xen.git
20 years agoVarious minor vmxassist cleanups.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 17:53:52 +0000 (17:53 +0000)]
Various minor vmxassist cleanups.

Signed-Off-By: Leendert van Doorn <leendert@watson.ibm.com>
20 years agoWe need to be more precise and restore the %eip.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 17:53:34 +0000 (17:53 +0000)]
We need to be more precise and restore the %eip.

Signed-Off-By: Leendert van Doorn <leendert@watson.ibm.com>
20 years agoHandle page overlapping copies.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 17:53:04 +0000 (17:53 +0000)]
Handle page overlapping copies.

Signed-Off-By: Leendert van Doorn <leendert@watson.ibm.com>
20 years agoQemu-dm dumps core with the pcnet device. This patches fixes it.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 17:52:37 +0000 (17:52 +0000)]
Qemu-dm dumps core with the pcnet device. This patches fixes it.

When pcnet_receive calls pcnet_poll, which polls the receive and the send
rings. Whenever there is an element in the send ring that is owned by
the Lance chip it will call pcnet_transmit and send it. When the element
is the endp(acket), pcnet_transmit will copy it out, send the packet
(qemu_send_packet) and then clear the owner bit. Somewherer along the
qemu_send_packet execution path, pcnet_recieve is called again, which
calls pcnet_poll and starts this whole process again. This very rapidly
leads to a stack overflow and crashes qemu.

The fix is simple, stop the recursion. Once the packet is copied into
qemu datatstructure (before qemu_send_packet is called!), the owner bit
on the ring element should be cleared.

Signed-Off-By: Leendert van Doorn <leendert@watson.ibm.com>
20 years agoFix preemption-check race in memory_op hypercall.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 17:02:08 +0000 (17:02 +0000)]
Fix preemption-check race in memory_op hypercall.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix writing to mmap'ed /dev/mem region mapped PROT_WRITE
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 16:51:55 +0000 (16:51 +0000)]
Fix writing to mmap'ed /dev/mem region mapped PROT_WRITE
and MAP_PRIVATE. This is in fact a generic Linux bug.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoInstall hotplug script(s) unconditionally.
cl349@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 14:20:12 +0000 (14:20 +0000)]
Install hotplug script(s) unconditionally.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agomerge?
cl349@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 14:17:08 +0000 (14:17 +0000)]
merge?

20 years agoRestore configurability of vif bring up script and passing arguments to the script.
cl349@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 14:15:49 +0000 (14:15 +0000)]
Restore configurability of vif bring up script and passing arguments to the script.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoMini-os fixes from Simon Kagstrom.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 10:04:42 +0000 (10:04 +0000)]
Mini-os fixes from Simon Kagstrom.

20 years agoMissing shadow files.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 10:04:11 +0000 (10:04 +0000)]
Missing shadow files.

20 years agoThis patch is to boot 32-bit VMX guest on the 64-bit host.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 09:43:32 +0000 (09:43 +0000)]
This patch is to boot 32-bit VMX guest on the 64-bit host.
Double-compile is used to make both 64-bit guest and 32-bit guest can
work, the shadow page-table uses current 64-bit shadow code's structure
to simulate 32-bit guest's 2-level page-table.

Signed-off-by: Chengyuan Li <chengyuan.li@intel.com>
Signed-off-by: Xiaohui Xin <xiaohui.xin@intel.com>
Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
20 years agoClean up tools/check error reporting. Remove logging.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 09:02:13 +0000 (09:02 +0000)]
Clean up tools/check error reporting. Remove logging.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoNew scripts I missed from previous security patch.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 08:06:59 +0000 (08:06 +0000)]
New scripts I missed from previous security patch.

20 years agoThis patch:
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 07:59:12 +0000 (07:59 +0000)]
This patch:

* adds a get_ssid ACM command that allows privileged domains to retrieve
types for either a given ssid reference or a given domain id (of a running
domain); this command can be used to extend access control into device
domains, e.g., to control network traffic currently moving through Domain
0 uncontrolled by the ACM policy

* adds a script getlabel.sh that allows users inside Dom0 to retrieve the
label for a given ssid reference or a given domain id (multiple labels
might map onto a single ssid reference)

* cleans up label-related code in tools/security by merging common
functions into labelfuncs.sh

* cleans up ACM code related to above changes (eventually approximating a
common coding style)

Signed-off-by Reiner Sailer <sailer@us.ibm.com>
Signed-off by Stefan Berger <stefanb@us.ibm.com>

20 years agoMake the check script log to /tmp instead of the working directory. This
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 07:55:45 +0000 (07:55 +0000)]
Make the check script log to /tmp instead of the working directory.  This
allows us to install from a read-only file system (such as a cdrom).

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
20 years agoI've found cramfs to be very useful for simplifying Xen testing with QEMU.
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 07:55:22 +0000 (07:55 +0000)]
I've found cramfs to be very useful for simplifying Xen testing with QEMU.

Also, I imagine cramfs is used frequently for initrds so it's probably a good
idea to support it out of the box.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
20 years agoThe attached patch makes xend clear the reboot request from the store
kaf24@firebug.cl.cam.ac.uk [Fri, 2 Sep 2005 07:53:46 +0000 (07:53 +0000)]
The attached patch makes xend clear the reboot request from the store
when rebooting a domain.  This prevents the lingering request from
being re-fulfilled, such as when executing a subsequent sysrq request.

Signed-off-by: Dan Smith <danms@us.ibm.com>
20 years agoFix save/restore when using grant tables in network, robust-ify netback
shand@ubuntu.eng.hq.xensource.com [Fri, 2 Sep 2005 04:19:34 +0000 (20:19 -0800)]
Fix save/restore when using grant tables in network, robust-ify netback
against bogus front-end drivers, plus various small cleanups. Note that
there is now only /one/ config option controlling the use of grant tables
with networking (not separate ones for rx and tx).

An outstanding issue is that console reconnect after a restore causes badness.

Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoFix build dependency for the XEN tpm driver.
shand@ubuntu.eng.hq.xensource.com [Thu, 1 Sep 2005 19:09:16 +0000 (11:09 -0800)]
Fix build dependency for the XEN tpm driver.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
20 years agoBetter Xen backtraces in debug builds (follow the
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 15:31:12 +0000 (15:31 +0000)]
Better Xen backtraces in debug builds (follow the
stack frame pointer).

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix network setup through hotplug on SUSE.
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:45:50 +0000 (10:45 +0000)]
Fix network setup through hotplug on SUSE.
Move the hotplug script from /etc/hotplug.d/xen-backend/backend.hotplug
to /etc/hotplug/xen-backend.agent, which seems to be a more common place
for the script.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
--HG--
rename : tools/examples/backend.hotplug => tools/examples/xen-backend.agent

20 years agoHook up sysrq for xencons.
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:38:57 +0000 (10:38 +0000)]
Hook up sysrq for xencons.
Having sysrq is regularly handy.  Following in the tradition set by the
various consoles for virtual ppc stuff, the sysrq key is ^O.
Signed-off-by: Jeremy Katz <katzj@redhat.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoFix closing of /var/log/xend.log fd
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:26:17 +0000 (10:26 +0000)]
Fix closing of /var/log/xend.log fd
Fix the problem where /var/log/xend.log was getting closed and then
/var/log/xend-debug was getting all the spam about the fd being closed
in the logging stuff.

Basically what was happening is that the main thread wasn't sticking
around and during its exit, the atexit handlers of the python logging
stuff got called closing the log fd.

Patch makes it so that we instead wait on the real server threads to
exit before shutting things down.

Signed-off-by: Jeremy Katz <katzj@redhat.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoremove random file
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:16:40 +0000 (10:16 +0000)]
remove random file
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agomerge?
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:16:14 +0000 (10:16 +0000)]
merge?

20 years agoCleanup domain listing.
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 10:08:53 +0000 (10:08 +0000)]
Cleanup domain listing.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoFix xm list for when info is requested for individual domains.
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 09:57:20 +0000 (09:57 +0000)]
Fix xm list for when info is requested for individual domains.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoAvoid code duplication between stringified and
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 09:37:54 +0000 (09:37 +0000)]
Avoid code duplication between stringified and
non-stringified asm macros, by stringifying on use
where that is needed.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoUse entry_get_paddr(l1e) to get the PFN
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 09:07:10 +0000 (09:07 +0000)]
Use entry_get_paddr(l1e) to get the PFN

Otherwise software used bits might get in the way.

Signed-off-by: Arun Sharma <arun.sharma@intel.com>
20 years agoTiny build fix for gcc 4.0.1
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 09:04:06 +0000 (09:04 +0000)]
Tiny build fix for gcc 4.0.1
Signed-off-by: Christopher Clark <christopher.clark@cl.cam.ac.uk>
20 years agoVTPM is off by default.
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 09:01:15 +0000 (09:01 +0000)]
VTPM is off by default.

20 years agoLinux 2.6 cleanups.
cl349@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 08:25:22 +0000 (08:25 +0000)]
Linux 2.6 cleanups.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoRemove curl checks. Patch from Anthony Liguori.
kaf24@firebug.cl.cam.ac.uk [Thu, 1 Sep 2005 08:11:16 +0000 (08:11 +0000)]
Remove curl checks. Patch from Anthony Liguori.

20 years agomerge?
shand@ubuntu.eng.hq.xensource.com [Wed, 31 Aug 2005 23:22:45 +0000 (15:22 -0800)]
merge?

20 years agoFix vtpm build when doing 'make dist' (particularly when not root).
shand@ubuntu.eng.hq.xensource.com [Wed, 31 Aug 2005 22:37:21 +0000 (14:37 -0800)]
Fix vtpm build when doing 'make dist' (particularly when not root).

Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoFix non-SMP build, broken by the previous changeset.
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 14:53:43 +0000 (14:53 +0000)]
Fix non-SMP build, broken by the previous changeset.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix DEBUG_PREEMPT builds. irq_disabled() macro must not use
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 14:20:56 +0000 (14:20 +0000)]
Fix DEBUG_PREEMPT builds. irq_disabled() macro must not use
smp_processor_id(). Use the underlying function instead.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoDisallow domain 0 access to certain I/O port ranges
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 12:56:42 +0000 (12:56 +0000)]
Disallow domain 0 access to certain I/O port ranges
(Master/Slave PICs, PIT, Speaker Control). Modifying
these can seriously harm Xen's health.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoextends xm info with xen version and various others information.
vh249@arcadians.cl.cam.ac.uk [Wed, 31 Aug 2005 10:39:53 +0000 (10:39 +0000)]
extends xm info with xen version and various others information.
exports version hypercall to userspace

Signed-off-by: Vincent Hanquez <vincent@xensource.com>
20 years agoWait a little bit for tty to appear.
cl349@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 10:24:43 +0000 (10:24 +0000)]
Wait a little bit for tty to appear.
There is a race condition that occurs after xend creates a domain.
The console client might be running before the console daemon has
noticed the new domain and setup a pty for it.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agomerge?
cl349@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 09:32:52 +0000 (09:32 +0000)]
merge?

20 years agoThis patch adjusts the plan9 loader to start "physical" ram at
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 08:45:10 +0000 (08:45 +0000)]
This patch adjusts the plan9 loader to start "physical" ram at
0x80000000 virtual rather than the virtual address of the kernel
image (which is higher).

Tim Newsham

20 years agoAdd back missing part of XenLinux acpi/boot.c.
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 08:43:57 +0000 (08:43 +0000)]
Add back missing part of XenLinux acpi/boot.c.
This fixes bug #76 (IRQ sharing fails under load).
Signed-off-by: Natasha Jarymowycz <natasha@us.ibm.com>
20 years agoUse MIN instead of MAX to avoid reading more than fits into the buffer.
cl349@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 08:43:31 +0000 (08:43 +0000)]
Use MIN instead of MAX to avoid reading more than fits into the buffer.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agomerge?
kaf24@firebug.cl.cam.ac.uk [Wed, 31 Aug 2005 08:27:34 +0000 (08:27 +0000)]
merge?

20 years agoMerge.
adsharma@los-vmm.sc.intel.com [Tue, 30 Aug 2005 20:36:49 +0000 (13:36 -0700)]
Merge.

20 years agomerge?
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 20:03:51 +0000 (20:03 +0000)]
merge?

20 years agoUse watch to detect new domains and avoid polling for dead domains.
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 20:02:59 +0000 (20:02 +0000)]
Use watch to detect new domains and avoid polling for dead domains.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoRaise VIRQ_DOM_EXC when a domain is destroyed from dom0.
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 20:01:23 +0000 (20:01 +0000)]
Raise VIRQ_DOM_EXC when a domain is destroyed from dom0.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoRemainder of Intel/IBM joint TPM Virtualization implementation for Xen.
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 19:53:53 +0000 (11:53 -0800)]
Remainder of Intel/IBM joint TPM Virtualization implementation for Xen.

Signed-off-by: Vinnie Scarlata <vincent.r.scarlata@intel.com>
Signed-off-by: Joseph Cihula <joseph.cihula@intel.com>
Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoTPM front-end and back-end implementation, and configuration updates.
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 19:48:08 +0000 (11:48 -0800)]
TPM front-end and back-end implementation, and configuration updates.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
Signed-off-by: Kylene Hall <kjhall@us.ibm.com>
Signed-off-by: Mahadevan Gomathisankaran <gmdev@iastate.edu>
Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoDon't clear console entries.
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 19:47:39 +0000 (19:47 +0000)]
Don't clear console entries.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agovTPM manager - Intel/IBM joint TPM Virtualization implementation for Xen.
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 19:39:25 +0000 (11:39 -0800)]
vTPM manager - Intel/IBM joint TPM Virtualization implementation for Xen.
This is a software implementation for people without h/w vTPM support.

Signed-off-by: Vinnie Scarlata <vincent.r.scarlata@intel.com>
Signed-off-by: Joseph Cihula <joseph.cihula@intel.com>
Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoProvide xend support for setting up TPM front- and back-end interfaces.
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 19:28:26 +0000 (11:28 -0800)]
Provide xend support for setting up TPM front- and back-end interfaces.

Signed-off-by: Steven Hand <steven@xensource.com>
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
20 years agoDocumentation about how to use the virtual TPM implementation.
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 19:19:52 +0000 (11:19 -0800)]
Documentation about how to use the virtual TPM implementation.

Signed-off-by: Steven Hand <steven@xensource.com>
Signed-off-by: Stefan Berger <stefanb@us.ibm.com>
20 years agoUnify access to mpt using macros.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 17:53:49 +0000 (17:53 +0000)]
Unify access to mpt using macros.
Also some code cleanup to x86_64 fault.c

Signed-off-by: Xin Li <xin.b.li@intel.com>
20 years agoAdd some more EXPORT_SYMBOLs for Xen-defined stuff.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 17:09:43 +0000 (17:09 +0000)]
Add some more EXPORT_SYMBOLs for Xen-defined stuff.
Based on a patch from Ross Mcilroy.

20 years agoSome fixes to the performance counters code.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 17:00:26 +0000 (17:00 +0000)]
Some fixes to the performance counters code.
Signed-off-by: Ross McIlroy <mcilrorc@dcs.gla.ac.uk>
20 years agoWhen generating an exception frame for injection into a
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 16:53:40 +0000 (16:53 +0000)]
When generating an exception frame for injection into a
guest, that frame's base should be aligned on a 16-byte
boundary (as is done by the hardware).

Signed-off-by: Jan Beulich <jbeulich@novell.com>
20 years agoFix x86/64 Xen register and stack dumps so they don't
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 16:44:05 +0000 (16:44 +0000)]
Fix x86/64 Xen register and stack dumps so they don't
wrap on an 80-column display.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
20 years agoFix x86/64 pagetable initialisation to not waste several
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 16:19:07 +0000 (16:19 +0000)]
Fix x86/64 pagetable initialisation to not waste several
megabytes of memory.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agomerge?
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 16:15:27 +0000 (16:15 +0000)]
merge?

20 years agoNew console transport and update xenconsoled.
cl349@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 16:14:53 +0000 (16:14 +0000)]
New console transport and update xenconsoled.
Add a new console interface using a seperate shared page and event channel
instead of passing the console input/output over control messages.
Update xenconsoled to use the new console interface.
Make xenconsoled garbage collect dying domains by subscribing to
the domain exception virq.
Signed-off-by: Robert Read <robert@xensource.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoFix user-space four- and five-argument hypercalls on
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 14:14:31 +0000 (14:14 +0000)]
Fix user-space four- and five-argument hypercalls on
x86/64. This fixes domU building on x86/64 (bug #196).
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix memory_op domid parameter in dom0 tools.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 13:56:08 +0000 (13:56 +0000)]
Fix memory_op domid parameter in dom0 tools.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix build failure due to missing unistd.h
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 13:16:05 +0000 (13:16 +0000)]
Fix build failure due to missing unistd.h
Signed-off-by: Michal Ostrowski <mostrows@watson.ibm.com>
20 years agoCross-compilation fixes.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 13:15:20 +0000 (13:15 +0000)]
Cross-compilation fixes.
Signed-off-by: Michal Ostrowski <mostrows@watson.ibm.com>
20 years agoMake mbootpack work on big-endian systems.
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 13:12:47 +0000 (13:12 +0000)]
Make mbootpack work on big-endian systems.
Signed-off-by: Michal Ostrowski <mostrows@watson.ibm.com>
20 years agoReplace dom_mem_op hypercall with memory_op hypercall with
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 11:53:47 +0000 (11:53 +0000)]
Replace dom_mem_op hypercall with memory_op hypercall with
better interface. Fix "mem=" parameter on x86/64 boot.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoThis patch export symbols in xenbus that is used by netfront
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 09:46:33 +0000 (09:46 +0000)]
This patch export symbols in xenbus that is used by netfront
and blkfront so that vbd and vnif can be made to modules.

Signed-off-by: Xiaofeng Ling <xiaofeng.ling@intel.com>
20 years agoAdd CONFIG_BLK_DEV_UB as unconfigured item to x86/64 dom0
kaf24@firebug.cl.cam.ac.uk [Tue, 30 Aug 2005 08:47:51 +0000 (08:47 +0000)]
Add CONFIG_BLK_DEV_UB as unconfigured item to x86/64 dom0
defconfig file. Spotted by Yongkang You at Intel.

20 years agoFix save/restore (for SMP guests in particularl); the blkfront suspend code
shand@ubuntu.eng.hq.xensource.com [Tue, 30 Aug 2005 03:20:55 +0000 (19:20 -0800)]
Fix save/restore (for SMP guests in particularl);  the blkfront suspend code
was failing to unbind its irq handler correctly which (on SMP) caused the
guest to BUG() on resume when trying to request its callfunc_irq. Plus a
few small cosmetic changes.

Signed-off-by: Steven Hand <steven@xensource.com>
/bin/bash: :x: command not found

20 years agoSync Makefile with xen-unstable.hg. This should fix x86-64 bit build.
adsharma@los-vmm.sc.intel.com [Mon, 29 Aug 2005 23:05:29 +0000 (16:05 -0700)]
Sync Makefile with xen-unstable.hg. This should fix x86-64 bit build.

20 years agoMerge.
adsharma@los-vmm.sc.intel.com [Mon, 29 Aug 2005 22:31:51 +0000 (15:31 -0700)]
Merge.

20 years agoMerge.
adsharma@los-vmm.sc.intel.com [Mon, 29 Aug 2005 20:19:24 +0000 (13:19 -0700)]
Merge.

20 years agoSimplify and fix SMP irq and ipi teardown/setup on suspend/resume.
cl349@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 17:40:47 +0000 (17:40 +0000)]
Simplify and fix SMP irq and ipi teardown/setup on suspend/resume.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoUpdate suspend/resume functions for xenbus.
cl349@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 17:15:56 +0000 (17:15 +0000)]
Update suspend/resume functions for xenbus.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoOnly hotplug cpus and save/restore vcpu context on SMP builds.
cl349@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 17:14:08 +0000 (17:14 +0000)]
Only hotplug cpus and save/restore vcpu context on SMP builds.
The for_each_cpu_mask macro has a slightly odd behaviour in the
non-CONFIG_SMP case, making the code unexpectedly incorrect.
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoFix page_is_ram() start-of-day checks, and the watch routine
kaf24@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 14:53:38 +0000 (14:53 +0000)]
Fix page_is_ram() start-of-day checks, and the watch routine
inside balloon driver.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoFix "bogus" check by moving it inside the length check.
cl349@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 13:58:46 +0000 (13:58 +0000)]
Fix "bogus" check by moving it inside the length check.
Signed-off-by: Ryan Harper <ryanh@us.ibm.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoAllocate a request-id space per vbd in blkfront.
kaf24@firebug.cl.cam.ac.uk [Mon, 29 Aug 2005 13:56:53 +0000 (13:56 +0000)]
Allocate a request-id space per vbd in blkfront.
This should fix the assertions people have been seeing
in domU blkfront where we run out of IDs (because they
were being shared by multiple vbds).

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMinor (partial) build fix for 2.4.
shand@ubuntu.eng.hq.xensource.com [Mon, 29 Aug 2005 05:50:04 +0000 (21:50 -0800)]
Minor (partial) build fix for 2.4.

Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoFix nasty tools race between barking xu_autoreap() and xpopen3.wait() -
shand@ubuntu.eng.hq.xensource.com [Mon, 29 Aug 2005 05:40:36 +0000 (21:40 -0800)]
Fix nasty tools race between barking xu_autoreap() and xpopen3.wait() -
save/restore now seems more robust from a tools pov at least.

Signed-off-by: Steven Hand <steven@xensource.com>
20 years agoblkfront cleanup and retab.
kaf24@firebug.cl.cam.ac.uk [Sun, 28 Aug 2005 14:21:08 +0000 (14:21 +0000)]
blkfront cleanup and retab.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoUpdate __vmread()
lcy@vtsmp-build.sh.intel.com [Sun, 28 Aug 2005 06:55:37 +0000 (14:55 +0800)]
Update __vmread()
To make it not to break the stack.

Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
Signed-off-by: Chengyuan Li <chengyuan.li@intel.com>
20 years agoCONFIG_X86_PM_TIMER is a no-op option when building Linux
kaf24@firebug.cl.cam.ac.uk [Sat, 27 Aug 2005 11:49:35 +0000 (11:49 +0000)]
CONFIG_X86_PM_TIMER is a no-op option when building Linux
for Xen, but make sure it doesn't break the build.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoDon't build pcbios.c for xen/i386. No good can come of it.
kaf24@firebug.cl.cam.ac.uk [Sat, 27 Aug 2005 09:15:17 +0000 (09:15 +0000)]
Don't build pcbios.c for xen/i386. No good can come of it.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoMerge.
adsharma@los-vmm.sc.intel.com [Fri, 26 Aug 2005 20:47:16 +0000 (13:47 -0700)]
Merge.

20 years agoRefactor sync_lazy_execstate_cpu() into the more sensible
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 17:57:09 +0000 (17:57 +0000)]
Refactor sync_lazy_execstate_cpu() into the more sensible
sync_vcpu_execstate().

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoClean up and fix hypercall macros in XenLinux: we must assume that
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 17:42:34 +0000 (17:42 +0000)]
Clean up and fix hypercall macros in XenLinux: we must assume that
all register parameters are clobbered on return. Furthermore,
multicall arrays are clobbered on return from a multicall (only the
result field can be trusted!).

These restrictions are enforced by debug builds of Xen in the
following ways:
 1. On completion of a multicall, the call array is overwritten
    with garbage before writing in the result fields.
 2. On completion of a hypercall, all parameter registers are
    overwritten with garbage.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoRemove the unused ia64 patch directory.
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 13:06:49 +0000 (13:06 +0000)]
Remove the unused ia64 patch directory.
Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agoThe patched attached enables x86_64 xenlinux with "late pin, early
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 11:02:14 +0000 (11:02 +0000)]
The patched attached enables x86_64 xenlinux with "late pin, early
unpin", which is already implemented for x86_32. Since we now only pin
the root rather than any of the other levels, the overall performance
became better especially with workloads that require heavy memory
management operations.

On 8-way x86_64 xenlinux (dom0) the kernel build was improved by about
10% (using make -j32). Even a small setup like a UP HT system, I see
about 3% performance gain with kernel build (make -j4).

Lmbench also shows improvements in fork/exec/sh:
Processor, Processes - times in microseconds - smaller is better
--------------------------------------------------------------------
Host    OS  Mhz   null null      open slct sig  sig  fork exec sh =20
                  call  I/O stat clos TCP  inst hndl proc proc proc
--------- ------------- ---- ---- ---- ---- ---- ---- ---- ---- ----=20
Linux 2.6.12- 3786 1.13 1.36 3.93 6.04 10.5 1.43 4.33 536. 1446 3614
Linux 2.6.12- 3786 1.13 1.36 3.91 6.03 10.4 1.44 4.38 346. 1050 2831

Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
20 years agoFix GDT allocation for SMP guests.
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 11:00:14 +0000 (11:00 +0000)]
Fix GDT allocation for SMP guests.

Signed-off-by: Jun Nakajima <jun.nakajima@intel.com>
20 years agoVnet update.
cl349@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 10:52:53 +0000 (10:52 +0000)]
Vnet update.
Fix the tools/vnet makefiles so that vnets will compile
and includes updates to the vnet implementation.
Vnet ids are increased to 128 bits. The vnet module will
insmod whether xen-br0 exists or not.
Signed-off-by: Mike Wray <mike.wray@hp.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoFix vnets in xm.
cl349@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 10:51:10 +0000 (10:51 +0000)]
Fix vnets in xm.
The removal of 'xm call' to call a xend interface function directly
made the xend vnet functions inaccessible. This patch adds those functions
to xm and fixes the persistence in XmVnet to use xenstore.
Signed-off-by: Mike Wray <mike.wray@hp.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
20 years agoThe patch removes broken, and very complicated malloc in
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 10:35:36 +0000 (10:35 +0000)]
The patch removes broken, and very complicated malloc in
favour of much simpler (and working) Xen's allocator
(xmalloc by Rusty).

Signed-off-by: Grzegorz Milos <gm281@cam.ac.uk>
20 years agoClean up and fix domain_pause and friends. Synchronous
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 09:29:54 +0000 (09:29 +0000)]
Clean up and fix domain_pause and friends. Synchronous
pause should not only wait for the running flag to clear,
but also for the scheduler lock to be released. Also get
rid of some unused sync_lazy_execstate functions.

Signed-off-by: Keir Fraser <keir@xensource.com>
20 years agomerge?
kaf24@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 09:05:43 +0000 (09:05 +0000)]
merge?

20 years agoUpdate xentop to use new name for Xen version function.
cl349@firebug.cl.cam.ac.uk [Fri, 26 Aug 2005 08:50:31 +0000 (08:50 +0000)]
Update xentop to use new name for Xen version function.
Signed-off-by: Josh Triplett <josht@us.ibm.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>