kaf24@firebug.cl.cam.ac.uk [Sun, 30 Oct 2005 09:45:49 +0000 (10:45 +0100)]
Fix floating-point corruption (a nasty race in fp task-switch
exception handling).
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sun, 30 Oct 2005 08:39:55 +0000 (09:39 +0100)]
Add a commented-out sanity check for lazy fpu switching.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 29 Oct 2005 08:51:35 +0000 (09:51 +0100)]
merge
jbulpin@wicket.uk.xensource.com [Sat, 29 Oct 2005 00:58:00 +0000 (01:58 +0100)]
Prevent event listener socket from being inherited by children.
This was passing down to dhclient which was keeping the 8001 port
open and preventing xend restarting.
Also fixed a case typo in setCloExec
Signed-off-by: James Bulpin <james@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 28 Oct 2005 16:51:37 +0000 (17:51 +0100)]
Remove next_io_page field from domain structure.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 28 Oct 2005 16:11:59 +0000 (17:11 +0100)]
Clean up the 'initial mapping' area properly on x86/64.
We must completely blow away mappings of the original
p2m table and the initrd. Those areas eventually get freed
to the memory allocator and other uses of that memory will
get really confused by spurious redundant mappings in the
init mapping area.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 28 Oct 2005 15:41:11 +0000 (16:41 +0100)]
Fix sizeof_vcpu_shift in x86/64 xen_entry.S.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 28 Oct 2005 08:48:46 +0000 (09:48 +0100)]
This patch provide local APIC support for vmx guest.
A configure option is also added to disable it, which is off by default.
Signed-off-by: Jiang Yunhong <yunhong.jiang@intel.com>
Signed-off-by: Li Xin <xin.b.li@intel.com>
Signed-off-by: Eddie Dong <eddie.dong@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 17:22:45 +0000 (18:22 +0100)]
xm help notes that domid can be used in place of domname.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 17:21:57 +0000 (18:21 +0100)]
No side effects in BUG_ON statements in Xen.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:47:53 +0000 (17:47 +0100)]
To allow kernel to handle radically different hypervisors in the future,
I thought a magic field in the start_info might be useful. I just chose
"Xen-<major>.<minor>", but feel free to change it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au> (authored)
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:43:29 +0000 (17:43 +0100)]
I've created a patch to get the mini-os to compile in Xen-unstable. The patch
is pretty big, but this is mostly because I've replaced hypervisor.h with the
hypercalls.h header file from Linux (which seems easier to maintain
anyway). I've ignored X86_64.
Signed-off-by: Simon Kagstrom <simon.kagstrom@bth.se>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:29:39 +0000 (17:29 +0100)]
Originally in qemu when a IDE dma transfer is started which is triggered
by access to 0xc000 port in VMX, qemu will blocked there until it
finished. And during that block, there are extra dom switches between
dom0 and idle domain, and VMX can't be resumed. By making real IDE dma
tranfer run in another thread, qemu will return and VMX resume on time
and there are fewer cycle wasted.
Signed-off-by: Ke Yu <ke.yu@intel.com>
Signed-off-by: Xiaowei Yang <xiaowei.yang@intel.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:27:04 +0000 (17:27 +0100)]
The attached patch fixes 2 issues with the scripts written for the Xen
access control module and makes the tools more self-explanatory.
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:24:06 +0000 (17:24 +0100)]
This patch
* adapts the Xen access control module to recent changes in the "unbound
event channel" interface and its usage
* adds checks that became necessary because of domain structures "hanging
around" in the domain list after domain destruction
This patch is tested on today's xen-unstable with Null, CHWALL, STE,
CHWALL_STE policies.
Signed-off by: Reiner Sailer <sailer@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:20:33 +0000 (17:20 +0100)]
Remove all binaries on make clean.
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:16:48 +0000 (17:16 +0100)]
Remove a non-portable and unused error.h, and use signal.h not
sys/signal.h (in tools/misc).
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:16:12 +0000 (17:16 +0100)]
Use standard int typedefs.
Signed-off-by: John Levon <levon@movementarian.org>
kaf24@firebug.cl.cam.ac.uk [Thu, 27 Oct 2005 16:09:01 +0000 (17:09 +0100)]
BUG_ON() is a macro very very similar to assert(), and it's a really bad
idea in general to put statements with side-effects inside such a construct
(for example the BUG_ON() could be compiled away for non-debug builds).
The patch below fixes this for the drivers/xen tree.
I chose the general construct to store the respective return values in a
local var instead of an if (...) BUG(); because it looks like most of these
BUG_ON()'s need to be replaced by real error handling when going towards
production releases anyway, and then the local variable is going to be
needed anyway. (and besides the if construct tends to get unreable due to
the unlikely() etc).
Please apply.
Greetings,
Arjan van de Ven
emellor@leeni.uk.xensource.com [Thu, 27 Oct 2005 09:53:02 +0000 (10:53 +0100)]
Fix Jerone Young's xenstat / xenstore patch:
Reject the removal of -Werror from the xentop Makefile.
Reject the #include of xs.h inside xentop -- there is no need for it.
Have XendDomainInfo write domain names into /local/domain (which used to happen,
but hasn't for a little while).
Remove the dead code from xentop.c.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
jeroney@localhost.localdomain [Thu, 27 Oct 2005 09:19:58 +0000 (10:19 +0100)]
* Enable xenstat to use xenstore
* Have xentop display names instead of DomIDs
as requested in Bugzilla#311 filed by Ian
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 15:59:13 +0000 (16:59 +0100)]
Fix the check for non-existence of the save/restore directories, by using &&
rather than -a. The former shortcuts at the script level, whereas the latter
does not, which means that the backtick-enclosed arguments are evaluated
regardless of the success of preceding tests.
Tidy the aforementioned test into a function of its own.
Add an implementation of usleep for those systems that only have sleep (Debian,
for example).
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 13:59:29 +0000 (14:59 +0100)]
Move exception handler for sxp.ParseError above that for Exception, as the
latter subsumes the former. Minor tidy-ups.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 13:58:13 +0000 (14:58 +0100)]
Break the write_dev function out from block into new block-common.sh, where it
can also be used by block-enbd and block-nbd.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 13:57:16 +0000 (14:57 +0100)]
Added copyright notice.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 13:55:32 +0000 (14:55 +0100)]
New xendomains init script, courtesy of Kurt Garloff:
It now allows the user to configure what he wants to be done on
xendomains stop: sysrq, migrate, save, or shutdown.
For this to be safe, a timeout has been introduced (optional).
On start, saved domains may be restored, and not yet started
domains from the auto start dir will be started in addition.
status reports missing domains, restart and reload work as well.
The init script is somewhat SUSE-ish, but it should work on LSB
compliant and on RH based distros as well.
I would appreciate this to be put into tools/examples/init.d/
Signed-off-by: Kurt Garloff <garloff@suse.de>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 12:41:00 +0000 (13:41 +0100)]
Check whether dom is NULL when we handle a watch. This might occur if the
domain has recently been removed from our list, but with a watch still in
flight. This should fix the intermittent crash in xenconsoled that is plaguing
us at the moment.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Wed, 26 Oct 2005 11:06:58 +0000 (12:06 +0100)]
Added comment.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
jeroney@localhost.localdomain [Tue, 25 Oct 2005 16:37:00 +0000 (17:37 +0100)]
Remove second decleration of .PHONY & cleanup formating
emellor@leeni.uk.xensource.com [Tue, 25 Oct 2005 15:00:49 +0000 (16:00 +0100)]
Fix format spec for tracing.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Tue, 25 Oct 2005 10:52:56 +0000 (11:52 +0100)]
NBD scripts, courtesy of Kurt Garloff:
For domain migration, unless you use a rootless setup, you need
to have network storage available. For the poor people like me that
have no SAN, one easy possibility to have it is to use nbd.
Find attached the block-nbd script (in /etc/xen/scripts/) and
an example config file. The block-nbd script is based on the -enbd one.
They have been used successfully in the Xen Tutorial at Linux Kongress.
Signed-off-by: Kurt Garloff <garloff@suse.de>
emellor@leeni.uk.xensource.com [Tue, 25 Oct 2005 10:24:01 +0000 (11:24 +0100)]
Merged.
emellor@leeni.uk.xensource.com [Tue, 25 Oct 2005 10:22:47 +0000 (11:22 +0100)]
Moved SuSE-specific getcfg call into a preiftransfer call. The semantics of
this aren't beautiful, but this gets the distro-specific code out of the body
of the network-bridge script and into a common place, while still preserving
the existing code behaviour.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
djm@kirby.fc.hp.com [Tue, 25 Oct 2005 03:00:35 +0000 (21:00 -0600)]
Oops, missed hg add of vti example file
djm@kirby.fc.hp.com [Tue, 25 Oct 2005 02:57:47 +0000 (20:57 -0600)]
Necessary change to support ia64/vti domain build in common code.
Signed-off-by Ke Yu <ke.yu@intel.com>
Signed-off-by Anthony Xu <anthony.xu@intel.com>
Signed-off-by Kevin Tian <kevin.tian@intel.com>
djm@kirby.fc.hp.com [Tue, 25 Oct 2005 02:56:53 +0000 (20:56 -0600)]
Image builder for ia64/vti domain, together with a example config file.
Signed-off-by Ke Yu <ke.yu@intel.com>
Signed-off-by Anthony Xu <anthony.xu@intel.com>
Signed-off-by Kevin Tian <kevin.tian@intel.com>
chris@ubuntu.eng.hq.xensource.com [Tue, 25 Oct 2005 02:04:30 +0000 (19:04 -0700)]
Don't run network stop script when doing xend stop.
It doesn't make sense to undo the network configuration when stopping
xend since that would cut off all VMs from the network. Undoing the
network setup should be an option independent of xend stop.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 17:41:27 +0000 (11:41 -0600)]
Code reordering for (partial) gcc3.4 compatibility
Signed-off-by Akio Takabe <takebe_akio@jp.fujitsu.com>
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 17:32:45 +0000 (11:32 -0600)]
Remove dup prologue lines
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 17:31:09 +0000 (11:31 -0600)]
Define VHPT constants from VHPT_SIZE_LOG2.
Prototypes added.
emellor@leeni.uk.xensource.com [Mon, 24 Oct 2005 17:24:07 +0000 (18:24 +0100)]
Change argument spec for introduce_domain argument parsing, as the MFN is a long.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 24 Oct 2005 17:21:38 +0000 (18:21 +0100)]
Remove recently added uses of errno, as these cause trouble on ia64 builds.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 16:46:39 +0000 (10:46 -0600)]
Need additional include for ia64
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 15:15:53 +0000 (09:15 -0600)]
Necessary change to make ia64 dom0 rework:
- All even channels on xen/ia64 shares one common interrupt vector
- xen_start_info is initialized only after xen_init, so adjust
sequence to avoid access before initialization
Signed-off-by Kevin Tian <kevin.tian@intel.com>
djm@kirby.fc.hp.com [Mon, 24 Oct 2005 15:08:13 +0000 (09:08 -0600)]
Merge
emellor@leeni.uk.xensource.com [Mon, 24 Oct 2005 14:18:31 +0000 (15:18 +0100)]
Removed some unnecessary imports.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 24 Oct 2005 13:12:42 +0000 (14:12 +0100)]
Added a timestamp to the logging output by trace_io.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Mon, 24 Oct 2005 13:11:07 +0000 (14:11 +0100)]
Do not create the console tty until domain_create_ring has succeeded. This
means that any reads attempted on that tty will not cause handle_tty_read to
segfault looking for an interface that does not exist.
Move the writing of the tty node so that it happens after console/limit has
been read, as it is cleaner if the writing is the last thing that happens.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
vhanquez@kneesa.uk.xensource.com [Mon, 24 Oct 2005 11:57:37 +0000 (12:57 +0100)]
compute the correct totalram_pages when the balloon increase or decrease the
reservation. memory reports should be more correct now.
From: Satoshi Uchida <s-uchida@ap.jp.nec.com>
Signed-off-by: Vincent Hanquez <vincent@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Mon, 24 Oct 2005 07:04:38 +0000 (08:04 +0100)]
Attached patch avoids "Bad L1 flags 80" for VMX domains. Thanks Ian for
the suggestions.
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
emellor@leeni.uk.xensource.com [Sun, 23 Oct 2005 21:45:15 +0000 (22:45 +0100)]
Change the semantics of GetDomainPath so that it always succeeds, regardless of
whether a domain has been introduced to the store. Added a separate message
XS_IS_DOMAIN_INTRODUCED and API for that (xs_is_domain_introduced) to determine
whether the domain has really been introduced. This change means that the
tools can determine the correct domain path earlier in the domain creation
process, which is particularly a factor with live migration, as it allows us
to create the devices earlier in the process, and unpause the new domain before
performing the introduce. Until recently we already had these features, but
the simplification of the interface between xend and xenstored caused breakage.
No longer clear out the domain path when a domain is introduced -- this was a
hack to work around the recent problematic semantics of GetDomainPath.
Do not write the contents of the info block to the store. All the configuration
info is written to the /vm path, and anything else in the info block is either
dealt with explicitly or is ephemeral and has no place in the store.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Sun, 23 Oct 2005 21:34:13 +0000 (22:34 +0100)]
Instead of writing errors to
/local/domain/0/backend/<devclass>/<dom>/<devid>/error, write them instead to
/local/domain/0/error/backend/<yada>. This is not the best place for them
perhaps, but it moves them out of the backend directory, on which the drivers
have a watch. This fixes the problem whereby writing an error will trigger a
watch, causing the error message to be written again, and repeat. Fixes bug
#286.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sun, 23 Oct 2005 15:54:51 +0000 (16:54 +0100)]
Enable TSC-offsetting capability to ensure that VMX
guest's TSC remains in sync with virtual-PIT interrupts.
This avoids the following message from full-virt Linux:
"Losing too many ticks!
TSC cannot be used as a timesource. =20
Possible reasons for this are:
You're running with Speedstep,
......"
Signed-off-by: Eddie Dong <eddie.dong@intel.com>
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
kaf24@firebug.cl.cam.ac.uk [Sun, 23 Oct 2005 15:51:47 +0000 (16:51 +0100)]
Fix an issue for passing arguement from control panel to deivce model
for some arguemnt like 'localtime', 'isa', device model need an argument
"-localtime", instead of "-localtime 1"
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Sun, 23 Oct 2005 10:51:15 +0000 (11:51 +0100)]
some fixes to VMX:
1) enable ACPI in VMXAssist by default.
2) do nothing when emulating wbinvd instruction in VMXAssist.
3) use MACRO in cpu number setting/getting code.
4) remove useless parameter flags from xc_vmx_build.
5) remove usage of dsi.v_start in xc_vmx_build when not handling
vmxloader elf.
6) unmap shared_page_va when destroy VMX domain.
7) change virtual_platform_def to vmx_platform.
8) remove useless code in vmx_setup_platform.
9) change parameter of vmx_setup_platform from vcpu to domain.
10) in xen HV, vmx domain get domain processor number from the reserved
E820 area.
11) in domain.c, don't define some functions when on i386 platform or no
CONFIG_VMX.
Also removed some ugly tailing space.
Signed-off-by: Xin Li <xin.b.li@intel.com>
emellor@leeni.uk.xensource.com [Sat, 22 Oct 2005 10:04:45 +0000 (11:04 +0100)]
Reintroduce the changes made by changeset 7455:
021324804fbd, which were
disabled by workaround 7468:
17a9f111fa93. We additionally need to set the
FD_CLOEXEC flag on the status fd given to SrvServer when spawning the network
script, as at least on some platforms this causes xend to fail to start
properly.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 22 Oct 2005 09:33:26 +0000 (10:33 +0100)]
Fix 'xm vcpu-list'. List cpumap as 'CPU Affinity'. Use
special descriptive strings for empty and full cpu maps.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 22 Oct 2005 07:35:03 +0000 (08:35 +0100)]
cpumap cleanups -- Python now deals with lists of cpus
rather than cpu bitmaps.
Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 22 Oct 2005 06:43:56 +0000 (07:43 +0100)]
Fix assertions in bitops.c.
Signed-off-by: Aravindh Puthiyaparambil
<aravindh.puthiyaparambil@unisys.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 22 Oct 2005 06:37:15 +0000 (07:37 +0100)]
This patch is to make xentop able to print vcpu usage for out of order
enabled vcpus in a domain. This really isn't an issue anymore with the
changes made to dom0 ops and xm/xend.
But, I provide it because:
- Josh noticed my previous patch called getvcpuinfo twice for each vpcu,
which was completely silly.
- this patch would be useful in the remote chance that domain vpcus are
enabled out of order.
Signed-off-by: Daniel Stekloff <dsteklof@us.ibm.com>
kaf24@firebug.cl.cam.ac.uk [Sat, 22 Oct 2005 06:35:36 +0000 (07:35 +0100)]
Two patches were applied to pygrub's setup.py to work around python2.2
limitations and only one is needed. Revert one of them.
Signed-Off-By: Muli Ben-Yehuda <mulix@mulix.org>
djm@kirby.fc.hp.com [Fri, 21 Oct 2005 19:58:39 +0000 (13:58 -0600)]
Backout incautious public/io header changes
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 19:07:54 +0000 (20:07 +0100)]
Disable Xen verbosity by default. At this point I think it
is best to disable 'scary' warning messages by default,
even though this will also quieten down genuine error
paths.
The ideal solution would be to introduce DPRINTK levels
(e.g., info, warning, error) and allow these to be selected
at compile and/or boot time. We could then enable error
messages until things have stabilised some more.
Signed-off-by: Keir Fraser <keir@xensource.com>
jrb44@plym.cl.cam.ac.uk [Fri, 21 Oct 2005 17:51:42 +0000 (18:51 +0100)]
Temporary workaround to xend start problem.
Signed-off-by: James Bulpin <james@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 17:29:26 +0000 (18:29 +0100)]
This is a follow up of PIC device model by Xiaofeng and me to move to
hypervisor using virtual weired interrupt line. With this patch the
performance of CPU2K improves 7%, kernel kuild improves 14% and
cyclesoak improves 12%. It is quite amazing!
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
Signed-off-by: Eddie Dong <eddie.dong@intel.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 17:19:38 +0000 (18:19 +0100)]
Tweak to pygrub build config detection.
Signed-off-by: Jeremy Katz <katzj@redhat.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 15:22:09 +0000 (16:22 +0100)]
Fix allocation of phys_to_machine_mapping array during
x86/64 linux boot. The original builder-provided array
must be made read-only in the initial mapping.
Signed-off-by: Keir Fraser <keir@xensource.com>
jrb44@plym.cl.cam.ac.uk [Fri, 21 Oct 2005 14:42:26 +0000 (15:42 +0100)]
Add biossums dependency to Makefile to avoid parallel build breakage.
Signed-off-by: James Bulpin <james@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 14:34:17 +0000 (15:34 +0100)]
Merged.
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 14:31:51 +0000 (15:31 +0100)]
Explicitly delete xc instance, which will hopefully solve the intermittent
failure to start up.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 13:33:46 +0000 (14:33 +0100)]
Fix early bootstrap of secondary VCPUs on x86/64.
Signed-off-by: Keir Fraser <keir@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 12:30:31 +0000 (13:30 +0100)]
Break proc paths out into separate include file, rather than inlining them.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 11:07:14 +0000 (12:07 +0100)]
Merged.
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 11:06:17 +0000 (12:06 +0100)]
Break common network code in xen-network-common.sh, and common vif hotplugging
code into vif-common.sh.
Fix the antispoofing in network-bridge to allow forwarding from physical
device to bridge, and vif0.0 to eth0.
Change the interface between netback driver and vif hotplugging scripts. The
mac address and bridge details are read from the store directly by the script,
so that the kernel driver does not need to get involved.
Remove an awful lot of cruft from vif-bridge. Move the iptables handling into
vif-common.sh, and fix it so that the new vif can get past the antispoofing
entries. Merge vif-nat and vif-route into this structure too.
Use ip instead of ifconfig to get necessary details.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 11:04:30 +0000 (12:04 +0100)]
Tidy xend-config.sxp, removing entries that haven't been used since the
hotplugging stuff was introduced (block-*, console-port-base, console-address)
and introducing entries for options that have been present for ages
(xend-{http,unix,relocation}-server, xend-unix-path, xend-relocation-address,
enable-dump). Remove vif-antispoof, as Vifctl no longer passes this option
down.
Made the xend-relocation-server default to no, as this is a potential security
hole and so must be turned on explicitly.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 10:29:51 +0000 (11:29 +0100)]
Work around internal error in some versions of x86/64 gcc
by manually hoisting gdt address calculation outside loop.
Signed-off-by: Keir Fraser <keir@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 10:22:48 +0000 (11:22 +0100)]
Removed xen.util.process, as no-one is using it, and it leaves defunct processes
behind.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 10:21:05 +0000 (11:21 +0100)]
Remove unused Vifctl.vifctl and Vifctl.set_vif_name. Remove the bridge and
antispoof parameters from Vifctl.network -- these are already handled by the
script, so it is redundant to have a separate override for these parameters
coming from outside. Don't use xen.util.process.runscript, because we don't
need to return the output, and therefore can just use os.spawnl.
Remove XendRoot.get_vif_antispoof, matching change above.
Prefix XendRoot.network_script_dir onto the return value for
XendRoot.get_network_script(), removing that burden from callers.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Fri, 21 Oct 2005 10:08:48 +0000 (11:08 +0100)]
Output the Xen changeset in the logs at start of day.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 09:46:30 +0000 (10:46 +0100)]
Upgrade tree to 2.6.12.6.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 09:24:35 +0000 (10:24 +0100)]
Update one file from qemu 0.71 for Cirrus Logic VGA device model.
Using the older device model running pinball game on windows with
full screen can cause guest crash.
Signed-off-by: yunhong.jiang@intel.com
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 09:20:38 +0000 (10:20 +0100)]
Send special keys that your graphical user interface intercepts at
low level, such as ctrl-alt-del in X Window.
Adds a command "sendkey" to deal with it in the monitor, you
can send keys to the vm. Use - to press several keys simultaneously.
Example:
sendkey ctrl-alt-f1
Signed-off-by: Dan Xu <dan.d.xu@intel.com>
Signed-off-by: Zhai, Edwin <edwin.zhai@intel.com>
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Fri, 21 Oct 2005 09:17:29 +0000 (10:17 +0100)]
Turn off some debug switches by default
These macros cause extra prints in /var/log/xend-debug, and
affect the performance.
Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 20 Oct 2005 20:37:15 +0000 (21:37 +0100)]
This patch to the Xen access control module (ACM) and tools:
1. adapts ACM hooks to the slightly changed event channel structure
2. introduces an ACM_GETDECISION command, which enables authorized
domains to retrieve policy decisions regarding the sharing of resources
(STE policy) from the Xen hypervisor
3. includes cleanup (warnings I found when applying analysis tools such
as beam or flawfinder to the ACM code)
The get_decision function is useful to enforce:
* the security policy on network traffic in the network backends in
domain 0; currently there is no enforcement in Dom0 and all packets flow
freely
* the security policy in block device backends to control which
domains can access which vdisk resources
I have added a small test program that shows how to use the get_decision
ACM interface call, it is in tools/security/get_decision.c and will be
compiled together with the policy tools. As usual, the ACM is unconfigured
until you switch on a security policy on in Config.mk.
Signed-off: Reiner Sailer <sailer@us.ibm.com>
jrb44@plym.cl.cam.ac.uk [Thu, 20 Oct 2005 18:37:41 +0000 (19:37 +0100)]
Avoid specifying bogus DISPLAY=None when spawning device model.
Signed-off-by: James Bulpin <james@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 20 Oct 2005 13:06:07 +0000 (14:06 +0100)]
Set antispoof default to yes, as this seems safer, and is the current behaviour
in VifCtl.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
emellor@leeni.uk.xensource.com [Thu, 20 Oct 2005 11:35:57 +0000 (12:35 +0100)]
Merged.
emellor@leeni.uk.xensource.com [Thu, 20 Oct 2005 11:34:50 +0000 (12:34 +0100)]
Fix the recently-introduced failure to preserve uuid across a guest reboot.
Store the uuid in string form inside XendDomainInfo, to aid this.
Add restart-squelching logic to ensure that runaway domains to restart
indefinitely. The logic here could be cleverer, say including a backoff for
retrying, in case a network block device has gone away, but for now this will
do to stop the pathological behaviour. Fixes bug #275.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 20 Oct 2005 11:10:14 +0000 (12:10 +0100)]
New function direct_kernel_remap_pfn_range(). Like
direct_remap_pfn_range but doesn't take a vma and implicitly
maps into init_mm.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 20 Oct 2005 10:52:27 +0000 (11:52 +0100)]
disable bogus touchpad device model, which cause annoying dmesg on 2.6 kernel
Signed-off-by: Edwin Zhai <edwin.zhai@intel.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 20 Oct 2005 10:27:51 +0000 (11:27 +0100)]
smp_call_function() must be usable as an expression
return an integer.
Signed-off-by: Keir Fraser <keir@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Thu, 20 Oct 2005 10:25:55 +0000 (11:25 +0100)]
Fix NMI race with context switch.
The machine used to auto reboot if an NMI was received in a critical
time window when context switching domains. There is a small time
window when the GDT may become unmapped (after CR3 is updated and
before setting GDTR with the new GDT during a domain context
switch. If an NMI is received during this time window a triple fault
is triggered causing the machine to auto reboot.
Bug found and original patch proposed by Jose Renato Santos
<jsantos@hpl.hp.com>.
Signed-off-by: Keir Fraser <keir@xensource.com>
tristan.gingold@bull.net [Thu, 20 Oct 2005 10:01:11 +0000 (06:01 -0400)]
Makes smp_call_function a real stmt if CONFIG_SMP is not defined.
This avoid a GCC warning.
emellor@leeni.uk.xensource.com [Wed, 19 Oct 2005 17:18:30 +0000 (18:18 +0100)]
This patch fixes the problem that VMX guest can not get brought up on
current tip (changeset 7425).
Signed-off-by: Xin Li <xin.b.li@intel.com>
emellor@leeni.uk.xensource.com [Wed, 19 Oct 2005 15:25:39 +0000 (16:25 +0100)]
Merged.
emellor@leeni.uk.xensource.com [Wed, 19 Oct 2005 15:24:54 +0000 (16:24 +0100)]
New network-bridge script and associated gubbins.
This is Kurt Garloff's reworked network-bridge script:
* we got rid of ifconfig
* it works for netdev != eth0
* arp on and off are symmetric as are ifdown and ifup
* ifup will be passed the ifcfg config file name if needed
(the ifup may otherwise figure that the veth0 hardware is
NOT the same as the original ${netdev} and not use the same
config -- this happens on SUSE. Charles Coffing tracked this
one down.)
Plus Kurt's avoid-dash patch:
the network setup scripts on SUSE have trouble with the bridge
name xen-br0; they don't expect the '-'.
Arguably this should be fixed.
But I assume there's more scripts out there which may not like it,
so I suggest the following patch to rename xen-br0 to xenbr0.
Plus Charles Duffy's patch to support multiple bridges:
The attached patch allows the network-bridge script to be used to
generate multiple bridges corresponding to different physical
interfaces. It adds a new parameter, "vifnum", used to refer both to
the loopback interface to be used and to set defaults regarding the
physical interface and bridge name.
Thus, if one wishes to start xenbr0 on eth0 and xenbr1 on eth1, one
need only call:
network-bridge start ## vifnum is 0 by default
network-bridge start vifnum=1
...well, that and set loopback.nloopbacks=2 in the Dom0 kernel
parameters.
Plus renaming of virtnum to vifnum in Charles' patch, as requested by Ian Pratt.
Plus a fix to DevController to allocate vif IDs starting from 0 (i.e. vif2.0
is now domain 2's first vif, as opposed to vif2.1 in the recent past).
Plus tidying up inside network-bridge using some helper variables.
Signed-off-by: Ewan Mellor <ewan@xensource.com>
kaf24@firebug.cl.cam.ac.uk [Wed, 19 Oct 2005 15:07:11 +0000 (16:07 +0100)]
pygrub's setup.py relies on distutils.UnixCCompiler.has_function(),
which does not exist with python2.2, causing the following build
error:
make[2]: Entering directory `/home/muli/xen/x86.hg/tools/pygrub'
CFLAGS=" -m32 -march=i686" python setup.py build
Traceback (most recent call last):
File "setup.py", line 15, in ?
if cc.has_function("ext2fs_open2"):
AttributeError: UnixCCompiler instance has no attribute 'has_function'
The following patch gets it to build, but is pretty ugly. A proper fix
would be to do the check for ext2fs_open2() in a way that is backward
compatible with python2.2.
Signed-Off-By: Muli Ben-Yehuda <mulix@mulix.org>
kaf24@firebug.cl.cam.ac.uk [Wed, 19 Oct 2005 14:51:14 +0000 (15:51 +0100)]
Declare ac_timer_init() properly in ac_timer.h
emellor@leeni.uk.xensource.com [Wed, 19 Oct 2005 13:00:37 +0000 (14:00 +0100)]
Merge.
emellor@leeni.uk.xensource.com [Wed, 19 Oct 2005 12:59:22 +0000 (13:59 +0100)]
Added diagnosis tool for broken domains.
Signed-off-by: Ewan Mellor <ewan@xensource.com>