# HG changeset patch
# User Alex Williamson <alex.williamson@xxxxxx>
# Date 1175875710 21600
# Node ID eb3e430242ac8732d665b0a1f929bc324bc9fcdd
# Parent 9ec7dadc98ba1eadca8fe9c38ae6d6dd41d6d29d
# Parent ef33477324f669037a1b35b719fa0a01e98ec2f3
merge with xen-unstable.hg
---
tools/libxc/xc_hvm_restore.c | 360 ---
tools/libxc/xc_linux_restore.c | 962
--------
.hgignore | 1
docs/Docs.mk | 1
docs/xen-api/Makefile | 3
docs/xen-api/coversheet.tex | 24
docs/xen-api/presentation.tex | 2
docs/xen-api/vm-lifecycle.tex | 2
docs/xen-api/wire-protocol.tex | 2
docs/xen-api/xenapi-coversheet.tex | 16
docs/xen-api/xenapi-datamodel-graph.dot | 55
docs/xen-api/xenapi-datamodel.tex | 261 --
docs/xen-api/xenapi.tex | 10
extras/mini-os/xenbus/xenbus.c | 2
linux-2.6-xen-sparse/arch/i386/kernel/head-xen.S | 8
linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c | 2
linux-2.6-xen-sparse/arch/i386/mach-xen/setup.c | 4
linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c | 12
linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S | 61
linux-2.6-xen-sparse/arch/x86_64/kernel/head-xen.S | 11
linux-2.6-xen-sparse/arch/x86_64/mm/init-xen.c | 4
linux-2.6-xen-sparse/drivers/xen/Kconfig | 13
linux-2.6-xen-sparse/drivers/xen/balloon/balloon.c | 38
linux-2.6-xen-sparse/drivers/xen/balloon/sysfs.c | 6
linux-2.6-xen-sparse/drivers/xen/blkfront/blkfront.c | 5
linux-2.6-xen-sparse/drivers/xen/core/machine_kexec.c | 1
linux-2.6-xen-sparse/drivers/xen/core/reboot.c | 4
linux-2.6-xen-sparse/drivers/xen/core/xen_proc.c | 5
linux-2.6-xen-sparse/drivers/xen/fbfront/xenfb.c | 1
linux-2.6-xen-sparse/drivers/xen/netfront/netfront.c | 6
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypercall.h | 8
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypervisor.h | 16
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h | 2
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/page.h | 14
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-2level.h | 32
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-3level.h | 60
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable.h | 107
linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h | 2
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/hypercall.h | 8
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h | 2
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/page.h | 12
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/pgtable.h | 163 -
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/setup_arch_post.h | 4
linux-2.6-xen-sparse/include/xen/hvm.h | 1
linux-2.6-xen-sparse/include/xen/hypercall.h | 24
tools/ioemu/hw/xen_platform.c | 8
tools/ioemu/target-i386-dm/exec-dm.c | 10
tools/ioemu/target-i386-dm/helper2.c | 70
tools/ioemu/vl.c | 84
tools/ioemu/xenstore.c | 14
tools/libxc/Makefile | 4
tools/libxc/ia64/xc_ia64_linux_restore.c | 25
tools/libxc/xc_domain_restore.c | 1086
++++++++++
tools/libxc/xc_hvm_save.c | 407 +--
tools/libxc/xenguest.h | 27
tools/libxc/xg_private.c | 15
tools/libxen/include/xen_host.h | 7
tools/libxen/include/xen_sr.h | 36
tools/libxen/include/xen_vdi.h | 15
tools/libxen/src/xen_common.c | 45
tools/libxen/src/xen_host.c | 12
tools/libxen/src/xen_sr.c | 105
tools/libxen/src/xen_vdi.c | 33
tools/libxen/test/test_bindings.c | 43
tools/python/xen/xend/XendAPI.py | 132 -
tools/python/xen/xend/XendCheckpoint.py | 21
tools/python/xen/xend/XendLocalStorageRepo.py | 17
tools/python/xen/xend/XendNode.py | 39
tools/python/xen/xend/XendPBD.py | 79
tools/python/xen/xend/XendQCoWStorageRepo.py | 23
tools/python/xen/xend/XendStateStore.py | 2
tools/python/xen/xend/XendStorageRepository.py | 13
tools/python/xen/xend/XendTask.py | 10
tools/python/xen/xend/image.py | 9
tools/python/xen/xm/main.py | 2
tools/xcutils/xc_restore.c | 34
unmodified_drivers/linux-2.6/Makefile | 1
unmodified_drivers/linux-2.6/balloon/Kbuild | 9
unmodified_drivers/linux-2.6/balloon/Makefile | 3
unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h | 22
unmodified_drivers/linux-2.6/mkbuildtree | 3
unmodified_drivers/linux-2.6/overrides.mk | 3
unmodified_drivers/linux-2.6/platform-pci/platform-pci.c | 10
unmodified_drivers/linux-2.6/platform-pci/xen_support.c | 9
xen/arch/x86/hvm/hvm.c | 113 -
xen/arch/x86/hvm/io.c | 227 +-
xen/arch/x86/hvm/irq.c | 15
xen/arch/x86/hvm/platform.c | 32
xen/arch/x86/hvm/svm/svm.c | 143 -
xen/arch/x86/hvm/svm/vmcb.c | 8
xen/arch/x86/hvm/vmx/vmx.c | 18
xen/common/domain.c | 8
xen/common/memory.c | 11
xen/common/schedule.c | 4
xen/drivers/char/ns16550.c | 2
xen/include/asm-ia64/xentypes.h | 5
xen/include/asm-powerpc/types.h | 4
xen/include/asm-x86/event.h | 7
xen/include/asm-x86/hvm/hvm.h | 14
xen/include/asm-x86/hvm/io.h | 3
xen/include/asm-x86/hvm/support.h | 5
xen/include/asm-x86/types.h | 4
xen/include/public/acm.h | 16
xen/include/public/foreign/Makefile | 2
xen/include/public/hvm/hvm_op.h | 8
xen/include/public/hvm/ioreq.h | 1
xen/include/xen/types.h | 2
107 files changed, 2671 insertions(+), 2805 deletions(-)
diff -r 9ec7dadc98ba -r eb3e430242ac .hgignore
--- a/.hgignore Fri Apr 06 10:06:30 2007 -0600
+++ b/.hgignore Fri Apr 06 10:08:30 2007 -0600
@@ -14,6 +14,7 @@
.*\.orig$
.*\.rej$
.*/a\.out$
+.*/Modules\.symvers$
.*/cscope\..*$
^cscope.*$
^[^/]*\.bz2$
diff -r 9ec7dadc98ba -r eb3e430242ac docs/Docs.mk
--- a/docs/Docs.mk Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/Docs.mk Fri Apr 06 10:08:30 2007 -0600
@@ -6,6 +6,7 @@ DOXYGEN := doxygen
DOXYGEN := doxygen
POD2MAN := pod2man
DOT := dot
+NEATO := neato
pkgdocdir := /usr/share/doc/xen
mandir := /usr/share/man
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/Makefile
--- a/docs/xen-api/Makefile Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/Makefile Fri Apr 06 10:08:30 2007 -0600
@@ -37,6 +37,9 @@ xenapi.dvi: $(TEX) $(EPS) $(EPSDOT)
%.eps: %.dot
$(DOT) -Tps $< >$@
+xenapi-datamodel-graph.eps: xenapi-datamodel-graph.dot
+ $(NEATO) -Goverlap=false -Tps $< >$@
+
.PHONY: clean
clean:
rm -f *.pdf *.ps *.dvi *.aux *.log $(EPSDOT)
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/coversheet.tex
--- a/docs/xen-api/coversheet.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/coversheet.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
@@ -20,7 +20,7 @@
\begin{center}
\resizebox{8cm}{!}{\includegraphics{\coversheetlogo}}
-\vspace{3cm}
+\vspace{2cm}
\begin{Huge}
\doctitle{}
@@ -37,9 +37,27 @@ Date: \datestring{}
\begin{tabular}{rl}
\docauthors{}
\end{tabular}
-
\end{Large}
\end{center}
+\vspace{.5cm}
+\begin{large}
+\textbf{Contributors:}
+\begin{multicols}{2}{
+Stefan Berger, IBM \\
+Daniel Berrang\'e, Red Hat \\
+Gareth Bestor, IBM \\
+Hollis Blanchard, IBM \\
+Mike Day, IBM \\
+Jim Fehlig, Novell \\
+Jon Harrop, XenSource \\
+Vincent Hanquez, XenSource \\
+John Levon, Sun Microsystems \\
+Jon Ludlam, XenSource \\
+Alastair Tse, XenSource \\
+Daniel Veillard, Red Hat \\
+Tom Wilkie, University of Cambridge}
+\end{multicols}
+\end{large}
\vfill
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/presentation.tex
--- a/docs/xen-api/presentation.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/presentation.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/vm-lifecycle.tex
--- a/docs/xen-api/vm-lifecycle.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/vm-lifecycle.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/wire-protocol.tex
--- a/docs/xen-api/wire-protocol.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/wire-protocol.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/xenapi-coversheet.tex
--- a/docs/xen-api/xenapi-coversheet.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/xenapi-coversheet.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
@@ -12,26 +12,24 @@
%
%% Document title
-\newcommand{\doctitle}{Xen Management API Draft}
+\newcommand{\doctitle}{Xen Management API}
\newcommand{\coversheetlogo}{xen.eps}
%% Document date
-\newcommand{\datestring}{25th August 2006}
+\newcommand{\datestring}{5th April 2007}
-\newcommand{\releasestatement}{Open Preview Release\\Comments are welcome!}
+\newcommand{\releasestatement}{Candidate for Release\\Comments are welcome!}
%% Document revision
-\newcommand{\revstring}{API Revision 0.4.3 (Draft for discussion)}
+\newcommand{\revstring}{API Revision 0.9.0}
%% Document authors
\newcommand{\docauthors}{
Ewan Mellor: & {\tt ewan@xxxxxxxxxxxxx} \\
Richard Sharp: & {\tt richard.sharp@xxxxxxxxxxxxx} \\
-David Scott: & {\tt david.scott@xxxxxxxxxxxxx} \\
-Jon Harrop: & {\tt jon.harrop@xxxxxxxxxxxxx}
-}
-\newcommand{\legalnotice}{Copyright \copyright{} 2006 XenSource, Inc.\\ \\
+David Scott: & {\tt david.scott@xxxxxxxxxxxxx}}
+\newcommand{\legalnotice}{Copyright \copyright{} 2006-2007 XenSource, Inc.\\ \\
Permission is granted to copy, distribute and/or modify this document under
the terms of the GNU Free Documentation License, Version 1.2 or any later
version published by the Free Software Foundation; with no Invariant Sections,
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/xenapi-datamodel-graph.dot
--- a/docs/xen-api/xenapi-datamodel-graph.dot Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/xenapi-datamodel-graph.dot Fri Apr 06 10:08:30 2007 -0600
@@ -1,18 +1,39 @@ digraph g{
-digraph g{
-node [ shape=box ]; session [ URL="session.html" ] task [ URL="task.html" ] VM
[ URL="VM.html" ] host [ URL="host.html" ] host_cpu [ URL="host_cpu.html" ]
network [ URL="network.html" ] VIF [ URL="VIF.html" ] PIF [ URL="PIF.html" ] SR
[ URL="SR.html" ] VDI [ URL="VDI.html" ] VBD [ URL="VBD.html" ] VTPM [
URL="VTPM.html" ] console [ URL="console.html" ] user [ URL="user.html" ] debug
[ URL="debug.html" ];
-session -> host [ label="this_host(1)" ]
-session -> user [ label="this_user(1)" ]
-host -> VM [ color="blue", arrowhead="crow", arrowtail="none" ]
-host -> host_cpu [ color="blue", arrowhead="crow", arrowtail="none" ]
-VIF -> VM [ color="blue", arrowhead="none", arrowtail="crow" ]
-VIF -> network [ color="blue", arrowhead="none", arrowtail="crow" ]
-PIF -> host [ color="blue", arrowhead="none", arrowtail="crow" ]
-PIF -> network [ color="blue", arrowhead="none", arrowtail="crow" ]
-SR -> VDI [ color="blue", arrowhead="crow", arrowtail="none" ]
-VDI -> VBD [ color="blue", arrowhead="crow", arrowtail="none" ]
-VDI -> VDI [ color="blue", arrowhead="none", arrowtail="crow" ]
-VBD -> VM [ color="blue", arrowhead="none", arrowtail="crow" ]
-VTPM -> VM [ label="backend(1)" ]
-VTPM -> VM [ color="blue", arrowhead="none", arrowtail="crow" ]
-console -> VM [ color="blue", arrowhead="none", arrowtail="crow" ]
+#
+# Copyright (c) 2006-2007 XenSource, Inc.
+#
+# Permission is granted to copy, distribute and/or modify this document under
+# the terms of the GNU Free Documentation License, Version 1.2 or any later
+# version published by the Free Software Foundation; with no Invariant
+# Sections, no Front-Cover Texts and no Back-Cover Texts. A copy of the
+# license is included in the section entitled
+# "GNU Free Documentation License" or the file fdl.tex.
+#
+
+digraph "Xen-API Class Diagram" {
+fontname="Verdana";
+
+node [ shape=box ]; session VM host network VIF PIF SR VDI VBD PBD user;
+node [shape=ellipse]; PIF_metrics VIF_metrics VM_metrics VBD_metrics
PBD_metrics VM_guest_metrics host_metrics;
+node [shape=box]; host_cpu console
+session -> host [ arrowhead="none" ]
+session -> user [ arrowhead="none" ]
+VM -> VM_metrics [ arrowhead="none" ]
+VM -> VM_guest_metrics [ arrowhead="none" ]
+VM -> console [ arrowhead="crow" ]
+host -> PBD [ arrowhead="crow", arrowtail="none" ]
+host -> host_metrics [ arrowhead="none" ]
+host -> host_cpu [ arrowhead="none" ]
+VIF -> VM [ arrowhead="none", arrowtail="crow" ]
+VIF -> network [ arrowhead="none", arrowtail="crow" ]
+VIF -> VIF_metrics [ arrowhead="none" ]
+PIF -> host [ arrowhead="none", arrowtail="crow" ]
+PIF -> network [ arrowhead="none", arrowtail="crow" ]
+PIF -> PIF_metrics [ arrowhead="none" ]
+SR -> PBD [ arrowhead="crow", arrowtail="none" ]
+PBD -> PBD_metrics [ arrowhead="none" ]
+SR -> VDI [ arrowhead="crow", arrowtail="none" ]
+VDI -> VBD [ arrowhead="crow", arrowtail="none" ]
+VBD -> VM [ arrowhead="none", arrowtail="crow" ]
+VTPM -> VM [ arrowhead="none", arrowtail="crow" ]
+VBD -> VBD_metrics [ arrowhead="none" ]
}
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/xenapi-datamodel.tex
--- a/docs/xen-api/xenapi-datamodel.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/xenapi-datamodel.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
@@ -512,7 +512,6 @@ Quals & Field & Type & Description \\
$\mathit{RO}_\mathit{run}$ & {\tt progress} & int & if the task is still
pending, this field contains the estimated percentage complete (0-100). If task
has completed (successfully or unsuccessfully) this should be 100. \\
$\mathit{RO}_\mathit{run}$ & {\tt type} & string & if the task has completed
successfully, this field contains the type of the encoded result (i.e. name of
the class whose reference is in the result field). Undefined otherwise. \\
$\mathit{RO}_\mathit{run}$ & {\tt result} & string & if the task has
completed successfully, this field contains the result value (either Void or an
object reference). Undefined otherwise. \\
-$\mathit{RO}_\mathit{run}$ & {\tt error\_code} & int & if the task has
failed, this field contains the error code. Undefined otherwise. \\
$\mathit{RO}_\mathit{run}$ & {\tt error\_info} & string Set & if the task has
failed, this field contains the set of associated error strings. Undefined
otherwise. \\
$\mathit{RO}_\mathit{run}$ & {\tt allowed\_operations} &
(task\_allowed\_operations) Set & Operations allowed on this task \\
\hline
@@ -826,38 +825,6 @@ Get the result field of the given task.
\noindent {\bf Return Type:}
{\tt
string
-}
-
-
-value of the field
-\vspace{0.3cm}
-\vspace{0.3cm}
-\vspace{0.3cm}
-\subsubsection{RPC name:~get\_error\_code}
-
-{\bf Overview:}
-Get the error\_code field of the given task.
-
- \noindent {\bf Signature:}
-\begin{verbatim} int get_error_code (session_id s, task ref self)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt task ref } & self & reference to the object \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-int
}
@@ -4952,6 +4919,27 @@ dmesg string
\vspace{0.3cm}
\vspace{0.3cm}
\vspace{0.3cm}
+\subsubsection{RPC name:~list\_methods}
+
+{\bf Overview:}
+List all supported methods.
+
+ \noindent {\bf Signature:}
+\begin{verbatim} (string Set) list_methods (session_id s)\end{verbatim}
+
+
+\vspace{0.3cm}
+
+ \noindent {\bf Return Type:}
+{\tt
+string Set
+}
+
+
+The name of every supported method.
+\vspace{0.3cm}
+\vspace{0.3cm}
+\vspace{0.3cm}
\subsubsection{RPC name:~get\_all}
{\bf Overview:}
@@ -9190,51 +9178,35 @@ Quals & Field & Type & Description \\
$\mathit{RO}_\mathit{run}$ & {\tt physical\_utilisation} & int & physical
space currently utilised on this storage repository (in bytes). Note that for
sparse disk formats, physical\_utilisation may be less than virtual\_allocation
\\
$\mathit{RO}_\mathit{ins}$ & {\tt physical\_size} & int & total physical size
of the repository (in bytes) \\
$\mathit{RO}_\mathit{ins}$ & {\tt type} & string & type of the storage
repository \\
-$\mathit{RO}_\mathit{ins}$ & {\tt location} & string & a string that uniquely
determines the location of the storage repository; the format of this string
depends on the repository's type \\
+$\mathit{RO}_\mathit{ins}$ & {\tt content\_type} & string & the type of the
SR's content, if required (e.g. ISOs) \\
\hline
\end{longtable}
\subsection{RPCs associated with class: SR}
-\subsubsection{RPC name:~clone}
-
-{\bf Overview:}
-Take an exact copy of the Storage Repository;
- the cloned storage repository has the same type as its parent
-
- \noindent {\bf Signature:}
-\begin{verbatim} (SR ref) clone (session_id s, SR ref sr, string loc, string
name)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt SR ref } & sr & The Storage Repository to clone \\ \hline
-
-{\tt string } & loc & The location string that defines where the new storage
repository will be located \\ \hline
-
-{\tt string } & name & The name of the new storage repository \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-SR ref
-}
-
-
-The ID of the newly created Storage Repository.
+\subsubsection{RPC name:~get\_supported\_types}
+
+{\bf Overview:}
+Return a set of all the SR types supported by the system.
+
+ \noindent {\bf Signature:}
+\begin{verbatim} (string Set) get_supported_types (session_id s)\end{verbatim}
+
+
+\vspace{0.3cm}
+
+ \noindent {\bf Return Type:}
+{\tt
+string Set
+}
+
+
+the supported SR types
\vspace{0.3cm}
\vspace{0.3cm}
\vspace{0.3cm}
\subsubsection{RPC name:~get\_all}
{\bf Overview:}
-Return a list of all the Storage Repositories known to the system
+Return a list of all the SRs known to the system.
\noindent {\bf Signature:}
\begin{verbatim} ((SR ref) Set) get_all (session_id s)\end{verbatim}
@@ -9248,7 +9220,7 @@ Return a list of all the Storage Reposit
}
-A list of all the IDs of all the Storage Repositories
+references to all objects
\vspace{0.3cm}
\vspace{0.3cm}
\vspace{0.3cm}
@@ -9608,13 +9580,13 @@ value of the field
\vspace{0.3cm}
\vspace{0.3cm}
\vspace{0.3cm}
-\subsubsection{RPC name:~get\_location}
-
-{\bf Overview:}
-Get the location field of the given SR.
-
- \noindent {\bf Signature:}
-\begin{verbatim} string get_location (session_id s, SR ref self)\end{verbatim}
+\subsubsection{RPC name:~get\_content\_type}
+
+{\bf Overview:}
+Get the content\_type field of the given SR.
+
+ \noindent {\bf Signature:}
+\begin{verbatim} string get_content_type (session_id s, SR ref
self)\end{verbatim}
\noindent{\bf Arguments:}
@@ -9637,70 +9609,6 @@ string
value of the field
-\vspace{0.3cm}
-\vspace{0.3cm}
-\vspace{0.3cm}
-\subsubsection{RPC name:~create}
-
-{\bf Overview:}
-Create a new SR instance, and return its handle.
-
- \noindent {\bf Signature:}
-\begin{verbatim} (SR ref) create (session_id s, SR record args)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt SR record } & args & All constructor arguments \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-SR ref
-}
-
-
-reference to the newly created object
-\vspace{0.3cm}
-\vspace{0.3cm}
-\vspace{0.3cm}
-\subsubsection{RPC name:~destroy}
-
-{\bf Overview:}
-Destroy the specified SR instance.
-
- \noindent {\bf Signature:}
-\begin{verbatim} void destroy (session_id s, SR ref self)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt SR ref } & self & reference to the object \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-void
-}
-
-
-
\vspace{0.3cm}
\vspace{0.3cm}
\vspace{0.3cm}
@@ -9828,73 +9736,6 @@ Quals & Field & Type & Description \\
\hline
\end{longtable}
\subsection{RPCs associated with class: VDI}
-\subsubsection{RPC name:~snapshot}
-
-{\bf Overview:}
-Take an exact copy of the VDI; the snapshot lives in the same Storage
-Repository as its parent.
-
- \noindent {\bf Signature:}
-\begin{verbatim} (VDI ref) snapshot (session_id s, VDI ref vdi)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt VDI ref } & vdi & The VDI to snapshot \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-VDI ref
-}
-
-
-The ID of the newly created VDI.
-\vspace{0.3cm}
-\vspace{0.3cm}
-\vspace{0.3cm}
-\subsubsection{RPC name:~resize}
-
-{\bf Overview:}
-Resize the vdi to the size.
-
- \noindent {\bf Signature:}
-\begin{verbatim} void resize (session_id s, VDI ref vdi, int
size)\end{verbatim}
-
-
-\noindent{\bf Arguments:}
-
-
-\vspace{0.3cm}
-\begin{tabular}{|c|c|p{7cm}|}
- \hline
-{\bf type} & {\bf name} & {\bf description} \\ \hline
-{\tt VDI ref } & vdi & The VDI to resize \\ \hline
-
-{\tt int } & size & The new size of the VDI \\ \hline
-
-\end{tabular}
-
-\vspace{0.3cm}
-
- \noindent {\bf Return Type:}
-{\tt
-void
-}
-
-
-
-\vspace{0.3cm}
-\vspace{0.3cm}
-\vspace{0.3cm}
\subsubsection{RPC name:~get\_all}
{\bf Overview:}
@@ -11988,7 +11829,7 @@ Quals & Field & Type & Description \\
$\mathit{RO}_\mathit{run}$ & {\tt uuid} & string & unique identifier/object
reference \\
$\mathit{RO}_\mathit{ins}$ & {\tt host} & host ref & physical machine on
which the pbd is available \\
$\mathit{RO}_\mathit{ins}$ & {\tt SR} & SR ref & the storage repository that
the pbd realises \\
-$\mathit{RO}_\mathit{ins}$ & {\tt device\_config} & (string $\rightarrow$
string) Map & a config string that is provided to the host's SR-backend-driver
\\
+$\mathit{RO}_\mathit{ins}$ & {\tt device\_config} & (string $\rightarrow$
string) Map & a config string to string map that is provided to the host's
SR-backend-driver \\
$\mathit{RO}_\mathit{run}$ & {\tt currently\_attached} & bool & is the SR
currently attached on this host? \\
\hline
\end{longtable}
diff -r 9ec7dadc98ba -r eb3e430242ac docs/xen-api/xenapi.tex
--- a/docs/xen-api/xenapi.tex Fri Apr 06 10:06:30 2007 -0600
+++ b/docs/xen-api/xenapi.tex Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
%
-% Copyright (c) 2006 XenSource, Inc.
+% Copyright (c) 2006-2007 XenSource, Inc.
%
% Permission is granted to copy, distribute and/or modify this document under
% the terms of the GNU Free Documentation License, Version 1.2 or any later
@@ -17,6 +17,7 @@
\usepackage{graphics}
\usepackage{longtable}
\usepackage{fancyhdr}
+\usepackage{multicol}
\setlength\topskip{0cm}
\setlength\topmargin{0cm}
@@ -40,17 +41,10 @@ remotely configuring and controlling vir
remotely configuring and controlling virtualised guests running on a
Xen-enabled host.
-~
-
-{\bf \large This document is an early draft for discussion purposes only.}
-
-~
-
\input{presentation}
\include{wire-protocol}
\include{vm-lifecycle}
-\include{todo}
\include{xenapi-datamodel}
\include{fdl}
diff -r 9ec7dadc98ba -r eb3e430242ac extras/mini-os/xenbus/xenbus.c
--- a/extras/mini-os/xenbus/xenbus.c Fri Apr 06 10:06:30 2007 -0600
+++ b/extras/mini-os/xenbus/xenbus.c Fri Apr 06 10:08:30 2007 -0600
@@ -178,6 +178,7 @@ static void release_xenbus_id(int id)
{
BUG_ON(!req_info[id].in_use);
spin_lock(&req_lock);
+ req_info[id].in_use = 0;
nr_live_reqs--;
req_info[id].in_use = 0;
if (nr_live_reqs == NR_REQS - 1)
@@ -335,6 +336,7 @@ xenbus_msg_reply(int type,
xb_write(type, id, trans, io, nr_reqs);
schedule();
+ remove_waiter(w);
wake(current);
rep = req_info[id].reply;
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/i386/kernel/head-xen.S
--- a/linux-2.6-xen-sparse/arch/i386/kernel/head-xen.S Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/head-xen.S Fri Apr 06 10:08:30
2007 -0600
@@ -142,7 +142,7 @@ ENTRY(cpu_gdt_table)
.quad 0x0000000000000000 /* 0xf0 - unused */
.quad 0x0000000000000000 /* 0xf8 - GDT entry 31: double-fault
TSS */
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
/*
* __xen_guest information
*/
@@ -180,18 +180,18 @@ ENTRY(cpu_gdt_table)
#endif
.ascii ",LOADER=generic"
.byte 0
-#endif /* CONFIG_XEN_COMPAT_030002 */
+#endif /* CONFIG_XEN_COMPAT <= 0x030002 */
ELFNOTE(Xen, XEN_ELFNOTE_GUEST_OS, .asciz, "linux")
ELFNOTE(Xen, XEN_ELFNOTE_GUEST_VERSION, .asciz, "2.6")
ELFNOTE(Xen, XEN_ELFNOTE_XEN_VERSION, .asciz, "xen-3.0")
ELFNOTE(Xen, XEN_ELFNOTE_VIRT_BASE, .long, __PAGE_OFFSET)
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET, .long, __PAGE_OFFSET)
#else
ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET, .long, 0)
-#endif /* !CONFIG_XEN_COMPAT_030002 */
+#endif
ELFNOTE(Xen, XEN_ELFNOTE_ENTRY, .long, startup_32)
ELFNOTE(Xen, XEN_ELFNOTE_HYPERCALL_PAGE, .long, hypercall_page)
ELFNOTE(Xen, XEN_ELFNOTE_HV_START_LOW, .long, HYPERVISOR_VIRT_START)
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c
--- a/linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/i386/kernel/time-xen.c Fri Apr 06 10:08:30
2007 -0600
@@ -1010,7 +1010,7 @@ static void stop_hz_timer(void)
singleshot.timeout_abs_ns = jiffies_to_st(j);
singleshot.flags = 0;
rc = HYPERVISOR_vcpu_op(VCPUOP_set_singleshot_timer, cpu, &singleshot);
-#ifdef CONFIG_XEN_COMPAT_030004
+#if CONFIG_XEN_COMPAT <= 0x030004
if (rc) {
BUG_ON(rc != -ENOSYS);
rc = HYPERVISOR_set_timer_op(singleshot.timeout_abs_ns);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/i386/mach-xen/setup.c
--- a/linux-2.6-xen-sparse/arch/i386/mach-xen/setup.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/i386/mach-xen/setup.c Fri Apr 06 10:08:30
2007 -0600
@@ -114,7 +114,7 @@ void __init machine_specific_arch_setup(
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &event);
if (ret == 0)
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &failsafe);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret == -ENOSYS)
ret = HYPERVISOR_set_callbacks(
event.address.cs, event.address.eip,
@@ -123,7 +123,7 @@ void __init machine_specific_arch_setup(
BUG_ON(ret);
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &nmi_cb);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret == -ENOSYS) {
static struct xennmi_callback __initdata cb = {
.handler_address = (unsigned long)nmi
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c
--- a/linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/i386/mm/hypervisor.c Fri Apr 06 10:08:30
2007 -0600
@@ -303,7 +303,7 @@ int xen_create_contiguous_region(
set_phys_to_machine((__pa(vstart)>>PAGE_SHIFT)+i,
INVALID_P2M_ENTRY);
}
- if (HYPERVISOR_multicall(cr_mcl, i))
+ if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
BUG();
/* 2. Get a new contiguous memory extent. */
@@ -312,7 +312,7 @@ int xen_create_contiguous_region(
success = (exchange.nr_exchanged == (1UL << order));
BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
BUG_ON(success && (rc != 0));
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
/* Compatibility when XENMEM_exchange is unsupported. */
if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
@@ -342,7 +342,7 @@ int xen_create_contiguous_region(
cr_mcl[i - 1].args[MULTI_UVMFLAGS_INDEX] = order
? UVMF_TLB_FLUSH|UVMF_ALL
: UVMF_INVLPG|UVMF_ALL;
- if (HYPERVISOR_multicall(cr_mcl, i))
+ if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
BUG();
if (success)
@@ -400,7 +400,7 @@ void xen_destroy_contiguous_region(unsig
INVALID_P2M_ENTRY);
out_frames[i] = (__pa(vstart) >> PAGE_SHIFT) + i;
}
- if (HYPERVISOR_multicall(cr_mcl, i))
+ if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
BUG();
/* 3. Do the exchange for non-contiguous MFNs. */
@@ -408,7 +408,7 @@ void xen_destroy_contiguous_region(unsig
success = (exchange.nr_exchanged == 1);
BUG_ON(!success && ((exchange.nr_exchanged != 0) || (rc == 0)));
BUG_ON(success && (rc != 0));
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
/* Compatibility when XENMEM_exchange is unsupported. */
if (HYPERVISOR_memory_op(XENMEM_decrease_reservation,
@@ -432,7 +432,7 @@ void xen_destroy_contiguous_region(unsig
cr_mcl[i - 1].args[MULTI_UVMFLAGS_INDEX] = order
? UVMF_TLB_FLUSH|UVMF_ALL
: UVMF_INVLPG|UVMF_ALL;
- if (HYPERVISOR_multicall(cr_mcl, i))
+ if (HYPERVISOR_multicall_check(cr_mcl, i, NULL))
BUG();
balloon_unlock(flags);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S
--- a/linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/arch/x86_64/kernel/entry-xen.S Fri Apr 06
10:08:30 2007 -0600
@@ -201,6 +201,19 @@ END(ret_from_fork)
END(ret_from_fork)
/*
+ * initial frame state for interrupts and exceptions
+ */
+ .macro _frame ref
+ CFI_STARTPROC simple
+ CFI_DEF_CFA rsp,SS+8-\ref
+ /*CFI_REL_OFFSET ss,SS-\ref*/
+ CFI_REL_OFFSET rsp,RSP-\ref
+ /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/
+ /*CFI_REL_OFFSET cs,CS-\ref*/
+ CFI_REL_OFFSET rip,RIP-\ref
+ .endm
+
+/*
* System call entry. Upto 6 arguments in registers are supported.
*
* SYSCALL does not save anything on the stack and does not change the
@@ -232,10 +245,7 @@ END(ret_from_fork)
*/
ENTRY(system_call)
- CFI_STARTPROC simple
- CFI_DEF_CFA rsp,PDA_STACKOFFSET
- CFI_REGISTER rip,rcx
- /*CFI_REGISTER rflags,r11*/
+ _frame (RIP-0x10)
SAVE_ARGS -8,0
movq %rax,ORIG_RAX-ARGOFFSET(%rsp)
/*
@@ -273,9 +283,7 @@ sysret_check:
*/
TRACE_IRQS_ON
XEN_UNBLOCK_EVENTS(%rsi)
- CFI_REGISTER rip,rcx
RESTORE_ARGS 0,8,0
- /*CFI_REGISTER rflags,r11*/
HYPERVISOR_IRET VGCF_IN_SYSCALL
/* Handle reschedules */
@@ -505,24 +513,16 @@ ENTRY(stub_rt_sigreturn)
CFI_ENDPROC
END(stub_rt_sigreturn)
-/*
- * initial frame state for interrupts and exceptions
- */
- .macro _frame ref
- CFI_STARTPROC simple
- CFI_DEF_CFA rsp,SS+8-\ref
- /*CFI_REL_OFFSET ss,SS-\ref*/
- CFI_REL_OFFSET rsp,RSP-\ref
- /*CFI_REL_OFFSET rflags,EFLAGS-\ref*/
- /*CFI_REL_OFFSET cs,CS-\ref*/
- CFI_REL_OFFSET rip,RIP-\ref
- .endm
-
/* initial frame state for interrupts (and exceptions without error code) */
-#define INTR_FRAME _frame RIP
+#define INTR_FRAME _frame (RIP-0x10); \
+ CFI_REL_OFFSET rcx,0; \
+ CFI_REL_OFFSET r11,8
+
/* initial frame state for exceptions with error code (and interrupts with
vector already pushed) */
-#define XCPT_FRAME _frame ORIG_RAX
+#define XCPT_FRAME _frame (RIP-0x18); \
+ CFI_REL_OFFSET rcx,0; \
+ CFI_REL_OFFSET r11,8
/*
* Interrupt exit.
@@ -599,8 +599,9 @@ retint_kernel:
#endif
CFI_ENDPROC
-END(common_interrupt)
-
+END(retint_check)
+
+#ifndef CONFIG_XEN
/*
* APIC interrupts.
*/
@@ -613,7 +614,6 @@ END(common_interrupt)
CFI_ENDPROC
.endm
-#ifndef CONFIG_XEN
ENTRY(thermal_interrupt)
apicinterrupt THERMAL_APIC_VECTOR,smp_thermal_interrupt
END(thermal_interrupt)
@@ -668,12 +668,16 @@ END(spurious_interrupt)
.macro zeroentry sym
INTR_FRAME
movq (%rsp),%rcx
+ CFI_RESTORE rcx
movq 8(%rsp),%r11
+ CFI_RESTORE r11
addq $0x10,%rsp /* skip rcx and r11 */
+ CFI_ADJUST_CFA_OFFSET -0x10
pushq $0 /* push error code/oldrax */
CFI_ADJUST_CFA_OFFSET 8
pushq %rax /* push real oldrax to the rdi slot */
CFI_ADJUST_CFA_OFFSET 8
+ CFI_REL_OFFSET rax,0
leaq \sym(%rip),%rax
jmp error_entry
CFI_ENDPROC
@@ -682,10 +686,14 @@ END(spurious_interrupt)
.macro errorentry sym
XCPT_FRAME
movq (%rsp),%rcx
+ CFI_RESTORE rcx
movq 8(%rsp),%r11
+ CFI_RESTORE r11
addq $0x10,%rsp /* rsp points to the error code */
+ CFI_ADJUST_CFA_OFFSET -0x10
pushq %rax
CFI_ADJUST_CFA_OFFSET 8
+ CFI_REL_OFFSET rax,0
leaq \sym(%rip),%rax
jmp error_entry
CFI_ENDPROC
@@ -799,6 +807,7 @@ paranoid_schedule\trace:
*/
ENTRY(error_entry)
_frame RDI
+ CFI_REL_OFFSET rax,0
/* rdi slot contains rax, oldrax contains error code */
cld
subq $14*8,%rsp
@@ -806,6 +815,7 @@ ENTRY(error_entry)
movq %rsi,13*8(%rsp)
CFI_REL_OFFSET rsi,RSI
movq 14*8(%rsp),%rsi /* load rax from rdi slot */
+ CFI_REGISTER rax,rsi
movq %rdx,12*8(%rsp)
CFI_REL_OFFSET rdx,RDX
movq %rcx,11*8(%rsp)
@@ -839,6 +849,7 @@ ENTRY(error_entry)
#endif
error_call_handler:
movq %rdi, RDI(%rsp)
+ CFI_REL_OFFSET rdi,RDI
movq %rsp,%rdi
movq ORIG_RAX(%rsp),%rsi # get error code
movq $-1,ORIG_RAX(%rsp)
@@ -1186,7 +1197,7 @@ END(int3)
ENTRY(overflow)
zeroentry do_overflow
-END(debug)
+END(overflow)
ENTRY(bounds)
zeroentry do_bounds
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/x86_64/kernel/head-xen.S
--- a/linux-2.6-xen-sparse/arch/x86_64/kernel/head-xen.S Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/arch/x86_64/kernel/head-xen.S Fri Apr 06
10:08:30 2007 -0600
@@ -33,9 +33,6 @@ startup_64:
startup_64:
ENTRY(_start)
movq $(init_thread_union+THREAD_SIZE-8),%rsp
- /* zero EFLAGS after setting rsp */
- pushq $0
- popfq
/* rsi is pointer to startup info structure.
pass it to C */
@@ -155,7 +152,7 @@ ENTRY(empty_zero_page)
ENTRY(empty_zero_page)
.skip PAGE_SIZE
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
/*
* __xen_guest information
*/
@@ -187,17 +184,17 @@ ENTRY(empty_zero_page)
.ascii "|supervisor_mode_kernel"
.ascii ",LOADER=generic"
.byte 0
-#endif /* CONFIG_XEN_COMPAT_030002 */
+#endif /* CONFIG_XEN_COMPAT <= 0x030002 */
ELFNOTE(Xen, XEN_ELFNOTE_GUEST_OS, .asciz, "linux")
ELFNOTE(Xen, XEN_ELFNOTE_GUEST_VERSION, .asciz, "2.6")
ELFNOTE(Xen, XEN_ELFNOTE_XEN_VERSION, .asciz, "xen-3.0")
ELFNOTE(Xen, XEN_ELFNOTE_VIRT_BASE, .quad, __START_KERNEL_map)
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET, .quad, __START_KERNEL_map)
#else
ELFNOTE(Xen, XEN_ELFNOTE_PADDR_OFFSET, .quad, 0)
-#endif /* !CONFIG_XEN_COMPAT_030002 */
+#endif
ELFNOTE(Xen, XEN_ELFNOTE_ENTRY, .quad, startup_64)
ELFNOTE(Xen, XEN_ELFNOTE_HYPERCALL_PAGE, .quad, hypercall_page)
ELFNOTE(Xen, XEN_ELFNOTE_L1_MFN_VALID, .quad,
_PAGE_PRESENT,_PAGE_PRESENT)
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/arch/x86_64/mm/init-xen.c
--- a/linux-2.6-xen-sparse/arch/x86_64/mm/init-xen.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/arch/x86_64/mm/init-xen.c Fri Apr 06 10:08:30
2007 -0600
@@ -54,7 +54,7 @@ struct dma_mapping_ops* dma_ops;
struct dma_mapping_ops* dma_ops;
EXPORT_SYMBOL(dma_ops);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
unsigned int __kernel_page_user;
EXPORT_SYMBOL(__kernel_page_user);
#endif
@@ -551,7 +551,7 @@ void __init xen_init_pt(void)
addr = page[pud_index(__START_KERNEL_map)];
addr_to_page(addr, page);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
/* On Xen 3.0.2 and older we may need to explicitly specify _PAGE_USER
in kernel PTEs. We check that here. */
if (HYPERVISOR_xen_version(XENVER_version, NULL) <= 0x30000) {
diff -r 9ec7dadc98ba -r eb3e430242ac linux-2.6-xen-sparse/drivers/xen/Kconfig
--- a/linux-2.6-xen-sparse/drivers/xen/Kconfig Fri Apr 06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/Kconfig Fri Apr 06 10:08:30 2007 -0600
@@ -235,13 +235,12 @@ choice
endchoice
-config XEN_COMPAT_030002
- bool
- default XEN_COMPAT_030002_AND_LATER
-
-config XEN_COMPAT_030004
- bool
- default XEN_COMPAT_030002_AND_LATER || XEN_COMPAT_030004_AND_LATER
+config XEN_COMPAT
+ hex
+ default 0xffffff if XEN_COMPAT_LATEST_ONLY
+ default 0x030004 if XEN_COMPAT_030004_AND_LATER
+ default 0x030002 if XEN_COMPAT_030002_AND_LATER
+ default 0
endmenu
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/balloon/balloon.c
--- a/linux-2.6-xen-sparse/drivers/xen/balloon/balloon.c Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/balloon/balloon.c Fri Apr 06
10:08:30 2007 -0600
@@ -48,13 +48,20 @@
#include <asm/hypervisor.h>
#include <xen/balloon.h>
#include <xen/interface/memory.h>
+#include <asm/maddr.h>
+#include <asm/page.h>
#include <asm/pgalloc.h>
#include <asm/pgtable.h>
#include <asm/uaccess.h>
#include <asm/tlb.h>
+#include <linux/highmem.h>
#include <linux/list.h>
#include <xen/xenbus.h>
#include "common.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
#ifdef CONFIG_PROC_FS
static struct proc_dir_entry *balloon_pde;
@@ -217,6 +224,7 @@ static int increase_reservation(unsigned
set_phys_to_machine(pfn, frame_list[i]);
+#ifdef CONFIG_XEN
/* Link back into the page tables if not highmem. */
if (pfn < max_low_pfn) {
int ret;
@@ -226,6 +234,7 @@ static int increase_reservation(unsigned
0);
BUG_ON(ret);
}
+#endif
/* Relinquish the page back to the allocator. */
ClearPageReserved(page);
@@ -271,9 +280,11 @@ static int decrease_reservation(unsigned
if (!PageHighMem(page)) {
v = phys_to_virt(pfn << PAGE_SHIFT);
scrub_pages(v, 1);
+#ifdef CONFIG_XEN
ret = HYPERVISOR_update_va_mapping(
(unsigned long)v, __pte_ma(0), 0);
BUG_ON(ret);
+#endif
}
#ifdef CONFIG_XEN_SCRUB_PAGES
else {
@@ -284,9 +295,11 @@ static int decrease_reservation(unsigned
#endif
}
+#ifdef CONFIG_XEN
/* Ensure that ballooned highmem pages don't have kmaps. */
kmap_flush_unused();
flush_tlb_all();
+#endif
balloon_lock(flags);
@@ -446,7 +459,7 @@ static struct notifier_block xenstore_no
static int __init balloon_init(void)
{
-#ifdef CONFIG_X86
+#if defined(CONFIG_X86) && defined(CONFIG_XEN)
unsigned long pfn;
struct page *page;
#endif
@@ -456,8 +469,12 @@ static int __init balloon_init(void)
IPRINTK("Initialising balloon driver.\n");
+#ifdef CONFIG_XEN
bs.current_pages = min(xen_start_info->nr_pages, max_pfn);
totalram_pages = bs.current_pages;
+#else
+ bs.current_pages = totalram_pages;
+#endif
bs.target_pages = bs.current_pages;
bs.balloon_low = 0;
bs.balloon_high = 0;
@@ -479,7 +496,7 @@ static int __init balloon_init(void)
#endif
balloon_sysfs_init();
-#ifdef CONFIG_X86
+#if defined(CONFIG_X86) && defined(CONFIG_XEN)
/* Initialise the balloon with excess memory space. */
for (pfn = xen_start_info->nr_pages; pfn < max_pfn; pfn++) {
page = pfn_to_page(pfn);
@@ -498,6 +515,14 @@ static int __init balloon_init(void)
subsys_initcall(balloon_init);
+static void balloon_exit(void)
+{
+ /* XXX - release balloon here */
+ return;
+}
+
+module_exit(balloon_exit);
+
void balloon_update_driver_allowance(long delta)
{
unsigned long flags;
@@ -507,6 +532,7 @@ void balloon_update_driver_allowance(lon
balloon_unlock(flags);
}
+#ifdef CONFIG_XEN
static int dealloc_pte_fn(
pte_t *pte, struct page *pmd_page, unsigned long addr, void *data)
{
@@ -524,6 +550,7 @@ static int dealloc_pte_fn(
BUG_ON(ret != 1);
return 0;
}
+#endif
struct page **alloc_empty_pages_and_pagevec(int nr_pages)
{
@@ -559,8 +586,13 @@ struct page **alloc_empty_pages_and_page
if (ret == 1)
ret = 0; /* success */
} else {
+#ifdef CONFIG_XEN
ret = apply_to_page_range(&init_mm, vaddr, PAGE_SIZE,
dealloc_pte_fn, NULL);
+#else
+ /* Cannot handle non-auto translate mode. */
+ ret = 1;
+#endif
}
if (ret != 0) {
@@ -576,7 +608,9 @@ struct page **alloc_empty_pages_and_page
out:
schedule_work(&balloon_worker);
+#ifdef CONFIG_XEN
flush_tlb_all();
+#endif
return pagevec;
err:
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/balloon/sysfs.c
--- a/linux-2.6-xen-sparse/drivers/xen/balloon/sysfs.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/balloon/sysfs.c Fri Apr 06 10:08:30
2007 -0600
@@ -29,9 +29,15 @@
*/
#include <linux/capability.h>
+#include <linux/errno.h>
#include <linux/stat.h>
+#include <linux/string.h>
#include <linux/sysdev.h>
#include "common.h"
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
#define BALLOON_CLASS_NAME "memory"
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/blkfront/blkfront.c
--- a/linux-2.6-xen-sparse/drivers/xen/blkfront/blkfront.c Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/blkfront/blkfront.c Fri Apr 06
10:08:30 2007 -0600
@@ -661,9 +661,10 @@ void do_blkif_request(request_queue_t *r
if (RING_FULL(&info->ring))
goto wait;
- DPRINTK("do_blk_req %p: cmd %p, sec %lx, "
+ DPRINTK("do_blk_req %p: cmd %p, sec %llx, "
"(%u/%li) buffer:%p [%s]\n",
- req, req->cmd, req->sector, req->current_nr_sectors,
+ req, req->cmd, (long long)req->sector,
+ req->current_nr_sectors,
req->nr_sectors, req->buffer,
rq_data_dir(req) ? "write" : "read");
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/core/machine_kexec.c
--- a/linux-2.6-xen-sparse/drivers/xen/core/machine_kexec.c Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/core/machine_kexec.c Fri Apr 06
10:08:30 2007 -0600
@@ -7,7 +7,6 @@
#include <xen/interface/kexec.h>
#include <linux/mm.h>
#include <linux/bootmem.h>
-#include <asm/hypercall.h>
extern void machine_kexec_setup_load_arg(xen_kexec_image_t *xki,
struct kimage *image);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/core/reboot.c
--- a/linux-2.6-xen-sparse/drivers/xen/core/reboot.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/core/reboot.c Fri Apr 06 10:08:30
2007 -0600
@@ -8,6 +8,10 @@
#include <asm/hypervisor.h>
#include <xen/xenbus.h>
#include <linux/kthread.h>
+
+#ifdef HAVE_XEN_PLATFORM_COMPAT_H
+#include <xen/platform-compat.h>
+#endif
MODULE_LICENSE("Dual BSD/GPL");
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/core/xen_proc.c
--- a/linux-2.6-xen-sparse/drivers/xen/core/xen_proc.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/core/xen_proc.c Fri Apr 06 10:08:30
2007 -0600
@@ -1,4 +1,5 @@
+#include <linux/module.h>
#include <linux/proc_fs.h>
#include <xen/xen_proc.h>
@@ -12,7 +13,11 @@ struct proc_dir_entry *create_xen_proc_e
return create_proc_entry(name, mode, xen_base);
}
+EXPORT_SYMBOL_GPL(create_xen_proc_entry);
+
void remove_xen_proc_entry(const char *name)
{
remove_proc_entry(name, xen_base);
}
+
+EXPORT_SYMBOL_GPL(remove_xen_proc_entry);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/fbfront/xenfb.c
--- a/linux-2.6-xen-sparse/drivers/xen/fbfront/xenfb.c Fri Apr 06 10:06:30
2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/fbfront/xenfb.c Fri Apr 06 10:08:30
2007 -0600
@@ -24,6 +24,7 @@
#include <linux/module.h>
#include <linux/vmalloc.h>
#include <linux/mm.h>
+#include <linux/mutex.h>
#include <asm/hypervisor.h>
#include <xen/evtchn.h>
#include <xen/interface/io/fbif.h>
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/drivers/xen/netfront/netfront.c
--- a/linux-2.6-xen-sparse/drivers/xen/netfront/netfront.c Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/drivers/xen/netfront/netfront.c Fri Apr 06
10:08:30 2007 -0600
@@ -1511,7 +1511,7 @@ static void netif_release_rx_bufs(struct
struct sk_buff *skb;
unsigned long mfn;
int xfer = 0, noxfer = 0, unused = 0;
- int id, ref;
+ int id, ref, rc;
if (np->copying_receiver) {
WPRINTK("%s: fix me for copying receiver.\n", __FUNCTION__);
@@ -1579,7 +1579,9 @@ static void netif_release_rx_bufs(struct
mcl->args[2] = 0;
mcl->args[3] = DOMID_SELF;
mcl++;
- HYPERVISOR_multicall(np->rx_mcl, mcl - np->rx_mcl);
+ rc = HYPERVISOR_multicall_check(
+ np->rx_mcl, mcl - np->rx_mcl, NULL);
+ BUG_ON(rc);
}
}
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypercall.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypercall.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypercall.h Fri Apr
06 10:08:30 2007 -0600
@@ -238,7 +238,7 @@ HYPERVISOR_memory_op(
static inline int
HYPERVISOR_multicall(
- void *call_list, int nr_calls)
+ multicall_entry_t *call_list, int nr_calls)
{
return _hypercall2(int, multicall, call_list, nr_calls);
}
@@ -261,7 +261,7 @@ HYPERVISOR_event_channel_op(
{
int rc = _hypercall2(int, event_channel_op, cmd, arg);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
struct evtchn_op op;
op.cmd = cmd;
@@ -301,7 +301,7 @@ HYPERVISOR_physdev_op(
{
int rc = _hypercall2(int, physdev_op, cmd, arg);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
struct physdev_op op;
op.cmd = cmd;
@@ -358,7 +358,7 @@ HYPERVISOR_suspend(
int rc = _hypercall3(int, sched_op, SCHEDOP_shutdown,
&sched_shutdown, srec);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = _hypercall3(int, sched_op_compat, SCHEDOP_shutdown,
SHUTDOWN_suspend, srec);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypervisor.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypervisor.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/hypervisor.h Fri Apr
06 10:08:30 2007 -0600
@@ -122,7 +122,13 @@ void xen_destroy_contiguous_region(
/* Turn jiffies into Xen system time. */
u64 jiffies_to_st(unsigned long jiffies);
-#include <asm/hypercall.h>
+#ifdef CONFIG_XEN_SCRUB_PAGES
+#define scrub_pages(_p,_n) memset((void *)(_p), 0, (_n) << PAGE_SHIFT)
+#else
+#define scrub_pages(_p,_n) ((void)0)
+#endif
+
+#include <xen/hypercall.h>
#if defined(CONFIG_X86_64)
#define MULTI_UVMFLAGS_INDEX 2
@@ -140,7 +146,7 @@ HYPERVISOR_yield(
{
int rc = HYPERVISOR_sched_op(SCHEDOP_yield, NULL);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = HYPERVISOR_sched_op_compat(SCHEDOP_yield, 0);
#endif
@@ -154,7 +160,7 @@ HYPERVISOR_block(
{
int rc = HYPERVISOR_sched_op(SCHEDOP_block, NULL);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = HYPERVISOR_sched_op_compat(SCHEDOP_block, 0);
#endif
@@ -172,7 +178,7 @@ HYPERVISOR_shutdown(
int rc = HYPERVISOR_sched_op(SCHEDOP_shutdown, &sched_shutdown);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = HYPERVISOR_sched_op_compat(SCHEDOP_shutdown, reason);
#endif
@@ -192,7 +198,7 @@ HYPERVISOR_poll(
set_xen_guest_handle(sched_poll.ports, ports);
rc = HYPERVISOR_sched_op(SCHEDOP_poll, &sched_poll);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = HYPERVISOR_sched_op_compat(SCHEDOP_yield, 0);
#endif
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/maddr.h Fri Apr
06 10:08:30 2007 -0600
@@ -158,7 +158,7 @@ static inline paddr_t pte_machine_to_phy
#define pfn_to_mfn(pfn) (pfn)
#define mfn_to_pfn(mfn) (mfn)
#define mfn_to_local_pfn(mfn) (mfn)
-#define set_phys_to_machine(pfn, mfn) BUG_ON((pfn) != (mfn))
+#define set_phys_to_machine(pfn, mfn) ((void)0)
#define phys_to_machine_mapping_valid(pfn) (1)
#define phys_to_machine(phys) ((maddr_t)(phys))
#define machine_to_phys(mach) ((paddr_t)(mach))
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/page.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/page.h Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/page.h Fri Apr 06
10:08:30 2007 -0600
@@ -44,12 +44,6 @@
foreign; \
})
#define HAVE_ARCH_FREE_PAGE
-
-#ifdef CONFIG_XEN_SCRUB_PAGES
-#define scrub_pages(_p,_n) memset((void *)(_p), 0, (_n) << PAGE_SHIFT)
-#else
-#define scrub_pages(_p,_n) ((void)0)
-#endif
#ifdef CONFIG_X86_USE_3DNOW
@@ -108,7 +102,7 @@ static inline unsigned long long pmd_val
static inline unsigned long long pmd_val(pmd_t x)
{
unsigned long long ret = x.pmd;
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret) ret = pte_machine_to_phys(ret) | _PAGE_PRESENT;
#else
if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
@@ -140,7 +134,7 @@ static inline unsigned long pgd_val(pgd_
static inline unsigned long pgd_val(pgd_t x)
{
unsigned long ret = x.pgd;
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret) ret = machine_to_phys(ret) | _PAGE_PRESENT;
#else
if (ret & _PAGE_PRESENT) ret = machine_to_phys(ret);
@@ -203,10 +197,10 @@ extern int page_is_ram(unsigned long pag
#endif
#define __KERNEL_START (__PAGE_OFFSET + __PHYSICAL_START)
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
#undef LOAD_OFFSET
#define LOAD_OFFSET 0
-#endif /* CONFIG_XEN_COMPAT_030002 */
+#endif
#define PAGE_OFFSET ((unsigned long)__PAGE_OFFSET)
#define VMALLOC_RESERVE ((unsigned long)__VMALLOC_RESERVE)
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-2level.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-2level.h
Fri Apr 06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-2level.h
Fri Apr 06 10:08:30 2007 -0600
@@ -36,8 +36,37 @@
#define pte_clear(mm,addr,xp) do { set_pte_at(mm, addr, xp, __pte(0)); }
while (0)
#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0)
-#define ptep_get_and_clear(mm,addr,xp) __pte_ma(xchg(&(xp)->pte_low, 0))
+#define pte_none(x) (!(x).pte_low)
+
+static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long
addr, pte_t *ptep)
+{
+ pte_t pte = *ptep;
+ if (!pte_none(pte)) {
+ if (mm != &init_mm)
+ pte = __pte_ma(xchg(&ptep->pte_low, 0));
+ else
+ HYPERVISOR_update_va_mapping(addr, __pte(0), 0);
+ }
+ return pte;
+}
+
+#define ptep_clear_flush(vma, addr, ptep) \
+({ \
+ pte_t *__ptep = (ptep); \
+ pte_t __res = *__ptep; \
+ if (!pte_none(__res) && \
+ ((vma)->vm_mm != current->mm || \
+ HYPERVISOR_update_va_mapping(addr, __pte(0), \
+ (unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+ UVMF_INVLPG|UVMF_MULTI))) { \
+ __ptep->pte_low = 0; \
+ flush_tlb_page(vma, addr); \
+ } \
+ __res; \
+})
+
#define pte_same(a, b) ((a).pte_low == (b).pte_low)
+
#define __pte_mfn(_pte) ((_pte).pte_low >> PAGE_SHIFT)
#define pte_mfn(_pte) ((_pte).pte_low & _PAGE_PRESENT ? \
__pte_mfn(_pte) : pfn_to_mfn(__pte_mfn(_pte)))
@@ -46,7 +75,6 @@
#define pte_page(_pte) pfn_to_page(pte_pfn(_pte))
-#define pte_none(x) (!(x).pte_low)
#define pfn_pte(pfn, prot) __pte(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
#define pfn_pmd(pfn, prot) __pmd(((pfn) << PAGE_SHIFT) | pgprot_val(prot))
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-3level.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-3level.h
Fri Apr 06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable-3level.h
Fri Apr 06 10:08:30 2007 -0600
@@ -99,6 +99,11 @@ static inline void pud_clear (pud_t * pu
#define pmd_offset(pud, address) ((pmd_t *) pud_page(*(pud)) + \
pmd_index(address))
+static inline int pte_none(pte_t pte)
+{
+ return !(pte.pte_low | pte.pte_high);
+}
+
/*
* For PTEs and PDEs, we must clear the P-bit first when clearing a page table
* entry, so clear the bottom half first and enforce ordering with a compiler
@@ -106,24 +111,50 @@ static inline void pud_clear (pud_t * pu
*/
static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t
*ptep)
{
- ptep->pte_low = 0;
- smp_wmb();
- ptep->pte_high = 0;
+ if ((mm != current->mm && mm != &init_mm)
+ || HYPERVISOR_update_va_mapping(addr, __pte(0), 0)) {
+ ptep->pte_low = 0;
+ smp_wmb();
+ ptep->pte_high = 0;
+ }
}
#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0)
static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long
addr, pte_t *ptep)
{
- pte_t res;
-
- /* xchg acts as a barrier before the setting of the high bits */
- res.pte_low = xchg(&ptep->pte_low, 0);
- res.pte_high = ptep->pte_high;
- ptep->pte_high = 0;
-
- return res;
-}
+ pte_t pte = *ptep;
+ if (!pte_none(pte)) {
+ if (mm != &init_mm) {
+ uint64_t val = pte_val_ma(pte);
+ if (__cmpxchg64(ptep, val, 0) != val) {
+ /* xchg acts as a barrier before the setting of
the high bits */
+ pte.pte_low = xchg(&ptep->pte_low, 0);
+ pte.pte_high = ptep->pte_high;
+ ptep->pte_high = 0;
+ }
+ } else
+ HYPERVISOR_update_va_mapping(addr, __pte(0), 0);
+ }
+ return pte;
+}
+
+#define ptep_clear_flush(vma, addr, ptep) \
+({ \
+ pte_t *__ptep = (ptep); \
+ pte_t __res = *__ptep; \
+ if (!pte_none(__res) && \
+ ((vma)->vm_mm != current->mm || \
+ HYPERVISOR_update_va_mapping(addr, __pte(0), \
+ (unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+ UVMF_INVLPG|UVMF_MULTI))) { \
+ __ptep->pte_low = 0; \
+ smp_wmb(); \
+ __ptep->pte_high = 0; \
+ flush_tlb_page(vma, addr); \
+ } \
+ __res; \
+})
static inline int pte_same(pte_t a, pte_t b)
{
@@ -131,11 +162,6 @@ static inline int pte_same(pte_t a, pte_
}
#define pte_page(x) pfn_to_page(pte_pfn(x))
-
-static inline int pte_none(pte_t pte)
-{
- return !pte.pte_low && !pte.pte_high;
-}
#define __pte_mfn(_pte) (((_pte).pte_low >> PAGE_SHIFT) | \
((_pte).pte_high << (32-PAGE_SHIFT)))
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable.h
--- a/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-i386/mach-xen/asm/pgtable.h Fri Apr
06 10:08:30 2007 -0600
@@ -210,9 +210,13 @@ extern unsigned long pg0[];
/* To avoid harmful races, pmd_none(x) should check only the lower when PAE */
#define pmd_none(x) (!(unsigned long)pmd_val(x))
+#ifdef CONFIG_XEN_COMPAT_030002
/* pmd_present doesn't just test the _PAGE_PRESENT bit since wr.p.t.
can temporarily clear it. */
#define pmd_present(x) (pmd_val(x))
+#else
+#define pmd_present(x) (pmd_val(x) & _PAGE_PRESENT)
+#endif
#define pmd_bad(x) ((pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER &
~_PAGE_PRESENT)) != (_KERNPG_TABLE & ~_PAGE_PRESENT))
@@ -252,36 +256,47 @@ static inline pte_t pte_mkhuge(pte_t pte
# include <asm/pgtable-2level.h>
#endif
-static inline int ptep_test_and_clear_dirty(struct vm_area_struct *vma,
unsigned long addr, pte_t *ptep)
-{
- if (!pte_dirty(*ptep))
- return 0;
- return test_and_clear_bit(_PAGE_BIT_DIRTY, &ptep->pte_low);
-}
-
-static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
unsigned long addr, pte_t *ptep)
-{
- if (!pte_young(*ptep))
- return 0;
- return test_and_clear_bit(_PAGE_BIT_ACCESSED, &ptep->pte_low);
-}
-
-static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned
long addr, pte_t *ptep, int full)
-{
- pte_t pte;
- if (full) {
- pte = *ptep;
- pte_clear(mm, addr, ptep);
- } else {
- pte = ptep_get_and_clear(mm, addr, ptep);
- }
- return pte;
-}
+#define ptep_test_and_clear_dirty(vma, addr, ptep) \
+({ \
+ pte_t __pte = *(ptep); \
+ int __ret = pte_dirty(__pte); \
+ if (__ret) { \
+ __pte = pte_mkclean(__pte); \
+ if ((vma)->vm_mm != current->mm || \
+ HYPERVISOR_update_va_mapping(addr, __pte, 0)) \
+ (ptep)->pte_low = __pte.pte_low; \
+ } \
+ __ret; \
+})
+
+#define ptep_test_and_clear_young(vma, addr, ptep) \
+({ \
+ pte_t __pte = *(ptep); \
+ int __ret = pte_young(__pte); \
+ if (__ret) \
+ __pte = pte_mkold(__pte); \
+ if ((vma)->vm_mm != current->mm || \
+ HYPERVISOR_update_va_mapping(addr, __pte, 0)) \
+ (ptep)->pte_low = __pte.pte_low; \
+ __ret; \
+})
+
+#define ptep_get_and_clear_full(mm, addr, ptep, full) \
+ ((full) ? ({ \
+ pte_t __res = *(ptep); \
+ if (test_bit(PG_pinned, &virt_to_page((mm)->pgd)->flags)) \
+ xen_l1_entry_update(ptep, __pte(0)); \
+ else \
+ *(ptep) = __pte(0); \
+ __res; \
+ }) : \
+ ptep_get_and_clear(mm, addr, ptep))
static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long
addr, pte_t *ptep)
{
- if (pte_write(*ptep))
- clear_bit(_PAGE_BIT_RW, &ptep->pte_low);
+ pte_t pte = *ptep;
+ if (pte_write(pte))
+ set_pte_at(mm, addr, ptep, pte_wrprotect(pte));
}
/*
@@ -418,6 +433,20 @@ extern void noexec_setup(const char *str
#define pte_unmap_nested(pte) do { } while (0)
#endif
+#define __HAVE_ARCH_PTEP_ESTABLISH
+#define ptep_establish(vma, address, ptep, pteval) \
+ do { \
+ if ( likely((vma)->vm_mm == current->mm) ) { \
+ BUG_ON(HYPERVISOR_update_va_mapping(address, \
+ pteval, \
+ (unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+ UVMF_INVLPG|UVMF_MULTI)); \
+ } else { \
+ xen_l1_entry_update(ptep, pteval); \
+ flush_tlb_page(vma, address); \
+ } \
+ } while (0)
+
/*
* The i386 doesn't have any external MMU info: the kernel page
* tables contain all the necessary information.
@@ -430,26 +459,11 @@ extern void noexec_setup(const char *str
*/
#define update_mmu_cache(vma,address,pte) do { } while (0)
#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
-#define ptep_set_access_flags(__vma, __address, __ptep, __entry, __dirty) \
- do { \
- if (__dirty) { \
- if ( likely((__vma)->vm_mm == current->mm) ) { \
- BUG_ON(HYPERVISOR_update_va_mapping(__address, \
- __entry, \
- (unsigned
long)(__vma)->vm_mm->cpu_vm_mask.bits| \
- UVMF_INVLPG|UVMF_MULTI)); \
- } else { \
- xen_l1_entry_update(__ptep, __entry); \
- flush_tlb_page(__vma, __address); \
- } \
- } \
+#define ptep_set_access_flags(vma, address, ptep, entry, dirty)
\
+ do { \
+ if (dirty) \
+ ptep_establish(vma, address, ptep, entry); \
} while (0)
-
-#define __HAVE_ARCH_PTEP_ESTABLISH
-#define ptep_establish(__vma, __address, __ptep, __entry) \
-do { \
- ptep_set_access_flags(__vma, __address, __ptep, __entry, 1); \
-} while (0)
#include <xen/features.h>
void make_lowmem_page_readonly(void *va, unsigned int feature);
@@ -508,6 +522,7 @@ direct_remap_pfn_range(vma,from,pfn,size
#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_DIRTY
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
+#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
#define __HAVE_ARCH_PTEP_SET_WRPROTECT
#define __HAVE_ARCH_PTE_SAME
#include <asm-generic/pgtable.h>
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h
--- a/linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h Fri Apr 06
10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-ia64/hypervisor.h Fri Apr 06
10:08:30 2007 -0600
@@ -55,7 +55,7 @@ extern int running_on_xen;
#include <xen/interface/event_channel.h>
#include <xen/interface/physdev.h>
#include <xen/interface/sched.h>
-#include <asm/hypercall.h>
+#include <xen/hypercall.h>
#include <asm/ptrace.h>
#include <asm/page.h>
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/hypercall.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/hypercall.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/hypercall.h Fri Apr
06 10:08:30 2007 -0600
@@ -241,7 +241,7 @@ HYPERVISOR_memory_op(
static inline int
HYPERVISOR_multicall(
- void *call_list, int nr_calls)
+ multicall_entry_t *call_list, int nr_calls)
{
return _hypercall2(int, multicall, call_list, nr_calls);
}
@@ -259,7 +259,7 @@ HYPERVISOR_event_channel_op(
{
int rc = _hypercall2(int, event_channel_op, cmd, arg);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
struct evtchn_op op;
op.cmd = cmd;
@@ -299,7 +299,7 @@ HYPERVISOR_physdev_op(
{
int rc = _hypercall2(int, physdev_op, cmd, arg);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (unlikely(rc == -ENOSYS)) {
struct physdev_op op;
op.cmd = cmd;
@@ -359,7 +359,7 @@ HYPERVISOR_suspend(
int rc = _hypercall3(int, sched_op, SCHEDOP_shutdown,
&sched_shutdown, srec);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (rc == -ENOSYS)
rc = _hypercall3(int, sched_op_compat, SCHEDOP_shutdown,
SHUTDOWN_suspend, srec);
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/maddr.h Fri Apr
06 10:08:30 2007 -0600
@@ -140,7 +140,7 @@ static inline paddr_t pte_machine_to_phy
#define pfn_to_mfn(pfn) (pfn)
#define mfn_to_pfn(mfn) (mfn)
#define mfn_to_local_pfn(mfn) (mfn)
-#define set_phys_to_machine(pfn, mfn) BUG_ON((pfn) != (mfn))
+#define set_phys_to_machine(pfn, mfn) ((void)0)
#define phys_to_machine_mapping_valid(pfn) (1)
#define phys_to_machine(phys) ((maddr_t)(phys))
#define machine_to_phys(mach) ((paddr_t)(mach))
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/page.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/page.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/page.h Fri Apr
06 10:08:30 2007 -0600
@@ -23,12 +23,6 @@
foreign; \
})
#define HAVE_ARCH_FREE_PAGE
-
-#ifdef CONFIG_XEN_SCRUB_PAGES
-#define scrub_pages(_p,_n) memset((void *)(_p), 0, (_n) << PAGE_SHIFT)
-#else
-#define scrub_pages(_p,_n) ((void)0)
-#endif
/* PAGE_SHIFT determines the page size */
#define PAGE_SHIFT 12
@@ -110,7 +104,7 @@ static inline unsigned long pmd_val(pmd_
static inline unsigned long pmd_val(pmd_t x)
{
unsigned long ret = x.pmd;
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret) ret = pte_machine_to_phys(ret) | _PAGE_PRESENT;
#else
if (ret & _PAGE_PRESENT) ret = pte_machine_to_phys(ret);
@@ -172,10 +166,10 @@ static inline pgd_t __pgd(unsigned long
#define __PAGE_OFFSET 0xffff880000000000
#endif /* !__ASSEMBLY__ */
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
#undef LOAD_OFFSET
#define LOAD_OFFSET 0
-#endif /* CONFIG_XEN_COMPAT_030002 */
+#endif
/* to align the pointer to the (next) page boundary */
#define PAGE_ALIGN(addr) (((addr)+PAGE_SIZE-1)&PAGE_MASK)
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/pgtable.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/pgtable.h Fri Apr
06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/asm/pgtable.h Fri Apr
06 10:08:30 2007 -0600
@@ -93,11 +93,6 @@ extern unsigned long empty_zero_page[PAG
#define pgd_none(x) (!pgd_val(x))
#define pud_none(x) (!pud_val(x))
-#define set_pte_batched(pteptr, pteval) \
- queue_l1_entry_update(pteptr, (pteval))
-
-extern inline int pud_present(pud_t pud) { return !pud_none(pud); }
-
static inline void set_pte(pte_t *dst, pte_t val)
{
*dst = val;
@@ -122,41 +117,6 @@ static inline void pgd_clear (pgd_t * pg
#define pud_page(pud) \
((unsigned long) __va(pud_val(pud) & PHYSICAL_PAGE_MASK))
-
-/*
- * A note on implementation of this atomic 'get-and-clear' operation.
- * This is actually very simple because Xen Linux can only run on a single
- * processor. Therefore, we cannot race other processors setting the 'accessed'
- * or 'dirty' bits on a page-table entry.
- * Even if pages are shared between domains, that is not a problem because
- * each domain will have separate page tables, with their own versions of
- * accessed & dirty state.
- */
-#define ptep_get_and_clear(mm,addr,xp) __pte_ma(xchg(&(xp)->pte, 0))
-
-#if 0
-static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long
addr, pte_t *xp)
-{
- pte_t pte = *xp;
- if (pte.pte)
- set_pte(xp, __pte_ma(0));
- return pte;
-}
-#endif
-
-struct mm_struct;
-
-static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned
long addr, pte_t *ptep, int full)
-{
- pte_t pte;
- if (full) {
- pte = *ptep;
- *ptep = __pte(0);
- } else {
- pte = ptep_get_and_clear(mm, addr, ptep);
- }
- return pte;
-}
#define pte_same(a, b) ((a).pte == (b).pte)
@@ -205,7 +165,7 @@ static inline pte_t ptep_get_and_clear_f
#define _PAGE_PROTNONE 0x080 /* If not present */
#define _PAGE_NX (1UL<<_PAGE_BIT_NX)
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
extern unsigned int __kernel_page_user;
#else
#define __kernel_page_user 0
@@ -318,6 +278,46 @@ static inline pte_t pfn_pte(unsigned lon
return __pte(pte);
}
+static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long
addr, pte_t *ptep)
+{
+ pte_t pte = *ptep;
+ if (!pte_none(pte)) {
+ if (mm != &init_mm)
+ pte = __pte_ma(xchg(&ptep->pte, 0));
+ else
+ HYPERVISOR_update_va_mapping(addr, __pte(0), 0);
+ }
+ return pte;
+}
+
+static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm, unsigned
long addr, pte_t *ptep, int full)
+{
+ if (full) {
+ pte_t pte = *ptep;
+ if (mm->context.pinned)
+ xen_l1_entry_update(ptep, __pte(0));
+ else
+ *ptep = __pte(0);
+ return pte;
+ }
+ return ptep_get_and_clear(mm, addr, ptep);
+}
+
+#define ptep_clear_flush(vma, addr, ptep) \
+({ \
+ pte_t *__ptep = (ptep); \
+ pte_t __res = *__ptep; \
+ if (!pte_none(__res) && \
+ ((vma)->vm_mm != current->mm || \
+ HYPERVISOR_update_va_mapping(addr, __pte(0), \
+ (unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+ UVMF_INVLPG|UVMF_MULTI))) { \
+ __ptep->pte = 0; \
+ flush_tlb_page(vma, addr); \
+ } \
+ __res; \
+})
+
/*
* The following only work if pte_present() is true.
* Undefined behaviour if not..
@@ -346,31 +346,29 @@ static inline pte_t pte_mkwrite(pte_t pt
static inline pte_t pte_mkwrite(pte_t pte) { __pte_val(pte) |= _PAGE_RW;
return pte; }
static inline pte_t pte_mkhuge(pte_t pte) { __pte_val(pte) |= _PAGE_PSE;
return pte; }
-struct vm_area_struct;
-
-static inline int ptep_test_and_clear_dirty(struct vm_area_struct *vma,
unsigned long addr, pte_t *ptep)
-{
- pte_t pte = *ptep;
- int ret = pte_dirty(pte);
- if (ret)
- set_pte(ptep, pte_mkclean(pte));
- return ret;
-}
-
-static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
unsigned long addr, pte_t *ptep)
-{
- pte_t pte = *ptep;
- int ret = pte_young(pte);
- if (ret)
- set_pte(ptep, pte_mkold(pte));
- return ret;
-}
+#define ptep_test_and_clear_dirty(vma, addr, ptep) \
+({ \
+ pte_t __pte = *(ptep); \
+ int __ret = pte_dirty(__pte); \
+ if (__ret) \
+ set_pte_at((vma)->vm_mm, addr, ptep, pte_mkclean(__pte)); \
+ __ret; \
+})
+
+#define ptep_test_and_clear_young(vma, addr, ptep) \
+({ \
+ pte_t __pte = *(ptep); \
+ int __ret = pte_young(__pte); \
+ if (__ret) \
+ set_pte_at((vma)->vm_mm, addr, ptep, pte_mkold(__pte)); \
+ __ret; \
+})
static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long
addr, pte_t *ptep)
{
pte_t pte = *ptep;
if (pte_write(pte))
- set_pte(ptep, pte_wrprotect(pte));
+ set_pte_at(mm, addr, ptep, pte_wrprotect(pte));
}
/*
@@ -403,6 +401,7 @@ static inline int pmd_large(pmd_t pte) {
/* to find an entry in a page-table-directory. */
#define pud_index(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
#define pud_offset(pgd, address) ((pud_t *) pgd_page(*(pgd)) +
pud_index(address))
+#define pud_present(pud) (pud_val(pud) & _PAGE_PRESENT)
/* PMD - Level 2 access */
#define pmd_page_kernel(pmd) ((unsigned long) __va(pmd_val(pmd) & PTE_MASK))
@@ -412,9 +411,13 @@ static inline int pmd_large(pmd_t pte) {
#define pmd_offset(dir, address) ((pmd_t *) pud_page(*(dir)) + \
pmd_index(address))
#define pmd_none(x) (!pmd_val(x))
+#ifdef CONFIG_XEN_COMPAT_030002
/* pmd_present doesn't just test the _PAGE_PRESENT bit since wr.p.t.
can temporarily clear it. */
#define pmd_present(x) (pmd_val(x))
+#else
+#define pmd_present(x) (pmd_val(x) & _PAGE_PRESENT)
+#endif
#define pmd_clear(xp) do { set_pmd(xp, __pmd(0)); } while (0)
#define pmd_bad(x) ((pmd_val(x) & ~(PTE_MASK | _PAGE_USER | _PAGE_PRESENT)) \
!= (_KERNPG_TABLE & ~(_PAGE_USER | _PAGE_PRESENT)))
@@ -468,25 +471,34 @@ static inline pte_t pte_modify(pte_t pte
#define update_mmu_cache(vma,address,pte) do { } while (0)
+/*
+ * Rules for using ptep_establish: the pte MUST be a user pte, and
+ * must be a present->present transition.
+ */
+#define __HAVE_ARCH_PTEP_ESTABLISH
+#define ptep_establish(vma, address, ptep, pteval) \
+ do { \
+ if ( likely((vma)->vm_mm == current->mm) ) { \
+ BUG_ON(HYPERVISOR_update_va_mapping(address, \
+ pteval, \
+ (unsigned long)(vma)->vm_mm->cpu_vm_mask.bits| \
+ UVMF_INVLPG|UVMF_MULTI)); \
+ } else { \
+ xen_l1_entry_update(ptep, pteval); \
+ flush_tlb_page(vma, address); \
+ } \
+ } while (0)
+
/* We only update the dirty/accessed state if we set
* the dirty bit by hand in the kernel, since the hardware
* will do the accessed bit for us, and we don't want to
* race with other CPU's that might be updating the dirty
* bit at the same time. */
#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
-#define ptep_set_access_flags(__vma, __address, __ptep, __entry, __dirty) \
- do { \
- if (__dirty) { \
- if ( likely((__vma)->vm_mm == current->mm) ) { \
- BUG_ON(HYPERVISOR_update_va_mapping(__address, \
- __entry, \
- (unsigned
long)(__vma)->vm_mm->cpu_vm_mask.bits| \
- UVMF_INVLPG|UVMF_MULTI)); \
- } else { \
- xen_l1_entry_update(__ptep, __entry); \
- flush_tlb_page(__vma, __address); \
- } \
- } \
+#define ptep_set_access_flags(vma, address, ptep, entry, dirty)
\
+ do { \
+ if (dirty) \
+ ptep_establish(vma, address, ptep, entry); \
} while (0)
/* Encode and de-code a swap entry */
@@ -506,6 +518,8 @@ extern int kern_addr_valid(unsigned long
#define DOMID_LOCAL (0xFFFFU)
+struct vm_area_struct;
+
int direct_remap_pfn_range(struct vm_area_struct *vma,
unsigned long address,
unsigned long mfn,
@@ -551,6 +565,7 @@ int touch_pte_range(struct mm_struct *mm
#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_DIRTY
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
+#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
#define __HAVE_ARCH_PTEP_SET_WRPROTECT
#define __HAVE_ARCH_PTE_SAME
#include <asm-generic/pgtable.h>
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/setup_arch_post.h
--- a/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/setup_arch_post.h
Fri Apr 06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/asm-x86_64/mach-xen/setup_arch_post.h
Fri Apr 06 10:08:30 2007 -0600
@@ -39,7 +39,7 @@ static void __init machine_specific_arch
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &failsafe);
if (ret == 0)
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &syscall);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret == -ENOSYS)
ret = HYPERVISOR_set_callbacks(
event.address,
@@ -50,7 +50,7 @@ static void __init machine_specific_arch
#ifdef CONFIG_X86_LOCAL_APIC
ret = HYPERVISOR_callback_op(CALLBACKOP_register, &nmi_cb);
-#ifdef CONFIG_XEN_COMPAT_030002
+#if CONFIG_XEN_COMPAT <= 0x030002
if (ret == -ENOSYS) {
static struct xennmi_callback __initdata cb = {
.handler_address = (unsigned long)nmi
diff -r 9ec7dadc98ba -r eb3e430242ac linux-2.6-xen-sparse/include/xen/hvm.h
--- a/linux-2.6-xen-sparse/include/xen/hvm.h Fri Apr 06 10:06:30 2007 -0600
+++ b/linux-2.6-xen-sparse/include/xen/hvm.h Fri Apr 06 10:08:30 2007 -0600
@@ -3,7 +3,6 @@
#define XEN_HVM_H__
#include <xen/interface/hvm/params.h>
-#include <asm/hypercall.h>
static inline unsigned long hvm_get_parameter(int idx)
{
diff -r 9ec7dadc98ba -r eb3e430242ac
linux-2.6-xen-sparse/include/xen/hypercall.h
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/linux-2.6-xen-sparse/include/xen/hypercall.h Fri Apr 06 10:08:30
2007 -0600
@@ -0,0 +1,24 @@
+#ifndef __XEN_HYPERCALL_H__
+#define __XEN_HYPERCALL_H__
+
+#include <asm/hypercall.h>
+
+static inline int
+HYPERVISOR_multicall_check(
+ multicall_entry_t *call_list, int nr_calls,
+ const unsigned long *rc_list)
+{
+ int rc = HYPERVISOR_multicall(call_list, nr_calls);
+
+ if (unlikely(rc < 0))
+ return rc;
+ BUG_ON(rc);
+
+ for ( ; nr_calls > 0; --nr_calls, ++call_list)
+ if (unlikely(call_list->result != (rc_list ? *rc_list++ : 0)))
+ return nr_calls;
+
+ return 0;
+}
+
+#endif /* __XEN_HYPERCALL_H__ */
diff -r 9ec7dadc98ba -r eb3e430242ac tools/ioemu/hw/xen_platform.c
--- a/tools/ioemu/hw/xen_platform.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/ioemu/hw/xen_platform.c Fri Apr 06 10:08:30 2007 -0600
@@ -29,16 +29,10 @@
extern FILE *logfile;
-static void platform_ioport_write(void *opaque, uint32_t addr, uint32_t val)
-{
- if (val == 0)
- qemu_invalidate_map_cache();
-}
-
static void platform_ioport_map(PCIDevice *pci_dev, int region_num,
uint32_t addr, uint32_t size, int type)
{
- register_ioport_write(addr, 1, 1, platform_ioport_write, NULL);
+ /* nothing yet */
}
static uint32_t platform_mmio_read(void *opaque, target_phys_addr_t addr)
diff -r 9ec7dadc98ba -r eb3e430242ac tools/ioemu/target-i386-dm/exec-dm.c
--- a/tools/ioemu/target-i386-dm/exec-dm.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/ioemu/target-i386-dm/exec-dm.c Fri Apr 06 10:08:30 2007 -0600
@@ -431,14 +431,8 @@ static inline int paddr_is_ram(target_ph
{
/* Is this guest physical address RAM-backed? */
#if defined(CONFIG_DM) && (defined(__i386__) || defined(__x86_64__))
- if (ram_size <= HVM_BELOW_4G_RAM_END)
- /* RAM is contiguous */
- return (addr < ram_size);
- else
- /* There is RAM below and above the MMIO hole */
- return ((addr < HVM_BELOW_4G_MMIO_START) ||
- ((addr >= HVM_BELOW_4G_MMIO_START + HVM_BELOW_4G_MMIO_LENGTH)
- && (addr < ram_size + HVM_BELOW_4G_MMIO_LENGTH)));
+ return ((addr < HVM_BELOW_4G_MMIO_START) ||
+ (addr >= HVM_BELOW_4G_MMIO_START + HVM_BELOW_4G_MMIO_LENGTH));
#else
return (addr < ram_size);
#endif
diff -r 9ec7dadc98ba -r eb3e430242ac tools/ioemu/target-i386-dm/helper2.c
--- a/tools/ioemu/target-i386-dm/helper2.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/ioemu/target-i386-dm/helper2.c Fri Apr 06 10:08:30 2007 -0600
@@ -135,9 +135,6 @@ void cpu_reset(CPUX86State *env)
{
int xcHandle;
int sts;
-
- /* pause domain first, to avoid repeated reboot request*/
- xc_domain_pause(xc_handle, domid);
xcHandle = xc_interface_open();
if (xcHandle < 0)
@@ -509,8 +506,11 @@ void __handle_ioreq(CPUState *env, ioreq
cpu_ioreq_xchg(env, req);
break;
case IOREQ_TYPE_TIMEOFFSET:
- cpu_ioreq_timeoffset(env, req);
- break;
+ cpu_ioreq_timeoffset(env, req);
+ break;
+ case IOREQ_TYPE_INVALIDATE:
+ qemu_invalidate_map_cache();
+ break;
default:
hw_error("Invalid ioreq type 0x%x\n", req->type);
}
@@ -597,6 +597,7 @@ int main_loop(void)
extern int suspend_requested;
CPUState *env = cpu_single_env;
int evtchn_fd = xc_evtchn_fd(xce_handle);
+ char qemu_file[20];
buffered_io_timer = qemu_new_timer(rt_clock, handle_buffered_io,
cpu_single_env);
@@ -604,52 +605,23 @@ int main_loop(void)
qemu_set_fd_handler(evtchn_fd, cpu_handle_ioreq, NULL, env);
- while (1) {
- if (vm_running) {
- if (shutdown_requested)
- break;
- if (reset_requested) {
- qemu_system_reset();
- reset_requested = 0;
- }
- if (suspend_requested) {
- fprintf(logfile, "device model received suspend signal!\n");
- break;
- }
- }
-
+ while (!(vm_running && suspend_requested))
/* Wait up to 10 msec. */
main_loop_wait(10);
- }
- if (!suspend_requested)
- destroy_hvm_domain();
- else {
- char qemu_file[20];
- ioreq_t *req;
- int rc;
-
- sprintf(qemu_file, "/tmp/xen.qemu-dm.%d", domid);
- xc_domain_pause(xc_handle, domid);
-
- /* Pull all outstanding ioreqs through the system */
- handle_buffered_io(env);
- main_loop_wait(1); /* For the select() on events */
-
- /* Stop the IDE thread */
- ide_stop_dma_thread();
-
- /* Make sure that all outstanding IO responses are handled too */
- if ( xc_hvm_drain_io(xc_handle, domid) != 0 )
- {
- fprintf(stderr, "error clearing ioreq rings (%s)\n",
- strerror(errno));
- return -1;
- }
-
- /* Save the device state */
- if (qemu_savevm(qemu_file) < 0)
- fprintf(stderr, "qemu save fail.\n");
- }
+
+ fprintf(logfile, "device model received suspend signal!\n");
+
+ /* Pull all outstanding ioreqs through the system */
+ handle_buffered_io(env);
+ main_loop_wait(1); /* For the select() on events */
+
+ /* Stop the IDE thread */
+ ide_stop_dma_thread();
+
+ /* Save the device state */
+ sprintf(qemu_file, "/tmp/xen.qemu-dm.%d", domid);
+ if (qemu_savevm(qemu_file) < 0)
+ fprintf(stderr, "qemu save fail.\n");
return 0;
}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/ioemu/vl.c
--- a/tools/ioemu/vl.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/ioemu/vl.c Fri Apr 06 10:08:30 2007 -0600
@@ -88,6 +88,7 @@
#include "exec-all.h"
+#include <xen/hvm/params.h>
#define DEFAULT_NETWORK_SCRIPT "/etc/xen/qemu-ifup"
#define DEFAULT_BRIDGE "xenbr0"
@@ -5886,7 +5887,8 @@ int set_mm_mapping(int xc_handle, uint32
void suspend(int sig)
{
- fprintf(logfile, "suspend sig handler called with requested=%d!\n",
suspend_requested);
+ fprintf(logfile, "suspend sig handler called with requested=%d!\n",
+ suspend_requested);
if (sig != SIGUSR1)
fprintf(logfile, "suspend signal dismatch, get sig=%d!\n", sig);
suspend_requested = 1;
@@ -5900,32 +5902,28 @@ static unsigned long last_address_index
static unsigned long last_address_index = ~0UL;
static uint8_t *last_address_vaddr;
-static int qemu_map_cache_init(unsigned long nr_pages)
-{
- unsigned long max_pages = MAX_MCACHE_SIZE >> PAGE_SHIFT;
- int i;
-
- if (nr_pages < max_pages)
- max_pages = nr_pages;
-
- nr_buckets = max_pages + (1UL << (MCACHE_BUCKET_SHIFT - PAGE_SHIFT)) - 1;
- nr_buckets >>= (MCACHE_BUCKET_SHIFT - PAGE_SHIFT);
+static int qemu_map_cache_init(void)
+{
+ unsigned long size;
+
+ nr_buckets = (((MAX_MCACHE_SIZE >> PAGE_SHIFT) +
+ (1UL << (MCACHE_BUCKET_SHIFT - PAGE_SHIFT)) - 1) >>
+ (MCACHE_BUCKET_SHIFT - PAGE_SHIFT));
fprintf(logfile, "qemu_map_cache_init nr_buckets = %lx\n", nr_buckets);
- mapcache_entry = malloc(nr_buckets * sizeof(struct map_cache));
- if (mapcache_entry == NULL) {
+ /*
+ * Use mmap() directly: lets us allocate a big hash table with no up-front
+ * cost in storage space. The OS will allocate memory only for the buckets
+ * that we actually use. All others will contain all zeroes.
+ */
+ size = nr_buckets * sizeof(struct map_cache);
+ size = (size + PAGE_SIZE - 1) & ~(PAGE_SIZE - 1);
+ mapcache_entry = mmap(NULL, size, PROT_READ|PROT_WRITE,
+ MAP_SHARED|MAP_ANONYMOUS, 0, 0);
+ if (mapcache_entry == MAP_FAILED) {
errno = ENOMEM;
return -1;
}
-
- memset(mapcache_entry, 0, nr_buckets * sizeof(struct map_cache));
-
- /*
- * To avoid ENOMEM from xc_map_foreign_batch() at runtime, we
- * pre-fill all the map caches in advance.
- */
- for (i = 0; i < nr_buckets; i++)
- (void)qemu_map_cache(((target_phys_addr_t)i) << MCACHE_BUCKET_SHIFT);
return 0;
}
@@ -6038,11 +6036,14 @@ int main(int argc, char **argv)
QEMUMachine *machine;
char usb_devices[MAX_USB_CMDLINE][128];
int usb_devices_index;
- unsigned long nr_pages, tmp_nr_pages, shared_page_nr;
- xen_pfn_t *page_array;
+ unsigned long ioreq_pfn;
extern void *shared_page;
extern void *buffered_io_page;
+#ifdef __ia64__
+ unsigned long nr_pages;
+ xen_pfn_t *page_array;
extern void *buffered_pio_page;
+#endif
char qemu_dm_logfilename[64];
@@ -6592,47 +6593,36 @@ int main(int argc, char **argv)
xc_handle = xc_interface_open();
- nr_pages = ram_size/PAGE_SIZE;
- tmp_nr_pages = nr_pages;
-
#if defined(__i386__) || defined(__x86_64__)
- if (ram_size > HVM_BELOW_4G_RAM_END) {
- tmp_nr_pages += HVM_BELOW_4G_MMIO_LENGTH >> PAGE_SHIFT;
- shared_page_nr = (HVM_BELOW_4G_RAM_END >> PAGE_SHIFT) - 1;
- } else
- shared_page_nr = nr_pages - 1;
-#endif
-
-#if defined(__i386__) || defined(__x86_64__)
-
- if ( qemu_map_cache_init(tmp_nr_pages) )
- {
+
+ if (qemu_map_cache_init()) {
fprintf(logfile, "qemu_map_cache_init returned: error %d\n", errno);
exit(-1);
}
+ xc_get_hvm_param(xc_handle, domid, HVM_PARAM_IOREQ_PFN, &ioreq_pfn);
+ fprintf(logfile, "shared page at pfn %lx\n", ioreq_pfn);
shared_page = xc_map_foreign_range(xc_handle, domid, PAGE_SIZE,
- PROT_READ|PROT_WRITE, shared_page_nr);
+ PROT_READ|PROT_WRITE, ioreq_pfn);
if (shared_page == NULL) {
fprintf(logfile, "map shared IO page returned error %d\n", errno);
exit(-1);
}
- fprintf(logfile, "shared page at pfn:%lx\n", shared_page_nr);
-
+ xc_get_hvm_param(xc_handle, domid, HVM_PARAM_BUFIOREQ_PFN, &ioreq_pfn);
+ fprintf(logfile, "buffered io page at pfn %lx\n", ioreq_pfn);
buffered_io_page = xc_map_foreign_range(xc_handle, domid, PAGE_SIZE,
- PROT_READ|PROT_WRITE,
- shared_page_nr - 2);
+ PROT_READ|PROT_WRITE, ioreq_pfn);
if (buffered_io_page == NULL) {
fprintf(logfile, "map buffered IO page returned error %d\n", errno);
exit(-1);
}
- fprintf(logfile, "buffered io page at pfn:%lx\n", shared_page_nr - 2);
-
#elif defined(__ia64__)
- page_array = (xen_pfn_t *)malloc(tmp_nr_pages * sizeof(xen_pfn_t));
+ nr_pages = ram_size/PAGE_SIZE;
+
+ page_array = (xen_pfn_t *)malloc(nr_pages * sizeof(xen_pfn_t));
if (page_array == NULL) {
fprintf(logfile, "malloc returned error %d\n", errno);
exit(-1);
@@ -6650,7 +6640,7 @@ int main(int argc, char **argv)
PROT_READ|PROT_WRITE,
BUFFER_PIO_PAGE_START >> PAGE_SHIFT);
- for (i = 0; i < tmp_nr_pages; i++)
+ for (i = 0; i < nr_pages; i++)
page_array[i] = i;
/* VTI will not use memory between 3G~4G, so we just pass a legal pfn
diff -r 9ec7dadc98ba -r eb3e430242ac tools/ioemu/xenstore.c
--- a/tools/ioemu/xenstore.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/ioemu/xenstore.c Fri Apr 06 10:08:30 2007 -0600
@@ -248,12 +248,8 @@ void xenstore_process_logdirty_event(voi
key = (key_t) strtoull(key_terminated, NULL, 16);
/* Figure out how bit the log-dirty bitmaps are */
- logdirty_bitmap_size = ((phys_ram_size + 0x20
- - (vga_ram_size + bios_size))
- >> (TARGET_PAGE_BITS)); /* nr of bits in map*/
- if (logdirty_bitmap_size > HVM_BELOW_4G_MMIO_START >> TARGET_PAGE_BITS)
- logdirty_bitmap_size +=
- HVM_BELOW_4G_MMIO_LENGTH >> TARGET_PAGE_BITS; /* still bits */
+ logdirty_bitmap_size = xc_memory_op(xc_handle,
+ XENMEM_maximum_gpfn, &domid) + 1;
logdirty_bitmap_size = ((logdirty_bitmap_size + HOST_LONG_BITS - 1)
/ HOST_LONG_BITS); /* longs */
logdirty_bitmap_size *= sizeof (unsigned long); /* bytes */
@@ -272,7 +268,7 @@ void xenstore_process_logdirty_event(voi
/* Double-check that the bitmaps are the size we expect */
if (logdirty_bitmap_size != *(uint32_t *)seg) {
- fprintf(logfile, "Log-dirty: got %lu, calc %lu\n",
+ fprintf(logfile, "Log-dirty: got %u, calc %lu\n",
*(uint32_t *)seg, logdirty_bitmap_size);
return;
}
@@ -304,7 +300,7 @@ void xenstore_process_logdirty_event(voi
fprintf(logfile, "Log-dirty: bad next-active entry: %s\n", act);
exit(1);
}
- logdirty_bitmap = seg + i * logdirty_bitmap_size;
+ logdirty_bitmap = (unsigned long *)(seg + i * logdirty_bitmap_size);
/* Ack that we've switched */
xs_write(xsh, XBT_NULL, active_path, act, len);
@@ -612,7 +608,7 @@ int xenstore_vm_write(int domid, char *k
path = xs_get_domain_path(xsh, domid);
if (path == NULL) {
- fprintf(logfile, "xs_get_domain_path(%d): error\n");
+ fprintf(logfile, "xs_get_domain_path: error\n");
goto out;
}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/Makefile
--- a/tools/libxc/Makefile Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxc/Makefile Fri Apr 06 10:08:30 2007 -0600
@@ -26,8 +26,8 @@ CTRL_SRCS-$(CONFIG_X86_Linux) += xc_ptra
GUEST_SRCS-y :=
GUEST_SRCS-y += xg_private.c
-GUEST_SRCS-$(CONFIG_MIGRATE) += xc_linux_restore.c xc_linux_save.c
-GUEST_SRCS-$(CONFIG_HVM) += xc_hvm_build.c xc_hvm_restore.c xc_hvm_save.c
+GUEST_SRCS-$(CONFIG_MIGRATE) += xc_domain_restore.c xc_linux_save.c
+GUEST_SRCS-$(CONFIG_HVM) += xc_hvm_build.c xc_hvm_save.c
# symlink libelf from xen/common/libelf/
LIBELF_SRCS := libelf-tools.c libelf-loader.c
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/ia64/xc_ia64_linux_restore.c
--- a/tools/libxc/ia64/xc_ia64_linux_restore.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxc/ia64/xc_ia64_linux_restore.c Fri Apr 06 10:08:30 2007 -0600
@@ -20,9 +20,6 @@ static unsigned long p2m_size;
/* number of 'in use' pfns in the guest (i.e. #P2M entries with a valid mfn) */
static unsigned long nr_pfns;
-/* largest possible value of nr_pfns (i.e. domain's maximum memory size) */
-static unsigned long max_nr_pfns;
-
static ssize_t
read_exact(int fd, void *buf, size_t count)
{
@@ -62,10 +59,10 @@ read_page(int xc_handle, int io_fd, uint
}
int
-xc_linux_restore(int xc_handle, int io_fd, uint32_t dom,
- unsigned long p2msize, unsigned long maxnrpfns,
+xc_domain_restore(int xc_handle, int io_fd, uint32_t dom,
unsigned int store_evtchn, unsigned long *store_mfn,
- unsigned int console_evtchn, unsigned long *console_mfn)
+ unsigned int console_evtchn, unsigned long *console_mfn,
+ unsigned int hvm, unsigned int pae)
{
DECLARE_DOMCTL;
int rc = 1, i;
@@ -85,12 +82,19 @@ xc_linux_restore(int xc_handle, int io_f
/* A temporary mapping of the guest's start_info page. */
start_info_t *start_info;
- p2m_size = p2msize;
- max_nr_pfns = maxnrpfns;
+ if (hvm) {
+ ERROR("HVM Restore is unsupported");
+ goto out;
+ }
/* For info only */
nr_pfns = 0;
+ if ( !read_exact(io_fd, &p2m_size, sizeof(unsigned long)) )
+ {
+ ERROR("read: p2m_size");
+ goto out;
+ }
DPRINTF("xc_linux_restore start: p2m_size = %lx\n", p2m_size);
if (!read_exact(io_fd, &ver, sizeof(unsigned long))) {
@@ -106,11 +110,6 @@ xc_linux_restore(int xc_handle, int io_f
/* needed for build domctl, but might as well do early */
ERROR("Unable to mlock ctxt");
return 1;
- }
-
- if (xc_domain_setmaxmem(xc_handle, dom, PFN_TO_KB(max_nr_pfns)) != 0) {
- errno = ENOMEM;
- goto out;
}
/* Get pages. */
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xc_domain_restore.c
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/libxc/xc_domain_restore.c Fri Apr 06 10:08:30 2007 -0600
@@ -0,0 +1,1086 @@
+/******************************************************************************
+ * xc_domain_restore.c
+ *
+ * Restore the state of a guest session.
+ *
+ * Copyright (c) 2003, K A Fraser.
+ * Copyright (c) 2006, Intel Corporation
+ * Copyright (c) 2007, XenSource Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc., 59 Temple
+ * Place - Suite 330, Boston, MA 02111-1307 USA.
+ *
+ */
+
+#include <stdlib.h>
+#include <unistd.h>
+
+#include "xg_private.h"
+#include "xg_save_restore.h"
+#include "xc_dom.h"
+
+#include <xen/hvm/ioreq.h>
+#include <xen/hvm/params.h>
+
+/* max mfn of the current host machine */
+static unsigned long max_mfn;
+
+/* virtual starting address of the hypervisor */
+static unsigned long hvirt_start;
+
+/* #levels of page tables used by the current guest */
+static unsigned int pt_levels;
+
+/* number of pfns this guest has (i.e. number of entries in the P2M) */
+static unsigned long p2m_size;
+
+/* number of 'in use' pfns in the guest (i.e. #P2M entries with a valid mfn) */
+static unsigned long nr_pfns;
+
+/* Live mapping of the table mapping each PFN to its current MFN. */
+static xen_pfn_t *live_p2m = NULL;
+
+/* A table mapping each PFN to its new MFN. */
+static xen_pfn_t *p2m = NULL;
+
+/* A table of P2M mappings in the current region */
+static xen_pfn_t *p2m_batch = NULL;
+
+static ssize_t
+read_exact(int fd, void *buf, size_t count)
+{
+ int r = 0, s;
+ unsigned char *b = buf;
+
+ while (r < count) {
+ s = read(fd, &b[r], count - r);
+ if ((s == -1) && (errno == EINTR))
+ continue;
+ if (s <= 0) {
+ break;
+ }
+ r += s;
+ }
+
+ return (r == count) ? 1 : 0;
+}
+
+/*
+** In the state file (or during transfer), all page-table pages are
+** converted into a 'canonical' form where references to actual mfns
+** are replaced with references to the corresponding pfns.
+** This function inverts that operation, replacing the pfn values with
+** the (now known) appropriate mfn values.
+*/
+static int uncanonicalize_pagetable(int xc_handle, uint32_t dom,
+ unsigned long type, void *page)
+{
+ int i, pte_last;
+ unsigned long pfn;
+ uint64_t pte;
+ int nr_mfns = 0;
+
+ pte_last = PAGE_SIZE / ((pt_levels == 2)? 4 : 8);
+
+ /* First pass: work out how many (if any) MFNs we need to alloc */
+ for(i = 0; i < pte_last; i++) {
+
+ if(pt_levels == 2)
+ pte = ((uint32_t *)page)[i];
+ else
+ pte = ((uint64_t *)page)[i];
+
+ /* XXX SMH: below needs fixing for PROT_NONE etc */
+ if(!(pte & _PAGE_PRESENT))
+ continue;
+
+ pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
+
+ if(pfn >= p2m_size) {
+ /* This "page table page" is probably not one; bail. */
+ ERROR("Frame number in type %lu page table is out of range: "
+ "i=%d pfn=0x%lx p2m_size=%lu",
+ type >> 28, i, pfn, p2m_size);
+ return 0;
+ }
+
+ if(p2m[pfn] == INVALID_P2M_ENTRY) {
+ /* Have a 'valid' PFN without a matching MFN - need to alloc */
+ p2m_batch[nr_mfns++] = pfn;
+ }
+ }
+
+
+ /* Allocate the requistite number of mfns */
+ if (nr_mfns && xc_domain_memory_populate_physmap(
+ xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
+ ERROR("Failed to allocate memory for batch.!\n");
+ errno = ENOMEM;
+ return 0;
+ }
+
+ /* Second pass: uncanonicalize each present PTE */
+ nr_mfns = 0;
+ for(i = 0; i < pte_last; i++) {
+
+ if(pt_levels == 2)
+ pte = ((uint32_t *)page)[i];
+ else
+ pte = ((uint64_t *)page)[i];
+
+ /* XXX SMH: below needs fixing for PROT_NONE etc */
+ if(!(pte & _PAGE_PRESENT))
+ continue;
+
+ pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
+
+ if(p2m[pfn] == INVALID_P2M_ENTRY)
+ p2m[pfn] = p2m_batch[nr_mfns++];
+
+ pte &= ~MADDR_MASK_X86;
+ pte |= (uint64_t)p2m[pfn] << PAGE_SHIFT;
+
+ if(pt_levels == 2)
+ ((uint32_t *)page)[i] = (uint32_t)pte;
+ else
+ ((uint64_t *)page)[i] = (uint64_t)pte;
+ }
+
+ return 1;
+}
+
+
+/* Load the p2m frame list, plus potential extended info chunk */
+static xen_pfn_t * load_p2m_frame_list(int io_fd, int *pae_extended_cr3)
+{
+ xen_pfn_t *p2m_frame_list;
+ vcpu_guest_context_t ctxt;
+
+ if (!(p2m_frame_list = malloc(P2M_FL_SIZE))) {
+ ERROR("Couldn't allocate p2m_frame_list array");
+ return NULL;
+ }
+
+ /* Read first entry of P2M list, or extended-info signature (~0UL). */
+ if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
+ ERROR("read extended-info signature failed");
+ return NULL;
+ }
+
+ if (p2m_frame_list[0] == ~0UL) {
+ uint32_t tot_bytes;
+
+ /* Next 4 bytes: total size of following extended info. */
+ if (!read_exact(io_fd, &tot_bytes, sizeof(tot_bytes))) {
+ ERROR("read extended-info size failed");
+ return NULL;
+ }
+
+ while (tot_bytes) {
+ uint32_t chunk_bytes;
+ char chunk_sig[4];
+
+ /* 4-character chunk signature + 4-byte remaining chunk size. */
+ if (!read_exact(io_fd, chunk_sig, sizeof(chunk_sig)) ||
+ !read_exact(io_fd, &chunk_bytes, sizeof(chunk_bytes))) {
+ ERROR("read extended-info chunk signature failed");
+ return NULL;
+ }
+ tot_bytes -= 8;
+
+ /* VCPU context structure? */
+ if (!strncmp(chunk_sig, "vcpu", 4)) {
+ if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ ERROR("read extended-info vcpu context failed");
+ return NULL;
+ }
+ tot_bytes -= sizeof(struct vcpu_guest_context);
+ chunk_bytes -= sizeof(struct vcpu_guest_context);
+
+ if (ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3))
+ *pae_extended_cr3 = 1;
+ }
+
+ /* Any remaining bytes of this chunk: read and discard. */
+ while (chunk_bytes) {
+ unsigned long sz = chunk_bytes;
+ if ( sz > P2M_FL_SIZE )
+ sz = P2M_FL_SIZE;
+ if (!read_exact(io_fd, p2m_frame_list, sz)) {
+ ERROR("read-and-discard extended-info chunk bytes failed");
+ return NULL;
+ }
+ chunk_bytes -= sz;
+ tot_bytes -= sz;
+ }
+ }
+
+ /* Now read the real first entry of P2M list. */
+ if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
+ ERROR("read first entry of p2m_frame_list failed");
+ return NULL;
+ }
+ }
+
+ /* First entry is already read into the p2m array. */
+ if (!read_exact(io_fd, &p2m_frame_list[1], P2M_FL_SIZE - sizeof(long))) {
+ ERROR("read p2m_frame_list failed");
+ return NULL;
+ }
+
+ return p2m_frame_list;
+}
+
+
+
+int xc_domain_restore(int xc_handle, int io_fd, uint32_t dom,
+ unsigned int store_evtchn, unsigned long *store_mfn,
+ unsigned int console_evtchn, unsigned long *console_mfn,
+ unsigned int hvm, unsigned int pae)
+{
+ DECLARE_DOMCTL;
+ int rc = 1, i, j, n, m, pae_extended_cr3 = 0;
+ unsigned long mfn, pfn;
+ unsigned int prev_pc, this_pc;
+ int verify = 0;
+ int nraces = 0;
+
+ /* The new domain's shared-info frame number. */
+ unsigned long shared_info_frame;
+ unsigned char shared_info_page[PAGE_SIZE]; /* saved contents from file */
+ shared_info_t *shared_info = (shared_info_t *)shared_info_page;
+
+ /* A copy of the CPU context of the guest. */
+ vcpu_guest_context_t ctxt;
+
+ /* A table containing the type of each PFN (/not/ MFN!). */
+ unsigned long *pfn_type = NULL;
+
+ /* A table of MFNs to map in the current region */
+ xen_pfn_t *region_mfn = NULL;
+
+ /* Types of the pfns in the current region */
+ unsigned long region_pfn_type[MAX_BATCH_SIZE];
+
+ /* A temporary mapping, and a copy, of one frame of guest memory. */
+ unsigned long *page = NULL;
+
+ /* A copy of the pfn-to-mfn table frame list. */
+ xen_pfn_t *p2m_frame_list = NULL;
+
+ /* A temporary mapping of the guest's start_info page. */
+ start_info_t *start_info;
+
+ /* Our mapping of the current region (batch) */
+ char *region_base;
+
+ xc_mmu_t *mmu = NULL;
+
+ /* used by debug verify code */
+ unsigned long buf[PAGE_SIZE/sizeof(unsigned long)];
+
+ struct mmuext_op pin[MAX_PIN_BATCH];
+ unsigned int nr_pins;
+
+ uint64_t vcpumap = 1ULL;
+ unsigned int max_vcpu_id = 0;
+ int new_ctxt_format = 0;
+
+ /* Magic frames in HVM guests: ioreqs and xenstore comms. */
+ uint64_t magic_pfns[3]; /* ioreq_pfn, bufioreq_pfn, store_pfn */
+
+ /* Buffer for holding HVM context */
+ uint8_t *hvm_buf = NULL;
+
+ /* For info only */
+ nr_pfns = 0;
+
+ if ( !read_exact(io_fd, &p2m_size, sizeof(unsigned long)) )
+ {
+ ERROR("read: p2m_size");
+ goto out;
+ }
+ DPRINTF("xc_domain_restore start: p2m_size = %lx\n", p2m_size);
+
+ if ( !hvm )
+ {
+ /*
+ * XXX For now, 32bit dom0's can only save/restore 32bit domUs
+ * on 64bit hypervisors.
+ */
+ memset(&domctl, 0, sizeof(domctl));
+ domctl.domain = dom;
+ domctl.cmd = XEN_DOMCTL_set_address_size;
+ domctl.u.address_size.size = sizeof(unsigned long) * 8;
+ rc = do_domctl(xc_handle, &domctl);
+ if ( rc != 0 ) {
+ ERROR("Unable to set guest address size.");
+ goto out;
+ }
+ rc = 1;
+ }
+
+ if(!get_platform_info(xc_handle, dom,
+ &max_mfn, &hvirt_start, &pt_levels)) {
+ ERROR("Unable to get platform info.");
+ return 1;
+ }
+
+ if (lock_pages(&ctxt, sizeof(ctxt))) {
+ /* needed for build domctl, but might as well do early */
+ ERROR("Unable to lock ctxt");
+ return 1;
+ }
+
+ /* Load the p2m frame list, plus potential extended info chunk */
+ if ( !hvm )
+ {
+ p2m_frame_list = load_p2m_frame_list(io_fd, &pae_extended_cr3);
+ if ( !p2m_frame_list )
+ goto out;
+ }
+
+ /* We want zeroed memory so use calloc rather than malloc. */
+ p2m = calloc(p2m_size, sizeof(xen_pfn_t));
+ pfn_type = calloc(p2m_size, sizeof(unsigned long));
+ region_mfn = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
+ p2m_batch = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
+
+ if ((p2m == NULL) || (pfn_type == NULL) ||
+ (region_mfn == NULL) || (p2m_batch == NULL)) {
+ ERROR("memory alloc failed");
+ errno = ENOMEM;
+ goto out;
+ }
+
+ if (lock_pages(region_mfn, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
+ ERROR("Could not lock region_mfn");
+ goto out;
+ }
+
+ if (lock_pages(p2m_batch, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
+ ERROR("Could not lock p2m_batch");
+ goto out;
+ }
+
+ /* Get the domain's shared-info frame. */
+ domctl.cmd = XEN_DOMCTL_getdomaininfo;
+ domctl.domain = (domid_t)dom;
+ if (xc_domctl(xc_handle, &domctl) < 0) {
+ ERROR("Could not get information on new domain");
+ goto out;
+ }
+ shared_info_frame = domctl.u.getdomaininfo.shared_info_frame;
+
+ /* Mark all PFNs as invalid; we allocate on demand */
+ for ( pfn = 0; pfn < p2m_size; pfn++ )
+ p2m[pfn] = INVALID_P2M_ENTRY;
+
+ if(!(mmu = xc_init_mmu_updates(xc_handle, dom))) {
+ ERROR("Could not initialise for MMU updates");
+ goto out;
+ }
+
+ DPRINTF("Reloading memory pages: 0%%\n");
+
+ /*
+ * Now simply read each saved frame into its new machine frame.
+ * We uncanonicalise page tables as we go.
+ */
+ prev_pc = 0;
+
+ n = m = 0;
+ while (1) {
+
+ int j, nr_mfns = 0;
+
+ this_pc = (n * 100) / p2m_size;
+ if ( (this_pc - prev_pc) >= 5 )
+ {
+ PPRINTF("\b\b\b\b%3d%%", this_pc);
+ prev_pc = this_pc;
+ }
+
+ if (!read_exact(io_fd, &j, sizeof(int))) {
+ ERROR("Error when reading batch size");
+ goto out;
+ }
+
+ PPRINTF("batch %d\n",j);
+
+ if (j == -1) {
+ verify = 1;
+ DPRINTF("Entering page verify mode\n");
+ continue;
+ }
+
+ if (j == -2) {
+ new_ctxt_format = 1;
+ if (!read_exact(io_fd, &max_vcpu_id, sizeof(int)) ||
+ (max_vcpu_id >= 64) ||
+ !read_exact(io_fd, &vcpumap, sizeof(uint64_t))) {
+ ERROR("Error when reading max_vcpu_id");
+ goto out;
+ }
+ continue;
+ }
+
+ if (j == 0)
+ break; /* our work here is done */
+
+ if (j > MAX_BATCH_SIZE) {
+ ERROR("Max batch size exceeded. Giving up.");
+ goto out;
+ }
+
+ if (!read_exact(io_fd, region_pfn_type, j*sizeof(unsigned long))) {
+ ERROR("Error when reading region pfn types");
+ goto out;
+ }
+
+ /* First pass for this batch: work out how much memory to alloc */
+ nr_mfns = 0;
+ for ( i = 0; i < j; i++ )
+ {
+ unsigned long pfn, pagetype;
+ pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
+ pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
+
+ if ( (pagetype != XEN_DOMCTL_PFINFO_XTAB) &&
+ (p2m[pfn] == INVALID_P2M_ENTRY) )
+ {
+ /* Have a live PFN which hasn't had an MFN allocated */
+ p2m_batch[nr_mfns++] = pfn;
+ }
+ }
+
+
+ /* Now allocate a bunch of mfns for this batch */
+ if (nr_mfns && xc_domain_memory_populate_physmap(
+ xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
+ ERROR("Failed to allocate memory for batch.!\n");
+ errno = ENOMEM;
+ goto out;
+ }
+
+ /* Second pass for this batch: update p2m[] and region_mfn[] */
+ nr_mfns = 0;
+ for ( i = 0; i < j; i++ )
+ {
+ unsigned long pfn, pagetype;
+ pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
+ pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
+
+ if ( pagetype == XEN_DOMCTL_PFINFO_XTAB)
+ region_mfn[i] = ~0UL; /* map will fail but we don't care */
+ else
+ {
+ if (p2m[pfn] == INVALID_P2M_ENTRY) {
+ /* We just allocated a new mfn above; update p2m */
+ p2m[pfn] = p2m_batch[nr_mfns++];
+ nr_pfns++;
+ }
+
+ /* setup region_mfn[] for batch map.
+ * For HVM guests, this interface takes PFNs, not MFNs */
+ region_mfn[i] = hvm ? pfn : p2m[pfn];
+ }
+ }
+
+ /* Map relevant mfns */
+ region_base = xc_map_foreign_batch(
+ xc_handle, dom, PROT_WRITE, region_mfn, j);
+
+ if ( region_base == NULL )
+ {
+ ERROR("map batch failed");
+ goto out;
+ }
+
+ for ( i = 0; i < j; i++ )
+ {
+ void *page;
+ unsigned long pagetype;
+
+ pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
+ pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
+
+ if ( pagetype == XEN_DOMCTL_PFINFO_XTAB )
+ /* a bogus/unmapped page: skip it */
+ continue;
+
+ if ( pfn > p2m_size )
+ {
+ ERROR("pfn out of range");
+ goto out;
+ }
+
+ pfn_type[pfn] = pagetype;
+
+ mfn = p2m[pfn];
+
+ /* In verify mode, we use a copy; otherwise we work in place */
+ page = verify ? (void *)buf : (region_base + i*PAGE_SIZE);
+
+ if (!read_exact(io_fd, page, PAGE_SIZE)) {
+ ERROR("Error when reading page (type was %lx)", pagetype);
+ goto out;
+ }
+
+ pagetype &= XEN_DOMCTL_PFINFO_LTABTYPE_MASK;
+
+ if ( (pagetype >= XEN_DOMCTL_PFINFO_L1TAB) &&
+ (pagetype <= XEN_DOMCTL_PFINFO_L4TAB) )
+ {
+ /*
+ ** A page table page - need to 'uncanonicalize' it, i.e.
+ ** replace all the references to pfns with the corresponding
+ ** mfns for the new domain.
+ **
+ ** On PAE we need to ensure that PGDs are in MFNs < 4G, and
+ ** so we may need to update the p2m after the main loop.
+ ** Hence we defer canonicalization of L1s until then.
+ */
+ if ((pt_levels != 3) ||
+ pae_extended_cr3 ||
+ (pagetype != XEN_DOMCTL_PFINFO_L1TAB)) {
+
+ if (!uncanonicalize_pagetable(xc_handle, dom,
+ pagetype, page)) {
+ /*
+ ** Failing to uncanonicalize a page table can be ok
+ ** under live migration since the pages type may have
+ ** changed by now (and we'll get an update later).
+ */
+ DPRINTF("PT L%ld race on pfn=%08lx mfn=%08lx\n",
+ pagetype >> 28, pfn, mfn);
+ nraces++;
+ continue;
+ }
+ }
+ }
+ else if ( pagetype != XEN_DOMCTL_PFINFO_NOTAB )
+ {
+ ERROR("Bogus page type %lx page table is out of range: "
+ "i=%d p2m_size=%lu", pagetype, i, p2m_size);
+ goto out;
+
+ }
+
+
+ if (verify) {
+
+ int res = memcmp(buf, (region_base + i*PAGE_SIZE), PAGE_SIZE);
+
+ if (res) {
+
+ int v;
+
+ DPRINTF("************** pfn=%lx type=%lx gotcs=%08lx "
+ "actualcs=%08lx\n", pfn, pfn_type[pfn],
+ csum_page(region_base + i*PAGE_SIZE),
+ csum_page(buf));
+
+ for (v = 0; v < 4; v++) {
+
+ unsigned long *p = (unsigned long *)
+ (region_base + i*PAGE_SIZE);
+ if (buf[v] != p[v])
+ DPRINTF(" %d: %08lx %08lx\n", v, buf[v], p[v]);
+ }
+ }
+ }
+
+ if (!hvm
+ && xc_add_mmu_update(xc_handle, mmu,
+ (((unsigned long long)mfn) << PAGE_SHIFT)
+ | MMU_MACHPHYS_UPDATE, pfn)) {
+ ERROR("failed machpys update mfn=%lx pfn=%lx", mfn, pfn);
+ goto out;
+ }
+ } /* end of 'batch' for loop */
+
+ munmap(region_base, j*PAGE_SIZE);
+ n+= j; /* crude stats */
+
+ /*
+ * Discard cache for portion of file read so far up to last
+ * page boundary every 16MB or so.
+ */
+ m += j;
+ if ( m > MAX_PAGECACHE_USAGE )
+ {
+ discard_file_cache(io_fd, 0 /* no flush */);
+ m = 0;
+ }
+ }
+
+ /*
+ * Ensure we flush all machphys updates before potential PAE-specific
+ * reallocations below.
+ */
+ if (!hvm && xc_finish_mmu_updates(xc_handle, mmu)) {
+ ERROR("Error doing finish_mmu_updates()");
+ goto out;
+ }
+
+ DPRINTF("Received all pages (%d races)\n", nraces);
+
+ if ( hvm )
+ {
+ uint32_t rec_len;
+
+ /* Set HVM-specific parameters */
+ if ( !read_exact(io_fd, magic_pfns, sizeof(magic_pfns)) )
+ {
+ ERROR("error reading magic page addresses");
+ goto out;
+ }
+
+ /* These comms pages need to be zeroed at the start of day */
+ if ( xc_clear_domain_page(xc_handle, dom, magic_pfns[0]) ||
+ xc_clear_domain_page(xc_handle, dom, magic_pfns[1]) ||
+ xc_clear_domain_page(xc_handle, dom, magic_pfns[2]) )
+ {
+ ERROR("error zeroing magic pages");
+ goto out;
+ }
+
+ xc_set_hvm_param(xc_handle, dom, HVM_PARAM_IOREQ_PFN, magic_pfns[0]);
+ xc_set_hvm_param(xc_handle, dom, HVM_PARAM_BUFIOREQ_PFN,
magic_pfns[1]);
+ xc_set_hvm_param(xc_handle, dom, HVM_PARAM_STORE_PFN, magic_pfns[2]);
+ xc_set_hvm_param(xc_handle, dom, HVM_PARAM_PAE_ENABLED, pae);
+ xc_set_hvm_param(xc_handle, dom, HVM_PARAM_STORE_EVTCHN, store_evtchn);
+ *store_mfn = magic_pfns[2];
+
+ /* Read vcpu contexts */
+ for (i = 0; i <= max_vcpu_id; i++)
+ {
+ if (!(vcpumap & (1ULL << i)))
+ continue;
+
+ if ( !read_exact(io_fd, &(ctxt), sizeof(ctxt)) )
+ {
+ ERROR("error read vcpu context.\n");
+ goto out;
+ }
+
+ if ( (rc = xc_vcpu_setcontext(xc_handle, dom, i, &ctxt)) )
+ {
+ ERROR("Could not set vcpu context, rc=%d", rc);
+ goto out;
+ }
+ rc = 1;
+ }
+
+ /* Read HVM context */
+ if ( !read_exact(io_fd, &rec_len, sizeof(uint32_t)) )
+ {
+ ERROR("error read hvm context size!\n");
+ goto out;
+ }
+
+ hvm_buf = malloc(rec_len);
+ if ( hvm_buf == NULL )
+ {
+ ERROR("memory alloc for hvm context buffer failed");
+ errno = ENOMEM;
+ goto out;
+ }
+
+ if ( !read_exact(io_fd, hvm_buf, rec_len) )
+ {
+ ERROR("error loading the HVM context");
+ goto out;
+ }
+
+ rc = xc_domain_hvm_setcontext(xc_handle, dom, hvm_buf, rec_len);
+ if ( rc )
+ ERROR("error setting the HVM context");
+
+ goto out;
+ }
+
+ /* Non-HVM guests only from here on */
+
+ if ((pt_levels == 3) && !pae_extended_cr3) {
+
+ /*
+ ** XXX SMH on PAE we need to ensure PGDs are in MFNs < 4G. This
+ ** is a little awkward and involves (a) finding all such PGDs and
+ ** replacing them with 'lowmem' versions; (b) upating the p2m[]
+ ** with the new info; and (c) canonicalizing all the L1s using the
+ ** (potentially updated) p2m[].
+ **
+ ** This is relatively slow (and currently involves two passes through
+ ** the pfn_type[] array), but at least seems to be correct. May wish
+ ** to consider more complex approaches to optimize this later.
+ */
+
+ int j, k;
+
+ /* First pass: find all L3TABs current in > 4G mfns and get new mfns */
+ for ( i = 0; i < p2m_size; i++ )
+ {
+ if ( ((pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) ==
+ XEN_DOMCTL_PFINFO_L3TAB) &&
+ (p2m[i] > 0xfffffUL) )
+ {
+ unsigned long new_mfn;
+ uint64_t l3ptes[4];
+ uint64_t *l3tab;
+
+ l3tab = (uint64_t *)
+ xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
+ PROT_READ, p2m[i]);
+
+ for(j = 0; j < 4; j++)
+ l3ptes[j] = l3tab[j];
+
+ munmap(l3tab, PAGE_SIZE);
+
+ if (!(new_mfn=xc_make_page_below_4G(xc_handle, dom, p2m[i]))) {
+ ERROR("Couldn't get a page below 4GB :-(");
+ goto out;
+ }
+
+ p2m[i] = new_mfn;
+ if (xc_add_mmu_update(xc_handle, mmu,
+ (((unsigned long long)new_mfn)
+ << PAGE_SHIFT) |
+ MMU_MACHPHYS_UPDATE, i)) {
+ ERROR("Couldn't m2p on PAE root pgdir");
+ goto out;
+ }
+
+ l3tab = (uint64_t *)
+ xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
+ PROT_READ | PROT_WRITE, p2m[i]);
+
+ for(j = 0; j < 4; j++)
+ l3tab[j] = l3ptes[j];
+
+ munmap(l3tab, PAGE_SIZE);
+
+ }
+ }
+
+ /* Second pass: find all L1TABs and uncanonicalize them */
+ j = 0;
+
+ for ( i = 0; i < p2m_size; i++ )
+ {
+ if ( ((pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) ==
+ XEN_DOMCTL_PFINFO_L1TAB) )
+ {
+ region_mfn[j] = p2m[i];
+ j++;
+ }
+
+ if(i == (p2m_size-1) || j == MAX_BATCH_SIZE) {
+
+ if (!(region_base = xc_map_foreign_batch(
+ xc_handle, dom, PROT_READ | PROT_WRITE,
+ region_mfn, j))) {
+ ERROR("map batch failed");
+ goto out;
+ }
+
+ for(k = 0; k < j; k++) {
+ if(!uncanonicalize_pagetable(xc_handle, dom,
+ XEN_DOMCTL_PFINFO_L1TAB,
+ region_base + k*PAGE_SIZE)) {
+ ERROR("failed uncanonicalize pt!");
+ goto out;
+ }
+ }
+
+ munmap(region_base, j*PAGE_SIZE);
+ j = 0;
+ }
+ }
+
+ if (xc_finish_mmu_updates(xc_handle, mmu)) {
+ ERROR("Error doing finish_mmu_updates()");
+ goto out;
+ }
+ }
+
+ /*
+ * Pin page tables. Do this after writing to them as otherwise Xen
+ * will barf when doing the type-checking.
+ */
+ nr_pins = 0;
+ for ( i = 0; i < p2m_size; i++ )
+ {
+ if ( (pfn_type[i] & XEN_DOMCTL_PFINFO_LPINTAB) == 0 )
+ continue;
+
+ switch ( pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK )
+ {
+ case XEN_DOMCTL_PFINFO_L1TAB:
+ pin[nr_pins].cmd = MMUEXT_PIN_L1_TABLE;
+ break;
+
+ case XEN_DOMCTL_PFINFO_L2TAB:
+ pin[nr_pins].cmd = MMUEXT_PIN_L2_TABLE;
+ break;
+
+ case XEN_DOMCTL_PFINFO_L3TAB:
+ pin[nr_pins].cmd = MMUEXT_PIN_L3_TABLE;
+ break;
+
+ case XEN_DOMCTL_PFINFO_L4TAB:
+ pin[nr_pins].cmd = MMUEXT_PIN_L4_TABLE;
+ break;
+
+ default:
+ continue;
+ }
+
+ pin[nr_pins].arg1.mfn = p2m[i];
+ nr_pins++;
+
+ /* Batch full? Then flush. */
+ if (nr_pins == MAX_PIN_BATCH) {
+ if (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0) {
+ ERROR("Failed to pin batch of %d page tables", nr_pins);
+ goto out;
+ }
+ nr_pins = 0;
+ }
+ }
+
+ /* Flush final partial batch. */
+ if ((nr_pins != 0) && (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0)) {
+ ERROR("Failed to pin batch of %d page tables", nr_pins);
+ goto out;
+ }
+
+ DPRINTF("\b\b\b\b100%%\n");
+ DPRINTF("Memory reloaded (%ld pages)\n", nr_pfns);
+
+ /* Get the list of PFNs that are not in the psuedo-phys map */
+ {
+ unsigned int count;
+ unsigned long *pfntab;
+ int nr_frees, rc;
+
+ if (!read_exact(io_fd, &count, sizeof(count))) {
+ ERROR("Error when reading pfn count");
+ goto out;
+ }
+
+ if(!(pfntab = malloc(sizeof(unsigned long) * count))) {
+ ERROR("Out of memory");
+ goto out;
+ }
+
+ if (!read_exact(io_fd, pfntab, sizeof(unsigned long)*count)) {
+ ERROR("Error when reading pfntab");
+ goto out;
+ }
+
+ nr_frees = 0;
+ for (i = 0; i < count; i++) {
+
+ unsigned long pfn = pfntab[i];
+
+ if(p2m[pfn] != INVALID_P2M_ENTRY) {
+ /* pfn is not in physmap now, but was at some point during
+ the save/migration process - need to free it */
+ pfntab[nr_frees++] = p2m[pfn];
+ p2m[pfn] = INVALID_P2M_ENTRY; // not in pseudo-physical map
+ }
+ }
+
+ if (nr_frees > 0) {
+
+ struct xen_memory_reservation reservation = {
+ .nr_extents = nr_frees,
+ .extent_order = 0,
+ .domid = dom
+ };
+ set_xen_guest_handle(reservation.extent_start, pfntab);
+
+ if ((rc = xc_memory_op(xc_handle, XENMEM_decrease_reservation,
+ &reservation)) != nr_frees) {
+ ERROR("Could not decrease reservation : %d", rc);
+ goto out;
+ } else
+ DPRINTF("Decreased reservation by %d pages\n", count);
+ }
+ }
+
+ for (i = 0; i <= max_vcpu_id; i++) {
+ if (!(vcpumap & (1ULL << i)))
+ continue;
+
+ if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
+ ERROR("Error when reading ctxt %d", i);
+ goto out;
+ }
+
+ if ( !new_ctxt_format )
+ ctxt.flags |= VGCF_online;
+
+ if (i == 0) {
+ /*
+ * Uncanonicalise the suspend-record frame number and poke
+ * resume record.
+ */
+ pfn = ctxt.user_regs.edx;
+ if ((pfn >= p2m_size) ||
+ (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ ERROR("Suspend record frame number is bad");
+ goto out;
+ }
+ ctxt.user_regs.edx = mfn = p2m[pfn];
+ start_info = xc_map_foreign_range(
+ xc_handle, dom, PAGE_SIZE, PROT_READ | PROT_WRITE, mfn);
+ start_info->nr_pages = p2m_size;
+ start_info->shared_info = shared_info_frame << PAGE_SHIFT;
+ start_info->flags = 0;
+ *store_mfn = start_info->store_mfn = p2m[start_info->store_mfn];
+ start_info->store_evtchn = store_evtchn;
+ start_info->console.domU.mfn = p2m[start_info->console.domU.mfn];
+ start_info->console.domU.evtchn = console_evtchn;
+ *console_mfn = start_info->console.domU.mfn;
+ munmap(start_info, PAGE_SIZE);
+ }
+
+ /* Uncanonicalise each GDT frame number. */
+ if (ctxt.gdt_ents > 8192) {
+ ERROR("GDT entry count out of range");
+ goto out;
+ }
+
+ for (j = 0; (512*j) < ctxt.gdt_ents; j++) {
+ pfn = ctxt.gdt_frames[j];
+ if ((pfn >= p2m_size) ||
+ (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ ERROR("GDT frame number is bad");
+ goto out;
+ }
+ ctxt.gdt_frames[j] = p2m[pfn];
+ }
+
+ /* Uncanonicalise the page table base pointer. */
+ pfn = xen_cr3_to_pfn(ctxt.ctrlreg[3]);
+
+ if (pfn >= p2m_size) {
+ ERROR("PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
+ pfn, p2m_size, pfn_type[pfn]);
+ goto out;
+ }
+
+ if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
+ ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
+ ERROR("PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
+ pfn, p2m_size, pfn_type[pfn],
+ (unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
+ goto out;
+ }
+
+ ctxt.ctrlreg[3] = xen_pfn_to_cr3(p2m[pfn]);
+
+ /* Guest pagetable (x86/64) stored in otherwise-unused CR1. */
+ if ( (pt_levels == 4) && ctxt.ctrlreg[1] )
+ {
+ pfn = xen_cr3_to_pfn(ctxt.ctrlreg[1]);
+
+ if (pfn >= p2m_size) {
+ ERROR("User PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
+ pfn, p2m_size, pfn_type[pfn]);
+ goto out;
+ }
+
+ if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
+ ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
+ ERROR("User PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
+ pfn, p2m_size, pfn_type[pfn],
+ (unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
+ goto out;
+ }
+
+ ctxt.ctrlreg[1] = xen_pfn_to_cr3(p2m[pfn]);
+ }
+
+ domctl.cmd = XEN_DOMCTL_setvcpucontext;
+ domctl.domain = (domid_t)dom;
+ domctl.u.vcpucontext.vcpu = i;
+ set_xen_guest_handle(domctl.u.vcpucontext.ctxt, &ctxt);
+ rc = xc_domctl(xc_handle, &domctl);
+ if (rc != 0) {
+ ERROR("Couldn't build vcpu%d", i);
+ goto out;
+ }
+ rc = 1;
+ }
+
+ if (!read_exact(io_fd, shared_info_page, PAGE_SIZE)) {
+ ERROR("Error when reading shared info page");
+ goto out;
+ }
+
+ /* clear any pending events and the selector */
+ memset(&(shared_info->evtchn_pending[0]), 0,
+ sizeof (shared_info->evtchn_pending));
+ for ( i = 0; i < MAX_VIRT_CPUS; i++ )
+ shared_info->vcpu_info[i].evtchn_pending_sel = 0;
+
+ /* Copy saved contents of shared-info page. No checking needed. */
+ page = xc_map_foreign_range(
+ xc_handle, dom, PAGE_SIZE, PROT_WRITE, shared_info_frame);
+ memcpy(page, shared_info, PAGE_SIZE);
+ munmap(page, PAGE_SIZE);
+
+ /* Uncanonicalise the pfn-to-mfn table frame-number list. */
+ for (i = 0; i < P2M_FL_ENTRIES; i++) {
+ pfn = p2m_frame_list[i];
+ if ((pfn >= p2m_size) || (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
+ ERROR("PFN-to-MFN frame number is bad");
+ goto out;
+ }
+
+ p2m_frame_list[i] = p2m[pfn];
+ }
+
+ /* Copy the P2M we've constructed to the 'live' P2M */
+ if (!(live_p2m = xc_map_foreign_batch(xc_handle, dom, PROT_WRITE,
+ p2m_frame_list, P2M_FL_ENTRIES))) {
+ ERROR("Couldn't map p2m table");
+ goto out;
+ }
+
+ memcpy(live_p2m, p2m, ROUNDUP(p2m_size * sizeof(xen_pfn_t), PAGE_SHIFT));
+ munmap(live_p2m, ROUNDUP(p2m_size * sizeof(xen_pfn_t), PAGE_SHIFT));
+
+ DPRINTF("Domain ready to be built.\n");
+ rc = 0;
+
+ out:
+ if ( (rc != 0) && (dom != 0) )
+ xc_domain_destroy(xc_handle, dom);
+ free(mmu);
+ free(p2m);
+ free(pfn_type);
+ free(hvm_buf);
+
+ /* discard cache for save file */
+ discard_file_cache(io_fd, 1 /*flush*/);
+
+ DPRINTF("Restore exit with rc=%d\n", rc);
+
+ return rc;
+}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xc_hvm_restore.c
--- a/tools/libxc/xc_hvm_restore.c Fri Apr 06 10:06:30 2007 -0600
+++ /dev/null Thu Jan 01 00:00:00 1970 +0000
@@ -1,360 +0,0 @@
-/******************************************************************************
- * xc_hvm_restore.c
- *
- * Restore the state of a HVM guest.
- *
- * Copyright (c) 2003, K A Fraser.
- * Copyright (c) 2006 Intel Corperation
- * rewriten for hvm guest by Zhai Edwin <edwin.zhai@xxxxxxxxx>
- *
- * This program is free software; you can redistribute it and/or modify it
- * under the terms and conditions of the GNU General Public License,
- * version 2, as published by the Free Software Foundation.
- *
- * This program is distributed in the hope it will be useful, but WITHOUT
- * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
- * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
- * more details.
- *
- * You should have received a copy of the GNU General Public License along with
- * this program; if not, write to the Free Software Foundation, Inc., 59 Temple
- * Place - Suite 330, Boston, MA 02111-1307 USA.
- *
- */
-
-#include <stdlib.h>
-#include <unistd.h>
-
-#include "xg_private.h"
-#include "xg_save_restore.h"
-
-#include <xen/hvm/ioreq.h>
-#include <xen/hvm/params.h>
-#include <xen/hvm/e820.h>
-
-/* max mfn of the whole machine */
-static unsigned long max_mfn;
-
-/* virtual starting address of the hypervisor */
-static unsigned long hvirt_start;
-
-/* #levels of page tables used by the currrent guest */
-static unsigned int pt_levels;
-
-/* A list of PFNs that exist, used when allocating memory to the guest */
-static xen_pfn_t *pfns = NULL;
-
-static ssize_t
-read_exact(int fd, void *buf, size_t count)
-{
- int r = 0, s;
- unsigned char *b = buf;
-
- while (r < count) {
- s = read(fd, &b[r], count - r);
- if ((s == -1) && (errno == EINTR))
- continue;
- if (s <= 0) {
- break;
- }
- r += s;
- }
-
- return (r == count) ? 1 : 0;
-}
-
-int xc_hvm_restore(int xc_handle, int io_fd,
- uint32_t dom, unsigned long max_pfn,
- unsigned int store_evtchn, unsigned long *store_mfn,
- unsigned int pae, unsigned int apic)
-{
- DECLARE_DOMCTL;
-
- /* A copy of the CPU context of the guest. */
- vcpu_guest_context_t ctxt;
-
- char *region_base;
-
- unsigned long buf[PAGE_SIZE/sizeof(unsigned long)];
-
- xc_dominfo_t info;
- unsigned int rc = 1, n, i;
- uint32_t rec_len, nr_vcpus;
- uint8_t *hvm_buf = NULL;
- unsigned long long v_end;
- unsigned long shared_page_nr;
-
- unsigned long pfn;
- unsigned int prev_pc, this_pc;
- int verify = 0;
-
- /* Types of the pfns in the current region */
- unsigned long region_pfn_type[MAX_BATCH_SIZE];
-
- /* The size of an array big enough to contain all guest pfns */
- unsigned long pfn_array_size = max_pfn + 1;
-
- /* Number of pages of memory the guest has. *Not* the same as max_pfn. */
- unsigned long nr_pages = max_pfn;
- /* MMIO hole doesn't contain RAM */
- if ( nr_pages >= HVM_BELOW_4G_MMIO_START >> PAGE_SHIFT )
- nr_pages -= HVM_BELOW_4G_MMIO_LENGTH >> PAGE_SHIFT;
- /* VGA hole doesn't contain RAM */
- nr_pages -= 0x20;
-
- /* XXX: Unlikely to be true, but matches previous behaviour. :( */
- v_end = (nr_pages + 0x20) << PAGE_SHIFT;
-
- DPRINTF("xc_hvm_restore:dom=%d, nr_pages=0x%lx, store_evtchn=%d, "
- "*store_mfn=%ld, pae=%u, apic=%u.\n",
- dom, nr_pages, store_evtchn, *store_mfn, pae, apic);
-
- if(!get_platform_info(xc_handle, dom,
- &max_mfn, &hvirt_start, &pt_levels)) {
- ERROR("Unable to get platform info.");
- return 1;
- }
-
- DPRINTF("xc_hvm_restore start: nr_pages = %lx, max_pfn = %lx, "
- "max_mfn = %lx, hvirt_start=%lx, pt_levels=%d\n",
- nr_pages, max_pfn, max_mfn, hvirt_start, pt_levels);
-
- if (mlock(&ctxt, sizeof(ctxt))) {
- /* needed for build dom0 op, but might as well do early */
- ERROR("Unable to mlock ctxt");
- return 1;
- }
-
-
- pfns = malloc(pfn_array_size * sizeof(xen_pfn_t));
- if (pfns == NULL) {
- ERROR("memory alloc failed");
- errno = ENOMEM;
- goto out;
- }
-
- if(xc_domain_setmaxmem(xc_handle, dom, PFN_TO_KB(nr_pages)) != 0) {
- errno = ENOMEM;
- goto out;
- }
-
- for ( i = 0; i < pfn_array_size; i++ )
- pfns[i] = i;
- for ( i = HVM_BELOW_4G_RAM_END >> PAGE_SHIFT; i < pfn_array_size; i++ )
- pfns[i] += HVM_BELOW_4G_MMIO_LENGTH >> PAGE_SHIFT;
-
- /* Allocate memory for HVM guest, skipping VGA hole 0xA0000-0xC0000. */
- rc = xc_domain_memory_populate_physmap(
- xc_handle, dom, (nr_pages > 0xa0) ? 0xa0 : nr_pages,
- 0, 0, &pfns[0x00]);
- if ( (rc == 0) && (nr_pages > 0xc0) )
- rc = xc_domain_memory_populate_physmap(
- xc_handle, dom, nr_pages - 0xa0, 0, 0, &pfns[0xc0]);
- if ( rc != 0 )
- {
- PERROR("Could not allocate memory for HVM guest.\n");
- goto out;
- }
-
-
- /**********XXXXXXXXXXXXXXXX******************/
- if (xc_domain_getinfo(xc_handle, dom, 1, &info) != 1) {
- ERROR("Could not get domain info");
- return 1;
- }
-
- domctl.cmd = XEN_DOMCTL_getdomaininfo;
- domctl.domain = (domid_t)dom;
- if (xc_domctl(xc_handle, &domctl) < 0) {
- ERROR("Could not get information on new domain");
- goto out;
- }
-
- prev_pc = 0;
-
- n = 0;
- while (1) {
-
- int j;
-
- this_pc = (n * 100) / nr_pages;
- if ( (this_pc - prev_pc) >= 5 )
- {
- PPRINTF("\b\b\b\b%3d%%", this_pc);
- prev_pc = this_pc;
- }
-
- if (!read_exact(io_fd, &j, sizeof(int))) {
- ERROR("HVM restore Error when reading batch size");
- goto out;
- }
-
- PPRINTF("batch %d\n",j);
-
- if (j == -1) {
- verify = 1;
- DPRINTF("Entering page verify mode\n");
- continue;
- }
-
- if (j == 0)
- break; /* our work here is done */
-
- if (j > MAX_BATCH_SIZE) {
- ERROR("Max batch size exceeded. Giving up.");
- goto out;
- }
-
- if (!read_exact(io_fd, region_pfn_type, j*sizeof(unsigned long))) {
- ERROR("Error when reading region pfn types");
- goto out;
- }
-
- region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_WRITE, region_pfn_type, j);
-
- for ( i = 0; i < j; i++ )
- {
- void *page;
-
- pfn = region_pfn_type[i];
- if ( pfn & XEN_DOMCTL_PFINFO_LTAB_MASK )
- continue;
-
- if ( pfn > max_pfn )
- {
- ERROR("pfn out of range");
- goto out;
- }
-
- if ( pfn >= 0xa0 && pfn < 0xc0) {
- ERROR("hvm restore:pfn in vga hole");
- goto out;
- }
-
-
- /* In verify mode, we use a copy; otherwise we work in place */
- page = verify ? (void *)buf : (region_base + i*PAGE_SIZE);
-
- if (!read_exact(io_fd, page, PAGE_SIZE)) {
- ERROR("Error when reading page (%x)", i);
- goto out;
- }
-
- if (verify) {
-
- int res = memcmp(buf, (region_base + i*PAGE_SIZE), PAGE_SIZE);
-
- if (res) {
-
- int v;
-
- DPRINTF("************** pfn=%lx gotcs=%08lx "
- "actualcs=%08lx\n", pfn,
- csum_page(region_base + i*PAGE_SIZE),
- csum_page(buf));
-
- for (v = 0; v < 4; v++) {
-
- unsigned long *p = (unsigned long *)
- (region_base + i*PAGE_SIZE);
- if (buf[v] != p[v])
- DPRINTF(" %d: %08lx %08lx\n", v, buf[v], p[v]);
- }
- }
- }
-
- } /* end of 'batch' for loop */
- munmap(region_base, j*PAGE_SIZE);
- n+= j; /* crude stats */
-
- }/*while 1*/
-
- xc_set_hvm_param(xc_handle, dom, HVM_PARAM_PAE_ENABLED, pae);
- xc_set_hvm_param(xc_handle, dom, HVM_PARAM_STORE_EVTCHN, store_evtchn);
-
- if ( v_end > HVM_BELOW_4G_RAM_END )
- shared_page_nr = (HVM_BELOW_4G_RAM_END >> PAGE_SHIFT) - 1;
- else
- shared_page_nr = (v_end >> PAGE_SHIFT) - 1;
-
- /* Ensure we clear these pages */
- if ( xc_clear_domain_page(xc_handle, dom, shared_page_nr) ||
- xc_clear_domain_page(xc_handle, dom, shared_page_nr-1) ||
- xc_clear_domain_page(xc_handle, dom, shared_page_nr-2) ) {
- rc = -1;
- goto out;
- }
-
- xc_set_hvm_param(xc_handle, dom, HVM_PARAM_STORE_PFN, shared_page_nr-1);
- xc_set_hvm_param(xc_handle, dom, HVM_PARAM_BUFIOREQ_PFN, shared_page_nr-2);
- xc_set_hvm_param(xc_handle, dom, HVM_PARAM_IOREQ_PFN, shared_page_nr);
-
- /* caculate the store_mfn , wrong val cause hang when introduceDomain */
- *store_mfn = (v_end >> PAGE_SHIFT) - 2;
- DPRINTF("hvm restore: calculate new store_mfn=0x%lx, v_end=0x%llx.\n",
- *store_mfn, v_end);
-
- if (!read_exact(io_fd, &nr_vcpus, sizeof(uint32_t))) {
- ERROR("error read nr vcpu !\n");
- goto out;
- }
- DPRINTF("hvm restore:get nr_vcpus=%d.\n", nr_vcpus);
-
- for (i =0; i < nr_vcpus; i++) {
- if (!read_exact(io_fd, &rec_len, sizeof(uint32_t))) {
- ERROR("error read vcpu context size!\n");
- goto out;
- }
- if (rec_len != sizeof(ctxt)) {
- ERROR("vcpu context size dismatch!\n");
- goto out;
- }
-
- if (!read_exact(io_fd, &(ctxt), sizeof(ctxt))) {
- ERROR("error read vcpu context.\n");
- goto out;
- }
-
- if ( (rc = xc_vcpu_setcontext(xc_handle, dom, i, &ctxt)) ) {
- ERROR("Could not set vcpu context, rc=%d", rc);
- goto out;
- }
- }
-
- /* restore hvm context including pic/pit/shpage */
- if (!read_exact(io_fd, &rec_len, sizeof(uint32_t))) {
- ERROR("error read hvm context size!\n");
- goto out;
- }
-
- hvm_buf = malloc(rec_len);
- if (hvm_buf == NULL) {
- ERROR("memory alloc for hvm context buffer failed");
- errno = ENOMEM;
- goto out;
- }
-
- if (!read_exact(io_fd, hvm_buf, rec_len)) {
- ERROR("error read hvm buffer!\n");
- goto out;
- }
-
- if (( rc = xc_domain_hvm_setcontext(xc_handle, dom, hvm_buf, rec_len))) {
- ERROR("error set hvm buffer!\n");
- goto out;
- }
-
- rc = 0;
- goto out;
-
- out:
- if ( (rc != 0) && (dom != 0) )
- xc_domain_destroy(xc_handle, dom);
- free(pfns);
- free(hvm_buf);
-
- DPRINTF("Restore exit with rc=%d\n", rc);
-
- return rc;
-}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xc_hvm_save.c
--- a/tools/libxc/xc_hvm_save.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxc/xc_hvm_save.c Fri Apr 06 10:08:30 2007 -0600
@@ -45,36 +45,10 @@
#define DEF_MAX_ITERS 29 /* limit us to 30 times round loop */
#define DEF_MAX_FACTOR 3 /* never send more than 3x nr_pfns */
-/* max mfn of the whole machine */
-static unsigned long max_mfn;
-
-/* virtual starting address of the hypervisor */
-static unsigned long hvirt_start;
-
-/* #levels of page tables used by the current guest */
-static unsigned int pt_levels;
-
/* Shared-memory bitmaps for getting log-dirty bits from qemu */
static unsigned long *qemu_bitmaps[2];
static int qemu_active;
static int qemu_non_active;
-
-int xc_hvm_drain_io(int handle, domid_t dom)
-{
- DECLARE_HYPERCALL;
- xen_hvm_drain_io_t arg;
- int rc;
-
- hypercall.op = __HYPERVISOR_hvm_op;
- hypercall.arg[0] = HVMOP_drain_io;
- hypercall.arg[1] = (unsigned long)&arg;
- arg.domid = dom;
- if ( lock_pages(&arg, sizeof(arg)) != 0 )
- return -1;
- rc = do_xen_hypercall(handle, &hypercall);
- unlock_pages(&arg, sizeof(arg));
- return rc;
-}
/*
** During (live) save/migrate, we maintain a number of bitmaps to track
@@ -291,9 +265,8 @@ int xc_hvm_save(int xc_handle, int io_fd
xc_dominfo_t info;
int rc = 1, i, j, last_iter, iter = 0;
- int live = (flags & XCFLAGS_LIVE);
- int debug = (flags & XCFLAGS_DEBUG);
- int stdvga = (flags & XCFLAGS_STDVGA);
+ int live = !!(flags & XCFLAGS_LIVE);
+ int debug = !!(flags & XCFLAGS_DEBUG);
int sent_last_iter, skip_this_iter;
/* The highest guest-physical frame number used by the current guest */
@@ -302,8 +275,8 @@ int xc_hvm_save(int xc_handle, int io_fd
/* The size of an array big enough to contain all guest pfns */
unsigned long pfn_array_size;
- /* Other magic frames: ioreqs and xenstore comms */
- unsigned long ioreq_pfn, bufioreq_pfn, store_pfn;
+ /* Magic frames: ioreqs and xenstore comms. */
+ uint64_t magic_pfns[3]; /* ioreq_pfn, bufioreq_pfn, store_pfn */
/* A copy of the CPU context of the guest. */
vcpu_guest_context_t ctxt;
@@ -330,122 +303,97 @@ int xc_hvm_save(int xc_handle, int io_fd
xc_shadow_op_stats_t stats;
- unsigned long total_sent = 0;
+ unsigned long total_sent = 0;
+
+ uint64_t vcpumap = 1ULL;
DPRINTF("xc_hvm_save: dom=%d, max_iters=%d, max_factor=%d, flags=0x%x, "
"live=%d, debug=%d.\n", dom, max_iters, max_factor, flags,
live, debug);
/* If no explicit control parameters given, use defaults */
- if(!max_iters)
- max_iters = DEF_MAX_ITERS;
- if(!max_factor)
- max_factor = DEF_MAX_FACTOR;
+ max_iters = max_iters ? : DEF_MAX_ITERS;
+ max_factor = max_factor ? : DEF_MAX_FACTOR;
initialize_mbit_rate();
- if(!get_platform_info(xc_handle, dom,
- &max_mfn, &hvirt_start, &pt_levels)) {
- ERROR("HVM:Unable to get platform info.");
+ if ( xc_domain_getinfo(xc_handle, dom, 1, &info) != 1 )
+ {
+ ERROR("HVM: Could not get domain info");
return 1;
}
-
- if (xc_domain_getinfo(xc_handle, dom, 1, &info) != 1) {
- ERROR("HVM:Could not get domain info");
+ nr_vcpus = info.nr_online_vcpus;
+
+ if ( mlock(&ctxt, sizeof(ctxt)) )
+ {
+ ERROR("HVM: Unable to mlock ctxt");
return 1;
}
- nr_vcpus = info.nr_online_vcpus;
-
- if (mlock(&ctxt, sizeof(ctxt))) {
- ERROR("HVM:Unable to mlock ctxt");
- return 1;
- }
/* Only have to worry about vcpu 0 even for SMP */
- if (xc_vcpu_getcontext(xc_handle, dom, 0, &ctxt)) {
- ERROR("HVM:Could not get vcpu context");
- goto out;
- }
-
- /* cheesy sanity check */
- if ((info.max_memkb >> (PAGE_SHIFT - 10)) > max_mfn) {
- ERROR("Invalid HVM state record -- pfn count out of range: %lu",
- (info.max_memkb >> (PAGE_SHIFT - 10)));
- goto out;
- }
-
- if ( xc_get_hvm_param(xc_handle, dom, HVM_PARAM_STORE_PFN, &store_pfn)
- || xc_get_hvm_param(xc_handle, dom, HVM_PARAM_IOREQ_PFN, &ioreq_pfn)
- || xc_get_hvm_param(xc_handle, dom,
- HVM_PARAM_BUFIOREQ_PFN, &bufioreq_pfn) )
- {
- ERROR("HVM: Could not read magic PFN parameters");
- goto out;
- }
- DPRINTF("saved hvm domain info:max_memkb=0x%lx, max_mfn=0x%lx, "
- "nr_pages=0x%lx\n", info.max_memkb, max_mfn, info.nr_pages);
-
- if (live) {
-
- if (xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_ENABLE_LOGDIRTY,
- NULL, 0, NULL, 0, NULL) < 0) {
+ if ( xc_vcpu_getcontext(xc_handle, dom, 0, &ctxt) )
+ {
+ ERROR("HVM: Could not get vcpu context");
+ goto out;
+ }
+
+ DPRINTF("saved hvm domain info: max_memkb=0x%lx, nr_pages=0x%lx\n",
+ info.max_memkb, info.nr_pages);
+
+ if ( live )
+ {
+ /* Live suspend. Enable log-dirty mode. */
+ if ( xc_shadow_control(xc_handle, dom,
+ XEN_DOMCTL_SHADOW_OP_ENABLE_LOGDIRTY,
+ NULL, 0, NULL, 0, NULL) < 0 )
+ {
ERROR("Couldn't enable shadow mode");
goto out;
}
- last_iter = 0;
-
DPRINTF("hvm domain live migration debug start: logdirty enable.\n");
- } else {
- /* This is a non-live suspend. Issue the call back to get the
- domain suspended */
-
- last_iter = 1;
-
- /* suspend hvm domain */
- if (suspend_and_state(suspend, xc_handle, io_fd, dom, &info, &ctxt)) {
+ }
+ else
+ {
+ /* This is a non-live suspend. Suspend the domain .*/
+ if ( suspend_and_state(suspend, xc_handle, io_fd, dom, &info, &ctxt) )
+ {
ERROR("HVM Domain appears not to have suspended");
goto out;
}
}
- DPRINTF("after 1st handle hvm domain nr_pages=0x%lx, "
+ last_iter = !live;
+
+ max_pfn = xc_memory_op(xc_handle, XENMEM_maximum_gpfn, &dom);
+
+ DPRINTF("after 1st handle hvm domain max_pfn=0x%lx, "
"max_memkb=0x%lx, live=%d.\n",
- info.nr_pages, info.max_memkb, live);
-
- /* Calculate the highest PFN of "normal" memory:
- * HVM memory is sequential except for the VGA and MMIO holes. */
- max_pfn = info.nr_pages - 1;
- /* If the domain has a Cirrus framebuffer and we haven't already
- * suspended qemu-dm, it will have 8MB of framebuffer memory
- * still allocated, which we don't want to copy: qemu will save it
- * for us later */
- if ( live && !stdvga )
- max_pfn -= 0x800;
- /* Skip the VGA hole from 0xa0000 to 0xc0000 */
- max_pfn += 0x20;
- /* Skip the MMIO hole: 256MB just below 4GB */
- if ( max_pfn >= (HVM_BELOW_4G_MMIO_START >> PAGE_SHIFT) )
- max_pfn += (HVM_BELOW_4G_MMIO_LENGTH >> PAGE_SHIFT);
+ max_pfn, info.max_memkb, live);
/* Size of any array that covers 0 ... max_pfn */
pfn_array_size = max_pfn + 1;
+ if ( !write_exact(io_fd, &pfn_array_size, sizeof(unsigned long)) )
+ {
+ ERROR("Error when writing to state file (1)");
+ goto out;
+ }
+
/* pretend we sent all the pages last iteration */
sent_last_iter = pfn_array_size;
/* calculate the power of 2 order of pfn_array_size, e.g.
15->4 16->4 17->5 */
- for (i = pfn_array_size-1, order_nr = 0; i ; i >>= 1, order_nr++)
+ for ( i = pfn_array_size-1, order_nr = 0; i ; i >>= 1, order_nr++ )
continue;
/* Setup to_send / to_fix and to_skip bitmaps */
to_send = malloc(BITMAP_SIZE);
to_skip = malloc(BITMAP_SIZE);
-
- if (live) {
+ if ( live )
+ {
/* Get qemu-dm logging dirty pages too */
void *seg = init_qemu_maps(dom, BITMAP_SIZE);
qemu_bitmaps[0] = seg;
@@ -462,44 +410,40 @@ int xc_hvm_save(int xc_handle, int io_fd
}
hvm_buf = malloc(hvm_buf_size);
- if (!to_send ||!to_skip ||!hvm_buf) {
+ if ( !to_send || !to_skip || !hvm_buf )
+ {
ERROR("Couldn't allocate memory");
goto out;
}
memset(to_send, 0xff, BITMAP_SIZE);
- if (lock_pages(to_send, BITMAP_SIZE)) {
+ if ( lock_pages(to_send, BITMAP_SIZE) )
+ {
ERROR("Unable to lock to_send");
return 1;
}
/* (to fix is local only) */
- if (lock_pages(to_skip, BITMAP_SIZE)) {
+ if ( lock_pages(to_skip, BITMAP_SIZE) )
+ {
ERROR("Unable to lock to_skip");
return 1;
}
analysis_phase(xc_handle, dom, pfn_array_size, to_skip, 0);
-
/* We want zeroed memory so use calloc rather than malloc. */
pfn_batch = calloc(MAX_BATCH_SIZE, sizeof(*pfn_batch));
-
- if (pfn_batch == NULL) {
+ if ( pfn_batch == NULL )
+ {
ERROR("failed to alloc memory for pfn_batch array");
errno = ENOMEM;
goto out;
}
- /* Start writing out the saved-domain record. */
- if (!write_exact(io_fd, &max_pfn, sizeof(unsigned long))) {
- ERROR("write: max_pfn");
- goto out;
- }
-
- while(1) {
-
+ for ( ; ; )
+ {
unsigned int prev_pc, sent_this_iter, N, batch;
iter++;
@@ -510,51 +454,56 @@ int xc_hvm_save(int xc_handle, int io_fd
DPRINTF("Saving HVM domain memory pages: iter %d 0%%", iter);
- while( N < pfn_array_size ){
-
+ while ( N < pfn_array_size )
+ {
unsigned int this_pc = (N * 100) / pfn_array_size;
int rc;
- if ((this_pc - prev_pc) >= 5) {
+ if ( (this_pc - prev_pc) >= 5 )
+ {
DPRINTF("\b\b\b\b%3d%%", this_pc);
prev_pc = this_pc;
}
- /* slightly wasteful to peek the whole array evey time,
- but this is fast enough for the moment. */
- if (!last_iter && (rc = xc_shadow_control(
+ if ( !last_iter )
+ {
+ /* Slightly wasteful to peek the whole array evey time,
+ but this is fast enough for the moment. */
+ rc = xc_shadow_control(
xc_handle, dom, XEN_DOMCTL_SHADOW_OP_PEEK, to_skip,
- pfn_array_size, NULL, 0, NULL)) != pfn_array_size) {
- ERROR("Error peeking HVM shadow bitmap");
- goto out;
- }
-
+ pfn_array_size, NULL, 0, NULL);
+ if ( rc != pfn_array_size )
+ {
+ ERROR("Error peeking HVM shadow bitmap");
+ goto out;
+ }
+ }
/* load pfn_batch[] with the mfn of all the pages we're doing in
this batch. */
- for (batch = 0; batch < MAX_BATCH_SIZE && N < pfn_array_size; N++){
-
+ for ( batch = 0;
+ (batch < MAX_BATCH_SIZE) && (N < pfn_array_size);
+ N++ )
+ {
int n = permute(N, pfn_array_size, order_nr);
- if (0&&debug) {
+ if ( 0 && debug )
DPRINTF("%d pfn= %08lx %d \n",
iter, (unsigned long)n, test_bit(n, to_send));
- }
-
- if (!last_iter && test_bit(n, to_send)&& test_bit(n, to_skip))
+
+ if ( !last_iter &&
+ test_bit(n, to_send) &&
+ test_bit(n, to_skip) )
skip_this_iter++; /* stats keeping */
- if (!((test_bit(n, to_send) && !test_bit(n, to_skip)) ||
- (test_bit(n, to_send) && last_iter)))
+ if ( !((test_bit(n, to_send) && !test_bit(n, to_skip)) ||
+ (test_bit(n, to_send) && last_iter)) )
continue;
/* Skip PFNs that aren't really there */
- if ((n >= 0xa0 && n < 0xc0) /* VGA hole */
- || (n >= (HVM_BELOW_4G_MMIO_START >> PAGE_SHIFT)
- && n < (1ULL << 32) >> PAGE_SHIFT) /* 4G MMIO hole */
- || n == store_pfn
- || n == ioreq_pfn
- || n == bufioreq_pfn)
+ if ( (n >= 0xa0 && n < 0xc0) /* VGA hole */
+ || (n >= (HVM_BELOW_4G_MMIO_START >> PAGE_SHIFT) &&
+ n < (1ULL << 32) >> PAGE_SHIFT) /* 4G MMIO hole */ )
continue;
/*
@@ -568,24 +517,27 @@ int xc_hvm_save(int xc_handle, int io_fd
batch++;
}
- if (batch == 0)
+ if ( batch == 0 )
goto skip; /* vanishingly unlikely... */
- /* map_foreign use pfns now !*/
- if ((region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_READ, pfn_batch, batch)) == 0) {
+ region_base = xc_map_foreign_batch(
+ xc_handle, dom, PROT_READ, pfn_batch, batch);
+ if ( region_base == 0 )
+ {
ERROR("map batch failed");
goto out;
}
/* write num of pfns */
- if(!write_exact(io_fd, &batch, sizeof(unsigned int))) {
+ if ( !write_exact(io_fd, &batch, sizeof(unsigned int)) )
+ {
ERROR("Error when writing to state file (2)");
goto out;
}
/* write all the pfns */
- if(!write_exact(io_fd, pfn_batch, sizeof(unsigned long)*batch)) {
+ if ( !write_exact(io_fd, pfn_batch, sizeof(unsigned long)*batch) )
+ {
ERROR("Error when writing to state file (3)");
goto out;
}
@@ -615,21 +567,23 @@ int xc_hvm_save(int xc_handle, int io_fd
DPRINTF("\r %d: sent %d, skipped %d, ",
iter, sent_this_iter, skip_this_iter );
- if (last_iter) {
+ if ( last_iter )
+ {
print_stats( xc_handle, dom, sent_this_iter, &stats, 1);
-
DPRINTF("Total pages sent= %ld (%.2fx)\n",
total_sent, ((float)total_sent)/pfn_array_size );
}
- if (last_iter && debug){
+ if ( last_iter && debug )
+ {
int minusone = -1;
memset(to_send, 0xff, BITMAP_SIZE);
debug = 0;
DPRINTF("Entering debug resend-all mode\n");
/* send "-1" to put receiver into debug mode */
- if(!write_exact(io_fd, &minusone, sizeof(int))) {
+ if ( !write_exact(io_fd, &minusone, sizeof(int)) )
+ {
ERROR("Error when writing to state file (6)");
goto out;
}
@@ -637,22 +591,22 @@ int xc_hvm_save(int xc_handle, int io_fd
continue;
}
- if (last_iter) break;
-
- if (live) {
-
-
- if(
- ((sent_this_iter > sent_last_iter) && RATE_IS_MAX()) ||
- (iter >= max_iters) ||
- (sent_this_iter+skip_this_iter < 50) ||
- (total_sent > pfn_array_size*max_factor) ) {
-
+ if ( last_iter )
+ break;
+
+ if ( live )
+ {
+ if ( ((sent_this_iter > sent_last_iter) && RATE_IS_MAX()) ||
+ (iter >= max_iters) ||
+ (sent_this_iter+skip_this_iter < 50) ||
+ (total_sent > pfn_array_size*max_factor) )
+ {
DPRINTF("Start last iteration for HVM domain\n");
last_iter = 1;
- if (suspend_and_state(suspend, xc_handle, io_fd, dom, &info,
- &ctxt)) {
+ if ( suspend_and_state(suspend, xc_handle, io_fd, dom, &info,
+ &ctxt))
+ {
ERROR("Domain appears not to have suspended");
goto out;
}
@@ -662,25 +616,30 @@ int xc_hvm_save(int xc_handle, int io_fd
(unsigned long)ctxt.user_regs.edx);
}
- if (xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_CLEAN, to_send,
- pfn_array_size, NULL,
- 0, &stats) != pfn_array_size) {
+ if ( xc_shadow_control(xc_handle, dom,
+ XEN_DOMCTL_SHADOW_OP_CLEAN, to_send,
+ pfn_array_size, NULL,
+ 0, &stats) != pfn_array_size )
+ {
ERROR("Error flushing shadow PT");
goto out;
}
/* Pull in the dirty bits from qemu too */
- if (!last_iter) {
+ if ( !last_iter )
+ {
qemu_active = qemu_non_active;
qemu_non_active = qemu_active ? 0 : 1;
qemu_flip_buffer(dom, qemu_active);
- for (j = 0; j < BITMAP_SIZE / sizeof(unsigned long); j++) {
+ for ( j = 0; j < BITMAP_SIZE / sizeof(unsigned long); j++ )
+ {
to_send[j] |= qemu_bitmaps[qemu_non_active][j];
qemu_bitmaps[qemu_non_active][j] = 0;
}
- } else {
- for (j = 0; j < BITMAP_SIZE / sizeof(unsigned long); j++)
+ }
+ else
+ {
+ for ( j = 0; j < BITMAP_SIZE / sizeof(unsigned long); j++ )
to_send[j] |= qemu_bitmaps[qemu_active][j];
}
@@ -688,61 +647,96 @@ int xc_hvm_save(int xc_handle, int io_fd
print_stats(xc_handle, dom, sent_this_iter, &stats, 1);
}
-
-
} /* end of while 1 */
DPRINTF("All HVM memory is saved\n");
+
+ {
+ struct {
+ int minustwo;
+ int max_vcpu_id;
+ uint64_t vcpumap;
+ } chunk = { -2, info.max_vcpu_id };
+
+ if (info.max_vcpu_id >= 64) {
+ ERROR("Too many VCPUS in guest!");
+ goto out;
+ }
+
+ for (i = 1; i <= info.max_vcpu_id; i++) {
+ xc_vcpuinfo_t vinfo;
+ if ((xc_vcpu_getinfo(xc_handle, dom, i, &vinfo) == 0) &&
+ vinfo.online)
+ vcpumap |= 1ULL << i;
+ }
+
+ chunk.vcpumap = vcpumap;
+ if(!write_exact(io_fd, &chunk, sizeof(chunk))) {
+ ERROR("Error when writing to state file (errno %d)", errno);
+ goto out;
+ }
+ }
/* Zero terminate */
i = 0;
- if (!write_exact(io_fd, &i, sizeof(int))) {
+ if ( !write_exact(io_fd, &i, sizeof(int)) )
+ {
ERROR("Error when writing to state file (6)");
goto out;
}
-
- /* save vcpu/vmcs context */
- if (!write_exact(io_fd, &nr_vcpus, sizeof(uint32_t))) {
- ERROR("error write nr vcpus");
- goto out;
- }
-
- /*XXX: need a online map to exclude down cpu */
- for (i = 0; i < nr_vcpus; i++) {
-
- if (xc_vcpu_getcontext(xc_handle, dom, i, &ctxt)) {
+ /* Save magic-page locations. */
+ memset(magic_pfns, 0, sizeof(magic_pfns));
+ xc_get_hvm_param(xc_handle, dom, HVM_PARAM_IOREQ_PFN,
+ (unsigned long *)&magic_pfns[0]);
+ xc_get_hvm_param(xc_handle, dom, HVM_PARAM_BUFIOREQ_PFN,
+ (unsigned long *)&magic_pfns[1]);
+ xc_get_hvm_param(xc_handle, dom, HVM_PARAM_STORE_PFN,
+ (unsigned long *)&magic_pfns[2]);
+ if ( !write_exact(io_fd, magic_pfns, sizeof(magic_pfns)) )
+ {
+ ERROR("Error when writing to state file (7)");
+ goto out;
+ }
+
+ /* save vcpu/vmcs contexts */
+ for ( i = 0; i < nr_vcpus; i++ )
+ {
+ if (!(vcpumap & (1ULL << i)))
+ continue;
+
+ if ( xc_vcpu_getcontext(xc_handle, dom, i, &ctxt) )
+ {
ERROR("HVM:Could not get vcpu context");
goto out;
}
- rec_size = sizeof(ctxt);
- DPRINTF("write %d vcpucontext of total %d.\n", i, nr_vcpus);
- if (!write_exact(io_fd, &rec_size, sizeof(uint32_t))) {
- ERROR("error write vcpu ctxt size");
+ DPRINTF("write vcpu %d context.\n", i);
+ if ( !write_exact(io_fd, &(ctxt), sizeof(ctxt)) )
+ {
+ ERROR("write vcpu context failed!\n");
goto out;
}
-
- if (!write_exact(io_fd, &(ctxt), sizeof(ctxt)) ) {
- ERROR("write vmcs failed!\n");
- goto out;
- }
}
if ( (rec_size = xc_domain_hvm_getcontext(xc_handle, dom, hvm_buf,
- hvm_buf_size)) == -1) {
+ hvm_buf_size)) == -1 )
+ {
ERROR("HVM:Could not get hvm buffer");
goto out;
}
- if (!write_exact(io_fd, &rec_size, sizeof(uint32_t))) {
+ if ( !write_exact(io_fd, &rec_size, sizeof(uint32_t)) )
+ {
ERROR("error write hvm buffer size");
goto out;
}
- if ( !write_exact(io_fd, hvm_buf, rec_size) ) {
+ if ( !write_exact(io_fd, hvm_buf, rec_size) )
+ {
ERROR("write HVM info failed!\n");
+ goto out;
}
/* Success! */
@@ -750,12 +744,11 @@ int xc_hvm_save(int xc_handle, int io_fd
out:
- if (live) {
- if(xc_shadow_control(xc_handle, dom,
- XEN_DOMCTL_SHADOW_OP_OFF,
- NULL, 0, NULL, 0, NULL) < 0) {
+ if ( live )
+ {
+ if ( xc_shadow_control(xc_handle, dom, XEN_DOMCTL_SHADOW_OP_OFF,
+ NULL, 0, NULL, 0, NULL) < 0 )
DPRINTF("Warning - couldn't disable shadow mode");
- }
}
free(hvm_buf);
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xc_linux_restore.c
--- a/tools/libxc/xc_linux_restore.c Fri Apr 06 10:06:30 2007 -0600
+++ /dev/null Thu Jan 01 00:00:00 1970 +0000
@@ -1,962 +0,0 @@
-/******************************************************************************
- * xc_linux_restore.c
- *
- * Restore the state of a Linux session.
- *
- * Copyright (c) 2003, K A Fraser.
- */
-
-#include <stdlib.h>
-#include <unistd.h>
-
-#include "xg_private.h"
-#include "xg_save_restore.h"
-#include "xc_dom.h"
-
-/* max mfn of the current host machine */
-static unsigned long max_mfn;
-
-/* virtual starting address of the hypervisor */
-static unsigned long hvirt_start;
-
-/* #levels of page tables used by the current guest */
-static unsigned int pt_levels;
-
-/* number of pfns this guest has (i.e. number of entries in the P2M) */
-static unsigned long p2m_size;
-
-/* number of 'in use' pfns in the guest (i.e. #P2M entries with a valid mfn) */
-static unsigned long nr_pfns;
-
-/* largest possible value of nr_pfns (i.e. domain's maximum memory size) */
-static unsigned long max_nr_pfns;
-
-/* Live mapping of the table mapping each PFN to its current MFN. */
-static xen_pfn_t *live_p2m = NULL;
-
-/* A table mapping each PFN to its new MFN. */
-static xen_pfn_t *p2m = NULL;
-
-/* A table of P2M mappings in the current region */
-static xen_pfn_t *p2m_batch = NULL;
-
-static ssize_t
-read_exact(int fd, void *buf, size_t count)
-{
- int r = 0, s;
- unsigned char *b = buf;
-
- while (r < count) {
- s = read(fd, &b[r], count - r);
- if ((s == -1) && (errno == EINTR))
- continue;
- if (s <= 0) {
- break;
- }
- r += s;
- }
-
- return (r == count) ? 1 : 0;
-}
-
-/*
-** In the state file (or during transfer), all page-table pages are
-** converted into a 'canonical' form where references to actual mfns
-** are replaced with references to the corresponding pfns.
-** This function inverts that operation, replacing the pfn values with
-** the (now known) appropriate mfn values.
-*/
-static int uncanonicalize_pagetable(int xc_handle, uint32_t dom,
- unsigned long type, void *page)
-{
- int i, pte_last;
- unsigned long pfn;
- uint64_t pte;
- int nr_mfns = 0;
-
- pte_last = PAGE_SIZE / ((pt_levels == 2)? 4 : 8);
-
- /* First pass: work out how many (if any) MFNs we need to alloc */
- for(i = 0; i < pte_last; i++) {
-
- if(pt_levels == 2)
- pte = ((uint32_t *)page)[i];
- else
- pte = ((uint64_t *)page)[i];
-
- /* XXX SMH: below needs fixing for PROT_NONE etc */
- if(!(pte & _PAGE_PRESENT))
- continue;
-
- pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
-
- if(pfn >= p2m_size) {
- /* This "page table page" is probably not one; bail. */
- ERROR("Frame number in type %lu page table is out of range: "
- "i=%d pfn=0x%lx p2m_size=%lu",
- type >> 28, i, pfn, p2m_size);
- return 0;
- }
-
- if(p2m[pfn] == INVALID_P2M_ENTRY) {
- /* Have a 'valid' PFN without a matching MFN - need to alloc */
- p2m_batch[nr_mfns++] = pfn;
- }
- }
-
-
- /* Alllocate the requistite number of mfns */
- if (nr_mfns && xc_domain_memory_populate_physmap(
- xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
- ERROR("Failed to allocate memory for batch.!\n");
- errno = ENOMEM;
- return 0;
- }
-
- /* Second pass: uncanonicalize each present PTE */
- nr_mfns = 0;
- for(i = 0; i < pte_last; i++) {
-
- if(pt_levels == 2)
- pte = ((uint32_t *)page)[i];
- else
- pte = ((uint64_t *)page)[i];
-
- /* XXX SMH: below needs fixing for PROT_NONE etc */
- if(!(pte & _PAGE_PRESENT))
- continue;
-
- pfn = (pte >> PAGE_SHIFT) & MFN_MASK_X86;
-
- if(p2m[pfn] == INVALID_P2M_ENTRY)
- p2m[pfn] = p2m_batch[nr_mfns++];
-
- pte &= ~MADDR_MASK_X86;
- pte |= (uint64_t)p2m[pfn] << PAGE_SHIFT;
-
- if(pt_levels == 2)
- ((uint32_t *)page)[i] = (uint32_t)pte;
- else
- ((uint64_t *)page)[i] = (uint64_t)pte;
- }
-
- return 1;
-}
-
-
-int xc_linux_restore(int xc_handle, int io_fd, uint32_t dom,
- unsigned long p2msize, unsigned long maxnrpfns,
- unsigned int store_evtchn, unsigned long *store_mfn,
- unsigned int console_evtchn, unsigned long *console_mfn)
-{
- DECLARE_DOMCTL;
- int rc = 1, i, j, n, m, pae_extended_cr3 = 0;
- unsigned long mfn, pfn;
- unsigned int prev_pc, this_pc;
- int verify = 0;
- int nraces = 0;
-
- /* The new domain's shared-info frame number. */
- unsigned long shared_info_frame;
- unsigned char shared_info_page[PAGE_SIZE]; /* saved contents from file */
- shared_info_t *shared_info = (shared_info_t *)shared_info_page;
-
- /* A copy of the CPU context of the guest. */
- vcpu_guest_context_t ctxt;
-
- /* A table containing the type of each PFN (/not/ MFN!). */
- unsigned long *pfn_type = NULL;
-
- /* A table of MFNs to map in the current region */
- xen_pfn_t *region_mfn = NULL;
-
- /* Types of the pfns in the current region */
- unsigned long region_pfn_type[MAX_BATCH_SIZE];
-
- /* A temporary mapping, and a copy, of one frame of guest memory. */
- unsigned long *page = NULL;
-
- /* A copy of the pfn-to-mfn table frame list. */
- xen_pfn_t *p2m_frame_list = NULL;
-
- /* A temporary mapping of the guest's start_info page. */
- start_info_t *start_info;
-
- /* Our mapping of the current region (batch) */
- char *region_base;
-
- xc_mmu_t *mmu = NULL;
-
- /* used by debug verify code */
- unsigned long buf[PAGE_SIZE/sizeof(unsigned long)];
-
- struct mmuext_op pin[MAX_PIN_BATCH];
- unsigned int nr_pins;
-
- uint64_t vcpumap = 1ULL;
- unsigned int max_vcpu_id = 0;
- int new_ctxt_format = 0;
-
- p2m_size = p2msize;
- max_nr_pfns = maxnrpfns;
-
- /* For info only */
- nr_pfns = 0;
-
- DPRINTF("xc_linux_restore start: p2m_size = %lx\n", p2m_size);
-
- /*
- * XXX For now, 32bit dom0's can only save/restore 32bit domUs
- * on 64bit hypervisors.
- */
- memset(&domctl, 0, sizeof(domctl));
- domctl.domain = dom;
- domctl.cmd = XEN_DOMCTL_set_address_size;
- domctl.u.address_size.size = sizeof(unsigned long) * 8;
- rc = do_domctl(xc_handle, &domctl);
- if ( rc != 0 ) {
- ERROR("Unable to set guest address size.");
- goto out;
- }
-
- if(!get_platform_info(xc_handle, dom,
- &max_mfn, &hvirt_start, &pt_levels)) {
- ERROR("Unable to get platform info.");
- return 1;
- }
-
- if (lock_pages(&ctxt, sizeof(ctxt))) {
- /* needed for build domctl, but might as well do early */
- ERROR("Unable to lock ctxt");
- return 1;
- }
-
- if (!(p2m_frame_list = malloc(P2M_FL_SIZE))) {
- ERROR("Couldn't allocate p2m_frame_list array");
- goto out;
- }
-
- /* Read first entry of P2M list, or extended-info signature (~0UL). */
- if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
- ERROR("read extended-info signature failed");
- goto out;
- }
-
- if (p2m_frame_list[0] == ~0UL) {
- uint32_t tot_bytes;
-
- /* Next 4 bytes: total size of following extended info. */
- if (!read_exact(io_fd, &tot_bytes, sizeof(tot_bytes))) {
- ERROR("read extended-info size failed");
- goto out;
- }
-
- while (tot_bytes) {
- uint32_t chunk_bytes;
- char chunk_sig[4];
-
- /* 4-character chunk signature + 4-byte remaining chunk size. */
- if (!read_exact(io_fd, chunk_sig, sizeof(chunk_sig)) ||
- !read_exact(io_fd, &chunk_bytes, sizeof(chunk_bytes))) {
- ERROR("read extended-info chunk signature failed");
- goto out;
- }
- tot_bytes -= 8;
-
- /* VCPU context structure? */
- if (!strncmp(chunk_sig, "vcpu", 4)) {
- if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
- ERROR("read extended-info vcpu context failed");
- goto out;
- }
- tot_bytes -= sizeof(struct vcpu_guest_context);
- chunk_bytes -= sizeof(struct vcpu_guest_context);
-
- if (ctxt.vm_assist & (1UL << VMASST_TYPE_pae_extended_cr3))
- pae_extended_cr3 = 1;
- }
-
- /* Any remaining bytes of this chunk: read and discard. */
- while (chunk_bytes) {
- unsigned long sz = chunk_bytes;
- if ( sz > P2M_FL_SIZE )
- sz = P2M_FL_SIZE;
- if (!read_exact(io_fd, p2m_frame_list, sz)) {
- ERROR("read-and-discard extended-info chunk bytes failed");
- goto out;
- }
- chunk_bytes -= sz;
- tot_bytes -= sz;
- }
- }
-
- /* Now read the real first entry of P2M list. */
- if (!read_exact(io_fd, p2m_frame_list, sizeof(long))) {
- ERROR("read first entry of p2m_frame_list failed");
- goto out;
- }
- }
-
- /* First entry is already read into the p2m array. */
- if (!read_exact(io_fd, &p2m_frame_list[1], P2M_FL_SIZE - sizeof(long))) {
- ERROR("read p2m_frame_list failed");
- goto out;
- }
-
- /* We want zeroed memory so use calloc rather than malloc. */
- p2m = calloc(p2m_size, sizeof(xen_pfn_t));
- pfn_type = calloc(p2m_size, sizeof(unsigned long));
- region_mfn = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
- p2m_batch = calloc(MAX_BATCH_SIZE, sizeof(xen_pfn_t));
-
- if ((p2m == NULL) || (pfn_type == NULL) ||
- (region_mfn == NULL) || (p2m_batch == NULL)) {
- ERROR("memory alloc failed");
- errno = ENOMEM;
- goto out;
- }
-
- if (lock_pages(region_mfn, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
- ERROR("Could not lock region_mfn");
- goto out;
- }
-
- if (lock_pages(p2m_batch, sizeof(xen_pfn_t) * MAX_BATCH_SIZE)) {
- ERROR("Could not lock p2m_batch");
- goto out;
- }
-
- /* Get the domain's shared-info frame. */
- domctl.cmd = XEN_DOMCTL_getdomaininfo;
- domctl.domain = (domid_t)dom;
- if (xc_domctl(xc_handle, &domctl) < 0) {
- ERROR("Could not get information on new domain");
- goto out;
- }
- shared_info_frame = domctl.u.getdomaininfo.shared_info_frame;
-
- if (xc_domain_setmaxmem(xc_handle, dom, PFN_TO_KB(max_nr_pfns)) != 0) {
- errno = ENOMEM;
- goto out;
- }
-
- /* Mark all PFNs as invalid; we allocate on demand */
- for ( pfn = 0; pfn < p2m_size; pfn++ )
- p2m[pfn] = INVALID_P2M_ENTRY;
-
- if(!(mmu = xc_init_mmu_updates(xc_handle, dom))) {
- ERROR("Could not initialise for MMU updates");
- goto out;
- }
-
- DPRINTF("Reloading memory pages: 0%%\n");
-
- /*
- * Now simply read each saved frame into its new machine frame.
- * We uncanonicalise page tables as we go.
- */
- prev_pc = 0;
-
- n = m = 0;
- while (1) {
-
- int j, nr_mfns = 0;
-
- this_pc = (n * 100) / p2m_size;
- if ( (this_pc - prev_pc) >= 5 )
- {
- PPRINTF("\b\b\b\b%3d%%", this_pc);
- prev_pc = this_pc;
- }
-
- if (!read_exact(io_fd, &j, sizeof(int))) {
- ERROR("Error when reading batch size");
- goto out;
- }
-
- PPRINTF("batch %d\n",j);
-
- if (j == -1) {
- verify = 1;
- DPRINTF("Entering page verify mode\n");
- continue;
- }
-
- if (j == -2) {
- new_ctxt_format = 1;
- if (!read_exact(io_fd, &max_vcpu_id, sizeof(int)) ||
- (max_vcpu_id >= 64) ||
- !read_exact(io_fd, &vcpumap, sizeof(uint64_t))) {
- ERROR("Error when reading max_vcpu_id");
- goto out;
- }
- continue;
- }
-
- if (j == 0)
- break; /* our work here is done */
-
- if (j > MAX_BATCH_SIZE) {
- ERROR("Max batch size exceeded. Giving up.");
- goto out;
- }
-
- if (!read_exact(io_fd, region_pfn_type, j*sizeof(unsigned long))) {
- ERROR("Error when reading region pfn types");
- goto out;
- }
-
- /* First pass for this batch: work out how much memory to alloc */
- nr_mfns = 0;
- for ( i = 0; i < j; i++ )
- {
- unsigned long pfn, pagetype;
- pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
- pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
-
- if ( (pagetype != XEN_DOMCTL_PFINFO_XTAB) &&
- (p2m[pfn] == INVALID_P2M_ENTRY) )
- {
- /* Have a live PFN which hasn't had an MFN allocated */
- p2m_batch[nr_mfns++] = pfn;
- }
- }
-
-
- /* Now allocate a bunch of mfns for this batch */
- if (nr_mfns && xc_domain_memory_populate_physmap(
- xc_handle, dom, nr_mfns, 0, 0, p2m_batch) != 0) {
- ERROR("Failed to allocate memory for batch.!\n");
- errno = ENOMEM;
- goto out;
- }
-
- /* Second pass for this batch: update p2m[] and region_mfn[] */
- nr_mfns = 0;
- for ( i = 0; i < j; i++ )
- {
- unsigned long pfn, pagetype;
- pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
- pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
-
- if ( pagetype == XEN_DOMCTL_PFINFO_XTAB)
- region_mfn[i] = ~0UL; /* map will fail but we don't care */
- else
- {
- if (p2m[pfn] == INVALID_P2M_ENTRY) {
- /* We just allocated a new mfn above; update p2m */
- p2m[pfn] = p2m_batch[nr_mfns++];
- nr_pfns++;
- }
-
- /* setup region_mfn[] for batch map */
- region_mfn[i] = p2m[pfn];
- }
- }
-
- /* Map relevant mfns */
- region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_WRITE, region_mfn, j);
-
- if ( region_base == NULL )
- {
- ERROR("map batch failed");
- goto out;
- }
-
- for ( i = 0; i < j; i++ )
- {
- void *page;
- unsigned long pagetype;
-
- pfn = region_pfn_type[i] & ~XEN_DOMCTL_PFINFO_LTAB_MASK;
- pagetype = region_pfn_type[i] & XEN_DOMCTL_PFINFO_LTAB_MASK;
-
- if ( pagetype == XEN_DOMCTL_PFINFO_XTAB )
- /* a bogus/unmapped page: skip it */
- continue;
-
- if ( pfn > p2m_size )
- {
- ERROR("pfn out of range");
- goto out;
- }
-
- pfn_type[pfn] = pagetype;
-
- mfn = p2m[pfn];
-
- /* In verify mode, we use a copy; otherwise we work in place */
- page = verify ? (void *)buf : (region_base + i*PAGE_SIZE);
-
- if (!read_exact(io_fd, page, PAGE_SIZE)) {
- ERROR("Error when reading page (type was %lx)", pagetype);
- goto out;
- }
-
- pagetype &= XEN_DOMCTL_PFINFO_LTABTYPE_MASK;
-
- if ( (pagetype >= XEN_DOMCTL_PFINFO_L1TAB) &&
- (pagetype <= XEN_DOMCTL_PFINFO_L4TAB) )
- {
- /*
- ** A page table page - need to 'uncanonicalize' it, i.e.
- ** replace all the references to pfns with the corresponding
- ** mfns for the new domain.
- **
- ** On PAE we need to ensure that PGDs are in MFNs < 4G, and
- ** so we may need to update the p2m after the main loop.
- ** Hence we defer canonicalization of L1s until then.
- */
- if ((pt_levels != 3) ||
- pae_extended_cr3 ||
- (pagetype != XEN_DOMCTL_PFINFO_L1TAB)) {
-
- if (!uncanonicalize_pagetable(xc_handle, dom,
- pagetype, page)) {
- /*
- ** Failing to uncanonicalize a page table can be ok
- ** under live migration since the pages type may have
- ** changed by now (and we'll get an update later).
- */
- DPRINTF("PT L%ld race on pfn=%08lx mfn=%08lx\n",
- pagetype >> 28, pfn, mfn);
- nraces++;
- continue;
- }
- }
- }
- else if ( pagetype != XEN_DOMCTL_PFINFO_NOTAB )
- {
- ERROR("Bogus page type %lx page table is out of range: "
- "i=%d p2m_size=%lu", pagetype, i, p2m_size);
- goto out;
-
- }
-
-
- if (verify) {
-
- int res = memcmp(buf, (region_base + i*PAGE_SIZE), PAGE_SIZE);
-
- if (res) {
-
- int v;
-
- DPRINTF("************** pfn=%lx type=%lx gotcs=%08lx "
- "actualcs=%08lx\n", pfn, pfn_type[pfn],
- csum_page(region_base + i*PAGE_SIZE),
- csum_page(buf));
-
- for (v = 0; v < 4; v++) {
-
- unsigned long *p = (unsigned long *)
- (region_base + i*PAGE_SIZE);
- if (buf[v] != p[v])
- DPRINTF(" %d: %08lx %08lx\n", v, buf[v], p[v]);
- }
- }
- }
-
- if (xc_add_mmu_update(xc_handle, mmu,
- (((unsigned long long)mfn) << PAGE_SHIFT)
- | MMU_MACHPHYS_UPDATE, pfn)) {
- ERROR("failed machpys update mfn=%lx pfn=%lx", mfn, pfn);
- goto out;
- }
- } /* end of 'batch' for loop */
-
- munmap(region_base, j*PAGE_SIZE);
- n+= j; /* crude stats */
-
- /*
- * Discard cache for portion of file read so far up to last
- * page boundary every 16MB or so.
- */
- m += j;
- if ( m > MAX_PAGECACHE_USAGE )
- {
- discard_file_cache(io_fd, 0 /* no flush */);
- m = 0;
- }
- }
-
- /*
- * Ensure we flush all machphys updates before potential PAE-specific
- * reallocations below.
- */
- if (xc_finish_mmu_updates(xc_handle, mmu)) {
- ERROR("Error doing finish_mmu_updates()");
- goto out;
- }
-
- DPRINTF("Received all pages (%d races)\n", nraces);
-
- if ((pt_levels == 3) && !pae_extended_cr3) {
-
- /*
- ** XXX SMH on PAE we need to ensure PGDs are in MFNs < 4G. This
- ** is a little awkward and involves (a) finding all such PGDs and
- ** replacing them with 'lowmem' versions; (b) upating the p2m[]
- ** with the new info; and (c) canonicalizing all the L1s using the
- ** (potentially updated) p2m[].
- **
- ** This is relatively slow (and currently involves two passes through
- ** the pfn_type[] array), but at least seems to be correct. May wish
- ** to consider more complex approaches to optimize this later.
- */
-
- int j, k;
-
- /* First pass: find all L3TABs current in > 4G mfns and get new mfns */
- for ( i = 0; i < p2m_size; i++ )
- {
- if ( ((pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) ==
- XEN_DOMCTL_PFINFO_L3TAB) &&
- (p2m[i] > 0xfffffUL) )
- {
- unsigned long new_mfn;
- uint64_t l3ptes[4];
- uint64_t *l3tab;
-
- l3tab = (uint64_t *)
- xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
- PROT_READ, p2m[i]);
-
- for(j = 0; j < 4; j++)
- l3ptes[j] = l3tab[j];
-
- munmap(l3tab, PAGE_SIZE);
-
- if (!(new_mfn=xc_make_page_below_4G(xc_handle, dom, p2m[i]))) {
- ERROR("Couldn't get a page below 4GB :-(");
- goto out;
- }
-
- p2m[i] = new_mfn;
- if (xc_add_mmu_update(xc_handle, mmu,
- (((unsigned long long)new_mfn)
- << PAGE_SHIFT) |
- MMU_MACHPHYS_UPDATE, i)) {
- ERROR("Couldn't m2p on PAE root pgdir");
- goto out;
- }
-
- l3tab = (uint64_t *)
- xc_map_foreign_range(xc_handle, dom, PAGE_SIZE,
- PROT_READ | PROT_WRITE, p2m[i]);
-
- for(j = 0; j < 4; j++)
- l3tab[j] = l3ptes[j];
-
- munmap(l3tab, PAGE_SIZE);
-
- }
- }
-
- /* Second pass: find all L1TABs and uncanonicalize them */
- j = 0;
-
- for ( i = 0; i < p2m_size; i++ )
- {
- if ( ((pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) ==
- XEN_DOMCTL_PFINFO_L1TAB) )
- {
- region_mfn[j] = p2m[i];
- j++;
- }
-
- if(i == (p2m_size-1) || j == MAX_BATCH_SIZE) {
-
- if (!(region_base = xc_map_foreign_batch(
- xc_handle, dom, PROT_READ | PROT_WRITE,
- region_mfn, j))) {
- ERROR("map batch failed");
- goto out;
- }
-
- for(k = 0; k < j; k++) {
- if(!uncanonicalize_pagetable(xc_handle, dom,
- XEN_DOMCTL_PFINFO_L1TAB,
- region_base + k*PAGE_SIZE)) {
- ERROR("failed uncanonicalize pt!");
- goto out;
- }
- }
-
- munmap(region_base, j*PAGE_SIZE);
- j = 0;
- }
- }
-
- if (xc_finish_mmu_updates(xc_handle, mmu)) {
- ERROR("Error doing finish_mmu_updates()");
- goto out;
- }
- }
-
- /*
- * Pin page tables. Do this after writing to them as otherwise Xen
- * will barf when doing the type-checking.
- */
- nr_pins = 0;
- for ( i = 0; i < p2m_size; i++ )
- {
- if ( (pfn_type[i] & XEN_DOMCTL_PFINFO_LPINTAB) == 0 )
- continue;
-
- switch ( pfn_type[i] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK )
- {
- case XEN_DOMCTL_PFINFO_L1TAB:
- pin[nr_pins].cmd = MMUEXT_PIN_L1_TABLE;
- break;
-
- case XEN_DOMCTL_PFINFO_L2TAB:
- pin[nr_pins].cmd = MMUEXT_PIN_L2_TABLE;
- break;
-
- case XEN_DOMCTL_PFINFO_L3TAB:
- pin[nr_pins].cmd = MMUEXT_PIN_L3_TABLE;
- break;
-
- case XEN_DOMCTL_PFINFO_L4TAB:
- pin[nr_pins].cmd = MMUEXT_PIN_L4_TABLE;
- break;
-
- default:
- continue;
- }
-
- pin[nr_pins].arg1.mfn = p2m[i];
- nr_pins++;
-
- /* Batch full? Then flush. */
- if (nr_pins == MAX_PIN_BATCH) {
- if (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0) {
- ERROR("Failed to pin batch of %d page tables", nr_pins);
- goto out;
- }
- nr_pins = 0;
- }
- }
-
- /* Flush final partial batch. */
- if ((nr_pins != 0) && (xc_mmuext_op(xc_handle, pin, nr_pins, dom) < 0)) {
- ERROR("Failed to pin batch of %d page tables", nr_pins);
- goto out;
- }
-
- DPRINTF("\b\b\b\b100%%\n");
- DPRINTF("Memory reloaded (%ld pages of max %ld)\n", nr_pfns, max_nr_pfns);
-
- /* Get the list of PFNs that are not in the psuedo-phys map */
- {
- unsigned int count;
- unsigned long *pfntab;
- int nr_frees, rc;
-
- if (!read_exact(io_fd, &count, sizeof(count))) {
- ERROR("Error when reading pfn count");
- goto out;
- }
-
- if(!(pfntab = malloc(sizeof(unsigned long) * count))) {
- ERROR("Out of memory");
- goto out;
- }
-
- if (!read_exact(io_fd, pfntab, sizeof(unsigned long)*count)) {
- ERROR("Error when reading pfntab");
- goto out;
- }
-
- nr_frees = 0;
- for (i = 0; i < count; i++) {
-
- unsigned long pfn = pfntab[i];
-
- if(p2m[pfn] != INVALID_P2M_ENTRY) {
- /* pfn is not in physmap now, but was at some point during
- the save/migration process - need to free it */
- pfntab[nr_frees++] = p2m[pfn];
- p2m[pfn] = INVALID_P2M_ENTRY; // not in pseudo-physical map
- }
- }
-
- if (nr_frees > 0) {
-
- struct xen_memory_reservation reservation = {
- .nr_extents = nr_frees,
- .extent_order = 0,
- .domid = dom
- };
- set_xen_guest_handle(reservation.extent_start, pfntab);
-
- if ((rc = xc_memory_op(xc_handle, XENMEM_decrease_reservation,
- &reservation)) != nr_frees) {
- ERROR("Could not decrease reservation : %d", rc);
- goto out;
- } else
- DPRINTF("Decreased reservation by %d pages\n", count);
- }
- }
-
- for (i = 0; i <= max_vcpu_id; i++) {
- if (!(vcpumap & (1ULL << i)))
- continue;
-
- if (!read_exact(io_fd, &ctxt, sizeof(ctxt))) {
- ERROR("Error when reading ctxt %d", i);
- goto out;
- }
-
- if ( !new_ctxt_format )
- ctxt.flags |= VGCF_online;
-
- if (i == 0) {
- /*
- * Uncanonicalise the suspend-record frame number and poke
- * resume record.
- */
- pfn = ctxt.user_regs.edx;
- if ((pfn >= p2m_size) ||
- (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
- ERROR("Suspend record frame number is bad");
- goto out;
- }
- ctxt.user_regs.edx = mfn = p2m[pfn];
- start_info = xc_map_foreign_range(
- xc_handle, dom, PAGE_SIZE, PROT_READ | PROT_WRITE, mfn);
- start_info->nr_pages = p2m_size;
- start_info->shared_info = shared_info_frame << PAGE_SHIFT;
- start_info->flags = 0;
- *store_mfn = start_info->store_mfn = p2m[start_info->store_mfn];
- start_info->store_evtchn = store_evtchn;
- start_info->console.domU.mfn = p2m[start_info->console.domU.mfn];
- start_info->console.domU.evtchn = console_evtchn;
- *console_mfn = start_info->console.domU.mfn;
- munmap(start_info, PAGE_SIZE);
- }
-
- /* Uncanonicalise each GDT frame number. */
- if (ctxt.gdt_ents > 8192) {
- ERROR("GDT entry count out of range");
- goto out;
- }
-
- for (j = 0; (512*j) < ctxt.gdt_ents; j++) {
- pfn = ctxt.gdt_frames[j];
- if ((pfn >= p2m_size) ||
- (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
- ERROR("GDT frame number is bad");
- goto out;
- }
- ctxt.gdt_frames[j] = p2m[pfn];
- }
-
- /* Uncanonicalise the page table base pointer. */
- pfn = xen_cr3_to_pfn(ctxt.ctrlreg[3]);
-
- if (pfn >= p2m_size) {
- ERROR("PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
- pfn, p2m_size, pfn_type[pfn]);
- goto out;
- }
-
- if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
- ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
- ERROR("PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
- pfn, p2m_size, pfn_type[pfn],
- (unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
- goto out;
- }
-
- ctxt.ctrlreg[3] = xen_pfn_to_cr3(p2m[pfn]);
-
- /* Guest pagetable (x86/64) stored in otherwise-unused CR1. */
- if ( (pt_levels == 4) && ctxt.ctrlreg[1] )
- {
- pfn = xen_cr3_to_pfn(ctxt.ctrlreg[1]);
-
- if (pfn >= p2m_size) {
- ERROR("User PT base is bad: pfn=%lu p2m_size=%lu type=%08lx",
- pfn, p2m_size, pfn_type[pfn]);
- goto out;
- }
-
- if ( (pfn_type[pfn] & XEN_DOMCTL_PFINFO_LTABTYPE_MASK) !=
- ((unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT) ) {
- ERROR("User PT base is bad. pfn=%lu nr=%lu type=%08lx %08lx",
- pfn, p2m_size, pfn_type[pfn],
- (unsigned long)pt_levels<<XEN_DOMCTL_PFINFO_LTAB_SHIFT);
- goto out;
- }
-
- ctxt.ctrlreg[1] = xen_pfn_to_cr3(p2m[pfn]);
- }
-
- domctl.cmd = XEN_DOMCTL_setvcpucontext;
- domctl.domain = (domid_t)dom;
- domctl.u.vcpucontext.vcpu = i;
- set_xen_guest_handle(domctl.u.vcpucontext.ctxt, &ctxt);
- rc = xc_domctl(xc_handle, &domctl);
- if (rc != 0) {
- ERROR("Couldn't build vcpu%d", i);
- goto out;
- }
- }
-
- if (!read_exact(io_fd, shared_info_page, PAGE_SIZE)) {
- ERROR("Error when reading shared info page");
- goto out;
- }
-
- /* clear any pending events and the selector */
- memset(&(shared_info->evtchn_pending[0]), 0,
- sizeof (shared_info->evtchn_pending));
- for ( i = 0; i < MAX_VIRT_CPUS; i++ )
- shared_info->vcpu_info[i].evtchn_pending_sel = 0;
-
- /* Copy saved contents of shared-info page. No checking needed. */
- page = xc_map_foreign_range(
- xc_handle, dom, PAGE_SIZE, PROT_WRITE, shared_info_frame);
- memcpy(page, shared_info, PAGE_SIZE);
- munmap(page, PAGE_SIZE);
-
- /* Uncanonicalise the pfn-to-mfn table frame-number list. */
- for (i = 0; i < P2M_FL_ENTRIES; i++) {
- pfn = p2m_frame_list[i];
- if ((pfn >= p2m_size) || (pfn_type[pfn] != XEN_DOMCTL_PFINFO_NOTAB)) {
- ERROR("PFN-to-MFN frame number is bad");
- goto out;
- }
-
- p2m_frame_list[i] = p2m[pfn];
- }
-
- /* Copy the P2M we've constructed to the 'live' P2M */
- if (!(live_p2m = xc_map_foreign_batch(xc_handle, dom, PROT_WRITE,
- p2m_frame_list, P2M_FL_ENTRIES))) {
- ERROR("Couldn't map p2m table");
- goto out;
- }
-
- memcpy(live_p2m, p2m, ROUNDUP(p2m_size * sizeof(xen_pfn_t), PAGE_SHIFT));
- munmap(live_p2m, ROUNDUP(p2m_size * sizeof(xen_pfn_t), PAGE_SHIFT));
-
- DPRINTF("Domain ready to be built.\n");
-
- out:
- if ( (rc != 0) && (dom != 0) )
- xc_domain_destroy(xc_handle, dom);
- free(mmu);
- free(p2m);
- free(pfn_type);
-
- /* discard cache for save file */
- discard_file_cache(io_fd, 1 /*flush*/);
-
- DPRINTF("Restore exit with rc=%d\n", rc);
-
- return rc;
-}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xenguest.h
--- a/tools/libxc/xenguest.h Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxc/xenguest.h Fri Apr 06 10:08:30 2007 -0600
@@ -38,32 +38,21 @@ int xc_hvm_save(int xc_handle, int io_fd
void (*qemu_flip_buffer)(int, int));
/**
- * This function will restore a saved domain running Linux.
+ * This function will restore a saved domain.
*
* @parm xc_handle a handle to an open hypervisor interface
* @parm fd the file descriptor to restore a domain from
* @parm dom the id of the domain
- * @parm p2m_size number of pages the guest has (i.e. number entries in P2M)
- * @parm max_nr_pfns domains maximum real memory allocation, in pages
* @parm store_evtchn the store event channel for this domain to use
* @parm store_mfn returned with the mfn of the store page
+ * @parm hvm non-zero if this is a HVM restore
+ * @parm pae non-zero if this HVM domain has PAE support enabled
* @return 0 on success, -1 on failure
*/
-int xc_linux_restore(int xc_handle, int io_fd, uint32_t dom,
- unsigned long p2m_size, unsigned long max_nr_pfns,
- unsigned int store_evtchn, unsigned long *store_mfn,
- unsigned int console_evtchn, unsigned long *console_mfn);
-
-/**
- * This function will restore a saved hvm domain running unmodified guest.
- *
- * @parm store_mfn pass mem size & returned with the mfn of the store page
- * @return 0 on success, -1 on failure
- */
-int xc_hvm_restore(int xc_handle, int io_fd, uint32_t dom,
- unsigned long max_pfn, unsigned int store_evtchn,
- unsigned long *store_mfn,
- unsigned int pae, unsigned int apic);
+int xc_domain_restore(int xc_handle, int io_fd, uint32_t dom,
+ unsigned int store_evtchn, unsigned long *store_mfn,
+ unsigned int console_evtchn, unsigned long *console_mfn,
+ unsigned int hvm, unsigned int pae);
/**
* This function will create a domain for a paravirtualized Linux
@@ -159,8 +148,6 @@ int xc_get_hvm_param(
int xc_get_hvm_param(
int handle, domid_t dom, int param, unsigned long *value);
-int xc_hvm_drain_io(int handle, domid_t dom);
-
/* PowerPC specific. */
int xc_prose_build(int xc_handle,
uint32_t domid,
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxc/xg_private.c
--- a/tools/libxc/xg_private.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxc/xg_private.c Fri Apr 06 10:08:30 2007 -0600
@@ -209,16 +209,6 @@ __attribute__((weak))
return -1;
}
-__attribute__((weak))
- int xc_hvm_restore(int xc_handle, int io_fd, uint32_t dom,
- unsigned long max_pfn, unsigned int store_evtchn,
- unsigned long *store_mfn,
- unsigned int pae, unsigned int apic)
-{
- errno = ENOSYS;
- return -1;
-}
-
__attribute__((weak)) int xc_get_hvm_param(
int handle, domid_t dom, int param, unsigned long *value)
{
@@ -227,11 +217,6 @@ __attribute__((weak)) int xc_get_hvm_par
__attribute__((weak)) int xc_set_hvm_param(
int handle, domid_t dom, int param, unsigned long value)
-{
- return -ENOSYS;
-}
-
-__attribute__((weak)) int xc_hvm_drain_io(int handle, domid_t dom)
{
return -ENOSYS;
}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/include/xen_host.h
--- a/tools/libxen/include/xen_host.h Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/include/xen_host.h Fri Apr 06 10:08:30 2007 -0600
@@ -436,6 +436,13 @@ xen_host_dmesg(xen_session *session, cha
/**
+ * List all supported methods.
+ */
+extern bool
+xen_host_list_methods(xen_session *session, struct xen_string_set **result);
+
+
+/**
* Return a list of all the hosts known to the system.
*/
extern bool
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/include/xen_sr.h
--- a/tools/libxen/include/xen_sr.h Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/include/xen_sr.h Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2006, XenSource Inc.
+ * Copyright (c) 2006-2007, XenSource Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
@@ -22,6 +22,7 @@
#include "xen_common.h"
#include "xen_pbd_decl.h"
#include "xen_sr_decl.h"
+#include "xen_string_set.h"
#include "xen_vdi_decl.h"
@@ -72,7 +73,7 @@ typedef struct xen_sr_record
int64_t physical_utilisation;
int64_t physical_size;
char *type;
- char *location;
+ char *content_type;
} xen_sr_record;
/**
@@ -169,20 +170,6 @@ xen_sr_get_by_uuid(xen_session *session,
/**
- * Create a new SR instance, and return its handle.
- */
-extern bool
-xen_sr_create(xen_session *session, xen_sr *result, xen_sr_record *record);
-
-
-/**
- * Destroy the specified SR instance.
- */
-extern bool
-xen_sr_destroy(xen_session *session, xen_sr sr);
-
-
-/**
* Get all the SR instances with the given label.
*/
extern bool
@@ -253,10 +240,10 @@ xen_sr_get_type(xen_session *session, ch
/**
- * Get the location field of the given SR.
- */
-extern bool
-xen_sr_get_location(xen_session *session, char **result, xen_sr sr);
+ * Get the content_type field of the given SR.
+ */
+extern bool
+xen_sr_get_content_type(xen_session *session, char **result, xen_sr sr);
/**
@@ -274,11 +261,10 @@ xen_sr_set_name_description(xen_session
/**
- * Take an exact copy of the Storage Repository; the cloned storage
- * repository has the same type as its parent
- */
-extern bool
-xen_sr_clone(xen_session *session, xen_sr *result, xen_sr sr, char *loc, char
*name);
+ * Return a set of all the SR types supported by the system.
+ */
+extern bool
+xen_sr_get_supported_types(xen_session *session, struct xen_string_set
**result);
/**
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/include/xen_vdi.h
--- a/tools/libxen/include/xen_vdi.h Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/include/xen_vdi.h Fri Apr 06 10:08:30 2007 -0600
@@ -338,21 +338,6 @@ xen_vdi_remove_from_other_config(xen_ses
/**
- * Take an exact copy of the VDI; the snapshot lives in the same
- * Storage Repository as its parent.
- */
-extern bool
-xen_vdi_snapshot(xen_session *session, xen_vdi *result, xen_vdi vdi);
-
-
-/**
- * Resize the vdi to the size.
- */
-extern bool
-xen_vdi_resize(xen_session *session, xen_vdi vdi, int64_t size);
-
-
-/**
* Return a list of all the VDIs known to the system.
*/
extern bool
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/src/xen_common.c
--- a/tools/libxen/src/xen_common.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/src/xen_common.c Fri Apr 06 10:08:30 2007 -0600
@@ -102,6 +102,9 @@ add_struct_value(const struct abstract_t
const char *),
const char *, xmlNode *);
+static xmlNode *
+add_container(xmlNode *parent, const char *name);
+
static void
call_raw(xen_session *, const char *, abstract_value [], int,
const abstract_type *, void *);
@@ -1290,6 +1293,48 @@ make_body_add_type(enum abstract_typenam
}
break;
+ case MAP:
+ {
+ const struct struct_member *member = v->type->members;
+ arbitrary_map *map_val = v->u.struct_val;
+ xmlNode *param_node = add_param_struct(params_node);
+ for (size_t i = 0; i < map_val->size; i++) {
+ enum abstract_typename typename_key = member[0].type->typename;
+ enum abstract_typename typename_val = member[1].type->typename;
+ int offset_key = member[0].offset;
+ int offset_val = member[1].offset;
+ int struct_size = v->type->struct_size;
+
+ switch (typename_key) {
+ case STRING: {
+ char **addr = (void *)(map_val + 1) +
+ (i * struct_size) +
+ offset_key;
+ char *key = *addr;
+
+ switch (typename_val) {
+ case STRING: {
+ char *val;
+ addr = (void *)(map_val + 1) +
+ (i * struct_size) +
+ offset_val;
+ val = *addr;
+ add_struct_member(param_node, key, "string", val);
+ break;
+ }
+ default:
+ assert(false);
+ }
+ break;
+ }
+ default:
+ assert(false);
+ }
+ }
+ }
+ break;
+
+
default:
assert(false);
}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/src/xen_host.c
--- a/tools/libxen/src/xen_host.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/src/xen_host.c Fri Apr 06 10:08:30 2007 -0600
@@ -740,6 +740,18 @@ xen_host_dmesg(xen_session *session, cha
bool
+xen_host_list_methods(xen_session *session, struct xen_string_set **result)
+{
+
+ abstract_type result_type = abstract_type_string_set;
+
+ *result = NULL;
+ xen_call_(session, "host.list_methods", NULL, 0, &result_type, result);
+ return session->ok;
+}
+
+
+bool
xen_host_get_all(xen_session *session, struct xen_host_set **result)
{
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/src/xen_sr.c
--- a/tools/libxen/src/xen_sr.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/src/xen_sr.c Fri Apr 06 10:08:30 2007 -0600
@@ -1,5 +1,5 @@
/*
- * Copyright (c) 2006, XenSource Inc.
+ * Copyright (c) 2006-2007, XenSource Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
@@ -65,9 +65,9 @@ static const struct_member xen_sr_record
{ .key = "type",
.type = &abstract_type_string,
.offset = offsetof(xen_sr_record, type) },
- { .key = "location",
- .type = &abstract_type_string,
- .offset = offsetof(xen_sr_record, location) }
+ { .key = "content_type",
+ .type = &abstract_type_string,
+ .offset = offsetof(xen_sr_record, content_type) }
};
const abstract_type xen_sr_record_abstract_type_ =
@@ -94,7 +94,7 @@ xen_sr_record_free(xen_sr_record *record
xen_vdi_record_opt_set_free(record->vdis);
xen_pbd_record_opt_set_free(record->pbds);
free(record->type);
- free(record->location);
+ free(record->content_type);
free(record);
}
@@ -140,37 +140,6 @@ xen_sr_get_by_uuid(xen_session *session,
bool
-xen_sr_create(xen_session *session, xen_sr *result, xen_sr_record *record)
-{
- abstract_value param_values[] =
- {
- { .type = &xen_sr_record_abstract_type_,
- .u.struct_val = record }
- };
-
- abstract_type result_type = abstract_type_string;
-
- *result = NULL;
- XEN_CALL_("SR.create");
- return session->ok;
-}
-
-
-bool
-xen_sr_destroy(xen_session *session, xen_sr sr)
-{
- abstract_value param_values[] =
- {
- { .type = &abstract_type_string,
- .u.string_val = sr }
- };
-
- xen_call_(session, "SR.destroy", param_values, 1, NULL, NULL);
- return session->ok;
-}
-
-
-bool
xen_sr_get_by_name_label(xen_session *session, struct xen_sr_set **result,
char *label)
{
abstract_value param_values[] =
@@ -239,6 +208,23 @@ xen_sr_get_vdis(xen_session *session, st
bool
+xen_sr_get_pbds(xen_session *session, struct xen_pbd_set **result, xen_sr sr)
+{
+ abstract_value param_values[] =
+ {
+ { .type = &abstract_type_string,
+ .u.string_val = sr }
+ };
+
+ abstract_type result_type = abstract_type_string_set;
+
+ *result = NULL;
+ XEN_CALL_("SR.get_PBDs");
+ return session->ok;
+}
+
+
+bool
xen_sr_get_virtual_allocation(xen_session *session, int64_t *result, xen_sr sr)
{
abstract_value param_values[] =
@@ -304,18 +290,18 @@ xen_sr_get_type(xen_session *session, ch
bool
-xen_sr_get_location(xen_session *session, char **result, xen_sr sr)
-{
- abstract_value param_values[] =
- {
- { .type = &abstract_type_string,
- .u.string_val = sr }
- };
-
- abstract_type result_type = abstract_type_string;
-
- *result = NULL;
- XEN_CALL_("SR.get_location");
+xen_sr_get_content_type(xen_session *session, char **result, xen_sr sr)
+{
+ abstract_value param_values[] =
+ {
+ { .type = &abstract_type_string,
+ .u.string_val = sr }
+ };
+
+ abstract_type result_type = abstract_type_string;
+
+ *result = NULL;
+ XEN_CALL_("SR.get_content_type");
return session->ok;
}
@@ -353,22 +339,13 @@ xen_sr_set_name_description(xen_session
bool
-xen_sr_clone(xen_session *session, xen_sr *result, xen_sr sr, char *loc, char
*name)
-{
- abstract_value param_values[] =
- {
- { .type = &abstract_type_string,
- .u.string_val = sr },
- { .type = &abstract_type_string,
- .u.string_val = loc },
- { .type = &abstract_type_string,
- .u.string_val = name }
- };
-
- abstract_type result_type = abstract_type_string;
-
- *result = NULL;
- XEN_CALL_("SR.clone");
+xen_sr_get_supported_types(xen_session *session, struct xen_string_set
**result)
+{
+
+ abstract_type result_type = abstract_type_string_set;
+
+ *result = NULL;
+ xen_call_(session, "SR.get_supported_types", NULL, 0, &result_type,
result);
return session->ok;
}
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/src/xen_vdi.c
--- a/tools/libxen/src/xen_vdi.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/src/xen_vdi.c Fri Apr 06 10:08:30 2007 -0600
@@ -508,39 +508,6 @@ xen_vdi_remove_from_other_config(xen_ses
bool
-xen_vdi_snapshot(xen_session *session, xen_vdi *result, xen_vdi vdi)
-{
- abstract_value param_values[] =
- {
- { .type = &abstract_type_string,
- .u.string_val = vdi }
- };
-
- abstract_type result_type = abstract_type_string;
-
- *result = NULL;
- XEN_CALL_("VDI.snapshot");
- return session->ok;
-}
-
-
-bool
-xen_vdi_resize(xen_session *session, xen_vdi vdi, int64_t size)
-{
- abstract_value param_values[] =
- {
- { .type = &abstract_type_string,
- .u.string_val = vdi },
- { .type = &abstract_type_int,
- .u.int_val = size }
- };
-
- xen_call_(session, "VDI.resize", param_values, 2, NULL, NULL);
- return session->ok;
-}
-
-
-bool
xen_vdi_get_all(xen_session *session, struct xen_vdi_set **result)
{
diff -r 9ec7dadc98ba -r eb3e430242ac tools/libxen/test/test_bindings.c
--- a/tools/libxen/test/test_bindings.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/libxen/test/test_bindings.c Fri Apr 06 10:08:30 2007 -0600
@@ -64,6 +64,7 @@ typedef struct
static xen_vm create_new_vm(xen_session *session, bool hvm);
static void print_session_info(xen_session *session);
+static void print_methods(xen_session *session);
static void print_vm_power_state(xen_session *session, xen_vm vm);
static void print_vm_metrics(xen_session *session, xen_vm vm);
@@ -159,6 +160,14 @@ int main(int argc, char **argv)
xen_session_login_with_password(call_func, NULL, username, password);
print_session_info(session);
+ if (!session->ok)
+ {
+ /* Error has been logged, just clean up. */
+ CLEANUP;
+ return 1;
+ }
+
+ print_methods(session);
if (!session->ok)
{
/* Error has been logged, just clean up. */
@@ -644,6 +653,40 @@ static void print_session_info(xen_sessi
}
+static int pstrcmp(const void *p1, const void *p2)
+{
+ return strcmp(*(char **)p1, *(char **)p2);
+}
+
+
+/**
+ * Print the list of supported methods.
+ */
+static void print_methods(xen_session *session)
+{
+ xen_string_set *methods;
+
+ if (!xen_host_list_methods(session, &methods))
+ {
+ print_error(session);
+ goto done;
+ }
+
+ printf("%zd.\n", methods->size);
+ qsort(methods->contents, methods->size, sizeof(char *), pstrcmp);
+
+ printf("Supported methods:\n");
+ for (size_t i = 0; i < methods->size; i++)
+ {
+ printf(" %s\n", methods->contents[i]);
+ }
+ fflush(stdout);
+
+done:
+ xen_string_set_free(methods);
+}
+
+
/**
* Print the metrics for the given VM.
*/
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendAPI.py
--- a/tools/python/xen/xend/XendAPI.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/XendAPI.py Fri Apr 06 10:08:30 2007 -0600
@@ -26,20 +26,22 @@ import time
import time
import xmlrpclib
-from xen.xend import XendDomain, XendDomainInfo, XendNode, XendDmesg
-from xen.xend import XendLogging, XendTaskManager
-
-from xen.xend.XendAPIVersion import *
-from xen.xend.XendAuthSessions import instance as auth_manager
-from xen.xend.XendError import *
-from xen.xend.XendClient import ERROR_INVALID_DOMAIN
-from xen.xend.XendLogging import log
-from xen.xend.XendNetwork import XendNetwork
-from xen.xend.XendTask import XendTask
-from xen.xend.XendPIFMetrics import XendPIFMetrics
-from xen.xend.XendVMMetrics import XendVMMetrics
-
-from xen.xend.XendAPIConstants import *
+import XendDomain, XendDomainInfo, XendNode, XendDmesg
+import XendLogging, XendTaskManager
+
+from XendAPIVersion import *
+from XendAuthSessions import instance as auth_manager
+from XendError import *
+from XendClient import ERROR_INVALID_DOMAIN
+from XendLogging import log
+from XendNetwork import XendNetwork
+from XendTask import XendTask
+from XendPIFMetrics import XendPIFMetrics
+from XendVMMetrics import XendVMMetrics
+
+import XendPBD
+
+from XendAPIConstants import *
from xen.util.xmlrpclib2 import stringify
from xen.util.blkif import blkdev_name_to_number
@@ -394,6 +396,17 @@ def valid_sr(func):
_check_ref(lambda r: XendNode.instance().is_valid_sr,
'SR', func, *args, **kwargs)
+def valid_pbd(func):
+ """Decorator to verify if pbd_ref is valid before calling
+ method.
+
+ @param func: function with params: (self, session, pbd_ref)
+ @rtype: callable object
+ """
+ return lambda *args, **kwargs: \
+ _check_ref(lambda r: r in XendPBD.get_all_refs(),
+ 'PBD', func, *args, **kwargs)
+
def valid_pif(func):
"""Decorator to verify if pif_ref is valid before calling
method.
@@ -479,6 +492,7 @@ classes = {
'VTPM' : valid_vtpm,
'console' : valid_console,
'SR' : valid_sr,
+ 'PBD' : valid_pbd,
'PIF' : valid_pif,
'PIF_metrics' : valid_pif_metrics,
'task' : valid_task,
@@ -488,6 +502,7 @@ autoplug_classes = {
autoplug_classes = {
'network' : XendNetwork,
'VM_metrics' : XendVMMetrics,
+ 'PBD' : XendPBD.XendPBD,
'PIF_metrics' : XendPIFMetrics,
}
@@ -774,7 +789,6 @@ class XendAPI(object):
'progress',
'type',
'result',
- 'error_code',
'error_info',
'allowed_operations',
'session'
@@ -809,10 +823,6 @@ class XendAPI(object):
task = XendTaskManager.get_task(task_ref)
return xen_api_success(task.result)
- def task_get_error_code(self, session, task_ref):
- task = XendTaskManager.get_task(task_ref)
- return xen_api_success(task.error_code)
-
def task_get_error_info(self, session, task_ref):
task = XendTaskManager.get_task(task_ref)
return xen_api_success(task.error_info)
@@ -843,6 +853,7 @@ class XendAPI(object):
host_attr_ro = ['software_version',
'resident_VMs',
+ 'PBDs',
'PIFs',
'host_CPUs',
'cpu_configuration',
@@ -870,7 +881,8 @@ class XendAPI(object):
('get_log', 'String'),
('send_debug_keys', None)]
- host_funcs = [('get_by_name_label', 'Set(host)')]
+ host_funcs = [('get_by_name_label', None),
+ ('list_methods', None)]
# attributes
def host_get_name_label(self, session, host_ref):
@@ -913,6 +925,8 @@ class XendAPI(object):
return xen_api_success(XendNode.instance().xen_version())
def host_get_resident_VMs(self, session, host_ref):
return xen_api_success(XendDomain.instance().get_domain_refs())
+ def host_get_PBDs(self, _, ref):
+ return xen_api_success(XendPBD.get_all_refs())
def host_get_PIFs(self, session, ref):
return xen_api_success(XendNode.instance().get_PIF_refs())
def host_get_host_CPUs(self, session, host_ref):
@@ -925,8 +939,6 @@ class XendAPI(object):
return xen_api_success(['pygrub'])
def host_get_sched_policy(self, _, host_ref):
return xen_api_success(XendNode.instance().get_vcpus_policy())
- def host_set_sched_policy(self, _, host_ref, policy):
- return xen_api_todo()
def host_get_cpu_configuration(self, _, host_ref):
return xen_api_success(XendNode.instance().get_cpu_configuration())
@@ -992,6 +1004,12 @@ class XendAPI(object):
return xen_api_success((XendNode.instance().uuid,))
return xen_api_success([])
+ def host_list_methods(self, _):
+ def _funcs():
+ return [getattr(XendAPI, x) for x in XendAPI.__dict__]
+
+ return xen_api_success([x.api for x in _funcs()
+ if hasattr(x, 'api')])
# Xen API: Class host_CPU
# ----------------------------------------------------------------
@@ -2061,8 +2079,8 @@ class XendAPI(object):
vif_ref = dom.create_vif(vif_struct)
xendom.managed_config_save(dom)
return xen_api_success(vif_ref)
- except XendError:
- return xen_api_error(XEND_ERROR_TODO)
+ except XendError, exn:
+ return xen_api_error(['INTERNAL_ERROR', str(exn)])
def VIF_destroy(self, session, vif_ref):
xendom = XendDomain.instance()
@@ -2169,7 +2187,7 @@ class XendAPI(object):
'other_config']
VDI_attr_inst = VDI_attr_ro + VDI_attr_rw
- VDI_methods = [('snapshot', 'VDI'), ('destroy', None)]
+ VDI_methods = [('destroy', None)]
VDI_funcs = [('create', 'VDI'),
('get_by_name_label', 'Set(VDI)')]
@@ -2233,8 +2251,6 @@ class XendAPI(object):
return xen_api_success_void()
# Object Methods
- def VDI_snapshot(self, session, vdi_ref):
- return xen_api_todo()
def VDI_destroy(self, session, vdi_ref):
sr = XendNode.instance().get_sr_containing_vdi(vdi_ref)
@@ -2349,8 +2365,8 @@ class XendAPI(object):
vtpm_ref = dom.create_vtpm(vtpm_struct)
xendom.managed_config_save(dom)
return xen_api_success(vtpm_ref)
- except XendError:
- return xen_api_error(XEND_ERROR_TODO)
+ except XendError, exn:
+ return xen_api_error(['INTERNAL_ERROR', str(exn)])
else:
return xen_api_error(['HANDLE_INVALID', 'VM', vtpm_struct['VM']])
@@ -2424,8 +2440,8 @@ class XendAPI(object):
console_ref = dom.create_console(console_struct)
xendom.managed_config_save(dom)
return xen_api_success(console_ref)
- except XendError, e:
- return xen_api_error([XEND_ERROR_TODO, str(e)])
+ except XendError, exn:
+ return xen_api_error(['INTERNAL_ERROR', str(exn)])
# Xen API: Class SR
# ----------------------------------------------------------------
@@ -2434,18 +2450,17 @@ class XendAPI(object):
'physical_utilisation',
'physical_size',
'type',
- 'location']
+ 'content_type']
SR_attr_rw = ['name_label',
'name_description']
SR_attr_inst = ['physical_size',
'type',
- 'location',
'name_label',
'name_description']
- SR_methods = [('clone', 'SR'), ('destroy', None)]
+ SR_methods = []
SR_funcs = [('get_by_name_label', 'Set(SR)'),
('get_by_uuid', 'SR')]
@@ -2456,15 +2471,10 @@ class XendAPI(object):
def SR_get_by_name_label(self, session, label):
return xen_api_success(XendNode.instance().get_sr_by_name(label))
- def SR_create(self, session):
- return xen_api_error(XEND_ERROR_UNSUPPORTED)
+ def SR_get_supported_types(self, _):
+ return xen_api_success(['local', 'qcow_file'])
# Class Methods
- def SR_clone(self, session, sr_ref):
- return xen_api_error(XEND_ERROR_UNSUPPORTED)
-
- def SR_destroy(self, session, sr_ref):
- return xen_api_error(XEND_ERROR_UNSUPPORTED)
def SR_get_record(self, session, sr_ref):
sr = XendNode.instance().get_sr(sr_ref)
@@ -2497,8 +2507,8 @@ class XendAPI(object):
def SR_get_type(self, _, ref):
return self._get_SR_attr(ref, 'type')
- def SR_get_location(self, _, ref):
- return self._get_SR_attr(ref, 'location')
+ def SR_get_content_type(self, _, ref):
+ return self._get_SR_attr(ref, 'content_type')
def SR_get_name_label(self, _, ref):
return self._get_SR_attr(ref, 'name_label')
@@ -2519,6 +2529,33 @@ class XendAPI(object):
sr.name_description = value
XendNode.instance().save()
return xen_api_success_void()
+
+
+ # Xen API: Class PBD
+ # ----------------------------------------------------------------
+
+ PBD_attr_ro = ['host',
+ 'SR',
+ 'device_config',
+ 'currently_attached']
+ PBD_attr_rw = []
+ PBD_methods = [('destroy', None)]
+ PBD_funcs = [('create', None)]
+
+ def PBD_get_all(self, _):
+ return xen_api_success(XendPBD.get_all_refs())
+
+ def _PBD_get(self, _, ref):
+ return XendPBD.get(ref)
+
+ def PBD_create(self, _, record):
+ if 'uuid' in record:
+ return xen_api_error(['VALUE_NOT_SUPPORTED',
+ 'uuid', record['uuid'],
+ 'You may not specify a UUID on creation'])
+ new_uuid = XendPBD.XendPBD(record).get_uuid()
+ XendNode.instance().save()
+ return xen_api_success(new_uuid)
# Xen API: Class event
@@ -2572,15 +2609,6 @@ class XendAPI(object):
def debug_get_record(self, session, debug_ref):
return xen_api_success({'uuid': debug_ref})
-
-
- def list_all_methods(self, _):
- def _funcs():
- return [getattr(XendAPI, x) for x in XendAPI.__dict__]
-
- return xen_api_success([x.api for x in _funcs()
- if hasattr(x, 'api')])
- list_all_methods.api = '_UNSUPPORTED_list_all_methods'
class XendAPIAsyncProxy:
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendCheckpoint.py
--- a/tools/python/xen/xend/XendCheckpoint.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/XendCheckpoint.py Fri Apr 06 10:08:30 2007 -0600
@@ -91,7 +91,7 @@ def save(fd, dominfo, network, live, dst
# more information.
cmd = [xen.util.auxbin.pathTo(XC_SAVE), str(fd),
str(dominfo.getDomid()), "0", "0",
- str(int(live) | (int(hvm) << 2) | (int(stdvga) << 3)) ]
+ str(int(live) | (int(hvm) << 2)) ]
log.debug("[xc_save]: %s", string.join(cmd))
def saveInputHandler(line, tochild):
@@ -187,7 +187,6 @@ def restore(xd, fd, dominfo = None, paus
assert console_port
nr_pfns = (dominfo.getMemoryTarget() + 3) / 4
- max_nr_pfns = (dominfo.getMemoryMaximum() + 3) / 4
# if hvm, pass mem size to calculate the store_mfn
image_cfg = dominfo.info.get('image', {})
@@ -202,27 +201,21 @@ def restore(xd, fd, dominfo = None, paus
pae = 0
try:
- l = read_exact(fd, sizeof_unsigned_long,
- "not a valid guest state file: pfn count read")
- p2m_size = unpack("L", l)[0] # native sizeof long
-
- if p2m_size > 16*1024*1024: # XXX
- raise XendError(
- "not a valid guest state file: pfn count out of range")
-
shadow = dominfo.info['shadow_memory']
- log.debug("restore:shadow=0x%x, _static_max=0x%x, _static_min=0x%x, "
- "p2m_size=0x%x.", dominfo.info['shadow_memory'],
+ log.debug("restore:shadow=0x%x, _static_max=0x%x, _static_min=0x%x, ",
+ dominfo.info['shadow_memory'],
dominfo.info['memory_static_max'],
- dominfo.info['memory_static_min'], p2m_size)
+ dominfo.info['memory_static_min'])
balloon.free(xc.pages_to_kib(nr_pfns) + shadow * 1024)
shadow_cur = xc.shadow_mem_control(dominfo.getDomid(), shadow)
dominfo.info['shadow_memory'] = shadow_cur
+ xc.domain_setmaxmem(dominfo.getDomid(), dominfo.getMemoryMaximum())
+
cmd = map(str, [xen.util.auxbin.pathTo(XC_RESTORE),
- fd, dominfo.getDomid(), p2m_size, max_nr_pfns,
+ fd, dominfo.getDomid(),
store_port, console_port, int(is_hvm), pae, apic])
log.debug("[xc_restore]: %s", string.join(cmd))
diff -r 9ec7dadc98ba -r eb3e430242ac
tools/python/xen/xend/XendLocalStorageRepo.py
--- a/tools/python/xen/xend/XendLocalStorageRepo.py Fri Apr 06 10:06:30
2007 -0600
+++ b/tools/python/xen/xend/XendLocalStorageRepo.py Fri Apr 06 10:08:30
2007 -0600
@@ -30,13 +30,13 @@ import struct
import struct
from xen.util import mkdir
-from xen.xend import uuid
-from xen.xend.XendError import XendError
-from xen.xend.XendVDI import *
-from xen.xend.XendTask import XendTask
-from xen.xend.XendStorageRepository import XendStorageRepository
-from xen.xend.XendStateStore import XendStateStore
-from xen.xend.XendOptions import instance as xendoptions
+import uuid
+from XendError import XendError
+from XendVDI import *
+from XendTask import XendTask
+from XendStorageRepository import XendStorageRepository
+from XendStateStore import XendStateStore
+from XendOptions import instance as xendoptions
MB = 1024 * 1024
@@ -58,8 +58,7 @@ class XendLocalStorageRepo(XendStorageRe
"""
XendStorageRepository.__init__(self, sr_uuid, sr_type,
- name_label, name_description,
- '/')
+ name_label, name_description)
self.state = XendStateStore(xendoptions().get_xend_state_path()
+ '/local_sr')
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendNode.py
--- a/tools/python/xen/xend/XendNode.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/XendNode.py Fri Apr 06 10:08:30 2007 -0600
@@ -22,17 +22,18 @@ import xen.lowlevel.xc
from xen.util import Brctl
-from xen.xend import uuid, arch
-from xen.xend.XendError import *
-from xen.xend.XendOptions import instance as xendoptions
-from xen.xend.XendQCoWStorageRepo import XendQCoWStorageRepo
-from xen.xend.XendLocalStorageRepo import XendLocalStorageRepo
-from xen.xend.XendLogging import log
-from xen.xend.XendPIF import *
-from xen.xend.XendPIFMetrics import XendPIFMetrics
-from xen.xend.XendNetwork import *
-from xen.xend.XendStateStore import XendStateStore
-from xen.xend.XendMonitor import XendMonitor
+import uuid, arch
+import XendPBD
+from XendError import *
+from XendOptions import instance as xendoptions
+from XendQCoWStorageRepo import XendQCoWStorageRepo
+from XendLocalStorageRepo import XendLocalStorageRepo
+from XendLogging import log
+from XendPIF import *
+from XendPIFMetrics import XendPIFMetrics
+from XendNetwork import *
+from XendStateStore import XendStateStore
+from XendMonitor import XendMonitor
class XendNode:
"""XendNode - Represents a Domain 0 Host."""
@@ -193,13 +194,14 @@ class XendNode:
saved_srs = self.state_store.load_state('sr')
if saved_srs:
for sr_uuid, sr_cfg in saved_srs.items():
+ log.error("SAved SRS %s %s", sr_uuid, sr_cfg['type'])
if sr_cfg['type'] == 'qcow_file':
self.srs[sr_uuid] = XendQCoWStorageRepo(sr_uuid)
- elif sr_cfg['type'] == 'local_image':
+ elif sr_cfg['type'] == 'local':
self.srs[sr_uuid] = XendLocalStorageRepo(sr_uuid)
# Create missing SRs if they don't exist
- if not self.get_sr_by_type('local_image'):
+ if not self.get_sr_by_type('local'):
image_sr_uuid = uuid.createString()
self.srs[image_sr_uuid] = XendLocalStorageRepo(image_sr_uuid)
@@ -207,6 +209,11 @@ class XendNode:
qcow_sr_uuid = uuid.createString()
self.srs[qcow_sr_uuid] = XendQCoWStorageRepo(qcow_sr_uuid)
+ saved_pbds = self.state_store.load_state('pbd')
+ if saved_pbds:
+ for pbd_uuid, pbd_cfg in saved_pbds.items():
+ pbd_cfg['uuid'] = pbd_uuid
+ XendPBD.XendPBD(pbd_cfg)
def network_create(self, record, persist = True, net_uuid = None):
@@ -280,6 +287,7 @@ class XendNode:
self.state_store.save_state('cpu', self.cpus)
self.save_PIFs()
self.save_networks()
+ self.save_PBDs()
self.save_SRs()
def save_PIFs(self):
@@ -291,6 +299,11 @@ class XendNode:
net_records = dict([(k, v.get_record_internal(False))
for k, v in self.networks.items()])
self.state_store.save_state('network', net_records)
+
+ def save_PBDs(self):
+ pbd_records = dict([(v.get_uuid(), v.get_record())
+ for v in XendPBD.get_all()])
+ self.state_store.save_state('pbd', pbd_records)
def save_SRs(self):
sr_records = dict([(k, v.get_record(transient = False))
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendPBD.py
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/tools/python/xen/xend/XendPBD.py Fri Apr 06 10:08:30 2007 -0600
@@ -0,0 +1,79 @@
+#============================================================================
+# This library is free software; you can redistribute it and/or
+# modify it under the terms of version 2.1 of the GNU Lesser General Public
+# License as published by the Free Software Foundation.
+#
+# This library is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public
+# License along with this library; if not, write to the Free Software
+# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+#============================================================================
+# Copyright (c) 2007 Xensource Inc.
+#============================================================================
+
+
+import uuid
+from XendLogging import log
+
+
+attr_inst = ['uuid',
+ 'host',
+ 'SR',
+ 'device_config']
+attr_ro = attr_inst + ['currently_attached']
+
+
+_all = {}
+
+
+def get(ref):
+ return _all[ref]
+
+
+def get_all():
+ return _all.values()
+
+
+def get_all_refs():
+ return _all.keys()
+
+
+def get_by_SR(sr_ref):
+ return [k for (k, v) in _all.items() if v.get_SR() == sr_ref]
+
+
+class XendPBD:
+ """Physical block devices."""
+
+ def __init__(self, record):
+ if 'uuid' not in record:
+ record['uuid'] = uuid.createString()
+
+ import XendAPI
+ for v in attr_inst:
+ setattr(self, v, record[v])
+ self.currently_attached = True
+ _all[record['uuid']] = self
+
+
+ def destroy(self):
+ if self.uuid in _all:
+ del _all[self.uuid]
+
+
+ def get_record(self):
+ import XendAPI
+ result = {}
+ for v in attr_ro:
+ result[v] = getattr(self, v)
+ return result
+
+
+for v in attr_ro:
+ def f(v_):
+ setattr(XendPBD, 'get_' + v_, lambda s: getattr(s, v_))
+ f(v)
diff -r 9ec7dadc98ba -r eb3e430242ac
tools/python/xen/xend/XendQCoWStorageRepo.py
--- a/tools/python/xen/xend/XendQCoWStorageRepo.py Fri Apr 06 10:06:30
2007 -0600
+++ b/tools/python/xen/xend/XendQCoWStorageRepo.py Fri Apr 06 10:08:30
2007 -0600
@@ -29,12 +29,13 @@ import struct
import struct
from xen.util import mkdir
-from xen.xend import uuid
-from xen.xend.XendError import XendError
-from xen.xend.XendVDI import *
-from xen.xend.XendTask import XendTask
-from xen.xend.XendStorageRepository import XendStorageRepository
-from xen.xend.XendOptions import instance as xendoptions
+import uuid
+import XendPBD
+from XendError import XendError
+from XendVDI import *
+from XendTask import XendTask
+from XendStorageRepository import XendStorageRepository
+from XendOptions import instance as xendoptions
XEND_STORAGE_NO_MAXIMUM = sys.maxint
XEND_STORAGE_QCOW_FILENAME = "%s.qcow"
@@ -72,7 +73,6 @@ class XendQCoWStorageRepo(XendStorageRep
sr_type = "qcow_file",
name_label = "QCoW",
name_description = "Xend QCoW Storage Repository",
- location = xendoptions().get_xend_storage_path(),
storage_max = XEND_STORAGE_NO_MAXIMUM):
"""
@keyword storage_max: Maximum disk space to use in bytes.
@@ -85,9 +85,9 @@ class XendQCoWStorageRepo(XendStorageRep
"""
XendStorageRepository.__init__(self, sr_uuid, sr_type, name_label,
- name_description, location,
- storage_max)
+ name_description, storage_max)
self.storage_free = 0
+ self.location = xendoptions().get_xend_storage_path()
self._refresh()
def get_record(self, transient = True):
@@ -98,8 +98,9 @@ class XendQCoWStorageRepo(XendStorageRep
'physical_utilisation': self.physical_utilisation,
'physical_size': self.physical_size,
'type': self.type,
- 'location': self.location,
- 'VDIs': self.images.keys()}
+ 'content_type': self.content_type,
+ 'VDIs': self.images.keys(),
+ 'PBDs': XendPBD.get_by_SR(self.uuid)}
if self.physical_size == XEND_STORAGE_NO_MAXIMUM:
stfs = os.statvfs(self.location)
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendStateStore.py
--- a/tools/python/xen/xend/XendStateStore.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/XendStateStore.py Fri Apr 06 10:08:30 2007 -0600
@@ -204,7 +204,7 @@ class XendStateStore:
if type(val) == dict:
for val_uuid in val.keys():
val_node = doc.createElement(key)
- if key == 'other_config':
+ if key in ['other_config', 'device_config']:
val_node.setAttribute('key', str(val_uuid))
val_node.setAttribute('value', str(val[val_uuid]))
else:
diff -r 9ec7dadc98ba -r eb3e430242ac
tools/python/xen/xend/XendStorageRepository.py
--- a/tools/python/xen/xend/XendStorageRepository.py Fri Apr 06 10:06:30
2007 -0600
+++ b/tools/python/xen/xend/XendStorageRepository.py Fri Apr 06 10:08:30
2007 -0600
@@ -22,8 +22,9 @@ import threading
import threading
import sys
-from xen.xend.XendError import XendError
-from xen.xend.XendVDI import *
+from XendError import XendError
+from XendVDI import *
+import XendPBD
XEND_STORAGE_NO_MAXIMUM = sys.maxint
@@ -34,7 +35,6 @@ class XendStorageRepository:
sr_type = "unknown",
name_label = 'Unknown',
name_description = 'Not Implemented',
- location = '',
storage_max = XEND_STORAGE_NO_MAXIMUM):
"""
@keyword storage_max: Maximum disk space to use in bytes.
@@ -49,7 +49,6 @@ class XendStorageRepository:
# XenAPI Parameters
self.uuid = uuid
self.type = sr_type
- self.location = location
self.name_label = name_label
self.name_description = name_description
self.images = {}
@@ -57,6 +56,7 @@ class XendStorageRepository:
self.physical_size = storage_max
self.physical_utilisation = 0
self.virtual_allocation = 0
+ self.content_type = ''
self.lock = threading.RLock()
@@ -68,9 +68,10 @@ class XendStorageRepository:
'physical_utilisation': self.physical_utilisation,
'physical_size': self.physical_size,
'type': self.type,
- 'location': self.location,
+ 'content_type': self.content_type,
'VDIs': self.images.keys()}
-
+ if not transient:
+ retval ['PBDs'] = XendPBD.get_by_SR(self.uuid)
return retval
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/XendTask.py
--- a/tools/python/xen/xend/XendTask.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/XendTask.py Fri Apr 06 10:08:30 2007 -0600
@@ -24,7 +24,7 @@ class XendTask(threading.Thread):
"""Represents a Asynchronous Task used by Xen API.
Basically proxies the callable object in a thread and returns the
- results via self.{type,result,error_code,error_info}.
+ results via self.{type,result,error_info}.
@cvar task_progress: Thread local storage for progress tracking.
It is a dict indexed by thread_id. Note that the
@@ -71,7 +71,6 @@ class XendTask(threading.Thread):
self.uuid = uuid
self.result = None
- self.error_code = ''
self.error_info = []
self.name_label = label or func.__name__
@@ -118,13 +117,11 @@ class XendTask(threading.Thread):
self.result = result['Value']
self.set_status(XEN_API_TASK_STATUS_TYPE[1])
else:
- self.error_code = result['ErrorDescription'][0]
- self.error_info = result['ErrorDescription'][1:]
+ self.error_info = result['ErrorDescription']
self.set_status(XEN_API_TASK_STATUS_TYPE[2])
except Exception, e:
log.exception('Error running Async Task')
- self.error_code = 'INTERNAL ERROR'
- self.error_info = [str(e)]
+ self.error_info = ['INTERNAL ERROR', str(e)]
self.set_status(XEN_API_TASK_STATUS_TYPE[2])
self.task_progress_lock.acquire()
@@ -144,7 +141,6 @@ class XendTask(threading.Thread):
'progress': self.get_progress(),
'type': self.type,
'result': self.result,
- 'error_code': self.error_code,
'error_info': self.error_info,
'allowed_operations': {},
'session': self.session,
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xend/image.py
--- a/tools/python/xen/xend/image.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xend/image.py Fri Apr 06 10:08:30 2007 -0600
@@ -426,11 +426,14 @@ class HVMImageHandler(ImageHandler):
# Execute device model.
#todo: Error handling
args = [self.device_model]
- args = args + ([ "-d", "%d" % self.vm.getDomid(),
- "-m", "%s" % (self.getRequiredInitialReservation() / 1024)])
+ args = args + ([ "-d", "%d" % self.vm.getDomid() ])
+ if arch.type == "ia64":
+ args = args + ([ "-m", "%s" %
+ (self.getRequiredInitialReservation() / 1024) ])
args = args + self.dmargs
if restore:
- args = args + ([ "-loadvm", "/tmp/xen.qemu-dm.%d" %
self.vm.getDomid() ])
+ args = args + ([ "-loadvm", "/tmp/xen.qemu-dm.%d" %
+ self.vm.getDomid() ])
env = dict(os.environ)
if self.display:
env['DISPLAY'] = self.display
diff -r 9ec7dadc98ba -r eb3e430242ac tools/python/xen/xm/main.py
--- a/tools/python/xen/xm/main.py Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/python/xen/xm/main.py Fri Apr 06 10:08:30 2007 -0600
@@ -591,7 +591,7 @@ class Shell(cmd.Cmd):
self.prompt = "xm> "
if serverType == SERVER_XEN_API:
try:
- res = server.xenapi._UNSUPPORTED_list_all_methods()
+ res = server.xenapi.host.list_methods()
for f in res:
setattr(Shell, 'do_' + f + ' ', self.default)
except:
diff -r 9ec7dadc98ba -r eb3e430242ac tools/xcutils/xc_restore.c
--- a/tools/xcutils/xc_restore.c Fri Apr 06 10:06:30 2007 -0600
+++ b/tools/xcutils/xc_restore.c Fri Apr 06 10:08:30 2007 -0600
@@ -21,37 +21,31 @@ main(int argc, char **argv)
unsigned int xc_fd, io_fd, domid, store_evtchn, console_evtchn;
unsigned int hvm, pae, apic;
int ret;
- unsigned long p2m_size, max_nr_pfns, store_mfn, console_mfn;
+ unsigned long store_mfn, console_mfn;
- if (argc != 10)
- errx(1, "usage: %s iofd domid p2m_size max_nr_pfns store_evtchn "
+ if ( argc != 8 )
+ errx(1, "usage: %s iofd domid store_evtchn "
"console_evtchn hvm pae apic", argv[0]);
xc_fd = xc_interface_open();
- if (xc_fd < 0)
+ if ( xc_fd < 0 )
errx(1, "failed to open control interface");
io_fd = atoi(argv[1]);
domid = atoi(argv[2]);
- p2m_size = atoi(argv[3]);
- max_nr_pfns = atoi(argv[4]);
- store_evtchn = atoi(argv[5]);
- console_evtchn = atoi(argv[6]);
- hvm = atoi(argv[7]);
- pae = atoi(argv[8]);
- apic = atoi(argv[9]);
+ store_evtchn = atoi(argv[3]);
+ console_evtchn = atoi(argv[4]);
+ hvm = atoi(argv[5]);
+ pae = atoi(argv[6]);
+ apic = atoi(argv[7]);
- if (hvm) {
- ret = xc_hvm_restore(xc_fd, io_fd, domid, max_nr_pfns, store_evtchn,
- &store_mfn, pae, apic);
- } else
- ret = xc_linux_restore(xc_fd, io_fd, domid, p2m_size,
- max_nr_pfns, store_evtchn, &store_mfn,
- console_evtchn, &console_mfn);
+ ret = xc_domain_restore(xc_fd, io_fd, domid, store_evtchn, &store_mfn,
+ console_evtchn, &console_mfn, hvm, pae);
- if (ret == 0) {
+ if ( ret == 0 )
+ {
printf("store-mfn %li\n", store_mfn);
- if (!hvm)
+ if ( !hvm )
printf("console-mfn %li\n", console_mfn);
fflush(stdout);
}
diff -r 9ec7dadc98ba -r eb3e430242ac unmodified_drivers/linux-2.6/Makefile
--- a/unmodified_drivers/linux-2.6/Makefile Fri Apr 06 10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/Makefile Fri Apr 06 10:08:30 2007 -0600
@@ -2,6 +2,7 @@ include $(M)/overrides.mk
obj-m += platform-pci/
obj-m += xenbus/
+obj-m += balloon/
obj-m += blkfront/
obj-m += netfront/
obj-m += util/
diff -r 9ec7dadc98ba -r eb3e430242ac unmodified_drivers/linux-2.6/balloon/Kbuild
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/unmodified_drivers/linux-2.6/balloon/Kbuild Fri Apr 06 10:08:30
2007 -0600
@@ -0,0 +1,9 @@
+include $(M)/overrides.mk
+
+obj-m = xen-balloon.o
+
+EXTRA_CFLAGS += -I$(M)/platform-pci
+
+xen-balloon-objs =
+xen-balloon-objs += balloon.o
+xen-balloon-objs += sysfs.o
diff -r 9ec7dadc98ba -r eb3e430242ac
unmodified_drivers/linux-2.6/balloon/Makefile
--- /dev/null Thu Jan 01 00:00:00 1970 +0000
+++ b/unmodified_drivers/linux-2.6/balloon/Makefile Fri Apr 06 10:08:30
2007 -0600
@@ -0,0 +1,3 @@
+ifneq ($(KERNELRELEASE),)
+include $(src)/Kbuild
+endif
diff -r 9ec7dadc98ba -r eb3e430242ac
unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h
--- a/unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h Fri Apr
06 10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/compat-include/xen/platform-compat.h Fri Apr
06 10:08:30 2007 -0600
@@ -25,6 +25,10 @@
#define NET_IP_ALIGN 0
#endif
+#if defined(_LINUX_SKBUFF_H) && !defined(CHECKSUM_HW)
+#define CHECKSUM_HW CHECKSUM_PARTIAL
+#endif
+
#if defined(_LINUX_ERR_H) && !defined(IS_ERR_VALUE)
#define IS_ERR_VALUE(x) unlikely((x) > (unsigned long)-1000L)
#endif
@@ -36,7 +40,7 @@
/* Some kernels have this typedef backported so we cannot reliably
* detect based on version number, hence we forcibly #define it.
*/
-#if defined(__LINUX_TYPES_H) || defined(__LINUX_GFP_H)
+#if defined(__LINUX_TYPES_H) || defined(__LINUX_GFP_H) ||
defined(_LINUX_KERNEL_H)
#define gfp_t unsigned
#endif
@@ -45,6 +49,14 @@
#define atomic_notifier_chain_register(chain,nb)
notifier_chain_register(chain,nb)
#define atomic_notifier_chain_unregister(chain,nb)
notifier_chain_unregister(chain,nb)
#define atomic_notifier_call_chain(chain,val,v)
notifier_call_chain(chain,val,v)
+#endif
+
+#if defined(_LINUX_MM_H) && defined set_page_count
+#define init_page_count(page) set_page_count(page, 1)
+#endif
+
+#if defined(__LINUX_GFP_H) && !defined __GFP_NOMEMALLOC
+#define __GFP_NOMEMALLOC 0
#endif
#if defined(_LINUX_FS_H) && LINUX_VERSION_CODE < KERNEL_VERSION(2,6,9)
@@ -71,9 +83,17 @@ void *kzalloc(size_t size, int flags);
#define end_that_request_last(req, uptodate) end_that_request_last(req)
#endif
+#if defined(_LINUX_CAPABILITY_H) && LINUX_VERSION_CODE < KERNEL_VERSION(2,6,16)
+#define capable(cap) (1)
+#endif
+
#if defined(_LINUX_KERNEL_H) && LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18)
extern char *kasprintf(gfp_t gfp, const char *fmt, ...)
__attribute__ ((format (printf, 2, 3)));
+#endif
+
+#if defined(_LINUX_SYSRQ_H) && LINUX_VERSION_CODE > KERNEL_VERSION(2,6,18)
+#define handle_sysrq(x,y,z) handle_sysrq(x,y)
#endif
/*
diff -r 9ec7dadc98ba -r eb3e430242ac unmodified_drivers/linux-2.6/mkbuildtree
--- a/unmodified_drivers/linux-2.6/mkbuildtree Fri Apr 06 10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/mkbuildtree Fri Apr 06 10:08:30 2007 -0600
@@ -15,6 +15,9 @@ XL=$C/../../linux-2.6-xen-sparse
for d in $(find ${XL}/drivers/xen/ -maxdepth 1 -type d | sed -e 1d); do
if ! echo $d | egrep -q back; then
+ lndir $d $(basename $d) > /dev/null 2>&1
+ fi
+ if ! echo $d | egrep -q ball; then
lndir $d $(basename $d) > /dev/null 2>&1
fi
done
diff -r 9ec7dadc98ba -r eb3e430242ac unmodified_drivers/linux-2.6/overrides.mk
--- a/unmodified_drivers/linux-2.6/overrides.mk Fri Apr 06 10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/overrides.mk Fri Apr 06 10:08:30 2007 -0600
@@ -4,7 +4,8 @@
#
# (i.e. we need the native config for things like -mregparm, but
# a Xen kernel to find the right headers)
-EXTRA_CFLAGS += -D__XEN_INTERFACE_VERSION__=0x00030202
+EXTRA_CFLAGS += -D__XEN_INTERFACE_VERSION__=0x00030205
+EXTRA_CFLAGS += -DCONFIG_XEN_COMPAT=0xffffff
EXTRA_CFLAGS += -I$(M)/include -I$(M)/compat-include
-DHAVE_XEN_PLATFORM_COMPAT_H
ifeq ($(ARCH),ia64)
EXTRA_CFLAGS += -DCONFIG_VMX_GUEST
diff -r 9ec7dadc98ba -r eb3e430242ac
unmodified_drivers/linux-2.6/platform-pci/platform-pci.c
--- a/unmodified_drivers/linux-2.6/platform-pci/platform-pci.c Fri Apr 06
10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/platform-pci/platform-pci.c Fri Apr 06
10:08:30 2007 -0600
@@ -208,14 +208,6 @@ static uint64_t get_callback_via(struct
((uint64_t)(pin - 1) & 3));
}
-/* Invalidate foreign mappings (e.g., in qemu-based device model). */
-static uint16_t invlmap_port;
-void xen_invalidate_foreign_mappings(void)
-{
- outb(0, invlmap_port);
-}
-EXPORT_SYMBOL(xen_invalidate_foreign_mappings);
-
static int __devinit platform_pci_init(struct pci_dev *pdev,
const struct pci_device_id *ent)
{
@@ -239,8 +231,6 @@ static int __devinit platform_pci_init(s
printk(KERN_WARNING DRV_NAME ":no resources found\n");
return -ENOENT;
}
-
- invlmap_port = ioaddr;
if (request_mem_region(mmio_addr, mmio_len, DRV_NAME) == NULL)
{
diff -r 9ec7dadc98ba -r eb3e430242ac
unmodified_drivers/linux-2.6/platform-pci/xen_support.c
--- a/unmodified_drivers/linux-2.6/platform-pci/xen_support.c Fri Apr 06
10:06:30 2007 -0600
+++ b/unmodified_drivers/linux-2.6/platform-pci/xen_support.c Fri Apr 06
10:08:30 2007 -0600
@@ -59,12 +59,3 @@ void xen_machphys_update(unsigned long m
}
EXPORT_SYMBOL(xen_machphys_update);
-void balloon_update_driver_allowance(long delta)
-{
-}
-EXPORT_SYMBOL(balloon_update_driver_allowance);
-
-void balloon_release_driver_page(struct page *page)
-{
-}
-EXPORT_SYMBOL(balloon_release_driver_page);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/hvm.c
--- a/xen/arch/x86/hvm/hvm.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/hvm.c Fri Apr 06 10:08:30 2007 -0600
@@ -131,7 +131,7 @@ void hvm_do_resume(struct vcpu *v)
switch ( p->state )
{
case STATE_IORESP_READY: /* IORESP_READY -> NONE */
- hvm_io_assist(v);
+ hvm_io_assist();
break;
case STATE_IOREQ_READY: /* IOREQ_{READY,INPROCESS} -> IORESP_READY */
case STATE_IOREQ_INPROCESS:
@@ -144,48 +144,6 @@ void hvm_do_resume(struct vcpu *v)
domain_crash_synchronous();
}
}
-}
-
-/* Called from the tools when saving a domain to make sure the io
- * request-response ring is entirely empty. */
-static int hvmop_drain_io(
- XEN_GUEST_HANDLE(xen_hvm_drain_io_t) uop)
-{
- struct xen_hvm_drain_io op;
- struct domain *d;
- struct vcpu *v;
- ioreq_t *p;
- int rc;
-
- if ( copy_from_guest(&op, uop, 1) )
- return -EFAULT;
-
- if ( !IS_PRIV(current->domain) )
- return -EPERM;
-
- d = rcu_lock_domain_by_id(op.domid);
- if ( d == NULL )
- return -ESRCH;
-
- rc = -EINVAL;
- /* Can't do this to yourself, or to a domain without an ioreq ring */
- if ( d == current->domain || !is_hvm_domain(d) || get_sp(d) == NULL )
- goto out;
-
- rc = 0;
-
- domain_pause(d); /* It's not safe to do this to running vcpus */
- for_each_vcpu(d, v)
- {
- p = &get_vio(v->domain, v->vcpu_id)->vp_ioreq;
- if ( p->state == STATE_IORESP_READY )
- hvm_io_assist(v);
- }
- domain_unpause(d);
-
- out:
- rcu_unlock_domain(d);
- return rc;
}
int hvm_domain_initialise(struct domain *d)
@@ -563,19 +521,13 @@ static hvm_hypercall_t *hvm_hypercall_ta
HYPERCALL(hvm_op)
};
-void hvm_do_hypercall(struct cpu_user_regs *pregs)
-{
- if ( unlikely(ring_3(pregs)) )
- {
- pregs->eax = -EPERM;
- return;
- }
-
+static void __hvm_do_hypercall(struct cpu_user_regs *pregs)
+{
if ( (pregs->eax >= NR_hypercalls) || !hvm_hypercall_table[pregs->eax] )
{
- gdprintk(XENLOG_WARNING, "HVM vcpu %d:%d did a bad hypercall %d.\n",
- current->domain->domain_id, current->vcpu_id,
- pregs->eax);
+ if ( pregs->eax != __HYPERVISOR_grant_table_op )
+ gdprintk(XENLOG_WARNING, "HVM vcpu %d:%d bad hypercall %d.\n",
+ current->domain->domain_id, current->vcpu_id, pregs->eax);
pregs->eax = -ENOSYS;
return;
}
@@ -641,20 +593,14 @@ static hvm_hypercall_t *hvm_hypercall32_
HYPERCALL(event_channel_op)
};
-void hvm_do_hypercall(struct cpu_user_regs *pregs)
-{
- if ( unlikely(ring_3(pregs)) )
- {
- pregs->rax = -EPERM;
- return;
- }
-
+static void __hvm_do_hypercall(struct cpu_user_regs *pregs)
+{
pregs->rax = (uint32_t)pregs->eax; /* mask in case compat32 caller */
if ( (pregs->rax >= NR_hypercalls) || !hvm_hypercall64_table[pregs->rax] )
{
- gdprintk(XENLOG_WARNING, "HVM vcpu %d:%d did a bad hypercall %ld.\n",
- current->domain->domain_id, current->vcpu_id,
- pregs->rax);
+ if ( pregs->rax != __HYPERVISOR_grant_table_op )
+ gdprintk(XENLOG_WARNING, "HVM vcpu %d:%d bad hypercall %ld.\n",
+ current->domain->domain_id, current->vcpu_id, pregs->rax);
pregs->rax = -ENOSYS;
return;
}
@@ -678,6 +624,37 @@ void hvm_do_hypercall(struct cpu_user_re
}
#endif /* defined(__x86_64__) */
+
+int hvm_do_hypercall(struct cpu_user_regs *pregs)
+{
+ int flush, preempted;
+ unsigned long old_eip;
+
+ if ( unlikely(ring_3(pregs)) )
+ {
+ pregs->eax = -EPERM;
+ return 0;
+ }
+
+ /*
+ * NB. In future flush only on decrease_reservation.
+ * For now we also need to flush when pages are added, as qemu-dm is not
+ * yet capable of faulting pages into an existing valid mapcache bucket.
+ */
+ flush = ((uint32_t)pregs->eax == __HYPERVISOR_memory_op);
+
+ /* Check for preemption: RIP will be modified from this dummy value. */
+ old_eip = pregs->eip;
+ pregs->eip = 0xF0F0F0FF;
+
+ __hvm_do_hypercall(pregs);
+
+ preempted = (pregs->eip != 0xF0F0F0FF);
+ pregs->eip = old_eip;
+
+ return (preempted ? HVM_HCALL_preempted :
+ flush ? HVM_HCALL_invalidate : HVM_HCALL_completed);
+}
void hvm_update_guest_cr3(struct vcpu *v, unsigned long guest_cr3)
{
@@ -963,12 +940,6 @@ long do_hvm_op(unsigned long op, XEN_GUE
guest_handle_cast(arg, xen_hvm_set_pci_link_route_t));
break;
- case HVMOP_drain_io:
- rc = hvmop_drain_io(
- guest_handle_cast(arg, xen_hvm_drain_io_t));
- break;
-
-
default:
{
gdprintk(XENLOG_WARNING, "Bad HVM op %ld.\n", op);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/io.c
--- a/xen/arch/x86/hvm/io.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/io.c Fri Apr 06 10:08:30 2007 -0600
@@ -289,8 +289,12 @@ static void set_reg_value (int size, int
long get_reg_value(int size, int index, int seg, struct cpu_user_regs *regs);
-static inline void set_eflags_CF(int size, unsigned long v1,
- unsigned long v2, struct cpu_user_regs *regs)
+static inline void set_eflags_CF(int size,
+ unsigned int instr,
+ unsigned long result,
+ unsigned long src,
+ unsigned long dst,
+ struct cpu_user_regs *regs)
{
unsigned long mask;
@@ -300,14 +304,28 @@ static inline void set_eflags_CF(int siz
mask = ~0UL >> (8 * (sizeof(mask) - size));
- if ((v1 & mask) > (v2 & mask))
- regs->eflags |= X86_EFLAGS_CF;
+ if ( instr == INSTR_ADD )
+ {
+ /* CF=1 <==> result is less than the augend and addend) */
+ if ( (result & mask) < (dst & mask) )
+ {
+ ASSERT((result & mask) < (src & mask));
+ regs->eflags |= X86_EFLAGS_CF;
+ }
+ }
else
- regs->eflags &= ~X86_EFLAGS_CF;
-}
-
-static inline void set_eflags_OF(int size, unsigned long v1,
- unsigned long v2, unsigned long v3,
+ {
+ ASSERT( instr == INSTR_CMP || instr == INSTR_SUB );
+ if ( (src & mask) > (dst & mask) )
+ regs->eflags |= X86_EFLAGS_CF;
+ }
+}
+
+static inline void set_eflags_OF(int size,
+ unsigned int instr,
+ unsigned long result,
+ unsigned long src,
+ unsigned long dst,
struct cpu_user_regs *regs)
{
unsigned long mask;
@@ -316,21 +334,32 @@ static inline void set_eflags_OF(int siz
size = BYTE;
ASSERT((size <= sizeof(mask)) && (size > 0));
- mask = ~0UL >> (8 * (sizeof(mask) - size));
-
- if ((v3 ^ v2) & (v3 ^ v1) & mask)
- regs->eflags |= X86_EFLAGS_OF;
-}
-
-static inline void set_eflags_AF(int size, unsigned long v1,
- unsigned long v2, unsigned long v3,
+ mask = 1UL << ((8*size) - 1);
+
+ if ( instr == INSTR_ADD )
+ {
+ if ((src ^ result) & (dst ^ result) & mask);
+ regs->eflags |= X86_EFLAGS_OF;
+ }
+ else
+ {
+ ASSERT(instr == INSTR_CMP || instr == INSTR_SUB);
+ if ((dst ^ src) & (dst ^ result) & mask)
+ regs->eflags |= X86_EFLAGS_OF;
+ }
+}
+
+static inline void set_eflags_AF(int size,
+ unsigned long result,
+ unsigned long src,
+ unsigned long dst,
struct cpu_user_regs *regs)
{
- if ((v1 ^ v2 ^ v3) & 0x10)
+ if ((result ^ src ^ dst) & 0x10)
regs->eflags |= X86_EFLAGS_AF;
}
-static inline void set_eflags_ZF(int size, unsigned long v1,
+static inline void set_eflags_ZF(int size, unsigned long result,
struct cpu_user_regs *regs)
{
unsigned long mask;
@@ -341,11 +370,11 @@ static inline void set_eflags_ZF(int siz
mask = ~0UL >> (8 * (sizeof(mask) - size));
- if ((v1 & mask) == 0)
+ if ((result & mask) == 0)
regs->eflags |= X86_EFLAGS_ZF;
}
-static inline void set_eflags_SF(int size, unsigned long v1,
+static inline void set_eflags_SF(int size, unsigned long result,
struct cpu_user_regs *regs)
{
unsigned long mask;
@@ -354,9 +383,9 @@ static inline void set_eflags_SF(int siz
size = BYTE;
ASSERT((size <= sizeof(mask)) && (size > 0));
- mask = ~0UL >> (8 * (sizeof(mask) - size));
-
- if (v1 & mask)
+ mask = 1UL << ((8*size) - 1);
+
+ if (result & mask)
regs->eflags |= X86_EFLAGS_SF;
}
@@ -379,10 +408,10 @@ static char parity_table[256] = {
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1
};
-static inline void set_eflags_PF(int size, unsigned long v1,
+static inline void set_eflags_PF(int size, unsigned long result,
struct cpu_user_regs *regs)
{
- if (parity_table[v1 & 0xFF])
+ if (parity_table[result & 0xFF])
regs->eflags |= X86_EFLAGS_PF;
}
@@ -454,7 +483,7 @@ static void hvm_mmio_assist(struct cpu_u
{
int sign = p->df ? -1 : 1;
int size = -1, index = -1;
- unsigned long value = 0, diff = 0;
+ unsigned long value = 0, result = 0;
unsigned long src, dst;
src = mmio_opp->operand[0];
@@ -575,31 +604,15 @@ static void hvm_mmio_assist(struct cpu_u
if (src & REGISTER) {
index = operand_index(src);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data & value;
+ result = (unsigned long) p->data & value;
} else if (src & IMMEDIATE) {
value = mmio_opp->immediate;
- diff = (unsigned long) p->data & value;
+ result = (unsigned long) p->data & value;
} else if (src & MEMORY) {
index = operand_index(dst);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data & value;
- set_reg_value(size, index, 0, regs, diff);
- }
- break;
-
- case INSTR_ADD:
- if (src & REGISTER) {
- index = operand_index(src);
- value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data + value;
- } else if (src & IMMEDIATE) {
- value = mmio_opp->immediate;
- diff = (unsigned long) p->data + value;
- } else if (src & MEMORY) {
- index = operand_index(dst);
- value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data + value;
- set_reg_value(size, index, 0, regs, diff);
+ result = (unsigned long) p->data & value;
+ set_reg_value(size, index, 0, regs, result);
}
/*
@@ -609,24 +622,55 @@ static void hvm_mmio_assist(struct cpu_u
*/
regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|
X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
- set_eflags_ZF(size, diff, regs);
- set_eflags_SF(size, diff, regs);
- set_eflags_PF(size, diff, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
+ break;
+
+ case INSTR_ADD:
+ if (src & REGISTER) {
+ index = operand_index(src);
+ value = get_reg_value(size, index, 0, regs);
+ result = (unsigned long) p->data + value;
+ } else if (src & IMMEDIATE) {
+ value = mmio_opp->immediate;
+ result = (unsigned long) p->data + value;
+ } else if (src & MEMORY) {
+ index = operand_index(dst);
+ value = get_reg_value(size, index, 0, regs);
+ result = (unsigned long) p->data + value;
+ set_reg_value(size, index, 0, regs, result);
+ }
+
+ /*
+ * The CF, OF, SF, ZF, AF, and PF flags are set according
+ * to the result
+ */
+ regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|X86_EFLAGS_AF|
+ X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
+ set_eflags_CF(size, mmio_opp->instr, result, value,
+ (unsigned long) p->data, regs);
+ set_eflags_OF(size, mmio_opp->instr, result, value,
+ (unsigned long) p->data, regs);
+ set_eflags_AF(size, result, value, (unsigned long) p->data, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
break;
case INSTR_OR:
if (src & REGISTER) {
index = operand_index(src);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data | value;
+ result = (unsigned long) p->data | value;
} else if (src & IMMEDIATE) {
value = mmio_opp->immediate;
- diff = (unsigned long) p->data | value;
+ result = (unsigned long) p->data | value;
} else if (src & MEMORY) {
index = operand_index(dst);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data | value;
- set_reg_value(size, index, 0, regs, diff);
+ result = (unsigned long) p->data | value;
+ set_reg_value(size, index, 0, regs, result);
}
/*
@@ -636,24 +680,24 @@ static void hvm_mmio_assist(struct cpu_u
*/
regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|
X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
- set_eflags_ZF(size, diff, regs);
- set_eflags_SF(size, diff, regs);
- set_eflags_PF(size, diff, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
break;
case INSTR_XOR:
if (src & REGISTER) {
index = operand_index(src);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data ^ value;
+ result = (unsigned long) p->data ^ value;
} else if (src & IMMEDIATE) {
value = mmio_opp->immediate;
- diff = (unsigned long) p->data ^ value;
+ result = (unsigned long) p->data ^ value;
} else if (src & MEMORY) {
index = operand_index(dst);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data ^ value;
- set_reg_value(size, index, 0, regs, diff);
+ result = (unsigned long) p->data ^ value;
+ set_reg_value(size, index, 0, regs, result);
}
/*
@@ -663,9 +707,9 @@ static void hvm_mmio_assist(struct cpu_u
*/
regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|
X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
- set_eflags_ZF(size, diff, regs);
- set_eflags_SF(size, diff, regs);
- set_eflags_PF(size, diff, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
break;
case INSTR_CMP:
@@ -673,16 +717,16 @@ static void hvm_mmio_assist(struct cpu_u
if (src & REGISTER) {
index = operand_index(src);
value = get_reg_value(size, index, 0, regs);
- diff = (unsigned long) p->data - value;
+ result = (unsigned long) p->data - value;
} else if (src & IMMEDIATE) {
value = mmio_opp->immediate;
- diff = (unsigned long) p->data - value;
+ result = (unsigned long) p->data - value;
} else if (src & MEMORY) {
index = operand_index(dst);
value = get_reg_value(size, index, 0, regs);
- diff = value - (unsigned long) p->data;
+ result = value - (unsigned long) p->data;
if ( mmio_opp->instr == INSTR_SUB )
- set_reg_value(size, index, 0, regs, diff);
+ set_reg_value(size, index, 0, regs, result);
}
/*
@@ -691,12 +735,24 @@ static void hvm_mmio_assist(struct cpu_u
*/
regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|X86_EFLAGS_AF|
X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
- set_eflags_CF(size, value, (unsigned long) p->data, regs);
- set_eflags_OF(size, diff, value, (unsigned long) p->data, regs);
- set_eflags_AF(size, diff, value, (unsigned long) p->data, regs);
- set_eflags_ZF(size, diff, regs);
- set_eflags_SF(size, diff, regs);
- set_eflags_PF(size, diff, regs);
+ if ( src & (REGISTER | IMMEDIATE) )
+ {
+ set_eflags_CF(size, mmio_opp->instr, result, value,
+ (unsigned long) p->data, regs);
+ set_eflags_OF(size, mmio_opp->instr, result, value,
+ (unsigned long) p->data, regs);
+ }
+ else
+ {
+ set_eflags_CF(size, mmio_opp->instr, result,
+ (unsigned long) p->data, value, regs);
+ set_eflags_OF(size, mmio_opp->instr, result,
+ (unsigned long) p->data, value, regs);
+ }
+ set_eflags_AF(size, result, value, (unsigned long) p->data, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
break;
case INSTR_TEST:
@@ -709,16 +765,16 @@ static void hvm_mmio_assist(struct cpu_u
index = operand_index(dst);
value = get_reg_value(size, index, 0, regs);
}
- diff = (unsigned long) p->data & value;
+ result = (unsigned long) p->data & value;
/*
* Sets the SF, ZF, and PF status flags. CF and OF are set to 0
*/
regs->eflags &= ~(X86_EFLAGS_CF|X86_EFLAGS_PF|
X86_EFLAGS_ZF|X86_EFLAGS_SF|X86_EFLAGS_OF);
- set_eflags_ZF(size, diff, regs);
- set_eflags_SF(size, diff, regs);
- set_eflags_PF(size, diff, regs);
+ set_eflags_ZF(size, result, regs);
+ set_eflags_SF(size, result, regs);
+ set_eflags_PF(size, result, regs);
break;
case INSTR_BT:
@@ -764,13 +820,14 @@ static void hvm_mmio_assist(struct cpu_u
}
}
-void hvm_io_assist(struct vcpu *v)
+void hvm_io_assist(void)
{
vcpu_iodata_t *vio;
ioreq_t *p;
struct cpu_user_regs *regs;
struct hvm_io_op *io_opp;
unsigned long gmfn;
+ struct vcpu *v = current;
struct domain *d = v->domain;
io_opp = &v->arch.hvm_vcpu.io_op;
@@ -788,10 +845,17 @@ void hvm_io_assist(struct vcpu *v)
p->state = STATE_IOREQ_NONE;
- if ( p->type == IOREQ_TYPE_PIO )
+ switch ( p->type )
+ {
+ case IOREQ_TYPE_INVALIDATE:
+ goto out;
+ case IOREQ_TYPE_PIO:
hvm_pio_assist(regs, p, io_opp);
- else
+ break;
+ default:
hvm_mmio_assist(regs, p, io_opp);
+ break;
+ }
/* Copy register changes back into current guest state. */
hvm_load_cpu_guest_regs(v, regs);
@@ -804,6 +868,7 @@ void hvm_io_assist(struct vcpu *v)
mark_dirty(d, gmfn);
}
+ out:
vcpu_end_shutdown_deferral(v);
}
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/irq.c
--- a/xen/arch/x86/hvm/irq.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/irq.c Fri Apr 06 10:08:30 2007 -0600
@@ -335,6 +335,21 @@ int is_isa_irq_masked(struct vcpu *v, in
return ((v->domain->arch.hvm_domain.vpic[isa_irq >> 3].imr &
(1 << (isa_irq & 7))) &&
domain_vioapic(v->domain)->redirtbl[gsi].fields.mask);
+}
+
+/*
+ * TODO: 1. Should not need special treatment of event-channel events.
+ * 2. Should take notice of interrupt shadows (or clear them).
+ */
+int hvm_local_events_need_delivery(struct vcpu *v)
+{
+ int pending;
+
+ pending = (vcpu_info(v, evtchn_upcall_pending) || cpu_has_pending_irq(v));
+ if ( unlikely(pending) )
+ pending = hvm_interrupts_enabled(v);
+
+ return pending;
}
#if 0 /* Keep for debugging */
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/platform.c
--- a/xen/arch/x86/hvm/platform.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/platform.c Fri Apr 06 10:08:30 2007 -0600
@@ -865,7 +865,7 @@ void send_pio_req(unsigned long port, un
if ( hvm_portio_intercept(p) )
{
p->state = STATE_IORESP_READY;
- hvm_io_assist(v);
+ hvm_io_assist();
return;
}
@@ -914,7 +914,7 @@ static void send_mmio_req(unsigned char
if ( hvm_mmio_intercept(p) || hvm_buffered_io_intercept(p) )
{
p->state = STATE_IORESP_READY;
- hvm_io_assist(v);
+ hvm_io_assist();
return;
}
@@ -939,6 +939,34 @@ void send_timeoffset_req(unsigned long t
if ( !hvm_buffered_io_send(p) )
printk("Unsuccessful timeoffset update\n");
+}
+
+/* Ask ioemu mapcache to invalidate mappings. */
+void send_invalidate_req(void)
+{
+ struct vcpu *v = current;
+ vcpu_iodata_t *vio;
+ ioreq_t *p;
+
+ vio = get_vio(v->domain, v->vcpu_id);
+ if ( vio == NULL )
+ {
+ printk("bad shared page: %lx\n", (unsigned long) vio);
+ domain_crash_synchronous();
+ }
+
+ p = &vio->vp_ioreq;
+ if ( p->state != STATE_IOREQ_NONE )
+ printk("WARNING: send invalidate req with something "
+ "already pending (%d)?\n", p->state);
+
+ p->type = IOREQ_TYPE_INVALIDATE;
+ p->size = 4;
+ p->dir = IOREQ_WRITE;
+ p->data = ~0UL; /* flush all */
+ p->io_count++;
+
+ hvm_send_assist_req(v);
}
static void mmio_operands(int type, unsigned long gpa,
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/svm/svm.c
--- a/xen/arch/x86/hvm/svm/svm.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/svm/svm.c Fri Apr 06 10:08:30 2007 -0600
@@ -131,66 +131,6 @@ static void svm_store_cpu_guest_regs(
}
}
-
-static inline int long_mode_do_msr_read(struct cpu_user_regs *regs)
-{
- u64 msr_content = 0;
- struct vcpu *v = current;
- struct vmcb_struct *vmcb = v->arch.hvm_svm.vmcb;
-
- switch ((u32)regs->ecx)
- {
- case MSR_EFER:
- msr_content = v->arch.hvm_svm.cpu_shadow_efer;
- break;
-
-#ifdef __x86_64__
- case MSR_FS_BASE:
- msr_content = vmcb->fs.base;
- goto check_long_mode;
-
- case MSR_GS_BASE:
- msr_content = vmcb->gs.base;
- goto check_long_mode;
-
- case MSR_SHADOW_GS_BASE:
- msr_content = vmcb->kerngsbase;
- check_long_mode:
- if ( !svm_long_mode_enabled(v) )
- {
- svm_inject_exception(v, TRAP_gp_fault, 1, 0);
- return 0;
- }
- break;
-#endif
-
- case MSR_STAR:
- msr_content = vmcb->star;
- break;
-
- case MSR_LSTAR:
- msr_content = vmcb->lstar;
- break;
-
- case MSR_CSTAR:
- msr_content = vmcb->cstar;
- break;
-
- case MSR_SYSCALL_MASK:
- msr_content = vmcb->sfmask;
- break;
- default:
- return 0;
- }
-
- HVM_DBG_LOG(DBG_LEVEL_2, "msr_content: %"PRIx64"\n",
- msr_content);
-
- regs->eax = (u32)(msr_content >> 0);
- regs->edx = (u32)(msr_content >> 32);
- return 1;
-}
-
static inline int long_mode_do_msr_write(struct cpu_user_regs *regs)
{
u64 msr_content = (u32)regs->eax | ((u64)regs->edx << 32);
@@ -242,52 +182,12 @@ static inline int long_mode_do_msr_write
break;
-#ifdef __x86_64__
- case MSR_FS_BASE:
- case MSR_GS_BASE:
- case MSR_SHADOW_GS_BASE:
- if ( !svm_long_mode_enabled(v) )
- goto gp_fault;
-
- if ( !is_canonical_address(msr_content) )
- goto uncanonical_address;
-
- if ( ecx == MSR_FS_BASE )
- vmcb->fs.base = msr_content;
- else if ( ecx == MSR_GS_BASE )
- vmcb->gs.base = msr_content;
- else
- vmcb->kerngsbase = msr_content;
- break;
-#endif
-
- case MSR_STAR:
- vmcb->star = msr_content;
- break;
-
- case MSR_LSTAR:
- case MSR_CSTAR:
- if ( !is_canonical_address(msr_content) )
- goto uncanonical_address;
-
- if ( ecx == MSR_LSTAR )
- vmcb->lstar = msr_content;
- else
- vmcb->cstar = msr_content;
- break;
-
- case MSR_SYSCALL_MASK:
- vmcb->sfmask = msr_content;
- break;
-
default:
return 0;
}
return 1;
- uncanonical_address:
- HVM_DBG_LOG(DBG_LEVEL_1, "Not cano address of msr write %x\n", ecx);
gp_fault:
svm_inject_exception(v, TRAP_gp_fault, 1, 0);
return 0;
@@ -596,6 +496,12 @@ static int svm_realmode(struct vcpu *v)
unsigned long eflags = v->arch.hvm_svm.vmcb->rflags;
return (eflags & X86_EFLAGS_VM) || !(cr0 & X86_CR0_PE);
+}
+
+static int svm_interrupts_enabled(struct vcpu *v)
+{
+ unsigned long eflags = v->arch.hvm_svm.vmcb->rflags;
+ return !irq_masked(eflags);
}
static int svm_guest_x86_mode(struct vcpu *v)
@@ -900,6 +806,7 @@ static struct hvm_function_table svm_fun
.paging_enabled = svm_paging_enabled,
.long_mode_enabled = svm_long_mode_enabled,
.pae_enabled = svm_pae_enabled,
+ .interrupts_enabled = svm_interrupts_enabled,
.guest_x86_mode = svm_guest_x86_mode,
.get_guest_ctrl_reg = svm_get_ctrl_reg,
.get_segment_base = svm_get_segment_base,
@@ -2013,22 +1920,14 @@ static inline void svm_do_msr_access(
case MSR_IA32_TIME_STAMP_COUNTER:
msr_content = hvm_get_guest_time(v);
break;
- case MSR_IA32_SYSENTER_CS:
- msr_content = vmcb->sysenter_cs;
- break;
- case MSR_IA32_SYSENTER_ESP:
- msr_content = vmcb->sysenter_esp;
- break;
- case MSR_IA32_SYSENTER_EIP:
- msr_content = vmcb->sysenter_eip;
- break;
case MSR_IA32_APICBASE:
msr_content = vcpu_vlapic(v)->hw.apic_base_msr;
break;
+ case MSR_EFER:
+ msr_content = v->arch.hvm_svm.cpu_shadow_efer;
+ break;
+
default:
- if (long_mode_do_msr_read(regs))
- goto done;
-
if ( rdmsr_hypervisor_regs(ecx, &eax, &edx) ||
rdmsr_safe(ecx, eax, edx) == 0 )
{
@@ -2060,15 +1959,6 @@ static inline void svm_do_msr_access(
case MSR_IA32_TIME_STAMP_COUNTER:
hvm_set_guest_time(v, msr_content);
pt_reset(v);
- break;
- case MSR_IA32_SYSENTER_CS:
- vmcb->sysenter_cs = msr_content;
- break;
- case MSR_IA32_SYSENTER_ESP:
- vmcb->sysenter_esp = msr_content;
- break;
- case MSR_IA32_SYSENTER_EIP:
- vmcb->sysenter_eip = msr_content;
break;
case MSR_IA32_APICBASE:
vlapic_msr_set(vcpu_vlapic(v), msr_content);
@@ -2276,7 +2166,7 @@ asmlinkage void svm_vmexit_handler(struc
unsigned long eip;
struct vcpu *v = current;
struct vmcb_struct *vmcb = v->arch.hvm_svm.vmcb;
- int inst_len;
+ int inst_len, rc;
exit_reason = vmcb->exitcode;
save_svm_cpu_user_regs(v, regs);
@@ -2385,8 +2275,13 @@ asmlinkage void svm_vmexit_handler(struc
inst_len = __get_instruction_length(v, INSTR_VMCALL, NULL);
ASSERT(inst_len > 0);
HVMTRACE_1D(VMMCALL, v, regs->eax);
- __update_guest_eip(vmcb, inst_len);
- hvm_do_hypercall(regs);
+ rc = hvm_do_hypercall(regs);
+ if ( rc != HVM_HCALL_preempted )
+ {
+ __update_guest_eip(vmcb, inst_len);
+ if ( rc == HVM_HCALL_invalidate )
+ send_invalidate_req();
+ }
break;
case VMEXIT_CR0_READ:
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/svm/vmcb.c
--- a/xen/arch/x86/hvm/svm/vmcb.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/svm/vmcb.c Fri Apr 06 10:08:30 2007 -0600
@@ -141,6 +141,14 @@ static int construct_vmcb(struct vcpu *v
disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_FS_BASE);
disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_GS_BASE);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_SHADOW_GS_BASE);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_CSTAR);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_LSTAR);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_STAR);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_SYSCALL_MASK);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_IA32_SYSENTER_CS);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_IA32_SYSENTER_ESP);
+ disable_intercept_for_msr((char *)arch_svm->msrpm, MSR_IA32_SYSENTER_EIP);
vmcb->msrpm_base_pa = (u64)virt_to_maddr(arch_svm->msrpm);
vmcb->iopm_base_pa = (u64)virt_to_maddr(hvm_io_bitmap);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/arch/x86/hvm/vmx/vmx.c
--- a/xen/arch/x86/hvm/vmx/vmx.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/arch/x86/hvm/vmx/vmx.c Fri Apr 06 10:08:30 2007 -0600
@@ -957,6 +957,13 @@ static int vmx_pae_enabled(struct vcpu *
return (vmx_paging_enabled(v) && (cr4 & X86_CR4_PAE));
}
+static int vmx_interrupts_enabled(struct vcpu *v)
+{
+ unsigned long eflags = __vmread(GUEST_RFLAGS);
+ return !irq_masked(eflags);
+}
+
+
static void vmx_update_host_cr3(struct vcpu *v)
{
ASSERT( (v == current) || !vcpu_runnable(v) );
@@ -1030,6 +1037,7 @@ static struct hvm_function_table vmx_fun
.paging_enabled = vmx_paging_enabled,
.long_mode_enabled = vmx_long_mode_enabled,
.pae_enabled = vmx_pae_enabled,
+ .interrupts_enabled = vmx_interrupts_enabled,
.guest_x86_mode = vmx_guest_x86_mode,
.get_guest_ctrl_reg = vmx_get_ctrl_reg,
.get_segment_base = vmx_get_segment_base,
@@ -2618,10 +2626,16 @@ asmlinkage void vmx_vmexit_handler(struc
}
case EXIT_REASON_VMCALL:
{
+ int rc;
HVMTRACE_1D(VMMCALL, v, regs->eax);
inst_len = __get_instruction_length(); /* Safe: VMCALL */
- __update_guest_eip(inst_len);
- hvm_do_hypercall(regs);
+ rc = hvm_do_hypercall(regs);
+ if ( rc != HVM_HCALL_preempted )
+ {
+ __update_guest_eip(inst_len);
+ if ( rc == HVM_HCALL_invalidate )
+ send_invalidate_req();
+ }
break;
}
case EXIT_REASON_CR_ACCESS:
diff -r 9ec7dadc98ba -r eb3e430242ac xen/common/domain.c
--- a/xen/common/domain.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/common/domain.c Fri Apr 06 10:08:30 2007 -0600
@@ -307,7 +307,7 @@ void domain_kill(struct domain *d)
domain_pause(d);
/* Already dying? Then bail. */
- if ( xchg(&d->is_dying, 1) )
+ if ( test_and_set_bool(d->is_dying) )
{
domain_unpause(d);
return;
@@ -453,7 +453,7 @@ void domain_pause_for_debugger(void)
struct vcpu *v;
atomic_inc(&d->pause_count);
- if ( xchg(&d->is_paused_by_controller, 1) )
+ if ( test_and_set_bool(d->is_paused_by_controller) )
domain_unpause(d); /* race-free atomic_dec(&d->pause_count) */
for_each_vcpu ( d, v )
@@ -553,13 +553,13 @@ void domain_pause_by_systemcontroller(st
void domain_pause_by_systemcontroller(struct domain *d)
{
domain_pause(d);
- if ( xchg(&d->is_paused_by_controller, 1) )
+ if ( test_and_set_bool(d->is_paused_by_controller) )
domain_unpause(d);
}
void domain_unpause_by_systemcontroller(struct domain *d)
{
- if ( xchg(&d->is_paused_by_controller, 0) )
+ if ( test_and_clear_bool(d->is_paused_by_controller) )
domain_unpause(d);
}
diff -r 9ec7dadc98ba -r eb3e430242ac xen/common/memory.c
--- a/xen/common/memory.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/common/memory.c Fri Apr 06 10:08:30 2007 -0600
@@ -173,17 +173,6 @@ int guest_remove_page(struct domain *d,
if ( test_and_clear_bit(_PGC_allocated, &page->count_info) )
put_page(page);
- if ( unlikely((page->count_info & PGC_count_mask) != 1) )
- {
- shadow_drop_references(d, page);
- /* We'll make this a guest-visible error in future, so take heed! */
- if ( (page->count_info & PGC_count_mask) != 1 )
- gdprintk(XENLOG_INFO, "Dom%d freeing in-use page %lx "
- "(pseudophys %lx): count=%lx type=%lx\n",
- d->domain_id, mfn, get_gpfn_from_mfn(mfn),
- (unsigned long)page->count_info, page->u.inuse.type_info);
- }
-
guest_physmap_remove_page(d, gmfn, mfn);
put_page(page);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/common/schedule.c
--- a/xen/common/schedule.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/common/schedule.c Fri Apr 06 10:08:30 2007 -0600
@@ -461,7 +461,11 @@ ret_t do_sched_op(int cmd, XEN_GUEST_HAN
if ( d == NULL )
break;
+ /* domain_pause() prevens any further execution in guest context. */
+ domain_pause(d);
domain_shutdown(d, (u8)sched_remote_shutdown.reason);
+ domain_unpause(d);
+
rcu_unlock_domain(d);
ret = 0;
diff -r 9ec7dadc98ba -r eb3e430242ac xen/drivers/char/ns16550.c
--- a/xen/drivers/char/ns16550.c Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/drivers/char/ns16550.c Fri Apr 06 10:08:30 2007 -0600
@@ -239,7 +239,7 @@ static void ns16550_init_postirq(struct
uart->irqaction.name = "ns16550";
uart->irqaction.dev_id = port;
if ( (rc = setup_irq(uart->irq, &uart->irqaction)) != 0 )
- printk("ERROR: Failed to allocate na16550 IRQ %d\n", uart->irq);
+ printk("ERROR: Failed to allocate ns16550 IRQ %d\n", uart->irq);
/* Master interrupt enable; also keep DTR/RTS asserted. */
ns_write_reg(uart, MCR, MCR_OUT2 | MCR_DTR | MCR_RTS);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-ia64/xentypes.h
--- a/xen/include/asm-ia64/xentypes.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-ia64/xentypes.h Fri Apr 06 10:08:30 2007 -0600
@@ -5,6 +5,11 @@ typedef unsigned long ssize_t;
typedef unsigned long ssize_t;
typedef unsigned long size_t;
typedef long long loff_t;
+
+typedef char bool_t;
+#define test_and_set_bool(b) xchg(&(b), 1)
+#define test_and_clear_bool(b) xchg(&(b), 0)
+
#endif /* !__ASSEMBLY__ */
#endif /* _ASM_IA64_XENTYPES_H */
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-powerpc/types.h
--- a/xen/include/asm-powerpc/types.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-powerpc/types.h Fri Apr 06 10:08:30 2007 -0600
@@ -70,5 +70,9 @@ typedef u64 dma64_addr_t;
typedef unsigned short xmem_bufctl_t;
+typedef int bool_t;
+#define test_and_set_bool(b) xchg(&(b), 1)
+#define test_and_clear_bool(b) xchg(&(b), 0)
+
#endif /* __ASSEMBLY__ */
#endif
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-x86/event.h
--- a/xen/include/asm-x86/event.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-x86/event.h Fri Apr 06 10:08:30 2007 -0600
@@ -35,12 +35,13 @@ static inline void vcpu_mark_events_pend
vcpu_kick(v);
}
+int hvm_local_events_need_delivery(struct vcpu *v);
static inline int local_events_need_delivery(void)
{
struct vcpu *v = current;
- return ((vcpu_info(v, evtchn_upcall_pending) &&
- !vcpu_info(v, evtchn_upcall_mask)) ||
- (is_hvm_vcpu(v) && cpu_has_pending_irq(v)));
+ return (is_hvm_vcpu(v) ? hvm_local_events_need_delivery(v) :
+ (vcpu_info(v, evtchn_upcall_pending) &&
+ !vcpu_info(v, evtchn_upcall_mask)));
}
static inline int local_event_delivery_is_enabled(void)
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-x86/hvm/hvm.h
--- a/xen/include/asm-x86/hvm/hvm.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-x86/hvm/hvm.h Fri Apr 06 10:08:30 2007 -0600
@@ -93,13 +93,15 @@ struct hvm_function_table {
* 1) determine whether paging is enabled,
* 2) determine whether long mode is enabled,
* 3) determine whether PAE paging is enabled,
- * 4) determine the mode the guest is running in,
- * 5) return the current guest control-register value
- * 6) return the current guest segment descriptor base
+ * 4) determine whether interrupts are enabled or not,
+ * 5) determine the mode the guest is running in,
+ * 6) return the current guest control-register value
+ * 7) return the current guest segment descriptor base
*/
int (*paging_enabled)(struct vcpu *v);
int (*long_mode_enabled)(struct vcpu *v);
int (*pae_enabled)(struct vcpu *v);
+ int (*interrupts_enabled)(struct vcpu *v);
int (*guest_x86_mode)(struct vcpu *v);
unsigned long (*get_guest_ctrl_reg)(struct vcpu *v, unsigned int num);
unsigned long (*get_segment_base)(struct vcpu *v, enum x86_segment seg);
@@ -190,6 +192,12 @@ hvm_pae_enabled(struct vcpu *v)
}
static inline int
+hvm_interrupts_enabled(struct vcpu *v)
+{
+ return hvm_funcs.interrupts_enabled(v);
+}
+
+static inline int
hvm_guest_x86_mode(struct vcpu *v)
{
return hvm_funcs.guest_x86_mode(v);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-x86/hvm/io.h
--- a/xen/include/asm-x86/hvm/io.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-x86/hvm/io.h Fri Apr 06 10:08:30 2007 -0600
@@ -147,9 +147,10 @@ extern void send_pio_req(unsigned long p
extern void send_pio_req(unsigned long port, unsigned long count, int size,
paddr_t value, int dir, int df, int value_is_ptr);
void send_timeoffset_req(unsigned long timeoff);
+void send_invalidate_req(void);
extern void handle_mmio(unsigned long gpa);
extern void hvm_interrupt_post(struct vcpu *v, int vector, int type);
-extern void hvm_io_assist(struct vcpu *v);
+extern void hvm_io_assist(void);
#endif /* __ASM_X86_HVM_IO_H__ */
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-x86/hvm/support.h
--- a/xen/include/asm-x86/hvm/support.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-x86/hvm/support.h Fri Apr 06 10:08:30 2007 -0600
@@ -228,7 +228,10 @@ void hvm_print_line(struct vcpu *v, cons
void hvm_print_line(struct vcpu *v, const char c);
void hlt_timer_fn(void *data);
-void hvm_do_hypercall(struct cpu_user_regs *pregs);
+#define HVM_HCALL_completed 0 /* hypercall completed - no further action */
+#define HVM_HCALL_preempted 1 /* hypercall preempted - re-execute VMCALL */
+#define HVM_HCALL_invalidate 2 /* invalidate ioemu-dm memory cache */
+int hvm_do_hypercall(struct cpu_user_regs *pregs);
void hvm_hlt(unsigned long rflags);
void hvm_triple_fault(void);
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/asm-x86/types.h
--- a/xen/include/asm-x86/types.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/asm-x86/types.h Fri Apr 06 10:08:30 2007 -0600
@@ -52,6 +52,10 @@ typedef unsigned long paddr_t;
typedef unsigned long size_t;
+typedef char bool_t;
+#define test_and_set_bool(b) xchg(&(b), 1)
+#define test_and_clear_bool(b) xchg(&(b), 0)
+
#endif /* __ASSEMBLY__ */
#if defined(__i386__)
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/public/acm.h
--- a/xen/include/public/acm.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/public/acm.h Fri Apr 06 10:08:30 2007 -0600
@@ -124,7 +124,7 @@ struct acm_policy_version
{
uint32_t major;
uint32_t minor;
-} __attribute__((packed));
+};
/* each buffer consists of all policy information for
@@ -145,12 +145,12 @@ struct acm_policy_buffer {
uint32_t secondary_policy_code;
uint32_t secondary_buffer_offset;
struct acm_policy_version xml_pol_version; /* add in V3 */
-} __attribute__((packed));
+};
struct acm_policy_reference_buffer {
uint32_t len;
-} __attribute__((packed));
+};
struct acm_chwall_policy_buffer {
uint32_t policy_version; /* ACM_CHWALL_VERSION */
@@ -162,7 +162,7 @@ struct acm_chwall_policy_buffer {
uint32_t chwall_conflict_sets_offset;
uint32_t chwall_running_types_offset;
uint32_t chwall_conflict_aggregate_offset;
-} __attribute__((packed));
+};
struct acm_ste_policy_buffer {
uint32_t policy_version; /* ACM_STE_VERSION */
@@ -170,7 +170,7 @@ struct acm_ste_policy_buffer {
uint32_t ste_max_types;
uint32_t ste_max_ssidrefs;
uint32_t ste_ssid_offset;
-} __attribute__((packed));
+};
struct acm_stats_buffer {
uint32_t magic;
@@ -179,7 +179,7 @@ struct acm_stats_buffer {
uint32_t primary_stats_offset;
uint32_t secondary_policy_code;
uint32_t secondary_stats_offset;
-} __attribute__((packed));
+};
struct acm_ste_stats_buffer {
uint32_t ec_eval_count;
@@ -188,7 +188,7 @@ struct acm_ste_stats_buffer {
uint32_t gt_denied_count;
uint32_t ec_cachehit_count;
uint32_t gt_cachehit_count;
-} __attribute__((packed));
+};
struct acm_ssid_buffer {
uint32_t len;
@@ -200,7 +200,7 @@ struct acm_ssid_buffer {
uint32_t secondary_policy_code;
uint32_t secondary_max_types;
uint32_t secondary_types_offset;
-} __attribute__((packed));
+};
#endif
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/public/foreign/Makefile
--- a/xen/include/public/foreign/Makefile Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/public/foreign/Makefile Fri Apr 06 10:08:30 2007 -0600
@@ -13,7 +13,7 @@ clean:
rm -f checker checker.c $(XEN_TARGET_ARCH).size
rm -f *.pyc *.o *~
-ifeq ($(CROSS_COMPILE),)
+ifeq ($(CROSS_COMPILE)$(XEN_TARGET_ARCH),$(XEN_COMPILE_ARCH))
check-headers: checker
./checker > $(XEN_TARGET_ARCH).size
diff -u reference.size $(XEN_TARGET_ARCH).size
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/public/hvm/hvm_op.h
--- a/xen/include/public/hvm/hvm_op.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/public/hvm/hvm_op.h Fri Apr 06 10:08:30 2007 -0600
@@ -70,12 +70,4 @@ typedef struct xen_hvm_set_pci_link_rout
typedef struct xen_hvm_set_pci_link_route xen_hvm_set_pci_link_route_t;
DEFINE_XEN_GUEST_HANDLE(xen_hvm_set_pci_link_route_t);
-/* Drain all outstanding qemu-dm IO responses from a domain's ioreq ring. */
-#define HVMOP_drain_io 5
-struct xen_hvm_drain_io {
- domid_t domid;
-};
-typedef struct xen_hvm_drain_io xen_hvm_drain_io_t;
-DEFINE_XEN_GUEST_HANDLE(xen_hvm_drain_io_t);
-
#endif /* __XEN_PUBLIC_HVM_HVM_OP_H__ */
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/public/hvm/ioreq.h
--- a/xen/include/public/hvm/ioreq.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/public/hvm/ioreq.h Fri Apr 06 10:08:30 2007 -0600
@@ -40,6 +40,7 @@
#define IOREQ_TYPE_XCHG 5
#define IOREQ_TYPE_ADD 6
#define IOREQ_TYPE_TIMEOFFSET 7
+#define IOREQ_TYPE_INVALIDATE 8 /* mapcache */
/*
* VMExit dispatcher should cooperate with instruction decoder to
diff -r 9ec7dadc98ba -r eb3e430242ac xen/include/xen/types.h
--- a/xen/include/xen/types.h Fri Apr 06 10:06:30 2007 -0600
+++ b/xen/include/xen/types.h Fri Apr 06 10:08:30 2007 -0600
@@ -19,8 +19,6 @@
#define LONG_MAX ((long)(~0UL>>1))
#define LONG_MIN (-LONG_MAX - 1)
#define ULONG_MAX (~0UL)
-
-typedef char bool_t;
/* bsd */
typedef unsigned char u_char;
_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog
|
Home