensource.com>

Merge with xen-ia64-unstable.hg

Copy vmlinuz to correct folder when make source on IA64 platform.
Signed-off-by: Yongkang You <yongkang.you@intel.com>

[ioemu] fix build - prototype for static function needs to be static.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

Allow vcpu_avail to be specified in the configuration file.  Rationalise our
use of online_vcpus and max_vcpu_id, keeping track of the configured value in
vcpus_number.

Fix HVM SMP configuration to use the XendDomainInfo.getVCpuCount() call that
is used everywhere else, rather than parsing it out of the device config.

Fix xm list to show the configured VCPU count when the VM is down.

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

Fix get_config_value fix

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

Fix get_config_value call.

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

[HVM] Save/Restore: don't clobber page types in shadow code
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[HVM] Save/restore: Sometimes the hg changeset is not available.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[HVM] Save/Restore: Various structures have changed
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[HVM] save restore: shadow fix

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
fix 2 shadow issues:
* restored guest's mem type_info are not correct
* remove annoying warnings when restore

enable smp:
turn on the vcpu when set_vcpu_context in restore as cpu hotplug is not feasible for HVM

[HVM] save restore: guest memory handling

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
add support for save/restore HVM guest memory

[HVM] save restore: device model support

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
qemu's save/restore including vga acceleration support

[merge] with xen-unstable

[linux] Fix set_fixmap for vdso mapping on i386 and make x86_64 version correcter.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

Revert 13389; libfsimage version does not need to change after all.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[IA64] Remove duplicate typedefs

These are now defined in xen/include/xen/types.h

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

merge with xen-unstable.hg

[IA64] Passing address of PSCB is not needed

Since the address of PSCB is decided by xenolinux now,
passing address of PSCB to xenolinux is not needed.

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

[IA64] Remove duplicate check is_running_on_xen

When these functions are called, it is running on XEN.

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

[IA64] Let Xen handle identity mapping for xenolinux region 7

During kernel build in dom0, half of the resulting slow reflections
are alt DTLB misses.  Since we know Linux identity maps region 7,
Xen can handle the miss itself.  This reduces the system time component
of the kernel build by nearly 15%.

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

[XEN] Simplify CR3 switching code (remove slow path that allows old
page tables to be writable in new page tables -- shadow mode cannot
handle the failure case and it is easy to work around this in the
guest anyway). Also remove broken shadow-refcount compat-mode case.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Fix deferred mmu op logic.
Signed-off-by: Keir Fraser <keir@xensource.com>

[linux] Update to linux-2.6.18.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
--HG--
rename : linux-2.6-xen-sparse/include/asm-i386/mach-xen/setup_arch_pre.h => linux-2.6-xen-sparse/include/asm-i386/mach-xen/setup_arch.h
rename : patches/linux-2.6.17/blktap-aio-16_03_06.patch => patches/linux-2.6.18/blktap-aio-16_03_06.patch
rename : patches/linux-2.6.17/fix-ide-cd-pio-mode.patch => patches/linux-2.6.18/fix-ide-cd-pio-mode.patch
rename : patches/linux-2.6.17/git-3566561bfadffcb5dbc85d576be80c0dbf2cccc9.patch => patches/linux-2.6.18/git-3566561bfadffcb5dbc85d576be80c0dbf2cccc9.patch
rename : patches/linux-2.6.17/git-4bfaaef01a1badb9e8ffb0c0a37cd2379008d21f.patch => patches/linux-2.6.18/git-4bfaaef01a1badb9e8ffb0c0a37cd2379008d21f.patch
rename : patches/linux-2.6.17/git-dbaab49f92ff6ae6255762a948375e4036cbdbd2.patch => patches/linux-2.6.18/git-dbaab49f92ff6ae6255762a948375e4036cbdbd2.patch
rename : patches/linux-2.6.17/i386-mach-io-check-nmi.patch => patches/linux-2.6.18/i386-mach-io-check-nmi.patch
rename : patches/linux-2.6.17/ipv6-no-autoconf.patch => patches/linux-2.6.18/ipv6-no-autoconf.patch
rename : patches/linux-2.6.17/linux-2.6.19-rc1-kexec-move_segment_code-i386.patch => patches/linux-2.6.18/linux-2.6.19-rc1-kexec-move_segment_code-i386.patch
rename : patches/linux-2.6.17/linux-2.6.19-rc1-kexec-move_segment_code-x86_64.patch => patches/linux-2.6.18/linux-2.6.19-rc1-kexec-move_segment_code-x86_64.patch
rename : patches/linux-2.6.17/net-csum.patch => patches/linux-2.6.18/net-csum.patch
rename : patches/linux-2.6.17/net-gso-5-rcv-mss.patch => patches/linux-2.6.18/net-gso-5-rcv-mss.patch
rename : patches/linux-2.6.17/net-gso-6-linear-segmentation.patch => patches/linux-2.6.18/net-gso-6-linear-segmentation.patch
rename : patches/linux-2.6.17/pmd-shared.patch => patches/linux-2.6.18/pmd-shared.patch
rename : patches/linux-2.6.17/rename-TSS_sysenter_esp0-SYSENTER_stack_esp0.patch => patches/linux-2.6.18/rename-TSS_sysenter_esp0-SYSENTER_stack_esp0.patch
rename : patches/linux-2.6.17/series => patches/linux-2.6.18/series
rename : patches/linux-2.6.17/x86-elfnote-as-preprocessor-macro.patch => patches/linux-2.6.18/x86-elfnote-as-preprocessor-macro.patch
rename : patches/linux-2.6.17/x86-put-note-sections-into-a-pt_note-segment-in-vmlinux.patch => patches/linux-2.6.18/x86-put-note-sections-into-a-pt_note-segment-in-vmlinux.patch
rename : patches/linux-2.6.17/x86_64-put-note-sections-into-a-pt_note-segment-in-vmlinux.patch => patches/linux-2.6.18/x86_64-put-note-sections-into-a-pt_note-segment-in-vmlinux.patch
rename : patches/linux-2.6.17/xen-hotplug.patch => patches/linux-2.6.18/xen-hotplug.patch
rename : patches/linux-2.6.17/xenoprof-generic.patch => patches/linux-2.6.18/xenoprof-generic.patch

Export XEN_TARGET_ARCH to buildconfig sub-Makefiles even if it is not
specified on the command line.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] A few emulator fixes.
Signed-off-by: Keir Fraser <keir@xensource.com>

[MINIOS] New make structure to support different architectures.
Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>

[HVM] i8254: Fix bogus use of current

The function pit_load_count incorrectly references current for
determining whether the channel number is zero.  This breaks
when starting a new guest because current points to dom0.

The fix is to explicitly pass the address for verification.

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

linux: Allow microcode updates of smaller than default size.
Signed-off-by: Kurt Garloff <kurt@garloff.de>

Microcode does not need to have the default size of 2000+48 bytes.
A corresponding patch has gone into Linux 2.6.19; this is the
port to Xen.

Signed-off-by: Kurt Garloff <kurt@garloff.de>
Reference: Patch to linux kernel from Sep 27 2006
> # User Shaohua Li <shaohua.li@intel.com>
> # Node ID 45898b908138b5d93c2cc7353f061ce54af145dc
> # Parent  f962eab7b82c9bf1a6da69571046e764f5128395
> [PATCH] x86 microcode: don't check the size
>
> IA32 manual says if micorcode update's size is 0, then the size is
> default size (2048 bytes). But this doesn't suggest all microcode
> update's size should be above 2048 bytes to me. We actually had a
> microcode update whose size is 1024 bytes. The patch just removed
> the check.
>
> Signed-off-by: Shaohua Li <shaohua.li@intel.com>
> Cc: Tigran Aivazian <tigran@veritas.com>
> Signed-off-by: Andrew Morton <akpm@osdl.org>
> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
>
> committer: Linus Torvalds <torvalds@g5.osdl.org> 1159370778 -0700

[XEND] Fix warning about 'cpus' option on domain creation.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

Add scheduler name to 'xm info' output.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>

linux/i386: enhance dump_fault_path() in the highpte case

As long as the pte page isn't really located in highmem, there is no
reason to not access it, in order to print the complete page table
hierarchy. A functionally similar patch will go to lkml for native
Linux.

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

[XEN] Fix PAE writable-pte fixup to only trigger on writes to upper
half the the PAE PTE.
Signed-off-by: Jan Beulich <jbeulich@novell.com>

Document vfb configuration.

Signed-off-by: Markus Armbruster <armbru@redhat.com>

[XEN] Include byteorder functions from Linux.
Signed-off-by: Keir Fraser <keir@xensource.com>

[LIBFSIMAGE] Fix makefile to track version number change
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[IA64] Paravirtualize syscall path in file fsys.S

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

Remove useless fsip_fs_free(). Delegate freeing of private data to its plugin
users. Miscellaneous tidy.

Signed-off-by: John Levon <john.levon@sun.com>

[XEN] Fix early-unshadow detection for 3- or 4-level guest pagetables.
Early-unshadow will unshadow whenever 2 zero values are written
to the same page; for PAE, one PTE takes 2 writes.  Only check
for the early unshadow when writing the low half of the PTE.

[IA64] Add fsys.S to sparse tree

From 2.6.16.33

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

[IA64] Remove dorfirfi completely

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

[IA64] Add privilege check back for hypercall

hypercall is allowed when cpl=2

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

[IA64] Change vCPU initialization to avoid domVTi privregs memory leak

1) This patch moved some processing from vcpu_initialise() and
   added a new function vcpu_late_initialise().
   It executes the following initializations for VCPU of
   dom0/domU.
    - Allocate the VHPT
    - Allocate the privregs area and assign these pages into
      guest pseudo physical address space.
    - Set the tlbflush_timestamp.

   It is executed in the following sequence.

   dom0:
     start_kernel()
       ->domain_create()
       ->alloc_vcpu(VCPU0)
         ->alloc_vcpu_struct(VCPU0)
         ->vcpu_initialise(VCPU0)
       ->vcpu_late_initialise(VCPU0)

       ->construct_dom0
         ->alloc_vcpu(othe VCPUs)
           ->alloc_vcpu_struct(other VCPUs)
           ->vcpu_initialise(other VCPUs)

     ia64_hypercall(FW_HYPERCALL_IPI)
       ->fw_hypercall_ipi(XEN_SAL_BOOT_RENDEZ_VEC)
         ->arch_set_info_guest(other VCPUs)
           ->vcpu_late_initialise(other VCPUs)

   domU:
     do_domctl(XEN_DOMCTL_createdomain)
       ->domain_create()

     do_domctl(XEN_DOMCTL_max_vcpus)
       ->alloc_vcpu(all VCPUs)
         ->alloc_vcpu_struct(all VCPUs)
         ->vcpu_initialise(all VCPUs)

     do_domctl(XEN_DOMCTL_setvcpucontext)
       ->set_info_guest(VCPU0)
         ->arch_set_info_guest(VCPU0)
           ->vcpu_late_initialise(VCPU0)

     ia64_hypercall(FW_HYPERCALL_IPI)
       ->fw_hypercall_ipi(XEN_SAL_BOOT_RENDEZ_VEC)
         ->arch_set_info_guest(other VCPUs)
           ->vcpu_late_initialise(other VCPUs)

2) This patch modified the domain_set_shared_info_va().
   Currently, initialization of arch.privregs->interrupt_mask_addr
   of all VCPUs is executed in domain_set_shared_info_va().
   However, allocation of privregs area is late by modified of 1).
   Therefore, this patch modified initialization of
   arch.privregs->interrupt_mask_addr to the following sequence.

   dom0 and domU:
     ia64_hypercall(FW_HYPERCALL_SET_SHARED_INFO_VA)
       ->domain_set_shared_info_va()
         Initialize interrupt_mask_addr of VCPU0

     ia64_hypercall(FW_HYPERCALL_IPI)
       ->fw_hypercall_ipi(XEN_SAL_BOOT_RENDEZ_VEC)
         ->arch_set_info_guest(other VCPUs)
           ->vcpu_late_initialise(other VCPUs)
           Initialize interrupt_mask_addr of other VCPUs

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

[HVM] save restore: vcpu context support

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
save/restore HVM vcpu context such as vmcs

[HVM] save restore: save restore dev in HV

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
save/restore all dev state in HV such as PIT/PIC/APIC

[HVM] save restore: new hyper-call

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
add a pair of hyper-call for hvm guest context

[HVM] save restore: frame work

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
frame work for HVM save restore in Control Panel

[linux config] enable CONFIG_SCSI_SATA_AHCI/CONFIG_BNX2 in dom0 config.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

[XEND] Change SrvServer to not require sockets to be non-blocking.

Set all spawned server threads to be daemonic and make sure their
sockets are closed properly using shutdown() and close(), otherwise
reload will complain about the "Address still in use."

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

[linux config] enable CONFIG_SCSI_SATA_AHCI in dom0 config.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

[linux] remove gratuitous differences with plain linux.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>

Solaris should used a netback of type 'SUNW_mac' and 'vif-vnic' as the
vif-script by default.

Signed-off-by: David Edmondson <dme@sun.com>

Allow xend configuration via Solaris SMF as well as the config file.

Signed-off-by: John Levon <john.levon@sun.com>

Rename XendRoot to XendOptions.

Signed-off-by: John Levon <john.levon@sun.com>

[XEND] Fix order of kernel argument construction to prevent arg repeat bug

Make sure user specified kernel args are appended at the end but also
make sure that we do not duplicate root= and ip=.

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

Enable lazy (on-demand) allocation of memory to a guest being restored; this
means that ballooned down domains only require as much memory as is currently
being used (rather than their max) when being restored from save, or when
being migrated.

Signed-off-by: Steven Hand <steven@xensource.com>

merge with xen-unstable.hg

Remove bogus sanity check from xc_linux_save.

(a guest from another host (or an earlier time) can quite validly have
a max_mfn bigger the current max on this host)

Signed-off-by: Steven Hand <steven@xensource.com>

Comment fix.

Signed-off-by: Steven Hand <steven@xensource.com>

[PYGRUB] Make the boot-choice timeout actually time out.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>

[HVM] Fix calculation of address size in long mode.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Allow stack-address-size to be specified differently from
regular address-size in the emulator.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] BSF/BSR set ZF not CF.
Signed-off-by: Keir Fraser <keir@xensource.com>

[MINIOS] Clean up final link stage of the build.

Allow extra application objects to be linked against a partial link of
minios with non-global names localised. For now we assume that global
names ill be assigned a xenos_ prefix in due course. We may decide to
choose a different prefix.

Those who wish to finish the link from an external Makefile can link
their own objects against mini-os.o (which must be first in link
order!) and specify the appropriate architecture linker script.

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

Fix clean build after cset 13345:648e58e31c26

Signed-off-by: Steven Hand <steven@xensource.com>

[IA64] Fix save/restore

Using the new hypercall path, scratch registers are not saved/restored.

So after returning from hypercall, r20 may contain garbage data.
This triggered General Exception.

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

[IA64] Fix multicall hypercall

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

[XEN] Emulate BSF/BSR. Fix ARPL emulation.
Fix building POPA emulation on x86/32.
Fix asm constraints for in-memory operands.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Emulate AMD-defined two-byte NOP opcodes with ModRM.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Emulate PUSHA/POPA/ARPL/BOUND.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Fix some e820 start-of-day issues by clipping all E820_RAM
regions to be page-sized and -aligned.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Emulate MUL/DIV. Tweak test suite build.
Signed-off-by: Keir Fraser <keir@xensource.com>

[TESTS] Fix native jump to 1MB absolute address in emulator test.
Fix typo in emulator comment.
Signed-off-by: Keir Fraser <keir@xensource.com>

[XEN] Clean up long division code, fix for C99-mandated
truncation-towards-zero.
Signed-off-by: Keir Fraser <keir@xensource.com>

[IA64] Implement hyper_get_psr

Mov from psr is used frequently by xeno.

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

[IA64] Fix xencomm_arch_hypercall_suspend

This hypercall is changed to new hypercall mechanism

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