xen.git
19 years agoPV-on-HVM: TSO support prior to 2.6.9 appears to be unreliable,
Ian Campbell [Fri, 10 Nov 2006 12:34:28 +0000 (12:34 +0000)]
PV-on-HVM: TSO support prior to 2.6.9 appears to be unreliable,
especially when firewalling is enabled.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years agoPV-on-HVM: Do not enable barriers in blkfront before 2.6.16
Ian Campbell [Fri, 10 Nov 2006 12:34:26 +0000 (12:34 +0000)]
PV-on-HVM: Do not enable barriers in blkfront before 2.6.16

Barrier support was rewritten after 2.6.15. Perhaps it would be
possible to implement support for old kernels as well but for now it
is easier to just disable them.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[LIBFSIMAGE] Don't leak FDs or RAM when closing filesystems
Tim Deegan [Fri, 10 Nov 2006 12:27:56 +0000 (12:27 +0000)]
[LIBFSIMAGE] Don't leak FDs or RAM when closing filesystems
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years ago[HVM] Remove HVM halt timer. It's no longer needed since interrupts
kfraser@localhost.localdomain [Fri, 10 Nov 2006 11:01:15 +0000 (11:01 +0000)]
[HVM] Remove HVM halt timer. It's no longer needed since interrupts
can wake it up now.

Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[HVM] Remove a duplicated range checking in vioapic_set_irq.
kfraser@localhost.localdomain [Fri, 10 Nov 2006 10:31:29 +0000 (10:31 +0000)]
[HVM] Remove a duplicated range checking in vioapic_set_irq.
Since irq range checking has already been done at the
beginning of this function.

Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[HVM] Clarify ioreq interactions between Xen and qemu-dm.
kfraser@localhost.localdomain [Fri, 10 Nov 2006 10:30:38 +0000 (10:30 +0000)]
[HVM] Clarify ioreq interactions between Xen and qemu-dm.

Mainly this involves placing appropriate barriers before
updating the shared state variable, and after reading from it. The
model is that it is state switches that drive the transfer of data to
and fro in the shared ioreq structure. So writers should wmb() before
updating the variable; readers should rmb() after seeing a state
change.

These barriers aren't actually required given the current code
structure since the communication flow is really driven by
event-channel notifications, which happen to provide a sufficient
barrier. However, relying on this for all time and on all
architectures seems foolish and the barriers have negligible cost
compared with the totoal ioreq round-trip latency. Also the model of
communications being driven by the shared-memory state variable more
closely matches other inter-domain protocols (where there is usually a
shared-memory producer index), and is hence a model we are familiar
with and likely to implement correctly.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LIBFSIMAGE] Further 64bit compile fix
Tim Deegan [Fri, 10 Nov 2006 09:50:35 +0000 (09:50 +0000)]
[LIBFSIMAGE] Further 64bit compile fix
64bit libext2fs can have __u64 defined as unsigned long long,
while a uint64_t is unsigned long, and GCC complains.

Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years ago[LIBFSIMAGE] Fix printf-format warning on x86/64.
kaf24@firebug.cl.cam.ac.uk [Fri, 10 Nov 2006 07:48:25 +0000 (07:48 +0000)]
[LIBFSIMAGE] Fix printf-format warning on x86/64.
Signed-off-by: Ben Thomas <ben@virtualiron.com>
19 years agoServer-side capitalisation fixes.
Ewan Mellor [Thu, 9 Nov 2006 15:43:24 +0000 (15:43 +0000)]
Server-side capitalisation fixes.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoThe attached patch implements more of the Xen-API functionality related
Ewan Mellor [Thu, 9 Nov 2006 15:27:32 +0000 (15:27 +0000)]
The attached patch implements more of the Xen-API functionality related
to the vTPM.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
19 years agoRename abstract_value.double_val -> float_val. All the other names match the
Ewan Mellor [Thu, 9 Nov 2006 11:40:28 +0000 (11:40 +0000)]
Rename abstract_value.double_val -> float_val.  All the other names match the
abstract type names in the doc, so this one should too.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoAdded a destroy method for each object. Based upon the code for VM.destroy
Ewan Mellor [Thu, 9 Nov 2006 11:37:35 +0000 (11:37 +0000)]
Added a destroy method for each object.  Based upon the code for VM.destroy
by Jim Fehlig.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoMore capitalisation fixes -- the on-wire method name should match the
Ewan Mellor [Thu, 9 Nov 2006 11:23:09 +0000 (11:23 +0000)]
More capitalisation fixes -- the on-wire method name should match the
documentation in terms of capitalisation.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoMinor tidying.
Ewan Mellor [Thu, 9 Nov 2006 11:11:38 +0000 (11:11 +0000)]
Minor tidying.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoReplace uses of the typename (e.g. xen_host) where its used as a variable name.
Ewan Mellor [Thu, 9 Nov 2006 11:02:48 +0000 (11:02 +0000)]
Replace uses of the typename (e.g. xen_host) where its used as a variable name.
This is an unusual style, and out of sync with the rest of the library.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[HVM] Change vlapic printk()s into appropriate gdprintk()s.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 18:14:53 +0000 (18:14 +0000)]
[HVM] Change vlapic printk()s into appropriate gdprintk()s.
Simplify apic_round_robin().
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Clean up two small APIC TDCR issues.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 17:53:05 +0000 (17:53 +0000)]
[HVM] Clean up two small APIC TDCR issues.

First, only save the bits that can be set, so that on a subsequent
read can't see MBZ bits as set. Next, be sure to initialize the
timer_divide_count, which is derived from the TDCR.  This avoids a
potential divide-by-zero elsewhere in the code.

The original patch was provided by Ben Thomas <ben@virtualiron.com>.
This modified version refactors the code slightly and renames
timer_divide_count to timer_divisor.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoAdd net-gso-5-rcv-mss.patch to series file. This was missed in
Ian Campbell [Thu, 9 Nov 2006 17:28:12 +0000 (17:28 +0000)]
Add net-gso-5-rcv-mss.patch to series file. This was missed in
11845:6677b97612a2

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
19 years ago[HVM] Clean up IOREQ state managemnet and evtchn notifications.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 17:23:58 +0000 (17:23 +0000)]
[HVM] Clean up IOREQ state managemnet and evtchn notifications.
Based on a patch from Xin Li <xin.b.li@intel.com>. Changed
significantly on the Xen side -- not as cut down as the original
patch: this one keeps the xen_event_channel 'API' unchanged.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Missed interface update belonging to previous changeset.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 16:50:44 +0000 (16:50 +0000)]
[XEN] Missed interface update belonging to previous changeset.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[BLK] Add support for barrier writes to blkfront/blkback.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 16:49:59 +0000 (16:49 +0000)]
[BLK] Add support for barrier writes to blkfront/blkback.

Protocol changes:
 * There is a new operation (BLKIF_OP_WRITE_BARRIER)
   to pass on barrier requests.
 * There is a new state (BLKIF_RSP_EOPNOTSUPP) to indicate
   unsupported operations (barrier writes may fail depending
   on the underlying block device).
 * A new xenstore node named "feature-barrier" indicates the
   backend is able to handle barrier writes.  The value can
   be 1 (all is fine) or 0 (underlying block device doesn't
   support barriers).

blkback changes:  Add "feature-barrier" node to indicate barrier
support, pass incoming barrier requests to the block layer using
submit_bio(WRITE_BARRIER, bio).  Some error handling fixes to
properly pass through barrier write failures, so the frontend
can turn off barriers then.

blkfront changes:  Check if the backend sets "feature-barrier", if
present switch to QUEUE_ORDERED_DRAIN mode.  Send off barrier
requests to the backend using the new BLKIF_OP_WRITE_BARRIER
operation.  Also some error handling for the EOPNOTSUPP case.

Background:  Barriers are needed to make journaling filesystems work
reliable.  For some requests they need order constrains to make the
transactions work correctly.  Barrier requests are used to pass that
ordering information to the block layer and/or to the device, so the
constrains are obeyed when reordering requests for better write
performance.

Signed-off-by: Gerd Hoffmann <kraxel@suse.de>
19 years agoMerge
Alastair Tse [Thu, 9 Nov 2006 16:20:58 +0000 (16:20 +0000)]
Merge

19 years ago[XM-TEST] Relax craete_mem*_pos checks to allow a 2MB less than
Alastair Tse [Thu, 9 Nov 2006 16:20:26 +0000 (16:20 +0000)]
[XM-TEST] Relax craete_mem*_pos checks to allow a 2MB less than
expected.

Now that xm list outputs the real system memory allocation, this can
be 1-2MB less than expected.

Signed-off-by: Alastair Tse <atse@xensource.com>
19 years agoMerge
Tim Deegan [Thu, 9 Nov 2006 16:06:43 +0000 (16:06 +0000)]
Merge

19 years ago[XEN] Console keystroke to reset shadows.
Tim Deegan [Thu, 9 Nov 2006 16:05:23 +0000 (16:05 +0000)]
[XEN] Console keystroke to reset shadows.
In debug builds, adds a keyhandler that blows away all shadows, to
force them to be rebuilt.
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years ago[XEND] Ignore dying domains when refreshing list of domains.
Alastair Tse [Thu, 9 Nov 2006 16:05:00 +0000 (16:05 +0000)]
[XEND] Ignore dying domains when refreshing list of domains.

Also cleanup stored configuration if managed domain is deleted.

Signed-off-by: Alastair Tse <atse@xensource.com>
19 years ago[HVM] Clarify the HVM e820 map. Also make it rather more
kfraser@localhost.localdomain [Thu, 9 Nov 2006 15:38:15 +0000 (15:38 +0000)]
[HVM] Clarify the HVM e820 map. Also make it rather more
conservative wrt the ACPI region. hvmloader should fix up
by freeing unused memory in the range 0xE0000-0x100000.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[PYGRUB] Fix a few bounds violations in pygrub's command-line editor.
Tim Deegan [Thu, 9 Nov 2006 14:23:24 +0000 (14:23 +0000)]
[PYGRUB] Fix a few bounds violations in pygrub's command-line editor.
Don't step off the end of the line, or delete characters that aren't there.

Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
Signed-off-by: Tim Deegan <Tim.Deegan@xensource.com>
19 years agoUse libfsimage for reading filesystem images.
Tim Deegan [Thu, 9 Nov 2006 14:11:11 +0000 (14:11 +0000)]
Use libfsimage for reading filesystem images.

Signed-off-by: John Levon <john.levon@sun.com>
19 years agoAdd libfsimage, a C library for reading files from filesystem images.
Tim Deegan [Thu, 9 Nov 2006 14:09:53 +0000 (14:09 +0000)]
Add libfsimage, a C library for reading files from filesystem images.
Initial support is provided for Solaris UFS, ext2 (both using libext2fs
and not), and reiserfs.

Signed-off-by: John Levon <john.levon@sun.com>
19 years ago[VMXASSIST] Fix mishandling of segment registers on entry into VM86_PROTECTED mode.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 14:00:40 +0000 (14:00 +0000)]
[VMXASSIST] Fix mishandling of segment registers on entry into VM86_PROTECTED mode.

The bug occurs when the guest is entering protected mode and reaches
the far jump to set up the 32-bit segment registers for the virtual
VMENTER we're about to perform. vmxassist, in protected_mode(), looks
at the segment registers, which might be 16-bit segments or might be
32-bit segment selectors at this stage (depending on whether the OS
has reloaded them since entering protected mode); and it tries to load
the segments from the GDT.  Unconditionally.  Even if the segment
register still actually contains a real mode 16-bit segment.  Whoops.

vmxassist already has code to detect 16-bit segments that survived
unmodified from a transition into and out of protected mode, and to
save and restore those appropriately.  It does this using
"saved_rm_regs", which get cleared on entry to protected mode, and
then set to the old segment value if we fail to set a given 32-bit
segment correctly.

The fix is to save the 16-bit segments *always*, on entry to protected
mode when %CR0(PE) is first set; and to clear the saved 16-bit segment
and set the 32-bit variant in oldctx whenever a 32-bit segment
descriptor is set during the transition to 32-bit CS.  Then, when we
finally do the VMENTER, we will set up the VMCS from only the 32-bit
segments, clearing the VMCS entries for segments that have not been
assigned valid 32-bit segments yet.

Tested on various RHEL-5 boot.isos, including ones which worked before
and ones which triggered the bug; all now boot correctly.

Signed-off-by: Stephen Tweedie <sct@redhat.com>
19 years ago[BLKTAP] Fix hotplug script to check if specified file exists.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 13:43:21 +0000 (13:43 +0000)]
[BLKTAP] Fix hotplug script to check if specified file exists.

Now users will see the following error:

[root@bebop ~]# xm create -c 128
Using config file "/etc/xen/128".
Error: Device 51713 (tap) could not be connected.
/tmp/128.img does not exist

Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
19 years agoAdd lost record format in xentrace formats file.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 13:30:16 +0000 (13:30 +0000)]
Add lost record format in xentrace formats file.
Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[IA64][VMX] Fix 3G memory limit for VTi domain.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 13:28:45 +0000 (13:28 +0000)]
[IA64][VMX] Fix 3G memory limit for VTi domain.
Signed-off-by : Zhang xiantao <xiantao.zhang@intel.com>

19 years ago[XENSTORE] Fix xenstore-ls program to connect via kernel xenbus
kfraser@localhost.localdomain [Thu, 9 Nov 2006 13:25:47 +0000 (13:25 +0000)]
[XENSTORE] Fix xenstore-ls program to connect via kernel xenbus
connection by default, like all other client utils.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoMerge
Alastair Tse [Thu, 9 Nov 2006 13:18:58 +0000 (13:18 +0000)]
Merge

19 years ago[XEND] Fix bug with managed domains not having their state updated.
Alastair Tse [Thu, 9 Nov 2006 13:16:55 +0000 (13:16 +0000)]
[XEND] Fix bug with managed domains not having their state updated.

Also added some protection against corrupted sxp configuration files.

Signed-off-by: Alastair Tse <atse@xensource.com>
19 years ago[XEND] Fix domain_lookup_nr to check managed_domains too.
Alastair Tse [Thu, 9 Nov 2006 12:38:12 +0000 (12:38 +0000)]
[XEND] Fix domain_lookup_nr to check managed_domains too.

Bug introduced when moving managed domains out of self.domains caused
domain_lookup_nr ignoring managed domains.

Signed-off-by: Alastair Tse <atse@xensource.com>
19 years ago[XEND][XM] Move system memory reporting to XendConfig rather than getSysMem.
Alastair Tse [Thu, 9 Nov 2006 12:36:20 +0000 (12:36 +0000)]
[XEND][XM] Move system memory reporting to XendConfig rather than getSysMem.

Remove the additional XMLRPC calls per domain just to get system
memory and instead make sure we update that value and return it via
the SXP that gets returned when xm list is called.

Signed-off-by: Alastair Tse <atse@xensource.com>
19 years ago[BLK] Expand number of fake-SCSI VBD volumes supported to 120.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 12:07:28 +0000 (12:07 +0000)]
[BLK] Expand number of fake-SCSI VBD volumes supported to 120.

Signed-off-by: Takanori Kasai <Kasai.Takanori@jp.fujitsu.com>
Signed-off-by: Tomonari Horikoshi <t.horikoshi@jp.fujitsu.com>
Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
19 years ago[XENOPROF] Oprofile user level samples for passive domains are being lost.
kfraser@localhost.localdomain [Thu, 9 Nov 2006 11:47:42 +0000 (11:47 +0000)]
[XENOPROF] Oprofile user level samples for passive domains are being lost.
The number of of lost samples is most significant when dom0 is idle.

From: joserenato.santos@hp.com
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoWrite the /local/domain/<domid>/vm node early in the startup process
Ewan Mellor [Wed, 8 Nov 2006 18:29:16 +0000 (18:29 +0000)]
Write the /local/domain/<domid>/vm node early in the startup process
(immediately after the domain is created).  QEMU needs this in place when it
starts up, in order to find the VNC password.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoIgnore dying domains on recreate, and log the fact.
Ewan Mellor [Wed, 8 Nov 2006 18:27:31 +0000 (18:27 +0000)]
Ignore dying domains on recreate, and log the fact.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoAdded test_bindings to the ignore list.
Ewan Mellor [Wed, 8 Nov 2006 15:26:22 +0000 (15:26 +0000)]
Added test_bindings to the ignore list.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[HVM] Avoid long delay between setting IRQ for PV drivers and
kaf24@firebug.cl.cam.ac.uk [Thu, 9 Nov 2006 09:04:41 +0000 (09:04 +0000)]
[HVM] Avoid long delay between setting IRQ for PV drivers and
servicing it. Should call service_ioapic() synchronously.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Reserve 2 pages for ACPI tables in e820.
kaf24@firebug.cl.cam.ac.uk [Wed, 8 Nov 2006 19:48:46 +0000 (19:48 +0000)]
[HVM] Reserve 2 pages for ACPI tables in e820.
This is required when appending SSDTs to the core
ACPI data.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Do not print grant-table warning if a page cannot be pinned
kfraser@localhost.localdomain [Wed, 8 Nov 2006 17:31:25 +0000 (17:31 +0000)]
[XEN] Do not print grant-table warning if a page cannot be pinned
because the (previous) owner is dying. This quietens down domain
destruction quite a bit.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[VMX] __vmread() and __vmwrite() now BUG on failure, rather than
kfraser@localhost.localdomain [Wed, 8 Nov 2006 16:43:50 +0000 (16:43 +0000)]
[VMX] __vmread() and __vmwrite() now BUG on failure, rather than
returning the error to the caller.

None of our vmcs reads/writes should fail. In fact many callers were
ignoring failures, which can hide bugs (in fact it hid one in
vmx_restore_dr(), which was incorrectly using __vmread() before
the vmcs was loaded).

__vmread_safe() is used by the vmcs dump function, which is the only
function which can legitimately fail a vmcs access.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Simplify relationship between VIOAPIC and VLAPICs.
kfraser@localhost.localdomain [Wed, 8 Nov 2006 15:11:18 +0000 (15:11 +0000)]
[HVM] Simplify relationship between VIOAPIC and VLAPICs.
It's not really dynamic since there is always exactly
one VLAPIC per VCPU.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[VMX] Check INTR_TYPE is NMI before doing physical NMI processing.
kfraser@localhost.localdomain [Wed, 8 Nov 2006 15:10:21 +0000 (15:10 +0000)]
[VMX] Check INTR_TYPE is NMI before doing physical NMI processing.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoAdd support for VM.destroy RPC in libxen.
jfehlig@jfehlig2.provo.novell.com [Wed, 8 Nov 2006 01:14:16 +0000 (18:14 -0700)]
Add support for VM.destroy RPC in libxen.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
19 years agoReplace uses of [ ... == ... ] with the POSIX-legal single-equals. Tidy up
Ewan Mellor [Wed, 8 Nov 2006 13:22:08 +0000 (13:22 +0000)]
Replace uses of [ ... == ... ] with the POSIX-legal single-equals.  Tidy up
external-device-migrate script in its handling of parameters.

Patch from Roland Illig <roland.illig@gmx.de>.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoDon't complain about name non-uniqueness if one domain is dying. This happens
Ewan Mellor [Wed, 8 Nov 2006 12:56:45 +0000 (12:56 +0000)]
Don't complain about name non-uniqueness if one domain is dying.  This happens
across a reboot regularly.  Inside _refresh, there is the occasion for new
domains to appear, if an existing domain reboots itself.  In this case, we
need to get the list of running domains from Xen again.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agocpu_weight is now an int, not a float.
Ewan Mellor [Wed, 8 Nov 2006 12:26:34 +0000 (12:26 +0000)]
cpu_weight is now an int, not a float.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoAdd a shutdownWatch field in the HVMImageHandler constructor, to avoid error
Ewan Mellor [Wed, 8 Nov 2006 12:25:56 +0000 (12:25 +0000)]
Add a shutdownWatch field in the HVMImageHandler constructor, to avoid error
messages if creation fails and the watch gets torn down.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoRefresh the domain list on domain_start or domain_create. We need to do this
Ewan Mellor [Wed, 8 Nov 2006 11:13:50 +0000 (11:13 +0000)]
Refresh the domain list on domain_start or domain_create.  We need to do this
so that the list of running domains is up-to-date for the name-uniqueness
check.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoSave the old domid when restarting, as it may get wiped out before we need to
Ewan Mellor [Wed, 8 Nov 2006 11:08:18 +0000 (11:08 +0000)]
Save the old domid when restarting, as it may get wiped out before we need to
log it (on restart failure).

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years ago[HVM] Remove bogus free of vlapic structure.
kfraser@localhost.localdomain [Wed, 8 Nov 2006 10:48:55 +0000 (10:48 +0000)]
[HVM] Remove bogus free of vlapic structure.
It is incorrect now that vlapic is embedded inside
the vcpu structure.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Add debug mode to xmalloc().
kfraser@localhost.localdomain [Wed, 8 Nov 2006 10:48:21 +0000 (10:48 +0000)]
[XEN] Add debug mode to xmalloc().

This detects programming mistakes such as:
 * double free
 * use after free
 * write beyond bounds of allocation

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years agoThis patch addresses bugzilla #649, enabling `xm list' to display the
Ewan Mellor [Wed, 8 Nov 2006 09:50:09 +0000 (09:50 +0000)]
This patch addresses bugzilla #649, enabling `xm list' to display the
actual amount of memory allocated by a guest.  Also this patch:
  -  makes it more apparent that 'xm mem-set' has no effect on HVM's.
  -  show's that memory is changing if 'xm list' is issued while guest is
ballooning (w/o touching balloon driver)

Signed-off-by: Tommie McAfee <tommie.mcafee@unisys.com>
19 years agoAdd the Japanese keymap for VNC Server.
kasai.takanori@jp.fujitsu.com [Wed, 8 Nov 2006 09:38:06 +0000 (09:38 +0000)]
Add the Japanese keymap for VNC Server.

Signed-off-by: Takanori Kasai < kasai.takanori@jp.fujitsu.com >
19 years agofix xm-test configure.ac to not define RD_PATH which prevented env override
Ewan Mellor [Thu, 2 Nov 2006 23:05:24 +0000 (23:05 +0000)]
fix xm-test configure.ac to not define RD_PATH which prevented env override

Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
19 years agoUpdate README.
Ewan Mellor [Wed, 1 Nov 2006 14:27:54 +0000 (14:27 +0000)]
Update README.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoFix comment.
Ewan Mellor [Wed, 1 Nov 2006 14:27:29 +0000 (14:27 +0000)]
Fix comment.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoAdd helper script for releasing xm-test binary ISOs.
Ewan Mellor [Wed, 1 Nov 2006 14:27:19 +0000 (14:27 +0000)]
Add helper script for releasing xm-test binary ISOs.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoBump xm-test version to 1.1 to accommodate new devices.
Ewan Mellor [Wed, 1 Nov 2006 08:42:02 +0000 (08:42 +0000)]
Bump xm-test version to 1.1 to accommodate new devices.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
19 years agoAllow the i386 ramdisk to be used on x86_64. Part of a patch from Yu Ping Y.
Ewan Mellor [Wed, 1 Nov 2006 09:21:27 +0000 (09:21 +0000)]
Allow the i386 ramdisk to be used on x86_64.  Part of a patch from Yu Ping Y.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
19 years agoCorrect copy-n-paste error.
Ewan Mellor [Wed, 1 Nov 2006 08:43:54 +0000 (08:43 +0000)]
Correct copy-n-paste error.

Also Whitespace cleanups.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
19 years agoAdd xvd devices to the initrd.
Ewan Mellor [Wed, 1 Nov 2006 08:41:24 +0000 (08:41 +0000)]
Add xvd devices to the initrd.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
19 years agoUpdate all block tests to use the Xen Virtual Block Device (xvd) rather than IDEor...
Ewan Mellor [Wed, 1 Nov 2006 08:39:45 +0000 (08:39 +0000)]
Update all block tests to use the Xen Virtual Block Device (xvd) rather than IDEor SCSI.

This patch also modifies
09_block_attach_and_dettach_device_check_data_pos.py to force / to be
read-write otherwise the mkdir fails.

Also Whistspace cleanups.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
19 years agoThe XenDomain class does not respect the arch defaults where
Ewan Mellor [Wed, 1 Nov 2006 08:38:22 +0000 (08:38 +0000)]
The XenDomain class does not respect the arch defaults where
as XmTestDomain does.  This part of the patch forces both HVM and
non-HVM tests to use XmTestDomain.  With appropriate changes to the
extraConfig arg.

On PPC we need to use the arch defaults to get the console setup
correctly.

Signed-off-by: Tony Breeds <tony@bakeyournoodle.com>
19 years agoThis fixes another missing part for x86-64 support in the xm testsuite.
Ewan Mellor [Wed, 1 Nov 2006 08:34:55 +0000 (08:34 +0000)]
This fixes another missing part for x86-64 support in the xm testsuite.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
19 years ago[XEN] Put the non-GPL license text in each public header file.
kaf24@localhost.localdomain [Wed, 8 Nov 2006 00:03:11 +0000 (00:03 +0000)]
[XEN] Put the non-GPL license text in each public header file.

This excludes arch-powerpc.h and xencomm.h which require
confirmation from the original authors that the existing GPL
text can be removed and replaced.

Original patch from: Mark Johnson <mark.johnson@sun.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] Fix 64-bit build.
kaf24@localhost.localdomain [Tue, 7 Nov 2006 23:14:09 +0000 (23:14 +0000)]
[XEN] Fix 64-bit build.

This required fiddling the asm constraints of the atomic bitops. It
seems gcc isn't entirely happy with "+m": the manual says that the
'+' modifier should be used only when a register constraint is
available.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] License clarification.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 18:13:12 +0000 (18:13 +0000)]
[XEN] License clarification.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[SVM] Fix an interrupt race window in the do_launch/vmrun/vmexit loop for AMD-V.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 17:48:18 +0000 (17:48 +0000)]
[SVM] Fix an interrupt race window in the do_launch/vmrun/vmexit loop for AMD-V.

There is also some comment cleanup in this patch.

This problem found by Virtual Iron (Dave Winchell), and patch also
provided by VI.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
Signed-off-by: Dave Winchell <dwinchell@virtualiron.com>
19 years ago[HVM] VIOAPIC and VPIC cleanups.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 17:46:40 +0000 (17:46 +0000)]
[HVM] VIOAPIC and VPIC cleanups.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] VLAPIC cleanups. vlapic tsruct is statically contained
kfraser@localhost.localdomain [Tue, 7 Nov 2006 15:48:10 +0000 (15:48 +0000)]
[HVM] VLAPIC cleanups. vlapic tsruct is statically contained
inside the vcpu structure.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM] Replace shared-memory PIC state with a set-irq-level hypercall.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 13:13:52 +0000 (13:13 +0000)]
[HVM] Replace shared-memory PIC state with a set-irq-level hypercall.
This simplifies the IRQ logic significantly and avoids the bogus
hvm_pic_assist() on domain resume path.

There is more work to be done here. At least:
 1. set-irq-level should really be set-interrupt-wire-level. Wire
    state needs to be distinguished from PIC (in particular, PIC IRR)
    state.
 2. Hypercalls can be batched in qemu and pushed down in one
    multicall.

Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[BLKTAP] No need for barrier when updating rsp_prod_pvt.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 12:31:17 +0000 (12:31 +0000)]
[BLKTAP] No need for barrier when updating rsp_prod_pvt.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[LIBXC] Fix memory alloc/dealloc routines to return error properly.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 11:54:52 +0000 (11:54 +0000)]
[LIBXC] Fix memory alloc/dealloc routines to return error properly.
Remove unnecessary decrease_reservation() from hvm builder.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[IA64] Build blktap driver by default.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 11:19:26 +0000 (11:19 +0000)]
[IA64] Build blktap driver by default.
Signed-off-by: Aron Griffis <aron@hp.com>
19 years ago[BLKTAP] ia64 support
kfraser@localhost.localdomain [Tue, 7 Nov 2006 11:17:39 +0000 (11:17 +0000)]
[BLKTAP] ia64 support
Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
19 years ago[BLKTAP] avoid race between tapdisk and xvd daemon with ufe_ring.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 11:14:52 +0000 (11:14 +0000)]
[BLKTAP] avoid race between tapdisk and xvd daemon with ufe_ring.
- Tapdisk updates rsp_cons by ioctl(). on the other hand xvd daemon
  reads rsp_cons in do_block_io_op() with RING_FULL().
  Copy request and memory barrier before updating rsp_cons.
- Tapdisk access req_prod_pvt with select(). on the other hand xvd
  daemon updates it in do_block_io_op().

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
19 years ago[BLKTAP] Various cleanups necessary for ia64 support.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 11:01:35 +0000 (11:01 +0000)]
[BLKTAP] Various cleanups necessary for ia64 support.
- remove magic number 8 and 64.
  page size / sector size is 8 on x86, however on ia64 page size is
  16kb.
  blkring size is 64 on x86. however it's 128 on ia64.
- replace 0xFFFF with INVALID_GRANT_HANDLE
- don't map io ring into user space with uncachable.
  kernel access the page with cacable and tapdisk uses memory barrier
  so that it isn't necessary.
- remove printk warning and inserted new line.
- remove unused variable, page.
- add one BUG()

Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp>
19 years agoEnhance HVM xentrace:
kfraser@localhost.localdomain [Tue, 7 Nov 2006 10:48:48 +0000 (10:48 +0000)]
Enhance HVM xentrace:
1) VMX xentrace data are store in current vcpu instead physical CPU.
2) Log PIO data in xentrace.

Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[XENAPI] Add 'image' field to xen_vbd_record, allowing an image to be
kfraser@localhost.localdomain [Tue, 7 Nov 2006 10:46:03 +0000 (10:46 +0000)]
[XENAPI] Add 'image' field to xen_vbd_record, allowing an image to be
associated with the vbd. xend supports this field and the client
bindings need it to support existing images that are not under
control of an SR.

Signed-off-by: Jim Fehlig <jfehlig@novell.com>
19 years ago[XEN] Remove unnecessary FORCE_CRASH() macro.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 10:41:35 +0000 (10:41 +0000)]
[XEN] Remove unnecessary FORCE_CRASH() macro.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[XEN] When balancing idlers per socket/core, do it one at a time.
Emmanuel Ackaouy [Tue, 7 Nov 2006 10:37:30 +0000 (10:37 +0000)]
[XEN] When balancing idlers per socket/core, do it one at a time.
Signed-off-by: Emmanuel Ackaouy <ack@xensource.com>
19 years ago[HVM][SVM] Turn off long mode (EFER.LMA) when CR0.PG==0.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 10:19:20 +0000 (10:19 +0000)]
[HVM][SVM] Turn off long mode (EFER.LMA) when CR0.PG==0.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM][VMX] Disable IA32e mode in VM_ENTRY_CONTROLS when guest tries
kfraser@localhost.localdomain [Tue, 7 Nov 2006 10:18:50 +0000 (10:18 +0000)]
[HVM][VMX] Disable IA32e mode in VM_ENTRY_CONTROLS when guest tries
to disable PG from LME mode.

Without this patch, the 64 Vista cannot boot.

Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
Signed-off-by: Xin Li <xin.b.li@intel.com>
19 years ago[NET] back: Fix wrap to zero in transmit credit scheduler.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 09:48:19 +0000 (09:48 +0000)]
[NET] back: Fix wrap to zero in transmit credit scheduler.
This could happen when credit_bytes == ~0UL (i.e., scheduling is 'disabled').
Signed-off-by: Kirk Allan <kallan@novell.com>
19 years ago[HVM] MAke HVM hypercall table NR_hypercalls entries big.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 09:35:57 +0000 (09:35 +0000)]
[HVM] MAke HVM hypercall table NR_hypercalls entries big.
Otherwise a buggy/malicious HVM guest can cause us to read
off the end of the table.
Signed-off-by: George Dunlap <gdunlap@xensource.com>
19 years agoTrivial change to look in correct location for qemu-dm when using the debug wrapper...
kfraser@localhost.localdomain [Tue, 7 Nov 2006 09:32:02 +0000 (09:32 +0000)]
Trivial change to look in correct location for qemu-dm when using the debug wrapper on a 64 bit system.
Signed-off-by: Ben Thomas <ben@virtualiron.com>
19 years agoPV-on-HVM: Fix PV-on-HVM drivers need maddr.h included for defn of maddr_t.
kfraser@localhost.localdomain [Tue, 7 Nov 2006 09:29:53 +0000 (09:29 +0000)]
PV-on-HVM: Fix PV-on-HVM drivers need maddr.h included for defn of maddr_t.
Signed-off-by: Takanori Kasai <Kasai.Takanori@jp.fujitsu.com>
Signed-off-by: Tsunehisa Doi <Doi.Tsunehisa@jp.fujitsu.com>
19 years ago[LINUX] Use pfn_to_page instead of relying in memory assumptions
kfraser@localhost.localdomain [Tue, 7 Nov 2006 09:28:18 +0000 (09:28 +0000)]
[LINUX] Use pfn_to_page instead of relying in memory assumptions

This makes usage agnostic wrt the memory model being used.

Signed-off-by: Glauber de Oliveira Costa <gcosta@redhat.com>
19 years ago[HVM] Fix 64-bit HVM domain creation.
kaf24@localhost.localdomain [Tue, 7 Nov 2006 00:02:52 +0000 (00:02 +0000)]
[HVM] Fix 64-bit HVM domain creation.
 1. vlapic must be initialised before VMX-specific init.
 2. various shadow tweaks (e.g., check d->vcpu[0] before use).
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[SVM] Strip down svm_do_launch() pending complete removal.
kaf24@localhost.localdomain [Mon, 6 Nov 2006 20:47:10 +0000 (20:47 +0000)]
[SVM] Strip down svm_do_launch() pending complete removal.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[SVM] Move VMCB construction to VCPU creation time.
kaf24@localhost.localdomain [Mon, 6 Nov 2006 20:23:08 +0000 (20:23 +0000)]
[SVM] Move VMCB construction to VCPU creation time.
Signed-off-by: Keir Fraser <keir@xensource.com>
19 years ago[HVM][SVM] Start moving SVM launch code to domain/vcpu initialisation.
kaf24@localhost.localdomain [Mon, 6 Nov 2006 18:57:33 +0000 (18:57 +0000)]
[HVM][SVM] Start moving SVM launch code to domain/vcpu initialisation.
This first patch is just enough to get SVM guests working again after
I broke SVM support earlier today.
Signed-off-by: Keir Fraser <keir@xensource.com>