smh22@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 14:57:42 +0000 (15:57 +0100)]
Avoid audit code during live migrate on 32bit since it's not required.
Signed-off-by: Steven Hand <steven@xensource.com>
smh22@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 14:54:44 +0000 (15:54 +0100)]
Tidy up audit code for PAE plus avoid its use during live migrate where
it is not strictly required.
Signed-off-by: Steven Hand <steven@xensource.com>
smh22@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 13:50:30 +0000 (14:50 +0100)]
Cleanup backend block device logic under save/restore; this should
fix the issues some people have seen with segfaults in vbd_translate
and/or user-space hangs.
Signed-off-by: Steven Hand <steven@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 13:37:24 +0000 (14:37 +0100)]
Remove shadow32 dead code (it's never built for other than
x86_32).
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 10:48:33 +0000 (11:48 +0100)]
Fix vmx guest issue of allowing accessing supervisor page
from user level program.
In shadow fault, we need to check U/S bit in error code.
It is just a fix for shadow32.c, for x86_64 code and
public code, it is already handled.
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 10:42:21 +0000 (11:42 +0100)]
Small fix for shadow mode: fix a p.t. limit check.
Signed-off-by:Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 10:40:16 +0000 (11:40 +0100)]
Clean up a previous qemu workaround where we use
/etc/xen/qemu-vgaram-bin as vgabios.
Signed-off-by: Ke Yu <ke.yu@intel.com>
kaf24@firebug.cl.cam.ac.uk [Tue, 29 Nov 2005 10:38:53 +0000 (11:38 +0100)]
Fix a race condition for multi-thread qemu dma, where vmx linux guests
show warning "dma interrupt lost" and dma becomes very slow.
root cause: In the time between set ide irq and set dma status, if guest
receive the irq and query the status, it will find the status is not
ready and therefore treat it as pseudo interrupt. Change the order of
set irq and set dma status fixes this issue.
Signed-off-by: Ke Yu <ke.yu@intel.com>
pl@us.ibm.com [Tue, 29 Nov 2005 01:00:10 +0000 (01:00 +0000)]
Fix minor traceback in xen-bugtool when send fails
rread@ubuntu.eng.hq.xensource.com [Mon, 28 Nov 2005 22:34:00 +0000 (15:34 -0700)]
The device sharing check races when more than one file backed vbd is
configured. Both of the files pick the same free loop device in the for
loop, and then one fails to losetup. This patch will now retry if
losetup fails until it succeeds or no more devices are available.
Signed-off-by: Robert Read <robert@xensource.com>
rread@ubuntu.eng.hq.xensource.com [Mon, 28 Nov 2005 19:16:27 +0000 (12:16 -0700)]
Make psmouse module builtin.
Signed-off-by: Robert Read <robert@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 18:48:54 +0000 (19:48 +0100)]
Bundle vcpu_time and vcpu_info structures together into a single
structure that is 64 bytes on x86. This ensures that indexing into
the array is fast (power-of-two size) and that accesses are cache
friendly (cache line size is usually 32 or 64 bytes).
Rename vcpu_data to vcpu_info, vcpu_time to vcpu_info.time.
Signed-off-by: Keir Fraser <keir@xensource.com>
dan@guaranine.beaverton.ibm.com [Mon, 28 Nov 2005 18:22:02 +0000 (18:22 +0000)]
Fix isConsoleDead() to actually run a command on the DomU, since a dead
console is no longer exposed when attempting to attach. This makes test
04_memset_smallmem again able to detect console death and restart
xenconsoled so that the rest of the tests can run on x86_64.
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 17:22:02 +0000 (17:22 +0000)]
Merged.
cl349@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 17:07:22 +0000 (18:07 +0100)]
Move mk-symlinks target into tools/Rules.mk.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 16:48:25 +0000 (16:48 +0000)]
Change uuid string format to match that outputted by Xen.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 16:37:15 +0000 (17:37 +0100)]
Hardcode many cpu features for x86/64 -- we know 64-bit
cpus have many features by default.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 15:24:14 +0000 (16:24 +0100)]
In alloc_vcpu_struct, after doing a memset on the new allocated vcpu, we
do a memcpy from idle0_vcpu.arch to arch of the new vcpu, which causes
monitor_pagetable is set to a none 0 value. For VMX guest which is
using external shadow mode, this is not what we really need. the
previous patch is OK for UP VMX guest, but failed on SMP VMX guest.
Signed-off-by: Xin Li <xin.b.li@intel.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 15:17:03 +0000 (16:17 +0100)]
Remove bogus extra unmap_domain_page() call in
free_monitor_pagetable().
Signed-off-by: Xin Li <xin.b.li@intel.com>
Michael.Fetterman@cl.cam.ac.uk [Mon, 28 Nov 2005 14:30:48 +0000 (15:30 +0100)]
merge
Michael.Fetterman@cl.cam.ac.uk [Mon, 28 Nov 2005 14:22:47 +0000 (15:22 +0100)]
merge
Keir did a similar change earlier, but my version removed some special case
code as well, now that __gpfn_to_mfn() can protect itself.
Michael.Fetterman@cl.cam.ac.uk [Mon, 28 Nov 2005 14:12:39 +0000 (15:12 +0100)]
minor printf and log message rework in the shadow code.
Michael.Fetterman@cl.cam.ac.uk [Mon, 28 Nov 2005 14:09:44 +0000 (15:09 +0100)]
Allow __gpfn_to_mfn() to automatically deal with translated domains != current.
Renamed gpfn_to_mfn_foreign() to get_mfn_from_pfn_foreign(), making it more
consistent with get_mfn_from_pfn().
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 13:08:53 +0000 (14:08 +0100)]
arch/xen/kernel/fixup.c -> arch/xen/i386/kernel/fixup.c
arch/xen/kernel/devmem.c -> drivers/xen/char/mem.c
arch/xen/kernel/smp.c deleted
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 28 Nov 2005 12:39:22 +0000 (13:39 +0100)]
Simplify boot_vcpu() implementation.
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 12:21:17 +0000 (12:21 +0000)]
Use StringIO to simplify string_iterator here. Altered the up-front
documentation.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 12:20:33 +0000 (12:20 +0000)]
Added README.xen-bugtool.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 11:05:07 +0000 (11:05 +0000)]
Change uses of \s within sed regexps -- some versions of sed do not support this
escape.
Allow the vif-nat and vif-route scripts to have netdev specified, rather than
hard-coding eth0.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 28 Nov 2005 01:47:28 +0000 (01:47 +0000)]
Added xen-bugtool, an application that collects various system logs and can
save them as a tarball, or submit them to a pre-existing bugzilla bug.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 13:09:46 +0000 (13:09 +0000)]
Set a console limit, in order to detect runaways.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 10:46:26 +0000 (10:46 +0000)]
Fix console handling; saveLog call was broken if ConsoleError occurred.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 01:28:11 +0000 (01:28 +0000)]
Remove keyword from call to xc.domain_setmaxmem; keyword support for this call
was removed with changeset 7955:
f72da8799418.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 01:06:44 +0000 (01:06 +0000)]
Merged.
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 01:06:20 +0000 (01:06 +0000)]
Change the console handling, to remove the loop and timeout if a console fails
to respond first time. For restored and unpaused domains there is no need for
the console to output data, so we cannot use this as a way to detect the
liveness of a console. Instead, a command must always be sent to the console,
which means that this failure can only be detected once the constructor of the
Console class has returned successfully. The __chewall loop still remains, in
order to detect runaway consoles and to clear out previous data from the
console before executing commands, but we no longer treat failure of that
method to read data as a failure of the console itself.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 01:02:36 +0000 (01:02 +0000)]
Use the getId method to tidy up simple call to domid.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 01:01:42 +0000 (01:01 +0000)]
Remove looping from handling of xm domid. This command should now work first
time after creation of a domain. Have the getRunningDomains method detect
failure of the xm list command, as indicated by the error code.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 00:59:37 +0000 (00:59 +0000)]
Mark 03_network_attach_detach as XFAIL. There are known bugs in the network
frontend device teardown.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 00:58:44 +0000 (00:58 +0000)]
Restore 05_create_noroot, as this test now seems to work, and remove
13_create_multinic from the XFAIL list, as this one works also.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 00:57:34 +0000 (00:57 +0000)]
Remove some Makefile cruft.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 00:56:34 +0000 (00:56 +0000)]
Added device-sharing checks for loopback-mounted files. The existing
check_sharing functions have been rejigged slightly so that the file-specific
stuff can share the code therein. This separates the actual test from the
generation of the error message, and introduces the canonicalise_mode function.
This also restores the broken teardown functionality for file: devices -- the
writing of XENBUS_PATH/node had been removed, but this is required for file
device teardown.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 27 Nov 2005 00:10:14 +0000 (00:10 +0000)]
Renable FAIL call that was disabled back when block-attach did not work.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sat, 26 Nov 2005 23:59:09 +0000 (23:59 +0000)]
Write the error message and the hotplug-status=busy in the same transaction, to
ensure that the message is picked up.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
smh22@firebug.cl.cam.ac.uk [Sat, 26 Nov 2005 11:17:35 +0000 (12:17 +0100)]
Fix formatting.
Signed-off-by: Steven Hand <steven@xensource.com>
smh22@firebug.cl.cam.ac.uk [Sat, 26 Nov 2005 11:13:39 +0000 (12:13 +0100)]
Fix save/restore and migrate on x86_64.
Signed-off-by: Steven Hand <steven@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 26 Nov 2005 10:32:57 +0000 (11:32 +0100)]
More cleanups to the pmd-shared patch.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 26 Nov 2005 09:43:27 +0000 (10:43 +0100)]
Clean the arch_memory_op() for-loop header.
Signed-off-by: Keir Fraser <keir@xensource.com>
smh22@firebug.cl.cam.ac.uk [Sat, 26 Nov 2005 09:24:08 +0000 (10:24 +0100)]
Fix x86_64 build.
Signed-off-by: Steven Hand <steven@xensource.com>
emellor@leeni.uk.xensource.com [Sat, 26 Nov 2005 01:21:55 +0000 (01:21 +0000)]
Merged.
emellor@leeni.uk.xensource.com [Sat, 26 Nov 2005 01:13:51 +0000 (01:13 +0000)]
Fix the unpause_basic test. The console will appear to have failed, given the
current Console.py API, because no bytes have come from it, but that's not a
failure -- an unpaused domain has no reason to output any more data. We need
to send a command to the domain before we can be sure whether it is alive or
dead, so move the FAIL call, allowing the console connection to throw
ConsoleError.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sat, 26 Nov 2005 01:08:26 +0000 (01:08 +0000)]
Fix the 06_block-list_checkremove test. The use of string.find was invalid
(should be testing against -1, not merely using if or if not), one block-list
command was missing, and the block-detach command is asynchronous (to give the
domain time to close down the device) so we need a time.sleep(1) call before
trying xm block-list.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sat, 26 Nov 2005 00:38:45 +0000 (00:38 +0000)]
Added a comment.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 23:19:23 +0000 (23:19 +0000)]
Also, make a small tweak to the 12_block_attach test so that we don't leave
a DomU running with a block device attached.
Signed-off-by: Dan Smith <danms@us.ibm.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 20:58:07 +0000 (20:58 +0000)]
Check when using physical devices that the device is not already in use. Using
a device from two guests, or from dom 0 and a guest, can destroy the filesystem,
and this check prevents that. This should ensure that the xm tests
11_block_attach_shared_dom0 and 12_block_attach_shared_domU should pass.
Closes bug #331.
Devices may be shared only if all uses are read-only. If anyone has read-write
access to a device, then no sharing is allowed.
The mode in the config file may now have an exclamation mark appended, to
indicate that devices may be shared (i.e. that this new check should be
bypassed). This supports some network block devices, but is clearly dangerous
and should only be used when you know what you are doing.
The mode specified in the config file is now written explicitly to the store,
To implement this, the blkback driver has been changed to create two entries
in /sys describing each device in use. This means that we can determine which
devices are in use without crawling through the store.
The physical-device node has been changed to give the major and minor of the
device, separated by a colon. This means that we do not need to pack these
numbers into one, removing the restriction to 8 bit minors that we had in place
before.
The mode specified in the config file is now written explicitly to the store,
rather than using the presence or absence of a read-only node. This supports
the write-sharing override above.
If the device is in use, a new hotplug status of "busy" is written to the
store, and a message is written to a new hotplug-error node. Xend uses these
things for diagnosis.
The block scripts do not need to handle online or offline events, so this
support has been removed. These scripts can be called twice, depending upon
the hotplug config, so they now identify the second run and do not attempt to
set up the device again.
The device directories in the store are cleaned out before new details are
written there. This prevents stale information from messing up block device
hotplugging. This will be one cause of "blkback: changing physical device
not supported" messages.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 17:43:35 +0000 (18:43 +0100)]
Implement new memory_op() XENMEM_machphys_mfn_list. Replaces old
privcmd Linux ioctl and includes an implementation for x86/64.
Signed-off-by: Keir Fraser <keir@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 17:12:12 +0000 (17:12 +0000)]
Only write a success node when coming online, not going offline. Delete the
interface from the bridge when going offline -- the comment stating that
interfaces are auto-removed does not seem to be true, at least some of the
time. Log failure of brctl delif and ifconfig down only at debug level, as
this script may be called more than once, depending upon the hotplug
configuration.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 17:09:49 +0000 (17:09 +0000)]
Only write a success node when coming online, not going offline.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 17:08:22 +0000 (17:08 +0000)]
Don't issue the iptables diagnostic when removing entries. Depending upon your
hotplug config, the offline event may come through to this script twice, and
the iptables -D will fail the second time.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 17:07:19 +0000 (17:07 +0000)]
Match rename of parameters to match the corresponding hotplug events. This
took place some time ago, so presumably not many people are using this scripts!
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 17:05:45 +0000 (17:05 +0000)]
Fix handling of devid_end when devid is an integer, but not present as a device.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 16:58:36 +0000 (17:58 +0100)]
In shadow-external mode, every VCPU must free its own
monitor pagetable.
Signed-off-by: Xin Li <xin.b.li@intel.com>
smh22@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 16:53:41 +0000 (17:53 +0100)]
Tolerate temporary inconsistency between type info and page contents on
(live) restore. This should now enable live migrate of fork-intensive (or
other page-table-intensive) workloads.
Signed-off-by: Steven Hand <steven@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 15:26:19 +0000 (15:26 +0000)]
check_hotplug fails on one of my machines running ubunto 4.10 "Warty
Warhog", due to misparsing the udevinfo version string. The problem is
caused by an old version of sed that does not appreciate the
regexp. This patch fixes it, and also works on debian unstable and
SLES9SP2. Please apply...
Signed-Off-By: Muli Ben-Yehuda <mulix@mulix.org>
Signed-Off-By: Orna Agmon <ladypine@ladypine.org>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 15:18:26 +0000 (15:18 +0000)]
Add help option on runtests.sh
Signed-off-by: Murillo Fernandes Bernardes <mfb@br.ibm.com>
dan@elm3b196.beaverton.ibm.com [Fri, 25 Nov 2005 15:17:37 +0000 (15:17 +0000)]
Add sanity checks that let us know if there are any problems before
we run the entire suite
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 15:17:18 +0000 (15:17 +0000)]
Fix stupid invalid syntax on network-attach tests.
Signed-off-by: Murillo Fernandes Bernardes <mfb@br.ibm.com>
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 11:19:09 +0000 (11:19 +0000)]
Merged.
emellor@leeni.uk.xensource.com [Fri, 25 Nov 2005 11:19:03 +0000 (11:19 +0000)]
Move XenbusState from the Linux-specific xenbus.h into a new file with the Xen
public interfaces.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
smh22@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 09:51:09 +0000 (10:51 +0100)]
Use page-sized batches in GETPAGEFRAMEINFO2 (and generally tidy up).
Signed-off-by: Steven Hand <steven@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 08:17:34 +0000 (09:17 +0100)]
Fix the issue of on x86_64, rhel4 can not get into runlevel 5.
Xserver will scan pci and the cirrus VGA driver will register
mmio each time, cause mmio array out of space.
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 08:15:08 +0000 (09:15 +0100)]
I cleaned up acm_ops.c and eliminated returns inside the switch
statement. When we need locks, we can place them now around the switch
statement.
I also included the comments from Rusty and now return -EPERM for denied
permission errors.
Signed-off: Reiner Sailer <sailer@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 25 Nov 2005 08:14:01 +0000 (09:14 +0100)]
Update vtpm driver following the recent changes to the
xenbus driver and its semantics.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 22:21:48 +0000 (23:21 +0100)]
Big reworking of SHARED_KERNEL_PMD logic. Includes several
bug fixes for PAE, and reverts my previous changeset that
broke non-pae.
Signed-off-by: Keir Fraser <keir@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 19:57:01 +0000 (19:57 +0000)]
Merged.
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 19:56:38 +0000 (19:56 +0000)]
Programmatically generate the xm help strings from the snippets, rather than
splicing the strings together manually.
Replace the word "bogus" with the command in question where it appears, as this
shows up in the Usage statements.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 19:53:09 +0000 (19:53 +0000)]
Added a count of the number of users of this block device, and refuse to close
down the frontend until that number becomes 0. There are no reasonable
semantics for hot-unplugging a block device beneath a mounted filesystem, and
this solution ensures that the guest is not crashed nor the filesystem corruptedby the actions of the administrator.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 19:01:16 +0000 (20:01 +0100)]
Remove CONFIG_PREEMPT configuration option from Xen kernel builds.
The option reduces overall system performance and makes little sense
when running over a preemptive hypervisor. Removing the option also
fixes some hypothetical races.
The forthcoming voluntary-preemption option will make much more sense
to enable when running over Xen.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 18:49:15 +0000 (19:49 +0100)]
Quieten down grant-table transfer failure when it is because
a domain is dying.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 18:40:18 +0000 (19:40 +0100)]
Another PAE fix for pgd_ctor.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 18:16:13 +0000 (19:16 +0100)]
More PAE PGD fixes. Ensure we do not pin PGDs with no kernel
PMD (Xen will not allow it). This is okay even across
suspend/resume -- such PGDs are completely empty and so
contain no PTEs that become garbage across suspend/resume.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 17:19:51 +0000 (18:19 +0100)]
Fix pgd allocation for PAE. We must handle the fact that
a PGD may get pinned under our feet by suspend/resume.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 16:13:46 +0000 (17:13 +0100)]
Fix 32b non-pae build.
Signed-off-by: Keir Fraser <keir@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 15:57:45 +0000 (15:57 +0000)]
Merged.
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 15:57:36 +0000 (15:57 +0000)]
Check whether a device is already configured before trying to configure it
again. Closes bug #397.
To perform this check in the same transaction as the writing of the device
entries has required a little bit of rejigging.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 15:55:49 +0000 (15:55 +0000)]
Support destruction of device by /dev/sda1 (for example) even when the
configuration just said sda1.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 24 Nov 2005 15:52:42 +0000 (15:52 +0000)]
Minor tidy, no semantic change.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 15:47:44 +0000 (16:47 +0100)]
Shootdown TLB entries across all VCPUs for SMP shadow mode.
Signed-off-by: Keir Fraser <keir@xensource.com>
smh22@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 15:32:38 +0000 (16:32 +0100)]
Re-enable tools support for live migration on PAE.
Signed-off-by: Steven Hand <steven@xensource.com>
smh22@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 15:30:40 +0000 (16:30 +0100)]
Quiet down live migrate on PAE domains.
Signed-off-by: Steven Hand <steven@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 14:38:33 +0000 (15:38 +0100)]
Clean up mark_dirty() implementation to check for log-dirty
status internally, with the shadow_lock held for consistency.
Callers no need to check log-dirty status themselves.
Also, add dirty logging to alloc_page_type/free_page_type, so
that we see page tables coming and going during live relocation.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 14:08:27 +0000 (15:08 +0100)]
Remove unused 'nr_recs' field from trace-buffer header
structure. This also happens to ensure correct t_rec
alignment on ia64. A better layout for trace buffers
will be coming soon.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 11:06:07 +0000 (12:06 +0100)]
Adds ac_timer based polling to the ns16550 UART driver. This is
useful when the interrupt line is not connected in hardware or the
mechanism to enable it is not readily available in the hypervisor.
Polling is only enabled when the UART IRQ is set to zero.
Signed-off-by: Alex Williamson <alex.williamson@hp.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 10:17:20 +0000 (11:17 +0100)]
This fixes a small race between when a domain is created and when xentop
tries to read it's info from the xenstore. Instead of sending a NULL
pointer for the name the name will be displayed as a ' ' until the next
refresh where xentop will then properly display the name.
Signed-off-by: Jerone Young <jyoung5@us.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 10:10:43 +0000 (11:10 +0100)]
xen/common/acm_ops.c, check for a NULL pointer and then
cheerfully dereferences it.
Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 10:06:26 +0000 (11:06 +0100)]
dom0_ops return -EACCES when the interface version in wrong. I think
that an acm failure is closer to being non-priv'ed, which returns
-EPERM. Wouldn't want dom0 tools to report they were the wrong version
just because ACM limited some operation, surely?
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (authored)
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 10:05:21 +0000 (11:05 +0100)]
Clean up some Xen files in the ioemulator.
1) fix indent
2) use fprintf(logfile, ...) instead of perror and printf
3) remove tailing spaces
Signed-off-by: Xin Li <xin.b.li@intel.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 24 Nov 2005 10:03:07 +0000 (11:03 +0100)]
Fix ia64 build. Two typos in header files.
Signed-off-by: Keir Fraser <keir@xensource.com>
rread@ubuntu.eng.hq.xensource.com [Thu, 24 Nov 2005 05:04:13 +0000 (22:04 -0700)]
Make the microcode driver builtin.
As a module it complains about unknown symbols sys_munlock and sys_lock.
rread@ubuntu.eng.hq.xensource.com [Thu, 24 Nov 2005 04:55:07 +0000 (21:55 -0700)]
Disable crufty modules still using cli and friends.
rread@ubuntu.eng.hq.xensource.com [Thu, 24 Nov 2005 04:12:38 +0000 (21:12 -0700)]
Compile IDE drivers as builtins instead of modules.
This makes the -xen config compatible with the distribution kernel configs,
but at the cost of some bloat when being used as a guest kernel.
emellor@leeni.uk.xensource.com [Wed, 23 Nov 2005 19:37:33 +0000 (19:37 +0000)]
Reverse accidental commits for changesets 8001 and 7996-7999.
The Mercurial Queues extensions suddenly looks a lot less useful than it first
seemed.
Signed-off-by: Ewan Mellor <ewan@xensource.com>