xen.git
18 years ago[HVM] Qemu rtl8139: correct ring-buffer wrapping.
Tim Deegan [Mon, 16 Jul 2007 13:27:12 +0000 (14:27 +0100)]
[HVM] Qemu rtl8139: correct ring-buffer wrapping.
The rx DMA should never overrun when it hits the end of a 64k buffer
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years ago[XEN] HAP: need to intercept CR0 even when using NPT
Tim Deegan [Mon, 16 Jul 2007 09:36:52 +0000 (10:36 +0100)]
[XEN] HAP: need to intercept CR0 even when using NPT
because we tinker with CR0.TS.  Thanks to Wei Huang for pointing this out.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years agoRemove unused function reserve_boot_memory() and alloc_boot_low_page().
Keir Fraser [Sat, 14 Jul 2007 09:26:02 +0000 (10:26 +0100)]
Remove unused function reserve_boot_memory() and alloc_boot_low_page().
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxen: Fix 32-bit boot.
Keir Fraser [Sat, 14 Jul 2007 07:19:17 +0000 (08:19 +0100)]
xen: Fix 32-bit boot.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoRemove hard tabs.
kfraser@localhost.localdomain [Fri, 13 Jul 2007 16:08:05 +0000 (17:08 +0100)]
Remove hard tabs.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86/32: Limit bootstrap 1:1 mapping to 1GB.
kfraser@localhost.localdomain [Fri, 13 Jul 2007 14:42:27 +0000 (15:42 +0100)]
x86/32: Limit bootstrap 1:1 mapping to 1GB.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoImplement domctl_sendtrigger_nmi for x86 hvm.
kfraser@localhost.localdomain [Fri, 13 Jul 2007 12:49:08 +0000 (13:49 +0100)]
Implement domctl_sendtrigger_nmi for x86 hvm.

This patch implements DOMCTL_SENDTRIGGER_NMI hypercall for x86 hvm
domains. We can manually inject NMI into a domain by using
'xm trigger [domain] nmi [vcpu]'.

Signed-off-by: Yosuke Iwamatsu <y-iwamatsu@ab.jp.nec.com>
18 years agoDo not clobber AMD TSC offset on real-mode switch.
kfraser@localhost.localdomain [Fri, 13 Jul 2007 12:46:49 +0000 (13:46 +0100)]
Do not clobber AMD TSC offset on real-mode switch.
Signed-off-by: David Lively <dlively@virtualiron.com>
Signed-off-by: Ben Guthro <bguthro@virtualiron.com>
18 years ago[Xend] Fix problem when destroying VMs using the Xen-API
kfraser@localhost.localdomain [Fri, 13 Jul 2007 12:43:29 +0000 (13:43 +0100)]
[Xend] Fix problem when destroying VMs using the Xen-API

When domains are destroyed/suspend-resume/migration using the Xen-API
things break due to the domid not having been reset. This patch fixes
this. This is the best place I found for fixing this problem. I could
not push this line into _stateSet() for the case of DOM_STATE_HALTED
and left a comment regarding this.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years ago[Build] Remove ACM option for NULL and STE+CHWALL policy
kfraser@localhost.localdomain [Fri, 13 Jul 2007 12:42:41 +0000 (13:42 +0100)]
[Build] Remove ACM option for NULL and STE+CHWALL policy
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years ago[ACM-security] Some fixes to tools.
kfraser@localhost.localdomain [Thu, 12 Jul 2007 15:07:28 +0000 (16:07 +0100)]
[ACM-security] Some fixes to tools.

- Allow multiple ChineseWallTypes in a VM labels
- check for surfacing exceptions in the python code
- check for array sizes in the XML DOM in python
- properly display the labels when doing 'xm list --label' in xm's
non-Xen-API mode
- rely on the security checking hooks in xend to check access to the
block interface rather than doing this in xm.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoFix PV-on-HVM driver build.
kfraser@localhost.localdomain [Thu, 12 Jul 2007 15:03:41 +0000 (16:03 +0100)]
Fix PV-on-HVM driver build.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[LIBXC] Fix xc_map_grant_refs() function.
kfraser@localhost.localdomain [Thu, 12 Jul 2007 13:14:29 +0000 (14:14 +0100)]
[LIBXC] Fix xc_map_grant_refs() function.

This patch fixes the xc_map_grant_refs() function, which would fail
because, instead of passing the malloc()-ed ioctl argument pointer to
the ioctl, the address of this pointer was being passed instead.

Signed-off-by: Derek Murray <Derek.Murray@cl.cam.ac.uk>
18 years agox86: Various cleanups around CR4 handling, cpu_possible_map, and VMX initialisation.
kfraser@localhost.localdomain [Thu, 12 Jul 2007 10:49:02 +0000 (11:49 +0100)]
x86: Various cleanups around CR4 handling, cpu_possible_map, and VMX initialisation.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[Docs] Documentation of extension of the Xen-API for managing security policies
kfraser@localhost.localdomain [Thu, 12 Jul 2007 09:06:44 +0000 (10:06 +0100)]
[Docs] Documentation of extension of the Xen-API for managing security policies

Provides an extension to the Xen-API documentation with the classes
and methods implemented for supporting of managing security policies
using the Xen-API.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoMerge with ia64 xen tree.
kfraser@localhost.localdomain [Thu, 12 Jul 2007 08:53:58 +0000 (09:53 +0100)]
Merge with ia64 xen tree.

18 years ago[IA64] Fix VTi domain shutdown
Alex Williamson [Wed, 11 Jul 2007 17:32:30 +0000 (11:32 -0600)]
[IA64] Fix VTi domain shutdown

Event channel destruction is now done earlier, this patch does
the corresponding modification in IA64 side. It fixes dom0 crash
when VTI destroyed.

Signed-off-by: Zhang Xin <xing.z.zhang@intel.com>
18 years agoConvert __init into __devinit in wakeup path.
kfraser@localhost.localdomain [Wed, 11 Jul 2007 16:57:47 +0000 (17:57 +0100)]
Convert __init into __devinit in wakeup path.

Need to ensure all the code slice in the wakeup path still
existing. For this purpose, we have to use __devinit instead
of __init, since the former is null for CONFIG_HOTPLUG while
the latter always causes related code to be free-ed after
first boot.

Later when adding __init to some function, be sure to consider
wakeup case and cpu hotplug!

Signed-off-by <Kevin.Tian@intel.com>

18 years agoProvide cpu hotplug support to Xen. Note this hotplug
kfraser@localhost.localdomain [Wed, 11 Jul 2007 16:48:10 +0000 (17:48 +0100)]
Provide cpu hotplug support to Xen. Note this hotplug
support is specific to PM, instead of for a run-time
single CPU hotplug which can be a separate task. See
embedded comment:

/*
 * XXX: One important thing missed here is to migrate vcpus
 * from dead cpu to other online ones and then put whole
 * system into a stop state. It assures a safe environment
 * for a cpu hotplug/remove at normal running state.
 *
 * However for xen PM case, at this point:
 *     -> All other domains should be notified with PM event,
 *        and then in following states:
 *             * Suspend state, or
 *             * Paused state, which is a force step to all
 *               domains if they do nothing to suspend
 *     -> All vcpus of dom0 (except vcpu0) have already beem
 *        hot removed
 * with the net effect that all other cpus only have idle vcpu
 * running. In this special case, we can avoid vcpu migration
 * then and system can be considered in a stop state.
 *
 * So current cpu hotplug is a special version for PM specific
 * usage, and need more effort later for full cpu hotplug.
 * (ktian1)
 */

Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoPull necessary Linux CPU hotplug logic into Xen. Due to
kfraser@localhost.localdomain [Wed, 11 Jul 2007 16:28:09 +0000 (17:28 +0100)]
Pull necessary Linux CPU hotplug logic into Xen. Due to
distinct difference on basic hierarchy and features,
this pull is done on function level.

Signed-off-by Kevin Tian <kevin.tian@intel.com>

18 years agoAdd HVM hardware feature suspend/resume.
kfraser@localhost.localdomain [Wed, 11 Jul 2007 16:23:09 +0000 (17:23 +0100)]
Add HVM hardware feature suspend/resume.
Signed-off-by: Ke Yu <ke.yu@intel.com>
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoProvide basic Xen PM infrastructure
kfraser@localhost.localdomain [Wed, 11 Jul 2007 14:47:14 +0000 (15:47 +0100)]
Provide basic Xen PM infrastructure

Basic infrastructure for Xen S3 support with a common CPU
context save/restore logic for both 32bit and 64bit.

Wakeup code is split into two parts:
    - the first locates after trampoline code, to share all the
      tricks on the latter, like relocation base and identy mapping
    - the 2nd part locates in xen code segment, to do the actual
      CPU context restore

Signed-off-by Ke Yu <ke.yu@intel.com>
Signed-off-by Kevin Tian <kevin.tian@intel.com>

18 years agohvm: Re-introduce LPT1 device in ACPI tables.
kfraser@localhost.localdomain [Wed, 11 Jul 2007 12:49:11 +0000 (13:49 +0100)]
hvm: Re-introduce LPT1 device in ACPI tables.

Once the presence of LPT1 is baked into a Windows guest, it gets upset
if it is subsequently removed. Hence we add the 'LTP1' device object
back into the ACPI DSDT, and remove IRQ 7 from the list of IRQs usable
by the PCI-ISA links.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAdd new domctl hypercall to expose current heap values. This
kfraser@localhost.localdomain [Wed, 11 Jul 2007 12:03:57 +0000 (13:03 +0100)]
Add new domctl hypercall to expose current heap values.  This
functionality is needed for probing how much memory is available in a
given node prior to VM creation.

Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[VTPM] Fixed typo in VTPM manager.
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:56:52 +0000 (10:56 +0100)]
[VTPM] Fixed typo in VTPM manager.

Replaces an instance of pthread_mutex_lock() that should be
pthreaad_mutex_unlock(). With thanks to our OpenTC partners at the
Technical University of Sofia, whose static analysis found this bug.

Signed-off-by: Derek G. Murray <Derek.Murray@cl.cam.ac.uk>
18 years ago[xend] Fix xm trigger command for inactive managed domains
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:54:06 +0000 (10:54 +0100)]
[xend] Fix xm trigger command for inactive managed domains
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoxend: Fix 'xm dumpcore' for inactive managed domains.
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:53:32 +0000 (10:53 +0100)]
xend: Fix 'xm dumpcore' for inactive managed domains.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years ago[xenstore] adds a -w (wide) flag to xenstore-ls to support seeing full
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:52:19 +0000 (10:52 +0100)]
[xenstore] adds a -w (wide) flag to xenstore-ls to support seeing full
contents of xenstore entries. There is a bit of code cleanup as well
(snprintf vs. sprintf), one formatting fix, and comments).

There is no change to the behavior of xenstore-ls without -w.

Signed-off-by: John Zulauf <john.zulauf@intel.com>
18 years ago[XM] Tools support for extensions of the Xen-API for managing security policies
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:49:43 +0000 (10:49 +0100)]
[XM] Tools support for extensions of the Xen-API for managing security policies

This patch adds a couple of new commands for using the Xen-API
extensions for security policies. Older tools are converted to support
going through the Xen-API for their operations rather than doing the
operations directly in their own code.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoMake QEMU consistently report write caching support for emulated IDE
kfraser@localhost.localdomain [Wed, 11 Jul 2007 09:48:15 +0000 (10:48 +0100)]
Make QEMU consistently report write caching support for emulated IDE
drives to fix a hang during SLES 9 HVM guest installation.

Without this, the SLES 9 installer kernels (32 and 64 bit) were
getting inconsistent information from QEMU as to whether the
(emulated) IDE drives support write caching (which they do).  So part
of the kernel thought write caching was enabled (and enabled the usage
of barrier writes) and part of it didn't, which triggered a bug in
which the same barrier write is submitted over and over again ...

Fixed by setting another bit in the WIN_IDENTIFY (IDE drive "identify"
command) response to indicate we really, truly support write caching.

Signed-off-by: David Lively <dlively@virtualiron.com>
Signed-off-by: Ben Guthro <bguthro@virtualiron.com>
18 years agovmx: Stupid typo in BUG_ON() stmt.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 22:42:13 +0000 (23:42 +0100)]
vmx: Stupid typo in BUG_ON() stmt.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agovmx: Never use physical addresses above 4GB for VMCS state on i386.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 22:37:42 +0000 (23:37 +0100)]
vmx: Never use physical addresses above 4GB for VMCS state on i386.

Thsi requires special allocation of the vlapic regs page, but does let
us get rid of some top-half writes to a few VMCS fields.

Assert a few more facts about the VMX_BASIC_MSR.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] Fixup physinfo
Alex Williamson [Tue, 10 Jul 2007 17:15:54 +0000 (11:15 -0600)]
[IA64] Fixup physinfo

Use max cpus per node to guess at sockets per node.  This avoids
averaging problems with offline cpus and nodes without cpus.  Also
fill in the cpu_to_node array.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years agoSimplify vmx host state setup code.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 14:45:44 +0000 (15:45 +0100)]
Simplify vmx host state setup code.
Signed-off-by: Xin Li <xin.b.li@intel.com>
18 years agoxenoprof: Fix initialisation. Much can be done at boot-time, as
kfraser@localhost.localdomain [Tue, 10 Jul 2007 14:41:26 +0000 (15:41 +0100)]
xenoprof: Fix initialisation. Much can be done at boot-time, as
indicated by the fact that the low-level functiosn are marked
'__init'.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agomerge with xen-unstable.hg
Alex Williamson [Tue, 10 Jul 2007 14:39:26 +0000 (08:39 -0600)]
merge with xen-unstable.hg

18 years ago[libxen] C-bindings for the extensions of the Xen-API for managing security policies
kfraser@localhost.localdomain [Tue, 10 Jul 2007 13:57:04 +0000 (14:57 +0100)]
[libxen] C-bindings for the extensions of the Xen-API for managing security policies
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agohvm: Fix memcpy_words() asm in ioemu.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 13:42:23 +0000 (14:42 +0100)]
hvm: Fix memcpy_words() asm in ioemu.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoRe-factor e820 truncation code and reintroduce clipping for 32-on-64
kfraser@localhost.localdomain [Tue, 10 Jul 2007 10:10:38 +0000 (11:10 +0100)]
Re-factor e820 truncation code and reintroduce clipping for 32-on-64
guests (to 166GB).
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoXen API specification improvements.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 09:22:41 +0000 (10:22 +0100)]
Xen API specification improvements.

Move the error list to be more prominent, and explicitly document
which generic error codes can be returned by all methods.

Mark host CPU flags/features return value as unspecified.

Document PERMISSION_DENIED.

Document error returns of session.login_with_password.

Signed-off-by: John Levon <john.levon@sun.com>
18 years agoxend: Fix xm pause/unpause for inactive managed domains.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 09:15:01 +0000 (10:15 +0100)]
xend: Fix xm pause/unpause for inactive managed domains.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoxend: Count CPUs rather than relying on a static
kfraser@localhost.localdomain [Tue, 10 Jul 2007 09:13:16 +0000 (10:13 +0100)]
xend: Count CPUs rather than relying on a static
nodes/sockets/cores/threads relationship.
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years agoxm: Improve help message for vif creation.
kfraser@localhost.localdomain [Tue, 10 Jul 2007 09:11:14 +0000 (10:11 +0100)]
xm: Improve help message for vif creation.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoSimplify Unisys ES7000 code in light of the fact we do not support
kfraser@localhost.localdomain [Tue, 10 Jul 2007 09:07:00 +0000 (10:07 +0100)]
Simplify Unisys ES7000 code in light of the fact we do not support
legacy boxes with very non-standard APIC setup.
From: Raj Subrahmanian <raj.subrahmanian@unisys.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxend: Missing last line from Xen_API security patch.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 18:28:54 +0000 (19:28 +0100)]
xend: Missing last line from Xen_API security patch.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoAdd accel option to vif xend config.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 17:10:48 +0000 (18:10 +0100)]
Add accel option to vif xend config.
Signed-off-by: Kieran Mansley <kmansley@solarflare.com>
18 years agohvm: Clean ups and fix MSR access functions.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 17:04:23 +0000 (18:04 +0100)]
hvm: Clean ups and fix MSR access functions.
Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] Support special guest optimizations in the hypervisor
Alex Williamson [Mon, 9 Jul 2007 15:22:58 +0000 (09:22 -0600)]
[IA64] Support special guest optimizations in the hypervisor

For PV add a new hypercall __HYPERVISOR_opt_feature, which lets dom0/domU
switch on/off special optimization features in the hypervisor.
For HVM this is done in set_os_type().

Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>
18 years ago[IA64] Rename NUM_MEM_DESC to NUM_EXTRA_MEM_DESC and reduce it.
Alex Williamson [Mon, 9 Jul 2007 14:33:02 +0000 (08:33 -0600)]
[IA64] Rename NUM_MEM_DESC to NUM_EXTRA_MEM_DESC and reduce it.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years ago[Xen-API] Extension of the Xen-API for managing Xen Security Policies
kfraser@localhost.localdomain [Mon, 9 Jul 2007 13:51:44 +0000 (14:51 +0100)]
[Xen-API] Extension of the Xen-API for managing Xen Security Policies

This patch implements extensions for managing security policies in
xend. The XSPolicy and ACMPolicy classes provide the interface for the
Xen-API and implement functionality for setting, updating and
activating of a Xen security policy as well as labeling of virtual
machines and resources such as block devices. Labeling of network
devices will follow.

The acmpolicy class implements a compiler for translating an XML
policy into their binary format and provides functionality for
comparing a current policy against a new one when changing/updating a
policy.

The xspolicyadmin class administers the policy of the system.

Some of the xend-internal code deals with transforming the labeling
information from the S-Expression format into the new Xen-API
format. This is similar to much of the other code that is already
there.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoIntroduce 'bootscrub' boolean cmdline option. Default on.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 13:30:46 +0000 (14:30 +0100)]
Introduce 'bootscrub' boolean cmdline option. Default on.

no-bootscrub will disable boot-time RAM scrubbing, speeding up boot
time significantly on big-memory systems.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAlso allow boolean cmdline params to be inverted in two other ways.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 13:29:53 +0000 (14:29 +0100)]
Also allow boolean cmdline params to be inverted in two other ways.

Now, a standard boolean (e.g., watchdog) can be inverted in three
ways:
 1. no-watchdog
 2. watchdog=no
 3. watchdog=off

Stacked inversions cancel each other: no-watchdog=no turns on the
watchdog.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAllow inversion of boolean cmdline parameters with 'no-' prefix.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 13:13:56 +0000 (14:13 +0100)]
Allow inversion of boolean cmdline parameters with 'no-' prefix.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoRemove no-pv-compat cmdline option.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 13:06:22 +0000 (14:06 +0100)]
Remove no-pv-compat cmdline option.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAdd support for building -mm, -tip and -rc upstream trees.
Ian Campbell [Mon, 9 Jul 2007 11:04:20 +0000 (12:04 +0100)]
Add support for building -mm, -tip and -rc upstream trees.

(note that not all of those contains Xen support just yet and that a
native kernel will be silently built in that case.)

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years agoAdd support to the tarball source type for using the ketchup tool.
Ian Campbell [Mon, 9 Jul 2007 10:57:07 +0000 (11:57 +0100)]
Add support to the tarball source type for using the ketchup tool.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years agoAdd hooks for a script to update kernel configuration after tree has been
Ian Campbell [Mon, 9 Jul 2007 10:29:39 +0000 (11:29 +0100)]
Add hooks for a script to update kernel configuration after tree has been
prepared. This is to support upstream sources which do not enable Xen in their
default configurations.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years agoRemove .hgskip. This was an ancient mercurial patch which was never
Ian Campbell [Mon, 9 Jul 2007 10:29:30 +0000 (11:29 +0100)]
Remove .hgskip. This was an ancient mercurial patch which was never
taken upstream.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years agoVMX: Fix function return code when a GP fault is generated on CR accesses.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 08:47:20 +0000 (09:47 +0100)]
VMX: Fix function return code when a GP fault is generated on CR accesses.
Signed-off-by: Xin Li <xin.b.li@intel.com>
18 years agoFix xm save command for a inactive managed domain.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 08:41:45 +0000 (09:41 +0100)]
Fix xm save command for a inactive managed domain.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoVMX: No VMExits on SYSENTER MSR accesses since VMX already has VMCS
kfraser@localhost.localdomain [Mon, 9 Jul 2007 08:39:51 +0000 (09:39 +0100)]
VMX: No VMExits on SYSENTER MSR accesses since VMX already has VMCS
fields for both host and guest state.
Signed-off-by: Xin Li <xin.b.li@intel.com>
18 years agoBetter error message for 'xm sysrq' on inactive domain.
kfraser@localhost.localdomain [Mon, 9 Jul 2007 08:38:35 +0000 (09:38 +0100)]
Better error message for 'xm sysrq' on inactive domain.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoSVM: Sync with VMX code changes.
Keir Fraser [Sat, 7 Jul 2007 10:21:34 +0000 (11:21 +0100)]
SVM: Sync with VMX code changes.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agovms: A cleanup to vmx.c:
Keir Fraser [Sat, 7 Jul 2007 10:08:57 +0000 (11:08 +0100)]
vms: A cleanup to vmx.c:
1) move some functions from vmx.h to vmx.c, since they are private.
2) adjust vmx_set_cr0 handling.
3) majorly coding style tidying.
Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoSpecify -fno-strict-aliasing in root build config file.
Keir Fraser [Sat, 7 Jul 2007 09:31:15 +0000 (10:31 +0100)]
Specify -fno-strict-aliasing in root build config file.

Fixes correctness issues with xenstored and gcc-4.2.

Original patch by Charles Coffing <ccoffing@novell.com>

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoMerge with PPC tree.
Keir Fraser [Sat, 7 Jul 2007 09:09:51 +0000 (10:09 +0100)]
Merge with PPC tree.

18 years ago[POWERPC][XEN] Stop sharing the MPIC between dom0 and Xen.
Hollis Blanchard [Fri, 6 Jul 2007 21:25:10 +0000 (16:25 -0500)]
[POWERPC][XEN] Stop sharing the MPIC between dom0 and Xen.
Xen is now the sole owner of the MPIC. Requires Xen Linux 2.6.18.
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
18 years ago[POWERPC][XEN] Increase foreign mapping size to accommodate Linux 2.6.18 usage.
Hollis Blanchard [Fri, 6 Jul 2007 21:21:39 +0000 (16:21 -0500)]
[POWERPC][XEN] Increase foreign mapping size to accommodate Linux 2.6.18 usage.
Signed-off-by: Christian Ehrhardt <ehrhardt@linux.vnet.ibm.com>
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
18 years ago[POWERPC][XEN] Quiet time debug messages.
Hollis Blanchard [Fri, 6 Jul 2007 21:19:36 +0000 (16:19 -0500)]
[POWERPC][XEN] Quiet time debug messages.
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
18 years ago[XEN][POWERPC] Merge with xen-unstable.hg.
Hollis Blanchard [Fri, 6 Jul 2007 21:01:04 +0000 (16:01 -0500)]
[XEN][POWERPC] Merge with xen-unstable.hg.
Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
18 years ago[PPC] Add missing '{' in conditional statement.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 15:34:51 +0000 (16:34 +0100)]
[PPC] Add missing '{' in conditional statement.
Signed-off-by: Jerone Young <jyoung5@us.ibm.com>
18 years agohvm: Add global default keymap setting
kfraser@localhost.localdomain [Fri, 6 Jul 2007 15:16:57 +0000 (16:16 +0100)]
hvm: Add global default keymap setting

Add global default keymap setting to be used when a VM does not have
a keymap configuration set.

Signed-off-by: Pat Campbell <plc@novell.com>
18 years agoExtended the physinfo sysctl to export NUMA cpu_to_node topology info.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 15:12:07 +0000 (16:12 +0100)]
Extended the physinfo sysctl to export NUMA cpu_to_node topology info.
Print this in 'xm info'.
Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoRevert 15471:7ac7f147241405af83e7a9d748cf7b01279734fc
kfraser@localhost.localdomain [Fri, 6 Jul 2007 14:01:20 +0000 (15:01 +0100)]
Revert 15471:7ac7f147241405af83e7a9d748cf7b01279734fc

Block-device specifiers in ioemu can contain colons, so skipping
always past the first colon is not a good idea. Better solutions are
in the pipeline to solve the blktap issues.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agodocs: update xm man page
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:45:02 +0000 (14:45 +0100)]
docs: update xm man page

- Fixed description of "Mem" column in "xm list" output.
- Added a bit of text for the credit scheduler.
- Described the --force option to block-detach.
- Made formatting and spelling more consistent.
- etc...

Signed-off-by: Charles Coffing <ccoffing@novell.com>
18 years ago[ACM] Support for initial unicode handling of strings in policies.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:43:51 +0000 (14:43 +0100)]
[ACM] Support for initial unicode handling of strings in policies.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
18 years agoExport debugger attached state to xc API.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:42:55 +0000 (14:42 +0100)]
Export debugger attached state to xc API.
Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years agoxend: Fix xm block/network-attach command for inactive managed domain
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:41:05 +0000 (14:41 +0100)]
xend: Fix xm block/network-attach command for inactive managed domain
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agoxend: Fix op_domain_sched_credit_set() in SrvDomain.py
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:39:21 +0000 (14:39 +0100)]
xend: Fix op_domain_sched_credit_set() in SrvDomain.py
This patch adds "cap" into op_domain_sched_credit_set().
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agovmx: Fix a typo.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:38:22 +0000 (14:38 +0100)]
vmx: Fix a typo.
Signed-off-by: Xin Li <xin.b.li@intel.com>
18 years agohvm: Fix vnc keymap for Japanese keyboards.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:37:47 +0000 (14:37 +0100)]
hvm: Fix vnc keymap for Japanese keyboards.
Signed-off-by: Takanori Kasai <kasai.takanori@jp.fujitsu.com>
Signed-off-by: Junko Ichino <ichino.junko@jp.fujitsu.com>
18 years agoUse short name format when reference to vcpu vmx union member.
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:36:34 +0000 (14:36 +0100)]
Use short name format when reference to vcpu vmx union member.
Signed-off-by: Xin Li <xin.b.li@intel.com>
18 years agoImprove xm sched-credit command for inactive managed domains
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:34:54 +0000 (14:34 +0100)]
Improve xm sched-credit command for inactive managed domains
Allows to change the scheduler parameters of inactive managed domains.
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agohvm: Allow blktap to be able to be booted as system volume for PV-on-HVM
kfraser@localhost.localdomain [Fri, 6 Jul 2007 13:22:59 +0000 (14:22 +0100)]
hvm: Allow blktap to be able to be booted as system volume for PV-on-HVM

When blktap was booting system volume, PV-on-HVM domain was not able
to be  started.

The configuration file is specified as follows.
disk = [ 'tap:aio:/xen/test/rhel5ga_full.img,hda,w' ]

The error occurred by the initialization of system volume in qemu-dm.
qemu: could not open hard disk image 'aio:/xen/test/rhel5ga_full.img'

Signed-off-by: Takanori Kasai <kasai.takanori@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[HVM] ROMBIOS: reboot when we see Ctrl+Alt+Del
Tim Deegan [Fri, 6 Jul 2007 10:17:01 +0000 (11:17 +0100)]
[HVM] ROMBIOS: reboot when we see Ctrl+Alt+Del
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years ago[IA64] Removed unused __get_free_pages() and __free_pages()
Alex Williamson [Thu, 5 Jul 2007 19:38:35 +0000 (13:38 -0600)]
[IA64] Removed unused __get_free_pages() and __free_pages()

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years ago[IA64] cleanup: remove dead code.
Alex Williamson [Thu, 5 Jul 2007 19:33:03 +0000 (13:33 -0600)]
[IA64] cleanup: remove dead code.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Minor reformating and comments for vmx_fault.c
Alex Williamson [Thu, 5 Jul 2007 19:30:25 +0000 (13:30 -0600)]
[IA64] Minor reformating and comments for vmx_fault.c

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Renames vmx_process.c into vmx_fault.c
Alex Williamson [Thu, 5 Jul 2007 19:24:42 +0000 (13:24 -0600)]
[IA64] Renames vmx_process.c into vmx_fault.c

Signed-off-by: Tristan Gingold <tgingold@free.fr>
--HG--
rename : xen/arch/ia64/vmx/vmx_process.c => xen/arch/ia64/vmx/vmx_fault.c

18 years ago[IA64] Makefile fix: do not always redo .S targets.
Alex Williamson [Thu, 5 Jul 2007 19:17:30 +0000 (13:17 -0600)]
[IA64] Makefile fix: do not always redo .S targets.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Clear the key part of cr.itir
Alex Williamson [Thu, 5 Jul 2007 19:15:44 +0000 (13:15 -0600)]
[IA64] Clear the key part of cr.itir

Clear the key part of cr.itir before itc.X in alt_dtlb_miss(), alt_itlb_miss()
and frametable_miss().  Preparation for using protection keys.

Signed-off-by: Dietmar Hahn <dietmar.hahn@fujitsu-siemens.com>
18 years ago[IA64] Fix soft lock up caused by xen_timer_interrupt()
Alex Williamson [Thu, 5 Jul 2007 19:04:00 +0000 (13:04 -0600)]
[IA64] Fix soft lock up caused by xen_timer_interrupt()

This patch intends to fix softlockup caused by xen_timer_interrupt().
This is caused by local_cpu_data->itm_next and stime_irq, itc_at_irq
inconsistency at CPU0 of hypervisor.  This patch sets stime_irq and
itc_at_irq every time in xen_timer_interrupt() to avoid this soft
lock up.

In other words, it is caused by competition of local_cpu_data->itm_next
and domain_itm in xen_timer_interrupt() and reprogram_timer() (more
specific vcpu_set_next_timer()).

For example:
 1) reprogram_timer() runs and set local_cpu_data->itm_next and set
    domain_itm as next itm.
 2) xen_timer_interrupt() called but following condition is not satisfied:
    while(time_after(ia64_get_itc(), local_cpu_data->itm_next)
    This skips stime_irq and itc_at_irq setting.
 3) goto 1)
 4) sometimes local_cpu_data->itm_next is rollback because
    ns_to_cycle()/IA64 is representing almost 32bit.
    (This occured at reprogram_timer())
 5) It causes soft lock up.
 6) Hypervisor returns to work(not hang).

To reproduce this issue, I do following configuration.

 1) boot Xen with pcpu=4 and Dom0 with vcpu=4
 2) boot domU1 with vcpu with vcpu-pin 0-1
 3) boot domU2 with vcpu with vcpu-pin 0-1
 4) run yes > /dev/null  2 process on domU1
 5) run nothing on domU2(to check softlock up occured or not)
 6) run kernel compile with -j4 on Dom0 continuously
 7) wait 4 or 8 hours to occur softlockup.

Signed-off-by: Atsushi SAKAI <sakaia@jp.fujitsu.com>
18 years ago[HVM] Shadow: avoid xen crash if guest uses special memory for pagetables
Tim Deegan [Thu, 5 Jul 2007 13:46:13 +0000 (14:46 +0100)]
[HVM] Shadow: avoid xen crash if guest uses special memory for pagetables
(just crash the guest and don't do any more PTE propagations).
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years ago[HVM] update Etherboot ROM: new version, stricter DHCP compliance.
Tim Deegan [Thu, 5 Jul 2007 09:19:25 +0000 (10:19 +0100)]
[HVM] update Etherboot ROM: new version, stricter DHCP compliance.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years agovmx: Write back modified interruptibility-info in enable_intr_window().
kfraser@localhost.localdomain [Wed, 4 Jul 2007 09:34:18 +0000 (10:34 +0100)]
vmx: Write back modified interruptibility-info in enable_intr_window().
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agohvm vmx: Support 'virtual NMI' feature of VMX.
kfraser@localhost.localdomain [Tue, 3 Jul 2007 17:46:00 +0000 (18:46 +0100)]
hvm vmx: Support 'virtual NMI' feature of VMX.
Signed-off-by: Haitao Shan <haitao.shan@intel.com>
Signed-off-by: Eddie Dong <eddie.dong@intel.com>
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxen: Fix event-channel destruction.
kfraser@localhost.localdomain [Tue, 3 Jul 2007 16:22:17 +0000 (17:22 +0100)]
xen: Fix event-channel destruction.

Lots of churn to fix a couple of bugs:

 1. If we tried to close an interdomain channel, and the remote domain
    is being destroyed, we immediately return success. But the remote
    domain will place our port in state 'unbound', not 'free'. Hence
    the port is effectively leaked.

 2. If two domains are being destroyed at the same time, and share an
    interdomain channel, the second to attempt the close may
    dereference an invalid domain pointer. Aiee!

Doing som ework to be able to destroy event-channel context much
earlier in domain death was the civilised thing to do.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[x86/32] Support two byte CMOVcc opcodes in 4G segment fixup.
Ian Campbell [Tue, 3 Jul 2007 16:04:50 +0000 (17:04 +0100)]
[x86/32] Support two byte CMOVcc opcodes in 4G segment fixup.

These are present in Debian Etch's libc6-i686 package.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years agox86: Clean up do_iret() hypercall.
kfraser@localhost.localdomain [Tue, 3 Jul 2007 15:02:31 +0000 (16:02 +0100)]
x86: Clean up do_iret() hypercall.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxenstore: Remove broken and unmaintained test code.
kfraser@localhost.localdomain [Tue, 3 Jul 2007 14:51:59 +0000 (15:51 +0100)]
xenstore: Remove broken and unmaintained test code.
Signed-off-by: Keir Fraser <keir@xensource.com>