patch ratelimit-blktap-wprintk

Close block devices when the pv drivers take over and flush the buffer cache.
- close and free the block devices in qemu when we switch to pv drivers in
  the guest
- use BLKFLSBUF to flush the buffer cache, both in qemu and in blkback

patch CA-15999-blkback-pause-unpause

patch CA-15586-blkback-close-bdev

patch CA-14804-fix-block-unplug-retries

patch blk-fix-sysfs-remove-race

patch CA-12483-fix-forced-shutdown-race

patch CA-9002-blktap-plug

patch CA-8806-blk-dont-reconnect-on-unplug

patch CA-7672-blk-shutdown.patch

Wakeup blktap thread once frontend is connected -- in case there are
already pending requests.  Do same for blkback.

patch blktap-shutdown-cleanup2

patch blktap-shutdown-cleanup

patch blk-name

Add user-space facing part of blkfront to blktap to provide a
device in the backend domain giving access to the blktap device.

patch blktap-ioprio.patch

patch backend-track-pages-with-page-flags.patch

patch get_user_pages_fast-support-foreign-pages.patch

sys_mlock and sys_munlock are needed to build the microcode driver as a module.
Index: tmp-mlock/mm/mlock.c
===================================================================

patch cdrom-sysctl-info.patch

patch nfs-allow-0-retransmits.patch

patch forward-port-bnx2i.patch

patch bnx2i-host-params

patch bnx2i-1.3.9

Avoid potential deadlock which wedges dom0.

patch fix-cfq-iscsi-live-lock.patch

patch open-iscsi-gfpkernel.patch

Add second location for acceleration config option for XenServer

patch forward-port-s2io.patch

REL_2.0.31.12965_LX-s2io.tar.gz taken from. Some files moved into
drivers/net/s2io/*.h due to exceedingly generic names.

patch bnx2x-disable-tpa

patch bnx2x-fix-paths

patch bnx2x-1.46.12

patch bnx2-cnic-over-bridge

patch bnx2-tune-interrupt-load

patch bnx2-1.8.2b

patch tg3-build-without-vlan-support.patch

patch tg3-tune-interrupt-load

patch tg3-3.92n

patch intel-net-driver-conflicting-names.patch

patch intel-net-driver-kcompat.patch

patch ixgbe-1.3.56.5.patch

patch ixgb-1.0.135.patch

patch igb-1.3.8.6.patch

patch e1000e-0.5.8.2.patch

patch e1000-8.0.6.patch

patch kexec-define-vmcore_elf_check_arch_cross.patch

only in patch2:
unchanged:

patch CA-14360-loadavg-not-uninterruptible

patch pci-ignore

Fix build failure with gcc 4.3

  LD      .tmp_vmlinux1
arch/x86/kernel/built-in.o: In function `timer_interrupt':
/data/work/linux-2.6.27/linux-2.6/arch/x86/kernel/time_32-xen.c:467: undefined reference to `__udivdi3'

patch xen-interface-xen_pcie_aer_op.patch

patch xen-interface-HVMOP.patch

patch XenSource

commit c6cb0e878446c79f42e7833d7bb69ed6bfbb381f
Author: Alexey Starikovskiy <astarikovskiy@suse.de>
Date:   Wed Jan 14 02:57:53 2009 +0300

    ACPI: EC: Don't trust ECDT tables from ASUS

    http://bugzilla.kernel.org/show_bug.cgi?id=9399
    http://bugzilla.kernel.org/show_bug.cgi?id=11880

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

commit 235c4a59278eb07e61d909f1f0c233733034a8b3
Author: Alexey Starikovskiy <astarikovskiy@suse.de>
Date:   Wed Jan 14 02:57:47 2009 +0300

    ACPI: EC: Limit workaround for ASUS notebooks even more

    References: http://bugzilla.kernel.org/show_bug.cgi?id=11884

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

commit 3e54048691bce3f323fd5460695273be379803b9
Author: Hannes Eder <hannes@hanneseder.net>
Date:   Sat Nov 29 07:21:29 2008 +0100

    ACPI: EC: fix compilation warning

    Fix the warning introduced in commit c5279dee26c0e8d7c4200993bfc4b540d2469598,
    and give the dummy variable a more verbose name.

      drivers/acpi/ec.c: In function 'acpi_ec_ecdt_probe':
      drivers/acpi/ec.c:1015: warning: ISO C90 forbids mixed declarations and code

Signed-off-by: Hannes Eder <hannes@hanneseder.net>
Acked-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

commit c5279dee26c0e8d7c4200993bfc4b540d2469598
Author: Alexey Starikovskiy <astarikovskiy@suse.de>
Date:   Wed Nov 26 17:11:53 2008 -0500

    ACPI: EC: Add some basic check for ECDT data

    One more ASUS comes with empty ECDT, add a guard for it...

    http://bugzilla.kernel.org/show_bug.cgi?id=11880

Signed-off-by: Alexey Starikovskiy <astarikovskiy@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>

[BLKTAP] If a frontend moves its request producer pointer backwards
over the backend's request consumer, blktap will currently count to
infinity trying to make the consumer catch up with the producer again.
During this time, it's impossible to shut down the tap device, and
this makes the toolstack very sad.  Fix it by deliberately breaking
the interface whenever we notice that this has happened.

Signed-off-by: Steven Smith <steven.smith@eu.citrix.com>

patch revert-linux-2.6.18-xen.hg-582.b29a06ba7a5f

patch revert-linux-2.6.18-xen.hg-598.f1d726b984a3

patch revert-linux-2.6.18-xen.hg-719.be9555ea5512

patch revert-linux-2.6.18-xen.hg-740.5012c470f875

# HG changeset patch
# User Keir Fraser <keir.fraser@citrix.com>
# Date 1237377065 0
# Node ID c0f2f398aa3cab195b632531447e95110e9d8282
# Parent  c1f0373ff44ea3f9ba3845e05faf3d27d7d52731
x86: Fix interaction of NTP and dom0->xen time updates

Don't discard NTP sync when updating Xen wallclock time from dom0,
as that's almost the first thing we do when we become synced.
Move the call to ntp_clear() into do_settimeofday(), which is the
only caller of __update_wallclock() that looks like it should break
NTP sync.

This fixes the timer chain that sets Xen's wallclock every minute when
dom0 is NTP synced, which in turn greatly improves wallclock accuracy
in PV domU.

Signed-off-by: Tim Deegan <Tim.Deegan@citrix.com>

patch forward-port-drivers-xen.patch

patch sync-drivers-xen.patch

patch linux-2.6.27.19-5.1.patch

patch build-system-integration.patch

patch hgignore.patch

Add patch queue.

Import 2.6.27