xen.git
19 years agomerge with xen-unstable.hg
awilliam@xenbuild.aw [Wed, 23 Aug 2006 17:11:27 +0000 (11:11 -0600)]
merge with xen-unstable.hg

19 years ago[XEN] Fix pirq_guest_unbind(). Remove unnecessary code
kfraser@localhost.localdomain [Wed, 23 Aug 2006 10:27:06 +0000 (11:27 +0100)]
[XEN] Fix pirq_guest_unbind(). Remove unnecessary code
(in fact flush_all_pending_eoi() was broken!).
Based on a patch from Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[IA64] Remove PIN processing for Domain-0 vcpu0
awilliam@xenbuild.aw [Tue, 22 Aug 2006 20:45:49 +0000 (14:45 -0600)]
[IA64] Remove PIN processing for Domain-0 vcpu0

This patch removes PIN processing for Domain-0 vcpu0 executed
at the time of Xen initialization.

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
19 years ago[IA64] Add stubs to Linux/IPF for the new hvm_op hypercall.
awilliam@xenbuild.aw [Tue, 22 Aug 2006 20:44:13 +0000 (14:44 -0600)]
[IA64] Add stubs to Linux/IPF for the new hvm_op hypercall.

Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
19 years ago[IA64] Add support for sending event channel interrupts to VT-i guest
awilliam@xenbuild.aw [Tue, 22 Aug 2006 20:42:05 +0000 (14:42 -0600)]
[IA64] Add support for sending event channel interrupts to VT-i guest

Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
19 years ago[IA64] HMVOP_get_param return parameter value
awilliam@xenbuild.aw [Tue, 22 Aug 2006 20:40:48 +0000 (14:40 -0600)]
[IA64] HMVOP_get_param return parameter value

HMVOP_get_param return parameter value within the provided
parameter struct for IPF

Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
19 years agomerge
Ian Campbell [Tue, 22 Aug 2006 17:51:02 +0000 (18:51 +0100)]
merge

19 years ago[XEN] Remove some accidentally commited debug code
Ian Campbell [Tue, 22 Aug 2006 17:50:22 +0000 (18:50 +0100)]
[XEN] Remove some accidentally commited debug code
from dom0 domain builder.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[SHADOW] Clean up the DOM0_SHADOW_OP interface.
kfraser@localhost.localdomain [Tue, 22 Aug 2006 17:50:01 +0000 (18:50 +0100)]
[SHADOW] Clean up the DOM0_SHADOW_OP interface.
Also bump the dom0 interface version for tools (but
not for dom0 kernel!).
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agomerge
Ian Campbell [Tue, 22 Aug 2006 17:48:00 +0000 (18:48 +0100)]
merge

19 years ago[XEN] Revert changes to dom0 domain builder made by cset 11226.
Ian Campbell [Tue, 22 Aug 2006 17:47:29 +0000 (18:47 +0100)]
[XEN] Revert changes to dom0 domain builder made by cset 11226.

We require that the dom0 domain builder still sets _PAGE_USER.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[HVM] Clean ups for PV-on-HVM drivers. In particular, platform-pci
kfraser@localhost.localdomain [Tue, 22 Aug 2006 16:16:58 +0000 (17:16 +0100)]
[HVM] Clean ups for PV-on-HVM drivers. In particular, platform-pci
driver can now handle multi-page hypercall stub areas.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LINUX] Update x86/64 ELF note patch to version accepted upstream.
Ian Campbell [Tue, 22 Aug 2006 14:55:43 +0000 (15:55 +0100)]
[LINUX] Update x86/64 ELF note patch to version accepted upstream.

Signed-off-by: Ian Campbell <ian.campbell@xesource.com>
19 years ago[x86_64] Guests no longer set _PAGE_USER on kernel mappings.
kfraser@localhost.localdomain [Tue, 22 Aug 2006 14:26:40 +0000 (15:26 +0100)]
[x86_64] Guests no longer set _PAGE_USER on kernel mappings.
This may allow guest kernels to be run outside ring 3 in future, and
also provides scope for optimisations today (e.g., using global bit on
user mappings).

Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[hvm] Export gpfn of store page to guest - not mfn.
Christian Limpach [Tue, 22 Aug 2006 14:13:07 +0000 (15:13 +0100)]
[hvm] Export gpfn of store page to guest - not mfn.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years agoMake add_to_bridge() always leave the device up even when it bails out due to the...
Anil Madhavapeddy [Tue, 22 Aug 2006 13:59:16 +0000 (14:59 +0100)]
Make add_to_bridge() always leave the device up even when it bails out due to the device already being present on the bridge.
This makes the script more idempotent if run twice (e.g. from hotplug and udev), since the second time the device will be taken down and not brought up again, leaving the host without networking.

Signed-off-by: Anil Madhavapeddy <anil@xensource.com>
19 years ago[LINUX] Support creating ELF note segments in the kernel ELF image.
Ian Campbell [Tue, 22 Aug 2006 11:20:43 +0000 (12:20 +0100)]
[LINUX] Support creating ELF note segments in the kernel ELF image.

i386 version is taken from 2.6.18-rc4-mm1 and x86_64 version has been
submitted for inclusion.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years agomerge.
Ian Campbell [Tue, 22 Aug 2006 10:34:46 +0000 (11:34 +0100)]
merge.

19 years ago[XEN] Debug builds check for correct usage of percpu areas.
kfraser@localhost.localdomain [Tue, 22 Aug 2006 10:33:59 +0000 (11:33 +0100)]
[XEN] Debug builds check for correct usage of percpu areas.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LINUX] Revert bits of cset 11202 since they break the -xenU build for x86/64.
Ian Campbell [Tue, 22 Aug 2006 10:30:13 +0000 (11:30 +0100)]
[LINUX] Revert bits of cset 11202 since they break the -xenU build for x86/64.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[XEN] Clean up some x86 bootstrap code. Replace some CPU iterators
kfraser@localhost.localdomain [Tue, 22 Aug 2006 10:19:48 +0000 (11:19 +0100)]
[XEN] Clean up some x86 bootstrap code. Replace some CPU iterators
with for_each_cpu() -- we want to ensure that per_cpu areas are
accessed only for cpus in cpu_possible_map.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoAdded option to xentrace to set the trace buffer size.
George Dunlap [Mon, 21 Aug 2006 16:05:11 +0000 (12:05 -0400)]
Added option to xentrace to set the trace buffer size.

Signed-off-by: George Dunlap <gdunlap@xensource.com>
19 years ago[XEN] Remove bogus assertion.
kfraser@localhost.localdomain [Mon, 21 Aug 2006 12:36:05 +0000 (13:36 +0100)]
[XEN] Remove bogus assertion.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoFix assertion.
Steven Hand [Mon, 21 Aug 2006 12:21:51 +0000 (13:21 +0100)]
Fix assertion.

Signed-off-by: Steven Hand <steven@xensource.com>
19 years ago[XEN] Revert patch 11213.
kfraser@localhost.localdomain [Mon, 21 Aug 2006 10:51:54 +0000 (11:51 +0100)]
[XEN] Revert patch 11213.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Fix maddr_from_mapped_domain_page().
kfraser@localhost.localdomain [Mon, 21 Aug 2006 10:39:27 +0000 (11:39 +0100)]
[XEN] Fix maddr_from_mapped_domain_page().
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Fix shadow2 integration with writable-pagetable logic.
kfraser@localhost.localdomain [Mon, 21 Aug 2006 09:28:02 +0000 (10:28 +0100)]
[XEN] Fix shadow2 integration with writable-pagetable logic.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Remove unnecessary printk from invalid opcode debug output.
kfraser@localhost.localdomain [Mon, 21 Aug 2006 08:01:53 +0000 (09:01 +0100)]
[XEN] Remove unnecessary printk from invalid opcode debug output.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[xend] Add 'xend_config_format' field to the Xend information block.
Christian Limpach [Sun, 20 Aug 2006 23:11:17 +0000 (00:11 +0100)]
[xend] Add 'xend_config_format' field to the Xend information block.
When running 'xm info' it appears thus:

 # xm info
 host                   : 10.13.4.252
 release                : 2.6.17-1.2157_FC5xen0
 version                : #1 SMP Tue Jul 11 23:30:05 EDT 2006
 machine                : x86_64
 ...snipped rest of data...
 cc_compile_domain      : build.redhat.com
 cc_compile_date        : Tue Jul 11 22:50:42 EDT 2006
 xend_config_format     : 2

It will also appear in the SEXPR returned when doing a GET on the
/xend/node path in XenD's HTTP service:

 (node
   (system Linux)
   (host 10.13.4.252)
   (release 2.6.17-1.2157_FC5xen0)
   (version '#1 SMP Tue Jul 11 23:30:05 EDT 2006')
   (machine x86_64)
   ...snipped rest of data...
   (cc_compile_domain build.redhat.com)
   (cc_compile_date 'Tue Jul 11 22:50:42 EDT 2006')
   (xend_config_format 2)
 )

The patch sets the format to value '2' to indicate that we're on the version
with the new style HVM configuration for CDROM devices. The old style having
implicitly been version '1'.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
19 years ago[qemu patches] Update patches for changeset 11209:9bb6c1c1890a.
Christian Limpach [Sun, 20 Aug 2006 23:01:11 +0000 (00:01 +0100)]
[qemu patches] Update patches for changeset 11209:9bb6c1c1890a.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu] hdparm tunable IDE write cache for HVM
Christian Limpach [Sun, 20 Aug 2006 22:59:34 +0000 (23:59 +0100)]
[qemu] hdparm tunable IDE write cache for HVM
qemu 0.8.2 has a flush callback to the storage backends, so now it is
possible to implement hdparm tunable IDE write cache enable/disable for
guest domains, allowing people to pick speed or data consistency on a
case by case basis.

As an added benefit, really large LBA48 IOs will now no longer be broken
up into smaller IOs on the host side.

From: Rik van Riel <riel@redhat.com>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu patches] Update patches for changeset 11206:fc3e7e65b953.
Christian Limpach [Sun, 20 Aug 2006 22:52:39 +0000 (23:52 +0100)]
[qemu patches] Update patches for changeset 11206:fc3e7e65b953.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu patches] Update patches for changeset 11206:fc3e7e65b953.
Christian Limpach [Sun, 20 Aug 2006 22:34:17 +0000 (23:34 +0100)]
[qemu patches] Update patches for changeset 11206:fc3e7e65b953.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu] Allow booting from more than one device.
Christian Limpach [Sun, 20 Aug 2006 22:33:28 +0000 (23:33 +0100)]
[qemu] Allow booting from more than one device.
The rombios supports trying to boot from more than one device and then
falling back.  Set 'boot=dc' in your config file to try booting first
from the CD and then the hard drive.

Based on a patch from: Jeremy Katz <katzj@redhat.com>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[XEN] More shadow2 cleanups -- primarily moving arch vcpu/domain
kaf24@localhost.localdomain [Sun, 20 Aug 2006 16:55:33 +0000 (17:55 +0100)]
[XEN] More shadow2 cleanups -- primarily moving arch vcpu/domain
fields into separate shadow2-specific structures.
Also rename shadow2_entry_points to shadow2_paging_mode.
Remove VCPUF_shadow2_translate_mode and replace with a better-named
field in vcpu.arch.shadow2
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Clarify x86 hypercall ABI in public header comments.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 16:07:54 +0000 (17:07 +0100)]
[XEN] Clarify x86 hypercall ABI in public header comments.
In particular: x86/64 *does* clobber its ragument registers
(just like i386). This happens in debug builds, and if a
continuation is created.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XENBUS] Allow user of xenbus_read_driver_state() to distinguish
kaf24@localhost.localdomain [Sat, 19 Aug 2006 13:28:10 +0000 (14:28 +0100)]
[XENBUS] Allow user of xenbus_read_driver_state() to distinguish
absence of state field in xenstore by returning XenbusStateUnknown
rather than XenbusStateClosed.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LINUX] Replace uses of CONFIG_XEN_PRIVILEGED_GUEST with is_initial_xendomain().
kaf24@localhost.localdomain [Sat, 19 Aug 2006 11:08:12 +0000 (12:08 +0100)]
[LINUX] Replace uses of CONFIG_XEN_PRIVILEGED_GUEST with is_initial_xendomain().
Signed-off-by: Jan Beulich <jbeulich@novell.com>
19 years ago[HVM][VMX] Don't update the EIP on debug register accesses
kaf24@localhost.localdomain [Sat, 19 Aug 2006 11:07:07 +0000 (12:07 +0100)]
[HVM][VMX] Don't update the EIP on debug register accesses

On debug register accesses, the EIP should not be updated.
Because of the way that lazy save/restore of the debug
registers is implemented, this initial debug register access
is lost.

Signed-off-by: Travis Betak <travis.betak@amd.com>
19 years ago[HVM] Fix SMBIOS entry point copy destination.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 11:06:36 +0000 (12:06 +0100)]
[HVM] Fix SMBIOS entry point copy destination.
Spotted by Xiaowei Yang <xiaowei.yang@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Fix receive path for auto-translated guests, so that
kaf24@localhost.localdomain [Sat, 19 Aug 2006 10:39:04 +0000 (11:39 +0100)]
[NET] front: Fix receive path for auto-translated guests, so that
driver balloon allowance is properly accounted.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Make some implicit header includes explicit.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 10:13:17 +0000 (11:13 +0100)]
[XEN] Make some implicit header includes explicit.
Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
19 years ago[NET] front: Fix tx buffer leak in teardown path. Fix rx-buffer
kaf24@localhost.localdomain [Sat, 19 Aug 2006 10:08:40 +0000 (11:08 +0100)]
[NET] front: Fix tx buffer leak in teardown path. Fix rx-buffer
cleanup: cannot free skbuffs until their memory is remapped by
multicall.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Fix rx buffer leak when tearing down an interface.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 09:58:07 +0000 (10:58 +0100)]
[NET] front: Fix rx buffer leak when tearing down an interface.
Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years ago[XENBUS] Cleanup correctly on suspend, unregister watches.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 09:21:02 +0000 (10:21 +0100)]
[XENBUS] Cleanup correctly on suspend, unregister watches.
Again needed for domU kexec, so the next kernel instance can register
watches without running into -EBUSY.

Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years ago[NET/BLK] back: Change the netback and blkback device drivers to allow
kaf24@localhost.localdomain [Sat, 19 Aug 2006 09:20:08 +0000 (10:20 +0100)]
[NET/BLK] back: Change the netback and blkback device drivers to allow
and handle frontend state transition from "Closing" to "Initialising".
Instead of destroying the backend device then they wait for a
reconnect from the frontend.  Needed for domU kexec support, to hand over
devices from one kernel instance to the next.

Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years ago[BALLOON] Make /proc/xen/ballon contain driver_pages (aka "driver headroom") uncondit...
kaf24@localhost.localdomain [Sat, 19 Aug 2006 09:17:48 +0000 (10:17 +0100)]
[BALLOON] Make /proc/xen/ballon contain driver_pages (aka "driver headroom") unconditionally.

Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years ago[XENBUS] A few cleanups.
kaf24@localhost.localdomain [Sat, 19 Aug 2006 09:15:01 +0000 (10:15 +0100)]
[XENBUS] A few cleanups.
  * move function prototypes to the top of the file.
  * uses the "struct bus_type" function pointers instead
    of overwriting the "struct device_driver" ones for
    xenbus_dev_probe() and xenbus_dev_remove().

Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years ago[IA64] makes some includes explicit
awilliam@xenbuild.aw [Fri, 18 Aug 2006 21:01:04 +0000 (15:01 -0600)]
[IA64] makes some includes explicit

Make some includes explicit.
Do not compile anymore vmx/mm.c: not used.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
19 years ago[IA64] perfc compile error patch
awilliam@xenbuild.aw [Fri, 18 Aug 2006 19:34:25 +0000 (13:34 -0600)]
[IA64] perfc compile error patch

Signed-off-by: Ken Hironaka <hironaka.ken@soft.fujitsu.com>
19 years ago[IA64] reformatting hyperprivop.S for 80-column readability
awilliam@xenbuild.aw [Fri, 18 Aug 2006 19:31:48 +0000 (13:31 -0600)]
[IA64] reformatting hyperprivop.S for 80-column readability

Minor reformatting to hyperprivop.S to make it readable
on 80-column displays, as requested by Linux coding style.
Minor formatting changes and typo corrections to get the
style more in line with the other .S files.

Signed-off-by: Al Stone <ahs3@fc.hp.com>
19 years ago[XEN] Remove shadow2 dependencies from common code.
kfraser@localhost.localdomain [Fri, 18 Aug 2006 16:59:26 +0000 (17:59 +0100)]
[XEN] Remove shadow2 dependencies from common code.
Disable debugtrace functionality even in debug builds:
it's currently only used by shadow2 maintainers.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Fix the APM BIOS to ensure that interrupts are enabled when it
kfraser@localhost.localdomain [Fri, 18 Aug 2006 15:20:45 +0000 (16:20 +0100)]
[HVM] Fix the APM BIOS to ensure that interrupts are enabled when it
executes an idling HLT. This allow us to re-enable the 'shutdown on
HLT-with-interrupts-disabled' heuristic.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Work around a Linux PAE issue where it sometimes sets
kfraser@localhost.localdomain [Fri, 18 Aug 2006 14:24:10 +0000 (15:24 +0100)]
[XEN] Work around a Linux PAE issue where it sometimes sets
the top half of a PTE without first clearing the bottom half
(to make the PTE not-present) hence the PTE fails validation
checks.

Also remove dead revalidation function after shadow2 checkin.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoFix Xen Rules.mk eval order so that debug=y always sets verbose=y.
kfraser@localhost.localdomain [Fri, 18 Aug 2006 14:22:54 +0000 (15:22 +0100)]
Fix Xen Rules.mk eval order so that debug=y always sets verbose=y.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[qemu patches] Update patches for changeset 11186:078bfd250677.
Christian Limpach [Thu, 17 Aug 2006 19:31:47 +0000 (20:31 +0100)]
[qemu patches] Update patches for changeset 11186:078bfd250677.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu] Support HVM guests with more than 3.75G memory.
Christian Limpach [Thu, 17 Aug 2006 19:31:13 +0000 (20:31 +0100)]
[qemu] Support HVM guests with more than 3.75G memory.
Changes are:
1) M2P table and e820 table are changed to skip address space from
HVM_RAM_LIMIT_BELOW_4G to 4G.
2) shared io page location, when less than HVM_RAM_LIMIT_BELOW_4G
memory, it's the last page of RAM as today, or it's the last page of
HVM_RAM_LIMIT_BELOW_4G RAM.
3) in qemu-dm address space from HVM_RAM_LIMIT_BELOW_4G to 4G are
stuffed with mfns starting from 4G, so the 1:1 mapping can still works.
This is ugly, but another limit check patch as changeset 10757 will
prevent qemu-dm to access this range.  This ugly stuffing will be
removed when the patch to remove 1:1 mapping from qemu-dm gets accepted
in the future.

Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years agoFix live migration w/ shadow2.
Steven Hand [Thu, 17 Aug 2006 19:30:05 +0000 (20:30 +0100)]
Fix live migration w/ shadow2.

Signed-off-by: Steven Hand <steven@xensource.com>
19 years ago[SVM] Fix 64-bit VMRUN.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 18:24:20 +0000 (19:24 +0100)]
[SVM] Fix 64-bit VMRUN.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Indentation cleanups (remove hard tabs).
kfraser@localhost.localdomain [Thu, 17 Aug 2006 18:17:50 +0000 (19:17 +0100)]
[XEN] Indentation cleanups (remove hard tabs).
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoMissing file from unmodified_drivers directory.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 18:17:30 +0000 (19:17 +0100)]
Missing file from unmodified_drivers directory.

19 years ago[XEN] Fix page-fault handler to fall back to shadow-mode fault path if
kfraser@localhost.localdomain [Thu, 17 Aug 2006 18:16:01 +0000 (19:16 +0100)]
[XEN] Fix page-fault handler to fall back to shadow-mode fault path if
writable pagetable logic fails to fix up a fault.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Improve correctness when there is no connection to backend driver.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 18:11:14 +0000 (19:11 +0100)]
[NET] front: Improve correctness when there is no connection to backend driver.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Allow driver to be built for an HVM guest.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 17:45:59 +0000 (18:45 +0100)]
[NET] front: Allow driver to be built for an HVM guest.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoFix 64bit shadow tracking
tdeegan@york.uk.xensource.com [Thu, 17 Aug 2006 17:39:10 +0000 (18:39 +0100)]
Fix 64bit shadow tracking
(The mask used to check for available shadows was off by 1 bit)

Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years agoAdd a small readme file explaining how to build PV-on-HVM drivers.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 16:10:13 +0000 (17:10 +0100)]
Add a small readme file explaining how to build PV-on-HVM drivers.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years agoUpdate hgignore list for PV-on-HVM drivers.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 16:09:21 +0000 (17:09 +0100)]
Update hgignore list for PV-on-HVM drivers.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[NET] front: Allow driver to be built against kernels which do not
kfraser@localhost.localdomain [Thu, 17 Aug 2006 16:08:19 +0000 (17:08 +0100)]
[NET] front: Allow driver to be built against kernels which do not
have extended checksum-offload info fields in teh skbuff structure.
Based on a patch from Steven Smith <ssmith@xensource.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[NET] front: Make netfront compile against kernels without the GSO patch.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:57:43 +0000 (16:57 +0100)]
[NET] front: Make netfront compile against kernels without the GSO patch.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[LINUX] Fix return value of gnttab_resume().
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:57:16 +0000 (16:57 +0100)]
[LINUX] Fix return value of gnttab_resume().
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Allow blkfront driver to be built for an HVM guest.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:33:42 +0000 (16:33 +0100)]
[HVM] Allow blkfront driver to be built for an HVM guest.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[HVM] Allow xenbus to run in an HVM guest.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:32:30 +0000 (16:32 +0100)]
[HVM] Allow xenbus to run in an HVM guest.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years agomerge
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:29:21 +0000 (16:29 +0100)]
merge

19 years ago[LINUX] Allow xenbus to be built as a module.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:28:47 +0000 (16:28 +0100)]
[LINUX] Allow xenbus to be built as a module.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years agoMerge
tdeegan@york.uk.xensource.com [Thu, 17 Aug 2006 15:26:34 +0000 (16:26 +0100)]
Merge

19 years ago[HVM] Avoid accessing start_info from xenbus as far as possible, since
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:22:08 +0000 (16:22 +0100)]
[HVM] Avoid accessing start_info from xenbus as far as possible, since
HVM domains do things slightly differently.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[HVM] Make grant tables functions work in PV-on-HVM drivers.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:21:08 +0000 (16:21 +0100)]
[HVM] Make grant tables functions work in PV-on-HVM drivers.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[HVM] Linux driver for the xen platform pseudo-PCI device.
kfraser@localhost.localdomain [Thu, 17 Aug 2006 15:08:01 +0000 (16:08 +0100)]
[HVM] Linux driver for the xen platform pseudo-PCI device.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years agoFix handling of PS bit in 64-bit pagetables
tdeegan@york.uk.xensource.com [Thu, 17 Aug 2006 15:05:51 +0000 (16:05 +0100)]
Fix handling of PS bit in 64-bit pagetables

Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years ago[XEN] Fix x86/64 bug where a guest application can crash the
kfraser@localhost.localdomain [Thu, 17 Aug 2006 11:08:26 +0000 (12:08 +0100)]
[XEN] Fix x86/64 bug where a guest application can crash the
guest OS by setting AC flag in RFLAGS. This wasn't being
cleared on entry to the guest kernel, causing unwanted faults
because the kernel runs in ring 3 on Xen.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[qemu patches] Update patches for changeset 11161:5c1021595e3c.
chris@kneesaa.uk.xensource.com [Thu, 17 Aug 2006 11:01:44 +0000 (12:01 +0100)]
[qemu patches] Update patches for changeset 11161:5c1021595e3c.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu] Add the xen platform device to the qemu PCI bus.
chris@kneesaa.uk.xensource.com [Thu, 17 Aug 2006 11:01:23 +0000 (12:01 +0100)]
[qemu] Add the xen platform device to the qemu PCI bus.
Useful functionality will come later.

Signed-off-by: Steven Smith <ssmith@xensource.com>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[NET] Fix ``rx->offset: 10, size: 4294967295'' bug.
ssmith@weybridge.uk.xensource.com [Thu, 17 Aug 2006 11:46:45 +0000 (12:46 +0100)]
[NET] Fix ``rx->offset: 10, size: 4294967295'' bug.

Signed-off-by: Steven Smith <sos22@cam.ac.uk>
19 years ago[qemu patches] Update patches for changeset 11158:a774cbd38187.
chris@kneesaa.uk.xensource.com [Thu, 17 Aug 2006 10:34:39 +0000 (11:34 +0100)]
[qemu patches] Update patches for changeset 11158:a774cbd38187.

Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[qemu] Default monitor/serial ports to /dev/null if -nographic.
chris@kneesaa.uk.xensource.com [Thu, 17 Aug 2006 10:34:13 +0000 (11:34 +0100)]
[qemu] Default monitor/serial ports to /dev/null if -nographic.
qemu-dm runs daemonized and the default for -nographic is to redirect
the monitor and serial ports to stdio which causes qemu-dm to busy
read from stdin.  Change the default for -nographic to redirect the
monitor and serial ports to /dev/null.

Based on a patch from: Steven Smith <sos22-xen@srcf.ucam.org>
Signed-off-by: Christian Limpach <Christian.Limpach@xensource.com>
19 years ago[NET] back: Initialise first fragment properly
kfraser@localhost.localdomain [Thu, 17 Aug 2006 10:32:39 +0000 (11:32 +0100)]
[NET] back: Initialise first fragment properly

The first fragment is used to store the pending_idx of the leading
txreq if it doesn't fit in the head area.  When it does fit into
the head we need to ensure that the first fragment contains a value
that is not equal to pending_idx as that's what we use to distinguish
between the two cases in a a number of places.

This patch sets the first fragment to ~0 which is not equal to any
valid pending_idx.  Without this initialisation, we may double-free
a pending_idx if the first fragment happened to contain a value
equal to it (this usually happened with pending_idx 0).

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
19 years ago[IA64] handle ld.s on guest tr mapped page (VTI)
awilliam@xenbuild.aw [Wed, 16 Aug 2006 20:28:57 +0000 (14:28 -0600)]
[IA64] handle ld.s on guest tr mapped page (VTI)

Windows does an ld.s on a tr mapped page.
Currently xen/ipf uses tc/vtlb to emulate guest TR,
that may cause guest ld.s on tr page to be deferred, it is not correct.
For trapping this ld.s intruction, xen/ipf always set machine dcr.dm=0.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
[Moved cr.dcr restore to only impact vti -> non-vti switch]
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
19 years ago[XEN] vga code cleanups and additions for other architectures.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 17:20:03 +0000 (18:20 +0100)]
[XEN] vga code cleanups and additions for other architectures.
Based on patches from Hollis Blanchard and Alex Williamson.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoFix broken .hgtags
tdeegan@york.uk.xensource.com [Wed, 16 Aug 2006 17:11:33 +0000 (18:11 +0100)]
Fix broken .hgtags
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years ago[IA64] efibootmgr support
awilliam@xenbuild.aw [Wed, 16 Aug 2006 16:40:17 +0000 (10:40 -0600)]
[IA64] efibootmgr support

This patch supports the efibootmgr on Domain-0.
This patch adds the following EFI runtime service emulations
for the efibootmgr support.
 - GetVariable
 - GetNextVariableName
 - SetVariable

Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
19 years ago[IA64] resolve VTI address wrap issue
awilliam@xenbuild.aw [Wed, 16 Aug 2006 16:21:13 +0000 (10:21 -0600)]
[IA64] resolve VTI address wrap issue

This patch resolves the address wrap issue in vtlb_purge and vhpt_purge

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
19 years agomerge
kfraser@localhost.localdomain [Wed, 16 Aug 2006 16:11:56 +0000 (17:11 +0100)]
merge

19 years agoTemporarily disable shutdown-on-HLT code.
tdeegan@york.uk.xensource.com [Wed, 16 Aug 2006 16:05:12 +0000 (17:05 +0100)]
Temporarily disable shutdown-on-HLT code.

19 years agoReplace shadow pagetable code with shadow2.
tdeegan@york.uk.xensource.com [Wed, 16 Aug 2006 16:02:35 +0000 (17:02 +0100)]
Replace shadow pagetable code with shadow2.

19 years ago[MINIOS] Fix the build after start_info interface changes.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:48:45 +0000 (16:48 +0100)]
[MINIOS] Fix the build after start_info interface changes.
Signed-off-by: Anil Madhavapeddy <anil@xensource.com>
19 years ago[IA64] enable credit schedule on VTI-domain
awilliam@xenbuild.aw [Wed, 16 Aug 2006 15:18:16 +0000 (09:18 -0600)]
[IA64] enable credit schedule on VTI-domain

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
19 years ago[LINUX] Define is_initial_xendomain() to zero for HVM drivers.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:16:32 +0000 (16:16 +0100)]
[LINUX] Define is_initial_xendomain() to zero for HVM drivers.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[LINUX] Update the hvm_op wrapper now that the return value of get_param is in
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:15:46 +0000 (16:15 +0100)]
[LINUX] Update the hvm_op wrapper now that the return value of get_param is in
the parameter structure rather than the hypercall return value.
Signed-off-by: Steven Smith <ssmith@xensource.com>
19 years ago[IA64] fix non-CONFIG_XEN sparse tree build
awilliam@xenbuild.aw [Wed, 16 Aug 2006 15:15:17 +0000 (09:15 -0600)]
[IA64] fix non-CONFIG_XEN sparse tree build

   This patch fixes some over-zealous code removal with the P==M
cleanup.  Some of the P==M code was shared with the bare metal kernel.
Also cleaned up a couple places we should have used CONFIG_XEN.

From: Juan Quintela <quintela@redhat.com>
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
19 years ago[NET] front: Fix indentation of a comment.
kfraser@localhost.localdomain [Wed, 16 Aug 2006 15:14:03 +0000 (16:14 +0100)]
[NET] front: Fix indentation of a comment.
Signed-off-by: Keir Fraser <keir@xensource.com>