]> xenbits.xensource.com Git - people/ssmith/nc2-2.6.27.git/log
people/ssmith/nc2-2.6.27.git
15 years agoGet the type of sframes right, and hence avoid a memory clobber in nc2/forklift-over-patchqueue
Steven Smith [Fri, 19 Jun 2009 14:54:27 +0000 (15:54 +0100)]
Get the type of sframes right, and hence avoid a memory clobber in
32 bit guests.

15 years agoAdd a .gitignore.
Steven Smith [Thu, 28 May 2009 15:30:36 +0000 (16:30 +0100)]
Add a .gitignore.

15 years agoPull in netchannel2 from 2.6.18.
Steven Smith [Thu, 28 May 2009 15:28:52 +0000 (16:28 +0100)]
Pull in netchannel2 from 2.6.18.

15 years agoV2 grant table support.
Steven Smith [Thu, 28 May 2009 14:02:36 +0000 (15:02 +0100)]
V2 grant table support.

15 years agopatch intel-hda-2.6.30 base-plus-patchqueue-97b1e293bee58f2cffa45260a51f349b2516de05
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch intel-hda-2.6.30

15 years agopatch remove-release-flr
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch remove-release-flr

15 years agopatch pass2-driver
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch pass2-driver

15 years agopatch snd-hda-intel-hp-elite-6930p-laptop-mode
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch snd-hda-intel-hp-elite-6930p-laptop-mode

15 years agopatch pciback-flr
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch pciback-flr

15 years agopatch xen-acpi-wmi
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch xen-acpi-wmi

15 years agopatch fix_domain_switch_code
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch fix_domain_switch_code

15 years agopatch sysrq-loglevel
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch sysrq-loglevel

15 years agopatch oom-debug-me-harder
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch oom-debug-me-harder

15 years agopatch oom-debugging
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch oom-debugging

15 years agoAdd extra debugging code to find out what's going on when unaccounted
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
Add extra debugging code to find out what's going on when unaccounted
space is used in swap.

15 years agopatch bridge-no-topology-change-when-no-stp.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bridge-no-topology-change-when-no-stp.patch

15 years agopatch bonding-vlan-fixes.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-vlan-fixes.patch

15 years agopatch debug-dump-skb-info-when-invalid
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch debug-dump-skb-info-when-invalid

15 years agopatch bonding-no-updelay-on-first-active-slave.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-no-updelay-on-first-active-slave.patch

15 years agopatch bonding-balance-slb-fixes3.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-balance-slb-fixes3.patch

15 years agopatch bonding-balance-slb-fixes2.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-balance-slb-fixes2.patch

15 years agopatch bonding-balance-slb-fixes.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-balance-slb-fixes.patch

15 years agopatch quiet-no-ufo
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch quiet-no-ufo

15 years agopatch promisc-bridging
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch promisc-bridging

15 years agopatch bonding-default-slb.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-default-slb.patch

15 years agopatch bonding-balance-slb.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bonding-balance-slb.patch

15 years agopatch bridge-locate-physical-device.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bridge-locate-physical-device.patch

15 years agoAdd a new ioctl to /proc/xen/privcmd which allows you to do certain
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a new ioctl to /proc/xen/privcmd which allows you to do certain
XENMEM operations on restricted fds.

15 years agoAdd a new ioctl to /proc/xen/privcmd which allows SCHEDOP_shutdown to
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a new ioctl to /proc/xen/privcmd which allows SCHEDOP_shutdown to
be applied on restricted fds.

15 years agoAdd a new ioctl to /proc/xen/privcmd which allows HVM operations to be
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a new ioctl to /proc/xen/privcmd which allows HVM operations to be
performed on restricted domains.

15 years agoAdd a new ioctl to /proc/xen/privcmd which allows domctls to be performed
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a new ioctl to /proc/xen/privcmd which allows domctls to be performed
without using the generic hypercall interface, so that they are available
on restricted fds.

This requires an unfortunate amount of fiddling with headers so that
XEN_GUEST_HANDLE_64 and uint64_aligned_t are available in kernel
space.

15 years agoAdd a RESTRICT ioctl to /proc/xen/privcmd, which allows a privcommand
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a RESTRICT ioctl to /proc/xen/privcmd, which allows a privcommand
file descriptor to be restricted to only work with one domain.

Certain difficult operations, like hypercalls, are prohibited completely
on restricted handles.

15 years agoAdd a RESTRICT ioctl to /dev/xen/evtchn, which allows an event channel
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add a RESTRICT ioctl to /dev/xen/evtchn, which allows an event channel
file descriptor to be restricted to only working with a particular domain.

15 years agopatch CA-13783-ipmi-fixes
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-13783-ipmi-fixes

15 years agopatch CA-14400-reduce-printk-levels
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-14400-reduce-printk-levels

15 years agopatch quiet-intel-rng
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch quiet-intel-rng

15 years agopatch quiet-sd-cache-info
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch quiet-sd-cache-info

15 years agopatch quieten-serio-i8042-warning.patch
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch quieten-serio-i8042-warning.patch

15 years agoIndex: tmp-ca4382/fs/fat/inode.c
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Index: tmp-ca4382/fs/fat/inode.c
===================================================================

15 years agopatch kernel-configuration
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch kernel-configuration

15 years agopatch create_config.sh-x86_64-sane-default
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch create_config.sh-x86_64-sane-default

15 years agopatch utility-guest
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch utility-guest

15 years agopatch config-create_config-script
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch config-create_config-script

15 years agoWatch the online node in the backend area, as well as the state node
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Watch the online node in the backend area, as well as the state node
in the frontend area, and fire the frontend state changed watch
whenever it changes.  This allows us to catch the case where a device
shuts down in a domU and then gets xm detach'd from in dom0.

Otherwise, the backend doesn't shut down correctly, since online was
set when the frontend shut down and we don't get another kick when it
becomes unset.

15 years ago__gnttab_dma_map_page can be called from a softirq (via the network
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
__gnttab_dma_map_page can be called from a softirq (via the network
transmit softirq for example) therefor gnttab_copy_grant_page needs to
take gntab_dma_lock in an interrupt safe manner.

15 years ago# HG changeset patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
# HG changeset patch
# User rread@ubuntu.eng.hq.xensource.com
# Node ID 86a2de4955f301841fb7d254e9dd5d801d108deb
# Parent  2a2de73f10ef15f396d54e31953e7bc4640e9252
[xen-xo] CA-259 fake out GET_IDLUN scsi ioctl

Signed-off-by: Andrew Peace <andrew.peace@xensource.com>
Signed-off-by: Robert Read <rread@xensource.com>
15 years agopatch netback-SET_NETDEV_DEV-before-register_netdev.patch
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch netback-SET_NETDEV_DEV-before-register_netdev.patch

15 years agopatch netback-thread
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch netback-thread

15 years agopatch netback-defensive.patch
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch netback-defensive.patch

15 years agoThere's no point in sending lots of little packets to a copying
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
There's no point in sending lots of little packets to a copying
receiver if we can instead arrange to copy them all into a single RX
buffer.  We need to copy anyway, so there's no overhead here, and this
is a little bit easier on the receiving domain's network stack.

15 years agoEnsure that packet csums are computed correctly when sending a GSO
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Ensure that packet csums are computed correctly when sending a GSO
packet to an interface which supports scatter-gather but not transmit
checksum offloads.

Signed-off-by: Steven Smith <ssmith@xensource.com>
15 years ago[NETBACK] Try to pull a minimum of 72 bytes into the skb data area
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
[NETBACK] Try to pull a minimum of 72 bytes into the skb data area
when receiving a packet into netback.  The previous number, 64, tended
to place a fragment boundary in the middle of the TCP header options
and led to unnecessary fragmentation in Windows <-> Windows
networking.

Signed-off-by: Steven Smith <ssmith@xensource.com>
15 years agoIt is possible for a frontend to generate a TSO request which doesn't
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
It is possible for a frontend to generate a TSO request which doesn't
actually need segmentation (i.e. with size < MTU).  Make sure this
doesn't crash the backend.

15 years agoThe Windows drivers push the network frontend to state Closed, then
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
The Windows drivers push the network frontend to state Closed, then
Initialised, then Closed again as part of device disable.  Make sure
the backend doesn't get stuck at closed.

15 years agoArrange that netback waits for the hotplug scripts to complete before
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Arrange that netback waits for the hotplug scripts to complete before
going to state Connected.  WHQL gets quite upset if it sends packets
which don't arrive, and that can happen if our hotplug scripts are
slow and don't hook the network interface up to the bridge in time.

15 years agopatch netback-watch-csum-offload
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch netback-watch-csum-offload

15 years agopatch disconnect-netback-on-close
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch disconnect-netback-on-close

15 years agoIt turns out that Windows occasionally generates packets in which the
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
It turns out that Windows occasionally generates packets in which the
IP and TCP headers are in different fragments.  Make sure that the
backends can handle this.

15 years agoAdd support to netback for delivering packets at a certain offset into
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add support to netback for delivering packets at a certain offset into
the page.

15 years agocommit 3976878f6d3c6faa1aed3a9d40ccf8c10647cd36
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
commit 3976878f6d3c6faa1aed3a9d40ccf8c10647cd36
Author: Ian Campbell <Ian.Campbell@citrix.com>
Date:   Fri Mar 27 15:44:31 2009 +0000

    blktap: cleanup usage of sysfs
    - do not deadock when writing to remove node
    - check __must_check function return values

15 years agopatch forward-port-block-drivers.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch forward-port-block-drivers.patch

15 years agoCA-27974: Fix blktap shutdown race due to improper event ordering.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
CA-27974: Fix blktap shutdown race due to improper event ordering.

Writing shutdown-done before switching device state to closed (6)
opens a remarkably small race window to fall through: The agent
removes the device directory just before the write to the 'state'
field will recreate it again. This in turn leads to xenbus failing to
remove the device, since removal is guided by directory existence.

With shutdown-done and connection state being rather independent,
trivially fixing event ordering to write shutdown-done last appears
safe but mandatory. Comment this tiny detail.

15 years agoprefer spin_lock_init to SPIN_LOCK_UNLOCKED.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
prefer spin_lock_init to SPIN_LOCK_UNLOCKED.

The former is prefered and keep lockdep happy.

Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
15 years agoCA-24604: Signal blktap kthread start as a state change notification to udev.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
CA-24604: Signal blktap kthread start as a state change notification to udev.

15 years agoCA-25742: Forward a shutdown-request="force" to userspace from blktap.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
CA-25742: Forward a shutdown-request="force" to userspace from blktap.

15 years agopatch blktap2
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blktap2

15 years agopatch blkback-pagemap
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blkback-pagemap

15 years agopatch CA-24784-resource-leak
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-24784-resource-leak

15 years agopatch CA-24267-blkback-unpause
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-24267-blkback-unpause

15 years agopatch CA-20346-blktap-vma-unmap
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-20346-blktap-vma-unmap

15 years agopatch blk-latency-stats
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blk-latency-stats

15 years agopatch ratelimit-blktap-wprintk
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch ratelimit-blktap-wprintk

15 years agoClose block devices when the pv drivers take over and flush the buffer cache.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
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

15 years agopatch CA-15999-blkback-pause-unpause
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-15999-blkback-pause-unpause

15 years agopatch CA-15586-blkback-close-bdev
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-15586-blkback-close-bdev

15 years agopatch CA-14804-fix-block-unplug-retries
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-14804-fix-block-unplug-retries

15 years agopatch blk-fix-sysfs-remove-race
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blk-fix-sysfs-remove-race

15 years agopatch CA-12483-fix-forced-shutdown-race
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-12483-fix-forced-shutdown-race

15 years agopatch CA-9002-blktap-plug
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-9002-blktap-plug

15 years agopatch CA-8806-blk-dont-reconnect-on-unplug
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-8806-blk-dont-reconnect-on-unplug

15 years agopatch CA-7672-blk-shutdown.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch CA-7672-blk-shutdown.patch

15 years agoWakeup blktap thread once frontend is connected -- in case there are
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
Wakeup blktap thread once frontend is connected -- in case there are
already pending requests.  Do same for blkback.

15 years agopatch blktap-shutdown-cleanup2
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blktap-shutdown-cleanup2

15 years agopatch blktap-shutdown-cleanup
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blktap-shutdown-cleanup

15 years agopatch blk-name
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blk-name

15 years agoAdd user-space facing part of blkfront to blktap to provide a
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
Add user-space facing part of blkfront to blktap to provide a
device in the backend domain giving access to the blktap device.

15 years agopatch blktap-ioprio.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch blktap-ioprio.patch

15 years agopatch backend-track-pages-with-page-flags.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch backend-track-pages-with-page-flags.patch

15 years agopatch get_user_pages_fast-support-foreign-pages.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch get_user_pages_fast-support-foreign-pages.patch

15 years agosys_mlock and sys_munlock are needed to build the microcode driver as a module.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
sys_mlock and sys_munlock are needed to build the microcode driver as a module.
Index: tmp-mlock/mm/mlock.c
===================================================================

15 years agopatch cdrom-sysctl-info.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch cdrom-sysctl-info.patch

15 years agopatch nfs-allow-0-retransmits.patch
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch nfs-allow-0-retransmits.patch

15 years agopatch forward-port-bnx2i.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch forward-port-bnx2i.patch

15 years agopatch bnx2i-host-params
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bnx2i-host-params

15 years agopatch bnx2i-1.3.9
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch bnx2i-1.3.9

15 years agoAvoid potential deadlock which wedges dom0.
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
Avoid potential deadlock which wedges dom0.

15 years agopatch fix-cfq-iscsi-live-lock.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch fix-cfq-iscsi-live-lock.patch

15 years agopatch open-iscsi-gfpkernel.patch
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
patch open-iscsi-gfpkernel.patch

15 years agoAdd second location for acceleration config option for XenServer
Steven Smith [Tue, 30 Jun 2009 11:55:48 +0000 (12:55 +0100)]
Add second location for acceleration config option for XenServer

15 years agopatch forward-port-s2io.patch
Steven Smith [Tue, 30 Jun 2009 11:55:47 +0000 (12:55 +0100)]
patch forward-port-s2io.patch