xen.git
18 years agox86: Code style in rdmsr_hypervisor_regs() is not consistent with
kfraser@localhost.localdomain [Mon, 18 Jun 2007 15:52:04 +0000 (16:52 +0100)]
x86: Code style in rdmsr_hypervisor_regs() is not consistent with
wrmsr_hypervisor_regs(). Sync them up.
Signed-off-by: Shane Wang <shane.wang@intel.com>
18 years agox86/64: Avoid bogus mbi pointer into relocated Xen address space.
kfraser@localhost.localdomain [Mon, 18 Jun 2007 15:48:05 +0000 (16:48 +0100)]
x86/64: Avoid bogus mbi pointer into relocated Xen address space.
Prevent similar bugs in future by poisoning the relocated bottom
megabyte.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Boot trampoline cleanups suggested by Xin Li.
kfraser@localhost.localdomain [Mon, 18 Jun 2007 15:47:16 +0000 (16:47 +0100)]
x86: Boot trampoline cleanups suggested by Xin Li.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agodocs: Fix docs install when graphvix is not installed.
Keir Fraser [Mon, 18 Jun 2007 11:11:36 +0000 (12:11 +0100)]
docs: Fix docs install when graphvix is not installed.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxend: Revert 15252:31ee1768e911 ("Make device detach wait for detach to complete").
Keir Fraser [Sat, 16 Jun 2007 10:18:44 +0000 (11:18 +0100)]
xend: Revert 15252:31ee1768e911 ("Make device detach wait for detach to complete").

Has a bad interaction with 12520 in destroyDevice(), where devid is
not an integer as expected.

I'm not sure which of these changesets is really the bogus one, but
this one went in later so it loses. :-)

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoMerge with xen-ia64-unstable
Keir Fraser [Sat, 16 Jun 2007 09:42:06 +0000 (10:42 +0100)]
Merge with xen-ia64-unstable

18 years agomerge with xen-unstable.hg
Alex Williamson [Fri, 15 Jun 2007 19:33:11 +0000 (13:33 -0600)]
merge with xen-unstable.hg

18 years ago[IA64] Fix dom0vp_expose_p2m()
Alex Williamson [Fri, 15 Jun 2007 18:38:35 +0000 (12:38 -0600)]
[IA64] Fix dom0vp_expose_p2m()

It exposes one more zero p2m page

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years ago[IA64] RBS is now in vcpu_guest_context_regs.
Alex Williamson [Fri, 15 Jun 2007 18:36:04 +0000 (12:36 -0600)]
[IA64] RBS is now in vcpu_guest_context_regs.

As there is not anymore local variable of type vcpu_guest_context_regs in
the hypervisor the rbs field (16KB) can be put in the structure.
Code to read/write the field added.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Add error checking to nvram store mkdir
Alex Williamson [Fri, 15 Jun 2007 17:34:34 +0000 (11:34 -0600)]
[IA64] Add error checking to nvram store mkdir

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[HVM] HAP tidying.
Tim Deegan [Fri, 15 Jun 2007 15:51:08 +0000 (16:51 +0100)]
[HVM] HAP tidying.
Tighten up locking discipline, dead/useless code and unnecessary VMEXITS.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years agokexec: Flush console on kexec_crash().
Keir Fraser [Fri, 15 Jun 2007 15:30:10 +0000 (16:30 +0100)]
kexec: Flush console on kexec_crash().
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] Create NVRAM save directory
Alex Williamson [Fri, 15 Jun 2007 15:25:09 +0000 (09:25 -0600)]
[IA64] Create NVRAM save directory

If you use RPM to install XEN, the directory(/usr/lib/xen/boot)
will be removed when you un-install the RPM. This patch create
that directory before NVRAM file creating.

Signed-off-by: Zhang Xin <xing.z.zhang@intel.com>
18 years agohvm: Add locking to platform timers.
Keir Fraser [Fri, 15 Jun 2007 14:30:49 +0000 (15:30 +0100)]
hvm: Add locking to platform timers.
Handy for correctness.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoMake Xen headers lint-clean, and fix xs_wire.h for Solaris compiler.
Keir Fraser [Fri, 15 Jun 2007 10:50:54 +0000 (11:50 +0100)]
Make Xen headers lint-clean, and fix xs_wire.h for Solaris compiler.
Signed-off-by: John Levon <john.levon@sun.com>
Signed-off-by: Ryan Scott <ryan.scott@sun.com>
18 years agohvm: Simplify timer<->vcpu/domain conversion in RTC and PIT timer
Keir Fraser [Fri, 15 Jun 2007 09:36:31 +0000 (10:36 +0100)]
hvm: Simplify timer<->vcpu/domain conversion in RTC and PIT timer
models. These timers are always bound to vcpu0, where a specific vcpu
matters.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxend: Replace tabs with 8 whitespaces in DevController.py
Keir Fraser [Fri, 15 Jun 2007 09:01:32 +0000 (10:01 +0100)]
xend: Replace tabs with 8 whitespaces in DevController.py
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agohvm: Fix multiplication overflow in hvm/pmtimer.c
Keir Fraser [Fri, 15 Jun 2007 08:42:39 +0000 (09:42 +0100)]
hvm: Fix multiplication overflow in hvm/pmtimer.c

Too many ACPI events (SCI) are raised on hvm because of
multiplication overflow.

FREQUENCE_PMTIMER=3579545
(1000000000ULL << 32) / FREQUENCE_PMTIMER = 0xae9a7b1663a
pmt_cycles_until_flip =~ 0x80000000
0xae9a7b1663a*0x80000000 = overflow!!!

Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
18 years agoFix compile errors in PV-on-HVM drivers because of missing asm/gnttab_dma.h.
Keir Fraser [Fri, 15 Jun 2007 08:32:22 +0000 (09:32 +0100)]
Fix compile errors in PV-on-HVM drivers because of missing asm/gnttab_dma.h.
Signed-off-by: KUWAMURA Shin'ya <kuwa@jp.fujitsu.com>
18 years agoUpdate Solaris balloon ioctl value
Keir Fraser [Fri, 15 Jun 2007 08:18:42 +0000 (09:18 +0100)]
Update Solaris balloon ioctl value

The balloon ioctl value should use a 32-bit value instead of a 16-bit
value to distinguish itself from other ioctls.

Signed-off-by: Ryan Scott <ryan.scott@sun.com>
18 years agoRemove bogus xenfb include; breaks Solaris build.
Keir Fraser [Fri, 15 Jun 2007 08:17:49 +0000 (09:17 +0100)]
Remove bogus xenfb include; breaks Solaris build.
Signed-off-by: John Levon <john.levon@sun.com>
18 years agoES7000: Fix ES7000-specific APIC code.
Keir Fraser [Fri, 15 Jun 2007 08:15:28 +0000 (09:15 +0100)]
ES7000: Fix ES7000-specific APIC code.
Signed-off-by: Raj Subrahmanian <raj.subrahmanian@unisys.com>
18 years agohvm: Reintroduce hpet_migrate_timers(), as it is not covered by vpt
Keir Fraser [Fri, 15 Jun 2007 08:08:19 +0000 (09:08 +0100)]
hvm: Reintroduce hpet_migrate_timers(), as it is not covered by vpt
logic, but fix so that only VCPU0 migrates the timers.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] Don't try to save nvram on PV domains
Alex Williamson [Thu, 14 Jun 2007 21:32:22 +0000 (15:32 -0600)]
[IA64] Don't try to save nvram on PV domains

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64] Revive ACPI power management for IA64.
Alex Williamson [Thu, 14 Jun 2007 21:29:52 +0000 (15:29 -0600)]
[IA64] Revive ACPI power management for IA64.

c/s 14552 took PM1a_EVT_BLK registers from Qemu into Xen.
Also support ACPI PM timer function.

Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
18 years ago[IA64] symlink kregs.h into libxc build
Alex Williamson [Thu, 14 Jun 2007 21:13:59 +0000 (15:13 -0600)]
[IA64] symlink kregs.h into libxc build

Suggested by Akio Takebe

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64] Fix dom0 load address
Alex Williamson [Thu, 14 Jun 2007 20:36:15 +0000 (14:36 -0600)]
[IA64] Fix dom0 load address

Calculate load address of dom0 image based on xen_pstart rather than
address from ELF image.

Signed-off-by: Jes Sorensen <jes@sgi.com>
18 years agohvm: Timer device model cleanups.
kfraser@localhost.localdomain [Thu, 14 Jun 2007 17:01:42 +0000 (18:01 +0100)]
hvm: Timer device model cleanups.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoUpdate docs for 'edd=' boot option.
kfraser@localhost.localdomain [Thu, 14 Jun 2007 17:01:07 +0000 (18:01 +0100)]
Update docs for 'edd=' boot option.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agohvm svm: Fix for BSOD when "migrating" from Intel to AMD.
kfraser@localhost.localdomain [Thu, 14 Jun 2007 14:54:57 +0000 (15:54 +0100)]
hvm svm: Fix for BSOD when "migrating" from Intel to AMD.

The Intel processor driver for Windows (XP and later at least) reads
an MSR that AMD doesn't have. This causes GP-fault in kernel mode,
which causes blue-screen-of-death on Windows. This prevents a
disk-image that was installed on Intel from starting on an AMD
processor.

This patch "fixes" the problem by allowing reads from the msr,
returning all zero's (which is a valid, if not particulary
meaningful, value for this register).

Signed-off-by: Mats Petersson <mats.petersson@amd.com>
18 years agohvm vmx: Various cleanups and clarifications around event injection.
Keir Fraser [Wed, 13 Jun 2007 21:24:28 +0000 (22:24 +0100)]
hvm vmx: Various cleanups and clarifications around event injection.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxenapi: Implement VM.set_VCPUs_at_startup and VM.set_VCPUs_max XenAPI
kfraser@localhost.localdomain [Wed, 13 Jun 2007 10:28:13 +0000 (11:28 +0100)]
xenapi: Implement VM.set_VCPUs_at_startup and VM.set_VCPUs_max XenAPI
methods in xend.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
18 years agoxenapi: Implement XenAPI method VBD.set_mode in python.
kfraser@localhost.localdomain [Wed, 13 Jun 2007 10:27:47 +0000 (11:27 +0100)]
xenapi: Implement XenAPI method VBD.set_mode in python.
Signed-off-by: Jim Fehlig <jfehlig@novell.com>
18 years agoxenapi: Fix segfault in libxenapi.
kfraser@localhost.localdomain [Wed, 13 Jun 2007 10:27:08 +0000 (11:27 +0100)]
xenapi: Fix segfault in libxenapi.

When calling xen_vbd_set_mode(), libxenapi attempted to convert enum
mode parameter to a string twice - resulting in segfault.  Removed
first conversion since conversion is taking place in
marshalling/demarshalling layer. Fixed similar double enum conversion
in other places as well.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
18 years agoxentop: Fix -b option (batch mode).
kfraser@localhost.localdomain [Wed, 13 Jun 2007 10:13:11 +0000 (11:13 +0100)]
xentop: Fix -b option (batch mode).

Some curses libraries don't work with xentop in batch mode, avoid
those calls. Avoid the over-verbose summary lines when batch mode is
requested.

Signed-off-by: Gary Pennington <gary.pennington@sun.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAllow make targets to be called via make -C
kfraser@localhost.localdomain [Wed, 13 Jun 2007 10:05:47 +0000 (11:05 +0100)]
Allow make targets to be called via make -C

Change PWD to CURDIR in new mk.linux-2.6-xen to allow make targets to
be called via a make -C

From: Ben Guthro <bguthro@virtualiron.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: ioapic_write() must check for weird redir entries with special
kfraser@localhost.localdomain [Wed, 13 Jun 2007 09:39:48 +0000 (10:39 +0100)]
x86: ioapic_write() must check for weird redir entries with special
delivery mode yet apparently valid vectoring information.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoMerge with xen-unstable.hg (staging)
Alex Williamson [Wed, 13 Jun 2007 02:41:08 +0000 (20:41 -0600)]
Merge with xen-unstable.hg (staging)

18 years ago[IA64] Fix xen build warning
Alex Williamson [Tue, 12 Jun 2007 22:54:19 +0000 (16:54 -0600)]
[IA64] Fix xen build warning

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64] libxc trivial fix compilation error.
Alex Williamson [Tue, 12 Jun 2007 21:23:55 +0000 (15:23 -0600)]
[IA64] libxc trivial fix compilation error.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years ago[IA64] Provide pxm_to_node and node_to_pxm translation functions
Alex Williamson [Tue, 12 Jun 2007 21:22:02 +0000 (15:22 -0600)]
[IA64] Provide pxm_to_node and node_to_pxm translation functions

Required for NUMA node table setup on sn2.

Signed-off-by: Jes Sorensen <jes@sgi.com>
18 years ago[IA64] Change virtual address of XEN UC indentity area.
Alex Williamson [Tue, 12 Jun 2007 21:20:06 +0000 (15:20 -0600)]
[IA64] Change virtual address of XEN UC indentity area.

This slightly simplifies the code and makes flexible map possible.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Slightly modify xen memory map to make room for IO space.
Alex Williamson [Tue, 12 Jun 2007 21:17:16 +0000 (15:17 -0600)]
[IA64] Slightly modify xen memory map to make room for IO space.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Use virtualized psr in vcpu context instead of real psr.
Alex Williamson [Tue, 12 Jun 2007 21:16:03 +0000 (15:16 -0600)]
[IA64] Use virtualized psr in vcpu context instead of real psr.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Support for multiple I/O port spaces
Alex Williamson [Tue, 12 Jun 2007 21:07:05 +0000 (15:07 -0600)]
[IA64] Support for multiple I/O port spaces

Linux has support for multiple I/O port spaces on a system.  Depending
on the system configuration, this might allow for 64k of I/O port space
per PCI bus.  The "extended" I/O port ranges are described to the OS by
_CRS methods on PCI root devices in ACPI namespace.  For instance, on my
system, /proc/iomem shows several entries like these:

80000000000-80003ffffff : PCI Bus 0000:00 I/O Ports 01000000-0100ffff
80100000000-80103ffffff : PCI Bus 0000:01 I/O Ports 02000000-0200ffff
80200000000-80203ffffff : PCI Bus 0000:0a I/O Ports 03000000-0300ffff
80300000000-80303ffffff : PCI Bus 0000:0f I/O Ports 04000000-0400ffff
...

These describe MMIO ranges that provide I/O port ranges per PCI bus.
My /proc/ioports then shows entries like these:

01000000-0100ffff : PCI Bus 0000:00
  01001000-010010ff : 0000:00:04.0
02000000-0200ffff : PCI Bus 0000:01
  02001000-02001fff : PCI Bus #02
    02001000-0200107f : 0000:02:07.0
      02001000-0200107f : tulip
    02001080-020010ff : 0000:02:06.0
      02001080-020010ff : tulip
    02001100-0200117f : 0000:02:05.0
    02001180-020011ff : 0000:02:04.0
      02001180-020011ff : tulip
  02002000-0200203f : 0000:01:02.1
  02002040-0200207f : 0000:01:02.0
    02002040-0200207f : e1000
03000000-0300ffff : PCI Bus 0000:0a
...

Xen currently has no concept of these extended I/O port space ranges and
prevents dom0 from doing the MMIO transactions required to access these
ports.  This patch solves the problem.  First we need to make
ioports_permit/deny_access() aware that multiple I/O port ranges with
different MMIO base address exist.  Next we need to provide a mechanism
to register new I/O port spaces, for this I've created the dom0vp op
IA64_DOM0VP_add_io_space.  This allows dom0 to do the work of finding
the ranges in ACPI namespace since we don't want to add that kind of
bulk to Xen.  Finally, dom0 needs to make use of this interface when new
ranges are found.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64][BUILDER] Define native protocol ABI for IA64 guests.
Alex Williamson [Tue, 12 Jun 2007 20:58:39 +0000 (14:58 -0600)]
[IA64][BUILDER] Define native protocol ABI for IA64 guests.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
18 years ago[IA64] Implement gnttab replace_and_unmap
Alex Williamson [Tue, 12 Jun 2007 14:48:02 +0000 (08:48 -0600)]
[IA64] Implement gnttab replace_and_unmap

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years ago[IA64] Compilation fix of c/s 15180:1f7a6456c330 of xen-unstable.hg
Alex Williamson [Tue, 12 Jun 2007 14:39:24 +0000 (08:39 -0600)]
[IA64] Compilation fix of c/s 15180:1f7a6456c330 of xen-unstable.hg

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years agox86: Initialise %ds when booting APs, otherwise data accesses go to
kfraser@localhost.localdomain [Tue, 12 Jun 2007 13:06:37 +0000 (14:06 +0100)]
x86: Initialise %ds when booting APs, otherwise data accesses go to
the wrong place.
Signed-off-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Gather BIOS EDD info during boot.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 13:03:09 +0000 (14:03 +0100)]
x86: Gather BIOS EDD info during boot.
Still needs plumbing to dom0.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoFix fully-virtualized core dumps for 32-bit guests.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:39:51 +0000 (11:39 +0100)]
Fix fully-virtualized core dumps for 32-bit guests.
Signed-off-by: Chris Lalancette <clalance@redhat.com>
18 years agotools: warn_unused_result build fixes.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:38:31 +0000 (11:38 +0100)]
tools: warn_unused_result build fixes.

Note that some of the existing error checking for asprintf was wrong
for Linux.  asprintf in glibc returns -1 on error, but leaves the
pointer uninitialized.  Only the BSDs zero out the pointer on error.

Also, while fixing these warnings I saw several error paths that were
incorrect.  This patch minimally fixes the warn_unused_result; more
complete error path cleanup will be a later patch.

Signed-off-by: Charles Coffing <ccoffing@novell.com>
18 years ago[IA64] When a domU is destroyed, it will fall into NVRAM saving
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:32:03 +0000 (11:32 +0100)]
[IA64] When a domU is destroyed, it will fall into NVRAM saving
path. But if there is no NVRAM file for the domU, the save operation
would fail. Then domU blocked by Xend's exception. This patch fixs
that issue.

Signed-off-by: Zhang Xin <xing.z.zhang@intel.com>
18 years agoFix xm vcpu-pin command for keyword 'all'.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:29:27 +0000 (11:29 +0100)]
Fix xm vcpu-pin command for keyword 'all'.

The keyword 'all' was supported as the 2nd argument of xm vcpu-pin
command since changeset 11619.  However, we cannot use the keyword
'all' since changeset 14481.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agotools: Clean up usage of const type qualifier.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:26:51 +0000 (11:26 +0100)]
tools: Clean up usage of const type qualifier.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Fix VBE EDID interrogation. Add a new boot option 'edid={no,force}'.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 10:04:27 +0000 (11:04 +0100)]
x86: Fix VBE EDID interrogation. Add a new boot option 'edid={no,force}'.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Fix time_resume() to notify all domains of wallclock change.
kfraser@localhost.localdomain [Tue, 12 Jun 2007 08:22:06 +0000 (09:22 +0100)]
x86: Fix time_resume() to notify all domains of wallclock change.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] White space cleanup from nvram patch
Alex Williamson [Mon, 11 Jun 2007 20:19:05 +0000 (14:19 -0600)]
[IA64] White space cleanup from nvram patch

This file is using space indenting.  No functional changes.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64] create free_irq()
Alex Williamson [Mon, 11 Jun 2007 19:59:48 +0000 (13:59 -0600)]
[IA64] create free_irq()

This isn't well tested, since it's not likely to get called, but
it is required to build.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years agomerge with xen-unstable.hg (staging)
Alex Williamson [Mon, 11 Jun 2007 19:46:42 +0000 (13:46 -0600)]
merge with xen-unstable.hg (staging)

18 years ago[IA64] Fix bug in xen_alloc_coherent()
Alex Williamson [Mon, 11 Jun 2007 17:21:31 +0000 (11:21 -0600)]
[IA64] Fix bug in xen_alloc_coherent()

xen_create_contiguous_region() needs an order, not a mask.  Also change
s/hwdev/dev/ for consistency.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
18 years ago[IA64] Fix typo of warning in __dom0vp_add_physmap()
Alex Williamson [Mon, 11 Jun 2007 17:17:33 +0000 (11:17 -0600)]
[IA64] Fix typo of warning in __dom0vp_add_physmap()

Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>
18 years ago[IA64] Avoid use of kernel type in user code.
Alex Williamson [Mon, 11 Jun 2007 17:05:27 +0000 (11:05 -0600)]
[IA64] Avoid use of kernel type in user code.

Signed-off-by: Tristan Gingold <tgingold@free.fr>
18 years ago[IA64] Fix p2m exposure.
Alex Williamson [Mon, 11 Jun 2007 17:02:05 +0000 (11:02 -0600)]
[IA64] Fix p2m exposure.

p2m table page doesn't belong to any domain so that special handling required.
Instead, dom_p2m is introduced and make all p2m page belong to dom_p2m.

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
18 years agoInclude Solaris kernel headers
kfraser@localhost.localdomain [Mon, 11 Jun 2007 14:50:35 +0000 (15:50 +0100)]
Include Solaris kernel headers

Add the Solaris kernel headers used be the userspace tools to the
distribution. The same setup as used by Linux is taken. We modify the
Linux kernel headers to also install in /usr/include/xen/sys/ -
without a kernel-neutral path, it can make using the headers very
difficult, and encourages errors.

Signed-off-by: John Levon <john.levon@sun.com>
18 years agoAdd suspend/resume to devices owned by Xen.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 14:44:48 +0000 (15:44 +0100)]
Add suspend/resume to devices owned by Xen.

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 agox86: Clean up interface to platform timers to extract more common code.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 14:10:23 +0000 (15:10 +0100)]
x86: Clean up interface to platform timers to extract more common code.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[QEMU] Fix up ioemu timer save/restore after version changes.
Tim Deegan [Mon, 11 Jun 2007 13:56:50 +0000 (14:56 +0100)]
[QEMU] Fix up ioemu timer save/restore after version changes.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years ago[XEN] Clean up locking/init code around log-dirty interfaces
Tim Deegan [Mon, 11 Jun 2007 13:38:46 +0000 (14:38 +0100)]
[XEN] Clean up locking/init code around log-dirty interfaces
to avoid deadlocks and make sure locks/functions are in place for
PV domains to be put in log-dirty mode if they're not already shadowed.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years ago[XEN] Make common log-dirty paging code and add HAP log-dirty support.
Tim Deegan [Mon, 11 Jun 2007 13:35:52 +0000 (14:35 +0100)]
[XEN] Make common log-dirty paging code and add HAP log-dirty support.
Signed-off-by: Wei Huang <wei.huang2@amd.com>
18 years ago[TOOLS] sync qemu and python view of save filenames after 15241:b0109d3dc3dd
Tim Deegan [Mon, 11 Jun 2007 10:37:10 +0000 (11:37 +0100)]
[TOOLS] sync qemu and python view of save filenames after 15241:b0109d3dc3dd
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
18 years agoxend: Fix duplicate random MAC generation.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 09:22:32 +0000 (10:22 +0100)]
xend: Fix duplicate random MAC generation.

Ensure that HVMImageHandler and NetifController see the same random
MAC address.

Signed-off-by: David Edmondson <dme@sun.com>
18 years agoxend: Make device detach wait for detach to complete.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 09:21:11 +0000 (10:21 +0100)]
xend: Make device detach wait for detach to complete.

*-detach will wait till device is detached or time'ed out.
XendConfig will be updated, if detached.

Signed-off-by: Max Zhen <max.zhen@sun.com>
18 years agoxend: Do not give up auto-ballooning if ballooning is proceeding okay.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 09:16:54 +0000 (10:16 +0100)]
xend: Do not give up auto-ballooning if ballooning is proceeding okay.
Under a loaded system, don't give up ballooning dom0 unless it's not
responding at all.

Signed-off-by: Ryan Scott <ryan.scott@sun.com>
18 years agoxend: Fix xm block-detach regression introduced by c/s 15157.
kfraser@localhost.localdomain [Mon, 11 Jun 2007 09:14:54 +0000 (10:14 +0100)]
xend: Fix xm block-detach regression introduced by c/s 15157.

Prior to this c/s, user was able to specify device name or id, e.g.
 xm block-detach dom xvdb
 xm block-detach dom 15728

Now, 'xm block-detach dom xvdb' silently fails.  xend.log does contain
the following

[2007-06-07 11:39:18 xend.XendDomainInfo 3775] DEBUG
(XendDomainInfo:519) devName=vbd/51712
[2007-06-07 11:39:18 xend.XendDomainInfo 3775] DEBUG
(XendDomainInfo:527) Could not find the device xvdb

This patch restores the previous behavior but retains the bugfix
introduced by c/s 15571.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
18 years agotools: Fix xenstored build error by checking vasprintf() return code.
Keir Fraser [Sun, 10 Jun 2007 18:58:22 +0000 (19:58 +0100)]
tools: Fix xenstored build error by checking vasprintf() return code.
Signed-off-by: Charles Coffing <ccoffing@novell.com>
18 years agohvm: e820 public header cleanup.
Keir Fraser [Sun, 10 Jun 2007 18:49:44 +0000 (19:49 +0100)]
hvm: e820 public header cleanup.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agodocs: Update for new 'vga=' boot option syntax.
Keir Fraser [Sun, 10 Jun 2007 18:29:32 +0000 (19:29 +0100)]
docs: Update for new 'vga=' boot option syntax.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Remove broken and unnecessary numa code from smpboot.c.
Keir Fraser [Sun, 10 Jun 2007 17:07:58 +0000 (18:07 +0100)]
x86: Remove broken and unnecessary numa code from smpboot.c.
This was trampling on random memory when node==NUMA_NO_NODE.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Fix save_all regs macro for new CONFIG_FRAME_POINTER define.
Keir Fraser [Sun, 10 Jun 2007 16:00:47 +0000 (17:00 +0100)]
x86: Fix save_all regs macro for new CONFIG_FRAME_POINTER define.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Fix log-level in printing e820 memory map.
Keir Fraser [Sun, 10 Jun 2007 15:12:51 +0000 (16:12 +0100)]
x86: Fix log-level in printing e820 memory map.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Allow selection of graphical video modes during boot.
Keir Fraser [Sun, 10 Jun 2007 14:58:08 +0000 (15:58 +0100)]
x86: Allow selection of graphical video modes during boot.

The 'vga=' boot option is extended as follows:

----------
 'vga=<mode-specifier>[,keep]' where <mode-specifier> is one of:

   'vga=ask':
      display a vga menu of available modes

   'vga=text-80x<rows>':
      text mode, where <rows> is one of {25,28,30,34,43,50,60}

   'vga=gfx-<width>x<height>x<depth>':
      graphics mode, e.g., vga=gfx-1024x768x16

   'vga=mode-<mode>:
      specifies a mode as specified in 'vga=ask' menu
      (NB. menu modes are displayed in hex, so mode numbers here must
           be prefixed with '0x' (e.g., 'vga=mode-0x0318'))

 The option 'keep' causes Xen to continue to print to the VGA console even
 after domain 0 starts to boot. The default behaviour is to relinquish
 control of the console to domain 0.
----------

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoioemu: save file name is snprintf()ed into a plenty big enough buffer.
kfraser@localhost.localdomain [Fri, 8 Jun 2007 18:22:08 +0000 (19:22 +0100)]
ioemu: save file name is snprintf()ed into a plenty big enough buffer.
From: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAvoid using /tmp for qemu state files since an unprivileged user
Steven Hand [Fri, 8 Jun 2007 16:37:58 +0000 (17:37 +0100)]
Avoid using /tmp for qemu state files since an unprivileged user
in dom0 could potentially exploit the fact.

Thanks to Dan Berrange <berrange@redhat.com>

Signed-off-by: Steven Hand
18 years agotools: Fix xm mem-max command.
kfraser@localhost.localdomain [Fri, 8 Jun 2007 10:21:21 +0000 (11:21 +0100)]
tools: Fix xm mem-max command.

When I tested xm mem-max and xm list, I saw the following error
message by xm list.  Because xm mem-max always changes
"memory_static_max", I think that the cause of the error message
is it.  This patch fixes it.

# xm list --long PVdomain.1 | grep mem
    (maxmem 1024)
    (memory 1024)
    (shadow_memory 0)
# xm mem-max PVdomain.1 512
Error: (22, 'Invalid argument')
Usage: xm mem-max <Domain> <Mem>

Set the maximum amount reservation for a domain.
# xm list --long PVdomain.1 | grep mem
Error: memory_dynamic_max must be less than or equal to
memory_static_max

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
18 years agohvm: Respect irqbase set by protected mode in mode switching with VMXAssist.
kfraser@localhost.localdomain [Fri, 8 Jun 2007 10:19:55 +0000 (11:19 +0100)]
hvm: Respect irqbase set by protected mode in mode switching with VMXAssist.

RHEL4U4 PAE SMP guest currently crashes, and we found changeset 15214
introduced it. This patch fixes it.

Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
18 years agox86: Print source of e820 memory map during boot. Fix Xen-e801 memmap
kfraser@localhost.localdomain [Thu, 7 Jun 2007 19:02:27 +0000 (20:02 +0100)]
x86: Print source of e820 memory map during boot. Fix Xen-e801 memmap
parsing. Get rid of unneeded e820_raw variable -- map straight onto
boot-trampoline e820 array.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agox86: Return to real-mode when booting primary CPU, and gather
kfraser@localhost.localdomain [Thu, 7 Jun 2007 18:41:59 +0000 (19:41 +0100)]
x86: Return to real-mode when booting primary CPU, and gather
memory-map information (from int15{e820,e801,88} and int12 bios
calls).

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoFix the README to match Xen 3.1
kfraser@localhost.localdomain [Thu, 7 Jun 2007 15:44:04 +0000 (16:44 +0100)]
Fix the README to match Xen 3.1
From: Ryan Riley <rileyrd@gmail.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoAdd backtrace support to xenoprof.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 13:30:21 +0000 (14:30 +0100)]
Add backtrace support to xenoprof.

Signed-off-by: Amitabha Roy <amitabha.roy@gmail.com>
Reviewed-by: Jose Renato G Santos <joserenato.santos@hp.com>
18 years agoioemu: Tiny indentation fix.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 10:15:24 +0000 (11:15 +0100)]
ioemu: Tiny indentation fix.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoxentop: Display blktap information.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 10:14:18 +0000 (11:14 +0100)]
xentop: Display blktap information.
Signed-off-by: Satoshi UCHIDA <s-uchida@ap.jp.nec.com>
18 years agoxenstat: Adding to scan blktap information.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 10:13:41 +0000 (11:13 +0100)]
xenstat: Adding to scan blktap information.
Signed-off-by: Satoshi UCHIDA <s-uchida@ap.jp.nec.com>
18 years agoFix the vnet module for Xen 3.1.x.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 10:06:00 +0000 (11:06 +0100)]
Fix the vnet module for Xen 3.1.x.
Signed-off-by: Robert Valentan <R.Valentan@solid-soft.at>
18 years agotools: Fix some type issues GCC 4.1.0 warnings.
kfraser@localhost.localdomain [Thu, 7 Jun 2007 10:02:23 +0000 (11:02 +0100)]
tools: Fix some type issues GCC 4.1.0 warnings.

FC5's gcc 4.1.0 can't make some files in tools/ due to its stronger
type checking.

From: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoqemu: Fix vnc port offset in xenstore
kfraser@localhost.localdomain [Wed, 6 Jun 2007 19:25:58 +0000 (20:25 +0100)]
qemu: Fix vnc port offset in xenstore

The new version of qemu uses 5900 instead of 0 as the port.
There is no need to offset it by 5900 when writing out the
console/vnc-port entry.

Signed-off-by: Ben Guthro <bguthro@virtualiron.com>
18 years ago[HVM] Prevent usb driver crashes in Windows
kfraser@localhost.localdomain [Wed, 6 Jun 2007 19:24:06 +0000 (20:24 +0100)]
[HVM] Prevent usb driver crashes in Windows

Use atomic updates to read/write usb controller data.
This can be done because:
    a) word copies on x86 are atomic
    b) The USB spec requires word alignment

This will need to be enhanced once USB 1.2 is supported.

Signed-off-by: Steve Ofsthun <sofsthun@virtualiron.com>
Update to copy 'longword'-sized atoms.

Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agopv-on-hvm: Allow Linux and Xen paths to be overridden by the environment.
kfraser@localhost.localdomain [Wed, 6 Jun 2007 16:49:39 +0000 (17:49 +0100)]
pv-on-hvm: Allow Linux and Xen paths to be overridden by the environment.
Signed-off-by: Ben Guthro <bguthro@virtualiron.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years agoFix pv-on-hvm mkbuildtree script for new repo layout.
kfraser@localhost.localdomain [Wed, 6 Jun 2007 16:32:58 +0000 (17:32 +0100)]
Fix pv-on-hvm mkbuildtree script for new repo layout.
Signed-off-by: Keir Fraser <keir@xensource.com>
18 years ago[IA64] Set NODES_SHIFT to something a bit more realistic
Alex Williamson [Wed, 6 Jun 2007 15:30:01 +0000 (09:30 -0600)]
[IA64] Set NODES_SHIFT to something a bit more realistic

Signed-off-by: Jes Sorensen <jes@sgi.com>