xen.git
20 years agoAvoid floating point in hash table implementation.
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 18:06:50 +0000 (19:06 +0100)]
Avoid floating point in hash table implementation.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMove dom0_op hypercall to the guest_handle interface (inside Xen).
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 17:26:41 +0000 (18:26 +0100)]
Move dom0_op hypercall to the guest_handle interface (inside Xen).

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMerged.
emellor@leeni.uk.xensource.com [Fri, 3 Mar 2006 14:56:34 +0000 (15:56 +0100)]
Merged.

20 years agoWait for the devices configured at start-up, before allowing the boot to
emellor@leeni.uk.xensource.com [Fri, 3 Mar 2006 14:56:03 +0000 (15:56 +0100)]
Wait for the devices configured at start-up, before allowing the boot to
proceed.  This avoids the race (bug #549) that would result in the infamous
VFS: Cannot open root device "sda1" or unknown-block(0,0) message even when
the hotplug layer had succeeded to configure the device.

Note that The Infamous Message has many possible root causes -- this change
only fixes one of them.  This particular root cause is characterised by
successful boots interleaved with unsuccessful ones, and diagnose.py showing
both front and backend drivers in the Connected state.

Closes bug #549.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoChange the way that reboot-timeouts are handled. Rather than refreshing the
emellor@leeni.uk.xensource.com [Fri, 3 Mar 2006 14:41:56 +0000 (15:41 +0100)]
Change the way that reboot-timeouts are handled.  Rather than refreshing the
list of domains each time XendDomainInfo.refresh is called, and spawning
threads for each shutdown still in progress, we watch the control/shutdown
node and spawn just the one thread immediately.  When the daemon starts (say
after a crash) all domains are checked for entries in the store indicating a
reboot in progress.

This avoids a large cost on the store and in threads when the list of domains
is refreshed.  This happens every time the server is queried, so the cost can
be quite high.  It's especially a high cost when using xm shutdown -a -w --
this operation should perform a lot better now.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoAdded further integrity checking, this time checking for duplicate directory
emellor@leeni.uk.xensource.com [Fri, 3 Mar 2006 14:37:28 +0000 (15:37 +0100)]
Added further integrity checking, this time checking for duplicate directory
entries and for orphaned nodes in the database.

Added two flags, -R and -L, to disable the recovery code and the remove of
/local at start-up.  This makes it much easier to analyse corrupted tdb files.

Added some missing talloc_free calls in the previous integrity checking code.

Removed the transaction handle from the trace_io message -- unfortunately,
the transaction is always null at this point, as it's not yet been looked up.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoCope when the filename returned by losetup for an existing loop device is not
emellor@leeni.uk.xensource.com [Fri, 3 Mar 2006 14:32:42 +0000 (15:32 +0100)]
Cope when the filename returned by losetup for an existing loop device is not
fully qualified.  In this case, the readlink -f call to canonicalise that name
fails.

Closes bug #553.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoDomain0 identifies IOAPIC by physical base address rather than BIOS-assigned identifier.
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 14:05:25 +0000 (15:05 +0100)]
Domain0 identifies IOAPIC by physical base address rather than BIOS-assigned identifier.
This seems a more reliable method since it seems that we always trust the base address
and we don't need to work around renumbering/rebasing as we do with APICIDs.

This is an alternative to the Yonah networking patch from Allen Kay at Intel.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoClean up PIT-thru-ioapic setup. We don't need the legacy missed-tick
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 13:10:35 +0000 (14:10 +0100)]
Clean up PIT-thru-ioapic setup. We don't need the legacy missed-tick
logic and so do not need to keep the PIT enabled thru the 8259 PIC.

Todo: arguable we should sync with x86/64 platform files, assuming they
do not throw out 'recent legacy' stuff (e.g., 32-bit only platforms
like P6). Getting rid of all the 82489DX and 486-era SMP stuff would be
great.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMinor enhancements to gdbbuild, the build script for gdbserver-xen.
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 10:17:09 +0000 (11:17 +0100)]
Minor enhancements to gdbbuild, the build script for gdbserver-xen.
 1. Use $MAKE if set
 2. Make mirror configurable from environment

From: Horms <horms@verge.net.au>

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoIt's not necessary to define a function for dumping virtual IOAPIC
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:58:42 +0000 (10:58 +0100)]
It's not necessary to define a function for dumping virtual IOAPIC
table entry, that makes the code hard to comply with condition
compiling of HVM_DBG_LOG.

Signed-off-by: Xin Li <xin.b.li@intel.com>
20 years agoAdd a description of the virtual TPM interface to the interface
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:57:37 +0000 (10:57 +0100)]
Add a description of the virtual TPM interface to the interface
document and explain some of the internals of the device driver.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
20 years agoThis patch undoes 8697 and makes xm reboot/shutdown work again. This code
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:56:54 +0000 (10:56 +0100)]
This patch undoes 8697 and makes xm reboot/shutdown work again. This code
does expose a bug in xenstored when it is killed & restarted, but that
is not caused by this code.

My sincere appology to Ke Yu for the long wait and assuming his code
was broken.  His code was perfectly fine. Its just that my test scripts
did something for a historical reason that I should have removed a long
time ago.

Signed-Off-By: Leendert van Doorn <leendert@watson.ibm.com>
20 years agoRaises an explicit error message when the domain config file specifies
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:56:13 +0000 (10:56 +0100)]
Raises an explicit error message when the domain config file specifies
an incorrect path to a file-backed VBD (disk image) or the file-backed
VBD does not exist.

Without this patch, if someone mis-types the path to the disk image,
he/she would have no indication as to what's wrong when the domain
fails to get created.

Signed-off-by: Khoa Huynh <khoa@us.ibm.com>
20 years agoVTPM fixes:
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:55:11 +0000 (10:55 +0100)]
VTPM fixes:
- fixes a problem in the vTPM backend where the timer was killed when
the timer function was run
- removes a variable from being put in the ring by the frontend
- reformatting due to indent tool running over some of the files

Has been tested with the vtpm-relevant parts of xm-test.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
20 years agoFix up some issues I found when porting PPC to the new common gdb stub code:
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:53:58 +0000 (10:53 +0100)]
Fix up some issues I found when porting PPC to the new common gdb stub code:
- cosmetic changes in the messages printed
- 'flags' must always be unsigned long.
- explicitly calling initialize_gdb() is not difficult. For x86 and ia64 I
placed this call immediately before do_initcalls(), since that's where it's
being called from now so we know it's safe. Architecture people can move it
earlier as appropriate.
- I don't understand all these ASSERT(!local_irq_is_enabled()) statements,
sometimes bracketing a single call like receive_command(). How exactly would
receive_command() manage to re-enable irqs? Also, a failing ASSERT would just
call into the stub again anways...
- initialize_gdb() was overcomplicated. serial_parse_handle() already handles
the parsing for us, and there's no need to panic there.

Signed-off-by: Hollis Blanchard <hollisb@us.ibm.com>
20 years agoVTPM_Tools: Fix error when closing only vtpm, and fix restore bug when
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:46:06 +0000 (10:46 +0100)]
VTPM_Tools: Fix error when closing only vtpm, and fix restore bug when
no dmis exist

The first bug prevented the TPM tests from running successfully more
than once in a row.

Signed-off-by: Vinnie Scarlata <vincent.r.scarlata@intel.com>
20 years agoIgnore external (user-provided) CFLAGS when building vmxassist and hvmloader.
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:44:40 +0000 (10:44 +0100)]
Ignore external (user-provided) CFLAGS when building vmxassist and hvmloader.
They do more harm than good.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoAdd a 'clear trap table' path to set_trap_table hypercall, taken when
kaf24@firebug.cl.cam.ac.uk [Fri, 3 Mar 2006 09:34:03 +0000 (10:34 +0100)]
Add a 'clear trap table' path to set_trap_table hypercall, taken when
the input table is NULL.

Also, no need for set_trap_table to take the per-domain biglock.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoCope if xenstore is down.
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 20:45:13 +0000 (21:45 +0100)]
Cope if xenstore is down.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoAdded hashtable implementation, to support the reachability check against the
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 20:44:49 +0000 (21:44 +0100)]
Added hashtable implementation, to support the reachability check against the
store.  This code is by Christopher Clark.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoDon't compile HVM_DBG_LOG into xen by default.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 20:37:14 +0000 (21:37 +0100)]
Don't compile HVM_DBG_LOG into xen by default.

Signed-off-by: Xin Li <xin.b.li@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
20 years agoFix libxc warnings with gcc-4.1 and glibc-2.3.90.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 20:36:03 +0000 (21:36 +0100)]
Fix libxc warnings with gcc-4.1 and glibc-2.3.90.

Signed-off-by: Charles Coffing <ccoffing@novell.com>
20 years agoMany fixes and cleanups for lomount:
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 20:35:17 +0000 (21:35 +0100)]
Many fixes and cleanups for lomount:
- Fixed several overflows, off-by-one, and uninitialized variables.
- Added well-defined exit codes.
- Proper handling of system()'s return value.
- Errors parsing partition table cause it to stop now.
- etcetera...

Tested on 32 and 64 bit, with valgrind, with physical disks and disk
images.

Signed-off-by: Charles Coffing <ccoffing@novell.com>
20 years agoChange debug level of vmx_io_instruction to DBG_LEVEL_IO.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 20:34:16 +0000 (21:34 +0100)]
Change debug level of vmx_io_instruction to DBG_LEVEL_IO.

From: Himanshu Raj

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoAdd empty files need to fix ia64 build.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 18:16:21 +0000 (19:16 +0100)]
Add empty files need to fix ia64 build.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoClean up use of wait_event_interruptible().
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 17:57:03 +0000 (18:57 +0100)]
Clean up use of wait_event_interruptible().

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agolinux-x86-64: eliminate duplicate exports
cl349@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 15:41:45 +0000 (15:41 +0000)]
linux-x86-64: eliminate duplicate exports
Symbols are exported from files shared with i386.

From: Jan Beulich <JBeulich@novell.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoRemove unnecessary include statements.
cl349@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 15:37:46 +0000 (15:37 +0000)]
Remove unnecessary include statements.

Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agolinux: clean up setup_arch_post.h
cl349@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 15:33:47 +0000 (15:33 +0000)]
linux: clean up setup_arch_post.h
- machine_specific_memory_setup() was unused on x86-64
- machine_specific_modify_cpu_capabilities() must not be __init (for CPU hotplug)
- machine_specific_modify_cpu_capabilities() should be called in similar ways on 32- and 64-bits

From: Jan Beulich <JBeulich@novell.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoWhile native code range-checks the requested address, possibly
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 15:06:51 +0000 (16:06 +0100)]
While native code range-checks the requested address, possibly
truncates the range, and then reads/writes page-wise until possibly
encountering a -EFAULT issue, Xen code accessed the whole range in a
single step, thus not allowing partially succeeding accesses.

From: Jan Beulich

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoA few changes to the new 'guest handle' interface:
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 13:43:24 +0000 (14:43 +0100)]
A few changes to the new 'guest handle' interface:

 DEFINE_GUEST_HANDLE -> define_guest_handle

 GUEST_HANDLE        -> guest_handle

 New __define_guest_handle allows handle name different from
 encapsulated type's name (useful for awkward typenames).

 Got rid of xen_ulong, now use guest_handle(ulong).

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMake sure that installed files have sensible permissions and are owned by the
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 10:41:48 +0000 (11:41 +0100)]
Make sure that installed files have sensible permissions and are owned by the
user running install. Also clean up the install script a little.

Signed-off-by: Horms <horms@verge.net.au>
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix problems when creating HVM guest on PAE host with >4G physical memory.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 10:03:34 +0000 (11:03 +0100)]
Fix problems when creating HVM guest on PAE host with >4G physical memory.

Signed-off-by: Yunhong Jiang yunhong.jiang@intel.com
Signed-off-by: Jun Nakajima jun.nakajima@intel.com
20 years agomerge
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 10:00:49 +0000 (11:00 +0100)]
merge

20 years agoFix printing of u64 value 'msr_content' to use PRIx64 format.
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 09:59:34 +0000 (10:59 +0100)]
Fix printing of u64 value 'msr_content' to use PRIx64 format.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMerge xen-ia64-unstable.hg
kaf24@firebug.cl.cam.ac.uk [Thu, 2 Mar 2006 09:49:19 +0000 (10:49 +0100)]
Merge xen-ia64-unstable.hg

20 years agoMerged.
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 02:21:38 +0000 (03:21 +0100)]
Merged.

20 years agoFix violation of C90 mixed-code-and-declarations restriction.
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 02:21:17 +0000 (03:21 +0100)]
Fix violation of C90 mixed-code-and-declarations restriction.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoRemoving out-of-date unused file.
vhanquez@kneesa.uk.xensource.com [Thu, 2 Mar 2006 02:14:45 +0000 (02:14 +0000)]
Removing out-of-date unused file.
current code available in io/xs_wire.h

Signed-off-by: Vincent Hanquez <vincent@xensource.com>
20 years agoFix printf.
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 01:35:23 +0000 (02:35 +0100)]
Fix printf.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoAdded a basic integrity checker, and some basic ability to recover from store
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 01:09:23 +0000 (02:09 +0100)]
Added a basic integrity checker, and some basic ability to recover from store
corruption, rather than just spewing error messages and exiting.

Added a xenstore-control executable, which sends commands to xenstored.
Currently, the only command is 'check', which triggers an integrity check.
(The integrity check is also triggered whenever a corrupted store is detected).

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoSilence the messages that are emitted when removing nodes that already have
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 01:01:17 +0000 (02:01 +0100)]
Silence the messages that are emitted when removing nodes that already have
been removed.  This is fine, in a cleanup script.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoAvoid a race between Xend removing the vm entries corresponding to a domain,
emellor@leeni.uk.xensource.com [Thu, 2 Mar 2006 00:59:28 +0000 (01:59 +0100)]
Avoid a race between Xend removing the vm entries corresponding to a domain,
and the hotplug scripts checking whether that domain is still alive.  This
triggers the "xenstore-read: couldn't read path /local/domain/107/vm" message
in /var/log/xen-hotplug.log, followed by "Device XYZ is mounted in a guest
domain and so cannot be mounted now".

Closes bug #514.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoFix problems with HVM PAE guests with >4GB memory.
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 22:41:43 +0000 (23:41 +0100)]
Fix problems with HVM PAE guests with >4GB memory.

Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
20 years agoAdd SVM 32bit msr support (combined both 32bit with 64bit functions)
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 22:39:43 +0000 (23:39 +0100)]
Add SVM 32bit msr support (combined both 32bit with 64bit functions)
and enable vlapic for 64bit.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoClean up SVM 32-bit exits.S file (remove all tsc/fpu).
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 22:38:18 +0000 (23:38 +0100)]
Clean up SVM 32-bit exits.S file (remove all tsc/fpu).

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoRevert two uses of CONFIG_PAGING_LEVELS to CPP rather than C predicate.
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 22:35:55 +0000 (23:35 +0100)]
Revert two uses of CONFIG_PAGING_LEVELS to CPP rather than C predicate.
Fixes PAE and 64-bit builds.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years ago[IA64] sparse update for 2.6.16-rc5
awilliam@xenbuild.aw [Wed, 1 Mar 2006 22:06:24 +0000 (15:06 -0700)]
[IA64] sparse update for 2.6.16-rc5

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
20 years ago[IA64] default to SMP
awilliam@xenbuild.aw [Wed, 1 Mar 2006 21:15:44 +0000 (14:15 -0700)]
[IA64] default to SMP

UP builds are not maintained on xen-unstable, we seem to be
running SMP well enough to make the transition.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
20 years ago[IA64] Turn on CONFIG_XEN_BLKDEV_BACKEND
awilliam@xenbuild.aw [Wed, 1 Mar 2006 21:12:30 +0000 (14:12 -0700)]
[IA64] Turn on CONFIG_XEN_BLKDEV_BACKEND

We relied on the Makefile for building this before, now we
need it set in our .config file.

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
20 years agomerge
awilliam@xenbuild.aw [Wed, 1 Mar 2006 19:47:25 +0000 (12:47 -0700)]
merge

20 years agomap_domain_page() now handles running on idle page tables.
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 17:34:06 +0000 (18:34 +0100)]
map_domain_page() now handles running on idle page tables.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoDisallow cmpxchg8b ptwr emulation for non-pae. Also a few cleanups.
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 17:33:36 +0000 (18:33 +0100)]
Disallow cmpxchg8b ptwr emulation for non-pae. Also a few cleanups.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years ago[IA64] Every processor call xen_early_setup (required on SMP-guest to set ivt).
awilliam@xenbuild.aw [Wed, 1 Mar 2006 17:01:54 +0000 (10:01 -0700)]
[IA64] Every processor call xen_early_setup (required on SMP-guest to set ivt).

Only BP set running_on_xen.
xensetup.S slightly compacted.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] warning cleanup
awilliam@xenbuild.aw [Wed, 1 Mar 2006 16:53:17 +0000 (09:53 -0700)]
[IA64] warning cleanup

Signed-off-by: Alex Williamson <alex.williamson@hp.com>
20 years agoRemove unused #include <sys/ioctl.h>.
emellor@leeni.uk.xensource.com [Wed, 1 Mar 2006 16:52:37 +0000 (17:52 +0100)]
Remove unused #include <sys/ioctl.h>.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years ago[IA64] VTI: updated vtlb, support_non_contiguous memory on vtidomain
awilliam@xenbuild.aw [Wed, 1 Mar 2006 15:29:00 +0000 (08:29 -0700)]
[IA64] VTI: updated vtlb, support_non_contiguous memory on vtidomain

Previously VTI-domain only supported contiguous memory,
this patch is intended to make VTI-domain support non-contiguous memory.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
20 years agoClean wrpt state when emulating CR3 write.
kaf24@firebug.cl.cam.ac.uk [Wed, 1 Mar 2006 13:49:31 +0000 (14:49 +0100)]
Clean wrpt state when emulating CR3 write.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoSVM patch to add 64bit hv support.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 21:57:38 +0000 (22:57 +0100)]
SVM patch to add 64bit hv support.
This patch only modifies svm files.
Tested with c/s 9015 with 32bit hv using UP Dom0, with UP linux and
winxpsp1 unmodified guests.
Tested with c/s 9015 with 64bit hv using UP Dom0, with 32bit and 64bit
UP linux and 32bit winxpsp1.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years ago[IA64] Unused fields in mm_struct commented out.
awilliam@xenbuild.aw [Tue, 28 Feb 2006 20:18:08 +0000 (13:18 -0700)]
[IA64] Unused fields in mm_struct commented out.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] Misc clean-up and warnings removal.
awilliam@xenbuild.aw [Tue, 28 Feb 2006 20:10:16 +0000 (13:10 -0700)]
[IA64] Misc clean-up and warnings removal.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] xen_timer_interval removed (was not used).
awilliam@xenbuild.aw [Tue, 28 Feb 2006 19:43:08 +0000 (12:43 -0700)]
[IA64] xen_timer_interval removed (was not used).

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] delivery_mask removed (was unused).
awilliam@xenbuild.aw [Tue, 28 Feb 2006 19:23:46 +0000 (12:23 -0700)]
[IA64] delivery_mask removed (was unused).

The field was renamed in arch-ia64.h for backward compatibility with xen linux.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] Found the lost memory
awilliam@xenbuild.aw [Tue, 28 Feb 2006 19:02:50 +0000 (12:02 -0700)]
[IA64] Found the lost memory

This patch can work around virtual memmap/discontig memory issues,
but memory gain may not be realized.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
20 years agoAllow tpmfront/tpmback to be built as modules.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 18:00:15 +0000 (19:00 +0100)]
Allow tpmfront/tpmback to be built as modules.

From: Jan Beulich

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoAllow netback to be built as a module.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 17:59:24 +0000 (18:59 +0100)]
Allow netback to be built as a module.

From: Jan Beulich

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoAllow blkback to be built as a module.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 17:56:00 +0000 (18:56 +0100)]
Allow blkback to be built as a module.

From: Jan Beulich

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years ago[IA64] cleanup warning in xen/ia64 (arch/ia64/xen)
awilliam@xenbuild.aw [Tue, 28 Feb 2006 17:29:30 +0000 (10:29 -0700)]
[IA64] cleanup warning in xen/ia64 (arch/ia64/xen)

These patches fix many issue (ex. panic dom0, oops domU/dom0...).
we will fix compile warnnings one step at a time.

Signed-off-by: Tsunehisa Doi <doi.tsunehisa@jp.fujitsu.com>
Signed-off-by: Kouya SHIMURA <kouya@jp.fujitsu.com>
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>
20 years ago[IA64] cleanup warning in xen/ia64 (arch/ia64/linux-xen)
awilliam@xenbuild.aw [Tue, 28 Feb 2006 17:26:43 +0000 (10:26 -0700)]
[IA64] cleanup warning in xen/ia64 (arch/ia64/linux-xen)

These patches fix many issue (ex. panic dom0, oops domU/dom0...).
we will fix compile warnnings one step at a time.

Signed-off-by: Tsunehisa Doi <doi.tsunehisa@jp.fujitsu.com>
Signed-off-by: Kouya SHIMURA <kouya@jp.fujitsu.com>
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>
20 years ago[IA64] cleanup warning in xen/ia64 (arch/ia64/linux)
awilliam@xenbuild.aw [Tue, 28 Feb 2006 17:24:03 +0000 (10:24 -0700)]
[IA64] cleanup warning in xen/ia64 (arch/ia64/linux)

These patches fix many issue (ex. panic dom0, oops domU/dom0...).
we will fix compile warnnings one step at a time.

Signed-off-by: Tsunehisa Doi <doi.tsunehisa@jp.fujitsu.com>
Signed-off-by: Kouya SHIMURA <kouya@jp.fujitsu.com>
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>
20 years ago[IA64] cleanup warning in xen/ia64 (include/asm-ia64)
awilliam@xenbuild.aw [Tue, 28 Feb 2006 17:21:54 +0000 (10:21 -0700)]
[IA64] cleanup warning in xen/ia64 (include/asm-ia64)

These patches fix many issue (ex. panic dom0, oops domU/dom0...).
we will fix compile warnnings one step at a time.

Signed-off-by: Tsunehisa Doi <doi.tsunehisa@jp.fujitsu.com>
Signed-off-by: Kouya SHIMURA <kouya@jp.fujitsu.com>
Signed-off-by: Masaki Kanno <kanno.masaki@jp.fujitsu.com>
Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com>
20 years agoAvoid using unexported add_timer_on() function in netback driver.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 17:15:18 +0000 (18:15 +0100)]
Avoid using unexported add_timer_on() function in netback driver.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years ago[IA64] Cleanup hypercall.h
awilliam@xenbuild.aw [Tue, 28 Feb 2006 17:13:55 +0000 (10:13 -0700)]
[IA64] Cleanup hypercall.h

Cleanup to hypercall.h of ia64 xenlinux, and remove used stuff.

Signed-off-by Kevin Tian <kevin.tian@intel.com>

20 years agoSketch a new interface for transferring hypercall arguments in memory.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 16:45:20 +0000 (17:45 +0100)]
Sketch a new interface for transferring hypercall arguments in memory.
Instead of manipulating guest virtual addresses, 'guest handles' are
passed across the hypercall interface, which may only be manipulated
via the interface exported by guest_access.h.

So far this has only been applied to the memory_op hypercall. The
interfaces are still subject to change. Other hypercalls can be updated
after the interface is agreed upon.

Also cleaned up the hypercall_create_continuation() interface to take
a format string and different-typed varargs (they do not all need to
be castable to longs).

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoClean up management of 'initial images' temporary buffers.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 10:58:45 +0000 (11:58 +0100)]
Clean up management of 'initial images' temporary buffers.
Avoid double-freeing boundary pages.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix tools Makefiles to respect CFLAGS. Disable stack-smash
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 10:41:54 +0000 (11:41 +0100)]
Fix tools Makefiles to respect CFLAGS. Disable stack-smash
protection when building vmxassist/hvmloader.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoAdd 'm' debug key to print free-memory info.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 10:07:18 +0000 (11:07 +0100)]
Add 'm' debug key to print free-memory info.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix a multi HVM domain issue (bugzilla bug #542):
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:18:34 +0000 (10:18 +0100)]
Fix a multi HVM domain issue (bugzilla bug #542):
        This issue can be reproduced on SMP platform, while the domain 0
is UP. The reason is, after finishing a dma request, the dma thread will
trigger the interrupt and then clear the call back function. When guest
get the interrupt , it will try to check the call back function, if it
is set, then it will trigger a dma request again. So if the checking for
the callbackfunction happens between trigger interrupt and clear
callback function on dma thread, it will cause NULL function calling.
After with domain 0 UP and smp platform, this situation can be
reproduced easily.

Signed-off-by: Yunhong Jiang <yunhong.jiang@intel.com>
20 years agoFix SVM cr0 handling.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:14:32 +0000 (10:14 +0100)]
Fix SVM cr0 handling.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoFix SVM PAE support.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:14:10 +0000 (10:14 +0100)]
Fix SVM PAE support.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoFix SVM guest pin timer migration logic.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:13:26 +0000 (10:13 +0100)]
Fix SVM guest pin timer migration logic.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoFix SVM vlapic logic.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:12:29 +0000 (10:12 +0100)]
Fix SVM vlapic logic.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years agoRemoving hvm_safe_block() from the SVM hlt function avoids SVM guest problems.
kaf24@firebug.cl.cam.ac.uk [Tue, 28 Feb 2006 09:11:33 +0000 (10:11 +0100)]
Removing hvm_safe_block() from the SVM hlt function avoids SVM guest problems.
This is a temporary solution.

Signed-off-by: Tom Woller <thomas.woller@amd.com>
20 years ago[IA64] fix SMP bug for vhpt
awilliam@xenbuild.aw [Mon, 27 Feb 2006 21:00:33 +0000 (14:00 -0700)]
[IA64] fix SMP bug for vhpt

vhpt_paddr & vhpt_pend are now per-cpu variables, since VHPT is allocated
per cpu.  This should really improve stability.

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] warning removal: NR_CPUS
awilliam@xenbuild.aw [Mon, 27 Feb 2006 20:31:10 +0000 (13:31 -0700)]
[IA64] warning removal: NR_CPUS

linux/threads nullified (moved to linux-null).

Signed-off-by: Tristan Gingold <tristan.gingold@bull.net>
20 years ago[IA64] SMP_HOST: Alloc vhpt from domheap
awilliam@xenbuild.aw [Mon, 27 Feb 2006 20:16:16 +0000 (13:16 -0700)]
[IA64] SMP_HOST: Alloc vhpt from domheap

It's more reasonable to alloc vhpt from domheap, instead of xenheap.

Signed-off-by: Anthony Xu <anthony.xu@intel.com>
20 years agoMerged.
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 17:44:38 +0000 (18:44 +0100)]
Merged.

20 years agoAdd support for configuring feature flags when starting guests.
Ian.Campbell@xensource.com [Mon, 27 Feb 2006 17:27:52 +0000 (17:27 +0000)]
Add support for configuring feature flags when starting guests.

Guest configuration files may now contain a features= declaration
listing the features which should be enabled in a | separated
list.

Signed-off-by: Ian Campbell <ian.campbell@xensource.com>
20 years agoUpdate to Linux 2.6.16-rc5.
cl349@firebug.cl.cam.ac.uk [Mon, 27 Feb 2006 16:56:08 +0000 (16:56 +0000)]
Update to Linux 2.6.16-rc5.

Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoCheck feature flag supervisor_mode_kernel in dom0 builder.
kaf24@firebug.cl.cam.ac.uk [Mon, 27 Feb 2006 16:55:22 +0000 (17:55 +0100)]
Check feature flag supervisor_mode_kernel in dom0 builder.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoReconstitute the ip parameter correctly. Closes bug #491.
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 16:26:16 +0000 (17:26 +0100)]
Reconstitute the ip parameter correctly.  Closes bug #491.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoAllow a guest to switch pagetables 'atomically'. That is, the new pagetables
kaf24@firebug.cl.cam.ac.uk [Mon, 27 Feb 2006 16:18:58 +0000 (17:18 +0100)]
Allow a guest to switch pagetables 'atomically'. That is, the new pagetables
do not need to be read-only in the old pagetables, and the old pagetables do
not need to be read-only in the new pagetables.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoUse cleverer sed pattern to avoid use of head -1.
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 16:16:28 +0000 (17:16 +0100)]
Use cleverer sed pattern to avoid use of head -1.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoBreak out the hotplug cleanup code into xen-hotplug-cleanup. Claim the block
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 16:16:04 +0000 (17:16 +0100)]
Break out the hotplug cleanup code into xen-hotplug-cleanup.  Claim the block
script's lock before doing the cleanup, to avoid a race that causes the message
"xenstore-list: could not list path backend/vbd/40" in /var/log/hotplug.log,
with a subsequent refusal to create the domain, after a domain reboot.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoMerged.
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 15:29:03 +0000 (16:29 +0100)]
Merged.

20 years agoMove files that have appeared in the wrong place due to incorrect patch -p setting.
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 15:28:57 +0000 (16:28 +0100)]
Move files that have appeared in the wrong place due to incorrect patch -p setting.

Signed-off-by: Ewan Mellor <ewan@xensource.com>
20 years agoxen/ia64 uaccess is still linux style with different declaration as
kaf24@firebug.cl.cam.ac.uk [Mon, 27 Feb 2006 15:22:02 +0000 (16:22 +0100)]
xen/ia64 uaccess is still linux style with different declaration as
common xen. It's meaningless to patch original file with many #ifdef
XEN, and so instead copy that file to create a new xen specific
uaccess file, with unused lines removed. May need render later.

Or else it breaks compilation on xen-unstable.

Signed-off-by Kevin Tian <kevin.tian@intel.com>

20 years agoFix a missed unmap in __entry().
kaf24@firebug.cl.cam.ac.uk [Mon, 27 Feb 2006 15:20:23 +0000 (16:20 +0100)]
Fix a missed unmap in __entry().
On x86_32p, after several creation/destroy of VMX domain, xen crashes
because we forget to unmap phys_table.

Signed-off-by: Xin Li <xin.b.li@intel.com>
20 years agoWhile testing TCP & UDP tests (coming soon), I tracked down a bug in the
emellor@leeni.uk.xensource.com [Mon, 27 Feb 2006 15:14:11 +0000 (16:14 +0100)]
While testing TCP & UDP tests (coming soon), I tracked down a bug in the
IP addressing scheme
used by xm-test: each machine running network tests is using the same IP
addresses, so two machines
in the same network can cross-pollinate/interfere with each other's tests.

This patch changes the IP addressing scheme to add some randomness. It's
gone through a couple of dozen
passes of xm-test, some on -bridge and some on -route, while xm-test was
running on another box on the same network.

Signed-off-by: Jim Dykman <dykman@us.ibm.com>